EXCHANGE RATE REGIMES AND ECONOMIC
PERFORMANCE: A COMPARATIVE ANALYSIS
BY
ESTHER OLUFUNMILA YO (AKIN LADE) ADEGBITE
B.Sc (Economics); Ibadan
M.Sc (Economics) Ibadan
A Thesis in the Department Of Economics
, Submitted to the Faculty of the Social Sciences In Partial
Fulfilment of the requirements for the degree of.
DOCTOR OF PHILOSOPHY
of the
UNIVERSITY OF IBADAN
1992/93 SESSION
•
UNIVERSITY OF IBADAN LIBRARY
IBAOAN UNIVERSITY LIBRARY
... 00_-,--'-1
I,C':".
NC. 6 ,0168
LC' . lH ·--l
~1Aj.; • H&
3851
·A3x
•
UNIVERSITY OF IBADAN LIBRARY
CERTIFICATION
We certifythat thisworkwas carriedout by Esther OIUfunmilayC/e) Adegbite in the
Department of Economics, University of Iba , lbadan.
T. Ademo 0 IDr
(Supervisor, and Chairman, Thesis Committe)
B.Sc (Economics) Jbadan, M.Sc (Economics), London,
MA; Ph. D. (Economics) Princeton:
Professor of Economics
Department of Economics
University of Ibadan, Ibadan.
..
Samue o. OLOFIN
(Member, Thesis Committee)
B.Sc. (Economics) Ibadan, M.Sc (Economics)
Ibadan, M.A.,Ph.D. (Economics) Princeton.
Member, Thesis Committee
Professor of Economics
Department of Economics
University of Ibadan, Ibadan.
(
Adedoyin SOYIBO
(Member, Thesis Committee)
B.Sc. (Mathematics), lbadan,
S.M. (Management Science/Operation Research) M.I. T.,
Ph.D. (Economics) Ibadan.
Reader in the Department of Economics
University of Ibadan.
UNIVERSITY OF IBADAN LIBRARY
DEDICATION
This Thesis is Dedicated to
My Parents
ELIJAH KOLADE AND ELIZABETH AGBEKE AKINLADE
For Inspiration and Support
My Husband
MATHEW OLUFEl\lI ADEGBITE
For Understanding and Endurance
,
My Children
ADEFUNMINIYI, ADEDEJI AND ADETOLA
For their forbearance .
•
UNIVERSITY OF IB
DAN LIBRARY
iii
ACKNOWLEDGEMENT
First I wish to thank all members of my thesis committee, Professor T. Ademola
Oyejide, Professor Samuel o. Olofin, and Dr. Adedoyin Soyibo. In particular I have to
thank Professor Oyejide my supervisor and Chairman of my thesis committee everything,
right from the choice of an appropriate title for my thesis through to getting financial support
to enable me carry out the study, Professor Oyejide was with me. What is more - he gave
me complete access to himself and to his books; for me, he removed all the "nightmares"
that my colleagues have as a result of supervisors that try to play "God". I wish someday
I would be in a position to reciprocate this kind fatherly support; if such a day never comes
I pray that the Almighty may never desert him, but may continue to shower him and all that
is his with abundant blessings.
,
I also wish to thank Professor Olofin profusely for his encouragement. Each time I
met with Professor Olofin he ensured that he boost my confidence, and I came out feeling
more confident, feeling more able to cope with the challenges ahead. I hope he will continue
to sharpen and utilize this psychological tool, which is of tremendrous value in building the
necessary crop of Nigerian academicians. May the good Lord continue to be with him and
bless him.
I also like to put on record Dr Soyibo's efforts to see that I complete this study
successfully. Dr Soyjbo never gave me a breathing space, whenever he saw I was relaxing,
UNIVERSITY OF IBADAN LIBRARY
iv
he kept prodding me, urging me on, encouraging me, advising me, and eventually when the
first draft of the thesis was ready he had to carry it with him outside the country while going
for a conference, just to ensure that he read it. I appreciate tremendously this effort to assist
a "student", and a "friend". May the good Lord continue to bless and guide him.
I must not forget to thank Dr. R.A. Olopoenia, my very good teacher both at the
undergraduate and graduate levels. It was Dr. Olopoenia who on sabbatical leave at
Makarere University, Kampala, Uganda, made my data-collection trip to Uganda a painless,
worthwhile, and fruitful one. I pray that he too will continue to meet kind generous and
considerate people like himself. May the Lord bless him real good.
I will also like to thank my husband, Olufemi Isola, "Omo ajilala eso", for his quiet
and dignified support. Many people expressed the fear that leaving my husband behind in
a city like Lagos for at least three full years might be a way of getting myself supplanted in
my matrimonial home. I thank God those fears never materialized. Furthermore I wish to
seize this opportunity to let him know that I have tremendous respect for his sense of
judgement and his industriousness.
And how do I thank my parents "enough" for everything, I'm sure I cannot thank
them "enough". They gave me all that was humanly possible. First, they kindly agreed to
take my children into their home for the three years the Ph.D program was supposed to have
lasted. Secondly th y never asked for anything for the upkeep of the children tight as things
were for them, many times they had to cope with the youthful excesses of my children, who
UNIVERSITY OF IBADAN LIBRARY
v
did not know that "grandpa" and "grandma" are no more as "strong" and energetic as daddy
and mummy. How many nights my parents couldn't sleep because one or the other of my
children was sick. May Jehovah bless them both and let them live long enough to "eat" the
fruits of their labour.
As is usual in works of this kind so many people contributed to make it a success,
so I will like to briefly mention a few people. In the department, Drs. Poloamina and
Ogwumike were particularly helpful in seeing me through the course-work. Dr. Ajakaiye
was invaluable in getting me access to NISER library. Dr Egwakhide was very helpful in
seeing me successfully complete the computer analysis. I wish to thank them all, I pray they
have more successes in all of their life' endeavours.
I must also thank.Dr Nwampah of the Bank of Ghana, Mrs Mbire of Department of
Economics Makarere University, Mr Stephen Ameyaw of the Bank of Ghana, Dr Kiggundu
formerly governor of the Bank of Uganda but now of the Department of Economics,
Makarere University, Dr Louis Kasekende of the Bank of Uganda - all of them assisted me
tremendously in getting data for the study.
I must never forget to thank Professor Olatunde Olatunji of the Linguistics
department University of Ibadan, who helped to ensure that I got accommodation at the
Postgraduate Hall of r~dence in the first two years of my program. I am indeed very
grateful. I wish also to express my sincere gratitude to Mrs. Eleyinmi of the Central Bank
of Nigeria who was of tremendous assistance in getting me literature from the Central Bank
•
UNIVERSITY OF IBADAN LIBRARY
vi
Library.
Mr. Tega Esabunor my mathematics "teacher" and my "brother" is one person I must
not forget to thank. May God continue to guard and guide him. I have to thank my friends
also for everything-Deola Adenikinju (Prince), Abdul Ganiyu Garba (professor), Jerome
(La Presido), John Adeyemo, Mr. Goke Omole, Mr. Falokun and Chidozie (professor).
Others that were of tremendous assistance include Dr. Raheem, Dr. Ogun and Dr.
Ogunkola.
Two other people I must mention are my uncle-in-law Mr. M. A. Adegbite who has
proved to be a father lito me and a great pillar of strength". I know that without his fatherly
wisdom I probably may not have had the privilege of coming for a Ph.d program in the first
instance. May the almighty Jehova continue to bless him and grant him long life.
My profound gratitude also goes to Professor Eno lnanga of the Department of
,
Economics, University of Ibadan for his constant encouragement and fatherly advice. May
God be with him.
My gratitude also goes to Mr. Balogun, Mr. Sam Adegoju, Mr. Kunle Ojo, Mr. Tony
Akinfe, Mr. John Alozie all of whom assisted with the different aspects of the computer
work and the word processing aspect of the thesis, andMr. Ezekiel Adegoke who has been
a very dependable younger brother. May God bless you all.
However as is usual in works of this kind I wish to say that I am responsible for all
errors and shortcomings found in this study.
UNIVERSITY OF IBADAN LIBRARY
vii
Finally I thank the African Economic Research Consortium who provided me with
the Research Grant for carrying out the work. I am profoundly grateful.
For everything - to JEHOVAH be the GLORY •
•
UNIVERSITY OF IBADAN LIBRARY
viii
ABSTRACT
In 1973 the industrial countries of the world abandoned the Bretton - Wood
adjustable - peg exchange rate system as a means of international payments, and embraced
a floating exchange rate system. By the beginning of the 1980's some developing countries
of the world joined the league of exchange rate-floaters. It was thought that a floating
exchange rate system is intrinsically superior to a fixed one because it not only insulates an
economy from the events in other economies but also provides automatic adjustment of the
trade balance and the balance of payments.
From the mid 1980's however there have been calls in the industrial countries for yet
a change in the international payments system from a floating one back to the Bretton-
Woods fixed system (Marris, 1984; Dunn, 1983) or to some other variant of a fixed system.
The questions then are - is there an ideal exchange rate regime? - is there reason to believe
that a given exchange rate regime enhances the performance of an economy better than
another? These questions form the focus of this study.
There have been several positions in the literature. While Mundell-Fleming (1960,
1962) maintain that a floating exchange system is better than a fixed one if a country tends
to depend more on monetary policy, but that a fixed exchange rate regime is ideal when
fiscal policy is the major instrument employed in an economy, Sohmen (1965) maintained
that a floating regime is superior whatever the more dominant economic policy (fiscal or
UNIVERSITY OF IBADAN LIBRARY
ix
monetary). Demberg (1970) maintained that the performance of an economy does not
depend on the exchange rate regime per-se but rather on the optimal mix of fiscal and
monetary policy. In the developing world there is fear that a floating exchange regime
would aggravate rather than reduce the problems of inflation. debt-service burden and
balance of payments disequilibrium (Olofin, Akinkugbe, Ajayi 1986). This study therefore
attempted to find out which of the positions in the literature really holds in the case of
developing African economies.
To find answers to the issues raised we chose three african economies who had
experienced both fixed and floating exchange rate systems. namely. Ghana. Nigeria and
Uganda. We built a model of each economy in the manner of Rhomberg (1964) and Tullio
1981. Each model has two versions. The shorter version has seven stochastic equa tions and
tries to capture the economy under a fixed system. while the longer version added two
additional stochastic equations to the first set and endogenizes exchange rates and interest
rates as obtains under a floating exchange system. Utilizing quarterly data for 1977 to 1990
for Nigeria and Ghana. and for 1981 to 1990 for Uganda and employing the Ordinary Least
Squares technique we estimated the shorter version of the model for the period 1977:1-
1990:4 and the longer version for the period 1986:4-1990:4 for Ghana and Nigeria. In the
case of Uganda we estimated the longer version for the period 1981:1 to 1990:4 and the
shorter version for 1987:2 to 1990:4.
Beyond the statistical tests of the individual equations and parameters. we attempted
UNIVERSITY OF IBADAN LIBRARY
x
to carry out rigorous tests of the validity of our model(s) through dynamic simulation. Thus
we solved our model(s) using the Time Series Processor (TSP) econometric Software
(Version 4.0) developed by Hall in 1983. When we solved each model using the Gauss -
Seidel iterative technique, each converged for each endogenous variable and for each year
demonstrating that each model is internally consistent. Utilizing different policy scenarios
we tried to find out the effects of monetary, fiscal and exchange rate policy changes on
internal sectors' macroaggregates of prices, real demand for money and money supply, as
well as on external sector's macroaggregates of exports, imports and the trade balance.
The results of our estimation exercises reveal that in Ghana a floating exchange rate
system does not fuel inflation as is suggested by casual empiricism; rather it is the money
supply that is the major propeller of domestic prices, given an exchange rate elasticity of
domestic prices of 2% which is statistically insignificant at the 5% level and a money supply
elasticity of domestic prices of 19% that is statistically significant at the 5% level. In Uganda
there is a remarkable pass through from nominal exchange rates onto prices which
contradicts Elbadawi's (1990) position, that it is not nominal exchange rates that fuel
inflation in Uganda but fiscal deficits. The exchange rate elasticity of domestic prices in
Uganda is 11% and this is statistically significant at the 5% level. However even in Uganda,
nominal money supply and nominal rates of interest proved to be greater propellers of prices
hence they have more dominant impact on inflation than the nominal exchange rate. In
Nigeria there is some degree of pass through from nominal exchange rates onto prices given
UNIVERSITY OF IBADAN LIBRARY
xi
an exchange rate elasticity of domestic prices of 5%, which is statistically significant at the
5% level. However as in Ghana and Uganda money supply was the greater propeller of
prices in Nigeria. What is more- the estimation results also showed that nominal exchange
rates in the three countries follow the money supply. This goes to show that the behaviour
of the money supply and hence monetary policy influences the direction and degree of
variability in nominal exchange rates under a floating system. Hence it shows that monetary
policy is crucial to the success of the floating exchange rate system. Further the money
supply was shown to vary in response to government fiscal deficits which makes fiscal
prudence or otherwise the major determinant of exchange rate movements.
For the simulation experiments we tried to fmd in what ways our endogenous
variables change if a given macroeconomic policy varies while the others are kept constant.
Thus we increased the rate of growth of government expendi ure while keeping monetary
policy and exchange rate policy constant. Similarly when we increased the rate of growth
of the money supply we assumed fiscal and exchange rate policies to be constant. Our results
show that in the long-run (over a period of at least ten years) a floating exchange rate
performs better than a fixed one in terms of ensuring expanded output which ensures
declining prices which in turn results in rising real demand for money and hence in rising
rates of interests. A floating exchange rate regime also expanded exports and higher positive
trade balance. Overall however the success of the floating system depends on coordinated
and prudent macroeconomic policies; in the words of Goldstein (1984) "the capacity of the
UNIVERSITY OF IBADAN LIBRARY
xii
exchange rate system per-se to do good or harm should not be overestimated... the
importance of discipline and coordinated macroeconomic policies for the successful
operation of floating exchange rate regime should not be underestimated".
,
UNIVERSITY OF IBADAN LIBRARY
xiii
TABLE OF CONTENTS
DEDICATION i
CERTIFICATION ii
ACKNOWLEDGEMENT iii
ABSTRACT vii
TABLE OF CONTENTS xiii
LIST OF TABLES xv
LIST OF CHARTS xviii
CHAPTER ONE: INTRODUCTION
1.1 The Problem 1
1.2 Need for the Study 14
1.3 Objectives of the Study 24
1.4 Scope and Organisation of the Study 27
CHAPTER TWO: REVIEW OF CASE ECONOMIES
2.1 Exchange Rates and Exchange Rate Policy
in Ghana: 1971-1990 33
2.2 Exchange Rates and Exchange Management in
Nigeria Before and After SAP 41
2.3 Exchange Rates and Exchange Rate Management
in Uganda 50
,
CHAPTER THREE: LITERATURE REVIEW
3.1 Theories of Exchange Rate Determination 59
3.1.1. Purchasing Power Parity Theory (PPP) 59
3.l.H. Interest Rate Parity Theory (IRP) 61
3.l.H!. Balance of Payments Approach (BOP) 62
3.1.IV. The Monetary Approach MP 64
3.l.V. Portfolio Balance Theory of Exchange Rate
Determina tion 67
3.2. Review of Models of the Workings of the Macroeconomy
under Different Exchange Rate Regimes 68
3.2.1. The Insular-Economy Model 71
3.2.11. The Mundell-Fleming Model (MFM) 73
3.2.H1. The Modified Mundell Fleming Model (MMFM) 80
3.3. Review of Empirical Literature 84
•
UNIVERSITY OF IBADAN LIBRARY
xiv
3.4. Summary 100
CHAPTER FOUR: METHODOLOGY, MODEL SPECIFICATION,
THEORITICAL MODEL VALIDA TION TECHNIQUES,
DATA AND DATA SOURCES.
4.1. Summary of Issues. 102
4.2. Analytical framework 103
4.3. Methodology 110
4.4. Model Specification 120
4.5. Theoretical Model Validation Techniques 128
4.6. Data and Data Sources. 130
CHAPTER FIVE: DESCRIPTIVE ANALYSIS OF DATA
5.1. The Trade, Current and Capital Account Balances 133
5.2. International Reserves and the Balance of Payments 142
5.3. Domestic Prices and Real Income 163
CHAPTER SIX: MODEL ESTIMATION AND EMPIRICAL VALIDA TION.
6.1. Model Estimation 173
6.2. Empirical Validation. 188
CHAPTER SEVEN: SIMULATION RESULTS
7.1. Simulation Results for NIgeria 214
'\ 7.2. Simulation Results for Ghana 220
7.3. Simulation Results for Uganda 226
CHAPTER EIGHT: SUMMARY, CONCLUSIONS AND RECOMMENDATIONS.
8.1. Summary of Model Estimation Results 236
8.2. Summary of Model Simulation Results 239
8.3. Limitations of the Study 245
8.4. Conclusions 248
8.5. Recommendations 249
8.6. Areas of further Research 253
REFERENCES
ACRONYM 255
Appendix 1: Definition of Terms 256
Appendix 2: Linear Estimation Results and Error Analysis 259
BIBLOGRAPHY • 231
UNIVERSITY OF IBADAN LIBRARY
xv
LIST OF TABLES
1.1. Per Capita Income of Selected African Countries
(in US Dollars) 17
1.2. Sectoral value Added for Selected African Countries
(in millions of Constant 1980, US Dollars, At
Producer Price) 18
1.3. Gross Domestic Product for Selected African
Countries (in US Dollars at Purchaser Value:
Current Price Series) 19
1.4. Resource Balance As Percentage of GDP For Selected
African Countries 21
1.5. Total External Debt for Selected African Countries
(in millions of Current US Dollars) 21
1.6. Per Capita GNP and Per Capita Debt in 1989, For
Selected African Countries (in US Dollars) 22
1.7. Average Annual Growth Rates of Exports, Imports
, And Consumer Price Index For Selected African
Countries (in Percentage) 22
1.8. Terms of Trade Index for Selected African Countries
(1980=100) 23
1.9. Rate Flexibility Under Six Alternative Exchange Rate
Arrangements 27
1.10. Features of Focus Countries for 1985. 28
4.1. Trade Shares of Ghana And Nigeria's Ten Major Trading
Partners 114
4.2. Trade Shares of Uganda's Eight Major
Trading Partners 115
UNIVERSITY OF IBADAN LIBRARY
xvi
5.1. Nigeria's External Balances 1981-1990
(In millions of Naira) 136
5.2. Ghana's External Balances 1981-1990
(In millions of Cedis) 139
5.3. Uganda's External Balances 1981-1990
(In millions of Ugandan Shillings) 140
5.4. International Reserves and Balance of Payments for
Focus Countries (1981: 1 - 1990:4) 143
6.1. Estimation Results for Dependent Variable -
Domestic Price (PD) 174
6.2. Estiomation Results for Dependent Variable -
Nominal Exchange Rate (EN) 177
6.3. Estimation Results for Dependent Variable -
Export Value (X) 179
6.4. Estimation Results for Dependent Variable-
Import Value (1M) 181
..
6.5. Estiomation Results for Dependent Variable -
International Reserves (RE) 183
6.6. Estimation Results for Dependent Variable -.
Real Demand for Money (RDM) 188
6.7. Estimation Results for Dependent Variable-
Domestic Rate of Interest (RD) 184
6.8. Estiomation Results for Dependent Variable-
Real Income (RI) 187
6.9. Estimation Results for Dependent Variable -
Nominal Money supply (MSS) 187
UNIVERSITY OF IBADAN LIBRARY
xvii
6.10. Nigeria: Root Mean Square
Simulation Errors 189
6.11. Nigeria: Theil's Inequality Coefficient And Its
Decomposition 189
6.12. Ghana : Root Mean Square
Simulation Errors 191
6.12. Ghana : Theil's Inequality Coefficient And Its
Decomposition 192
6.13. Uganda: Root Mean Square
Simulation Errors 193
6.11. Uganda: Theil's Inequality Coefficient And Its
Coefficient And its Decomposition 194
7.1. Nigeria: Change in Fiscal Policy 215
7.2. Nigeria: Change in Monetary Policy 216
7.3. Nigeria: Change In Exchange Rate Policy 217
, 7.4. Ghana: Change In Fiscal Policy 222
7.5. Ghana: Change In Monetary Policy 223
7.6. Ghana: Change In Exchange Rate Policy 224
7.7. Uganda: Change In Fiscal Policy 228
7.8. Uganda: Change In Monetary Policy 229
7.9. Uganda: Change In Exchange Rate Policy 230
UNIVERSITY OF IBADAN LIBRARY
xviii
LIST OF CHARTS
4.l. Internal:and'External Sector's Disequilibrium 109
5.1 Nigeria: Import Growth 149
5.2 Nigeria: Export Growth 150
5.3 Nigeria: Export and Import
Share in GDP 151
5.4 Nigeria: Total Foreign Sector
Share in GDP 152
5.5 Ghana: Export Growth 155
5.6 Ghana: Import Growth 156
5.7 Ghana: Export and Import
Share in GDP 157
5.8 Ghana: Total Foreign Sector
Share in GDP 158
'5.9 Uganda: Export Growth 159
5.10 Uganda: Import Growth 160
5.1l. Ugandq: Export and Import
Share in GDP 161
5.12 Uganda: Total Foreign Sector
Share in GDP 162
5.13 Nigeria: Growth Rate of
Domestic Prices 165
5.14 Nigeria: Real Income Growth 166
5.15 Ghana: Growth Rate in Domestic Prices 168
5.16 Ghana: Real Income Growth 169
•
5.17 Uganda: Growth Rate of Domestic Prices 171
UNIVERSITY OF IBADAN LIBRARY
xix
5.18. Uganda: Real Income Growth 172
o
6.1. Nigeria: Actual/Simulated chart of Domestic prices 195
6.2. Nigeria: Real Demand for Money (RDM)
Actuall/Sirnulated 196
6.3. Nigeria: Money Supply (MSS) Actual/Simulated 197
6.4. Nigeria: Imports (1M) Actual/Simulated 198
6.5. Nigeria: Exports (X) Actual/Simulated 199
6.6. Nigeria: Real Income Actual/Simulated 200
6.7. Ghana: Actual/Simulated Chart for
Domestic Price (Pd) 201
6.8. Ghana: Real Demand for Money (RDM) Actual/Simulated 202
6.9. Ghana: Money Supply (MSS) Actual/Simulated 203
,
6.10. Ghana: Imports (IM) Actual/Simulated 204
6.11. Ghana: Exports (X) Actual/Simulated 205
6.12. Ghana: International Reserves Actual/Simulated 206
6.13. Uganda: Actual/Simulated Chart of Domestic Prices (PD) 207
6.14. Uganda: Real Demand for Money. Actual/Simulated 208
6.15. Uganda: Imports (IM). Actual/Simulated 209
6.16. Uganda: Exports (X). Actual/Simulated 210
6.17. Uganda: Real Income. Actual/Simulated. 211
6.18. Uganda: Nominal Exchange Rates Actual/Simulated 212
NIVERSITY OF IB
DAN LIBRARY
CHAPTER ONE
INTRODUCTION
"... In the quarter Century Since World war II The Growth in Prosperity-or
In some Instance - The Aspiration for Prosperity - of Peoples Around The
World Has Become Increasingly Dependent Upon The International Flows
of Commerce "
Kvasnicka (1969)
"The Specification of Equilibrium in The World Economy Depends Oil The
Exchange Rate Regime ".
Frenkel and Razin (1987)
1.1 The Problem
Trade and investment flows at the international level have become virtually indispensable
to the economic growth and prosperity of nations today. This is more so in the developing
countries of the world, where revenues from import tariffs and export duties "Constitute a
, major source of government finance and where imports of intermediate goods and machinery
oil the industrialisation wheel of these economies. That exchange rates - the means of
effecting these international flows should be of tremendous importance to the world
economy is underscored by Frenkel and Razins's assertion quoted above.
Exchange rates have been variously defined in the literature. Exchange rate is said to be
the relative price of two national monies' or, the relative price of two national outputs' or the
relative price of tradeables to non-tradeables that results in simultaneous equilibrium in the
internal and external sectors of an economy.' Whatever definition one adopts the important
UNIVERSITY OF IBADAN LIBRARY
2
issue is that exchange rates influence directly trade flows and indirectly capital flows
between nations. Given the above therefore the factors determining the level of exchange
rate, its degree of variability, its medium and long-run behaviour etc become of crucial
importance, to economists, policyrnakers, and investors alike.
The world has known several exchange rate systems beginning with the pre-World War-I
fixed gold standard. During the first world war the international payments system in form
of the fixed gold standard collapsed. The fixed gold standard exchange rate regime was one
in which national currencies were tied to specific weights of gold, and were convertible into
gold, and hence exchanged with each other depending on the gold content of each currency.
Though exchange rates were fixed under the gold standard, there were however, margins of
variation known as "gold-points." When a given country's exchange rate depreciated to such
, a level that it paid arbitrageurs to buy gold in domestic money and export it to foreign
country to earn foreign exchange, the exchange rate was said to have reached the "gold-
export" point. When on the other hand the exchange rate had appreciated so much that it
paid to convert domestic currency into foreign exchange and then use the proceed to
purchase gold from foreign country for the purpose of importing into the domestic economy.
then the exchange rate was said to have reached the "gold-import" point. At the gold-export
point attempt is made by arbitrageurs and/or the monetary authorities to protect the value of
domestic currency and prevent its depreciating unduly. At the gold-import point attempt is
made to prevent domestic currency from appreciating unduly and thereby losing its
UNIVERSITY OF IBADAN LIBRARY
3
competitiveness. The gold import and export points were fun tions of the available supply
of gold (as each currency had full gold convertibility) and the cost of transporting gold",
Mining gold and transporting it were very cumbersome, As mentioned earlier this gold
\ standard broke down during the first World War.
After the First World War attempt to put in place an international payments system was
left in the hands of individual governments. Most countries, then chose exchange rates in
such a way that they did not acknowledge the fact that an exchange rate involves more than
one currency. Exchange stabilization was carried out as an act of national sovereignty. There
was no regard for the resulting interrelationship of currency values in comparison with the
cost and price levels. Stabilization was perceived in terms of gold only and not in terms of
other currencies. Britain in 1925 and France in 1926 were the major culprits whose exchange
stabilization attempts led to considerable disequilibrium in the exchange rate system. S
This "national-sovereignty" in exchange rate policy led to exchange rates being
manipulated for beggar-my-neighbour purposes as evidenced by the massive devaluation of
the 1930s. Though any given devaluation improved the balance of trade of the country
concerned and increased its volume of domestic employment-these gains were necessarily
at the expense of other countries and thus invited retaliation. Thus by the onset of the Second
World War it was abundantly clear that a "new" and more-closely-regulated exchange
arrangement had to be put in place. This desire could not be satisfied during the war and so
•
nations had to wait until after the Second World War to put a new international payments
UNIVERSITY OF IBADAN LIBRARY
4
system in place.
In 1944 at Bretton Woods nations of the world met to design a new exchange rate system.
First they formed an international clearing house known as the International Monetary Fund
(IMP). In the light of the experience of the 1920s and 1930s the architects of the Bretton
Woods system could not place much confidence in national governments to unilaterally
determine their own exchange rates and decide "if" and "when" change was needed and in
"what direction." What is more-the architects of Bretton Woods did not have much faith in
the market either, because they thought reliance on the market would generate uncertamties
with respect to the exchange rate level and variability which would necessarily imply
increased risk in international exchange, and thus reduction in the volume of international
trade. The system put in place at Bretton Woods variously known as the "adjustable-peg",
the "adjustable-parity" or "par-value system" was a fixed exchange rate system determined
by national governments subject to the approval of the IMF. The system allowed a one
percent margin of variation for any exchange rate (half a percentage point either side of the
parity).
Under the Bretton Woods system nations fixed their exchange rates in terms of gold, and
in some instances in terms of the United State's dollar. When a new monetary unit known
as the Standard Drawing Right (SDR) was evolved in the late 1960s nations also had to
redefine the exchange parity of their currencies both in terms of gold, the SDR, and for some
countries in terms of the dollar as well. The United State of America, had as of 1945 agreed
UNIVERSITY OF IBADAN LIBRARY
5
to convert holdings of dollars by foreign countries into gold. Governments were required to
notify the IMF before any change could be effected in their exchange rates. Also nations
were prohibited from making large exchange rate changes without the consent of the IMF.
These requirements reflect the fact that the architects of Bretton Woods gave a vote of no-
confidence in national governments when it came to managing their exchange system for the
good of the international community. What is more, the IMF also insisted that the only
justification for an exchange rate change was a "fundamental disequilibrium'" in a country's
balance of payments. What constituted a fundamental disequilibrium was not speIt out, it
was left to the discretion of the Ir..'1F.Temporary deficits were to be financed by drawing
down reserves with the IMP.
Three main problems plagued this adjustable peg system and finally led to its demise. In
the first instance, U.S.A. 's commitment to convert dollar holdings of other countries into
gold, could not be indefinitely sustained. In a way this commitment led to a veiled
perception of the degree of available international liquidity. Thus the commitment left
unresolved the problem of providing in some reasonably rational way for an adequate
secular growth in intemationalliquidity.
Secondly the role of exchange rate in balance of payments adjustments created a lot of
troubles. Because of its political implications, most nations running balance of payments
deficits refused to devalue their currencies, because devaluation would often mean, cut in
•
real wages, shift in production from domestic consumption market to export markets,
UNIVERSITY OF IBADAN LIBRARY
6
restraint on monetary and fiscal expansion etc. Thus nations usually postponed. devaluation
of their exchange rates until it was too late. The result then became excessi ely biting. Thus
the issue of exchange rate in the adjustment process became in the words of Coombs "a
political problem of appalling complexity ."7
Even for surplus countries needing to make adjustments the fear of harming their export-
industries made them shy away from revaluation. What is more - the IMF agreement itself
failed to specify on whom the onus for taking corrective action rests - whether the surplus
or the deficit country. Thus the Bretton Woods system led to appreciable surpluses in the
balance of payments of some countries and considerable deficits in those of others. As
countries felt the need for some exchange rate adjustment without bearing the full brunt of
its results, the search for alternative regimes began.
Some economists advocated what is known as "Crawling - Peg system." A crawling
, peg system is essentially an adjustable peg system, the difference being that instead of
waiting for two years for instance to make a 2% change in exchange parity, the country
needing to adjust would be adjusting every three months by one-quarter of one percentage
point, and in two years an adjustment of two percentage points in parity would have been
achieved. This system it was thought reduced considerably the adjustment costs. Some
countries actually practiced this system in the dying years of the Bretton Woods system (i.e.
late 1960s-early 1970s) and some still do today.
Another regime advocated and practiced by some countries was the "Gliding -
•
UNIVERSITY OF IBADAN LIBRARY
7
Parities" system". This system involved changing the parity of an exchange rate continuously
at very short time intervals and at very small rates. For instance the exchange rate may be
varied weekly by some .05 percentage point all throughout the year. It was felt that this
system reduced speculation against a currency as uncertainties regarding exchange rate
movements were reduced. This system was also a variant of the adjustable-peg.
By 1971 the stress of inflationary pressures being transmitted from countries with
huge deficits to surplus countries, along with huge balance of payments deficits gave the first
major crack to the adjustable - peg system (in whatever version). In 1971 the United State
of America announced its intention to discontinue convertibility of dollars into gold for
foreign countries holding dollars. By December of that year the Smithsonian agreement was
reached by the major industrial countries of the world, realigning their rates by declaring
new parities. In spite of the above attempt, by mid 1973 most industrial countries of the
,
world started to float their currencies. A floating currency essentially implies that the
exchange rate of a currency is determined in the market by forces of demand for, and supply
of, that currency.
Long before 1973, there have been calls for the adoption of floating exchange rate
regime, as being intrinsically "better" or "superior" to any fixed regime". Three main
reasons were advanced in favour of floating rates. The first is that "exchange rates are
prices" and as such they should be determined by forces of demand and supply and not by
administrative fiat. iecondly it was argued that the par-value or adjustable peg system
UNIVERSITY OF IBADAN LIBRARY
8
necessarily exposed any given economy to the real and monetary disturbances of other
economies; floating rates however would supposedly insulate the economy from such
external shocks. Thirdly, the insulation property of floating rates, would mean that, countries
could now enjoy autonomy in their design and pursuit of monetary and fiscal policies. Thus
the authorities of any given country would be free to pursue the maximization of its people's
welfare free from unwanted shackles imposed by other economies.
Against the adoption of floating exchange rate system were three main arguments. First
that a floating exchange rate creates uncertainty and hence risk, this risk tend to discourage
international trade. Though the risk can be covered by hedging operations, the price of
hedging, itself, is an additional cost of international trade and hence reduces the volume of
trade. Secondly, floating rates, it was argued, would lead to constant shifts in labour and
other resources between production of nontradable goods and production of tradable goods
as the exchange rate is used to adjust the balance of payments. Such constant shift in labour
and other resources would tend to aggravate frictional unemployment, and could be quite
wasteful if the exchange market condition that called for them are temporary. Thirdly it was
argued that the exchange rate could not be relied upon to promote adjustment. Any
appreciable movement of the exchange rate in a given direction was likely to generate
anticipation of a further movement in the same direction. This would likely lead to
speculative capital tran•sfer of a disequilibrating type. Such kind of capital transfer would
intensify the initial disequilibrium and produce" explosive conditions of instability". II Thi=
UNIVERSITY OF IBADAN LIBRARY
9
third argument was buttressed by the floating experience of France between 1924-1926. In
spite of the arguments against floating exchange rate system however, developed industrial
countries of the world still embraced the system in 1973 having been disillusioned with the
Bretton Woods system.
After almost a decade of floating exchange-rate experience in the de eloped
industrial countries of the world, the developing countries of the world too joined the
floating exchange-rate-regime nations. Among the fifteen such developing countries
identified by Quirk et aI., (1987)12 were nine African countries, namely Gambia, Ghana,
Guinea, Nigeria, Sierra Leone, South Africa, Uganda, Zaire and Zambia. The following
events led these African countries to the adoption of this international payments system by
the beginning of the 1980·s.
, Throughout the 19705, Africa, particularly sub-Saharan Africa, experienced a
deterioration in all major economic and social indicators. The GDP grew at an average
annual rate of only 3% percent between 1970 and 1980, thereafter the rate persistently
declined, and by 1983 it was negative. On the other hand population consistently rose at
about 3 percent per annum all throughout the 19705 and into the 1980s, consequently per
capita income declined and by the mid 1980s it was already below its 1970 level.
One of the major factors responsible for the poor perfonnance of these African
economies was their poor production structure. Production in most African economies was
heavily dominated ~ export - oriented agriculture e.g 94% of Uganda's exports was made
UNIVERSITY OF IBADAN LIBRARY
11
themselves in a debt-trap.
With the advent of the debt crisis came the need to drastically cut down on imports.
The drastic reduction in imports soon led to the collapse of the socio-economic
infrastructures, machineries broke down for want of spare-parts, health and educational
facilities got into a state of disrepair, development programmes were virtually halted. Soon
industries had to retrench their staff, as most operated at less tha 50% capa city. Massive
unemployment resulted. Besides, basic goods became scarce and this provide. a good
excuse for prices to shoot up.
What is more-nature itself seemed to have conspired against these economies as most
of them were ravaged by drought in the early years of the 1980s. The drought led to
population displacements, as food shortage became acute and the need to import food
became critical. Thus in the words of Adedeji "... the economic and social transformation
of Africa which started in the 1960s is not only coming to a halt but is gradually reversing" 13 •
Given such bleak prospect, the United Nations Organization, the World Bank, th
International Monetary Fund and their different agencies came together to seek solution to
the "developmental crises" in Africa. These bodies jointly and severally came to the
following conclusions among others - first, that export earnings of African countries have
to be enhanced, second, that the international monetary arrangement must be improved upon,
third that there should be increased flow of resources to Africa from the developed industrial
countries of the orld, including increases in multilateral concessionary financing
UNIVERSITY OF IBADAN LIBRARY
12
arrangements. Lastly that measures must be found to curb the persistent upward movement
in interest rates in the world market.
To deal with the first two propositions, exchange rate liberalization was introduced
to these countries within a broader package of relief measures known as Structural
Adjustment Programmes (SAP). The African countries that were most adversely affected
by the tripartite problems of worsening terms of trade, worsening climatic conditions and
heavy external debt burden opted for the structural adjustment programmes with the floating-
exchange-rate system implicit in the package. Among \:>thers, the floating exchange-rate
system is supposed to help find the realistic values of these countries' currencies. Gi ven that
the currencies of most of these economies were overvalued, a realistic exchange rate for
them, implicitly imply a devaluation of their currencies. Such devaluation supposedly would
encourage exports and discourage imports. Enhanced revenue from exports were also
"
supposed to encourage increased search for potentially exportable commodities thus helping
to diversify the export base of these economies. Increased domestic prices of imports, it was
hoped, would discourage imports while encouraging the inward search for raw materials and
other inputs for domestic industries. Such increased dependence on local resources would
promote economic self-reliance and promote sustainable pattern of consumption.
Furthermore, it was argued that, substantial depreciation that usually followed the
floating of an overvalued currency would encourage repatriation into the economy of
earnings by economic agents that have substantial foreign exchange earnings abroad. It
UNIVERSITY OF IBADAN LIBRARY
13
would also discourage the existence of parallel market for foreign exchange, which had
thrived in these economies in the days of exchange control and had been ready cond its for
capital flight.
In addition to the nine African countries enumerated by Quirk et al (1987) as having
adopted floating exchange rate system, were other developing countries of Asia and
America. These were, Bolivia, Dominican Republic, Jamaica, Lebanon, Phillipines and
Uruguay; of the nine African countries that adopted floating rates, Zambia had since
abandoned the system. Uganda abandoned the system in 1987 but returned to it years after.
One interesting point about the issue of exchange rate regimes, is that since about the
mid-1980s there have been calls from the developed countries of the world requesting for
yet a change in the international payments system from floating-exchange-rate system. In
the words of Devara jan and Rodrik (1991) "... by the mid-1980s the tide turned. Floa ting
,
exchange rates began to lose much of their lustre in the eyes of industrial-coun crypolicy-
makers. The wide gyration of the dollar during the 1980s and the short-term volatility of the
key-currencies eroded confidence in markets' ability to foster adjustment with no (or little)
tears. The Europeans linked their currencies tighter and proposals to limit flexibility became
widespread. In many parts of the developing world, exchange rate flexibility became another
name for inflation."
Also recently some IMP Staff" reporting on the reports of the Group of Ten
(representing the industrial countries participating in the General Arrangements to Borrow)
UNIVERSITY OF IBADAN LIBRARY
14
and the Group of Twenty Four on International Monetary Affairs (an intergovernmental
group) said "...Prior to the advent of floating some of its supporters anticipated that
stabilizing speculation would smooth exchange rate movements and prevent abrupt increases
in variability. Thirteen years of experience has proved otherwise".
The questions then arise - is there an ideal exchange rate regime? Is there any reason
to believe that one exchange rate regime enhances the performance of an economy better
than the other? Given that the goals of most governments as contained in their stabilization
policies are the attainment of high rate of growth in output and in employment, maintenance
of fairly stable price levels and balance in the international payments account; is there an
exchange rate regime that ensures the attainment of these stabilization objectives better than
the other? These are the questions that form the main focus of this study. We hope to answer
, these questions using data for three, of the nine developing African countries enumerated
in the list above, i.e. Ghana, Nigeria and Uganda.
1.2. Need For The Study
Given the four major problems faced by African countries since the 1970s right up
to the 19808 namely, their poor production base, the hostile international economic
environment as evidenced by declining terms of trade, the incidence of natural disasters such
as droughts and cyclones, and lastly the debt fatigue, some African countries sought solution
to these problems by the adoption of the IMF/World Bank (IMF/WB) supported structural
UNIVERSITY OF IBADAN LIBRARY
15
adjustment programs. These countries were Gambia, Ghana, Guinea, Nigeria, Sierra-Leone,
South-Africa, Uganda, Zaire, and Zambia. A very key and crucial element of these structural
adjustment packages is the change of exchange rate regimes from fixed to floating. Among
others, adoption of floating exchange rate regime by these crises-ridden economies was
supposed to help achieve the foIIowing; first, a massive devaluation of their hitherto
overvalued currencies", second, an improvement in exports performance consequent on the
massive devaluation, third, an improvement in the trade balance given a reduction of
imports, fourth, broadening of the production base given that higher import prices would
force domestic producers to produce 10caIIy, goods that were previously imported. In
addition high import prices would also force domestic producers to source raw-materials
locally, thus encouraging the development of backward linkages in the economy. As exports
become diversified given the boost in exporter-income arising from the massive devaluation
,
of domestic currencies, the economies would no longer be monocuItural, thus having greater
bargaining power in the international market place and thus being in a position to stem the
tide of declining terms of trade.
Whether these hopes have materialised after about half a decade to almost a decade
of the adoption of floating exchange rate system is worthy of investigation. Also, where the
above improvements in the economic performance of these countries have not materialized,
it is important to find out whether the structures put in place are actually the necessary
frames for such improv ments to take place. The tables below reveal what has obtained since
UNIVERSITY OF IBADAN LIBRARY
16
the beginning of the 1980s.
Table 1.1 reveals that in the first half of the 1980s per capita income for all the
countries except Gambia and Uganda was on the decline as they all recorded negative
growth rates of per capita income. Even after the adoption of floating exchange rate reg me
which was implicit in the structural adjustment packages, some countries still experienced
declining per capita income notably, Nigeria, Sierra-Leone, Zaire, and Zambia. The case of
Nigeria is particularly striking as its per capita income declined from a high of
US$1117.0 in 1982 to a low of US$368.0 in 1987, thus making her qualify for the
United nations "Low-Income-Economies" classification. That per capita income was
declining is no surprise given the performance of the sectoral value-added and
hence of the GDP as .shown in Tables 1.2 and 1.3 respectively. Table 1.2 reveals that in
, the agricultural sector, performance kept on deteriorating in Ghana and Nigeria in the first
half of the '80s, as evidenced by declining growth rates of value-added in this sector. In
Guinea and Sierra-Leone the agricultural sector was almost stagnant as evidenced by growth
rates of less than 1% i.e 0.3. and 0.7 respectively.
UNIVERSITY OF IBADAN LIBRARY
17
Table 1.1
Per Capita Income(GNPf of Selected African COl!fltries
(In US Dollars)
AVERAGE ANNUAL
GROWTH RATE(%)
1980 1981 1982 1983 1984 1985 1986 1987 1988 1980-85 1987
GAMBIA 384 341 318 293 252 228 211 219 240 0.7 2.4
GHANA 404 408 375 345 366 369 391 393 390 -3.8 0.7
GUiNEi 303 307 299 293 299 315 3\0 316 430 -1.4 3.3
NIGERIA \021 1090 1117 987 902 950 704 368 250 -6.0 -7.7
S. LEONE 317 367 382 388 366 352 325 300 220 -1.3 -2.2
UGANDA 280 220 240 220 220 230 230 260 250 0.1 0.1
ZAIRE 434 399 343 294 204 160 149 153 260 -3.9 -1.8
ZAMBIA 598 718 658 571 472 381 253 232 390 -4.1 -2.4
c
Source: I. V orld '( 989) Afncan Econonuc and Financial Data
2. World Bank (1991) World Bank Development Report
Note: Growth rates spanning more than one year are calculated using Least-square method: single year growth rates are percentage
changes.
