the impact of real exchange rates on economic … · that poorly managed exchange rates can be...
TRANSCRIPT
THE IMPACT OF REAL EXCHANGE RATES ON ECONOMIC GROWTH: A CASE
STUDY OF SOUTH AFRICA.
BY
KIN SIBANDA
A DISSERTATION SUBMITTED IN FULFILLMENT OF THE REQUIREMENTS OF A
MASTER OF COMMERCE
DEGREE IN ECONOMICS
DEPARTMENT OF ECONOMICS
FACULTY OF MANAGEMENT AND COMMERCE
UNIVERSITY OF FORT HARE
SOUTH AFRICA
NOVEMBER 2012
SUPERVISOR: PROFESSOR R NCWADI
ABSTRACT
This study examined the impact of real exchange rates on economic growth in South Africa. The
study used quarterly time series data for the period of 1994 to 2010. The Johansen cointegration
and vector error correction model was used to determine the impact of real exchange on
economic growth in South Africa. The explanatory variables in this study were real exchange
rates, real interest rates, money supply, trade openness and gross fixed capital formation. Results
from this study revealed that real exchange rates, gross fixed capital formation and real interest
rates have a positive long run impact on economic growth, while money supply and trade
openness have a