,
UNIVERSITY OF IBADAN LIBRARY
18
Table 1.2
SECTORAL VALUE ADDED FOR SELECTED AFRICAN COUNTRIES
(In millions of constant 1980 US Dollars al Producer Price)
AVERAGE
ANNUAL
GROWTH RATE
(%)
1980 1981 1982 1983 , 1984 1985 1986 1987 1990-85 1987
Agriculture
GAMBIA 63 56 69 83 78 il2 91 94 7.6 3.3
GHANA 2575 2509 2372 2207 2421 2436 2517 2517 -1.3 0.0
GUINEA 748 757 765 755 698 805 - - 0.3 -
NIGERIA 26186 21875 22547 22565 21063 24070 24070 26292 -1.5 -7.7
S. LEONE 334 337 348 345 346 346 344 343 0.7 -0.3
UGANDA 1216 1299 1430 1503 l343 1294 1231 1266 1.3 2.9
ZAIRE 2916 3040 3102 3164 3257 3362 3575 3686 2.5 3.1
ZAMBIA 552 597 527 572 304 625 681 674 2.1 -1.0
Industrial
GAMBIA 27 40 43 41 40 40 46 49 5.8 7.2
GHANA 528 443 368 324 353 416 447 498 -5.6 11.3
GUINEA 400 388 693 415 404 387 - - 0.1 -
NIGERIA 41890 37645 36837 31383 29418 32253 31361 29889 -6.1 -4.7
S. LEONE 222 216 216 186 203 198 191 183 -2.5 -4.1
UGANDA 76 77 77 84 89 81 76 89 2.2 14.2
ZAIRE 2969 3134 2134 3180 3490 3571 3685 3719 3.8 0.7
ZAMBIA 1605 1684 1684 1623 1587 1600 1574 1585 -0.6 0.7
I
Sources: I. World BankjUNDP (1989)African Economic and Financial Data
2. World Bank (l99i)World Development Report
Even after the adoption of floating exchange rates value-added in the agricultural sector still
went on declining in some countries i.e. Nigeria, Sierra-Leone, and Zambia. The performance of
the industrial sector is also crucial here, given the fact that it is now an undisputable fact that rapid
industrialisation is a key element of economic growth and development. In the first half of the 1980s
the industrial sector was on the decline in four out
UNIVERSITY OF IBADAN LIBRARY
19
T-ab-le-1.3
GROSS DOMESTIC PRODUCT FOR SOME SELECTED AFRICAN COUNTRIES
(In millions of Dollars at Purchasers Values: Currem Prices
. (CURRENT PRICE SERIES)
AVERAGE ANNUAL
GROWTH RATE (~)
1980 1981 1982 1983 1984 1985 1986 1980-85 1987
GAMBIA 237 236 207 213 210 237 275 4.8 5.8
GHANA 4445 4222 4036 4057 4504 5727 5075 -0.6 4.4
GUINEA 176410 1756 1749 1912 1982 1916 2166 0.9 5.9
NIGERIA 3312 95229 93130 89770 90232 46613 24840 -3.2 -4.6
S. LEONE 1101 1192 1335 1486 1375 1229 895 0.7 -2.0
UGANDA 1724 1370 1893 2283 3797 4442 3895 3.6 4.1
ZAIRE 10281 9089 9065 7708 4462 5566 5774 1.0 2.7
2034
ZA.>.fBIA 3885 4014 3872 3343 2597 1773 -0.2 -0.2
Sources: 1.World Bank/UNDP (1989) African Economic and Financial Data
2. World Bank (1991)World Development Report
of the eight countries in Table 1.2 i.e. Ghana, Nigeria, Sierra-Leone, and Zambia. Even after
the adoption of floating rates Nigeria and Sierra-Leone still experienced a high rate of de-
industrialization as evidenced by negative growth rates of over four percent(4%) in each
country.
The GDP for Ghana, Nigeria and Zambia recorded declining growth rates hich is
to be expected given the performance in the industrial and agricultural sectors of these
UNIVERSITY OF IBADAN LIBRARY
20
economies. The GDP in Nigeria, Sierra-Leone and Zambia still continued to decline up to
1987. Table 1.4 is a reflection of the extent of these countries need for foreign assistance.
The "Resource-Balance" measures the divergence between gross domestic saving and gross
domestic investment as a fraction of GDP. This balance was also negative for all the years
in all the countries. It is not surprising therefore that the external debts table (Table 1.5)
reveals a continuously increasing level of external debt for all the countries. Table 1.6 shows
that for some countries the per capita debt was greater than the per capita income notably
Nigeria, Sierra-Leone and Zambia.
Taking a close look at the external sector's aggregates on which exchange rates
impact directly, it is clear that their performance had not been different from that of the rest
of the economy. Table 1.7 reveals that for almost all the countries, exports were declining
in the years 1980-85 while they started improving from 1986 for the countries of Gambia,
,
Ghana and Zambia all of which had adopted floating exchange rates by that time. Table 1.7
also sows that imports declined continuously in the first half of the 1980s, this decline
continued well into the era of floating exchange rates for Nigeria, Sierra-Leone and Zaire.
The extent of the hostility of the international economic environment in the decade of the
'80s is reflected in the declining terms of trade of the countries as shown in Table 1.8. The
case of Nigeria is particularly remarkable where the terms of trade index declined from
108.7 in 1981 to 43.9 in 1986.
UNIVERSITY OF IBADAN LIBRARY
21
Table 1.4
RESOURCE BALANCE AS PERCENTAGE OF GDP FOR SELECTED
AFRICAN COUNTRIES
AVERAGE ANNUAL
GROWTH RATE(%)
1980 1981 1982 1983 1984 1985 1986 1987 1980-85 1986-87
GAMBIA -21.4 -3.2 -19.1 -12.4 -12.2 -16.3 -16.0 -24.9 -19.0 -20.5
GHANA -0.7 -0.6 0.4 -3.1 -0.3 -1.9 1.4 -3.0 -1.0 -0.8
GUINEA 5.8 3.2 2.6 2.0 3.6 3.6 -2.6 0.4 3.5 -1.1
NIGERIA 9.1 -3.5 -6.1 -3.4 1.7 3.3 -1.8 4.1 0.2 1.1
S. LEONE -14.1 -15.1 -10.6 -10.8 -1.3 -0.4 -0.3 3.3 -8.7 1.5
UGANDA -1.4 -6.1 -3.5 -3.4 -0.3 -2.3 -2.6 -7.2 -3.0 -4.9
ZAIRE -1.2 -5.6 -5.7 -2.7 -3.5 -4.6 1.0 3.2 -1.2 -1.1
ZAMBIA -4.0 -12.5 -8.8 -1.2 -3.8 -2.3 -4.4 4.6 -3.4 0.1
Source: World Bank/UNDP (1989) African Economic and Financial Data.
Table 1.5
, TOTAL EXTERNAL DEBT FOR SELECTED AFRICAN COUNTRIES
(In million of current US S)
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989
GAMBIA 137 176 208 212 230 245 271 319 - -
GHANA 1312 1460 1397 1600 1900- 2176 2656 3124 3099 3078
GUINEA 1117 1356 1356 1368 1293 1466 1568 1784 2563 2176
NIGERIA 8888 12039 12908 18586 18664 19522 24470 28714 20178 32832
S. LEONE 443 480 512 531 437 540 595 659 T2i 1657
UGANDA 730 768 922 1019 1043 1157 1260 1405 1925 1800
ZAIRE 4824 4916 4750 5111 5066 5885 6932 8630 8475 8843
ZAMBIA 3253 3624 3705 3784 3847 4641 5625 6400 6498 6874
-
Sources: I. W Bank/UNDP (1989) African Economic and Financial Data
2. World Bank (1991) World Development Report
UNIVERSITY OF IBADAN LIBRARY
22
Table 1.6
PER CAPITA GNP AND PER CAPITA DEBT IN 1989
FOR SELECTED AFRICAN COUNTRIES
(In US Dollars)
PER CAPITAL GNP PER CAPITA DEBT
GAMBIA 240 -
GHANA 390 214
GUINEA 430 389
NIGERIA 250 289
S. LEONE 220 264
UGANDA 250 108
ZAIRE 260 256
ZAMBIA 390 881
Source: World Bank (1991) World Development Report
Table 1.7
AVERAGE ANNUAL GROWTHRATES, OF EXPORTS, IMPORTS
AND CONSUMER PRICE INDEX FOR SELECTED AFRICAN COUNTRIES IN PERCENTAGES
EXPORTS IMPORTS CONSUMER PRICE
-- INDEX
1980-85 1987 1980-85 1987 1980-85 1987
GAMBIA -2.5 32.2 -5.9 40 13.6 23.5
GHANA -10.3 29.4 -11.1 63.! 57.3 39.8
GUINEA 2.6 - 6.2 - - -
NIGERIA -12.5 -12.6 -13.4 -2.5 19.9 10.2
S. LEONE -7.7 -20.8 -32.9 -2.9 51.7 178.7
UGANDA - - - - - 238.1
ZAIRE 3.8 -13.3 -5.9 -5.9 47.3 19.4
ZAMBIA -3.7 IJ.3 -11.0 9.1 19.6 43.0
Source: World .Bank/UNDP (1989) African Economic and Financial Data.
UNIVERSITY OF IBADAN LIBRARY
23
Table 1.8
TERMS OF TRADE INDEX FOR SELECTED AFRICAN COUNTRIES 1980-1987
AVERAGE ANNUAL
GROWTH RATE (%)
1980 1981 1982 1983 1984 1985 1986 1987 1980-85 1986-87
GAMBIA 100.0 104.0 86.0 91.6 92.4 109.7 109.3 98.4 0.5 -10.0
GHANA 100.0 .81.4 73.1 88.1 98.1 90.3 97.2 89.2 0.7 -8.2
GUINEA - - - - - - - - - -
NIGERIA 100.0 108.7 100.9 96.6 97.2 88.5 43.9 48.1 -2.8 9.7
S. LEONE 100.0 93.1 92.1 94.0 98.6 94.6 93.1 88.5 -0.2 -4.9
UGANDA 100.0 80.6 88.6 84.1 99.7 96.1 116.2 57.4 1.3 -42.0
ZAIRE 100.0 83.5 78.9 83.5 84.4 82.1 80.4 73.0 -0.5 -9.2
ZAMBIA 100.0 79.8 70.9 77.9 70.1 71.5 71.4 79.3 -5.5 11.1
Source: World Bank,IUNDP (1989) African Economic and Financial DaIJl
That there is urgent need for investigating the role of the exchange rate regime in the
, performance of these economies is further buttressed by the fact that while income was
declining as shown in Table 1.1 and 1.3, the rate of growth inflation was rising as shown in
columns six and seven of Table 1.7_In some countries, the rate of flow of domestic prices
was as high as 50% e.g. Ghana and Sierra-Leone. Even after the adoption of floating rates
inflation still remained high as evidenced by an annual growth rate in consumer price index
of 238% in Uganda and 90% in Zaire. What role the exchange regime played in this
inflation-behaviour is worthy of investigation. Given that one of the major reasons for opting
for the floating exchange rate regime is to help these countries resume on the growth path
UNIVERSITY OF IBADAN LIBRARY
24
by adjusting underlying export and import structures the results so far as indicated from all
the tables give no cause for cheer; it then becomes quite timely and quite relevant to find out
if really there is an "ideal" exchange rate regime that ensures "external" equilibrium,
"internal balance" and economic growth.
Furthermore, today, the World Bank staff and the IMF staff are confronted largely
with problems arising from exchange rate behaviour - such problems include those of
misalignment of real exchange rates, effects of exchange rate policy on the agricultural
sector, effect of devaluation on the ability of a developing country to service its debts etc.
H has thus become difficult if not impossible to tackle macroeconomic problems in the
developing countries without addressing the exchange rate issue. 16
In spite of the seeming overwhelming importance of the exchange rate policy to
developing countries there have been very few empirical studies that have attempted 0
"
investigate the problem in all its ramifications. This study is an attempt to fill that gap.
1.3. Objectives of the Study
Prior to the 1980s, concern of economists was whether monetary policy or fiscal policy
was more effective in achieving the goals of stabilization (i.e. high rate of growth in output,
full employment, fairly stable price level and balance of payments equilibrium). Since the
1980s however the focus has shifted from monetary and fiscal policy per se; the question
now asked is whether the objectives of stabilization can best be achieved under a fixed
exchange rate regime, 0 under a floating one, irrespective of the policy instruments adopted.
UNIVERSITY OF IBADAN LIBRARY
25
The answer to this question in essence is what this study is about.
We hope to employ an eclectic model that takes cognisance of both monetary and real
factors in the determination of the exchange rate. The model shall borrow from the works
of Rhomberg (1964), Tullio (1981), Khan and Knight (1981, 1986) Jebuni, et al (1990).
There shall be two versions of the model, the first assumes exchange rates to be fixed as
obtains under a fixed exchange rate regime. The second is an extension of the first and adds
two stochastic equations to the first version; these two additional equations endogenise
exchange rates and interest rates as obtains under a floating exchange rate regime.
Using quarterly data for the years 1977-1990 for Ghana and Nigeria we hope to
specifically do the following:
(1) Estimate our model for each country for all the years 1977-1990 under the
assumption of fixed exchange rates in order to be able to observe the behaviour of
imports, exports, the trade balance, the net foreign reserves, the net capital inflow,
and the growth rate of domestic prices.
(2) Estimate the expanded model for each country for only the years a gi en
country has floated. .-'
(3) Keeping fiscal policy constant, while varying the money supply we hope to
simulate the effects of monetary policy change on trade-flows, net foreign
reserves, balance of payments and inflation under the two versions of the
•
model i.e under a fixed exchange rate regime as well as under a floating
UNIVERSITY OF IBADAN LIBRARY
26
regime.
(4) Keeping monetary policy constant while varying the level of government
expenditure, we also hope to simulate the effects of this policy change on the
external sector's macroaggregates, and on inflation rates under both fixed
and floating exchange rate regimes.
(5) With the results from (3) and (4) above examining the performance of the
external sector, of real income and of inflation both under a fixed as well as
under a floating exchange rate regime when different policy instruments are
applied.
For Uganda the study shall cover the period 1981 to 1990 for which there is data. The
extended model shall be estimated first, because in the available data series the period of
floating comes first i.e. 1982-1987. Thus, result of the estimation of the extended model
, shall form the benchmark. Then the shorter model shall be estimated for the third quarter of
1987 through to fourth quarter of 1990.
Similar simulation exercises as in the case of Ghana and Nigeria shall also be conducted.
From all of the above, we hope to be able to come to a definite conclusion as to which
exchange rate regime has proved more useful in achieving external balance and maintaining
stable prices irrespective of policy instruments employed. And as such which exchange rate
regime is to be preferred.
UNIVERSITY OF IBADAN LIBRARY
27
1.4. Scope and Organisation of the study
In this study we shall collapse all the known exchange ra e regimes into two broad
categories - fixed and floating. The rational for this is not far fetched. The main yardstick
for this division is that while the adjustable-peg, the crawling-peg, the gliding-parities and
the gold standard all have formal parities the free-float and the managed-float do not. Hence
we hope to lump all those with formal-parity together as fixed exchange rate system, and
those without as floating exchange system Table 1.9 adapted Katz (1992) shows the
characteristics of the different exchange regimes.
Table 1.9
Rate Flexibility Under Six Alternative Exchange Rate
Arrangement
Exchange Rate System
Characteristics Free float Managed Gold Adjustable- peg Crawling- peg Gliding-
- float standard parties
Formal Parity None None Formal Formal Formal Formal
Formal Conditions None None Last resort Fundamental Any Any
for change in parity Disequilibrium Disequilibrium Disequili-
brium
Exchange Values in Flexible Flexible Fixed Fixed Less Fixed Less Fixed
the short Run
Desired Exchange Flexible Flexible Fixed Internationally Snghtly Slightly
Value in the Long agreed Flexible Flexible
Run Flexibility
Source: Adaptea fTom Katz S.I. !1972 ) RThe Case for the Par Vaiue System:
1972" Essays in International Finance, No. 92, March 1972 (Princeton
University: International Finance Section).
The rational for the choice of the three countries of Ghana, Nigeria and Uganda is not far
•
fetched. In the first instance the three countries have experienced both fixed (in almost all
UNIVERSITY OF IBADAN LIBRARY
28
its varieties) and floating, exchange rate regimes. Secondly they are all chronic debtors
needing some growth oriented economic package to get out of the woods. Thirdly, they all
have high degree of openness", and they are all monocultural economies, whereby a greater
percentage of their export earnings comes from only a single commodity. Table 1.10 below
gives the general features of the countries used in the study.
Table 1.10
Features of Focus Countries for 1985
Country Per Capita GDP" at Degree of Degree of Export Concentration"
Official Exchange Openness" Import , (Percentage)Rate Dependence"
Cocoa (65.2) Petroleum
GHANA 467 24.4 13.9 (98.4)
NIGERIA 854 44.0 14.9 Coffee (93.5)
UGANDA 129 41.0 23.3
, Source: Adapted from Table 6 pp 35-36 of Quirk P.J., Christensen B.V., Hug, K-M, Sasaki T.
(1987) "Floating Exchange Rates in Developing Countries: Experience With
Auction And Interbank Markets." Occasional Paper No. 53, May 1987
(IMP).
Notes:
·Per capita GDPin U.S. dollars
'Degree of Openness is measured by the ratio of total trade in goods and
non-factor services to
'Degree of Import Dependence, is the ratio ofImport of goods and non-factor
services to GDP
dExport Concentration is a measure of the country's ability to earn foreign
exchange, and it is in percentage .
•
UNIVERSITY OF IBADAN LIBRARY
29
In this study we shall be concerned with both real exchange rates and nominal exchange
rates. This is because real exchange rates reflect better the degree of a country's
competitiveness in the international markets, than do nominal exchange rates.
We shall study the economies of Ghana and Nigeria over the period of 1977-1990 and
over 1981-1990 for Uganda. We shall use quarterly data on exports, imports, trade balance
net foreign reserves, net capital inflows and prices, to enable us have fifty-six observation
points for the countries of Ghana and Nigeria and forty points for Uganda. For Ghana and
Nigeria who are the most recent floaters, there shall be 17 observation points of floating
experience. The number of years of fixed or floating-experience would defer for each
country, becau e the onset of floating regime was not uniform across countries. We hope
however that this slight variation in timing will not detract from the necessary economic
insights we hope to gain from the study.
The current chapter is the introductory chapter which describes the problem we hope to
look into, why there is need for such an exercise, and the goals we hope to attain from the
study. Chapter two provides a review of the case economies. Chapter three is a review of the
literature on exchange rates in general and on the performance of the macro economy under
floating-and/or under fixed-exchange rate regimes in particular. While chapter four describes
the methodology of the study, chapter five is a descriptive analysis of the data. Chapter six
gives the result of the model(s) estimations and empirical validation of the models, while
chapter seven discusses the result of the simulation exercises. Chapter eight gives a summary
UNIVERSITY OF IBADAN LIBRARY
30
of the findings, the conclusions, recommendations and areas for further research.
ENDNOTES
1. Boughton J.M. (1988): "The Monetary Approach To Exchange Rates: What Now
Remains?" Essays In International Finance, No. 171 Oct. 1988, (Princeton
University, International Finance Section) P. 5.
2. Branson W.H. (1977): "Asset Markets And Relative Prices In Exchange Rate
Determination". Reprints In International Finance No. 20 (princeton University,
International Finance Section) p. 5.
3. Edwards S. (1988): "Exchange Rate Misalignment in Developing Countries. (World
Bank Occasional Paper, No. 2[New Series) p.4.
4. See Machlup F. (1940): "The Theory of Foreign Exchanges" ECONOMICA, Vol.
7, PP.23-49.
5. Cooper R.N. (1987): "The Gold Standard: Historical Facts and Future Prospects" to
be found in Cooper R.N. (1987). The International Monetary System: Essays In
World Economics (MIT, U.S.A).
6. See IMP (1944): "Articles of Agreement" (Bretton Woods, July 1944) Article IV
Section 5).
7. Coombs C.A. (1976): "The Arena oflnternational Finance (New York: John Wiley
& Sons).
8. Katz S. I. (1972): "The Case for the Par-Value System: 1972" Essays In International
Finance." No. 92, March 1972 (Princeton University: International Finance Section).
9. Cooper R.N. (1969): "Flexing The International Monetary System The Case For
Gliding Parties" in Johnson H.G. and Swoboda A.k. (eds) The Economics of
Common currencies. (London: George Allen & Unwin Ltd) Chap. 14. pp. 229-243 .
•
UNIVERSITY OF IBADAN LIBRARY
31
10. Among others See (i) Friedman M. (1953): "The Case for Flexible Exchange Rates"
in Friedman (M). Essays In Positive Economics (Chicago, University of chicago
Press) pp. 157-203.
(ii) Sohmen E. (1965)
Flexible Exchange Rates (Chicago: The University of Chicago Press).
(iii) Johnson H.G. (1973): "The Case for Flexible Exchange Rates in
Johnson H.G. Further Essays in Monetary Economics (Harvard:
Harvard Univ. press) pp. 198-228.
11. Nurkse. R. (1944): "International Currency Experience (Geneva, League of Nations).
12. Quirk, 0.1., Christensen, B.V., Huh, K.M. and Sasaki, T. (1987): "Floating
Exchange Rates In Developing Countries" (Washington. D.c; IMF Occasional Paper
No. 53).
13. Adedeji, A. (1984): "The African Economic And Social Crisis: An Agenda For
Action By Africa And The International Community" (paper Delivered At The 1984
Second Regular Session Of The Economic And Social Council Of The United
Nations, Geneva) 10-12 July 1984. .
14. Corckett A. and Goldstein M. (1987): "Strengthening the International Monetary
System: Exchange Rates Surveillance and Objective Indicators. (Washington D.c.
IMP Occasional Paper NO. 50).
15. (i) Oyejide T.A., Soyode A. and Kayode M.O. (1985): "Nigeria And The IMF
(Nigeria, University of Ibadan Press). p.71.
(ii) Ajayi S.I. (1989): "Issues of Overvaluation And Exchange Rate Adjustments
In Nigeria (Washington D.C. Economic Development Institute, World
Bank).
(iii) Nguyuru H.I. Lupumba (1990): "The Appropriate Exchange Rate
Policy For African Countries: Some Policy Issues" (paper Presented
At The Eastern and Southern Africa Economic Association
conference - Nairobi) June 2-3, 1990 .
•
UNIVERSITY OF IBADAN LIBRARY
32
16. Edward S. and Ahamed, L. (1986): "Economic Adjustment And Exchange Rates In
Developing countries. (A National Bureau Of Economic Research Conference
Report. Chicago University Press, Chicago) p.1
17. A managed float is a system where though the exchange rate is determined by forces
of demand and supply, government sometimes intervenes in the market to prevent
excessive exchange rate variability.
18. Bosco. L. (1987): "Determinants of Exchange Rate Regimes in LDCs: Some
Empirical Evidence". Economic Notes I 1987. pp122.
UNIVERSITY OF IBADAN LIBRARY
CHAPTER TWO
REVIEW OF CASE ECONOMIES
"The Most Important Market In This Economy Is The Market For Foreign
Exchange"
Elbadawi (1990)
"The Naira Exchange Rate Has Since 1986 EmergedAs Perhaps The Most Dynamic
and Controversial Instrument of Economic Policy In Nigeria"
Omoruyi (1989)
The positions of Elbadawi and Omoruyi cited above succinctly illustrate the
importance of exchange rate policy and hence exchange rate regimes in sub-Saharan African
countries generally and in the focus countries of this study in particular, especially since the
onset of structural adjustment programme, with its implicit exchange liberalization strategy.
Below we attempt to provide a brief background information regarding the economies of
Ghana, Nigeria and Uganda with particular emphasis on their exchange rate management
policies since the 1970s.
2.1 Ex~han~ Rates and Exchange Rate Policy in Ghana 1971-1990
Prior to 1983 Ghana adopted a fixed exchange rate system of the Bretton-Woods
variety. Between 1967 and 1970 the exchange rate was 1.20 per $1. By 1971 the cedi ( )
was devalued to 1.28 per $1. By 1972 the exchange rate appreciated to 1.15 per $1 and
remained so until June 1978.
Prior to 19~7 the pound sterling was Ghana's intervention currency. However from
UNIVERSITY OF IBADAN LIBRARY
34
1977 the US dollar became Ghana's intervention currency and had remained so ever since.
Between June to August 1978 the exchange rate was made to crawl, according to Ameyaw'
from 1.15 per $1 in June to 1.30 per $1 in July and to 1.60 per $1 by August 1978. From
August 1978 the monetary authorities decided to fix the exchange rate permanently once
again. Thus the exchange rate was fixed at 2.75 to $1 until 1983.
Throughout the 1970s Ghana's inflation rate far exceeded those of her trading
partners. According to Chhibber and Shafik (1989) the Ghanaian inflation rate between 1973
and 1980 was about 60.3% per annum. Thus the rea) exchange rate during this period
continuously appreciated, in fact some economists have claimed that between 1970 - 1982
the real exchange rate in Ghana appreciated by over 500%. The real appreciation of the cedi
encouraged black rnarketeering in foreign exchange as parallel market premium soared with
increasing appreciation of the cedi. The parallel market premium is said to have risen from
8% in 1976 to 4264% in 1982 (Dordunoo, 1992)
The above scenario encouraged capital flight throughout the 1970s and early 1980s.
In addition to increasing appreciation of the domestic currency, other factors that contributed
to capital flight were the inappropriate policy stance of government, and the overdependence
of the economy on a regulatory framework. For example the policy stance of government
in the early 1980s that (i) Bank deposits of individuals greater than 50,000 be frozen
pending investigation for tax liability or fraud and (ii) Requirement that business
transactions exceeding 1,000 (cedis) be conducted by cheque. Similarly the continuous
•
UNIVERSITY OF IBADAN LIBRARY
35
ceiling on deposit rates kept bank deposits rates ridiculously low while lending rates were
exorbitantly high. Domestic saving and investment were impaired by this regulatory
framework in addition to encouraging capital flight. What is more - the lowered level of
investment affected production both for domestic consumption and for export. Borrowing
from the banking system to finance imports became unprofitable in the face of exorbitant
lending rates. Thus, imports of essential consumer goods and producer inputs also fell
sharply. This had negative multiplier effects on the economy, as production in both the
agricultural and industrial sectors declined sharply thus exerting greater pressure on
inflation.
Declining production, declining income, and increased external borrowing soon
contrived to make Ghana one of the greatest debtors in the world in the manner of sister
..• African countries like Nigeria and Uganda. Inability of Ghana to service its external debts
as, and, when due, led to the drying up of external credit lines. By 1983 Ghana's external
debt was put at some $600 million, and so it decided to overhaul her economy and start all
over. Thus in 1983, Ghana adopted an IMF/World Bank supported Economic Recovery
Program (ERP), one of the cornerstones of this program like others of its kind was the
adoption of a flexible exchange rate system in order to be able to correct for the
overvaluation of the domestic currency, encourage exports and discourage non-essential
imports. Other elements of the ERP were trade liberalization, public-enterprises reform,
•
fiscal reform such as reduction of subsidies and rationalisation of the public expenditure
UNIVERSITY OF IBADAN LIBRARY
36
program.
In April 1983 when the ERP was adopted the exchange rate was devalued from its
level of 2.75 per $1 to 14.53 per $1 a depreciation, of about 81% in one fell swoop. As
part of the adoption of a flexible exchange rate system were some new forex measures. The
first was the Bonus and Surcharge Scheme, whereby exporters were granted bonuses on
forex receipts and importers were surcharged for forex payments. Under this scheme exports
was divided into two categories, the first category was that which allowed an exporter to
obtain a bonus of 9.9 times the domestic value of forex earned, and the second category was
that which entitled the exporter to 7.5 times the domestic value of forex earned. For instance
if an individual has exported goods worth $1,000, and if the exchange rate is 14.53 cedi to
$1, then instead of earning 14,530, the exporter would earn (14,530 x 9.9) cedis, if the item
'falls under the 9.9 category".
As for the surcharge, anytime an individual or firm imports through the banking
system he is surcharged. In effect there was a multiple exchange rate system with a weighted
average of 30 to $1. This multiple exchange rate system was rather cumbersome to handle,
so by October 1983 the system was abolished and the exchange rates were unified to 30
per $1. By 1984, a real exchange rate rule under the framework of purchasing power parity
was put in place, whereby the exchange rate was adjusted quarterly, based on the inflation
differential between Ghana and her major trading partners. This real exchange rate rule was
•
in place until the second quarter of 1986, by which time the exchange rate has reached 90
UNIVERSITY OF IBADAN LIBRARY
37
per $1, the government then fixed the exchange rate at 90 per dollar until September 1986
when government decided to·float the exchange rate independently.
During the period April 1983 - August 1986 the black market premiun remained
high standing at about 183% which indicated the rationing of forex from official sources.
Imports remained controlled within the framework of an annual import programme. There
was a scheme for producers to utilize Special Unnumbered Licenses (SULS) which were
mainly a channel for remittances from Ghanaians abroad (Dordunoo, 1992). In September
1986 the government decided to float the exchange rate independently, using the Dutch
Auction System (DAS). Government opted for independent floating of the exchange rate
supposedly to depoliticize the issue of exchange rate adjustment, and, secondly, to enable
the authorities ensure an exchange rate that is in line with the fundamental forces of
demand and supply (i.e an equilibrium exchange rate). Further it was believed that
floating the exchange rate minimized capital flight and prevented foreign ex hange
dealers (mainly the commercial banks) from collusive behaviour.
When the auction system started in September 1986 there were in fact two-tiers of
exchange rates in the system. On the first-tier or the first-window the exchange rate was
administratively determined. It was pegged at 90 to $1. This exchange rate was used for
the importation of" essential goods and services". Such essential goods and services included
crude oil imports, import of petroleum products, drugs and loan - amortization. On the
•
second exchange window the exchange rate floated, being determined by the amount of
UNIVERSITY OF IBADAN LIBRARY
38
forex supplied by the BOG, and that, demanded by the dealers on behalf of their customers.
Licenses were issued to determine on which window an individual or firm could transact
business. There were three categories of licenses - 'A' 'S' and 'G'. The 'A' license category
qualified an individual for participation in the auction. The 'G' license category qualified
the holder to source his forex needs on the first window. While an'S' license meant that the
holder should source for his forex resources autonomously (i.e outside the official window
and outside the auction market).
By February 1987 the two windows were merged and all items qualified to come
under the "Auction" system. By January 1989 import-licensing system was completely
abandoned. Import declaration forms which remained in place were meant only for statistical
purposes. While the two exchange windows lasted i.e between September 1986 to February
1987, the surrender of exchange earnings to the Bank of Ghana was effected at two rates.
Earnings from exports of cocoa and residual oil products were to be surrendered at window
1 exchange rate, while all other transactions were at window 2 exchange rate. Transactions
on window 2 made up about 66% of all external payments and receipts.
The auctioning system adopted in 1986 was of the 'retail' type wherein the BOG was
empowered to sell forex to "final users" only. Thus the commercial banks (or dealers) had,
a very limited intermediary role to play in the system. Their main functions were to collate
the bids for auction funds from their clients and then channel these bids to the BOG. Every
Friday at the BOG each •dealer bid on behalf of its customers. The point where the supply
UNIVERSITY OF IBADAN LIBRARY
39
of forex by the BOG becomes exhausted becomes the "marginal rate". In the very first
auction this marginal rate was adopted as the forex 'price' at which every bidder bought
forex. Subsequently however the pricing system was changed to the Dutch Auction System
(DAS) whereby each bidder bought forex at the rate at which he bid. The DAS remained
in use till April 1990 when the Modified Dutch Auction System (MDAS) was introduced.
As an epilogue one may point out here that by April 1991 an "interbank forex market
was put in place whereby banks traded in forex among themselves, the sources of their forex
being the fraction of their export proceeds that the BOG allows them to keep, and other forex
resources from autonomous sources. By March 1992 the BOG stopped the auction system
completely leaving only the "interbank market" in place. Under this interbank market system
the BOG maintains what it calls a Dealing Desk wherein banks needing more forex
resources than they can get from fellow dealers buy from the BOG at the Dealing Desk. Also
under the platform of the "Dealing Desk, the BOG buys and sells forex at prices close to
those obtainable at the interbank market and at the parallel marker'.
In order to ensure that a single bank does not comer the whole interbank market the
BOG had an "opening-balance" rule which stipulated the maximum amount of forex any
bank must have as its opening balance, any excess above the "opening-balance" required by
the BOG has to be sold to other banks, anyone violating this rule was banned from
purchasing from the BOG's Dealing Desk whenever the need arose.
Within the period of this study however i.e between 1977-1990 it is important to
UNIVERSITY OF IBADAN LIBRARY
40
point out an interesting development. In February 1988 the "Foreign Exchange Bureau De-
Change System was approved in principle. By April 1988 the bureaux came into operation.
The bureaux were meant to cater for the needs of small buyers and sellers of forex, and, to
try and, absorb the parallel market into the official market thereby expanding the officially
available forex, and for purposes of control.
According to Chhiber and Shaffik (1989), and Fosu (1990) non-traditional exports
have been given a boost since the onset of floating rates. This improved performance of non-
traditional exports is said to be due to the boost in exporters income as a result of the
devalued currency. Secondly the fact that exporters could now retain a significant fraction
of their export earnings also encouraged exports. Also Chhibber and Shaffik maintain that
whereas inflation rate in Ghana between 1973 - 1980 was about 60.3% annually, by 1985
,
(two years after the onset of ERP) the inflation rate came down to some 10% per annum, and
even though it has risen to almost 30% by 1988, inflation rate is still far less than what
obtained prior to the adoption of the ERP with its attendant flexible exchange rate system.
Though import bills have gone up, a lot of the increased import is in the productive
sector of the economy hence it is potentially useful for economic growth and development.
And the increased imports (if Khan and Knights' 1986 thesis holds) will definitely improve
export performance as most productive sectors, even agricultural sector, depend on imported
inputs and equipment. •Also fiscal performance is said to have improved in Ghana, given the
inflow of adjustment grants and aids.
UNIVERSITY OF IBADAN LIBRARY
41
2.2 Exchange Rates and Exchange Rate Management in Nigeria Before and After SAP
In the 1970s Nigeria "adopteda single currency peg. As of 1973 it pegged its currency
to the US dollar and the exchange rate was "H0.65 to $1, and when the dollar was devalued
in 1973 Nigeria also devalued the naira in sympathy. By 1974 however it became obvious
that pegging to the US dollar which then had become very weak was inadvisable. Thus the
country adopted a new system whereby it pegged either to the US dollar or the pound
sterling depending on which one proved stronger in the world forex market.
As a result of the boom in petroleum prices between 1973 and 1974, Nigeria had
huge and persistent external surpluses. Inflationary pressures mounted and the naira started
to appreciate relative to the US dollar and soon the naira became overvalued. Appreciating
, the naira was thought. to be helpful in that it was thought to make "import-substitution-
industrialization" realizable, and further, that, naira appreciation was anti-inflation. In 1976,
half-hearted attempts were made at depreciating the naira, and later, it was decided to peg
the naira to a basket of currencies. A basket made up of the currencies of seven major
trading partners of Nigeria was chosen, and between 1976 and 1985 Nigeria experimented
with the basket peg. Though the naira was stable it was however considered overvalued in
both nominal and real terms (relative to the US dollar). By 1980 the exchange rate stood at
#0.54 per $1.
In the 1970s Nigeria maintained a regime of trade and exchange controls. There were also
several administrative devices to pep up the trade controls. Among these controls were.
UNIVERSITY OF IBADAN LIBRARY
42
(i) The requirement for "approval" by importers from the Central Bank of Nigeria in
order to purchase forex.
(ii) Compulsory surrender of forex receipts by importers.
(Hi) Channelisation of all forex transactions through the authorized dealers (i.e
commercial and merchant banks.
Administrative and quantitative rules to pep up the trade and exchange controls included,
import licensing, form 'M', and Comprehensive Import Supervision scheme.
As a result of the 1973/74 oil boom and the attendant surge in Nigeria's international
reserves, the country opted for forex management through portfolio diversification (Ojo,
1990). Given that this period coincided with the adoption of floating rates by the developed -
, industrialized countries of the world, there was a general instability in the global monetary
" system. So, Nigeria had to distribute its forex assets in order to reduce the risk of capital
loss. To achieve the aim of asset diversification an Investment Management Committee was
established in December 1974 charged with the responsibility of investing in securities of
Varying degrees of maturity denounciated in major international currencies. By the end of
1976 the Central Bartleof Nigeria's external assets were held in nine convertible currencies.
From February 1978 to August 1986 the exchange rate of the naira was based on import-
weighted basket of currencies supplemented by such factors like "reserve-level" "cross-rate-
considerations", "relative-rates-of-inflation" and discretional judgement of the perceived
relative strengths of various currencies of major trading partners" (Ojo, 1990). The actual
UNIVERSITY OF IBADAN LIBRARY
43
determination of the naira exchange rate appeared to have been largely influenced by the
level of forex inflow and reserves. As a rule, naira exchange rate was appreciated with an
increased inflow of forex, but a corresponding depreciation never took place with a reduced
forex flow. On the whole, because the exchange rate policy encouraged appreciation rather
than depreciation of the naira, it became counter productive, undermining the incentives to
produce for export, encouraging imports, sustaining overdependence of manufacturers on
imports and encouraging the growth of parallel forex markets and capital flight. By 1985 the
exchange rate stood at #0.89 per $l.
The Second-Tier Foreign Exchange Market (SFEM) came into operation on 26th of
September 1986. In this market, the naira was floated independently. The second-tier
exchange rate applied to all transactions except debt-service payments, embassy expenses,
,
subscriptions to international organisations, and settlement, of transitional or pre-SFEM
transactions. Dealers (made up mainly of commercial and merchant banks) were required
to bid for foreign exchange subject to a maximum of 10% of available forex by big banks
and 7% for small banks. Their bidding would be at prices that appeared to them to
approximate the market perception of demand for and supply of forex. Under this
arrangement some banks did not win at the auctions. So the banks were classified into 'big'
, 'medium' and 'small' and each category was entitled to a maximum of 5%,2% and 1.5%
respectively. In the first.two auctions, the exchange rate was determined by averaging the
successful bid rates. After the first two biddings the average-rate method was replaced with
UNIVERSITY OF IBADAN LIBRARY
44
the marginal-rate-pricing-method, which, it was believed, had the advantage of reducing the
continuous depreciation of the naira. By April 1987 when the exchange rate continued its
downward slide a supposedly more professional system was introduced i.e the Dutch
Auction System. Under this system dealers were allocated funds at their bid rates. The
objective was to reduce the propensity for high bids. The result was multiplicity of bid rates,
but the official exchange rate was determined at the margin. Further the bidding period was
revised from once a week to once in two weeks.
By July 1987 the first tier market (where the exchange rate was fixed by administrative
fiat) and the second-tier market were merged to produce just one exchange market known
as FEM (Foreign Exchange Market). The essence of the merger was to eliminate the implicit
subsidy on first tier transactions, rationalise the allocation of forex, and eliminate multiple
..
currency practices. There was introduced in 1988 what was known as the "autonomous"
market or the interbank market. Under this arrangement banks were free to source for
foreign exchange independently of the central bank, and sell to each other at rates
determined by themselves subject however to a maximum spread of one per cent. The
objective was to encourage non-oil foreign exchange inflow to banks and thereby relieve the
pressure on the official market. By January 1989 the "autonomous market" and the FEM
were merged, and the Inter-bank Foreign Exchange market (lFEM) was born. The essence
of this merger was to uaify the rates in the official and the autonomous segments of the
market in order to eliminate distortions in the exchange market.
UNIVERSITY OF IBADAN LIBRARY
45
The operations in lFEM were conducted on a daily basis, the exchange rate was
determined by using one or a combination of the following
(i) Weighted average of all quotations submitted by banks:(i.e individual bank's quotation
weighted by amount demanded).
(ii) Simple average of all quotations submitted by banks
Hi) Highest and lowest banks' quotations provided the latter does not deprecia te by more
than 2% when compared with the rate that emerges in (ii) above, and
(iv) Intelligence reports on exchange rate movements during the previous day both in
interbank market and in some world financial centres. (Olisadebe 1991).
From 1989 for purposes of forex allocation all dealing banks were classified into four
categories ranging from biggest to smallest. The classification was based on the weighted
,
average of a number of factors which included
(a) the level of a bank's net assets
(b) level of its net profits
(c) level of its loans and advances
(d) level of its deposit liabilities.
The proportion of total allocations that went to each group was a function of, the total
number of banks falling into a particular group, the amount of forex government was willing
to put into the market, asd the stock of contingent liabilities of the central bank.
At the onset of lFEM the practice whereby dealers bought forex on behalf of their clients
UNIVERSITY OF IBADAN LIBRARY
46
was abolished. So that dealers now bought forex on their own account. After the allocation
has been taken up, each bank's account with the Central Bank was debited with the naira
equivalent of the allocations. For this purpose the CBN's selling rate which was usually
0.5% above the market rate was applied. On a daily basis dealers submit bids on their own
account to the CBN. The bids were arranged in descending order of value, the bid rate that
cleared the market represented the marginal rate. Though the bid rates influenced the
determination of the CBN's "effective" rates, the criteria for determining the naira exchange
rate were not unique. However, once the rate was determined the CBN's selling rate was
regarded by the banks as the "base" rate.
When the exchange rate was floated in 1986 some complementary measures had to be
, taken, which among others were
(i) the dissolution in December 1986, of the Commodity Boards which had hitherto
been responsible solely for marketing non-oil agricultural exports.
(ii) abolition of imports and export licensing.
(Hi) establishment of foreign exchange bureaux in 1989 to handle small cale forex
transactions based on non-official funds.
The first of these measures was meant to encourage individual exporters. More
importantly the creation of "Foreign Currency Domicilliary Accounts", and the permission
granted exporters to•retain 100% of their forex receipts were aimed at reinforcing the effects
of the dissolution of the commodity boards. The third measure was aimed at reducing
UNIVERSITY OF IBADAN LIBRARY
47
pressure on official forex resources and reducing illegal forex practices. Institutional
facilities to further encourage·the promotion of non-oil exports included
(i) New Shipping Policy under the aegis of the newly established National Maritime
Authority. The policy was based on a 40-40-20 formula whereby Nigerian vessels
were allowed to lift 40% of the value of trade while trading partners and third party
vessels were allowed to lift 40% and 20% respectively. t
This policy was meant to enhance Nigeria's forex earnings in the maritime trade.
(ii) Other export incentives were the Rediscounting and Refinancing Facility (RRF)
meant to induce authorized dealers (mainly commercial and merchant banks) to go
into export financing. Loans given to exporters by the dealers can be refinanced by
the CBN at its minimum Rediscount rate thus providing a good margin to the dealers
, who lend at commercial rates.
(iii) Also established was the Nigerian Export Credit Guarantee and Insurance
Corporation (NEXIM) which was meant to provide further incentives to exporters
through credit guarantee, loan guaranteeand insurance. To keep forex expenditure
within reasonable limits, a strict forex budget control system was designed.
As earlier mentioned, in August 1989 in order to further combat the issue of excessive
parallel market premium, the government approved the establishment of "Bureaux. De
Change". These are institutions specifically licensed to undertake forex transactions on a
relatively small scale. The• bureaux were believed to be capable of broadening the scope of
UNIVERSITY OF IBADAN LIBRARY
48
legitimate foreign exchange transactions by providing free access to forex to small users in
a convenient and informal manner. On 14th December 1990, the Dutch Auction System was
reintroduced to check foreign exchange hoarding by banks, and submission of multiple bids
(which inflated the demand for forex). With the reintroduction of DAS banks were
regrouped into six categories with the largest banks being entitled to 4.70% each, of total
forex available and the smallest banks being entitled to 0.45%. The marginal rate continued
to be used to determine the official rate i.e the rate that exhausted the supply of forex. This
rate was marked up by 1% to determine the effective rate.
In order to re-align the demand for with the supply of forex certain complementary
measures were taken between 1986 and 1990. Among others these included the lowering of
prescribed ceiling on credit expansion in January 1989. The commercial and merchant banks
cash reserve requirements and the Central Bank's Minimum Rediscount Rate (MRR) were
"
revised upwards. Further, banks and other financial institutions, were, from April 1989,
prohibited from granting loans denominated in naira on the guarantee of foreign assets,
failure to comply with this regulation meant the offending bank had to deposit an amount
equal to the principal of loan so granted to the CBN. In June 1989 all deposits of the Federal
and State governments and those of their agencies were required to be transferred from the
commercial banks to the CBN. By July 1989 deposits so transferred totalled N8.5 billion.
From August 1989 stabilization securities were issued by CBN, and, banks thought to
possess excess liquidi~ are forced to purchase such stabilization securities. On the supply
UNIVERSITY OF IBADAN LIBRARY
49
side government continues to encourage non-oil exports and earnings therefrom. As an
epilogue one might add that from March 1992 government had adopted the use of forex
bureaux rate as its own selling rate and authorized dealers are free to purchase their forex
needs from the CBN on a fortnightly basis.
On the very first quarter of floating in 1986 the exchange rate depreciated by about 63%
to stand at N3.6 per $1. By the time the forex bureaux were established in 1989 the exchange
rate had further tumbled, depreciating by some 52% from its early floating period and some
82% from the pre-float era to stand at N7.5 per $1. By the end of 1990 the exchange rate was
N8.3470 per $1. As an epilogue one might add that by the end of 1992 exchange rate had
reached the rate of N20.5 per $1.
Government officials have since the onset of floating rates maintained that non-oil
exports have expanded. Olisadebe (1991) claimed that the share of non-oil exports to total
exports increased from 3.2% in 1981-85 to 6% in 1986-1990. Ojo (1990) maintained that
the value of non-oil exports averaged N2.7 billion between 1987-1989 as against NO.4 bill
between 1984-1986. Growth in the value of imports in dollar terms has been much less than
its growth pre-floating, although in naira terms imports value increased substantially. Ojo
further maintained that the structure of imports has shifted slightly from consumer goods
towards capital goods and raw materials. Official foreign exchange reserves have increased
from N3.3 billion H1 1988 to N13.8 billion in 1989. At this level the reserve level could
cover 3.8 months of forex commitments at the prevailing rate of disbursement as against 1.3
UNIVERSITY OF IBADAN LIBRARY
I
50
months at the end of 1988.
Olisadebe has further maintained that the current account deficit as a percentage of GDP
which averaged 5.6% per annum in the years 1981-1985 has now improved to an average
surplus of 3.4% in 1986 - 1990. To what extent these improvements are real or illusory will
be determined in this study.
2.3. Exchange Rates and Exchange Rate Management in Uganda.
In the 1970s Uganda operated a fixed exchange rate system. The exchange rate by
the mid-1970s was seven Ugandan shillings to one United states dollar (Ush7 per $1). At
this time there was no active exchange rate management policy. There were however three
series of external shocks in the 1970's which Uganda ought to have reacted to, in terms of
~
its exchange management policy, but which, it failed to react to. These were the oil crisis of
1974, whereby the price of crude petroleum shot-up by over 400%, then another oil-price
hike in 1978, and then the commodity price collapse of 1979, whereby the price of coffee
(the main export crop of Uganda) came tumbling down.
In spite of these external shocks the exchange rate remained fixed at Ush7 per $1. What
is more - there were political upheavals in Uganda in the 1970s some of which brought
sanctions against Uganda by the international community. The political upheavals in Uganda
involved the expulsi~n of the major operators of the economy from the country. This
expulsion, without adequate replacement, led to declining production from 1972- 1979.
UNIVERSITY OF IBADAN LIBRARY
51
Inflation rate soared, exchange rate remained fixed and the economy had to rely heavily on
the country's external reserves to sustain the exchange rate at Ush7 to $1. Sanctions against
Uganda led to the blockage of capital inflow in terms of aids, grants and loans at
concessionary rates. It also led to the drying up of credit - lines, and imports had to be on
cash-and-carry basis.
In time real GDP declined to negative levels, budget deficit was increasing at an annual
rate of23% and was financed by simply printing new money. Triple digit inflation resulted,
and soon a thriving parallel market for foreign exchange known as the "Kibanda" resulted.
Before the end of the decade of the 1970s the differential between the official exchange rate
and the parallel market rate was about 1:304•
, In the face of acute foreign exchange shortage in the official sector, recourse was had
to external borrowing at commercial rate. In time the country got cut in a debt-web.
Meanwhile efforts at home were concentrated at fighting the parallel market. In 1979
there was a civil war, and in 1980 there was another political crisis both succeeded in
weakening the economy the more.
As external debt piled up, external reserves got depleted to unbearably low levels,
gross domestic product fell, per capita income fell, inflation rate soared, the government then
decided in 1981 to take a drastic step to arrest the situation. Thus in 1981 the government
devalued the exchange tIlte. Between June 1981 and 1982 greater efforts were put at making
the exchange rate system more flexible hence series of devaluations took place during this
UNIVERSITY OF IBADAN LIBRARY
52
period, with the exchange rate moving from Ush8.12 to $1 in May 1981 to Ush76.75 per $1
in June 1981 and to Ush80.52 per $1 in September 1981 and Ush86.03 per $1 by June 1982.
By 1982 the government decided to float the exchange rate independently. The
government adopted a two-tier foreign exchange system whereby at one level foreign
exchange was administratively fixed by government and this rate was used for
government imports and external loan repayment, while at the second level forex was bought
and sold according to the forces of demand and supply. The rate at the second window was
made as high as that in the pa"rallel market in an attempt to eliminate the parallel market.
This attempt however did not succeed. Attempts to eliminate the parallel market did not
succeed for two reasons. First the officially generated forex resources was not sufficient to
meet total demand for. forex in the economy. Secondly the bidding system was slower
,
relative to the parallel market system due to bureaucracy.
In order to promote exports government decided that exporters could retain their
forex earnings. However these exporters preferred to siphon such forex earnings to the
parallel market rather than to the official market. What is more - government usually
guaranteed loans to exporters of coffee, teat and a few other export commodities. However
revenues from these exports do not fmd their way to the BOU rather they found their way
to the parallel market. So in order to bring exporters to the official market the Ugandan
government permitted Jhe opening of "Foreign Currency Domicilliary Accounts" and
allowed the financing of imports through these accounts.
UNIVERSITY OF IBADAN LIBRARY
53
When the bidding sessions started in June 1982 it was financed by the IMF while the
BOU supervised the weekly auction. In 1984 the IMF standby arrangement from which the
auction was financed came to an end and Ugandan monetary authorities had to fund the
forex market themselves. While the IMF standby arrangement lasted it pumped in about $2m
weekly into the auction market. Between June 1981 and December 1984 the exchange rate
moved from Ush76.75 per $1 to UshS51.57 per $1, a depreciation rate of about 86%.
At the onset of floating exchange market in 1982 there were two official exchange
rate windows. Window 1was where the exchange rate was administravely fixed, and the rate
obtainable here was used to fmance essential imports, such as, crude oil and petroleum
products; it was also used for external loan repayments. Window 2 was the auction system
where the rate determined by market forces prevailed. By the end of 1984 the two windows
" were merged to produce one door.
By June 1985 the exchange rate had reached Ush.600 per $1. By October of the same
year the government went back to operating two windows once again, on one window, the
exchange rate was determined by administrative fiat on the other it floated independently.
By June 1986 the exchange rate on the first window was Ush1,400 to $1 and that on the
second window was UshS,OOOto $1. According to Dr. Kiggundu', the dual exchange rate
system was problematic in one great respect, and that is, a lot of "non-priority" imports
found their way into the first window of the official exchange rate market. Thus the BOU
lost money. By May· 1987 there was about Ush850 billion in circulation. So a currency
UNIVERSITY OF IBADAN LIBRARY
54
reform was put in place and a new Ugandan Shilling was introduced. The old Ugandan
shilling exchanged for the new one at the rate of 100 old shillings to 1 new shilling. Thus
money supply came down to Ush.8.5 billion. To impact prices a 30% conversion tax was
introduced so that the money supply came down to new Ush.5.95 billion, then the exchange
rate was fixed at new Ush60 to $1. For two months prices actually fell, but government
expenditure just started to skyrocket. The exchange rate remained at Ush60 : $1 until June
1988.
By June 1988 the World Bank and the IMP had mounted so much pressure on the
monetary authorities calling for a devaluation. So a series of devaluations started in July
1988, and by June 1989 the exchange rate has depreciated to Ush200 per $1. frm Ire
1989 in order to siphon money from circulation the monetary authorities decided to sell
, forex at the rate of Ush 400 : $1 as an interim measure under what was known as a Special
Import Program. Thus while the official rate still remained at Ush 200 per $1, the special
Import Rate was Ush400 to $1. This move helped to siphon about Ush12 billion from
circulation. Once again prices went down. As pressures on the monetary authorities from the
Bretton woods institutions asking for further devaluation continued, the exchange rate was
fixed again at Ush300 per $1. Thereafter, a real exchange rate rule was put in place based
on purchasing power parity framework. This rule required that the monetary authorities
adjust the Ugandan exchange rate in line with the inflation differential between uganda and
her five major trading• partners. Following this rule the exchange rate moved to about
UNIVERSITY OF IBADAN LIBRARY
55
Ush380 per $1 by April 1990. By December 1990 following the real exchange rate rule the
exchange rate stood at some Ush540 per $1.
As an epilogue one might point out that between 1990 and 1991 the exchange rate·
management was no longer based on relative prices, rather the official exchange rate was
negotiated between the IMFfWB on the one hand and the monetary authorities of Uganda
on the other hand, in order to bring the official exchange rate up to the more depreciated
level obtainable at the forex bureaux. By 1991 the bureaux rate was about Ushl,300 per $1
while the official exchange rate was Ush960 per $1. By beginning of 1992 the Ugandan
authorities decided to go back to floating the exchange rate once again through the auction
system. The commercial banks bid on behalf of their customers at the BOU; however, there
is a floor rate (or price) below which the BOU does not allow this authorized dealers to go.
While the official exchange rate has depreciated to Ushl,l00 per $1 there is still a premium
maintained by the bureaux who sell at Ushl,300 per $1.
The return, of the exchange rate from floating to fixed in 1987 was said to be a result
of the thinking of those at the helm of monetary affairs in 1987, that the fixed exchange rate
was a catalyst to rapid growth and development of local industries such as soap and sugar
factories. Under the fixed system, priority industries were put on "Open General License
(ooL) so that they could get foreign exchange automatically (though this allocation system
had the snag that could (and did) result into rents for a privileged few rather than
improving economic efficiency). Another reason for abandoning the floating rate in 1987
UNIVERSITY OF IBADAN LIBRARY
56
was that the floating rate went along with trade liberalization; trade liberalization allowed
foreign goods to flood the domestic market, which not only destroyed local entrepreneurial
incentives but also negated the effects of changing structure of relative prices which floating
regime is supposed to bring about. Further the monetary authorities believed that the
multiplier effects of free trade and liberalized exchange rates were taking place abroad rather
than in the domestic economy, so that while the foreign economies "got the jobs" the
domestic economy "got the goods". It is interesting to note however that some
people have argued that in fact the 1988 and 1989 years witnessed impressive growth of the
GDP. Elbadawi (1990) claimed that the GDP growth rate over this period was 6% annually,
while inflation rate was about 150% per annum during these two years. The unduly high
inflation rate Elbadawi maintained was not due to whatever was happening to the official
exchange rate, rather it was due to government fiscal indiscipline. Some people have argued
that the export-stimulation goal of floating rates has not materialized in Uganda. However
others contest this claim. For instance Obidegwu" maintained that non-traditional exports
have expanded in Uganda growing at about 30% per annum. Such non-traditional exports
according to Obidegwu now bring in some $6Om to $7Om annually. What is more - there
is great potential for growth in this non-traditional exports especially for products such as
vanilla, chilleys and pineapples and even in the traditional exports too such as tea and cotton.
UNIVERSITY OF IBADAN LIBRARY
57
ENDNOTES
1. Mr. Ameyaw, S is the Head of Research Department, Bank of Ghana, the report is
from the interview he granted the Researcher while on data collection trip in Accra
in May 1992.
2. See Endnote 1.
3. So far. this account has benefitted tremendously from an interview with Dr. Louis
Kasekende, The Deputy Director. Research Department Bank of Uganda. The
Interview was granted while this Researcher was on a data-gathering trip to Uganda
in June 1992.
4. See Mbire and Watuwa (1992): "Exchange Rate Policy And Inflation: The Case of
Uganda" (Report Prepared For The African Economic Research Consortium
Workshop May 1992, Nairobi, Kenya).
5. Dr. Kiggundu was the governor of the Bank of Uganda from 1986-1990. This part
of the report is from the interview he granted this Researcher while on a data-
gathering trip to Kampala in June 1992.
6. Dr. C. Obidegwu is the World Bank Resident Economist in Uganda. What is here
reported is from an interview he granted this researcher while she was on a data-
collection trop to Kampala in June 1992.
•
UNIVERSITY OF IBADAN LIBRARY
CHAPTER THREE
LITERATURE REVIEW
"Theory is Inspired By History But Good Theorists
Do Not Always Have Long memories. "
Kenen (1985)
This chapter is devoted to a survey of the several theories of exchange rate
determination, and a survey of the models of the workings of the macroeconomy under
different exchange rate regimes. The chapter also considers the various empirical
investigations of the validity of exchange rate determination theories, and the different
macroeconomic models of the workings of the economy under different exchange regimes.
The earliest theories of exchange rate determination focused on, the current account
balance, price elasticities and propensities. Today focus has shifted to the bond and money
markets as reflected in the monetary approach and the asset market (or portfolio balance)
approach to exchange rate determination. Of the very early theories of exchange rate
determination the most enduring has been the purchasing power parity (Ppp) theory. Others
are the Interest Rate Parity Theory and the Balance of Payments Approach. Below we briefly
review each of these theories in turn .
•
UNIVERSITY OF IBADAN LIBRARY
59
3.1: Theories of Exchange -RateDetennination
3.1.1 Purchasing Power Parity Theory (PPP)
The PPP has its antecedents in the 19th century works of Ricardo, Mill, Goshen and
Marshall (Dornbusch, 1988). In fact Myhnnann (1976) and Einzig (1970) have even traced
the PPP back to eighteenth century Sweden and 16th century Spain. Whatever or whoever
be its forebears the theory had a rebirth during World War I when Gustav Casell (1916,
1918) turned it into a paradigm and christened it.
In essence the PPP maintains that across countries there is commodity arbitrage, such
that prices of commodities are equal across countries when expressed in one given currency.
Thus the exchange rate.is what IIequates the prices of traded goods in alternative currencies."
In its strong form, the PPP makes no allowances for tariffs, transport costs and other
restrictions, hence
Pi = ep', --------------------------- (3.1)
where Pi = Domestic currency price of good i
pt = Foreign currency price of good i
e = Nominal Exchange rate in units of domestic currency per unit
of foreign currency
Summing over all commodities in each economy gives
UNIVERSITY OF IBADAN LIBRARY
60
p*= f(p*.· ..·.··..·..·p*j··.···· ..Pn*). ·.·.. ·..·(3.3)
p = eP* ~ (3.4)
Equation (3.4) implies that
e = P 3.5
p*
Expressing the price ratios in just one currency (say the domestic currency) gives
p = 1 3.6
eP*
In its weak form the PPP doctrine admits of impediments
to trade, and this is represented as a constant, as shown in equation (3.7).
e= P --------------------- (3.7)
P*
,Where e is the constant representing all obstacles to trade.
Several studies have attempted to refute or validate the PPP theory. Among others
Yaeger (1958), Balassa (1964), Harrod (1930) have looked into the issue; the conclusion has
been that as a short-run theory of exchange rate determination, the PPP is inadequate;
however, in the long run, it could provide a "useful though not exact description of exchange
rate behaviour" (Dornbusch 1988 p.265) .
•
UNIVERSITY OF IBADAN LIBRARY
61
3.l.n Interest Rate Parity Theory (lRP)
Traced to the 1923 work of I.M. Keynes - A Treatise on Money, the IRP theory of
exchange rate determination maintains that, a significant force on both the spot and forward
exchange rates, in addition to speculation, is the movement of short-term capital for the
purpose of interest arbitrage. Such movement ensures that a country with a higher domestic
rate of interest relative to her trading partners' will witness continuous inflow of funds as
arbitrageurs try to take advantage of a higher rate of return on their funds. Such movement
of funds continue until the exchange rate in the low-interest country appreciates so much as
to ensure equilibrium in the capital market, whereby returns from both domestic and foreign
investment are equal. (Kindleberger, 1939; Jasay, 1958; Austen, 1961; and Sohmen, 1965).
Given that one unit of domestic currency will with a domestic interest rate of idgrow to
i
eell by time 't' (where e is the basis of the natural logarithm) when invested at home; then
if capital movements between countries are unrestricted, funds can alternatively be
transferred abroad. Given spot and forward exchange rates of ro and rt respectively, and a
foreign rate of interest of if, and assuming initially that transfer is costless, covered interest
arbitrage will yield
ift
(rJrJe units of domestic currency at time t for every
units invested abroad. If 0 is the expected percentage change in the spot exchange rate '1'
period from now,
•
UNIVERSITY OF IBADAN LIBRARY
62
Le
f> = rl - ro, :.(3.8)
ro.t
then, in equilibrium returns from investment at home and broad would be equal,
t(id-i
= r
)
rJro 1+tf> = e (3.9)
Equation (3.9) can be reduced to
f> = id - i, (3.1O)
Thus the spot exchange rate 't' period from now (ro.t), is determined by the interest
differential between the domestic economy (iJ and the foreign country 4 (Sohmen, 1965).
The underlying assumption in the IRP theory is that there are no capital controls, so
that capital is perfectly mobile and securities denominated in different currencies are perfect
'Substitutes. Transfer of funds across national boundaries is costless. These assumptions
seriously limit the usefulness of IRP as an adequate explanation of the behaviour of the
exchange rate in the short, medium and long run (lsard, 1987).
3.1.ill: Balance of Payments Approach - (BOP)
The BOP approach to exchange rate determination maintains that the exchange rate
between two currencies is determined by the demand for and supply of each country's
currency brought about by the need to make payments to each other for goods and services
UNIVERSITY OF IBADAN LIBRARY
63
exchanged (Wilby, 1981). The country that demands more of its trading partners goods and
services than the trading partner does of its own, drives the value of the foreign currency up
and that of the domestic currency down (i.e. foreign currency appreciates while domestic
currency depreciates). In the process of settling the payments and receipts for imports and
exports the exchange rate is determined.
There are two variants of the BOP approach to exchange rate determination. The first
variant emphasizes changes in the terms of trade as being the major determinant of shifts in
the balance of payments and hence of the determination of equilibrium level of the
exchange rate. This variant analyzes the importance of elasticities of demand for imports and
exports, as well as the elasticities of export supply, as being critical in determining
"desirable" terms of trade. Proponents of this variant include Machlup (1956) and Harberler
(1949).
,
The other variant of the BOP approach to exchange rate determination emphasizes
the importance of the relationship of domestic output to domestic absorption in the
determination of the current account balance and hence of exchange rate movement. This
variant maintains that expenditure reduction in the face of trade deficit is crucial to restoring
external balance. Thus, in order to avoid inflationary pressures in the face of trade account
deficit, an economy must reduce the level of its domestic absorption to that of domestic
output. Among others Alexander (1952), was one of the proponents of this variant.
The two variants of the BOP approach were however merged by Alexander (1959)
UNIVERSITY OF IBADAN LIBRARY
64
Tsiang (1961) and others. They showed that for both internal and external balance the
process of adjusting the balance of payments and hence of determining equilibrium exchange
rates must necessarily involve both expenditure switching (i.e. switching domestic
expenditures away from imports to home produced goods) and expenditure reduction in a
country running trade deficits.
The BOP approach has being criticized for narrowly focusing on the international
transactions in goods and services while overlooking international transactions in financial
assets which are included in the capita) account.
3.1.IV The Monetary Approach (MP)
The monetary approach to exchange rate determination contends that demand for foreign
exchange arises as a result of attempts to equilibrate the net stock demands for monies of
different countries. It asswnes that a country's price level is pegged to the world price level
and must move rigidly with it - in other words that the law of one price holds (Wilby, 1981;
Boughton, 1988).
Thus the monetary approach posits that disparities in nominal money supply relative to
real demand for money cause shifts in prices which disturb the maintenance of purchasing
power parity. Attempts to bring about equilibriwn in the money market therefore leads to
changes in the exchange rate. The exchange rate is fmally decided when the nominal money
•
supply has equalled money demand and parity has been restored in the two trading countries.
UNIVERSITY OF IBADAN LIBRARY
65
The real demand for money it should be noted, is however, a function of real income as well
as of nominal interest rates. Hence the monetary approach equations are essentially four
e= p-i>· .3.11
p = m' - ley + hi 3.12
P• = m • - ley. + hi' 3 13
e = m - m" + h(i - i*) - key - y'). .. 3.14
where • indicates the foreign country's variables.
e = nominal exchange rate
m = the nominal money supply
y = real income
i = nominal interest rate
e = long-run nominal exchange rate
k, h = adjustment operators.
(All variables are in small letters indicating that they
are in logarithm fonn.)
Bearing in mind that the real demand for money is a
function of real income, y, and nominal interest 'I';
equation 3.14 is the exchange rate equation, and it maintains that relative changes in money
supply, in interest rates and in real income affect the exchange rate. Thus an increase in
money supply not mate ed by an equiproportionate increase in money demand will all things
UNIVERSITY OF IBADAN LIBRARY
66
being equal lead to an equiproportionate increase in the exchange rate (i.e. exchange rate
depreciation). Increase in domestic real income however leads to a fall in domestic prices
and hence to an appreciation of the exchange rate. Higher interest rates relative to a
country's trading partners' result in excess money supply as real demand for money falls
thus bringing about exchange depreciation.
There are empirical and theoretical weaknesses inherent in this approach. First, according
to Dornbusch (1980) there is ample evidence that the assumption of PPP is not warranted,
••... not only does the short-term exchange rate deviate from a PPP path but there are also
cwnulative deviations from that path that show substantial persistence". Using the case of
the United States of America, 1973 - 1979 as an example, Isard, (1987) maintained that, in
spite of the USA running an inflationary rate that was at least one percentage point less than
, those of her five major trading partners, and as such ought to have had a real exchange
appreciation of about one percentage point over the period 1973 - 79, if PPP holds, the USA
rather had a real exchange depreciation of over 2 percentage points over the period. Other
econometric evidence reveal that the monetary approach is an unsatisfactory theory of
exchange rate determination (lsard, 1987) Essentially the monetary approach is a partial
equilibrium approach hinging on the premise that exchange rates are determined by the
conditions of stock equilibrium in the asset markets.
UNIVERSITY OF IBADAN LIBRARY
67
3.1.V Portfolio Balance Theory of Exchange Rate Detennination
This approach is also known as Asset market approach and has a long list of
advocates (Branson, 1977; Boyer, 1977; de Macedo, 1978; Allen and Kenen, 1980; and
Adler and Dumas. 1983;) among others.
The portfolio balance theory of exchange rate determination maintains that exchange rate
is determined in the short-run equilibration process of financial markets. given supplies of
domestic and foreign assets. Money is regarded as one of several fmancial assets in an
economy. The exchange rate, though determined in the process of the demand for and supply
of fmancial assets, is itself a principal determinant of the current account in the balance of
payments. Under a system of floating rates the current account is the net rate of
accumulation of foreign assets whose accumulation moves the exchange rate. This gives a
, dynamic system of exchange rate adjustment which includes asset markets. the current
account. and foreign asset accumulation.
The portfolio approach assumes that if there are J income earning assets in an economy
plus one money stock then in all there are (1+1) fmancial assets and for two economies there
are (2J+2) assets. Demand and supply for all these assets determine their rates of return and
hence their prices and since the exchange rate is the price of money, and money is one of the
financial assets in the portfolios then the exchange rate is also determined by the demand for
and supply of the fmancial assets in the two economies.
The weaknesses of tl\is approach are the same as those of previous approaches which it
UNIVERSITY OF IBADAN LIBRARY
68
incorporates. For instance portfolio balance theory assumes the existence of the law of one
price, as well as of interest rate parity. which assumes that securities denominated in
different currencies are perfect substitutes. The existence of purchasing power parity is still
being debated in the literature, while the reality of capital controls make perfect
substitutability of securities of different currencies impossible.
3.2: Review of Models of the Workings of the Macroeconomy
Under Different Exchange Rate Regimes
Irrespective of how exchange rates are determined one issue that has bothered
economists and policy makers alike, since the end of the second world war has been thatof
deciding the "ideal" exchange rate regime for all nations of the world as a whole, in order
,
to ensure domestic internal balance, external balance, and global balance (global balance as
reflected in improvements in global welfare as a whole).
In between the two world wars there was no unique exchange rate system. Prior to the
first world war nations of the world maintained a fixed exchange rate system known as the
"fixed gold standard" During the first world war this system broke down. Criticised for
being too rigid and too dependent, in that no nation could pursue completely autonomous
monetary, fiscal, or other macroeconomic policies without interference from real and
monetary events of the rest of the world, the system could no longer survive when the world
•
was no longer united. After the first world war some nations went to freely fluctuating
UNIVERSITY OF IBADAN LIBRARY
...,
69
exchange rates, others still tried, half-heartedly, the fixed gold standard system but it did not
work. There have been myriads of arguments for and against the fixed as well as the floating
exchange rate regimes. Different models were developed which tried to highlight the
performance of an open economy under one or the other of those two exchange rate
arrangements. Among the earliest models in this area was the Mundell-Fleming model
(MFM) (Mundell, 1961, 1962, 1968, Fleming, 1962). The essence of the Mundell - Fleming
model was to show that under a system of floating exchange rates monetary policy tends to
be quite effective in improving economic performance. Under a system of fixed exchange
rate regime however it is only fiscal policy that can be relied upon to uplift the economy.
Dernberg (1970) however hypothesized that the efficacy of both monetary and fiscal
policy does not depend on the type of exchange rate system in use. He went on to show that
given the right policy mix, effects of monetary and fiscal policies could be identical under
,
the different exchange rate regimes. This study attempts to investigate whether the Demberg
hypothesis is valid or not. Below is a review of a few of the models of the workings of the
macroeconomy under different exchange rate regimes.
3.2.1. The Insular-Economy-Model
Prior to the Mundell-Fleming-Model was the model termed by Kenen (1985) the Insular-
Economy-Model. The Insular-Economy-Model essentially derives from the Keynesian
income determination odel which was basically a closed economy model. However the
UNIVERSITY OF IBADAN LIBRARY
70
closed economy was opened in the works of Machlup (1943) Metzler (1943) and Meade
(1951) to produce the Insular-Economy-Model. This model essentially adopts a balance of
payments approach to exchange rate determination. It essumes away the capital account and
concerns itself with the trade and current account balances only. The model assumes that
wages are fixed over the period relevant to policy. Money supply a potent policy instrument
is assumed to be exogenous. All balance of payments flows are sterilized by government
policy actions or open market sales and purchases of securities. Other than money, there is
only one asset in the financial market which is treated implicitly,
In the insular-economy-model total demand for the output Y, of the economy is made
up of private domestic expenditure Z, government expenditure G, and the balance in the
import - export trade or trade balance, B.
Y = Z + G + B (3.15)
The private domestic expenditure Z, is itself a function of
domestic income Y, the nominal rate interest r, as well as real wealth A{P,
Letting A{P = a, then
Z = Z(Y,r,a) (3.16)
O<Zy<l
Zr<O
Za>O
The trade balance, B, depends on domestic private expenditure, Z, foreign expenditure Z*
UNIVERSITY OF IBADAN LIBRARY
71
and the terms of trade P/EP* where P is the domestic price level, p* the foreign price level
and E the nominal exchange rate in terms of units of domestic currency per unit of foreign
currency
Letting P/EP· = t, then
B = B(Z, Z·, t) (3.17)
-1<Bz<O
Bz* >0
Bt<O
The trade balance is assumed to be negatively related to the terms of trade, B, < 0, so that
devaluation will lead to reduced volume and value of exports (Marshall - Lerner - Robinson
condition).
Output supply is assumed to be inversely related to the real wage rate. At lower wage
" rates producers employ more resources and output rises.
y = a (PfW) (3.18)
In the monetary sector of the insular economy model, money supply is exogenously
determined. Demand for real money balances is however a function of domestic income,
Y, nominal rate of interest, r, as well as real wealth, •a', (or AlP)
M/P = M(y, r, a) (3.19)
M'y>O
M'r<O •
UNIVERSITY OF IBADAN LIBRARY
72
M'a~ 1
Finally foreign exchange can only be accumulated in this economy via the trade
balance. given that there is no capital account. hence
EF = B (3.20)
where F is the foreign exchange reserve of the economy. Given equations (3.15) to (3.20)
the economy's performance will vary depending on the exchange rate regime in place.
A: Flexible Exchange Regime
Under a regime of flexible rates equation (3.20) determines the exchange rate. Given
that there is no capital account. the balance of payments consists only of the trade balance
and the exchange rate must adjust to keep this balance at zero level i.e.
B(Z. Z*. t) = 0 (3.21)
, Once the trade balance is zero then equation (3.15) reduces to
Y = Z + G (3.22)
This implies that once the trade balance is zero then all other macroeconomic aggregates
become independent of the exchange rate. Expansionary fiscal policy given a constant
monetary policy stance will lead to increased output and income
dY 1
--- - --- > 0 (3.23)
dG l-Zy
A similar result is got from varying monetary policy while keeping fiscal policy
constant. •
UNIVERSITY OF IBADAN LIBRARY
73
B: Fixed Exchange Regime
Under a regime of fixed rates there is a leakage into imports from increased domestic
expenditures so that expansionary fiscal or monetary policy has an effect on output, that is
less than what can be had under a regime of flexible rates. Expansionary fiscal policy while
monetary policy is constant will from equation (3.15) result in
dY 1
--- - ------------ (3.24)
dO (l-Z'y)(1+B'z)
and
1 1
< ------------- ....................(.3..24B)
(l-Z'y) (1+B'z)(1-Z'y)
r
Further a rise or fall in foreign output and price also impact on domestic output via the trade
balance.
Equation (3.24B) implies that the expansionary effect of fiscal policy under fixed
rates is less than under floating rates. Since a similar result is got even for expansionary
monetary policy then the conclusion of the insular - economy - model is that an economy
performs better under a system of flexible rates than under a system of fixed rates provided,
there are no capital flows.
3.2.11: Mundell - Heming Model (MFM)
Described as the "workhorse of traditional open economy macroeconomics (Frenkel
UNIVERSITY OF IBADAN LIBRARY
74
and Razin, 1987), this model has been extended in several different directions over the years.
There has been the portfolio specification of capital mobility in the traditional MFM by
McKinon (1969), Floyd (1969), Frenkel and Rodriguez (1975). There has also been the
introduction of debt - revaluation effects into the main MFM by Boyer (1977) Rodriguez
(1979), and the introduction of expectations and its effects on the performance of the
macroeconomy under different exchange regimes (Dornbusch, 1976; Kouri, 1976; Mussa,
1976).
The Mundell - Fleming model with some adaptations remain the backbone of
macroeconomic models of the exchange rate. The assumptions of this model include:
(j) Fixed domestic prices, which by implication means that the exchange rate is equal
to the terms of trade
(ii) Full or perfect capital mobility, which thereby allows for assumption (iii)
(Hi) Perfect substitutability of domestic securities with foreign securities
(iv) Interest rate parity existence, given that arbitrageurs are assumed to ensure that
interest rate differentials across countries are quickly eliminated.
(v) Real income or output is demand determined.
Under this system an increase in money supply not backed by an equiproportionate
increase in money demand leads to falling interest rates which result in incipient capital
outflow. As capital flows out the domestic nominal exchange rate depreciates, and assuming
elasticities condition: are satisfied demand will shift towards home produced goods in the
UNIVERSITY OF IBADAN LIBRARY
75
face of a depreciating nominal exchange rate. The induced increase in output leads to a rise
in income and hence in the demand for money. The rise in money demand continues until
it equals the money supply and the rate of interest rises until it equals the foreign rate of
interest. In the end the economy settles at a higher level of output but a lower real exchange
rate (i.e. a depreciated real exchange rate).
An expansion in demand for home output as a result of fiscal policy stance such as an
increase in government expenditure or an exogenous shift in demand, leads to an increase
in income which is accompanied by an increase in money demand and hence a rise in
interest rates. As interest rates rise there follows incipient capital inflow which result in
appreciation of the nominal exchange rate. As the nominal exchange rate appreciates the
country loses its competitiveness hence a deterioration will occur in the current account
balance and this dampens or offsets the expansion.
The MFM has the following system of equations. First is the income equation
wherein national income Y, is the sum of private domestic expenditure X, government
expenditure S, and the trade balance or net exports, B.
Y=X+S+B .............. (3.25)
Total expenditure Z, is the sum of private plus government expenditures
Z = X + S (3.26)
Given the money stock M, then income velocity, V, in the economy is
V = YfM • (3.27)
UNIVERSITY OF IBADAN LIBRARY
76
While private disposable income Ntis
N=Y-T ... '.... (3.28)
where T is the lump-sum tax.and is itself a function of the level of income
T = TOO ...... (3.29)
Private expenditure X, then depends on the disposable income N, and the rate of interest, R.
x = X(Nt R) ................. (3.30)
O<~< 1
Xr<O
The rate of interest, R is a function of the income velocity of money
R = R(V) (3.31)
~>O
While the trade balance depends on total expenditure, Z, and the nominal exchange rate Ft
where the exchange rate is in terms of units of domestic currency per units of foreign
currency
B = B(Zt F) (3.32)
0< -s, < 1
Bf>O
Finally net capital inflow is a function of the rate of interest
C = C(r) ....... (3.33)
Given the set of eq&tions (3.15) to (3.33) therefore, the economy behaves in the following
UNIVERSITY OF IBADAN LIBRARY
77
ways under different exchange regimes.
Al:Flexible Exchange Regime
Under a flexible exchange rate system increased government expenditure
representing expansionary fiscal policy in the face of a constant monetary policy will lead
to increased output as shown below:
1
dY = ------------------------------- > 0 (3.34)
ds 10 (l-Xn)(l-Ty)-(Xr-Cr)Rv/m
where the subscripts 10 after dy/ds implies that the exchange rate system is flexible while
the monetary policy is constant (or fixed). The increased output will however be dampened
as increased capital inflows in response to rising interest rate, appreciates the exchange rate
thus reducing the economy's competitiveness and resulting in a worsening current account.
~
ds 10 -----> Otas Cr ---> 00
An increase in money supply while fiscal policy remains constant results in reduced
income velocity of money which implies reduction in the rate of interest. The reduced rate
of interest will generate increased domestic investment expenditure but it also leads to
increased capital outflow, which depreciates the nominal exchange rate thus helping to keep
external transaction•s in balance and the increased output and income arising from increased
UNIVERSITY OF IBADAN LIBRARY
78
investment expenditure is sustained
RvY(Cr-Xr) 1·
~ = ------------ ---------------------- > 0..(3. 35)
elm 11 ~ l-Xn(1-Ty)+Rv(Cr-Xr)
M2
where the subscripts 11 after dyjdrn implies that the exchange system is flexible and
monetary policy is also varying
B 1: Fixed Exchange Regime
Expansionary fiscal policy in the form of increased government expenditure will lead
to increased output and income. The increased income means increased tax revenue,
however this leakage into the tax revenue falls considerably short of the increase in output
I+Bz
~ = > 0 ... (3.36)
dsOO 1-(1+Bz)[Xn(1 +Ty)] +XrRv
M-
where the subscripts 00 after dy/ds imply that the exchange regime is fixed and monetary
policy is also fixed (or constant).
When monetary policy is employed, the rise in money supply reduces income
velocity thereby reducing rate of interest. The reduced rate of interest generate increased
expenditure, a lot of which leaks into imports thus leading to a deterioration of the trade
balance and dampening the output expansion. The increased capital outflow arising from
decreased rate of interest is not allowed to depreciate the nominal exchange rate as would
•
UNIVERSITY OF IBADAN LIBRARY
79
have happened under flexible rates as the government is committed to keeping the exchange
rate at a fixed level. Hence the effect of an expansionary monetary policy under fixed rate
is
-XrRvY 1
> 0 ..(3.37)
01 ~ 1 XrRv
Sz+-I- Xn(1-Ty~-----
It is important to note that equation (3.37) is less than equation (3.35) implying that
the rise in output and income as a result of expansionary monetary policy is less under fixed
rates than under floating ones. While equation (3.36) is greater than equation (3.34)
implying that the rise in output and income with an expansionary fiscal policy is higher
under fixed than under floating rates, provided capital is perfectly mobile.
The conclusion' of the MFM therefore is that floating rates are ideal for uplifting the
economy if monetary instnunents are employed, while fixed rates are more effective if fiscal
..•
instruments are employed (Fleming, 1962).
The weaknesses of the MFM are many. First it assumes that purchasing power parity
holds, and there is no conclusive evidence that purchasing power parity holds either in the
short or long run (Marwah and Bodkin, 1984; Balassa, 1964). Second the MFM also
assumes interest rate parity, which by implication means perfect mobility of capital. That,
there is interest rate parity, is, itself, a subject of controversy (Isard, 1987) .
•
UNIVERSITY OF IBADAN LIBRARY
80
3.2.III. The Modified Mundell-Fleming Model (MMFM)
This is an eclectic approach to the several modification of the Mundel- Fleming
model in the manner of Frenkel and Razin (1987). The modifications introduced into the
traditional MFM are in three respects. (i) The modified Mundel - Fleming model assumes
continuous full-employment (ii) It assumes complete price flexibility and (Hi) It explicitly
recognises the budget constraints that influence government policies.
In the MMFM,production of output, y, in an economy, gives rise to income from labour
yO and income from
capital v
y = yO +v ; (3.38)
The disposable income Z, is a function of labour income, yo, permanent real income
streams, 'a', and a lump sum tax T
Z= yo + a- T (3.39).
Asset preferences between money and equities are characterised by the demand for real
balances L. The demand for real balances L, is a fraction d, of the real wealth, W, and what
this fraction d, will be, depends on the rate of interest, r .
L = d (r) W (3.40)
Real wealth, W, is the sum of real money balances, M, and real equity holdings, aIr
W = M + 8/r (3.41)
Saving is proportional to the excess of target wealth, W over actual wealth, W
UNIVERSITY OF IBADAN LIBRARY
81
S='AcW-W)
Target wealth is itself proportional to labour income, yO,less lump sum tax, T
VI = 6(r) (y°_T) (3.43)
from equation «3.40) and (3.41) real money balances M, equals
M=~ (3.44)
1-5
so that actual wealth becomes
W = air (3.45)
1-5
substituting equations (3.43) and (3.45) into the savings function,(3.42) gives
.
S = 'A <b(r)(Y°-T)- air (3.46)
1-5
Equation (3.46) essentially implies that saving is a function of the rate of interest, permanent
real income streams, and the tax level. If government is operating a balanced budget then
total government expenditure 0, should be equal to the sum of its tax revenue T and its
holding of real income streams
o = T+g (3.47)
If government is a debtor then g is negative.
If purchasing power parity holds then the general price level in the domestic economy P,
should be the same 1s that in the foreign economy p. in domestic currency terms ..
UNIVERSITY OF IBADAN LIBRARY
82
P = eP* (3.48)
Where e is the nominal exchange rate. The trade balance surplus B, would be the excess of
domestic output over domestic absorption
B = Y - (Z-S)-G) (3.49)
Where (Z-S) is private expenditure. The current account surplus C, is the trade balance
surplus plus the service account surplus
C = B + (a + g - v) (3.50)
B + (a + g - v) = S + (T + g - G) (3.51)
Equation (3.51) maintains that the current account surplus is also equal to the rate of saving
plus the government budget surplus. Since under flexible rates the current account surplus
equals capital account.deficit, then the rate of capital outflow is fully specified by the saving
function, equation (3.46)
... The world supply of securities Q is equal to the number of claims to income from
capital in each country, Y and v: respectively, less their respective governments' holdings
of claims g and g*. It is assumed that the foreign country has similar behavioural equations
Q Ii Y + v: -g - g" (3.52)
For equilibrium in the short and long run, both under floating and fixed exchange regimes
the following conditions must hold
UNIVERSITY OF IBADAN LIBRARY
83
Short Run Equilibrium Conditions
The short run equilibrium conditions are given by
M = e> (afr) r: (3.53)
I="e>*
..-
M = e> (a*/r) (3.54)
I="e>*
Equation (3.53) and (3.54) maintain that actual money balance held in each country depend
both on the domestic and the foreign proportion of real wealth held 0 and the 0*respecti vely
and also on the rate of interest (r) assumed to be the same in both countries given interest
parity.
Total world saving will therefore be
Sfr, w) + S* (r, w") = 0 (3.55)
Long Run Equilibrium Condition
(l-e»W + (l-e>*)W* = Qjr (3.56)
Equation (3.56) implies that in the long run whether a country adopts a floating or a fixed
exchange rate regime there can not be any autonomy in a country's use of monetary and
fiscal policies because capital mobility implies that all policies have repercussions
throughout the world and floating rates not withstanding inflation rates are interdependent.
This conclusion. it should be noted removes the bottom from the arguments of the advoca tes
•
UNIVERSITY OF IBADAN LIBRARY
84
of floating exchange rates that its main advantage lies in its insulating an economy from
disturbances of variations in-foreign monetary and fiscal polices.
3.3 Review of Empirical Literature
There have been several empirical studies attempting to find out the performance of
the macroeconomy under different exchange rate regimes.
Rhomberg (1964) looked at the Canadian economy under fixed and under floating
exchange rate systems, Rhomberg's study was based on the Mundell - Fleming model.
Canada had in the period 1950 - 1962 adopted a floating exchange rate system. Prior to 1950
Canada maintained a fixed exchange rate system of the Bretton woods variety. By May
1962, Canada went back from a floating to a fixed exchange system. Rhomberg maintained
that this continuous shift of exchange system by Canada is a reflection of the "difficulties
,
encountered by an internationally dependent economy in its attempt to pursue independent
monetary and fiscal policies"
Rhomberg built a model of the Canadian economy with three variants. The first
version was an economy with freely fluctuating exchange rates, the second was an economy
with a fixed exchange rate system but where changes in the balance of payments are
neutralized and not allowed to influence the money supply. The third is a situation of fixed
exchange rates but where changes in the balance of payments impact on the money supply,
thus making the money supply an endogenous variable.
UNIVERSITY OF IBADAN LIBRARY
85
Employing seventeen behavioural equations and using the Limited Information Method,
Rhomberg tried to determine the behaviour of real income, balance of payments, domestic
prices, exports and imports under fixed and under floating exchange regimes. Rhomberg
divided the economy into three sectors, the real sector made up of the consumption function
the investment function and the income equation; the foreign trade sector, which was a
sector hinging mainly on the balance of payments identity and being linked to the real sector
through import demand and export supply equations. The third sector was the financial
sector where interest rates are determined. The financial sector is linked to the real sector
through the effects of long-term interest rates on investments and linked to the foreign trade
sector through the effect of changes in interest rates on capital movement.
Rhomberg found that what the response of macroaggregates would be under any
, exchange rate regime would depend on the structure of the economy. Mainly
macroeconomic performance under any exchange rate regime would depend on the marginal
ratio of imports to gross domestic product, the ratio of real demand for money to gross
domestic product and the speculative balance relative to the rate of interest. Rhomberg
hypothesized that if responsiveness of capital movements to interest rate differential exceed
a certain threshold - value, then the amount of capital attracted as a consequence of
budgetary expansion with a constant money supply will exceed the induced increase in
imports and therefore result in a gain in reserves (rather than a loss, under a fixed regime,
as Mundell-Fleming •Model postulated) and in currency appreciation (rather than
UNIVERSITY OF IBADAN LIBRARY
86
depreciation. under a flexible regime. as MFM maintained). This part of Rhomberg's results
contrasts with the Mundell Flemming one given the "threshold condition" in-built by
Rhomberg. He concluded that by 1960 the Canadian economy had reached such threshold
and hence the observed appreciation of the Canadian dollar in spite of a floating exchange
rate system and continuous fiscal expansion.
Rhomberg found the price elasticity of imports to be -1 and the interest elasticity of
money demand to be -0.2 under a floating exchange rate system. On the whole he concluded
that in the absence of the "threshold condition". increasing government expenditure as a
means of raising output and employment was more effective under fixed than under floating
rates. while increasing money supply to increase output and employment was more effective
under floating than under fixed rates. These conclusions are similar to those of Munde 1and
Fleming.
One interesting conclusion of Rhomebrg's however. is that a country operating a
fixed exchange rate system can enjoy "financial insulation" if it wishes. Financial insulation
is a situation whereby variations in capital movements do not affect income and employment
because government neutralizes completely the effects of such capital movements on the
money supply. Under a floating exchange rate regime. Rhomberg maintained. financial
insulation is not possible.
Tullio (1981) also studied the performance of the Italian economy under floating and
under fixed exchange rate regimes. In the 1970s. the Italian economy was characterized by
UNIVERSITY OF IBADAN LIBRARY
87
high inflation rate, large depreciations of the exchange rate, unemployment and unstable
growth pattern of real GDP. Tullio termed these features "vicious-circle features". He
therefore attempted to find out the role of discrete exchange rate changes in Italy and their
effects on prices, wages, trade flows and growth of real GDP, under specific policy rules.
He also wanted to fmd out the role of fiscal and monetary policy under different exchange
rate regimes. Tullio's model was along the modified Mundell-Fleming framework.
Tullio employed a 29-equation model of the Italian economy specified in continuous
time, and estimated with quarterly data for the third quarter of 1960 to the fourth quarter of
1978, using Wymer's (1976) program. Tullio's simulations of fiscal and monetary policies
were performed separately under the assumption of flexible rates and of fixed rates to gain
useful insights into the.efficacy of monetary and fiscal policies in stimulating real growth
under the two exchange rate regimes. Tullio estimated two versions of his model, the first
for the period 1960 - 1978 with the exchange rate assumed to be exogenous (as obtains under
a fixed regime) the second for the period 1973 - 1978 with the exchange rate assumed to be
endogenous (as obtains under a floating system).
Tullio found that a 10% devaluation or depreciation of the exchange rate led to a 3%
rise in domestic prices after one year. Domestic prices then remain at the new level for the
next five years because real output falls as a result of the depreciation, and this fall in real
output checks the growth of domestic prices. Real exports grow by about 2% as a result of
the 10% exchange rate depreciation, while aggregate imports fall by about 2%. The price of
UNIVERSITY OF IBADAN LIBRARY
88
imports in domestic currency grew by the full amount of the devaluation i.e. 10%, while
prices of exports in domestic currency grew but at a lesser speed than import prices. The fall
in imports and the rise in exports led to rising net international reserves. Tullio thus
concluded that the exchange rate is a powerful instrument of external adjustment despite the
fact that it worsens the problem of inflation.
For monetary policy simulation, the money supply was made to grow at an annual rate
of 30% as against the actual rate of 20% implying that a shock was introduced into the
system of 10% growth rate of money supply. It was found that after 10 years the price level
had risen by 10% under fixed rates and by 43% under floating rates. The different long - run
effect on prices under the two systems was said to be due to the higher rate of growth of
import prices under flexible rates. A 10% once and for all increase in government
,
expenditure had the effect of raising domestic prices; this rise was as high as 15% by the
10th year under floating regime and by only 7% in ten years under fixed rates. Also the
GDP rose more, under fixed, than under floating rates. Similarly the result of his fiscal
policy simulation showed the economy to have performed better under a fixed exchange
regime than under a floating one.
Tullio thus reached the conclusion that the economy performed better under fixed
than under floating rates irrespective of the policy instrument used (monetary or fiscal). He
however, added that t•he poorer performance under a floating system may not be due to th
exchange rate per - se, but it did occur that the economy performed poorer during the era of
UNIVERSITY OF IBADAN LIBRARY
89
floating, than fixed, rates.
Bruno and Sussman· (1979) also studied the economy of Israel under different
exchange rate regimes with a view to ftnding the effects on the current account and the rate
of inflation of large devaluations (under ftxed exchange rate regimes) and cumulative small
ones (under floating regimes). They came to the conclusion that large devaluations impact
more on the current account and on inflation, because of the awareness on both the p rt of
decision makers and other economic agents that certain goals must be achieved as a result
of the devaluation. Whereas the small frequent changes may not have the desired effects
because of the illusion that government and major decision makers are now free of
employing ftscal and monetary policy to achieve the targets of external balance and stable
prices.
Otani and Park (1976) built a model of the Korean economy wherein the money
stock and real income were endogenized in what otherwise was a monetary model. They
were interested in finding out the interplay between the monetary sector and the real sector
of the Korean economy. They tested the validity of their model against data for the period
1962-1974 - using quarterly data. Their study tried to examine the effects of alternative
policy instruments - exchange rates, monetary policy and ftscal policy - on the
macroeconomic aggregates of output, prices and the balance of payments.
In the model the economy was divided into three sectors - monetary sector, real
sector (where output supply could be primary or non-primary) and the foreign trade sector.
UNIVERSITY OF IBADAN LIBRARY
90
In the foreign trade sector import demand was assumed to consist mainly of demand for
intermediate goods and raw materials rather than demand for consumption goods. Primary
output referred to agricultural products. and its supply was assumed to be exogenous. while
non-primary output supply was influenced by the price level. real wages. rate of capacity
utilisation. and import prices.
Using the two - stage least squares estimator. the results of the estimation of the
model included the following:
(i) the current rate of inflation is heavily dependent on past rates of inflation
rather than on current variables such as current rate of interest. current real
income. etc.
(ii) the demand function for intermediate-goods imports was elastic with respect
to real output in the non-primary sector. while the supply of real output in
this sector was found to be
inelastic with respect to changes in the prices of imported inputs. TIlls shows that
the substitution between intermediate imports and domestic inputs is very low.
(iii) The simulation exercise revealed that a change in exchange rate policy has
considerable influence on the general price level. real income. volume of
intermediate imports and the money multiplier; thus finding exchange rate
to be a powerfulj trument in influencing the level of economic activity.
Khan and Knight (1981) tried to find out the qualitative effects of stabilization or
UNIVERSITY OF IBADAN LIBRARY
91
adjustment policies on the macroeconomic aggregates of output, balance of payments and
inflation in developing countries. They pooled a time-series cross-sectional data of 29
developing countries of which Ghana and Kenya were included. Using 8 annual observations
on each variable for each country, they used the Full Information Maximum Likelihood
(FIML) estimator that allows nonlinear constraints to be placed on parameters both within
and across equations.
They used the change in the stock of international reserves as a proxy for the balance
of payments, and they did not disaggregate the balance of payments into its current and
capital account balances. The balance of payments, they assumed, is a positive function of
the excess demand for nominal money balances, and a negative function of the deviation of
domestic price level from its purchasing power parity equilibrium.
A log ~- Alog 1;= al log Mdt-log ~_I
- ~ log Pt_I-Iog(ECt-I)P·f(t-,J-Bo ... (3.57)
where
Alog R, = Change in log of international reserves
Alog 1;= Change in log of nominal exchange rate
(log Mdt - log ~_I) = log of Excess demand for money
[log Pt_1- log Et_, Pft_l].= Purchasing Power Parity deviation.
The rate of infla ion was taken to be a positive function of excess supply of real
UNIVERSITY OF IBADAN LIBRARY
92
money balances· and a negative function of the deviation of domestic prices from their
purchasing power parity level.
Khan and Knight came up with the following conclusions.
(i) That excess supply of real money balances results in an increase in the rate of
inflation, the excess money supply elasticity inflationary rate is 0.33 . This parameter
was significant at 1 percent level.
(ii) Income elasticity of demand for money was found to be 1.2, a result similar .
to the one generally obtained for developing countries.
(iii) The elasticity of the balance of payments to changes in money supply was found to
be 0.49.
(iv) The output elasticity of excess money supply was found to be 0.043, and this was
significant at the 1% level.
In their sensitivity analysis a 10% increase in money supply led to an increase in
domestic price of about 5.5% after 4 years, under a fixed exchange rate system. The inflation
induced by the money supply shock kept the country's exchange rate overvalued for quire
some time. Eight years after the monetary expansion, the price level was still some 2.5%
above its pre-shock level. The output effect was smaller, output having only increased by
III of 1% in the second year after the expansion. In all Khan and Knight concluded that
stabilization programs designed to achieve quick results on the balance of payments via
sharp deflation are likely to have significant and undesirable impact on output employment,
UNIVERSITY OF IBADAN LIBRARY
93
and factor incomes, particularly in the short run.
Olofm, Akinkugbe and Ajayi (1986) in an ex-ante simulation analysis of the impact
of devaluation on the Nigerian economy employed a model with 104 stochastic equations,
and 58 exogenous variables. Known as CEAR-MAC-IV model, the model was estimated
using annual data for the years 1961-1981. Their study tried to anticipate the possible effects
of a floating exchange rate system with its attendant exchange rate depreciation on the
Nigerian economy. Thus, utilizing different rates of devaluation they tried to see how the
trade balance and the balance of payments would fare. Olofin et al maintained that the use
of devaluation as a means of changing relative price structure and thus improving the current
account as well as the overall balance of payments in a developing deficit economy is not
,unambiguously treated in the economic literature. They claimed that both the elasticities
approach, absorption approach and even monetary approach failed to specify by how much
a country can devalue in order to improve her external position.
They then went ahead to carry out some ex-ante simulations of the effects of different
percentages of devaluation combined with different levels of reduction of government
expenditure. In all they had five scenarios. The first was a situation where all exogenous
variables were following their trend path without any exogenous variable being consciously
adjusted. This scenario provided the baseline solution against which they measured the
results of the four other scenarios. The four other scenarios were those of a 50% devaluation
being accompanied first with a 20% reduction in the government expenditure, and then in
UNIVERSITY OF IBADAN LIBRARY
94
the third scenario being accompanied by a 10% reduction in government expenditure. The
fourth scenario was that of a 60% devaluation being accompanied by a 20% reduction in
government expenditure and in the fourth scenario by a 10% reduction in government
expenditure. They pegged the rate of inflation at 30% and capital expenditure at N4 billion.
Olofin et al came up with the conclusion that all devaluations irrespective of the
percentage decline in expenditure, worsened the trade balance and the balance of payments.
For instance in the first year after the 50% devaluation the balance of trade deteriorated by
about 59% compared to the baseline solution. This poorer performance of the external sector
Olofin et a1 maintained is due to the fact that a bulk of imports mainly machinery and
transport equipment as well as raw materials can hardly be adjusted. Also devaluation and
expenditure reductio little impact on exports which comprised mainly of petroleum,
-whose export volume and foreign denominated export price are both exogenously fixed. The
export value only change in the second year of the simulation (1987) by only .00 1%. Though
import of food and other consumer goods declined, the decline was more than neutralized
by a rise in import bill for capital goods. The level of foreign reserves therefore declined
significantly. In terms of the domestic sector the devaluations and reductions in expendi tures
led to declining growth rates in some sectors of the economy such as mining and quarrying,
building and construction. Overall there was a negative growth rate of GNP.
In another study Olofm, Osagie and Phillips (1985) estimated a model of the Nigerian
economy which they also simulated for different conceivable rates of devaluation. They
UNIVERSITY OF IBADAN LIBRARY
95
found that a 40%. devaluation of the naira in 1986 would move the debt service payments of
Nigeria from N6.8 bill in 1986 to N9.6 bill by 1995, representing an increase in the debt
service ratio from 76.11 % to a paralyzing 97.7% by 1995.
It can be concluded that the results of Olofm et al (1985) and Olofin et al (1986) reveal
that the performance of the Nigerian economy under a situation of decreasing value of the
exchange rates will lead to poorer performance of the economy in terms trade balance,
balance of payments, reserve level and debt repayment burden. It will be fruit~ul and
worthwhile to compare the results of these ex-ante simulations with those which we hope
to carry out in this study, and find out how much of he Olofin et al's results have
materialized. Since the onset of floating rates and the attendant exchange rate depreciation.
Jebuni, Sowa and Tutu (1991) in their study of the Ghanaian economy attempted to
fmd out the relationship between the growth rate of real output and the exchange rate. They
were interested in the real exchange rate rather than nominal. They came to the conclusion
that the exchange rate exerts a positive influence on real output and that it could also exert
a negative influence. They found that exports are positively related to both the real effective
exchange rate and total output. They maintained that since devaluations in Ghana had always
been part of a package, they are usually accompanied by a considerable amount of external
assistance which usually ensured an improved supply of forex, and forex availability exerts
a positive influence on imports and hence on exports and real output.
Chibber and Shaffik (1989) in their study of Ghana reached the conclusion that the
UNIVERSITY OF IBADAN LIBRARY
96
widely held belief that devaluation of the official exchange rate leads to growing domestic
inflation especially in African countries undergoing economic restructuring programmes is
wrong. They contended that there is no direct relationship between the official exchange rate
and inflation, because the level of domestic prices has often adjusted to the parallel market
rates before the official exchange rate devaluation takes place. If anything, Chibber and
Shafik maintained, the continuous and continuing depreciation of the exchange rate under
a floating system has had a positive effect on the budget. The revenue improvement they
contended ensured increased output hence declining domestic prices. The revenue
improvement carne primarily from two channels.
(i) From higher grants and aids disbursed at a more depreciated exchange rate.
(ii) From reduction in subsidies to importers due to elimination of exchange
, over-valuation.
Thus the devaluation did not result in increasing budget deficit. This position of Chibber
and Shaffik shall be explored in this study.
Olopoenia (1986) also, in an ex-ante study of floating exchange rates in Nigeria
found that an equilibrium exchange rate can only emerge in Nigeria by the merger of the
official exchange market with the parallel market given that the exchange premium between
official market rates and the parallel market rate was as high as 160%.
Ajakaiye (1991) iri a study of the effects of exchange rate depreciation on sectoral
prices came to the oonclusion that even though the essence of exchange depreciation
UNIVERSITY OF IBADAN LIBRARY
97
generally is to alter the structure of relative prices in favour of non-tradables, so that the
prices of tradeables rise relative to non tradeables, the industries most greatly impacted by
the depreciation in the Nigerian economy are precisely those in the non-tradeables sector,
so that prices in the nontradeables sector keep rising and thus making the relative price
structure unfavourable towards home goods (or nontradeables) still. Ajakaiye then
maintained that "dichotomizing the effects of exchange rate depreciation on prices in terms
of traded and non-traded goods is irrelevant and misleading in an indecomposable economy
like Nigeria which also depends tremendously on imported inputs". In other words the
purported change in the structure of relative prices that the adoption of floating rate is
supposed to bring and which is supposed to help stimulate exports, stimulate domestic
production and reduce imports may not materialize.
, Other studies on the exchange rate and the macroeconomy include Olisadebe (1991).
Olisadebe studied the Nigerian economy both under fixed and under floating exchange
regimes and came to the conclusion that if one took cognisance of certain limitations
inherent in attempting to decide the impact of exchange rate policy alone on the economy
in the face of several changing policy instruments then one can say that "the exchange rate
has had generally positive impact on the economy". Real output which declined by an
average of 4% in the first years of floating rates (1981-1985) rose by about 4.8% in the first
five years of floating rates (1986-1990), industrial and agricultural sectors' output increased
by 9.5% and 6.1 % respectively between 1986-1990 compared to an average of 2.8% and
UNIVERSITY OF IBADAN LIBRARY
98
1.2% respectively in the last five years pre - floating. The current account balance as a
percentage of GDP recorded an average deficit of 5.6% in the last five years of fixed
exchange rate system (1981-1985) contrasting with an average surplus of 3.4% in 1986-
1990. This reflected the drop in the volume of imports and the higher relative contribution
of non-oil exports to total exports. The contribution of non-oil exports to total exports
averaged 6.0% during 1986-1990 indicating an average increase of 2.8% points over the
level in the last five years of fixed rates (1981-1995). Along with trade and payments
liberalization the floating exchange rate policy has succeeded in eliminating the incidence
of payments arrears. Proportion of export earnings repatriated by exporters increased from
73.7% in 1987 to 98.5% in 1990. Diversification of non-oil exports has proceeded smoothly
with the exportation of hitherto relatively unknown exports such as peugeot cars, asbesto .
.. cement, charcoal and empty bottles. The parallel market premium has been reduced from
160.6% obtainable in the years of fixed exchange rates (1981-1985) to 54.3% in 1986-1990.
This reflects the extent to which the misalignment of the naira exchange rate and distortions
in the forex market have been corrected by the adoption of a floating exchange rate regime
in Nigeria.
In his study Ojo (1990) concluded that since the onset of floating rates there has been a
boost to non-oil exports and to the balance of payments. However the persistent deprecia lion
of the naira exchange rate is a new problem along with the instability of this exchange rate,
both coming in the wake of floating exchange rates. Ojo argued that the main problem lies
UNIVERSITY OF IBADAN LIBRARY
99
with the weak export base in the midst of a strong demand for foreign exchange. Given the
uncertainty with regards to efforts to expand the export base, and the overwhelming debt
burden, Ojo contended that the national development strategies have to be reviewed with a
view to reducing the dependence of the economy, on external trade in the long-run.
The adoption of a floating exchange system has resulted in a general but positive
movement on the balance of payments position and this has had positive effects on the
official forex reserves. The official forex reserves, according to Ojo rose from a level of
N3.3 bill in 1988 to N3.8 bill in 1989 while the reserve level could by 1990 cover 3 months
imports. In this study we hope to find out how much of this supposed improvement in
reserve level is real and how much is illusory.
Elbadawi (1990) had in his study of the Ugandan economy found that the official
exchange rate does not directly impact on the rate of inflation in Uganda. Rather he found
..
that the parallel market exchange rate is the major determinant of domestic prices in the
Ugandan economy. His finding is similar to that of Chibber and Shaffik for the Ghanaian
economy, and even for the Ugandan economy Elbadawi's position has been borne out by the
results of Watuwa and Mbire (1992). Hyuba (1991) has gone a step further to posit that the
high variability observed in the movement of domestic prices in Uganda since the onset of
floating rates is due to the fact that domestic prices are largely influenced by the parallel
market exchange rate a market that is known to be highly volatile. Another key factor of the
transmission of inflationary process in Uganda is governments fiscal deficit - Elbadawi
UNIVERSITY OF IBADAN LIBRARY
100
(1990), Watuwa and Mbire (1992).
3.4 SUMMARY
Unlike Rhomberg (1964), Tullio (1981) in his study of the Italian economy under fixed
and floating rates found that both monetary and fiscal policies perform better under a system
of fixed rates than under floating one. Though he was quick to add that the performance may
not necessarily be because of the adoption of floating rates, per se, however the fact remains
that such poorer performance of macroeconomic aggregates was observed under a floating
regime rather than under a fixed one. In other words, Tullio admitted that some other
variables might have accounted for the poor performance of the Italian economy under
floating rates other than the adoption of floating rates.
In their ex ante analysis of devaluations in Nigeria, Olofin, Osagie and Phillips' (1985)
fmdings imply that in terms of debt repayment burden and external sector's aggregates the
economy would be worse off under floating than under fIXed rates.
Further Chibber and Shafik had concluded in their study of the Ghanaian economy that
the accelerated inflationary rate observed post floating has been primarily a "monetary
phenomenon, reflecting as well underlying weakness in the financial system which need to
be tackled to sustain the reforms". In this study we would like to find out the response of
prices to the exchange rate and to monetary movements and fmd out which one plays the
more dominant role .
•
UNIVERSITY OF IBADAN LIBRARY
101
ENDNOTES.
1. Kenen. P.B. (1985) "Macroeconomic Theory and Policy: How the Closed Economy
was Opened" in Handbook of International Economics (North Holland Publishers)
Chapter 13.
2. See Chapter One, Section 1. above.
3. See Balsaas. B. (1987) "Effects of Exchange Rate Changes in Developing
Countries". World Bank Discussion Paper, Report No. DRD 291. May, 1987. Pg.3.
4. See: Rhomberg, R (1964). itA Model of the Canadian Economy Under fixed and
fluctuating Exchange Rates" Journal of Political Economy. Vol. LXXX. No. 1 Feb.
1964. pp.l
,
•
UNIVERSITY OF IBADAN LIBRARY
CHAPTER FOUR
METHODOLOGY, MODEL SPECIFICATION, THEORETICAL
MODEL VALIDATION TECHNIQUES
DATA AND DATA SOURCES
"The raison d'etre of the economics profession is not
just to satisfy our intellectual curiosity, but to provide
the analytical and empirical basisfor better policies"
Marris (1984).
4.1 Summary of Issues
, So far from the literature review the following issues or testable hypotheses have
arisen
(i) That market determined exchange rate fuel inflation. This implies that there
is a pass through from exchange rates to domestic prices, such pass through
being heightened under market determined rates. .
(ii) That market determined exchange rates follow the money supply, so that
excessive money supply leads to excessive exchange rate depreciation and
hence responsible monetary policy becomes crucial for efficient exchange
system performance.
UNIVERSITY OF IBADAN LIBRARY
103
(iii) That altering the structure of relative prices in the export sector through exchange
rate depreciation increases export volume and value.
(iv) That imports are crucial to export performance such that expanding imports
lead to increasing exports while declining imports compress exports.
. (v) That changing the structure of relative prices in the import sector through
nominal exchange rate depreciation helps improve the trade balance to the
extent that it reduces imports.
(vi) That imports in developing countries depend critically on import - capacity
(i.e availability of foreign exchange) as reflected in the flow of international
reserves.
(vii) That import - capacity depends on export performance as the level of
international reserves depends significantly on net-exports (or trade balance).
(viii) That fiscal deficits rather than nominal exchange rate depreciation is the major
source of inflation in developing countries.
(ix) That macroeconomic aggregates perform better under
floating than under fixed exchange rate regimes.
4.2 Analytical Framework
In any open economy aggregate production (Y) is essentially for two purposes -
domestic use or/and exp•ortation (X) Output utilized in the domestic economy can either be
UNIVERSITY OF IBADAN LIBRARY
104
for private consumption (C,) for pri~ate investment (I,) or for government consump .on and
investment (G). Thus total aggregate production would be the sum,
Y = Cp + I, + G + X ------------------------------- (4.1)
However, home or domestic consumption can itself be supplemented by imports, so
that aggregate domestic absorption (A) is the sum of the domestic goods used up at home
plus imports.
A = C, + I,+ G + 1M ----------------------------- (4.2)
Subtracting equation (4.2) from (4.1) gives
Y - A = X - 1M ------------------------------ (4.3)
Whenever domestic absorption is greater than total domestic production then it
implies that imports are greater than exports A> Y ====> 1M> X.
Whenever such a situation arises then the country's external and internal sectors are said to
..• be in disequilibrium. More specially the trade balance (TB (which is the net of exports over
imports) is in deficit.
TB = X - 1M < 0 when 1M> X ---------------------- (4.4)
The demand for imports depends on the level of aggregate output or aggregate expenditure
or income (Y) and the price of imports in domestic currency terms (EPim), where E is the
exchange rate in terms of units of domestic currency per unit of foreign currency, and Pim
is the foreign denominated price of imports. As for exports, for a small open economy its
supply depends on its domestic price (EPx) (where Px is the foreign denominated price of
UNIVERSITY OF IBADAN LIBRARY
105
the exports and E, the nominal exchange rate) and the level of aggregate production in the
economy.
So that, taking PUn and P x exogenous, than,
x - 1M = TB = f (y,E) -------------------------- (4.5)
dTB dTB
---- < 0 ----- > 0
dY dE
The first inequality implies that when income is rising import expenditure would be rising
too so that the net balance (X-1M) would be falling. The second inequality indicates that
when the nominal exchange rate is rising, earnings from exports rise too (in domestic
currency terms) thus export volume and value rise too, and the net balance (X-1M) rises. So
,
far we have assumed some given state of tariffs and other impediments to trade.
Whenever the TB is in deficit there are two main instruments for effecting a change.
The first is to change the level of absorption (A) the other is to change the exchange rate in
order to change the structure of relative prices in the economy.
For any or both of these instruments to be successful certain elasticities' conditions with
respect to exports and imports must be met. The elasticities condition is known as the
Marshal - Lerner - Robinson condition. Expressed in terms of supply elasticities (E) and
demand elasticities (11)•for exports and imports, the Marshal- Lerner - Robinson condition•
requires that
UNIVERSITY OF IBADAN LIBRARY
106
aTB = d [X, Px (tl-1) + IM.E Pim (nimj] ---------- 4.6
where
aTB = change in the trade balance
d = Percentage depreciation
~ = Original volume of exports
Px = Price of exports
Pim = Price of imports
1M = Volume of imports
If trade was originally balanced then equation (4.6) simplifies to
aTB = ~.Px (tlx + flim- 1) ----------------------(4.7)
If the exchange rate is the instrument employed to correct for a deficit in the trade balance,
then the exchange rate has to be devalued. The devaluation will improve the trade balance
if
(fix + 'lim- 1) > 0 --------------------------------(4.8)
Inequality (4.8) implies that provided the sum of the demand elasticities for exports and
imports is greater than unity, then a devaluation. will improve the trade balance
'lx + 'lim > 1 ---------------------------------------(4.9)
Inequality (4.9) is the Marshal - Lerner - Robinson condition.
When the exchange rate is devalued as a result of a deficit in the trade balance then
the relative price of i ports to that of domestic goods rises. The rise in the relative price of
UNIVERSITY OF IBADAN LIBRARY
107
imports switchesdemand away from imports and towards domestic goods (which includes
both domestic goods consumed at home and those exported). The increased demand for
domestically produced goods raises their production (Y) while a reduction in imports
reduces absorption (A) until the excess of A over Y is eliminated and both the external and
internal sectors become is balanced again.
The level of absorption (A) can also be consciously reduced through a cut in
government expenditures.
In chart 4.1 below the horizontal distance between the AA line and the YY line measures the
level of deficit or surplus on the trade account for a given exchange rate.
Any point to the South East of Q within the QYA segment represent the excess of absorption
over domestic production and hence a deficit in the trade balance. From any point within the
segment QYA equilibrium can be attained either by reducing consumption (i.e bringing
down the level of A to equal Y) or by moving the nominal exchange rate, E upwards toward
E". This exchange rate movement has the effect of increasing Y, for as the relative price of
imports rises while that of domestic goods falls, thereby increasing demand for domestic
goods, the increased production of domestic goods. Changing the exchange rate involves
what are known as expenditure-switching policies, while cutting the level of absorption
involves expenditure - reducing policies. Under floating exchange rates the exchange rate
automatically depreciates In response to a deficit in the trade balance, so that equilibrium in
the external sector is restored via the exchange rate as demand for imports fall while that
UNIVERSITY OF IBADAN LIBRARY
108
for domestic goods rises.
Under fixed rates the government is obliged to defend the par value of the exchange
rate. Hence to prevent an exchange rate depreciation in response to a deficit in the trade
balance the government draws down on the level of international reserves to finance the
deficits. In addition government cuts down the level of absorption both through indirect
measures that may result in a cut in private expenditures and through a reduction in money
supply that may result in a cut in private expenditures as well.
If initially the money supply was rising in response to increased aggregate demand
in the economy (which led to the excess of absorption over production in the first instance)
then the following will happen in the economy. Under a system of fixed exchange rates the
increased supply of money will further aggravate the increased demand for imports and the
•• government will be forced to use the international reserve to finance the deficit. The use of
the international reserve to pay for the excess of imports will lead to a reduction in the
monetary base of the economy and hence a fall in the money supply. As the money supply
falls, demand for money falls and the level of absorption and hence as imports fall too.
Under a system of floating rates however the increased demand for foreign exchange
to finance increased imports raises the price of foreign exchange, hence the exchange rate
automatically depreciates. Exchange rate depreciation alters the structure of relative prices
in the economy in favour of domestic goods but worsens the terms of trade and reduces the
gains from trade. The increased demand for domestic goods raises domestic production and
UNIVERSITY OF IBADAN LIBRARY
109
CHART 4.1
INTERNAL AND EXTERNAL
SECTORS' DISEQUILIBRIUM
E
E* - - - - - - - - - - --- - - - Q
A
A,Y
•
UNIVERSITY OF IBADAN LIBRARY
· 110
hence domestic income. As income rises demand for money also rises thus pushing up the
rate of interest The rise in the rate of interest dampens domestic investment expenditure,
and also generate incipient capital inflow which leads to an appreciation of the domestic
currency and corrects for the deterioration in the terms of trade that would have arisen from
the initial exchange rate depreciation. Thus under a system of floating rates equilibrium is
finally restored in the exchange rate, the demand for money and the rate of interest.
4.3 Methodology
This study covers three developing countries of Africa, namely, Ghana, Nigeria and
Uganda; all of which in the wake of chronic economic fatigue adopted economic
restructuring packages that included as a key element the adoption of floating exchange rate
regimes. The study shall cover the years 1977 - 1990 for the countries of Ghana and Nigeria,
and the years 1981 - 1990 for Uganda. The difference in the time coverage for Uganda is due
to unavailability of data for the years 1977 - 1980. The data for all the countries shall be on
quarterly basis. The choice of 1977 was to enable us have at least half a decade glimpse of
the era of purely fixed rates when the authorities had not started any serious tinkering with
their economies, as it was from 1982 that some serious tinkering started.
The case of Uganda proved very interesting and aptly reflected the dilemma of
African governments on the issue of the role of exchange rate regime in their countries'
economic growth and development. At the beginning of the 1980s, Uganda adopted a
UNIVERSITY OF IBADAN LIBRARY
· 111
floating exchange rate regime, in 1987 it abandoned this system and went back to the fixed
regime (of the Bretton Wrods type) and by 1992 abandoned the fixed exchange regime once
again and went back to floating system. Uganda proved unique in several respects too. First
there was no reliable data for the first four years of the study 1977 - 1980, secondly data on
some key variables could not be got from the standard international publications such as the
International Financial Statistics and so heavy reliance had to be placed on direct data from
the Bank of Uganda and secondary data from the Statistics Department, Ministry of finance
and Economic Planning, especially from the Department's Key Economic Indicators.
In this study we hope to use an eclectic model of the Dornbusch's General Model of
the Exchange Rate (1980) variety. The model takes cognisance of both monetary and real
factors in the determination of exchange rates. The model borrows heavily from the works
of Rhomberg (1964), Tullio (1981), Khan and Knight (1986). There are two versions to our
model. There is a short version wherein there are seven stochastic variables and the level of
International Reserves is treated as a stochastic variable as obtains under a fixed exchange
rate system. There is the longer version which adds two additional stochastic variables to the
original set, while eliminating international reserves as a stochastic variable, the longer
version treats it as an identity. This longer version endogenizes the exchange rate and the
domestic rate of interest as obtains under a floating exchange rate regime.
The shorter model was estimated for the years 1977-1990 for Ghana and Nigeria. The
solution of this version 0 the model was used as the benchmark or baseline solution against
UNIVERSITY OF IBADAN LIBRARY
112
which effects of policy variations on the macroeconomy were measured. The longer model
was estimated for only the years each country floated i.e. 1986 : 4- 1990 : 4. In the case of
Uganda the longer model was estimated for the period 1981 : 1 - 1990 : 4, and the shorter
model for 1987 : 2 - 1990: 4. It should be noted however, that in the process of estimation
a few quarters were lost due to the presence of lagged variables in the system.
In calculating the unit price of imports (pIM), the unit price of exports (PX), and the
total foreign price (PF) for each focus country we took cognisance of each country's major
trading partners. For Nigeria and Ghana, ten trading partners make up at least 70% of their
import and export trade and for Uganda eight trading partners. We calculated the share of
each trading partner in the import/export trade of each focus country. In order to have the
wholesale price index for each trading partner in dollars, we deflated each trading partner's
wholesale price index by the country's dollar, exchange rate, we then weighted the dollar
wholesale price index for each country by its share in the export or import trade of the focus
country (depending on whether we were calculating unit price of export or of imports) then
we summed over all trading partners. Thus for any given focus country its unit price of
export was calculated using the formula
PJt = Kx USA • WP1USA + ~WPIJ_'~j ------------(4.10)
j_IE USA
and for import unit price
Pim = KIM USA· WPI A + 9r.WPlj• KIMj ---------(4.11)
j-IEUSAj
UNIVERSITY OF IBADAN LIBRARY
·113
where Kx USA = Share of USA in the focus country's export trade
WPIUSA= Wholesale price index of USA
WPIj = Wholesale price Index of a given trading partner j.
EUSAj= The Exchange Rate Index between country j and the USA.
Kxj = Share of country j in the export trade of the focus country.
KIMUSA= Share of USA in the focus country's import trade
KIMj = Share of country j in focus country's import trade.
Note that for Uganda, j = 1 -------------------------- 7
And for Ghana and Nigeria j = 1 ---------------------- 9
Because Ghana and Nigeria had nine major trading partners
less the USA while Uganda had seven major trading partners
less the USA.
The total foreign price faced by each focus country is
P, = Px + Pim ------------------------------(4.12)
2
For Ghana the ten major trading partners were USA, UK, Japan, Germany,
Netherlands, Switzerland, France, Spain, Italy and Canada. For Nigeria, it was the same set
of countries as for Ghana except Canada that was not a major trading partner of Nigeria
instead Belgium proved to be a partner in Nigeria's foreign trade.
The table below shows the weighted trade shares of each trading partner for each
country.
UNIVERSITY OF IBADAN LIBRARY
114
TABLE 4.1
TRADE SHARES OF GHANA AND NIGERIA 'S
TEN MAJOR TRADING PARTNERS
Focus Country: Ghana Nigeria
Trading partners Export Import Export Import
United Kingdom .2049 .3252 .0452 .2313
United States .2043 .1798 .4416 .1560
Japan .1177 .0722 .0026 .1041
Gennany .1604 .1232 .1171 .0573
Netherlands .1177 .0293 .0945 .0389
Switzerland .1129 .0293 .0055 .1349
Spain .0352 .0108 .0580 .1349
France .0264 .0453 .1322 .0709
Italy .1750 .1227 .0809 .0709
Canada .0029 .0357 nil nil
Belgium nil nil .0225 .0338
Note that the ten trading partners for Ghana and Nigeria, make up about 75% of her export
and import trade. For Uganda, her eight major trading partners make up about 75% of their
export and import trade; attempts to add two more co1J!ltries to give Uganda ten major
trading partners as the other focus countries was not worth the effort because the next two
trading partners added less than half of one percent between them to the import and export
trade of Ugandan. Table 4.2 shows the distribution of trade between Uganda's eight trading
partners.
UNIVERSITY OF IBADAN LIBRARY
'115
TABLE 4.2
TRADE SHARE OF UGANDA'S EIGHT MAJOR TRADI G PARTNERS
Focus Country: Uganda
Trading Partners Export Import
United Kingdom .1597 .3506
United States .2954 .0673
Japan .0661 .1128
Germany .0831 .2024
Netherlands .0921 .0473
Spain ,.1098 .0207
, France .1299 .0664
Italy .0632 .1324
As mentioned earlier our study made use of quarterly variables, however there were
some variables such as the Gross Domestic Product (GDP) that were in annual form and so
had to be spliced through interpolation. To interpolate annual variables to their quarterly
components we used one of Friedman's (1962) approaches. In interpolating a variable that
is of interest to us, we selected another variable whose quarterly values are known and
whose intrayear movements are believed to be comparable to those of the variable of
UNIVERSITY OF IBADAN LIBRARY
116
interest, then we superimposed the intrayear movements in the selected variable on the
variable of interest, and then we eliminated the differences between the year to year
movements. In imposing the intrayear movements of a chosen variable on a variable of
interest we used "geometric operations". More specifically the following formula was
used for the interpolation.
= = !!i .~ -----------(4.13)
~
Where a = The series whose quarterly values are not known
b = The series whose quarterly values are known
at = The interpolated value of i in the series a
T3j = Trend value of 'i' in the series a
,
Tb, = Trend value of 'i' in the series b
i = i--4-.
The variables that were interpolated were (i) Gross Domestic Product (Y) (ii)
Government Revenue (GR) (Hi) Government Expenditure (GE) and (iv) Current Account
Balance (CUB). Variables believed to exhibit comparable intrayear movements to the above
focus variables were
(i) Imports - for Gross Domestic Product
(ii) Exports - for Government Revenue
UNIVERSITY OF IBADAN LIBRARY
'117
(iii) Money supply - for Government Expenditure
(iv) International Reserves - for Current Account Balance
The basis for the above assumptions is grounded in economic theory. Imports for
instance, has been proved to be a function of income such that the higher the level of income
the higher the marginal propensity to import (Oyejide 1992). Also as income falls imports
also tend to fall either due to government regulations restraining imports in the face of
declining income or due to a fall in people's purchasing power as income falls.
Similarly t the level of government revenue was assumed to be reasonably tracked by
the value of exports. This is because empirical research results have shown that for
developing countries, particularly primary - producing ones the major source of revenue for
government is export - earnings, usually from one major primary commodity. For instance
in the case of Ghana, cocoa is the major export commodity and earnings from Cocoa make
up a considerable fraction of government revenue, while the same applies to Uganda whose
main export is coffee. As for Nigeria over 90% of government revenue comes from
petroleum exports (Pinto, 1988; Olopoenia, 1991). Thus in interpolating we assumed that
the intrayear movements in the value of exports are comparable with that of government
revenue.
As for government expenditure several studies Elbadawi (1990) Mbire and Watuwa
(1992) among others have shown it to be closely related to money supply especially in
developing African economies. it has been shown that almost always increases in money
UNIVERSITY OF IBADAN LIBRARY
'118
supply are in response to financing government's expenditure.
Finally, movements in international reserves often are in response to the current
account requirements in parti ular to the trade balance component of the current account as
discussed in (4:2). Thus current account deficit is often eliminated by drawing down reserves
while a surplus in the current account is reflected in rising level of international reserves.
Thus intrayear movements in the level of international reserves were superimposed on the
current account balance.
The capital account balance was treated as a residual. Given the change in the level
of international reserves (DRE), the capital account balance CAB, is the excess of this
change in reserves over the current account balance, CUB.
DRE = CUB + CAB ----------------------------- (4.14)
DRE - CUB = CAB ------------------------------(4.15)
The models were estimated using the Ordinary Least Squares (OLS) method. We
estimated the shorter version for the years 1977:1-1990:4 for Ghana and Nigeria. This
shorter version is the model of the economy under fixed exchange rate system. The longer
version was estimated for only the years each country floated i.e 1986:4 to 1990:4; on the
assumption that the parameters of the first estimate remain constant
As pointed out earlier the case of Uganda was slightly different. For the ten years
covered for Ugandan, i operated a floating exchange rate system for a greater proportion of
the period. So we used the solution to the longer version of the model as the benchmark
UNIVERSITY OF IBADAN LIBRARY
119
solution against which we measured the changes in the variables under different policy
stance. Thus we estimated the longer version of the model for the years 1981: 1-1990:4 and
the shorter version for 1987:2-1990:4 for Uganda.
We were also interested in real income and how its value at any time 't' is
determined. First we calculated real income by deflating the gross domestic product (GDP)
by the domestic price level (PD). We then calculated the trend in real income, which we
termed RIS, by the formula.
RIS = RIo + get) ------------------------------4.16
Where RIo is real income at initial period, and 'g' was generated in the system. For
Nigeria 'g' was found to be -3.25; for Ghana -0.29, and, for Uganda 0.74.
We then experimented with two ways in which previous level of real income was
, likely to affect current real income. First, we assumed that the current level of real income
would be affected by the difference between the actual real income in the previous period
and its trend level in that period.
i.e O""'J1 = ~(RI(t-I) - RI (HV) ------------4 •17
RIt = RIt-l + A.-----------------------4.17b
"'1t = (RIS (H) - RI(HV) --------------4.17 c
We designated ~ = (RISt_l - RI(H)as RIST
Then we also experimented with the difference between the actual real income level
and its mean level in previous period impacting on the current level of income.
UNIVE
SITY OF IBADAN LIBRARY
120
i.e Rl, = f (RI(I-I)- Ri) -----------------4.18
more specifically,
Rl, = ~-1 - 5 ----------------------------4.18b
~ = RIr-r -RJ--------,----4.J8c
We designated 5 = ~-1 - RI as RISD
With respect to inflation we assumed expectations to be static so that the previous
rate of growth of prices (one quarter away) is expected to be its rate of growth in the current
quarter.
i.e XI = PD(I_l)- PD(I_2) --------------4.19
PD(t-2)
We were also interested in two types of real exchange
rates.•, the real exchange rate in the export sector which is
the ratio of the price of exports to that of domestic goods
EPx/Pd which we termed ETOT, the real exchange rate, in the
import sector, which is the ratio of import price to that of
the domestic price level EPimjP d, which we termed RPIM.
4.4 Model Specification
In this study we shall divide each focus country's economy into three sectors. The
first sector is the Price and MQnetary Sector, the second is the Foreign Trade Sector and the
UNIVERSITY OF IBADAN LIBRARY
121
third is the Real Sector.
We shall further assume that each economy is small hence it is a price taker both in
its import and export market. We further assume that neither imports nor exports are perfect
substitutes for domestic non-traded goods. Also that the consumer is rational and hence
maximizes utility subject to a budget constraint. Thus when relative price shift makes some
goods relatively more expensive less of such goods would be bought. We also assumed that
the supplier maximizes profit subject to a cost constraint. Thus given a cost structure the
higher the price the more of such goods would be produced.
Given the above assumptions therefore. the structure of our model. in general
functional form. is as shown below; we allow the data to determine the specific form. (Note
that Appendix I contains the definitions of all variables and terms).
I: Price and Monetary Sector
P~t) = f(RI(t).MSSI• EN(I)'RO(I)'X(l)'Pd(I-tJ 4.20
RDM(I) = f(RI(t)tGEl' RDw RO~I-tJ 4.2 i'
MSS(I) = f(RI(l» GDt• DePI• . 4.22
RO(t) = f(EMSS(t). Xt•Yet).RO(t-l' 4.23
In the price and monetary sector. equation (4.20) maintains that the domestic price level (pJ
depends on the level of real income (RI) nominal money supply (MSS) nominal exchange
rate (EN) nominal rat of interest (RD) inflationary expectations (x) and its own lagged
value Pd(t-l)' 'A - prior' we expect an inverse relationship between domestic prices and
UNIVERSITY OF IBADAN LIBRARY
122
output such that expansion of real income should lead to declining prices. Increased money
supply is expected to increase the level of domestic prices while a depreciating nominal
exchange rate is expected to lead to rising prices. Inflationary expectations which should
lead to increased demand for money is expected to also lead to rising prices. Given that
prices tend to be rigid downwards there is a threshold of the previous price level below
which the current price will not fall according to equation (4.20).
Equation (4.21) maintains that the real demand for money (RDM) is a function of
real income (RI)t government expenditure (GE) the rate of interest (RD) and its own lagged
value (RDM..l). It is expected that the higher the level of income the higher would be the
real demand for money given the classical position that money is demanded for transaction
purposes. Government expenditure is a major force determining the level of economic
activity in both developed and developing economies, hence the level of government
expenditure is expected to have a positive relationship with the real demand for money. As
for domestic rate of interest given the keynesian speculative motive for holding money we
expect an inverse relationship between the real demand for money and the rate of interest,
so that, as the rate of interest rises the real demand for money falls.
The money supply equation (4.22) maintains that money supply (MSS) depends on
the level of real income (RI) the government deficit (GD) and the domestic credit to the
private sector (DCP J. The higher the level of change in reserves the more money is
available to sustain increased level of domestic activity. As the level of real income rises
UNIVERSITY OF IBADAN LIBRARY
"123
the more the amount of money needed to sustain the level of transactions in the economy,
hence the higher the money supply. The domestic rate of interest (RD) is supposed to depend
on the excess money supply (EMSS) according to equation (4.23). The higher the excess
••• money supply the lower the rate of interest becomes. Also the domestic rate of interest
depends on inflationary expectations (n) nominal income (Y) and its own lagged value (R~.
J As inflationary expectation rises.demand for money rises too, thus pushing up the rate
of interest. Also the higher the level of income the higher the demand for money and hence
the higher the rate of interest.
II Foreign Trade Sector
~t) = f(RIS(!), ETOTt.2, IMo."DPX(t.J 4.24
IM(t) = f(RDM(t), RPIM(t)t DRE(tJ 4.25
" RE(I) = f(EMSS(t.l)' TB(HJ 4.26
EN(t) = f(RDF(t), RE(H)' Y(t)tMSSJ 4.27
The foreign trade sector is linked to the monetary sector through the domestic price
level which is one of the determinants of relative export price ETOT, and relative import
price RPIM. Further in equation 4.24 the value of exports (X) depends on the trend in real
income (RIS) the relative price in the export sector or real exchange rate in the export sector
with a two period lag (ETOTt.J the value of imports in the previous period (IM..) and the
domestic price of exports with a two period lag (DPx..J. As the trend in real income rises
export volume and hence value tend to rise. Similarly as the relative price of exports
UNIVERSITY OF IBADAN LIBRARY
(EPX/Pd == ETOT) with a two period lag, rises, the volwne and value of exports tend to rise
too because increased relative price of exports encourage increased export production. The
value of imports also determine the volwne of exports supply to the extent that imports in
the. previous period (lM.-l enter into the current production for exports. The domestic price
of exports also determine the quantity supplied in the light of our asswnption that the
producer is a profit maximizer, hence the higher the domestic export price the higher the
supply and hence the higher the export value.
Equation (4.25) maintains that imports (1M) depend on excess real demand for
money t (RDMEJ the money supply and the real demand for money link the monetary sector
with the foreign trade sector. Thus according to equation 4.25 imports depend on the real
demand for money RD~t the level of real income (RI) relative price of imports RPIM) and
change in reserves (DRE). While equation 4.26 the reserve level is influenced by the excess
money supply in the economy. The higher the excess money supply the lower the reserve
level is expected to be as the level of reserve is usually drawn down, and monetized to
increase the money supply and equation 4.27 the nominal exchange rate depends among
others on the nominal money supply. Real income (RI) the relative price of imports or the
real exchange rate in the imp~:>rtsector (RPIM = EPjJpd) and the import - capacity as
reflected in the change in reserves at a given time (t). The relationship is positive as between
imports and real income, and between imports and change in reserves, but negative as
between imports and its relative price.
UNIVERSITY OF IBADAN LIBRARY
125
The reserve level itself (RE) according to equation (4.26) depends on the excess
money supply in the economy EMSS level of government deficits with a one period lag
"
(GD,_t) and the trade balance (or net-export) also with a one period lag. Reserves and
government deficit have an inverse relationship, while it has a direct relationship with net
exports (or the trade balance) Nominal exchange rates according to equation (4.27) depends
on the interest rate differential (RDF) (i.e the differential between domestic and foreign rates
of interest (RD - RF) and also on the level of international reserves with a one period lag
(RJ;-l)' It also depends on nominal income (Y) and nominal money supply (MSS). While
.
there is an inverse relationship between nominal exchange rate and the interest rate
differential because the higher this differential the more appreciated the exchange rate gets.
There is similarly an inverse relationship between nominal exchange rates and the reserve
,level as well, the higher the level of international reserves the more appreciated the exchange
rate becomes also an inverse relationship exists between the nominal exchange rate and the
level of income (Y) because the higher the level of income the more appreciated the nominal
exchange rate becomes. As for nominal exchange rate and the money supply, if money
supply rises without a corresponding rise in the demand for money then the rate of exchange
depreciates.
III The Real Sector
RI(I) = f(RDME(t); RISDt, PD(,), I~_l) (4.28)
Equation 4.28 maintains that in the real sector the level of real income (RI) is determined
UNIVERSITY OF IBADAN LIBRARY
126
by the of excess real demand for money, (RDMEJ the differential between actual real
income in the previous period and the long run mean income (RISDJ the level of d estic
prices (PD) and the value of imports with a one period lag I~'l' While real income has a
positive relationship with excess real demand for money RDME. real income differential
(RISD) and imports (lM(I.lJ it has a negative relationship with prices. The identities and
definitional equations of our system are
TB(I) - X(I).- IM(I)' 4.29
GD(I) ii GE(I)'- GR(t)t 4.30
TDC(I) - DCPw- DCG(I) 4.31
EMSS(I) - MSS(I'I)'- RDM(I), 4.32
CAB(I) ii DRE(t) - CUB(I) 4.33
,
EN(I) - EN(I)+ 0 4.34
RD(I) - RD(t)+ 0 4.35
RE(I) - RE(I)+ 0 4.36
Identity (4.29) maintains that the trade balance (TB) is the excess of export value (X)
over import value (1M). Identity (4.30) is the government deficit (GD) which equal the
excess of government expenditure (GE) over government revenue (GR) which equals the
excess of government &xpenditure (GE) over government revenue (GR). Total domestic
credit (TDC) is according to identity (4.31) exactly equal to the sum of domestic credit to
UNIVERSITY OF IBADAN LIBRARY
127
the private sector (DCP) and domestic credit to government (DCa).
Our definitional equation (4.32) maintains that excess money supply in the current
period '1' (EMSSt) is the excess of money supply in the previous period MSS(t_,) over the
actual real demand for money in '1' (RDMJ. The capital account balance (CAB) is
according to identity (4.33) the residual of the change in reserves after taking away the
current account balance. The nominal exchange rate (EN) is fixed under a fixed exchange
system according to identity (4.34), while the nominal rate. of interest in the domestic
economy (RD) is also fixed under a fixed exchange rate system according to identity (4.35).
Lastly the level of international reserves (RE) is non-varying under a floating exchange rate
system according to the last identity (4.36).
Beyond the statistical tests of the individual equations and parameters rigorous test
" of the validity of a model through dynamic simulation is essential (pindyc1c and Rubinfeld
1981). The essence of such simulation is to enable us determine whether the model is
reasonable in reflecting economic reality, and to reveal the economic policy control
properties of the model so as to be able to determine the extent to which one can rely on the
model for forecasting purposes. Also simulation is useful in suggesting problem areas that
can be improved for a better specification of the model. Thus we attempted to validate our
models through dynamic simulation .
•
UNIVERSITY OF IBADAN LIBRARY
. 128
4.5 Theoretical Model Validation Techniques
In this study we were concerned with how wel our model tracked the historical data
generally and in particular how well it captured turning-points. Also we were concerned with
the predictive ability of each of our model(s). Thus in orcer to determine how well our
model(s) tracked the historical data we analysed the Root Mean Square Simulation Error
(RMSE) for each of our endogenous variables. For any given endogenous variable Z at time
'1' its Root Mean square simulation error is defined as
........ (4.36)
RMSE = T
.L ('45- '48)2
T t=1
i.e. the square root of the sum of squared deviations of
actual values of Z, from the simulated values Z, divided by the total number of periods
involved. The Root Mean Square Simulation Error is a measure of the deviation of the
simulated variable from its actual time path. The magni,tude of this error is evaluated by
comparing it with the average size of the variable in question (SRMSE). The smaller the size
both of RMSE and of SRMSE the better the model.
We also constructed charts reflecting the simulated and the actual values of our
endogenous variables for all the sample countries to reveal pictorially how well turning-
points are tracked. To test the predictive ability of our model(s) we were interested in the
Theil's inequality coefficient. The Theil's inequality coefficient is supposed to range from
•
UNIVERSITY OF IBADAN LIBRARY
-129
o to unity. When this coefficient equals 0 then it implies that the simulated and actual values
of the endogenous variable are exactly equal and hence there is a perfect fit.
i.e. U = 0 ?;S=?;& for all t
(Where U = Theil's inequality coefficient).
When the Theil's inequality coefficient equals unity then it shows that the predictive
ability of the model is as bad as it possibly could be (pindyck and Rubinfeld, 1981)
U = 1 ?;S ?;I for almost all t.
The highest value tolerable for a Theil's inequality coefficient is 0.3. The
decomposition of the Theil's Inequality into its characteristic sources - i.e. Bias Proportion,
Variance Proportion and Covariance Proportion is also important in determining the validity
of a model. The bias proportion is a reflection of systematic error, since it measures the
extent to which the average values of the simulated and actual series deviate from each
other. Thus the closer this bias proportion is to zero the better. A larger value of the bias
proportion is a cause for concern since it means that systematic error is present.
The variance proportion is a reflection of the ability of the model to replicate the
degree of variability in the endogenous variable of interest. If the variance proportion is
large then this means that the actual series has fluctuated considerably while the simulated
series shows little fluctsation or vice-versa. Thus the closer the variance proportion is to zero
the better.
UNIVERSITY OF IBADAN LIBRARY
130
The covariance proportion however measur unsystematic error - i.e it represent the
remaining error after deviations from average values and average variabilities have been
accounted for. The closer the covariance proportion is to unity the better, because, given that
predictions cannot be perfectly correlated with actual outcomes, and we do not want bias
errors and variability errors, then whatever errors we observe should only be unsystematic.
In other words whenever the Theil's Inequality is greater than zero, then the ideal
distribution of the errors should be,
for U > 0, then, U" = 0, UV = 0, U" = 1
where U" = Bias Proportion, U" = Variance Proportion and UC Covariance
Proportion
, Results of the estimation, validation and simulation exercises are presented in
chapters six and seven.
4.6 DATA AND DATA SOURCES
For this study we collected data on the following variables
(i) Gross domestic Product, (Y)
(ii) Money supply - (MSS)
(Hi) Domestic Prices - (PD)
(iv) International R•eserves - (RE)
(v) Domestic Rates of Interest - (RD)
UNIVERSITY OF IBADAN LIBRARY
(vi) Value of Imports - (IM)
(vii) Value of Exports - (X)
(viii) Nominal Exchange Rates - (EN)
(ix) Unit value of Imports - (pIM)
(x) Unit Value of Exports - (PX)
(xi) Foreign Prices - (PF)
(xii) Home goods' Prices (i.e Non Traded Goods' Prices) - (PH)
(xiii) Domestic Credit to Government - (DCG)
(xiv) Domestic Credit to the Private Sector - (DCP)
..• (xv) Trade Balance - (TB)
(xvi) Current Account Balance - (CUB)
(xvii) Capital Account Balance - (CAB)
(xviii) Change in Reserves - (DRE)
(xix) Government Expenditure - (GE)
(xx) Government Revenue - (GR)
(xxi) Foreign Rates oflnterest (Using USA's as proxy) - (RF)
For each country most of the data were got from the International financial statistics
(IFS) of the IMP (Interrjational Monetary Fund), the Balance of payments yearbook also of
the IMP, and the Government Finance Statistics Yearbook. Among others the following
were collected from the IFS' various years, - International Reserves, Domestic Rates of
UNIVERSITY OF IBADAN LIBRARY
-132
Interest, Nominal Exchange Rates, Foreign Rate of Interest. (i.e USA's). In the case of
Nigeria the following additional variables were also got from the IFS,
- Money supply, Domestic Credit to Government, Domestic Credit to the Private Sector,
Export Value, Import Value.
However data on Nigeria's, Government Revenue and Government Expenditure were
from the Central Bank of Nigeria's Publications." In particular the Statistical Bulletin of the
Central Bank of Nigeria proved invaluable. Similarly, Ghanaian data on Current Account
Balance and Government Revenue as well as Government Expenditure were got directly
from the Bank of Ghana, Research Department.
As for Uganda the majority of her data were got direct from the Bank of Uganda. The
Bank of Uganda's "Monetary-Survey of 1992 provided data on Domestic Credit to
Government, Domestic Rate of Interest and Nominal Exchange Rates. Data on Uganda's
GDP, were collected from "Key Economic Indicators" of the Statistics Department of the
Ministry of Finance and Economic Planning .
•
UNIVERSITY OF IBADAN LIBRARY
CHAPTER FIVE
DESCRIPTIVE ANALYSIS OF DATA
"When moving from the PRECISEformulations and GENERAL PRESCRIPTIONS
of economic theory to the real world of policy, it is necessary to remember that we
leave the realm of SYSTEM and PRECISION for a world of APPROXIMA TION,
which operates not by LA W but by TENDENCY"
Scammell (1974)
Some of the key economic aggregates on which the exchange rate impact directly are
the value of a country's exports, the value of imports, the trade balance, the current account
,
and the level of international reserves. Also the exchange rate could impact on domestic
prices and indirectly on the level of real income. In this chapter we describe with the aid of
appropriate charts and tables the behaviour of these macroeconomic aggregates over the
.
period covered by the study both under fixed and under floating exchange rate regimes.
5.1 The Trade, Current and Capital Account Balances
One of the key aggregates of interest to this study is the trade balance (TB) and its
behavioural pattern during th~ period of study. In Nigeria the trade balance in the early
years of the 1980s was almost always negative. Out of the 12 quarters of 1981 to 1983 the
TB was negative for 9 of the quarters. This probably reflects the squandermania attitude
UNIVERSITY OF IBADAN LIBRARY
·134
which resulted from the oil boom of the 1970s and the difficulty in stemming the tide when
there was a reversal of fortune. This negative trend in the trade balance at this period in time
probably explains the adoption of "austerity measures" in 1982 in order to put a brake on
import-growth and reduce the level of government expenditures. Among others the
"austerity measures" included the banning of certain categories of imports and the reduction
of Basic Travel Allowance (BTA) along with other trade controls. The "austerity measures"
were officially known as Economic Stabilization Act (1982).
Apparently. the effect of the austerity measures on merchandise trade only came to
be felt from 1984 when the trade balance was in surplus and this surplus trend continued
until 1990. It is interesting however to note that the current account balance however was
-negative for a considerable part of 1984 - 1990 in spite of the positive balance on the trade
account. In fact out of the 8 quarters of 1986-1987 the current account balance was negative
for 7 quarters. this was in spite of the Structural Adjustment Programme (SAP) that had been
put in place since the last quarter of 1986. It is only in the last three years of the study (i.e
1988 - 1990) that the current account balance became positive. Given the fact that the trade
balance had exhibited a positive trend since 1984. then the negative balances on the current
account for an the years of 1984 - 1987 (except for 1985) can only reflect the increasing
dependence of Nigeria on the imported skilled labour and other services of her trading
partners. such that the s•ervices balance ensured that the current account balance was
negative. In fact out of the 16 quarters of 1981-1984 the current account recorded deficit
UNIVERSITY OF IBADAN LIBRARY
. 135
balances for 15 quarters. This probably partly explains the need for SAP in 1986.
The capital account balance was positive from 1981 up to 1984:4 reflecting the
inflow of external loans. However, from 1984:2 the balance on this account started a
declining trend and by the first quarter of 1985 it had become negative. The negative trend
on the capital account continued until the first quarter of 1986, thereafter the tide changed
again. These negative trend seems to reflect the drying up of external credit lines as the level
of Nigeria's external loan skyrocketed, and also reflect the huge repayment burden that
Nigeria faced. The inflow of capital peaked in 1983, when for that year alone capital inflow
was some Nl.7 billion, an increase of about 55% over the 1982 inflow of Nl.l billion. As
, credit lines dried up capital inflow also started a downward trend from a level of N2.95
billion in the 4th quarter of 1983 to NO.32 billion in the 4th quarter of 1984. By the first
quarter of 1985 there was no inflow at all, rather there was a net outflow of NO.51 billion.
With the inception of the structural adjustment program however capital flowed in again, as
the country was deluged with all kinds of structural adjustment loans from both the World
Bank and IMF and other multilateral financial institutions. A considerable proportion of the
old (outstanding) loans were rescheduled by pushing their repayment farther into the future .
•
UNIVERSITY OF IBADAN LIBRARY
136
TableS. 1
Nigeria's External Balances 1981 - 1990
(In millions of Naira)
Period Balance of Current Account Capital Account
Trade Balance Balance
1981 : 1 1239 -3415 3705
: 2 - 372 -4117 3440
: 3 -1577 - 4384 305
:4 -239 - 3998 2476
1982: 1 - 700 - 2161 848
: 2 - 182 - 2294 2139
: 3 - 304 - 2919 2909
:4 404 - 4875 5081
1983 : 1 -984 - 5349 5468
: 2 316 - 3518 3116
: 3 - 407 - 5349 5466
:4 - 267 - 3136 2954
1984 : 1 1494 - 2115 2130
: 2 997 - 1653 1891
: 3 868 - 553 297
: 4 965 45 318
1985: 1 957 565 - 509
, : 2 353 901 - 1043
: 3 614 911 - 1187
:4 1357 2215 - 1487
1986: 1 424 6278 - 5039
: 2 598 - 1883 1174
: 3 651 - 8281 8432
:4 792 - 2999 2703
1987: 1 2477 - 875 1721
:2 3278 - 1592 2013
: 3 3522 - 1023 2027
:4 4605 - 295 776
1988: 1 2951 680 915
:2 4198 722 - 203
:3 5085 1051 237
:4 201 1448 1289
1989: 1 8839 868 4067
:2 10022 2832 - 2390
:3 6489 4332 - 3036
:4 7444 6735 - 3354
1990 : 1 8459
• 17393 -10635:2 8908 25897 -21119
:3 17714 34396 -30200
:4 31862 41782 -38468
Sources: Central bank of Nigeria - Statistical Bulletin (Various Issues)
UNIVERSITY OF IBADAN LIBRARY
137
By the second quarter of 1989 the capital flow pattern changed again. There was
from the second quarter of 1989 an outflow of capital reflecting the expectations of the IMF
and the WB and other creditors that after three years of structurally adjusting the economy,
Nigeria should by then be able to pay its debts. What is more, repayment on the rescheduled
debts had by 1989 started falling due. Thus for seven of the eight quarters of 1989-1990
there was net capital outflow. Table 5.1 bear out these facts.
In Ghana the Economic Recovery Program (ERP) took off in 1983. Prior to 1983,
especially in the 1970's the performance of the Ghanaian economy had been particularly
poor. This is reflected in the fact that of the 8 quarters of 1977 and 1978 the Trade Balance
(TB) was negative for 6 of the quarters. Similarly the TB was negative for all the four
" quarters of 1981. After the onset of the Economic Recovery Program in 1983 there was a
marked improvement in the merchandise trade as reflected in the TB for 1983 up to 1986.
Of the 16 quarters in this period the trade balance was only negative for about 5 quarters.
From 1986:4 when the country adopted a floating exchange system the performance of the
TB was not particularly impressive compared to the early years of adoption of ERP. In fact
of the 17 quarters of floating exchange rate experience covered in this study (i.e 1986:4 -
1990:4) the TB was negative for about 9 quarters. This is quite in contrast to what obtained
inunediately after the onset ofERP wherein for the 8 quarters of 1983 and 1984, the TB was
•
positive for 6 quarters. The Current Account Balance (CUB) even proved to be a greater
•
enigma than the Trade Balance. The CUB was negative for all the 24 quarters before the
UNIVERSITY OF IBADAN LIBRARY
138
onset of ERP (i.e 1977:1 - 1982:4) except for 2 quarters. In spite of the onset of ERP in
1983 and floating exchange rate system in 1986, the CUB continued to be negative
throughout, i.e. from 1983:1 to 1990:4. The phenomenon seems a little hard to explain. It
probably reflects the complete dependence of the Ghanaian services sector on external
services in terms of skilled labour, shipping and maritime services insurance services etc.
In contrast with the CUB performance the Capital Account Balance of the Ghanaian
economy proved to be positive throughout the 56 quarters covered by the study except for
4 quarters. This capital Account Balance (CAB) reflects the inflow of capital in form of
loans and grants prior to the onset ofERP in 1983 and in form of structural adjustment loans
from the IMF and the World Bank since the onset of ERP. That there has not been any
,
marked outflow of capital probably reflects how sympathetic Ghanaian external creditors
had been with her. Table 5.2 shows Ghana's external balances.
As for Uganda it embarked on an Economic Recovery Program (ERP) in 1982 to
correct for the deteriorating economic performance of the 1970s. Unfortunately, meaningful
data are not available on the '70s for Uganda. Thus the study covers only the ten year period
1981-1990. During this ten year period the merchandise trade of Uganda revealed an
interesting trend. Just after the adoption of the ERP in 1982 the TB performed beautifully
as revealed by the positive balance on the trade account for almost all the 16 quarters
immediately following the adoption of the ERP. However from 1987
UNIVERSITY OF IBADAN LIBRARY
139
Table 5.2
Ghana's External Balances 1981-1990 (In Millions of cedis)
Period Balance of Current Account Capital Account
Trade Balance Balance
19&1 : 1 -30 -236
:2 201-29 -218 118
: 3 -29 -802 794
:4 -28 -1160 1152
1982 : 1 302 -998 1013
:2 84 -794 827
: 3 243 -489 460
:4 -166 -205 160
1983 : 1 -126 -863 1531
:2 503 -1416 1843
:3 392 -1513 1493
:4 2510 -1384 1196
1984 : 1 -2199 -1669 4018
:2 1309 -2114 4849
:3 1406 -1958 2664
:4 2.726 -2753 6863
1_ 5 : 1 2333 -5099 1~O31
:2 354 -6339 9166
:3 -2092 -7429 9935
:4 -4530 -8508 11279
1986 : 1 3266 -11050 26329
:2 3545 -7365 -1136
:3 4537 -7224 12077
:4 -3242 -7609 15727
1987 : 1 -2500 -18706 55529
:2 5082 -24584 24138
:3 5764 -13156 -35162
:4 -3247 -15066 11238
1988 : 1 5680 -15500 20736
:2 -3886
:3 -16301 23970-4400 -11461 22752
:4 5174 -13537 24605
1989 : 1 -19150 13844 20736
:2 -16934 -13875 15495
:3 -20836 -13017 • 10959.: 4 -19680 -25245 71490
1990 : 1 30831 -39962 54403
:2 39834 -51029 54064
:3 33473 -62677 67735
:4 30539 -71108 86108
Source: Bank of Ghana: Research Department's "Balance of Payments Documents"
UNIVERSITY OF IBADAN LIBRARY
" 140
Table 5.3
Urganda' s Externa 1Balances 1981- 1990
(In muuons ot uzandan srunmzsj.
Period Trade Balance Current Account Car.ital Account
Balance Ba ance
1981 :1 1872 29 252
:2 471 4 -157
:3 -16 9 937
:4 -6175 14 3353
1982 :1 -1613 -7 -6
:2 318 -63 3926
:3 5441 -102 -1838
:4 -3255 -66 -4085
1983:1 -2745 -6700 10580
:2 2272 -13328 14069
:3 1214 -11751 9316
:4 6655 -11124 10116
1984 :1 4971 3275 -25973
:2 14910 19361 -20659
:3 11003 24112 -26258
:4 -5340 38524 -26724
1985 :1 12539 29559 -38368
:2 11760 44578 -27274
:3 26367 75009 -39852
:4 10776 51746 -77175
J986 :1 34 565627 -570817
:2 508 24903 17657
:3 860 189760 -160227
:4 580 72240 -157500
1987 :1 -637 3184 . -41838
:2 -3385 866 -1778
:3 -3464 -1795 9027
:4 -4255 -44319 46149
1988 :1 -3038 -24038 23462
:2 -3497 -12551 11921
:3 -9465 -35041 4131
:4 -6511 -20380 193151
1989 :1 -4106 -10116 4997
:2 -8103 -38543 42206
:3 -6759 -49311 15511
:4 -31124 -58712 61369
1990 :1 -26084 -123302 127949
:2 -22843 -130136 130278
:3 -31825 -122857 122087
:4 -538~2 -114886 116272
Source: 1M!<" s Balance of Payments Yearbook (Vanous Years)
UNIVERSITY OF IBADAN LIBRARY
-141
when the country went back to a fixed exchange rate system the Trade Balance (TB) became
negative and remained so until the end of the study period. (i.e 1987-1990). Table 5.3
reveals these points.
It is interesting to note that except for 1983 the Current Account Balance (CUB) was
also positive for all the years that the TB was positive; what is more, as the TB turned
negative from 1987, the CUB also became negative in particular from 1987:3 up till 1990:4
(see Table 5.3). This trend probably implies that the merchandise trade is the. more
predominant in Uganda's merchandise and services trade. Thus the CUB seems to move in
the same direction as the TB, signifying that the services balance is not as dominant as the
merchandise trade balance.
, The capital Account Balance (CAB) between 1981-1983 was more positive than
negative probably reflecting the inflow of loans and grants in an attempt to assist Uganda
restructure its economy. From 1984 however there seemed to have been a reversal as capital
flowed out for almost all the quarters of 1984:1 to 1987:2. This is probably a reflection of
Uganda's attempts to repay some of its huge outstanding external debts. From 1987:3
however up to 1990:4 there seemed to have been continuous capital inflow again. This
shows how the IMF and the WB have responded to the distress call of the United Nation
especially the Economic Commission for Africa for an urgent and substantial inflow of
capital into distressed Af can economies (Adedeji, 1985). The IMF standby arrangement
which Uganda used in funding its exchange market between 1982-1984 ended in 1984. The
UNIVERSITY OF IBADAN LIBRARY
t' .
142
increasing pressure of debt repayment in the next ten quarters must have revealed the need
for development aids and grants, the response by the developed countries is reflected in the
capital account balance of Uganda for the period 1987:3 - 1990:4. These facts are shown in
Table 5.3.
5.2 International Reserves and the Balance of Payments
The behaviour of Nigeria's international reserve level is worthy of examining. By
the first quarter of 1981 the level of international reserves in Nigeria peaked to stand at N5.9
billion, thereafter, it started a downward trend, and by 1983 it stood at some NO.7 billion
w,hich is only some 12% of its peak level. By 1984 the level of international reserves picked
up again, but very marginally. By 1986 the increase in the level of international reserves
became noticeable and by 1989 it had even surpassed its peak level of N5.9 billion of the
first quarter of 1981. By 1990 fourth quarter the level of international reserves (RE) was
some N31.1 billion. It is important to take note however that the level of RE is here reduced
to its naira equivalent. Thus the naira value of RE may be reflecting more the depreciation
of the naira exchange rate rather than the rate of reserve accumulation. Thus, in dollar terms,
the growth of RE may not be as outstanding as it seems. In fact a quick conversion of the
level of RE into dollars reveals that at its peak in 1981 the level of RE was some US $9.7
billion, whereas in 1989 the highest level was only US $1.8 billion and by the fourth quarter
of 1990 it was US $3.7
UNIVERSITY OF IBADAN LIBRARY
143
Table 5.4
In tematto•n aI Reserves and Balance 0fP ayments f,or FOCllS
Countries (IY~I:I - I 'JYU:4)
Nigeria Ghana Uganda
(In millions of Naira) (In millions of Cedis) (In millions of
Ugandan Shillings)
Period
RE BOP RE BOP RE BOP
1981 I 5887 7(1) 506 -35 409 281
2 5211 -676 406 -100 255 -154
3 3879 -1332 415 9 1201 946
4 2357 -1522 407 -8 4568 3367
19821 1044 -1313 432 -25 4555 -12
2 889 -155 465 33 8419 3863
3 879 -10 436 - 6478 -1941
4 1086 207 392 -
2494 2328 -4150
19831 1205 119 1062
-402 670
6208 3880
2 803 1487 425 6949 741
3 925 122 1467 -20 4514 -2434
4 743 -182 1279 -188 3506 -1008
19841 758 15 3627 2348 32754 -29248
2 996 238 6362 2735 31455 -1298
3 759 -237 7067 705 292(1) -2247
4 1122 363 11177 4110 41009 11800
19851 1178 56 181(1) 6932 32199 -88(1)
2 1035 -143 2(1)36 2827 49504 17304
3 7fIJ -275 23442 2506 84660 35156
4 1488 728 2f1J14 2572 59231 -25428
19861 2727 1239 41293 15279 54040 -5191
2 2018 2(1) 32791 -8502 96600 42560
3 2169 151 37644 4853 1261440 29540
4 1873 -296 54761 4117 880 29540
19871 2719 846
.3146 82585 36824 2226 -386542 427 82140 -445 1314 -912
3 4144 998 33821 -48319 1446 132
4 4624 480 29993 -3828 3276 1830
19881 4859 235 35228 5235 2700 -576
2 3933 -926 42897 7669 2070 -630
3 3112 -821 33711 -9186 84fIJ 63904 2953 -159 44779 11068 7395 -1065
19891 7888 4935 47952 3173 2277 -5118
2 8330 442 49572 1620 5940 3663
3 7626 1296 47520 -2052 2140 -3800
4 13007 3381 93771 46251 4794 2654
1990 1 19765 6758 108212 14441 9442 4648
2 23543 3778 111246 3034 9584 142
3 27738 4195 116304 5057 8814 770
4 31058 3320 131305 15001 10200 1386
Sources: (1) Central Bank of Nigeria's Statistical Bulletin (Various years).
(2) Bank of Ghana, Research Department's 'Balance of Payments Docwnents
(3) IMF's Balance of Payme•nts Yearbook (various years)
UNIVERSITY OF IBADAN LIBRARY
144
billion which is only about 38% of its 1981 peak level.
Be that as it may, it is important to note that the level of RE by the late 1980s was
quite an improvement over what obtained in the last four years of fixed exchange rates (i.e
1982 - 1985). As for the balance of payments it was in deficit for eight of the twelve
quarters of 1981 -1983; and for four out of the eight quarters of 1985 - 1986. By 1987 the
balance of payments became positive and remained so till the end of 1990. Table 5.4 reveals
these facts.
The behaviour of the International Reserve level (RE) in Ghana is also of great
interest. Beginning from a low level of about ¢154 million (cedis) in the first quarter of
1977, the level of international reserves rose gradually until the second quarter of 1980,
thereafter it started a gradual decline until the onset of the Economic Recovery Programme
in 1983, when the level of reserves suddenly shot up by over 270% between 1982:4 and
..• 1983: 1. The rise in the level of RE continued up until the second quarter of 1987, which was
three quarters into the floating exchange rate system. Between 1987:2 and 1987:3 the level
ofreserves fell by over 142%. However, from 1988:4 the level ofreserves started an upward
climb again beginning with ¢44.8 billion in 1988:4 the RE level stood at some ¢131 billion
by 1990:4 an increase of over 290%.
It is important to note that as in the case of Nigeria the performance of the RE in
dollar terms may not be as remarkable as it looks if one takes cognisance of the enormous
depreciation of the cedi from 1983 when the ERP started. Hence a lot of the growth in the
•
UNIVERSITY OF IBADAN LIBRARY
145
level of reserves 'might simply be reflecting the extent of cedi exchange rate depreciation
rather than anything else. For instance the highest level of RE, pre-1983 was about ¢999
million (which obtained in 1980:2) if converted into US dollar at the then cedi exchange rate
of¢2.75 per $1, it becomes about US $363 million. However the value of RE after the onset
of ERP by 1983:4 of ¢1278 million only amounts to some $88 million at the on-going
exchange rate of ¢14.5 per $1. Be that as it may, the level of RE grew at an average quarterly
rate of 38.9% in the last 17 quarters before floating exchange rates, and at an average
quarterly rate of 13.7% in the first 17 quarters since floating. Note that the last 17 quarters
of fixed rates was characterized by what can be termed a jumping - peg system, which is as
flexible as any floating system except that it is administratively determined. Table 5.4
contains necessary information about Ghana's international reserves' behaviour.
As for the balance of payments it was more negative than positive for the period
1977 to 1982. In fact of the 24 quarters making up this period the balance of payments was
negative for 14 quarters. After the onset of the economic recovery program however, the
balance of payments position became positive, and it remained positive for the next 12
quarters except two. In the year following the onset of floating rates however the balance
of payments was negative for three of the four quarters of that year (i.e 1987). From 1988
however, up to the end of the study period (1990:4) the balance of payments remained
positive except for one quarter. This performance is quite remarkable and probably has
implication for the worthiness or otherwise of a floating exchange rate regime. This Ghanian
UNIVERSITY OF IBADA
 LIBRARY
146
performance compares favourably with the Nigerian situation where the balance of payments
had also been positive from the second quarter of floating exchange system till the end of
the study period.
In Uganda, the level of international reserves showed no discernible pattern as
between the era of floating rates and that of fixed one. Except for one exceptional quarter
during the floating exchange era when the level of international reserves skyrocketed it was
otherwise oscillating both under fixed and floating systems, within moderate amplitudes.
The performance of the balance of payments was also mixed. There were as many incidents
of negative balances as there were of positive onest both under floating and under fixed
exchange regimes. Table 5.4t columns six and seven reveal these points.
In Nigeria since the onset of the structural adjustment program (SAP) and its main
,
baby - floating exchange rate system, there has been a marked improvement in export
performance. In fact by the tenth quarter of floating exchange rates the growth rate of
exports as between the 9th and 10th quarter was about' 150%. Though there were still
negative growth rates of exports in some quarters after floating, these were not as marked
as before the floating exchange system. Before the floating system there were periods when
the growth rate of exports was as poor as - 60%. On the average, exports grew at a quarterly
average of 3.4% in the last 17 quarters of fixed rates (i.e 1982:2 - 1986:3) and at 27.8% in
the first 17 quarters offlpating rates (i.e 1986:4 - 1990:4) chart 5.2 bears testimony to these
facts.
UNIVERSITY OF IBADAN LIBRARY
147
As for imports chart 5.1 reveals its performance. Import growth improved markedly
after the onset of floating rates. This is a surprise in the light of the fact that one of the
purposes of floating the exchange rate is to ensure a depreciation of the naira which was to
help discourage imports. However, because of the easier access of actual importers to forex
in the era of floating rates and also less trade controls, import growth rate was sometimes as
high as 70%. All throughout the four years of floating period covered by this study, import
performance seemed to have improved compared to the last four years pre-floating. On the
whole imports grew at an average quarterly rate of - 1.2% in the last 17 quarters of fixed
rates and at an average of 16.7% in the first 17 quarters of floating.
Since import performance has been improving while export performance has also
been improving it is interesting to check for the net performance. Chart 5.3 reveals the share
of export as a proportion of GDP and the share of imports as a proportion of GDP. The gap
in these ratios reveal to what extent the external sector has improved or worsened since the
onset of floating rates. Chart 5.3 shows the last 17 quarters before the onset of floating rates
ie 1982:3 - 1986:3, and the first 17 quarters after floating i.e 1986:4 - 1990:4. This chart
reveals that imports as a proportion of GDP has shown a very gradual increase. Exports as
a proportion of GDP has had more fluctuations but had been consistently higher than imports
share in GDP. What is more, as the floating exchange system progressed export share seems
to be increasing thus the gap between export and import share in GDP has been a positive
and increasing one. On the average export share in GDP was 4.7% in the last four years of
UNIVERSITY OF IBADAN LIBRARY
148
fixed exchange rate system while import share in GDP was 3.5%, making a net excess of
exports over imports as share of GDP of 1.2%. However in the first four years of floating
exchange system export share in GDP increased to 8% while imports share increased to
3.8%. thus there is a net excess of export share relative to import share in GDP of 4.2% an
increase of over 200% compared with the fixed exchange rate era.
As for the total foreign sector share in GDP, this has also been growing as shown in
chart 5.4. Total foreign sector share in GDP is the sum of export value and import value
relative to the GDP. The total foreign sector share in GDP has markedly increased since the
.ninth q,uarter of floating rates i.e 1988:4. On the whole total foreign sector share in GDP rose
to 11.8% after floating, compared with 8.2% in the last 17 quarters of pre-floating era.
The performance of the external sector of the Ghanaian economy can like the
Nigerian case be examined further. Chart 5.5 below reveals the behaviour of exports both
pre and after floating. Export value recorded an average quarterly growth rate of 22.6% in
the last 17 quarters of fixed rates (i.e if one discountenances 1983: 1 and 1984: 1 when the
growth rates were clearly exceptional). Exports however grew at an
•
UNIVERSITY OF IBADAN LIBRARY
149
Chart 5.1
Nigeria: Import Growth
100 __ ---------------------------- __
50
,
-so
1 3 5 7 9 11 13 15 17 1921 23 2S27 29 31 33
Time: Fixed Exchange Rate Era Floating Exchange Rate Era
UNIVERSITY OF IBADAN LIBRARY
150
Chart 5.2
~--N--i-g-e-r-i-a-:~ ---E-x-p--o-r-t--G--r-o-w--t-h----
150
100
~ 50
-50
1 3 5 7 9 11 13 15 17 1921 23 25 27 29 31 33
Time: Fixed Exchange Rate Era Floating Exchange Rate Era
•
UNIVERSITY OF IBADAN LIBRARY
151
Chart 5.3
Nigeria: Export and Import Share In GDP
20 ~--------------------------------------~
15
5 •..•••••
1 3 5 7 9 11 13 15 1719 21 23 2S 27 29 31 33
Time: Pixed ExchangeRate Era Ploating ExchangeRate Era
..• XGDP ...-IMGDP
UNIVERSITY OF IBADAN LIBRARY
Chart 5.4
~ ~-----N-ig-er-ia-:-T-o-ta-l -fo-re-ig-n--E-x-p-or-t -S-ha-r-e-In--G-D-P-----
20
~oo IS
10
1 3 5 7 9 11 13 15 17 19 21 23 2S 27 29 31 33
Time: Fixed Exchange Rate Era Floating Exchange Rate Era
UNIVERSITY OF IBADAN LIBRARY
153
average quarterly rate of 9.75 % in the first 17 quarters of floating rates. One thing to take
note of, is that, the last three years of fixed rates coincided with the onset of the economic
recovery program, and even though the exchange rate was not made to float immediately
after the onset of ERP, it was a jumping-peg system hence the exchange system had a lot of
flexibility in the last years before floating. If one however compares what obtained before
the onset of the jumping-peg system (i.e when the exchange rate was permanently fixed)
then one can see that there was a marked improvement since the onset of a more flexible
exchange rate system. For instance in the period 1978:2 to 1982:4 the value of exports grew
I
at just a quarterly average rate of 6.5%. However at the onset of jumping - peg system this
growth rate rose to 22.6% and even after the euphoria of the adoption of ERP had died, the
growth rate of exports still remained higher than the pre-ERP era standing at 9.75%.
As for imports, its performance since floating has not been impressive when
compared with the jumping - peg era, but when compared with the permanently fixed
exchange rate era, it has been remarkable. Import value grew at a rate of 6.4 % in the era of
permanently fixed rates (1978: 1 to 1982:2) and grew at 12.4% in the era of floating rates.
In the jumping - peg era imports grew at 15.7%. Even though we do not regard the period
1983:4 to 1986:3 as a period of floating exchange rate per-se, the system of jumping-peg has
an in-built flexibility, t•he result of which can not be ignored. Chart 5.6 reveals the
performance of imports in the Ghanaian economy.
In order to know the net performance of exports we compared the share of export in
UNIVERSITY OF IBADAN LIBRARY
154
GDP to tIlilt of imports in GDP. Chart 5.7 shows the gap between these two ratios, and
shows that for the eight quarters of 1989 - 1990, import share in GDP surpassed export
share. For the earlier periods there was no disceinible relationship as these ratios fluctuated
such that sometimes export share suroassed import share in GDP, at times it is vice-versa.
Whatever be the case the total foreign sector share in GDP rose from a meager 1.5% in
1982:3 to 9.8% in 1986:3 and by 1988:2 it reached 19% as revealed in chart 5.8.
In trying to analyze the performance of the external sector of the Ugandan economy
under fixed and under floating exchange rates, we found that, exports value grew at a
quarterlyav age of about 10% in the floating exchange era 1982:1 - 1987:1 and at about
7% in the fixed era (i.e 1987:2 - 1990:4 except for the quarters 1987:3 and 1988:3, that were
clearly exceptional). As for imports value, it grew at a quarterly average of 14.6% in the era
,of floating rates and at an average of 19.7% in the era of fixed rates. This is to be expected
as exchange rate appreciation in the era of fixed rates encouraged more importation. Also
the improved performance of imports in the fixed exchange era may also reflect conscious
government efforts to promote imports, believed to be crucial to domestic industrial
production and export expansion. Charts 5.9 and 5.10 reveal the performances of exports
and imports in the era of floating and of fixed rates. On the whole the net performance is
such that in the era of floating rates share of exports
UNIVERSITY OF IBADAN LIBRARY
155
Chart 5.5
Ghana: Export rowth
600~---------------------------- __
500
400
300
200
100
1 3 5 7 9 11 13 15 17 1921 23 25 27 29 31 33
Time: Fixed Exchange Rate Era Floating Exchange Rate Era
•
UNIVERSITY OF IBADAN LIBRARY
156
Chart 5.6
~-----G--h-a--n-a-:--I-m--p-o--r-t-G--r-o-w--t-h---~
250
200
150
100
50
-50~~~~~~~~~~~_~~~~~~~
1 3 5 7 9 11 1315 17192123 2527 29 3133
Time: Fixed ExchangeRite Bfa Floating Exchange Rate Era
•
UNIVERSITY OF IBADAN LIBRARY
--~- --:::::::=-~------~
157
Chart 5.7
Ghana: Import and Export Share in GDP
12 ~------------------------------------~
10
8
4
2
,
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33
Time: Fixed Exchange Rate Era Floating Exchange Rate Era
-ll-XGDP ...•. IMGDP
•
UNIVERSITY OF IBADAN LIBRARY
-- -------~.-=-- ----
Chart 5.8
Ghana: Total Foreign Export Share in GDP
20 ~------------------------------------~
15
~
CoI 10
, 5
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 31 33
Time: Fixed ExchangeRate Era Floating ExchangeRate Era
UNIVERSITY OF IBADAN LIBRARY
159
Chart 5.9
Uganda: Export Growth
400 ~--------------------------------~
300
200
100
,
-100
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Time: Floating Exchange Rate Era Fixed Exchange Rate Era
•
UNIVERSITY OF IBADAN LIBRARY
160
Chart 5.10
Uganda: Import Growth
wOr-------------------------------~
IS0
100
so
-so
,
-100
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Time: Floating Exchange Rate Era Fixed Exchange Rate Era
UNIVERSITY OF IBADAN LIBRARY
161
Chart 5.11
Uganda: Export and Import Share in GDP
20 ~--------------------------------------
15
5
,
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Time: Floating Exchange Rate Era Fixed Exchange Rate Era
..• XGDP ..•..IMGDP
UNIVERSITY OF IBADAN LIBRARY
162
Chart 5.12
Uganda: Total Foreign Sector Share In GDP
35 ~------------------------------------~
30
25
20
IS
10
,
5
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Time: Fixed Exchange Rate Era Floating Exchange Rate Era
•
UNIVERSITY OF IBADAN LIBRARY
163
in GDP was consis ntly higher than that of imports in GDP. However throughout the period
of fixed rates the trend reversed and import share in GDP became consistently higher than
export share. On the whole the total foreign sector share in GDP fell in the era of ftxed rates.
Charts 5.11 and 5.12 reveal these points.
5.3 Domestic Prices and Real Income
The performance of domestic prices is another phenomenon worthy of investigation.
In Nigeria the quarterly growth rate of domestic prices except for two quarters when it
fluctuated widely. has not been markedly higher than what it was before the onset of floating
rates; while it was 4.2 % average in the last 17 quarters of fixed rates it was 6.1 % in the first
17 quarters of floating. Chart 5.13 shows the performance of domestic prices in Nigeria .
..•
The level of real income in Nigeria was highest in the first quarter of 1978. recording
a level of N683 million and a quarterly growth rate of over 60% in the first quarter of 1978.
By 1983;2 real income has fallen to as low as 38% of its 1978: 1 peak level and by 1984:2
it has even fallen further to 22% of 1978: 1. At the eve of the structural adjustment program
real income was stilt some 22% of its 1978: 1 peak. and stood at N151 million. By the fourth
quarter of 1986 real income showed considerable improvement recording a quarterly growth
rate of over 81% over the previous quarter. From 1987: 1 to 1990:4 real income hovered in
the neighbourhood of N 50 million to N400 million. Although it has not reached the N600
million pea of the 1970s it at least has not fallen below N200 million as obtained in the last
UNIVERSITY OF IBADAN LIBRARY
164
four years before floating. What is more - the average quarterly growth rate of real income
in the first seventeen quarters of floating exchange experience is 8.2% as compared with 3%
in the last 17 quarters of fixed exchange system. Chart 5.14 shows the behaviour of real
income in the last 17 quarters of fixed exchange system and the first 17 quarters of floating
experience.
,
•
UNIVERSITY OF IBADAN LIBRARY
165
Chart 5.13
Nigeria: Growth Rate of OomestJc Prices
40 ~----------------------------------~
30
20
, 10
o ~--------~~--+_-_---------------4~
1 3 5 7 9 11 13 15 17 19 21 23 2S 27 29 31 33
Time: Fixed Exchange Rate Era Floating Exchange Rate Era
UNIVERSITY OF IBADAN LIBRARY
166
Chart 5.14
Nigeria: Real Income Growth
100 __ -----------------------------,
50
,
-so
1 3 S 7 9 11 13 15 17 19 21 23 25 27 29 31 33
Time: Fixed Exchange Rate Era Floating Exchange Rate Era
UNIVERSITY OF IBADAN LIBRARY
167
The domestic price level (Pd) in Ghana grew at a quarterly average rate of 10% in the last
17 quarters of fixed rates and at 8.5% since the onset of floating rates; in the period before
the jumping-peg system (i.e 1978: 1 to 1982:2) domestic prices grew at the rate of 15.9%.
The decline in the growth rate of domestic prices both in the jumping-peg and in the floating
exchange era is quite surprising. From the theoretical literature one expects increased
depreciation of the domestic currency to generate increased growth rate of domestic prices.
However a possible explanation for declining price growth rate in the face of increasing
depreciation of the currency, is that the era of increasing exchange rate deprcci tion is also
one of increased productivity as trade controls are relaxed in the face of a floating exchange
system. Access to forex by actual importers is also easier. More importantly in the case of
Ghana (according to Fosu, 1990) there is the flow of structural adjustment loans and grants
all of which aid increased productivity. However domestic prices declined in the face of
.•
increased output. Chart 5.15 shows the rate of growth of domestic prices for the period
1982:3 to 1990:4. Except for a few exceptions the growth rate after the onset of floating
seems reasonable enough. Of course if translated into annual growth rates these rates are
quite high but relative to what obtained before the onset of ERP they are much lower. As for
real income, there was a marked improvement in its level after the onset of floating exchange
regime. Chart 5.16 reveals the performance of real income in Ghana.
UNIVERSITY OF IBADAN LIBRARY
168
Chart 5. 1.5.
Ghana: Growth Rate In Domestic Prices
70
60
50
40
30
20
~
10
, 0
-10
-20
-30
-40
1 3 5 7 9 11 13 15 17 19 21 23 2S 27 29 31 33
Time: Fixed Exchange Rate Era Floating Exchange Rate Era
UNIVERSITY OF IBADAN LIBRARY
169
Chart 5.16
100G--h-a--n-a--:--R--e-a-l--I-n-c--o-m--e---G--r-o-w--th
so
-so
1 3 5 7 9 11 13 15 17 19 21 23 2S 27 29 31 33
Time: Fixed Exchange Rate Era Floating Exchange Rate Era
•
UNIVERSITY OF IBADAN LIBRARY
170
The domestic price level in Uganda grew at a quarterly average rate of about 17.3%
in the floating exchange rate era and at 16.27% in the fixed era. Except for an exceptional
fall in 1988: 1 the rate of growth of domestic prices was not too far ahead under the floating
exchange system. A possible explanation for this close performance of domestic prices
during floating and fixed exchange regime was a jumping peg system whose flexibility was
almost as high as that of the floating system. Chart 5.17 shows the pattern of growth of
domestic prices in the Ugandan economy.
The behaviour of real income in absolute terms is shown in chart 5.18. Chart 5.18
reveals that real income grew at a quarterly average of 42.2% in the floating era and at an
average of 14.2% of the fixed exchange rate era. The behaviour of real income probably
reflects the fact that output expanded more in the era of floating rates, may be due to the
,
removal of rigidities in the whole economy in general as the era of floating rates was also
characterized by trade liberalization, less government control of public enterprises and
parastatals.
•
UNIVERSITY OF IBADAN LIBRARY
171
Chart 5.17
Uganda: Growth Rate of Domestic Prices
200 ~----------------------------------~
150
100
50
o ~~~----------------~---------=~~~
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Time: Floating Exchange Rate Era Fixed Exchange Rate Era
•
UNIVERSITY OF IBADAN LIBRARY
172
Chart 5.18
Uganda: Real Income Growth
700 ...--------------------,
600
500
400
300
100
,
-100
1 3 5 7 9 11 13 15 17 19 21 23 25 27 29
Time: Floating Exchange Rate Era Fixed Exchange Rate Era
•
UNIVERSITY OF IBADAN LIBRARY
CHAPTER SIX
MODEL ESTIMATION AND EMPIRICAL V LIDATION
"It is important to remember that mathematics is a
Tool not an End ill itself, and that if ideas are to
influence policy-makers they must be translatedfrom
equations into the English Language. "
Marris (1984)
6.1 Model Estimation
Each model was estimated in both the direct linear and the log-linear forms. After
estimation we solved the models in these two form and carried out simulation-error analysis
, to enable us determine which form (direct linear or log-linear) better tra ked our data. We
decided on the basis of the simulation-error analysis to utilize the log-linear form for further
analysis. Appendix 2 contains the result of the direct linear estimation.
In estimating we employed the Ordinary Least Squares (OLS) technique and
corrected for first order serial correlation where necessary. The models are block-recursive
and their properties make the OLS an appropriate estimation procedure (pindyck and
Rubinfeld, 1981). Below we discuss the log linear results on the basis of issues highlighted
in 4.1 above.
The first hypothesis we attempted to investigate is, that market determined exchange
rates fuel inflation. This implies that there is a significant pass through from exchange rates
UNIVERSITY OF IBADAN LIBRARY
- - - - --- --~---
174
to domestic prices which is heightened under market determined exchang rates. Table 6.1
gives an indication of whether this hypothesis holds or not. (Note that for all the tables the
figures in brackets are the 't' statistics).
TABLE 6.1
Estimation Result For Dependent
Variable· Domestic Prices (pD)
Country Dependent COIISla INDEPENDENT VARIABLES
Variable nt
LogRI Log Log Log Log It, Log Rl DW
MSS EN RD PD,.,
Nigeria LogPD 0.44 .0.22 0.22 0.05 0.06 0.13 0.74 .98 \.91
(0.98) (2.21) (2.36) (0.55) (2.26) (9. 0)
Ghana l..o&PD 0.39 .o.J5 0.19 0.02 0.17 0.18 0.85 .98 1.88
(·2. 6) (2.S6) (0.56) (2.25) C4.70) (\4.70)
Uganda l..o&PD 2.16 .0.17 0.16 0.11 0.49 0.10 0.82 .99 1.89
(·3.68) (4.70) (3.12) (3.96) (2.34) (l6.90)
Note the figures in bracket lire the 'r' Statistics.
From table 6.1 we see that in Nigeria the exchange rate of domestic prices is 5%, and
the relationship is statistically significant at the 5% level. However the money supply
elasticity of domestic prices is more pronounced, standing at 22% while the elasticity of
domestic prices with respect to inflationary expectations is 13%. All the relationships were
statistically significant at the 5% level. These results show that in Nigeria there is indeed
some pass through from nominal exchange rates onto prices, but the impact is not as
pronounced as that of mo ey supply or inflationary expectations or that of real income on
prices. Though real income had the effect of lowering prices whenever it (real income)
UNIVERSITY OF IBADAN LIBRARY
175
rises.
In Ghana there does not seem to be any pass through from nominal exchange rates
onto domestic prices, given an exchange rate elasticity of domestic prices of only 2% which
is also not statistically significant at the 5% level. Nominal money sup ly in Ghana,
however, impacted significantly on domestic prices at the 5% level and so does real income.
The money supply elasticity of domestic prices was 19% and real income elasticity - 35%.
Similarly, as in Nigeria expectations with respect to inflation influenced current prices with
an elasticity of 18% the relationship also proved to be statistically significant. Nominal
domestic rate of interest also proved to be a powerful propeller of prices in Ghana. These
Ghanaian results confirm the contention of Chibber and Shaffik (1989) that in Ghana it is
not so much nominal exchange rate depreciation that fuels inflation as it is reckless increases
, in money supply often engendered by reckless fiscal deficits.
In Uganda there is a remarkable pass through from exchange rates unto domestic
prices. The exchange rate elasticity of domestic prices in Uganda is 11% and this is
statistically significant at the 5% level. This result seems to contradict Elba wi (1990)
wherein he maintained that in Uganda it is excessive fiscal deficits that fuel inflation rather
than official exchange rate depreciation. It is however important to note that the impact of
nominal money supply on domestic prices is much more remarkable. The money supply
elasticity of domestic prices in Uganda is 49% and this is statistically significant at the 5%
level. Another varia le that seems to impact considerably on domestic prices is the domestic
UNIVERSITY OF IBADAN LIBRARY
176
rate of interest. One is therefore tempted to conclude that Elbadawi (1990) may be right to
the extent that both money supply and domestic rates of interest proved to be more powerful
propellers of prices in Uganda, than the nominal exchange rate. Then one might say
government monetary policy action which often is taken to finance its fiscal deficits is the
major source of inflation and not nominal exchange rates.
It is important to note that real income had an in erse relationship with domestic
prices for all three countries indicating that as output rises, prices fall; the relationship
though significant for Ghana and Uganda was insignificant for Nigeria, (at the 5% level).
The result for Ghana and Uganda uphold the position in the literature that the higher the
level of output the lower the price level would be (Dornbusch 1980, Edwards 1988).
The second issue we investigated is, that market determined exchange rates follow
the money supply. If this hypothesis is valid then monetary policy is extremely important
for efficient performance of a floating exchange rate regime. Table 6.2 reveals the result of
this test.
•
UNIVERSITY OF IBADAN LIBRARY
177
TABLE 6.2
Estimation Result For Dependent
Variable - Nominal Exchange Rate (EN)
Country Dependent Constant INDEPENDENT VARIABLES
Variable
LogRDF, Log RE,., logY, Log MSS, Rl DW
Nigeria Log EN -2.02 0.37 -0.40 -0.05 0.85 .95 2.10
(0.93) (,-4.13) (-0.34) (3.01)
Ghana Log EN -1.48 ·0.53 -0.50 -0.18 0.54 .85 1.95
(1.20) (-2.71) (-4.49) (2.04)
Uganda Log EN -1.64 O.32w -0.49 -0.13 0.37 .89 1.91
(1.41) (-4.75) (-3.52) (2.97)
In Nigeria nominal exchange rate seems quite responsive to changes in the money
supply. The nominal money supply elasticity of exchange rates is 85% and this relationship
,
is stati tically significant at the 5% level. Another variable that impacted significantly on
nominal exchange rates is international reserves (with a one period lag). The reserve
elasticity of exchange rate is 40% in Nigeria. The relationship is inverse indicating that the
higher the level of reserves the lower (or more appreciated) the exchange rate becomes.
Nominal income however is insignificant in Nigeria in the determination of nominal
exchange rates. Both the Ghanaian and Ugandan results follow the Nigerian pattern in this
respect. The Ghanaian and Ugandan results indicate that indeed the nominal exchange rote
follows the money supply. The money supply elasticity of nominal exchange rates is 54%
•
in Ghana and 37% in Uganda.
UNIVERSITY OF IBADAN LIBRARY
178
These results show that prudent and disciplined monetary policy is very crucial to
the success of a floating exchange rate regime in these countries. One interesting point to
note is the impact of interest rate differential on nominal exchange rates. The theoretical
expectation is that the higher the interest rate differential (i.e domestic rate of interest minus
foreign rate of interest) the more appreciated the domestic currency becomes, because a
higher domestic rate of interest relative to (the foreign one) should encourage capital inflow,
from the three countries this relationship turned out with the wrong sign (positive instead
of being negative) and was also statistically insignificant at the 5% level . The probabl
reason for this situation is that in the literature it is often assumed that domestic securities
are peffect substitutes for foreign securities, so that a higher domestic rate of interest should
increase investment in domestic securities by foreign residents. However, in reality, the
'securities of our focus countries are not considered as perfect substitutes for foreign
securities, so that no matter how high domestic rate of interest gets it is not likely to generate
any appreciable capital inflow and hence no exchange rate appreciation will follow a rising
domestic rate of interest as theory postulates.
A third hypothesis we looked at is that altering the structure of relative prices in the
export sector by depreciating nominal exchange rates will generate in reased export value
given a two period lag (Khan 1981, Jebuni et al 1991) . Table 6.3 reveals the result of this
test.
•
UNIVERSITY OF IBADAN LIBRARY
179
TABLE 6.3
Estimation Result For Dependent
Variable - Export Value (X)
Country Dependent Constant INDEPENDENT VARIABLES
Variable
Log RIS, Log ETOT'.l Log IM,.I Log Rl DW
DPX,.l
Nigeria Log X, -49.41 0.70 -1.01 0.38 1.69 .92 2.10
(3.79) (-2.57) (3.85) (3.80)
Ghana Log X, -10.10 0.77 0.43 0.78 0.28 .96 1.90
(5.55) (4.31) (2.14) (2.23)
Uganda Log X, -8.04 0.67 -0.78 0.56 0.57 .93 2.10
(5.32) (-4.92) (5.80) (5.72)
From table 6.3 we found that the elasticity of exports with respect to changes in the
i'elative export price (ETOTt•2) in Ghana is 43% and this is statistically significant at the 5%
level. This reveals the fact that altering the structure of relative prices in the export sector
given a two period lag can be quite important in increasing export value. For Nigeria and
Uganda however the result came out with the wrong sign (Negative instead of Positive), The
suspicion here is that for these two Countries (Nigeria and Uganda) a two-period lag is not
the appropriate lag structure, probably some other lag structure will elicit the appropriate
respon,se.
Another way by which we tried to see the impact of nominal exchange rate
movement on exports is th•rough the actual price of exports in domestic currency terms given
UNIVERSITY OF IBADAN LIBRARY
180
a two period lag (DP~-2 ). As shown in table 6.3 the elasticity of exports with respect to its
domestic price given a two period lag is 169% for Nigeria. 28% for Ghana and 57% for
Uganda. The relationship was statistically significant at the 5% level in the three countries.
This shows that nominal exchange rate depreciation which enhances income in the export
sector. encourages increased export supply. It is important to note that real income had a
positive and significant impact on exports in the three countries.
A fourth issue we investigated is. that. imports are crucial to export performance. so
that compressing imports leads to reduced exports and expanding imports. leads to increased
exports as well. Table 6.3 also contain the required information. For the three focus
countries import impacted significantly on exports (at the 5% level). The import elasticity
of exports is 38% in Nigeria, 78% in Ghana and 56% in Uganda.
"
The fifth hypothesis considered is. that. changing the structure of relative prices in
the import sector in favour of the domestic goods. leads to declining imports. (Machlup,
1939. 1940; Alexander 1959.1961; Tsiang 1961. Sohmen 1965; Dornbusch 1980; Mussa
1976). Table 6.4 contain the relevant information on this score .
•
UNIVERSITY OF IBADAN LIBRARY
181
TABLE 6.4
Estimation Result For Dependent
Variable - Import Value (1M)
Country Dependent Constant INDEPENDENT VARIABLES
Variable
RDM, LogRI, RPIM Log DRE, Rl DW
Nigeria Log 1M, -1.15 0.37 0.33 0.12 0.28 .86 1.99
(3.34) (3.14) (0.50) (3.47)
Ghana Log 1M, -0.16 0.25 0.30 0.03 0.41 .98 1.89
(2.84) (3.45) (0.62) (2.10)
Uganda Log 1M, -16.07 0.28 0.84 -3.11 O.SS .93 1.97
(2.63) (11.34) (-1.39) (2.58)
For Ghana and Nigeria the relationship between imports and relative prices (RPI:M)
, turned out with the wrong sign. Even for Uganda wherein the relationship turned out with
the right sign, the relationship proved to be statistically insignificant (at the 5% level). It was
also statistically insignificant for Nigeria and Ghana too. However when we looked at the
real income elasticity of imports (also from Table 6.4 ) we found the relationship to be
statistically significant and with the right signs tQP; The real income elasticity of imports
is 33% for Nigeria, 30% for Ghana and 84% for Uganda. TIllS result goes to confirm the
contention of Osagie (1985) and Ajakaiye (1985) that for developing countries like Nigeria,
altering the relative price structure in the import sector is not likely to affect imports,
•
because import in these countries depends more on real income than on prices: One possible
UNIVERSITY OF IBADAN LIBRARY
182
reason for this observed insensitivity of imports to relative price change is the complete
dependence of almost all sectors of developing economies on imported inputs and
machineries, so that irrespective of rising relative import prices certain level of imports has
to be maintained otherwise these dependent economies would completely collapse.
That imports in developing countries also largely depend on import-capacity rather
than prices (Hemphill,1974; Moran 1984, Ndulu 1991) is the sixth hypothesis we attempted
to validate from Table 6.4. Taking the change in the level of international reserves as a
proxy for import-capacity we tried to find the nature and degree of relationship between
imports and change in international reserves. We found that the elasticity of imports with
respect to changes in the international reserve level is 28% for Nigeria, 41% for Ghana and
58.% for Uganda and the relationship is significant at the 5% level for all the countries. This
result has implications for export promotion as shown in the next hypothesis.
The seventh hypothesis we investigated is that, net exports is a major determinant
of import-capacity. We investigated this hypothesis through the relationship of net exports
( or trade balance) to the international reserve level as shown in Table 6.5. Table 6.5 shows
that the trade balance elasticity of reserves is 28% for Nigeria, 36% for Ghana, and 44% for
Uganda (given a one-period lag in the trade balance i.e. TBI_I). The relationship is
statistically significant at the 5% level in all the three countries. These results confirm that
net exports is a crucial dete inant of import - capacity to the extent that it impacts reserves
significantly and the change in reserves is according to hypothesis six, a major determinant
UNIVERSITY OF IBADAN LIBRARY
183
of imports. If we combine the results of hypothesis six and seven with that of hypothesis
four, it becomes obvious that the relationship between imports and exports is a cyclical one
and it could be a virtuous or vicious circle depending on the take-off point. So, it is
important for policy makers to take cognisance of this interwoveness in designing policies
that affect exports and imports.
TABLE 6.5
Estimation Result For Dependent
Variable - International Reserves (RE)
Country Dependent Constant INDEPENDENT
Variable VARIABLES
2
, EMSSI Log Log R DW
GDI.• TBI.•
Nigeria Log REI 1.21 -0.24 0.33 -0.24 .91 1.98
(-2.48) (3.14) (-2.48)
Ghana Log REI 1.48 -0.58 0.30 -1.10 .86 1.88
(-4.62) (3.45) (-2.56)
Uganda Log REI 0.12 -0.32 0.84 -0.41 .81 1.87
(-2.36) (11.34) (-2.36)
•
UNIVERSITY OF IBADAN LIBRARY
184
TABLE 6.7
Estimation Result For Dependent
Variable - Real Demand For Money (RDM)
Country Dependent Constant INDEPENDENT VARlABLES
Variable
LogY, Log LogRD, Log Rl DW
GJ; RDM, .•
Nigeria L-ogRDM, 3.75 0.02 0.17 -1.26 0.25 .89 1.99
(0.27) (2.72) (-2.03) (2.03)
Ghana LogRDM, 2.59 0.54 0.21 -1.59 0.54 .86 1.95
(5.48) (2.83) (-2.86) (6.92)
Uganda LogRDM, 3.75 0.59 0.18 -2.01 0.57 .96 1.93
(2.23) 2.77 (-3.24) (6.34)
That government deficit is a major source of excessive money supply increases is
,indirectly revealed in table 6.6 where the money supply is significantly impacted by
government deficit. The government deficit elasticity of money supply is 76% for Nigeria,
85% for Ghana, and 71% for Uganda all are statistically significant at 5% level.
The money supply on the other hand significantly impacts the international level as
revealed on table 6.5 where there is and inverse relationship between excess money supply,
and the reserve level. This shows that often reserves drawn down are monetized thus
increasing the money supply, and often is for purpose of financing government deficits.
The drawing down reserve to finance government deficit is a major source of money
supply increases, and, gi en the extent of passthrough between money supply and domestic
UNIVERSITY OF IBADAN LIBRARY
185
prices as shown under the first issue discussed, it follows that fiscal irresponsibility is a
major source of inflation in these three focus countries. That government spending i a prime
determinant of economic activity in the focus countries is shown in table 6.7 where the
government expenditure elasticity of the real demand for money is 17%,21 % and 18% in
Nigeria, Ghana and Uganda respectively, all statistically significant at the 5% level.
Table 6.8 reveals that the main determinant of the domestic rate of interest in
Nigeria and Ghana is the money supply. When there is excess money supply the domestic
rate of interest falls. Thus the relationship between the domestic rate of interest and excess
money supply is negative and significant at the 5% level for Nigeria and Ghana. Also
significant for the domestic rate of interest in Nigeria is the inflationary expectations and the
past level of the rate of interest. for Ghana the inflationary expectations impacts remarkably
on the domestic rate of interest but the level of the rate seems to have no bearing on its
present level. In Uganda the supply seems to have no significant impact on the domestic rote
of interest but nominal income and inflationary expectations do.
In Table 6.9, real income is shown to respond insignificantly to demand pressures
in Nigeria and Ghana. This probably reflects the fact that there are structural rigidities in
these economies that make it difficult for output to increase promptly and in proportion to
increased demand. He ce the relationship between real income and excess real demand for
money came out positive but insignificant in the two countries. Imports however impacted
significantly on real income in the three countries, while the differential between previous
UNIVERSITY OF IBADAN LIBRARY
186
real income and the mean real income also proved significant. In Uganda, demand pressure
impacts real income significantly given an excess real demand for money elasticity of real
income of 68% which was significant ant 5% level.
In all we can conclude that for Nigeria and Uganda there is indeed some degree of
pass through between nominal exchange rates and domestic prices. However even in these
economies nominal exchange rates are not the most forceful in propelling domestic prices.
rather the money supply and inflationary expectations are. In Ghana nominal exchange rates
do not seem to impact domestic prices at all. the most important explanatory variable in this
respect is the money supply. Also market determined exchange rates seem to follow the
money supply. This implies that exchange rate depreciation occurs with undue increases in
the money supply. Therefore there is need for prudent and disciplined monetary policy for
the success of a flexible exchange rate regime. As for our ninth hypothesis ie. that
macroeconomic indicators improve on the adoption of a floating exchange rate system we
need the results of the simulation exercises to be able to test this hypothesis. Hence this issue
shall be discussed in the next chapter .
•
UNIVERSITY OF IBADAN LIBRARY
187
Table 6.8
Estimation Result For Dependent
Variable - Domestic Rate of Interest (RD)
Country Dependent Constant INDEPENDENT VARIABLES
Variable
Log LogTI, LogY, LogRD"1 R2 DW
EMMS,
Nigeria LogRD, -1.98 -0.29 0.06 0.02 0.52 .96 1.99
(-2.67) (2.66) (0.43) (4.15)
Ghana Log RD, 2.07 -0.16 0.19 0.15 0.05 .80 1.93
(-2.02) (1.22) (2.18) (0.04)
Uganda Log RD, -0.02 0.05 0.54 0.54 0.16 .86 1.97
(1.61) (3.48) (3.48) (2.75)
Table 6.9
Estimation Result For Dependent
Variable - Real Income (RI)
Country Dependent Constant INDEPENDENT VARIABLES
Variable
Log Log RISD"I LogPD, Log 1M, R2 DW
RD~
Nigeria Log RI, 0.02 0.11 0.59 -0.32 0.31 .88 1.93
(0.65) (3.24) (-2.50) (3.39)
Ghana Log RI, 3.96 0.02 0.27 -0.15 0.16 .85 \.97
(0.39) (4.11 ) (-2.42) (2.67)
Uganda Log RI. 12.27 0.68 0.85 -0.18 0.55 .87 1.95
(4.40) (2.65) (-2.19) (3.82)
•
UNIVERSITY OF IBADAN LIBRARY
188
TABLE 6.6
Estimation Result For Dependent
Variable - Nominal Money Supply (MSS)
Country Dependent Constant INDEPENDENT VARIABLES
Variable
Log RI, LogTDC"l Log DIU; Rl DW
Nigeria Log MSS, 0.19 0.22 0.76 0.14 .89 1.98
(3.14) (16.90) (3.15)
Ghana Log MSS, 0.62 0.07 0.85 0.23 .95 2.01
(0.21) ( 16.28) (2.75)
Uganda Log MSS, 8.37 0.16 0.71 0.31 .95 1.97
(5.99) (13.43) (3.12)
, 6.2 Empirical Model Validation
We solved each of our models using the Time Series Processor (TSP) econometric
software (Version 4.0) developed by Hall in 1983. When we solved each model using the
Gauss-Seidel iterative technique (Ortega and Rheinboldt, 1951) each converged for each
endogenous variable and for each year. This demonstrates that each model is internally
consistent. Below is the result of the model validation exercise for each country.
The root mean square simulation errors for all the endogenous variables in the
Nigerian model are as shown in table 6.10. The Theil's inequality coefficient as well as its
decomposition for all the endogenous variables are shown in table 6.11.
UNIVERSITY OF IBADAN LIBRARY
189
Table 6.10
NIGERIA: ROOT MEAN SQUARE SIMULATION ERRORS
Dependent Variable RMSE
Domestic Prices 0.0600
Real Demand for Money 0.1030
Money Supply 0.1600
Domestic Rate of Interest 1.5300
Exports 0.2400
Imports 0.2800
International Reserves 1.6480
Nominal Exchange Rates 0.5200
Real Income 0.2240
Table 6.11
NIGERIA: THEIL'S INEQUALITY COEFFICIENT AND ITS DECOMPOSITION
Dependent Theil's Bias Variance Covariance
Variable Inequality Proportion Proportion Proportion
PD 0.00014 0.000057 0.00798 0.99145
RDM 0.00050 0.002070 0.06521 0.93272
MSS 0.00027 0.000100 0.04093 0.95897
RD 0.16399 0.142300 0.11426 0.74344
X 0.00084 0.000030 0.06558 0.93439
1M 0.00124 0.000080 0.02775 0.97224
RE 0.26309 0.171990 0.16397 0.66404
EN 0.04248 0.002220 0.00330 0.99448
RI 0.00149 0.000020 0.88370 0.91161
UNIVERSITY OF IBADAN LIBRARY
190
For six of the nine endogenous variables the RMSE is less than 0.3, while it is about
0.5 for one of the variables. For the two other variables the RMSE is greater than 1, i.e RD
and RE respectively. And if we compare the value of the RMSE with the average size of the
variables in question we get SRMSE of 0.1228 and 0.0003 respectively which indicate that
the error is marginal. As for how well the model captures turning-points, charts 6.1 to 6.0
reveal that the Nigerian model has performed well in this respect. The Domestic Price Chart
reveals that often the simulated values of this variable exactly equal the actual values and
all turning-points are perfectly tracked. The same goes for all the other endogenous variables
too.
The values of the Theil's Inequality coefficient are small with the largest being
0.26039. The smallness of these values indicate the ahsence of systematic bias. The
decomposition of the Theil's inequality coefficient into its component parts indicate that
only a small part of the error is due to bias for all endogenous variables except for domestic
rate of interest and international reserves. But given that the value of the Theil's inequality
coefficient for each of these variables is itself small (i.e. less than 0.3) then the fact that 0.17
of the error is due to bias as in the case of ~E, should not be too much cause for worry.
Generally the distribution of the errors is in favour of the covariance proportion for most of
the variables computed, which indicates the degree of unsystematic error. Since the
covariance proportion represents the remaining error after deviations from average values
•
and average variabilities have been accounted for, its accounting for the highest proportion
UNIVERSITY OF IBADAN LIBRARY
191
of the errors shows that the model on Nigeria performed beautifully as systematic error is
minimal.
As for Ghana the root mean square simulation errors for all the endogenous variables
in the Ghanaian model are as shown in table 6.12.
Table 6.12
GHANA: ROOT MEAN SQUARE SIMULATION ERRORS
Dependent Variable RMSE
Domestic Prices 0.1278
Real Demand for Money 0.1850
Money Supply 0.1334
Domestic Rate of Interest 0.0649
Exports 0.2668
Imports 0.2195
International Reserves 0.3065
Nominal Exchange Rates 0.0845
Real Income 0.1850
The root mean square simulation error is quite small for all the endogenous variables.
The largest value of the RMSE is 0.3065 for RE, when we compare this RMSE to the
average size of the variable in question (i.e International Reserves) we get an SR¥SE of
only 0.00011, which shows that RMSE of 0.3065 is not too bad. The low values of the
RMSE show that the simulated variables, did not depart significantly from the actual time
path of these variables• thus revealing that our model was able to track the underlying
UNIVERSITY OF IBADAN LIBRARY
192
economi realities.
As for how well the turning points are captured charts 6.7 to 6.12 depict how well our
model did this. The charts reveal that all movements in our endogenous variables are almo t
perfectly tracked by our model.
The Theil's inequality coefficient as well as its decomposition are shown in table
6. 13 for all the endogenous variables.
Table 6.13
GHANA: THEIL'S INEQUALITY COEFFICIENT AND ITS DECOMPOSITION
Dependent Theil's Bias Variance Covariance
Variable Inequality Proportion Proportion Proportion
PD 0.0004 0.0010 0.0025 0.9965
RDM 0.0021 0.0023 0.0273 0.9724
, MSS 0.0056 0.0000 0.0031 0.9969RD 0.0004 0.0367 0.4282 0.5353
X 0.0017 0.0001 0.0095 0.9904
1M 0.0006 0.0008 0.0031 0.9961
RE 0.1154 0.0444 0.1284 0.8338
EN 0.0002 0.0011 0.0002 0.9987
RI 0.0010 0.0011 0.1001 0.8989
For all the endogenous variables the Theil's inequality coefficient is quite small the
largest being 0.1154. Also the error proportions arein favour of the covariance proportion
implying that whatever error is observed is not systematic, so that just as in the case of
•
UNIVERSITY OF IBADAN LIBRARY
193
Nigeria our model of the Ghanaian economy has performed well in capturing underlying
realities f the Ghanaian economy.
For Uganda the root mean square simulation errors (RMSE) for all the endogenous variables
in the Ugandan model are as shown in Table 6. ]4 below.
Table 6. ]4
Uganda: Root Mean Square Simulation Errors.
Dependent Variable RMSE
Domestic Prices 0.1308
Real Demand for Money 0.4855
Money Supply 0.1324
Domestic Rate of Interest 0.1211
Export 0.0704
Import 0.0636
International Reserve 0.3973
Nominal Exchange Rate 0.4258
Real Income 0.0759
The RMSE is quite small for all the endogenous variables indicating the absence of
systematic error in the model. We constructed charts of the actual versus simulated
behaviour of the variables. Charts 6.13 to 6.18 reveal the extent to which our model tracks
reality and show that our model of the Ugandan economy is an adequate reflection of reality.
Also we used the Theil's inequality coefficient and its decomposition to measure the
predictability of our model. Table 6.15 below reveals the result of the exercise .
•
UNIVERSITY OF IBADAN LIBRARY
194
TABLE 6.15
Uganda: Theil's Inequality Coefficient And Its Decomposition
Dependent Theil's Bias Variance Covariance
Variable Inequality Proportion Proportion Proportion
PD 0.00022 0.00039 0.0033 0.99620
RDM 0.01722 0.00009 0.00320 0.99671
MSS 0.00014 0.00033 0.0033 0.99630
RD 0.00122 0.00453 0.07438 0.92109
X 0.2151 0.04080 0.0176 0.94150
1M 0.23880 0.04670 0.0115 0.94170
RE 0.00225 0.00001 0.09180 0.90819
EN 0.00538 0.00005 0.04930 0.95065
RI 0.08260 0.03966 0.01945 0.94090
Table 6.15 reveals that our model is able to accurately predict reality. The highest
value of the Theil's inequality coefficient permissible is 0.3 while the highest value in our
model is 0.23880 for imports, and the bias proportion is also very small. The covariance
proportion is highest indicating that any available errors at all are unsystematic. Results of
the simulation exercises are discussed in chapter seven .
•
UNIVERSITY OF IBADAN LIBRARY
19 S.
CHART 6.1
Nigeria: Act.ua l.'Si ni u l a ted Chart.
Dow.estic Pcice{PD).
800
700
600
500
400
300
200
100 •
o ~T~rn111TTr2' 4 2 4- 2 4- 2 ·1 2 4 2 4 2 4- 2Tr~
1979 :1000 :1003 :1003 :1003 :too.1 :1003 :1006 :100'3 :1006 :1009 :1Q90: 1:·3
o .\ctual + Simulated
;
UNIVERSITY OF IBADAN LIBRARY
•... ••....... •.... •.... ••....0 Iv w ...
0 0 0 0 o
I I
. /; .--'.
ilc
I-' .
.. •...
:::~'-
t5 I'J
t
•.
.O
...
~ •
.. •.....
55N
C..Il •• ~.•.::.. ..•... •...
65
e
:..
.
;.
J•N'..
".-
t.o N
.O.J ••.•..
~~
10'
• SQ*
UNIVERSITY OF IBADAN LIBRARY
Iv W •••. <J'Ic C'Io o 0 o "'1C C
----'-T-----r: -------,----,I----~
r o:»I:
",. ;:0
o ,..- -f '-',-
Q'\
g ';~-.: ;..~ .•.• -J (.oJn I'"
W.. -'-'-:.... ~r C,,. I-r_
~7v ~r.
+ -"-"' .•. 3'''''''-c
i~ r-~U-:I8~'"C....... ;-'--'.•.... .--
'~~
,
-. Iv
:..J •• '.~•...
•
UNIVERSITY OF IBADAN LIBRARY
~\~. 198
CHART6.4 i
Nigel'j a: Irn po r t.s (I~f)
ACTUAL/SI~ULATED I
J :\- --- --. ---- -.- .- -------. -.------ ----------------- I
12
11
10
9
--- 8If.
":~::, 7Ul
..oc: 6
~
'-' 5
4
/~f~\~. ~~ \~ ,.,
J "~•t..4.".'L;f::/ t3~"-tb~':2 ",+ ;Ge-r'"f/'sV,: ;'
o I Ii' i I I , I Iii ii' i I i Iii r i j i , Iii Iii I i Iii i , i I I I , I I •
o Act.ual Simulated .-
;
UNIVERSITY OF IBADAN LIBRARY
..:.. ".J 199
CHART 6.5
Nigeria: Exports X.
ACTUAL/SH'iULATED
SO r
I
45
40
35
,......
'" 30"0
~
r.!
III 25
;j
..0c:
E-< 20
15
10
"
:; •
~~~
0
Il Act.ua l + Simulaled
UNIVERSITY OF IBADAN LIBRARY
200
CHART 6.6
Nigeria: Heal Inc<.:1118 (HI)
ACTUAL/sn"ULATED
600
550
500
450
400 ~\
350 ~~
300
I
250 I
200
150
100
1978 n-rrrn ~:!~J ~ J l J 2 JTrmTTITITrrT~TrrrrTrrrrrrr;-J~:4
1979 :lUEIO :Hillll :19.83 :19B:i :11).8.1:100~ :19.8<3 :1VIl'J :1008 :1{}.39 :1n90 : 1: :j
o Actual + Simulated
UNIVERSITY OF IBADAN LIBRARY
201
-, CHART 6.7
Ghana: Act.u aL'Sirn u la t.ed Char Ls
Domestic price (Pd )
4.5
4
3.5
3
'"ff.'
"0
c: 2.5
c::
Ul
;:!
..0c: 2
f'<
'-"' 1 15'"
~,
151 £11' /r K~ij1_!ng'1 dr~~'~
0.5 ~ II-'
I ~ - ii •
01 _,I I ',I ' I ' I ' I ' f1TTTTTTT I ] 1 I 1 ]1 I I Tl ,I I 1 I I I'Ti il' I-T'TT--,-J
19 (8 : -i~ 2 ~ 4 ~ 2:1 4 ~ 2 ~ 4 :1 2:1 -1 ; 2 _: 4 :, 2, 4 ~ 2 4 ~ 2:: 4 :, 2:, 4 :: 2 ~ 4 2:: oj :, 2: .:1
1979 :1000 :1001> :1913.3 :11}133:11}13a :1985 :1118 :lllil7, :11).38 :H){39 : RC'lG : 1: 3
C Aclual + Simulated
UNIVERSITY OF IBADAN LIBRARY
202
.I. ".,'
CHART 6.8
Ghana: Real Dcru a n d For Money (RD~I)
ACTUAL/SIMULATED
120 •.-----------
110 f- er
100 f- \
\
90 1 1
80
70
60
•
50
.to V '\\'\ JfaB.30 1':1 ")," •.·a~7
20
1; .1
o Act ua l + Simulated
UNIVERSITY OF IBADAN LIBRARY
203
CHART 6.9
Ghana: Money Supply OvlSS)
ACTUALI/SIMULATED
280 '
260
240
220
200
180
,...,
If.
"0
c: 160
d i
III
:J 140 i
.t0:;
E 120•..-.<....
100 jye
80 1
::~ I'I~~~.
i
~
20 ~~..eEi~!j3T~a~~ Ii
o '~I jIfTTTl---;-1 'J " 'I "---;-1 'I 'J "'1---;-' 'I -., TIT' i ' 1 'TIl' J .1978 . 4: 2 4 < 2:, 4 ~ 2 ~ 4 ~ 2:1 4 , 2:' 4 ;, 2, -1:! 2:, 4 2' 4:! 2:; 4 ~ 2:~ 4 2:! -1 :! 2: 4
19f'9 :11}110:l00ll :19.83 :H!l33 :11l8J :11.}B'3 :111136 :1P[l2- :lHB8 :19H9 :11);90 : 1: 3
fJ Actual + Si m ul a t.ed
UNIVERSITY OF IBADAN LIBRARY
204
o·
CHART 6.10
Ghana: Irn po r t.s (11\·1)
ACTUAL/ SI I~ULATED
t,
l
~
I
~~. r~ •
[I. l~~
~a-m· • I
'1 . tFPl¥ffi+f' EfE''f'lmlY~~.Q~n~~.~j ~I I~I ~I III IUI TI InI 1JJ Tt IGI ,T. ~I IT[i D~-~iiI
:l Aclual + Sim u la t.ed
,
UNIVERSITY OF IBADAN LIBRARY
:: ':" -; 205 I
CHART 6.11 I
Ghana: Exno r t.s rX)
ACTOAL/SIMU LAt~D
110
100
I
90 I
80
70
'"Ul'
'1j
.:: 60
to::
Ul
0='
~ 50r...
'-'
40 I
30 I •
20 ~ pt'
10 ~ o ,.,1"'.1
L: -rr:r 7;~~'¥o 9i;; 9 T'i"71 ~ - I 'rn ~ T -pr~1>~- ..1 2 ~ .t 2 J 4 2 ~ 4 : .2:, 4 4.2 4: 2: 4 2 4nr:r
1979 :1000 :H}Bl! :11).83 :H1{33 :19.8J :1003 :1008 :1U8~ :19118 :1009 :1R90 : 1: 3
(] ACt.IWJ + Si m ul a tvd
UNIVERSITY OF IBADAN LIBRARY
. ,',;
206
CHART 6.12
Ghana: lnternational Re.ser'v es (RE)
ACTUAL/5 H1ULATSD
L50 rl----------------------------------------------------------------~ I
lclO
130 t
L20 r;/qJ
11U F;;,.I
100
90
80
70
60
50
·10 .
30
20 t~1~ . •
10 r-orf ~ I
oI~m~-rrl ~ ..". --' - , ~ Z -l 2 <12rrr~Trrrrrr~--I -.
L979 :LOOO:1003 :1003 :1003 :100;1 :100'3 :1006 :100<3:1008 :19£Hl:U190 : 1: 3
o .\ctnnl + Simulated
UNIVERSITY OF IBADAN LIBRARY
·:07
CHART 6.13
Uganda: Act.u a LzSim ul at.e d
Domestic prices (PD)
80 ,.---------------------------------------------------------------~
70
! '
60 I
i
i
,.... 50 I
Ul
'0
~~
Ul
;l 40
.o
!
c
E-<
'- I
30 l! I
I
el' I20 .
II ~/A-d ' I
10 ~. &~- ..., jl •
! ~~/.
o , . -~, ,.. - ""~L- lF~F-~=@~o==p='1=~~ - ,- 'I 1 I 1- - .
19U31~I34 :11 :1ros .l: j~ jb~;!1 1~B~ :11 :lttlo ~T
:2 :2 :4 :2 :4 :2 :4 :2 :4 :2 :4 :2 :4 :2 :4
'! Ael ual Simulated
i
I
UNIVERSITY OF IBADAN LIBRARY
•.... w .•.. 0• 0'. --J
0 ~ 0 0 0 C 0
0 0 0 0 0 0 0 c
.- I I I r-----l
-0
CO
W•... Clfq
~
~
~ ,r.-~~ .
.-
~
.. ~
:J ./'~
_~_-~o--l f2;C'~D o):. :
() ~~ >
::c-
:c- "
;J w- ~1f7 ';;;~ :;:0H -f~V l:0\
~(D
r ~~ +=
O~
(j) ~("1'
~
r- :
- o
:J,
~
"J
I-:rJ
0
~
IV
'-::/' 0CO0
~
CD
~
loJ
UNIVERSITY OF IBADAN LIBRARY
... ...• 209 '" .
CHART 6.15
T-T bIUT .'}·11d a .. Iru po iLs (I~[)
ACTUAL/SIMULATED
80 I
I
..•.
70 1 1
60
50
'"U"l '
'0
~r.i
:P. 40
.0=a'
~
'-' 30
I
20 I- /"
~'
~~
10 .
0
I L __ I L I L_
i 9n9 :11 i{nO :11
;2 :0: 4 :2 : 4 :2 :4 •
0 Actual + Simulaled
UNIVERSITY OF IBADAN LIBRARY
_..-. 210
CHART6.16
UgdlHla: Exports X
ACTUAL/ SJ t·~IILATED
90
I
80 . I
70
60
,-...
Vi
't:
~ 50e
Vi
~
..c0:: 40
f-<
'-'
30
20 r- ei
I
10
l I0
I 1-9-83-I1-,I --I ,I--I~_Ir-I,rI-~I ~I--I~T-I r~~~I~b1f,-~--~- <W~~~~~F?=T~i·?,'~h,~,:~2l,~__ J: ~4,~&'•~'~l'~~':;h'~'~l'~l~'~bb
o Actual + Sim ul a tr-d
UNIVERSITY OF IBADAN LIBRARY
211
CHART 6.17
£? a n d a : Rea 1 Ll co III e (R r )
- ACTUAL/Sn:ULATED'
2" ~ ---I
2.6
I
2.4
2.2
2
1.8
""' \
III \' It~ t
'tl 1.6
~~
III 1.4
;:l I ~~~ II '~/".
.0a
E-< 1.2
'-"
1
0.6
r tI
\~~ -
0.6 g
0.4
0.2
0
•
,-' Act.u a l SimuJa t e d
UNIVERSITY OF IBADAN LIBRARY
212
l {~c~.} 1 li ~; : Nun 1.in <31 Ex C:11 d }1g C:' 1.': l \ \.
7,CTliP.L/ SJ !~ULhED '
1.8
1.7
1.6
1.5 /
1.4
1.3 II
1.2
,-.. 1.1
fIj
't:i
I=i 1 jJ \
It
fIj
;:! 0.9
0
.!:i 0.8
t-<
'-' 0.7
0.6 I')
I ICA/~(f'..u
..,,'...---'-05 ~ /*';1" .i>:
0.4 10'
0.3
0.2 ~?.10/
0.1 ,---I L"'I' r:':"-i--l. .../ fI '\x~r-",w.f \ / /\ " gLe--B '¥ :"\~,--'<""" 3.,----E--J;v.'~( •
0
:1ili3:>j : I11 I;i 2£16I : I'1 :1~"I7: II1 i rIuns' : 1 :1~I 9 I : I11 :1~I',~o:' ;:.
.2 :4 :2 :·1 :2 :4 :2 :4 :2:4 2:4
[J Actual + Si m ul a tr d
UNIVERSITY OF IBADAN LIBRARY
CHAPTER SEVEN
SIMULATION RESULTS
"The analysis of idealized polar cases cannot pretend to be
more than a pedagogical exercise, to gain a firsthand
impression of the forces that are at work in the real world,
every attempt to make the system approach reality more
closely, creates complications, which SOO1l threaten to drown
economic insight in a plethora of unmanageable formulas. "
Sohmen (1965)
As pointed out in the previous chapter we tried to answer some 'what if questions
by carrying out dynamic ex-post simulation experiments under different policy scenarios.
These experiments allow us investigate the ninth hypothesis highlighted in chapter 4 section
I. Therewere three basic scenarios we considered for each country. The first is a situation
where the rate of growth of government expenditure increases over and above its actual
growth rate given that monetary and exchange rate policies remain unchanged. The second
is a situation where the rate of growth of money supply increased above its actual growth
rate, while fiscal and exchange rate policies remain constant. The third scenario is the
situation where the exchange rate policy changes, while fiscal and monetary policies remain
constant. For the first two scenarios (i.e changing fiscal and monetary policy environment)
we used the solution for the shorter version of our model (i.e the version of the model
covering the period 1917: 1 to 1990:4 wherein exchange rate is assumed to be fixed) as the
baseline solution for Ghana and Nigeria. For the exchange rate policy change, we used the
simulation result for the longer version of our model as the baseline solution for Ghana and
UNIVERSITY OF IBADAN LIBRARY
213
Nigeria, (i.e th version of the model covering the period 1986:4 1990:4 wherein exchange
o
rate and interest rates are assumed to be variable). This we did because for the exchange rate
policy change what we did was to assume that rather than adopt a floating exchange rate
system in the 4th quarter of 1986, suppose Ghana and Nigeria had simply depreciated their
currencies by 80% and left the exchange rate fixed what would have happened to our
endogenous variables relative to what happened to them as a result of floating.
In the case of Uganda we used the solution to the longer version of the model as the
baseline solution (i.e the version of the model covering 1981: 1 to 1990:4, wherein exchange
rates and interest rates are assumed to be variables). This is because, of the ten years studied
(1981 -1990). Uganda adopted a floating exchange rate system for six of those years. Thus
deviations against the result of this time period was used to analyze the effects of monetary
and fiscal policy changes. However since the change to a fixed exchange rate regime
occurred in the 2nd quarter of 1987 we attempted to investigate what would happen to
macroeconomic aggregates if instead of adopting a jumping-peg system in 1987:2, the
country had merely adopted a crawling peg system appreciating its currency at the rate of
5% per quarter beginning from 1987:2
In all we were interested in six main endogenous variables, three for the internal
sector and three for the external sector, these were domestic prices, real demand for money
•
and real income for internal balance and export value, import value and the trade balance for
the external sector. The results of these exercises for the three countries are discussed in the
o
UNIVERSITY OF IBADAN LIBRARY
214
following sections.
o
7.1 Simulation Results for Nigeria
In the case of Nigeria we tried to find out how the macroeconomic aggregates in the
Nigerian economy would perform if (i) rate of growth of government expenditure increases
from its quarterly average of 3.73% to 5% (ii) if money supply growth changed from its
actual quarterly average of 4.7% to 7%. These two policy changes were simulated separately
for the period 1981: 1 - 1990:4 to enable us have a long-run view of the situation on hand.
(iii) what would have happened if instead of floating the exchange rate independently in
1986:4 it had simply been depreciated from its 1986:3 level by 80% and allowed to remain
fixed till 1990:4. Below are the results of these experiments.
UNIVERSITY OF IBADAN LIBRARY
215
Table 1.1
Nigeria: Change in Fiscal Policy
Endogenous Average Size of Average Size Under Average Size Under
Variable Baseline Solution Fixed Exchange Floating Exchange
System System
PO 294.6300 294.4000 239.2800
POM 95.8500 95.8100 103.2400
1M 3651.1500 3653.8200 3106.6300
X 6880.4000 6893.3300 1221.1100
TB 3229.2500 3239.5100 3520.5400
RI 319.2300 321.6000 336.7300
Deviations From Baseline Solution
(Simulated Solution-Baseline Solution)
Endogenous Deviations Under Fixed Deviations Under Flo-
Variable System anng System
PO -0.2300 -55.3500
ROM 0.0200 1.3900
1M 2.6100 55.4800
X 12.9300 346.1100
TB 10.2600 291.2900
RI 2.3100 11.5000
Percentage Deviation
Endogenous Fixed Exchange Flouting Exchange
Variable System System
PO -0.0100 -18.1800
ROM 0.0200 1.1100
lM 0.0700 1.5200
X 0.1800 5.0400
TB 0.3100 9.0200
RI 0.1400 5.4800
•
UNIVERSITY OF IBADAN LIBRARY
216
Table 7.2
Nigeria' Change in Monetary Policy
Endogenous Average Size of Average Size Un- Average Size Under
Variable Baseline Solution cler Fixed Rates Floating Rates
PO 294.6300 311.8300 267.3800
ROM 95.8500 93.5400 99.4900
1M 3651.1500 3730.5000 3779.1300
X 6880.4000 6772.3700 7051.D300
TB 3229.2500 3041.8700 3271.9000
R1 319.2300 313.3200 371.7400
Deviations From Ba>oeline Solution
(Simulated Solution-Baseline Solution)
Endogenous Deviations Under Deviations Under Flo-
Variable Fixed System ating System
PO 17.2000 -27.2500
ROM -2.3100 -3.6400
1M 79.3500 127.9800
X -108.0300 170.6300
TB 187.3800 42.6500
RI -5.9100 52.5200
Percentage Deviation
ElIdogenous Fixed Exchange Floating Exchange
Variable System System
PO 5.8400 -9.2500
ROM -2.4100 3.7900
1M 2.1700 3.5100
X -1.5700 2.4800
TB -5.8000 1.3200
RI -1.8500 16.4500
•
UNIVERSITY OF IBADAN LIBRARY
217
Table 7.3
Nigerill. Change in Exchange Rate Policy
Endogenous Average Average Size Deviation From Percentage
Variable Size of of Simulated Baseline Solution Deviation
Baseline Solution
Solution
PO 812.1900 548.9000 -263.2900 -32.4200
ROM 52.3200 83.0000 30.6800 58.6400
1M 6644.1000 6857.8000 213.7000 3.2200
X 17852.3500 16265.0000 -1587.1100 -8.8900
TB 11208.2500 9407.2000 -1801.0500 -16.0700
RI 305.5900 336.1000 30.5100 9.9800
Table 7.1 reveals the effects of changing fiscal policy while monetary and exchange
rate policy remain fixed. The table is in three parts the first reveals the average size of our
endogenous variables over the 40 quarters both in the baseline solution and in the simulated
solution both under fixed and under floating exchange system. The second part reveals the
deviations between the baseline solution and the simulated solution while the third part
indicates the percentage deviation. It is interesting to note that increasing the rate of growth
of government expenditure from its actual quarterly rate of 3.73 % to 5% would lead under
a floating exchange rate system to a decline in domestic prices of over 18% and improve
export performance by about 5% of its baseline performance while the external sector (as
proxied by the trade balance) would improve by over 9%. On the whole real income would
rise by over 5% from its baseline level. Of course under a fixed exchange system the
performance of all the macrqeconomic aggregates improve too but very marginally. For
UNIVERSITY OF IBADAN LIBRARY
218
instance while the trade balance improved by only 0.31 % real income rises by only 0.74%
under ftxed rates as against 9.02% for trade balance and 5.48% for real income under
floating rates.
Table 7.2 shows the effects of introducing a monetary shock into the system. Here
instead of the actual average quarterly growth rate of 4.7% the money supply was made to
grow at 7%. As in the case of ftscal policy this monetary policy shock under a regime of
floating rates would in ten years lead to a decline of over 9% in domestic prices. Imports
would improve by over 3% while overall the trade balance performance would improve
marginally by over 1% but real income improves substantially by over 16%. Under a regime
of fixed rates however the sustained increase in money supply growth would by the end of
ten years lead to a permanent increase in domestic prices of over 5% and real demand for
money would fall by over 2%. The trade balance performance worsens by over 5% while
real income performance worsens also.
In other words monetary impulses under a regime of fixed rates would in Nigeria translate
to rising prices, worsening trade balance and worsening real income.
We also attempted to find out what would have happened to our economic aggregates
if rather than float the currency independently in the 1986:4 the Nigeria had simply
depreciated its currency by 80% from its 1986:3 level, and left the exchange rate ftxed at this
new rate till 1990:4. The result of this experiment proved quite interesting. Table 7.3 shows
•
that if such a situation had taken place, there would have been a decline of over 30% in
UNIVERSITY OF IBADAN LIBRARY
219
domestic prices from their baseline level, real demand for money would have risen by over
55% and real income too would have improved considerably by well over 9%.
However, the external sector's performance would have worsened as export
performance declines by over 8% and the external balance (as proxied by the trade balance)
declines by over 16%. This poorer performance of the external sector may be due to the fact
that given a fixed nominal exchange rate the declining domestic price level may not be
sufficient to compensate for any decline in domestic price of exports, as a result of decline
in world price of exports so that the real exchange rate appreciates. Alternatively even the
domestic price exports may fall if the nominal exchange rate is fixed and the world price of
exports is falling and as shown in table 6.3, the domestic export price with two period lag
(DP~_~ is a major determinant of exports in Nigeria.
In sum each of the three experiments for Nigeria indicate that in an ex-post situation,
the floating exchange rate system improves the performance of the Nigerian economy better
than a fixed one; either one is looking at the external or internal sector of the economy.
That prices are falling under a floating exchange system whereas they are rising
under a fixed one under monetary policy change seems quite strange. However the fact that
import value expands faster under the floating system is a pointer to the reason for declining
domestic prices. In the Nigerian economy almost all sectors are import dependent (be it
agricultural, manufacturing etc). Under a system of floating rates the market system ensures
that actual importers have reasonable access to foreign exchange resources, This easy'
UNIVERSITY OF IBADAN LIBRARY
220
foreign exchange access, ensures increased imports and hence increased domestic production
and from table 6.1 increased real income leads to falling prices. Under a regime of fixed
rates and its implicit forex rationing system there is a lot of crisis in the allocative
mechanism,such that actual importers never get direct access to forex resources, those who
do, get forex by virtue of "power patronage" or "political patronage". Thus, the actual
importer buys from these middlemen at exorbitant prices. This restricts imports (especially
of intermediate inputs and raw materials) thereby restricting output in the whole economy,
and hence rising domestic prices. Whereas, given the neoclassical position of the market
being more efficient than any allocative system, a system of floating rates ensures expanded
imports and expanded exports, as well as expanded domestic output hence the fall in
,domestic prices.
7.2 Simulation Results for Ghana
As pointed out earlier our simulation exercises were carried out under three
scenarios. The first is under a change in fiscal policy environment while monetary and
exchange policy environment remained constant. As a proxy for fiscal policy change we
used a change in the rate of growth of government expenditure. More specifically we
assumed that government expenditure had grown at an average quarterly rate of 16% as
against the actual rate of rowth of 14%. Also for the second scenario we assumed that the
rate of growth of money supply can proxy for monetary policy change. Thus we tried to find
UNIVERSITY OF IBADAN LIBRARY
. 221
out what would have happened if money supply had grown at an average quarterly rate of
14% as against the actual growth rate of 12%. These experiments covered the period 1981: 1
- 1990:4 in order to enable us glimpse the long-run effects of using a particular exchange rate
system.
For the exchange rate policy change, we attempted to investigate what would happen
to macroeconomic aggregates in the Ghanaian economy if instead of floating its exchange
rate in the fourth quarter of 1986, Ghana simply depreciated its currency by 80'%. Below are
the results.
•
UNIVERSITY OF IBADAN LIBRARY
222
-Tab-le -7.4
Ghana: Change in Fiscal Policy
Endogenous Average Size of Average Size Under Average Size Under
Variable Baseline Solution Fixed Exchange Floating Exchange
System System
PO 1421.5400 1364.6700 1321.0000
ROM 49.7300 50.8700 51.7500
1M 30851.6700 31240.400 31542.7500
X 26214.0800 27079.1400 27742.3600
TB 4637.5900 4161.2600 3800.3900
RI 398.9500 415.9100 428.9500
Deviations From Baseline Solution
(Simulated Solution-Baseline Solution)
Endogenous Deviations Under Fixed Deviations Under Ao-
Variable System ating System
PO -56.6700 -100.5400
ROM 1.1400 2.0200
1M 388.7300 691.0800
X 865.0600 1528.:!800
TB 476.3300 8372000
RI 16.9600 30.0000
Percentage Deviation
Endogenous Fixed Exchange Floating Exchange
Variable System System
PO -4.0000 -7.0700
ROM 2.3000 4.0700
1M 1.2600 2.2400
X 3.3000 5.8300
TB 10.2700 18.0500
RI 4.2500 7.5200
•
UNIVERSITY OF IBADAN LIBRARY
223
Table 7.5
Ghana: Change in Monetary Policy
Endogenous Average Size of Average Size Un- Average Size Under
Variable Baseline Solution der Fixed Floating Exchange
Exchange System System
PO 1421.S400 1662.0000 1579.0000
ROM 44.7300 44.8900 465600
1M 30851.6700 29201.1100 29768.7900
X 26214.0800 22559.7700 23815.4900
TB -46375900 -6641.3400 -5953.3000
RI 398.9500 327.2200 351.9100
Deviations From Baseline Solution
(Simulated Solution-Baseline Solution)
Endogenous Deviations Under Deviations Under Flo-
Variable Fixed System ating System
PO 240.4600 157.4600
ROM -4.8400 -3.1700
1M -1650.5600 -1082.8800
X -3654.3100 -2398.5900
TB -2003.7500 -1315.7100
RI -71.7300 -47.0500
Percentage Deviation
Endogenous Fixed Exchange Floating Exchange
Variable System System
PO 16.9000 11.0800
ROM -9.7300 -6.3800
1M 5.3500 -35100
X -13.9400 -9.5100
TB -43.2100 -28.3700
RI -17.9800 -11.7900
•
UNIVERSITY OF IBADAN LIBRARY
224
-Tab-le 7.6
Ghana: Change in Exchange Rate Policy
Endogenous Average Size Average Size of Deviation From Percentage
Variable of Baseline Simulated Solu- Baseline Solution Deviation
Solution tion
PO 3138.7000 3005.6000 -83.1000 -2.62
ROM 64.0000 65.0000 1.0000 1.53
1M 86424.1000 87150.0600 725.9600 0.84
X 72562.5000 74151.6100 1589.1100 2.19
TB -13861.6000 -2998.4500 873.1500 6.29 2.83
RI 443.0000 455.5400 12.5400
Table 7.4 reveals that if in Ghana, the rate of growth of government expenditure had
increased by 2% quarterly over and above its actual rate of growth then within ten years the
level of real income would have risen by 4.22% under a fixed exchange rate system and by
7.52% under a floating exchange system. Level of domestic price would have fallen as a
result of increased output. The fall in domestic prices would have been some 4% under a
fixed exchange rate system and about 7.07% under a floating one.
The increased rate of government expenditure while generating increased imports
and hence increased import value would also have led to higher exports as declining
domestic prices lead to increase in the real exchange rate (i.e. depreciation of the real
exchange rate (ETOT)) thus encouraging exports, and from table 6.3 the real exchange rate
ETOT with a two period lag has significant impact on exports. Over a period of ten years
•
the trade balance would have improved by some 10.25 % under a fixed exchange rate system
and some 18.05% under a floating exchange rate system. Overall the economy would have
UNIVERSITY OF IBADAN LIBRARY
225
performed better under a floating exchange rate system.
From table 7.5 it is clear that increase of 2% in the average quarterly rate of growth
of money supply over and above its actual level would have in ten years led to increases in
the rate of growth of domestic prices of about 16.9% under a fixed exchange rate regime and
about 11.08% under a floating one. The increased rate of price growth given a monetary
shock reflects the fact that monetary impulses in Ghana translate faster into higher prices
than into higher output. As prices rise the real exchange rate appreciates and with it export
volume and value fall too. The higher prices also reduce the level of real income thereby
reducing the level of imports as real income is crucial to imports (see table 6.4). In the end,
import value falls by some 5.35% under a fixed exchange rate regime while it falls by about
3.51% under a floating system. Export value also falls by some 13.94% under a fixed
exchange rate system and some 9.51 % under a floating one.
The differential in the performance of export value and import value under the two
exchange rate systems is a reflection of the fact that even though the real income is falling
under the two systems it is falling less under the floating system, hence the level of decline
in imports is less under a floating system. Given that the export sector itself depends on
imported inputs the fall in export value is less under a floating system then under a fixed
one. What is more- as in Nigeria the use of the market system in the allocation of foreign
exchange resources proves IllPre effective by allowing actual producers access to forex so
that level of imports is higher under floating system than under fixed. Under a fixed
UNIVERSITY OF IBADAN LIBRARY
· 226
exchange rate system the forex rationing system is so wrought with crisis that often the forex
does not get into the hands of importers but into those of the middlemen who then sell to
actual importers at exorbitant prices, hence reduced level of imports under a system of fixed
rates.
On the whole, though increased rate of growth of money supply in Ghana generates
increased growth rate of domestic prices the rate of price growth is less under a floating
system because the rate of expansion of output under a floating system is higher than under
a fixed one. The decline in the trade lance and similarly the decline in real income are
also less under a floating system than under a fixed one.
,
Table 7.6. reveals a situation where exchange rate is depreciated by 80% from its
1986: 3 level rather than adopting an independently floating system. The table reveals that
this exchange policy would have by 1990:4 reduced domestic prices by some 2.65%,
increased real income by some 2.62% while the trade balance would have improved by some
6.29%. This situation however does not provide a long run picture as the period covers only
four and a quarter years.
7.3 Simulation Results for •Uganda
As in the case of Ghana and Nigeria we performed simulation experiments under
three different scenarios. The first was under the assumption that government expenditure
increased at an average quarterly rate of 18.5% as against its actual rate of 17.43%. The
UNIVERSITY OF IBADAN LIBRARY
. 227
second was under the assumption that money supply grew at an average quarterly rate of
19% as against the actual growth rate of 17%. The third was that rather than adopt a fixed
exchange rate system of the jumping-peg variety in the second quarter of 1987, Uganda
merely adopted a crawling peg system whereby it appreciated its exchange rate by 5%
quarterly beginning from its 1987: 1 level.
The difference in the Ugandan experiments compared with Ghana and Nigeria is that
the solution to the longer version of our model (i.e the version wherein exchange rates and
interest rates are floating) was used as the baseline solution in the Ugandan case for fiscal
and monetary policy changes. This is because, of the ten years studied for Uganda (1981: 1-
,
1990:4) it floated her currency for the greater part of the period. In analyzing the exchange
policy change however we used solution to the shorter version of our model (i.e the version
where exchange rate is fixed) as the baseline. This is because the period covered by the
exchange rate change (i.e 1987:2 - 1990:4), is the period of fixed rates covered by the shorter
version of our model. Table 7.7, 7.8 and 7.9 contain the-results of these experiments .
•
UNIVERSITY OF IBADAN LIBRARY
228
Table 1.7
U!!.anda: Chanl!e in Fiscal Policy
Endogenous Average Size of Average Size Under Average Size Under
Variable Baseline Solution Fixed Exchange Floating Exchange
System System
PO 19212.8600 18355.5600 18119.51
ROM 139.7900 149. 1000 145.19
1M 22151.4300 24028.1500 23664.58
X 21058.1000 21193.6500 21481.98
TB -2098.1300 -2234.5000 -2182.60
RI 1538.0000 1601.6100 1518.29
Deviations From Baseline Solution
(Simulated Solution-Baseline Solution)
Endogenous Deviations Under Fixed Deviations Under ne-
Variable System ating System
PO -851.3000 -493.2800
ROM 9.4000 5.4000
1M 810.7200 501.1500
X 734.9500 423.2800
TB -135.1100 -83.8100 .
RI 69.6100 40.2900
Percentage Deviation
Endogenous Fixed Exchange Floating Exchange
Variable System System
PO -4.4600 -2.5100
ROM 6.6600 3.8800
1M 3.1600 2.1900
X 3.4900 2.0100
TB -6.4100 4.0000
RI 4.5300 2.6200
o
UNIVERSITY OF IBADAN LIBRARY
229
Table 7.8
Uganda: Change in Monetary Policy
Endogenous Average Size of Average Size Un- Average Size Under
Variable Baseline Solution der Fixed Rates Floating Rates
PO 19212.8600 11667.2900 11513.1300
ROM 139.7900 212.5400 212.7700
fM 23157.4300 30023.6100 30049.0800
X 21058.7000 27536.3600 27666.9200
TB -2098.7300 -2487.2500 -2382.1600
Rf 1538.0000 2082.1400 2084.1400
Deviations From Baseline Solution
(Simulated Solution-Baseline Solution)
Endogenous Deviations Under Deviations Under Flo-
Variable Fixed SYSlCm ating System
PO -7545.5700 -7699.9800
ROM 72.7500 72.9800
fM 6866.1800 6891.6500
X 6477.6600 6608.2200
rn ·JSB200 ·2U4JOO
Rf 544.1400 546.1400
Percentage Deviation
Endogenous Fixed Exchange Floating Exchange
Variable System System
PO -39.2700 -40.0800
ROM 52.0400 52.2100
fM 29.6500 29.7600
X 30.7600 31.3800
TB 18.5100 13.3800
Rf 35.3800 35.5100
•
UNIVERSITY OF IBADAN LIBRARY
· 230
-Tab-le 7.9
Uganda: Change in Exchange Rate POlicy
Endogenous Average Size Average Size Deviation From Percentage
Variable of Baseline of Simulated Baseline Solution Deviation
Solution Solution
PO 37429.13 48401.57 10972.4400 29.3200
ROM 223.25 125.06 -98.1900 -43.9800
1M 20963.59 18733.07 -2230.5200 -10.6400
X 10356.77 7978.86 -2377.9100 -22.9600
TB -10606.82 -10754.21 -147.3900 -1.4000
RI 1425.68 1244.62 -181.0600 -12.7000
Table 7.7 reveals that if the rate of government expenditure in Uganda were to
increase by some 1% over and above its actual average quarterly growth rate, then within
a space of ten years this would bring about a permanent decline in prices (domestic) of about
2% an improvement in real income also of about 2% and a worsening of the trade balance
by about 4% under a floating exchange system while it would lead to a decline in domestic
prices of over 4%, a rise in real income of over 4% and a worsening of the trade balance by
over 6% under a fixed exchange system. This worsening of the trade balance under a fixed
exchange rate system in spite of declining domestic prices is to be expected as the decline
in domestic prices may not be sufficient to outweigh any decline in international price of
export, so that the real exchange rate would be appreciating. Alternatively given that
nominal exchange rates are fixed, domestic price of exports (DPX) may be declining; if the
world price of exports (px) is declining, and given that DPX with a two period lag is crucial
to exports (see Table 6.3) export value may be declining or at least may not rise, as fast as
UNIVERSITY OF IBADAN LIBRARY
231
imports. Under a floating exchange regime however, nominal exchange rates rise to negate
declining internationally given export price (Px) so that decline in export value is reduced,
as domestic price of exports falls less than under a fixed rate system
Table 7.8 reveals the effect of introducing a quarterly shock of about 2% into the
system in form of increased growth of money supply. This increased money supply growth
expands output thereby leading to declining prices both under fixed and under floating
exchange systems. The declining domestic price level leads to increased real income by
about 35.51 % under a floating system and about 35.38% under a fixed system. Both imports
and exports increase relative to the baseline level but the growth in exports is not sufficient
,
to ensure a positive trade balance hence the trade balance declines under the two exchange
systems but the decline is greater under a fixed system than under a floating one.
Thus while the trade balance performance worsens by some 13% under a floating
rate system it worsens by some 18% under a fixed one. The worsening of this trade balance
is probably due to structural rigidities in the economy such that export supply may not be
able to respond adequately to relative price incentives while import demand may also not
be able to respond appropriately to relative price signals. Whatever be the case while
domestic prices are falljngunder the two exchange systems they are falling a little faster
under a floating system. And while real income is rising under the two systems it is, rising
a little faster under a floating one, and while trade balance is getting worse under the two
systems it is getting worse at a faster rate under the fixed exchange system.
UNIVERSITY OF IBADAN LIBRARY
232
TIle third scenario depicts the impact of adopting a crawling peg in 1987: 1 whereby
the exchange rate is allowed to appreciate at the rate of 5% quarterly from its level of Ush
1400 per dollar, instead of the sudden jump from Ush 1400 per dollar in 1987:1 to Ush 60
per dollar in 1987:2. The results of this experiment shows that such an appreciation would
have led to rising prices as the domestic price level would have risen by over 29% by
1990:4. The appreciation of the real exchange rate would have led to increased demand for
imports and reduced demand for home produced goods, as home produced goods would
have become relatively dearer than foreign goods. Though resources would in the short - run
flow to the domestic - goods production because of the relatively higher prices in the
medium to long run the lack of adequate demand in this sector would have stiffled increased
production of domestic goods thus generating rising domestic prices in the medium to - long
term. In the external sector exchange rate appreciation would also have a stiffling effect on
the growth of exports, thus export performance would have worsened. As resources flowing
from exports are needed to pay for imports (see Table 6.4) worsening export performance
would also have led to declining imports though imports would expectedly decline at a lower
rate than exports. Hence exports would have declined by 22.96%, imports by 10.64% and
real income by 12.7%. However, the fact that Uganda put in place a jumping-peg
arrangement wherein the over-appreciated currency had by 1988:1 become grossly
depreciated and the depreciation continued until 1990:4, giving a picture more of a floating
system than of a fixed one, accounted for the actual rise in real income in the actual trend
UNIVERSITY OF IBADAN LIBRARY
233
movement.
On the whole the results of our simulation experiments confirm or validate the
hypothesis that macroeconomic aggregates perform better under a system of floating rates
than under a fixed one.
•
UNIVERSITY OF IBADAN LIBRARY
CHAYrER EIGHT
SUMMARY, CONCLUSIONS AND RECOMME DATIONS
" Partial Conclusions Are Handicapped By The
Impossibility of Completely Isolating The Forces Under
Analysis, Conclusions of General Validity Are Limited,
Because Of The Impossibility of Assigning Lasting Weights
To The Various Forces"
Kindleberger (1956)
This study has attempted to answer the question - which exchange regime enhances
economic performance - a floating or a fixed one? In answering this question we have used
two main criteria. The first is - does the exchange rate regime help or hinder macroeconomic
policy in pursuit of fundamental domestic economic objectives of stable prices, and high
growth rate of real income? The second is - how effective is the system in ensuring external
sector equilibrium? In order to evaluate our exchange rate regimes using these two criteria
we have carried out our analysis at two levels. First we modelled the sample economies and
then estimated our model(s). Each country's model had two versions. The first version
attempts to capture the economy under a fixed exchange rate system, hence in this version
the exchange rate is treated as exogenous while the level of international reserves is treated
as stochastic (as obtains under fixed exchange systems). The second version of the model
attempts to capture the economy under a floating system, hence in this version the exchange
rate is treated as endogenous (a•s obtains under a floating system). The first version of the
model was estimated for all the years covered by the study 1977: 1 - 1990:4, for the countries
UNIVERSITY OF IBADAN LIBRARY
235
of Ghana and Nigeria. The second version of the model was estimated for only the years
these countries floated - 1986:4 to 1990:4. In estimating the second version of the model all
the parameters of the first version were assumed to remain constant, so that all that is
changing in the system is the nominal exchange rate and its companion, the domestic rate
of interest. Thus any differences in the behaviour of the system can only be attributable to
the exchange rate system. Having estimated these two versions of the models, we solved the
models; the solution to the first version (i.e fixed exchange rate versions) was used as
baseline solution for the evaluation of the macroeconomic aggregates under changing policy
environment.
At the second level we carried out dynamic simulation exercises under changing
monetary, fiscal and exchange policy environment. For each policy change examined we
,
assumed the other policies remained constant. It is important to note that for Uganda the
situation was slightly different. Having floated for the greater part of the period covered by
the study (1981:1 - 1990:4) the solunon to the version of the model designed for floating
system was employed as the baseline solution and it was estimated for the ten years 1981: 1 -
1990:4. The fixed exchange system version of the model was only estimated for the years
1987:2 - 1990:4.
UNIVERSITY OF IBADAN LIBRARY
236
8.1 SUMMARY OF MODEL ESTIMATION RESULTS
The results of the estimation reveal that, indeed there is some degree of pass through
from nominal exchange rates unto domestic prices in Nigeria and Uganda, the relationship
being more pronounced in Uganda where the exchange rate elasticity of domestic prices is
11% and it is significant at the 5% level. In Nigeria the exchange rate elasticity of domestic
prices is only 5% and it is also significant at the 5% level. In Ghana there doesn't seem to
be any pass through from nominal exchange rates onto prices as the relationship between the
two variables is statistically insignificant. The major propeller of prices in Ghana is the
money supply. Increases in nominal money supply increase prices. Thus while it may be true
that market exchange rates may contribute substantially to inflation in Uganda, it is not
likely to have a tremendous impact on inflation in Nigeria; and in Ghana it is not likely to
..
affect inflation at all. As pointed out earlier, the Ghanaian result confirms Chibber and
Shaffik's (1989) position concerning exchange market and inflation in Ghana, while the
Ugandan situation does not seem to support Elbadawi's '(1990) hypothesis with respect to
inflation and exchange rates in Ugandan.
In respect of the impact of monetary policy on exchange rates the estimation resu ts
confirm that indeed nominal exchange rates follow the money supply. The money supply
elasticity of nominal exchange rates is 85% in Nigeria, 54% in Ghana and 37% in Uganda,
all are significant at the 5%• level. This result goes to show that the success of a floating
exchange rate regime depends very, very, crucially on responsible, prudent, and judicious
UNIVERSITY OF IBADAN LIBRARY
237
monetary policy stance. Reckless increases of the money supply which often takes place in
these countries in response to reckless government spending aggravate nominal exchange
over - depreciation.
That imports are very crucial to exports is shown under the third hypothesis tested,
the result of which showed that the import elasticity of exports is 38% for Nigeria, 78% for
Ghana and 56% for Uganda. This result confirms Jebuni, Sowa and Tutu's (1991) position
for Ghana that the rise in exports is a result of inflow of structural adjustment loans which
expanded imports considerably. It also confirms the position of Khan and Knight (1986),
Moran (1989), that imports are crucial to export expansion in developing countries.
That changing the structure of relative prices in the export sector in favour of exports
through nominal exchange depreciation increases exports is confirmed by the Ghanaian
res,ult where the relative price ( with a two period lag) elasticity of exports is 43% and is
significant at the 5% level. Though the results came out with the wrong signs for Nigeria and
Uganda we suspect that, the problem here, lies with the lag - structure, rather than with the
validity of the position per - se; because a look at the domestic price of exports reveals that
the higher this price gets (as nominal exchange rate gets depreciated) the higher exports get
too.
That imports depend on exports, to the extent that net exports (or trade balance) is
a major determinant of reserves, and change in reserves is crucial to imports is shown in
tables 6.4 and 6.5. Table 6. reveals that the change in reserve elasticity of imports is 28%
UNIVERSITY OF IBADAN LIBRARY
·238
in Nigeria, 41 % in Ghana, and 58% in Uganda and all are statistically significant at the 5%
level. And table 6.5 reveals that reserve depends significantly on net export (or the trade
balance, with a one period lag, TBt_I). The trade balance elasticity of reserves is 28% for
Nigeria, 36% for Ghana and 44% for Uganda all statistically significant at the 5% level.
Finally the estimation results reveal that fiscal deficit is the major culprit for fuelling
inflation in all the three countries. Table 6.6 reveals that government deficit (GD) is a major
determinant of the money supply. The elasticity of nominal money supply with respect to
change in government deficit is 76% in Nigeria, 85% in Ghana and 71% in Uganda. What
is more, the money supply is shown to impact significantly on domestic prices (see Table
6.1). Thus prudent fiscal management policy is also crucial to the success of any exchange
rate regime, (be fixed or floating) as reckless fiscal spending engender undue increases in
the money supply, and since it has been shown that nominal exchange rates follow the
money supply, undue increases in the money supply, lead to undue exchange rate
depreciation.
In all, the estimation results show that exchange market movements are not the sole
factor responsible for inflation in the focus countries rather poor monetary and fiscal
management is. Also, that, overall output can be enhanced by enhancing imports because
the economies are heavily import - dependent. What is more - export performance is itself
dependent on import perfonhance and there is a symbiotic relationship between them so that
any policy that tries to promote exports while reducing imports may not work. Rather, the
UNIVERSITY OF IBADAN LIBRARY
policy target should be, promoting both exports and imports but ensuring a faster rate of
growth for exports than imports to ensure a positive external balance.
As for the position in the literature concerning capital mobility based on exchange
rate differential, the estimation results show that capital inflow in these countries do not
respond to interest rate differential between these countries and their trading parmers(which
are often developed economies) as the focus countries securities are not considered perfect
substitutes for those of their trading partners. Hence the exchange rate appreciation that is
supposed to result from maintaining domestic interest rate structure that is higher than that
of their trading partners never materialized.
8.2 Summary of Model Simulation Results
The dynamic simulation exercises reveal the following: (Note that the ex-post
simulation carried out under floating regimes for Nigeria and Ghana for the years 1981: 1-
1990:4 is known as "backcast simulation").
(i) In the long run (say over a period of ten years) a floating exchange rate system is
more effective in expanding output and curbing inflation in Nigeria when fiscal
policy is employed. This is revealed by a 5.48% permanent increase in real income
after ten years and an 18.78% reduction in domestic prices under a floating
•
exchange rate regime as compared with a 0.74% increase in real income and a 0.07%
reduction in domestic prices under a fixed exchange system when government
UNIVERSITY OF IBADAN LIBRARY
240
expenditure growth rate is increased by a quarterly rate of 1.25% over and above
its actual growth rate.
(ii) External performance is also better under a floating exchange regime than under a
fixed one, this is revealed by expansion of the trade balance by 0.3% under fixed
rates as against 9.02% under floating rates.
(iii) In the long run, a floating exchange rate system is more effective in expanding
output and curbing inflation than a fixed one when monetary impulses are changed.
This is revealed by a 16.45% increase in real income and a decrease of 9.25% in
domestic prices over a ten year period under a floating exchange rate regime as
against a reduction of 1.85% in real income and an increase of 5.84% in domestic
prices after ten years under a fixed exchange rate system with the introduction of a
2.3% increase in the average quarterly growth rate of money supply.
(iv) Under a floating system, use of monetary policy also proved more effective in
improving external sector's performance than under a fixed exchange system as
evidenced by an expansion in the trade balance of + 1.32% under floating rates as
against a decrease of - 5.80% under a system of fixed rates.
Results III and IV for Nigeria confirm the Mundell- Fleming's (1961,1962) as well
as Rhomberg's (1964) position that the floating exchange system performs better
than a fixed one hen monetary policy instruments are employed. However the
Nigerian result with respect to use of fiscal instruments contradicts the Mundell -
UNIVERSITY OF IBADAN LIBRARY
"241
Fleming one that a fixed exchange rate regime is more efficient than a floating one
when fiscal instruments are employed. In the case of Nigeria the floating system
proved superior both under fiscal and monetary impulses. This Nigerian result also
contradicts Tullio's (1981) position that the Italian economy performed better under
a fixed exchange regime both when fiscal and when monetary instruments are
applied.
(v) If rather than float her currency independently in 1986:4 Nigeria had simply
depreciated the currency by 80% from its 1986:3 level and left the exchange rate
fixed at the new level then by 1990:4 the level of domestic prices would have been
less than what they were by 32.4% . A considerable fall in domestic prices which
would have led to increased real income by 9.98%. However the external sector's
performance would have worsened as evidenced by a reduction in the trade balance
of 16.07%. Such worsening of the trade balance would also have had effects on the
current account balance, the balance of payments and the country's ability to
liquidate her debts.
(vi) In Ghana a floating exchange rate system is more effective in expanding real income
and reducing prices (which usually follows increased output) when fiscal expansion
is introduced, as evidenced by an expansion of real income by 7.52% and a decline
of domestic prices by 107% under floating rates as against an expansion in real
income of 4.25% and a decline of 4% in domestic prices under fixed rates. When
UNIVERSITY OF IBADAN LIBRARY
·242
the rate of growth of money supply is increased by 2% 0 er and above its actual
growth rate.
(vii) The performance of the external sector is also enhanced under a floating system than
under a fixed one when fiscal policy is used as shown by an expansion in the trade
balance of 18.05% under a floating system as against an expansion of 10.27% under
a fixed system when the rate of growth of government expenditure is increased over
and above its actual growth rate by a quarterly average of 2%. Result (VI) and. (VII)
above confirms Chibber and Shaffik's (1989) position that the onset of floating
exchange rates is likely to have proved beneficial to Ghana to the ex rent that it
improved fiscal balance given the inflow of structural adjustment grants at a more
enhanced domestic rate. If fiscal expansion has the long run impact that results (vi)
and (vii) imply, then an improved fiscal balance can only be beneficial to the
Ghanaian economy.
(viii) A floating system is less problematic in terms of worsening income an rising
domestic prices when monetary impulses translate faster into increased prices than
they do into increased output. This is borne out by the fact that given the shock of
a 2% average quarterly increase in the rate of growth of money supply in Ghana, real
income declined by 11.79% under floating rates within ten years as against 17.98%
decline under fixed rates. Also domestic price level rose by 11.08% under floating
system while it rose by 16.9% under fixed rates.
UNIVERSITY OF IBADAN LIBRARY
243
(ix) In terms of external performance a floating exchange rate system also proved less
problematic in terms of worsening trade balance when monetary impulses generate
lower real income and lower exports. This is shown by a decline in the trade balance
of 28.37% under floating rates as compared with a decline of 43.21 % under fixed
rates.
What is more - results VI - IX for Ghana imply that a floating exchange system is
superior to a fixed one irrespective of whether monetary or fiscal policy instruments
are used, this reinforces the Nigerian result.
(x) If rather than float her exchange rate independently in 1986:4 the Ghanaian
monetary authorities had simply depreciated the cedi by 80% from its 1986:3 level
, and left it at the new level till 1990:4 then, domestic prices would have fallen by
2.68% while real income would have risen by 6.29% and trade balance would have
improved by 6.29%.
Result (x) however provides a medium-run view asagainst a long-run view provided
by results (vi-ixj.a long-run view is better in taking decisions that would have long-
run implications.
(xi) In Uganda a fixed exchange rate system is more effective in raising real income and
reducing domestic price level when fiscal policy is employed, than a floating system.
This conclusion is borne out by the fact that an increase in the average quarterly
growth rate of government expenditure by 1.08% led to a decline in domestic prices
NIVERSITY OF IBADAN LIBRARY
244
of 4.46% under fixed system and an increase in real income 'Of 4.53% as against a
decline of 2.57% in domestic prices under a floating system and an increase of
2.62% in real income. This result ryhmes with Mundell - Fleming and Rhomberg's
position, that the economy performs better under a fixed system provided fiscal
instruments are employed.
(xii) A floating system seems less problematic in terms of external sector performance
even when fiscal policy is employed. This is revealed by a decline of 4% in the trade
balance under a floating system as compared with a decline of 6.47% under a fixed
system. Thus for Uganda under fiscal impulses the external sector performance will
be enhanced under floating than under fixed exchange rate regime.
(xiii) A floating exchange system is more effective in Uganda when monetary policy is
used in terms of increased real income and reduced level of domestic prices as
evidenced by a growth of 35.51 % in real income under floating rates and a decline
of 40.08 % in domestic prices over a ten year period, as compared with increased real
income of 35.38% under a system of fixed rates and 39.27% decline in domestic
prices when a shock of 2% in quarterly growth rate of money supply is introduced
into the system.
(xiv) A floating exchange rate regime also proves less problematic in terms of external
sector performance ,;,hen monetary policy is utili sed as shown by the fact that while
the trade balance worsened by some 13.5% under floating rates it worsened by some
UNIVERSITY OF IBADAN LIBRARY
245
18.38% under fixed rates.
It is important to point out one fact with regards to result (xiii). The differences in
the performance of these internal balance variables are so small under the two
systems because the fixed exchange rate system adopted in Uganda from 1987:2 is
more of a jumping peg and hence had almost the same degree of flexibility as the
floating one, hence the closeness of the performance of the macroeconomic
aggregates under the two systems. In spite of the above however, the floating system
still proved slightly superior.
(xv) If Uganda had appreciated her currency by 5% quarterly beginning from 1987: 1
instead of the sharp appreciation of 95.7% as between 1987: 1 and 1987:2, then the
level of domestic prices would have fallen by 12.7% while the trade balance would
have worsened by 1.4% by 1990:4.
In all it can be concluded that the result of our ninth hypothesis is, that, floating
exchange rate system enhances macroeconomic perfonnarice (in terms of higher growth rate
of income, higher exports, more positive trade balance) better than a fixed exchange system.
However this result can only obtain in the long - run when we are considering a period of
about 10 years.
•
8.3 Limitations of the Study
It is important to point out some shortcomings implicit in this study. First, is the fact
UNIVERSITY OF IBADAN LIBRARY
. 246
that it has been shown that in primary producing economies like Nigeria, Ghana and Uganda,
import demand is affected by non-price factors or restrictions other than those associated
with export earnings, thus, a term is usually required to represent such non-price-factors
(Khan and Knight 1986). Such a term is missing from our model(s). Second, imports were
lumped together as a homogenous whole, whereas we know that imports have three major
components, capital imports, consumer imports, and intermediate imports. A distinction
between these three categories of imports can prove very useful in deciding how to handle
the problem of increasing imports in the face of dwindling foreign exchange earnings.
Thirdly, the study has failed to incorporate comprehensively the parallel market for
foreign exchange, given that several studies (Elbadawi, 1990, Olopoenia, 1991, Chibber and
Shafik, 1989) have shown the parallel market to be a significant force in the economies of
our sample countries. However we do know that the bulk of the forex available in the focus
countries is in the hands of government, as government, through its agent is usually the sole
exporter of the major export commodity. For instancein Nigeria the Nigerian National
Petroleum Corporation is the sole exporter of petroleum, the major source of forex, in spite
of SAP. Thus the official forex market accounts for a significant proportion of the forex
market share in these countries. Besides, data on parallel market transactions is usually
considered suspect, because those who deal in the parallel market (because of its
illegitimacy) do not always come out with transparent responses during the data-collection
•
exercises. So that even published data on parallel market exchange rates in these economies
UNIVERSITY OF IBADAN LIBRARY
·247
are not considered completely dependable.
Further, the use of "experts" in the aggregated form is a serious shortcoming, to the
extent that it limits the insights that can be had into the specific impact of exchange-regime
change on specific forms of exports (traditional and non-traditional, oil and non-oil e.t.c).
The use of 'exports' in its aggregated form was informed by the ready availability of data
in this form, completely disaggregated exports for all the years covered by the study and for
all the countries was not available.
Another serious limitation of this study worthy of emphazing is the use of nominal
variables as against real ones. While we recognize the greater relevance of real variables we
were however constrained to utilize nominal ones because of the numerous problems of
trying to get real variables. First was the problem of getting an "appropriate" base year for
all the countries, then there was the problem of getting an appropriate deflator for all the
countries. The issue of deflator also brought in the problem of deciding on identical basket
of goods and services for all the countries to go into the deflator. And finally was the
problem of aggregation, to be sure that we had identical variables that we were deflating.
Because of all these problems inherent in the use of real variable we decided to use nominal
ones. Also, it has proved fairly difficult to isolate completely exchange rate effects from
general trade liberalization effects and financial liberalization effects on the macroaggregates
of interest, because almos all the policy instruments were changing simultaneously.
Finally, the study suffers from a major defect and that is, that to all intents and
UNIVERSITY OF IBADAN LIBRARY
·248
purposes it is a partial equilibrium analysis and hence Kindleberger's assertion cited at the
beginning of this chapter aptly applies here.
8.4 Conclusions
We can conclude this study in the words of Goldstein (1984) that "the capacity of
the exchange rate system per se to do good or harm should not be overestimated ... the
importance of discipline and coordinated macroeconomic policies for the successful
operation of floating rates should not be underestimated". In evaluating our exchange rate
regimes we noted in particular to what extent a given ERR promotes the achievement of
domestic stabilization objectives of stable prices and increased growth rate of income.
Using the growth rate of real income as a measure of employment growth we can conclude
from the results of our simulation exercises that a floating rate system appears better than
a fixed one in the long-run in promoting macroeconomic performance. In the six
experiments of the effects of monetary and fiscal polity changes on real income, and
domestic prices and real demand for money, for the three countries, the floating exchange
system proved superior in five of the six cases, the only time the fixed system proved
superior was in Uganda's case under a fiscal policy change. It is important to note that even
in situations where the economy was performing poorer as a result of the policy change the
rate of decline of the relevarh variables were lower under floating than under fixed system.
Secondly we evaluated the exchange rate regimes in terms of their ability to enhance
UNIVERSITY OF IBADAN LIBRARY
249
the external sectors' performance. Employing this criterion we discovered that for the six
experiments involving the use of fiscal and monetary policy instruments in the three
countries, the floating exchange system proved superior in all the six instances: even when
the external sector's performance was worsening, as evidenced by a declining trade balance,
the rate of decline was less under floating system than under fixed, this was the situation in
Uganda both when monetary policy was varied and when fiscal policy was varied.
Of course, we are aware that the floating exchange rate system has been noted for its
extreme exchange rate variability which producers say is not good for planning. It is
important to note however that as some economists have contended, a stable exchange rate
system depends more on stable macroeconomic policies at the national level rather than on
the form of the exchange rate regime itself. A system of stable exchange rates can itself be
jeopardized as much by insufficient as by excessive exchange rate flexibility. Exchange rate
variability is important to the extent that it impinges upon or facilities the achievement of
the ultimate targets of economic policy. And if the floating exchange rate system proves
potentially more efficient (in the long-run) in facilitating the achievement of these economic
objectives then the floating exchange rate system can only be skillfully nurtured in
developing countries to ensure that these countries get the best from the system.
8.5 Recommendations
•
An exchange rate system must not be thought of as possessing all the solutions to all
UNIVERSITY OF IBADAN LIBRARY
250
the economic problems that a country might face. In particular developing African countries
like Nigeria. Ghana and Uganda are known to face serious structural problems of which
exchange rate regimes cannot offer any viable solution - (Oyejide, 1990).
That casual empiricism seems to suggest that there has been high inflation rates. high
unemployment rates, slow growth in productivity in the latter half of the 1980s in Ghana.
Nigeria and Uganda may not after all be due to the adoption of floating rates. for in the
words of Goldstein "no exchange rate regime would have emerged unscathed from the
combination of shocks •... structural and institutional changes ... " of those years. What is
more - the period 1986:4-1990:4 provides only a medium-run view when compared for
instance with a period of ten years.
, As mentioned earlier there are structural rigidities that these economies have had to
grapple with which may themselves impinge on the performance of any exchange rate
system. The greatest of such structural problems is supply rigidities especially of goods and
of forex resources. Such supply shortages can however be remedied through increased
imports and increased domestic output. Increased imports become important as Khan and
Knight have shown that. even exports (the major forex earner) depend critically on imports
in developing countries. Our results have borne this out as shown by an import elasticity of
exports of 38% in Nigeria. Thus if import of intermediate inputs is so crucial to output
expansion (both domesti output and exports) then imports of the right type must be
encouraged. In Nigeria to increase supply a more aggressive promotion of non-oil
UNIVERSITY OF IBADAN LIBRARY
251
production must be encouraged. The execution of the Liquified Natural Gas project and the
Petrochemical and Oso Condensate must be pursued with seriousness in Nigeria.
For all the three countries studied a major problem has been the inadequacy of forex
supply to the forex market both from official and non-official sources. The official source
has been constrained by huge external debt service payments. This huge external debt
repayment burden leaves little left for supporting investment in these countries. This is a
situation that must be improved upon if indeed these countries must break free from the
chain of indebtedness and poverty. The IMF and WB can help use their positions to help the
countries obtain reduced rates of external loan repayment. Nigeria for instance spends as
high as 45 % of its export earnings on loan repayment; this is suicidal for any country, the
debt service rate has to be considerably reduced. On the issue of capital inflow and the
..
expected exchange rate appreciation that is supposed to follow such inflow in order to
temper the excessive exchange rate depreciation, this has not materialized more for socio,
polical, psychological as well as economic reasons. In the tirst place for any foreign investor
to be attracted to an economy (be it for investing in financial or real ass est) he must have the
confidence that the investment plus the returns are realizable. In a situation of polical strife,
turmoil and instability no rational investor would be willing to take the risk of losing all by
investing in such an uncertain environment. Also the lack of transparency in governance as
well as the lack of accoursability which often inform policy-reversal, (in order to forster
some private ends) disrupts planning, generates uncertainties and risks that can not easily
UNIVERSITY OF IBADAN LIBRARY
'252
be hedged, and so it discourages the inflow of capital. Thus in spite of continuously rising
domestic rates of interest in the three focus economies, no appreciable capital inflow has
followed (as economic theory stipulates). In the end the rising domestic rates of interest
proved counter productive as it discouraged domestic investment even by domestic
investors.
One would therefore wish to suggest that those in positions of authority must be
made to appreciate the interwoveness of policies and economics and cannot denigrate one
while trying to promote the other.
Finally, given that the governmental authorities in these countries believe in the
efficacy of the floating exchange rate system, hence the return of Uganda that had in 1987
abandoned the system, back to it in 1992, the system however requires skillful and dogged
maneuvering to make it work well. So far, the objective of diversification of exports is still
miles away from being achieved in all countries. thus each of these economies is still prone
to external shock in its present state. Foreign exchange leakages through non-repatriation
of export proceeds persists and smuggling still goes on unabated. Therefore, new strategies
must be designed that take exp icit cognisance of the fact that managing the exchange rate
is a complex question involving issues in economics, politics and psychology .
•
UNIVERSITY OF IBADAN LIBRARY
·253
8.6 Areas For Further Research
Some economists have maintained that a model's result is as good as its assumptions.
Certainly the assumptions on which our model(s) were based influenced our conclusions.
For instance we assumed that under a floating exchange system the level of international
reserves is no longer stochastic and hence we treated it as an identity. certainly this kind of
assumption influenced our result; so it will be worthwhile to relax some of our assumptions
and see to what extent and in what directions our macroaggregates change.
Secondly the issue of capacity utilization under different exchange rate regimes has
generated some debate. While some argue that capacity utilization expand under floating
rates some claim it does not. It would therefore be worthwhile to inbuild capacity utilization
rate explicitly into our model and see how this affects our results. Lastly it would be
.•interesting to carry out a comprehensive research on good-governance and economic
performance, to enable us see to what extent some of our macroaggregates have been
determined not by economic variables per-se but by socio-political and psychological
factors.
•
UNIVERSITY OF IBADAN LIBRARY
254
ENDNOTES
CP. Kindleberger (1956) cited in Machlup F. (1956) "The Terms of Trade Effects of
Devaluation Upon Real Income And The Balance of Tn de.
KYKLOS Vol. 19, 1956.
Oyejide T.A. (1990) Supply Response In The Context of Structural
Adjustment In Sub-Saharan Africa. (African Economic
Consortium - Special Paper I) .
•
UNIVERSITY OF IBADAN LIBRARY
255
ACRONn1S
l. IMF = International Monetary Fund
2. WB = The World Bank
3. FOREX = Foreign Exchange
4. SSA = Sub-Saharan Africa
5. LDC = Less Developed Countries
6. BOU = Bank of Uganda
7. BOG = Bank of Ghana
8. CBN = Central Bank of Nigeria
9. IMFfWB = International Monetary Fund Cum World Bank
10. SAP = Structural Adjustment Programme
11. ERP = Economic Restructuring Package
12. ERR = Exchange Rate regime
l3. BOP = Balance of Payments
•
UNIVERSITY OF IBADAN LIBRARY
· 256
APPENDIX I
Definition of Terms
1. MSS = Nominal Money Supply
2. DCG = Domestic Credit to Government
3. DCP = Domestic Credit to the Private Sector
4. RE = International Reserves in Domestic Currency Terms
5. DRE = Change in International Reserve
6. CUB = Current Account Balance
7. CAB = Capital Account Balance
8. TB = Trade Balance
9. X = Value of Exports in Domestic Currency Terms
1"0. 1M = Value of Imports in Domestic Currency Terms
11. GD = Government Deficit
12. GR = Government Revenue
13. GE = Government Expenditure
14. RDF = Interest rate differential between domestic and foreign rates of
interest.
15. RD = Domestic Rate of Interest
16. RF = Foreign Rate of Interest
17. RIST = Real Income differential between trend in real income in the previous
UNIVERSITY OF IBADAN LIBRARY
257
period and actual income in the last period (RIS(H) -RI(I_I)
18. RIS = Real Income Trend
19. RI = Real Income
20. RDME = Excess Real Demand For Money (RD~ - MSS(I_I)
21. RDM = Real Demand For Money which is Nominal Money Supply
deflated by the domestic price level (MSSjPDL).
22. PD = Domestic Price Index
23. 1tr= Expected rate of inflation at time t, P~-1-P<\-2
P~-2
24. PF = Foreign Price Index
25. EMSS = Excess Money Supply i.e. (MSS(I_I)- RDM(,»
0
26. RISD = Real Income Differential as Between Actual Real
Income in the Previous Period and Mean Real
Income (RI(H) - RI)
27. RI = Mean Real Income Over The Study Period
28. E = Nominal Exchange Rate Index
29. EN = Nominal Exchange Rate
30. EPF = Foreign Price Index in Domestic Currency Terms
31. RER = Real Exchange Rate (i.e. EpfjPd)
•
32. RER = Trend in Real Exchange Rate (EPf/Pd)'
33 RER = The mean Real Exchange Rate over the Sample period
UNIVERSITY OF IBADAN LIBRARY
258
34. RD = Domestic Rate of Interest
35. DPx = Export Price In Domestic Currency Terms (i.e. EPx).
36. ETOT = Real Exchange Rate In The Export Sector or Relative Price of Exports
to that of domestic price level in domesticcurrency terms (EPXjPD)
37. RPIM = Relative Price of Imports to That of Home Goods i.e., EPirn/Ph)
38. RID = Real Income Differential as Between Trend in Real Income and
Actual weal Income in the Current Period (RIS-RIJ
39. y = Nominal Income or Gross Domestic Product.
40. RI = Gross Domestic Product deflated by the comsumer
price Index
41. TOT = Terms of Trade (Pim/Px) .
, •
•
UNIVERSITY OF IBADAN LIBRARY
259
APPENDIX 2
Linear Estimation results.
And Error - Analysis
NIGERIA
AI, PDt = -0.47 - 0.65RI(tl + .18MSS(t)
(-2.83) (1.25)
+9.66EN(t) + .12RD(t)+.091t(t)
(2.82) (0.57) (2.01)
+0.78 PD (H)
(5.56)
R'2 = .94
DW = 1.90
A2, RD~t) = 54.6 + .16 Y(t)+ .05GE(t)
(0.41) (2.06)
-2.15 RD(t) + 0.57RDM«t.l)
(-3.48) (6.11)
R,2 = .82
DW = 2.10
A3, MSS(t) = 1976.12 + 3.25 RI(I)+ .67TDC(,)
(2.59) (16.48)
+2.22 DREt
(5.31)
R'2 = .82
DW = 1.72
A4, RD(t)= 4.99 + 0.12 EMSS(,) •+ . 171t(t)
(2.05) (0.73)
-.40 Y, + 0.59RD
"1
(-0.41) (2.82)
UNIVERSITY OF IBADAN LIBRARY
260
R2 = .84
DW = 1.92
A5, ~t) = -2625.35 + 4.06 RIS(t)
(0.45)
-9.64 ETOT(t.2)+ 0.921M(t.I)
(-1.67) (2.87)
+ 0.46DPX(t.2)
(2.32)
R2 = .89
DW = 2.20
A6, IM(t)-221.33 +0.46RISt + 2.64RPIM(J
(2.19) (2.09)
0.29DREt + .68RDM(t)
(0.48) (1.84)
R2 = .84
DW =2.04
A7, RE(t)= 822.59 - 0.05 GDt.1
(-0.96)
0.62TB(t.1)- 0.21EMSS(t)
(2.24) (-2.48)
R2 = .62
DW = 1.93 •
A8, ENt = -2.02 + 0.38 RDFt - O.4RE(t.l)
(2.93) (-4.13)
+ .05Yt + 0.85 MSSt
(0.34) + (3.01)
UNIVERSITY OF IBADAN LIBRARY
261
R2 = .75
DW = 2.10
A9 Rl, = 197.90 + .09RD~
(0.19)
+ 0.80RISD - .30PD + .09IM(I.I)
(5.19) (-2.41) (0.76)
R2 = .62
DW = 1.90
(please note that the figures in parenthesis are the t
statistics while DW stands for the Durbin Watson and R2
adjusted regression coefficient).
As pointed out in chapters four and six we solved our model in its linear and
logarithmic forms, and then carried out the error-analysis in order to decide which form
tracks the data better, as indicated by the form that has the. least error. Table Al reveals the
result of this exercise for Nigeria .
•
UNIVERSITY OF IBADAN LIBRARY
262
TABLE Al
ERROR ANALYSIS OF NIGERIA'S LINEAR AND LOGARITHMIC ESTIMATIONS
Error Type Root Mean Square RMSE Relative to
Error Average size of BIAS
(RMSE) variable
SRMSE
Dependent Linear Log Linear wg Linear Log
Variable Form Form Form Form Form Form
PD 22.35 0.0006 8.7759 0.0235 0.1479 0.0005
RDM 9.40 0.1030 9.2610 0.1010 -1.2860 0.0046
MSS 3778.60 0.1600 16.0680 0.0006 222.5600 0.0016
RD 1.41 1.1530 7.3280 0.0050 0.4800 1.1480
X 2221.50 0.2400 34.8800 0.0037 -1.2143 0.0010
IM 693.70 0.2800 18.3730 0.0070 -11.2460 0.0007
RE 1802.90 1.6464 32.1610 0.0290 390.000 -1.3961
EN 0.57 0.5200 7.1670 0.0840 0.2000 -0.5160
RI 460.40 0.224 125.5890 0.0610 -132.7000 0.0009
No matter what type of error one looks at it is ~lear that for Nigeria the linear
specification contains much more error than the logarithmic form. Therefore given that the
log form has less error, we decided to carry out further analysis on our model based on its
logarithmic form. For instance whereas the root mean square error for domestic prices is
only 0.06 under the logarithmic form it is 22.35 under the linear form. Also whereas the
degree of bias in the import eq•uation is only 0.0007 under the log estimation it is -11.246
under the linear estimation.
UNIVERSITY OF IBADAN LIBRARY
263
GHANA
A 10, PDt = 69.87 - 0.35RI(t) + .05MSS(t)
(-0.96) (2.21)
+ .03EN(,) + 0.32RD(t) + 0.191tr
(0.74) (0.96) (2.04)
+ .81PD(,_I)
(8.43)
R2 = .96
DW = 1.80
All, RDM. = 80.65 + .05Y(,) + .12GE(,)
(0.40) (2.82)
-1.59RDt + 0.80 RDM.-l
(-2.86) (5.42).
R2 = .84
DW = 1.87
A12, MSS, = 6181.88 + 0.23RI(,)
(0.72)
+.87TDC(,) + .26DRE('_1)
(8.64) (2.61)
R2 = .98
DW· = 1.90
A13, RD(I)= 2.23 - .08EMSS(,)
(-0.94)
-0.571tr + . 12Y, + .97RD(I.l)
(-1.47) (2.34) (2.56)
R2 •= .70
DW = 1.90
UNIVERSITY OF IBADAN LIBRARY
264
A14. ~t) = -1556.85 + .20RIS(t) + 2.28ETOT(t_2)
(0.39) (6.72)
+0.59IM._1+ .25DP~_2
(12.88) (1.71)
R2 = .89
DW = 1.97
A15. IM(t)= -2258.37 + 0.46RI(,)
(2.34)
+ 0_25RPIM(,) + 0.34DRE(,) + .18RDM(,)
(0.84) (3.78) (1.96)
A16. RE(t)= -276.04 - 2.21GD(t_l)
(-2.00)
24.66TB(,_l) - 0.45EMSS(,)
(3.93) (-1.85)
R2 = .85
DW =2.10
A17. ENt = -38.24 + 7.4RDF,
(13.88)
0.05RE(I_l)+ O.llY, + .18MSS,
(0.71) (3.55) (2.74)
R-2 = .85
DW = 1.95
•
A18, RI. = 396.00 + 0.09RDME(,)
(2.49),
+ .82RISD(I'1)- 0.12PD, + 0.23IM.
00.00) (2.10) (2.43)
UNIVERSITY OF IBADAN LIBRARY
R2 = .70
DW = 1.87
As in the case of Nigeria we carried out the error analysis of these two forms of our
model specification (direct linear and log-linear) by solving the model in two forms and
investigating the simulation errors. Table A2 reveals the result of this exercise .
•
UNIVERSITY OF IBADAN LIBRARY
TABLEA2
GHANA: ERROR ANALYSIS OF NIGERIA' S LINEAR AND
LOGARITHMIC ESTIMATIONS
Error Type Root Mean Square Error RMSE Relative to
(RMSE) Average size of variable BIAS
SRMSE
Dependent Direct Log Direct Log Direct Log Form
Variable Linear Form Form Linear Form Linear Form
Form
PD 188.217 0.1278 0.6193 J.OOO4 3.0478 0.0004
RDM 11.3044 0.1850 0.1652 0.0027 -4.0957 0.0002
MSS 7930.2870 0.1334 0.1334 0.00001 72.8623 0.4100
RD 1.3535 0.0649 0.0785 0.0037 -0.1623 -0.0124
X 4692.7627 0.2668 2.5978 0.00002 30.1661 0.0003
1M 3593.7600 0.2195 0.1349 0.00001 -35.2349 0.0006
RE 57742.1125 0.3065 2.13172 0.00001 -27112.67 0.6545
EN 26.4398 0.0845 0.3172 0.0010 7.2940 0.0027
R..I 56.8089 0.1850 0.1425 0.00046 -3.6954 -0.0060
It is clear from table A2 that the log specification is better than the direct linear
specification as the inherent errors are smaller. The root ean square simulation error is
almost always less than unity under the log specification and the same holds for the bias but
the root mean square errors were considerably large while the bias too was also large under
the linear specification. Thus we conducted further analysis on the log form estimation
result.
•
UNIVERSITY OF IBADAN LIBRARY
267
UGANDA
A19 P~I) = -lO250.79 - 35.08 RI(I)+ .31MSS(,)
(-2.53) (11.09)
+5.79EN(,) + 829.49RD(,) + O.lO 1t(I)
(2.01) (0.97) (2.35)
0.67PD(I.l)
(5.55)
R2 = .95
DW = 2.10
A20, RDM(,) = -30.47 + .53Y(,) + .42RDM(I'I)
(2.66) (6.13)
+ 0.19 GE(,) - lO.80RD(t)
(2.35) (-1.82)
Ii2 =.89
DW = 1.63
A21, MSS(,) = -7425.93 + 658.41RI(,) +0.73TDC(,)
(0.06) (13.44)
+2.04DRE(I.l)
(O.lO)
iP =.90
DW = 1.93
A22, RD(I)= 11.41 + 1.50EMSS(,) + .211t(t)
(0.57) 2.34)
-.02Y(t) + .84RD(I.l)
(0.96) (9.73)
iP = .84
DW = 1.83
UNIVERSITY OF IBADAN LIBRARY
268
A23, X. = -91313.28 + 5309.58RIS + 27.64ETOT(t.2)
(4.76) (2.98)
+0.561~.J + 0.21DPX(t.2)
(2.22) (3.81)
iP = .86
DW = 1.97
A24, I~ = 127496.5 + 1424.39 RF
(3.83)
-1394.2RPIM + .60DRE + .41RDM(t)
(-1.79) (6.54) (2.14)
iP = .84
DW = 1.76
A25, REt=-3486.92 - 265.36GDt.J
(4.60)
+ 0.57TBt-J - 054EMSS(t)
(3.59) (-1.64)
R2 = .87
DW = 2.20
A26, EN! =69.67 + 21.67RDF(t)
(3.90)
- .01 RE(t-J)- 15.46Y(t) + .46MMSS(t)
(-0.97) (-0.67) (2.66)
R? = .89
DW = 1.84
+A27, RIt = 15.70 + .03RDME(t)
(2.12)
+ .40RISD(t_J)- 0.32PD(,) + .45~
(2.81) (-3.34) (2.33)
R2 = .70
DW = 2.10 •
As in the cases of Nigeria and Ghana we also solved the
Ugandan model in both the direct linear and log-linear
forms. Table A3 contains the result of our error - analysis.
UNIVERSITY OF IBADAN LIBRARY
. 269
TABLEA3
UGANDA: ERROR ANALYSIS OF NIGERIA'S LINEAR AND
LOGARITHMIC ESTIMATIONS
ErrorType Root Mean Square RMSE Relative to
Error (RMSE) Average size of BIAS
variablSRMSE
Dependent Direct Log Direct Log Direct Log
Variable LinearFonn Fonn Linear Form Linear Fonn
Fonn Fonn
PD 22621.55 0.1308 1.3937 0.000008 15293.3 -0.0025
RDM 71.46 0.4855 0.4328 0.0029 26.73 0.0047
MSS 9609.11 0.1324 0.0899 0.000001 3.12 0.0023
RD 21.45 0.1211 0.7460 0.0042 14.58 -0.0081
X 1371.23 0.0704 0.78859 0.0008 -8527.34 +0.2975
1M 13065.85 0.6365 0.6593 0.0003 30.60 +0.1774
RE 1607.75 0.3973 0.0795 0.0002 -0.12 -0.0005
EN 845.10 0.4258 2.1827 0.0010 -551.10 0.0009
RI 4.38 0.0759 0.3014 0.0224 0.006 +0.3638
Table A3 reveals that whatever type of error one considers the direct linear form contains
more error than the log-linear form e.g. while the root mean square error is .4855 for the real
demand for money (RDM) under the log-form, it is 71.46 under the direct linear form. The Table
justifies our carrying out further analysis in the study on the basis of the log form results .
•
UNIVERSITY OF IBADAN LIBRARY
BIBWGRAPHY
Adler, M. and Dumas, B. (1983):
"International Portfolio choice and Corporation Finance: A Survey".
Journal of Finance, 38,
pp. 925-984.
Agmon, T. (1980)
"Inflation and Foreign Exchange Rates Under
Production and Monetary Uncertainty"
Journal of Finance and Quantitative Analysis
Nov. 1980 15(4) pp. 969-72.
Ajayi, S.I. (1986):
Issues of Overvaluation and Exchange Rate Adjustment In Nigeria Economic
Development Institute of The World Bank (Washington D.C.).
Ajayi, S.I. (1989):
. Exchange Rate Adjustment in Selected African
Countries. Economic Development Institute of the
World Bank, (Washington D.C.).
Alexander, S.S. (1958):
"Effects of a Devaluation on the Trade Balance".
IMF Staff Papers, April 1952,2, pp.263-278.
Alexander, S.S. (1959): .
"Effects of a Devaluation: A simplified Synthesis of
Elasticities od Abortion Approaches". American
Economic Review March, 1959,49, pp. 32-42.
Austen, J.H. (1961):
"Monetary Policy and the Forward Exchange Market",
The Journal of Finance 14, Dec. 1961.
Aliber, R.Z. (ed) 1974:
National Monetary Policies and the International
Financial Syste~ (Chicago: The Uni ersity of
Chicago Press).
UNIVERSITY OF IBADAN LIBRARY
271
Argy, V. and Porter, MG. (1972):
"The Forward Exchange Market and the Effect of
Domestic and External Disturbances Under Alternative
Exchange Rate Systems". IMP Staff papers, Nov
1972, 19 (3) pp. 503-532.
Balassa, B. (1964):
"The purchasing Power Parity Doctrine: A
Reappraisal". Journal of Pol. Eco. Vol LXXII
No.6, Dee, 1964 pp 584-596.
Balassa, B. (1974):
"Estimating the Shadow Price of Foreign Exchange
in Project Appraisal". Oxford Econ. Paper
July, 26 (2) pp. 147-168.
Balassa, B. (1987):
Effect of Exchange Rate Changes in Developing
Countries World Bank Discussion paper No.DRD 291,
May 1987.
Bilson, J.F.O. (1978):
The Monetary Approach to the Exchange Rate:
Some Empirical Evidence. IMP staff paper March
25 (1) pp. 48-75.
Black, S.W. (1973):
"International Money Markets and Flexible
Exchange Rates. Studies in International
Finance NO.32 (International Finance Section
Princeton).
Boughton, J. M. (1988):
"The Monetary Approach to Exchange Rates: What
Now Remains?". Essays in International Finance
No. 171, October .
•
UNIVERSITY OF IBADAN LIBRARY
272
Boyer, RS. (1977):
"Devaluation and Portfolio Balance". America Econ.
Rev. 67, pp. 54-63.
Boyer, RS. (1978):
"Financial Policies in an Open Economy". Economica
45 pp. 39-57.
Branson, W.H. 1977:
"Asset Markets and Relative Prices in Exchange
Rate Determination" in Reprints in International
Finance, No. 20, (Princeton University: Interna-
tional Finance Section).
Branson, W.H. and Henderson, D.W. 1985:
"The specification and Influence of Asset
Markets" in Handbook of International Economics,
vol. Il edited by Jones, R.W. and Kenen, P.B.
(Elsevier Science publishers. B.V.).
Branson, W.H. (1986):
Stabilization, Stagflation and Investment
Incentives: The Case of Kenya, 1979-1980 in
Edwards, S. and Ahamed, L. (eds); Economic
Adjustment and Exchange Rates in Developing
Countries (NBER Conference Report, Chicago:
University of Chicago Press) chapter 8,
pp.267-294.
Black, S.W. (1977):
"Exchange Policies for Less Development Countries
in a World of Floating Rates" Essays in
International Finance No.119 (International
Finance Section, Princeton).
Black, S.W. (1982): .
"The use of Monetary Policy for Internal and
External balance in Ten Industrial Countries"
in Frenkel. 1. (ed.) Exchange Rates and
UNIVERSITY OF IBADAN LIBRARY
273
International Macroeconomics (NBER, Cambridge)
Boughton (1988) - The monetary approach.
Canterberry, E.R. (1965):
Foreign Exchange, Capital Flows and Monetary
Polley. Studies in International Finance
No. 15,(International Finance Section:
Princeton).
Casell, O. (1918):
HAbnormal Deviations in International Exchanges".
Economic JournaJ Sept. 1918.
Casell, O. (1992):
Money and Foreign Exchange After 1914,
(New York).
Chibber, A. and Shafik, N. (1990):
"Exchange Reform, parallel markets and Inflation
in Africa: The case of Ghana". (paper Presented
at the World Bank Seminar, Nairobi, kenya,
June, 1-3 1990).
Collier, Pard Joshi, Y. (1991):
"Exchange Rate Policy in Developing Countries"
Oxford Review of Econ Policy vol. 5 No.3.
Coobms, C.A. (1976):
The Arena of International Finance (New York)
John Wiley & Sons).
Cooper, R.N. (1986):
"Dealing with the Trade Deficit in a Floating
Rate System", Brooklings paper Econ. Act.
1986 (1) P .195-207.
Cooper, R. R. (1987):
The International Monetary System: Essays in
World Economics (The MIT Press, 1987).
UNIVERSITY OF IBADAN LIBRARY
274
Corden, W.M. (1977):
Inflation, Exchange Rate and the World Economy
(Oxford: Clarendon Press).
Corker, R. (1989):
"External Adjustment and the Strong Yen".
IMP Staff Papers 36(2) June, pp. 464-495.
Crockett, A. and Goldstein, M. (1987):
Strenghtening the International Monetary System:
Exchange Rates Surveillance and Objective
Indicators. Occasional paper No.50
(Washington DC IMP).
Davidson, J. E. H. (1985):
"Econometric modelling of the Sterling Effective
Exchange Rate". Rev. Econ. Stud. April 52(2)
pp. 231-250.
Day, W. H. L. (1974):
"Domestic Credit and Money Ceiling Under Alternative
Exchange Rate Regimes". IMP Staff Papers Sept. 26(3)
pp.490-512.
Dervis, K., De Melo, J. and Robinson, S. (1981):
A General Equilibrium Analysis of Foreign Exchange
Shortages and Adjustment Mechanisms in a Developing
Economy. Staff Working Paper, No.443 (World Bank)
Jan.
de Macedo, 1. B. (1982):
"Portfolio Diversification Across Currencies"
in Cooper, R. N. Kenen, P. B. et al (eds.).
The International Monetary System Under Flexible
Exchange ~tes: Global, Regional and National
(Ballingeer, Cambridge) pp. 69-100.
UNIVERSITY OF IBADAN LIBRARY
275
Dordunoo, G. K. (1992)
"The Nature and Working of the Foreign Exchange
Auction market in Ghana" (paper Presented at the
Workshop of the African Ecomonic Research
Consortium, Nairobi. Kenya. May 1992).
Dornbusch. R (1976):
"Exchange Rate Expectations and Monetary Policy"
Journal of International Econs. Aug. 1976,6(3)
pp.231-244.
Dornbusch. R (1980):
"Exchange Rate Economics: Where Do We Stand?"
Brookings Paper on Econ. Act. (1) pp. 143-185.
Dornbusch, R (1988):
Exchange Rates and Inflation (The MIT Press,
Massachusetts).
Dornbusch. Rand F. Leslie, C. Helmers (1988):
The open Economy: Tools for Policy Makers in
Developing Countries. (Oxford Univ.
Press. 1988).
Donkas, J. (1985):
"The Rationality of Money Supply Expectations and
the Canadian - U.S. Exchange Rate Response to Money
Supply Announcement" Finance Review May, 20(2)
pp. 180-194.
Driskill, R (1980):
Exchange Rate Dymanics, Portfolio Balance and
Related Pric . American Economic Review Sept. 70(4)
pp.776-783.
Dervis, K. and Robinson, S. (1982):
A General Equilibrium Analysis of the causes
of a Foreign Exchange Crisis: The Case of
UNIVERSITY OF IBADAN LIBRARY
276
Turkey. Welwirtsch Arch. 118(2) pp.259-280.
Dervis, K. de Melo, J. and Robinson, S. (1981):
A General Equilibrium Analysis of Foreign
Exchange shortages in Developing Economy,
Staff Working Paper No.443, Jan. (World Bank).
Dunn, R. M. (1983):
"The Many Disappointments of Flexible Exchange
Rates" Essays in International and Finance
(Princeton University, International Finance
Section) No. 154.
Eaton, and Turnovsky, S. 1. (1982):
"Covered Interest Parity, Uncovered Interest
Parity and Exchange Rate Dynamics". Econ.
Jour. sept 93 (371) pp 555-575.
Edwards, S. (1983):
"Floating Exchange Rates, Expectations and New
Information". J. Monetary Econ. May, 11(3)
pp.321-336.
Edward, S. (1984):
"The Demand for International Reserves and
Monetary Equilibrium: Some Evidence from
Developing Countries" Rev. Econs and Statistics.
Aug. 66(3) pp. 495-500.
Edward, S. and Ahamed, L. (eds.) (1986):
Economic Adjustment and Exchange Rate in
DevelopLng Countries (NBER Conference Report:
Chicago: The Univ. of Chicago Press).
Edward, S. (1988): .
Exchange Rate Misalignment in Developing Countries.
World Bank Occasional Papers No.2, New series.
UNIVERSITY OF IBADAN LIBRARY
277
Eichengreen, B. 1. (1981):
A Dynamic Model of Tariffs, output and Employment
Under Flexible Exchange Rates. Journal Int. Econs.
Aug. 11(3) pp. 341-359.
Elbadawi, I. A. (1990):
"Inflationary Process. Stabilization and the
Role of the Public Expenditures in Uganda".
Countries Economics Department, World Bank, .
July 1990.
Enders K. and Herberg, H. (1983):
"The Dutch Disease: Causes, Consequences Cures
and Calmatives" Welwirtsch Arch, 119(3) pp. 473-497.
Feltenstein. A; Goldstein, M. and Schadler, S.M. (1979):
"A Multilateral Exchange Rate Model for Primary
Producing Countries" IMF Staff Papers Sept. 26(3)
pp.543-582.
Fleming, J.M. (1962):
Domestic Financial Policies Under Fixed and Under
Floating Exchange Rate. IMF Staff Papers Nov.
vol. 9. pp. 369-479.
Flood, R P. and Garber, P.M. (1984):
Collapsing Exchange Rate Regimes: Some Linear
Examples. Journal of Internal Econs. Aug. 17(1 va)
pp. 1-13.
Floyd, 1. E. (1964):
"International Capital Movements and Monetary
Equilibrium" American Econ. Rev., Sept., 59
pp. 472-492.
Frankel J.A. (1979): .
"On ThetMark: A Theory of Floating Exchange
Rate Based on Real Interest Differentials".
American Econ. Rev. Sept. 69(4) pp. 610-622.
UNIVERSITY OF IBADAN LIBRARY
278
Frenkel. I.A. (1976):
"A dynamic Analysis of the Balance of Payments
in a Model of Accumulation" in Frenkel J.A. and
Johnson H.G. (eds.), The Monetary Approach to the
Balance of Payments (London: Goerge Allen &
Unwin Ltd) chap. 5.
Frenkel. J.A. and Mussa, M. (1980):
"The Efficiency of Foreign Exchange Markets and
Measures of Turbulences". Amer. Econ. Rev.
May 70(2) pp. 374-381.
Frenkel. J.A. and Mussa, M. (1985):
Asset Markets, Exchange Rate and the Balance of
Payments in RW. Jones and P.B. Kenen (eds.)
Handbook of International Econimics, vol. 2
(Amsterdam. North Holland, 1985) pp.640-747.
Frenkel. lA. and Razin, A. (1987):
"The Mundell-Fleming Model A Quarter Century Leter:
A Unified Exposition" IMF Staff Papers, Dee, 34(4)
pp.576-620.
Frenkel. lA. and Rodrignez, C.A. (1975):
"Portfolio Equilibrium and the Balance of Payments:
A Monetary Approach". American Econ. Rev. 65
Sept. 1975 pp. 647-688.
Friedman. M. (1953):
Essays in Positive Economics (Chicago: Univ.
of Chicago Press).
Friedman, M. (1962):
"The Interpolation of Time Series by Related
Series" Journal of the American Statistical
Associatoion Dec., pp. 729-757.
UNIVERSITY OF IBADAN LIBRARY
279
Goldstein, M. (1984):
"Wither the Exchange Rate System"
Finance and Department, 21(2) June.
Hahn, F.H. (1977):
"The Monetary Approach to the Balance of Payment".
Journal of International Econ. Vol. 7,
pp.231-249.
Harberger, A.C. (1950):
"Currency Depreciation, Income and the Balance
of Trade". Journal of Pol. Econ. vol.58
pp.47-6O.
Harberler, G. (1949):
"The Market for Foreign Exchange and the Stability
of the Balance of Payments: A Theoretical Analysis".
Kyklos, 3, pp 193-218.
Helleiner, G. K. (1986):
"Balance of Payments, Experience and Growth Prospects
in Developing Countries: A Synthesis." World Devel.
special issue, Aug. 14(8) pp.877-909.
HempHill, J. (1974):
"The Effects of Foreign Exchange Receipts on
Imports of Developing countries". .
IMP Staff papers vol XXI No.3, November,
pp.637-693.
Hyuha, M. (1991):
"Reflections on Forex Bureau/Parallel and
official Exchange market (paper Presented at the Ministry
of Planning and Economic Development Seminar,
Kempala).
IMP, (1977) - The Monetary Approach to the Balance of Payment
(Washington DC IMP).
UNIVERSITY OF IBADAN LIBRARY
280
Jasay, A. E. (1958):
"Bank Rate or Forward Exchange Policy" Banca
Nationale de LavOTOquarterly Review XLIV, March,
1958.
Jebuni, C. D.; Sowa, N.K and Tutu, K.A. (1991):
"Exchange Rate Policy and Macroeconomic
Performance in Ghana" Research Paper No.6
(African Economic Research Consortium,
Nairobi, Kenya).
Johnson, J. G. (1972):
"The Case for Flexible Exchange Rates" in Johnson
H.G. Further Essay in Monetary Economics (Allen and
Unwin, London) pp? 198-222.
Johnson, H. G. (1976):
"The monetary Approach to Balance of Payments Theory"
in Frankel, lA. and Johnson, H.G. (eds.)
The Monetary Approach to the Balance of Payments
(London: George Allen & Unwin Ltd).
Johes, M. (1982):
"Automatic Output Stability and the Exchangement
A Multicountry Analysis", Review of Econ. Stud.
Jan. 79(1) pp. 91-107.
Kenen, P.B. (1975):
"Floats, Glides and Indicators: A comparison
of Methods for Changing Exchange Rates". Jour.
Inter. Econ. May 5(2) pp. 107-15 1.
Kenen, P. B. (1985):
"Macroeconomic Theory and Policy" in Jones R.W.
and Kene~ P.B;Handbook of International Economics,
vo1.2 (Amsterdam: North Holland) chapter 13,
pp. 625-677.
UNIVERSITY OF IBADAN LIBRARY
281
Keynes, J. M. (1923):
A Tract on Monetary Refonn (London: Cambridge University
Press).
Kindlerberger, C. P. (1937):
International Short-tenn Capital Movements (New
York; Columbia University Press).
Khan, M.S. (1977):
"The Determination of the Balance of Payments and
Income in Developing Countries" in IMP (1977).
The Monetary Approach To The Balance Of Payments
(Washington DC. IMF). Chapter 10, PP. 243-274.
Khan, M.S. (1986):
"Exchange Rate Policies of Developing Counters in
the context of External Shocks" Pakistan Dev. Rev.
Autumn 25(3) pp. 403-421.
Khan and Ki986):
"Imports Compression and Export performance
in Developing countries" World Bank Discussion paper
No.DRD 197, October.
Kouri, P.J.K. (1976):
"The Exchange Rate and the Balance of Payment in the
Short-run and in the Long-run: A Monetary Approach"
Scandinavian Journal of Econs. 78, pp. 280-304.
ouri,P.J.K. and Peter, M. (1974):
International Capital Flow and Portfolio
Equilibrium" Jour. of Pol. Econ. 82,pp. 443-467.
Kvasnicka, J.G. (1969):
"Special Drawing Rights" in Federal reserve Bank
of Chicago ( 984) Readings in International Finance
UNIVERSITY OF IBADAN LIBRARY
282
Macede de J.B. (1982):
"Portfolio Diversification Across Currencies"
in Cooper R.N.; Kanen, P.B. de Macede J.B.and
Ypersele, J. (eds)
The International Monetary System Under Flexible Exchange
Rates: Global, Regional, and National (Balinger, Cambridge)
pp.69-100.
Machlup, F. (1939):
"The Theory of Foreign Exchanges" Economica, 6, pp. 375-397.
Machlup, F. (1943):
International Trade and the National Income Multiplier
(philadelphia).
Machlup, F. (1956):
"The Terms of Trade Effects of Devaluation, UJX>nReal
Income and the Balance of Trade" KYKLOS, 9,417-452.
Makdisu, S. A. (1978):
"Flexible Exchange Rate Policy in an Open Developing
Economy: the Lebanese Experience 1950-1974". World
Development vol. 6, pp. 991-1003.
Marston, R C. (1982):
"Wages Prices and the choice Between Fixed and
Flexible Exchange Rates." Canadian Journal of Econs,
15(1) pp. 87-103.
Marris, S. (1984):
"Managing the World Economy: Will we Ever Learn"
Essays in International Finance (Princeton
University, International Finance Section)
xe.iss.
Marston, R C. (1985):
"Stabilization Politicies in Open Economies" in Jones
RW. and Kenen P.B. (eds.) Handbook of International
Economics, vol. 2, (Amsterderm: North Holland)
UNIVERSITY OF IBADAN LIBRARY
283
chap. 17. pp. 859-916.
Marwah, K. and Bodkin R. G. (1984):
"A Model of the Canadian Global Exchange Rate: A
Test of the 1970s". Journal of Policy Modelling
6(4) pp 455-483.
Mayer, T. (1989):
Economic Structure, the Exchange Rate and Adjustment
in the General Equilibrium Approach". IMP Staff
Papers 36(2), June 1989, pp. 435-463.
Meade, J. E. (1951):
The Balance of Payments (London: Oxford Univ.
Press).
Meade, J. E. (1955):
"The Case for Variable Exchange Rates" Three Bank
Review, 27,. pp. 3-28.
Metzler, L.A. (1942):
"Unemployment Equilibrium in International Trade"
Econometrica, 10 pp. 97-112.
Metzler, L.A. and Laursen, S. (1950):
"Flexible Exchange Rates and the Theory of
Employment". Review of Econs and S'tatistics
32, pp. 281-299.
Mbire and Watuwa (1992):
"Exchange Rate Policy and Inflation: The Case
of Uganda" (Report Prepared for the African
Economic research Consortium, Nairobo, Kenya).
Mckinnon, R. I. (1969):
"Portfolio Balance and International Payments
Adjustment" is Mundell R.A. and Swoboda. A. K.
(eds) Monetary Problems of the International
Economy (Chicago: Univ. of Chicago Press)
UNIVERSITY OF IBADAN LIBRARY
284
pp. 199-234.
Minford, P., Brech, M. and Mathews, K. (1980):
"A Relational Expectations Model of the U.K. Under
Floating Exchange Rates" European Econ. Review
Vol. 14, pp. 182-219.
Moran, C. (1989):
"Imports Under a Foreign Exchange Constraint"
The World Bank Review vol 3.No.2 pp 279-295.
Mundell. R A. (1960):
"The Monetary Dynamics of International Adjustment
Under Fixed and Flexible Rates" Quarterly Journal
of Econs vol. 74. May. pp. 227-257.
Mundell. R A. (1961):
"Flexible Exchange Rates and Employment Policy"
.. Canadian Journal of Econs. 27. Nov. 1961, pp. 509-517 .
Mundell, R. A. (1963):
"Capital Mobility and Stabilization Policy Under
Fixed and Flexible Exchange Rates Nov. 1963.29
pp.475-485. .
Mussa, M. (1976):
"The Exchange Rate. the Balance of Payments and
Monetary and Fiscal Policy Under a Regime of
Controlled Floating" Scandinavian Journal of
Eoons. 78. pp. 229-248.
Ndulu, B.J. (1991):
"Growth and Adjustment in Sub-Saharan Africa"
in Chibber, A. and Fischer. S. (eds.)
Economic Reform in Sub-Saharan Africa (A World
UNIVERSITY OF IBADAN LIBRARY
285
Bank Symposium).
Nguyuru, H.I. Lipumba (1990):
"The Appropriate Exchange Rate Policy for
African Countries: Some Policy Issues"
(paper Presented at the Eastern and Southern
African Economic Association's Conference
Nairobi, Kenya, June 2-3, 1990.
Obstfeld, M. (1981):
"Macroeconomic Policy, Exchange rate Dynamics and
Optimal Asset Accumulation" Journal of Political
Econ, 89, pp. 1142-1161.
Ochieng, E. O. (1988):
"Foreign Exchange Autioning in Uganda" Paper
Presented at the Seminar of the Department of
Economics ..Markarere University, Kampala).
Ojo, M. O. (1990):
The Management of Foreign Exchange Resources Under
Nigeria's Structural Adjustment Programme".
Economic and Finance Review, Central Bank of
Nigeria, vol 28, No.2, pp. 31-45.
Olisadebe, E. U. (1991):
"Appraisal of Resent Exchange Rate Policy measures
in Nigeria". Economic and Finance Review, Central
Bank of Nigeria. Vol 29, No.2.
Olofm, Akinkugbe and Ajayi (1986):
"Some Theoretical and Empirical Issues in
Conne tion with the Devaluation of the Naira".
The Nigerian Journal of Economic and Social Studies
July, 1986.
UNIVERSITY OF IBADAN LIBRARY
286
Olopoenia, R A. (1991):
"Fiscal Response to Oil Wealth and Balance of
Payments performance in Nigeria 1970-1989"
(Report Presented to the African Economic Research
Consortium, Nairobi, Kenya, Dec.).
Omoruyi, S.1. (1989):
"Structural Adjustment and National Development:
Deriving and Ensuring a Fair and Viable Exchange
Rate for the Naira" Economic and Financial Review
Central Bank of Nigeria, vol 27, No 3.
Osagie, E. (1985):
"Exchange Rate Policy in Relation to the current
Debate on the IMF Loan" (paper Presented at
Nigerian Institute of Social and Economic Reasearch
Seminar, Oct 1985).
Otani, I. and Park, Y. C. (1976):
"A Monetary Model of the Korean Economy". IMP
Staff Papers 23(1) March 1976, pp. 164-199.
Oyejide, T. A., Soyede, A. and Kayode, M.O. (1985):
Nigeria and IMF. (Nigeria: Ibadan University
Press). .
Oyejide, T. A. (1987):
"Resources Exports, Adjustment Problems and
Liberalization Prospects in Nigeria". Paper
Presented for Ford foundation Project on
Trade Policy and the Developing World.
Oyejide, T.A. (1990):
"Supply Response in the Context of Structural
Adjustment in Sub-Saharan Africa" Special Paper I
Africa Economic Research Consortium, Nairobi,
Kenya.
UNIVERSITY OF IBADAN LIBRARY
287
Oyejide, T.A. (1992):
"Impact of Price-Based and Quantity Based
Import Control measures in Nigeria" in Foreign
Trade Reforms and Development Strategy.
Pindyck, R.S. and Rubinfield, D.L. (1981):
Econometric Models and Economic Forecasts
(McGraw Hill, London).
Pinto, B. (1988):
"Nigerian During and After the Oil Boom:
A Policy Comparison with Indonesia"
The World Bank Economic Review, vol 1, No.3.
Rhomberg, R. (1964):
"A Model of the Canadian Economy Under Fixed and
Flunctuating Exchange Rates" Journal of Political
Economy vol LXXX, No 1, Feb, 1964.
Rodriguez, C. A. (1979):
Short and Long-Run Effects of Monetary and
Fiscal Policies Under Flexible Exchange Rates
and Perfect Capital Mobility, America, Econ,
Rev., 69, pp. 179-182.
Sachs, J. (1980):
"Wages Flexible Exchange Rates and Macroeconomic
Policy". Quarterly Journal, Econ, June 94(4)
pp.731-747.
Sachs, J. (1987):
"Trade and Exchange Rate Politicies in Growth
Oriented Adjustment Programmes: in Corbe,
Goldstein and Khema(eds.) Growth Oriented
Adjustment ,Programmes (lMF and The World Bank).
UNIVERSITY OF IBADAN LIBRARY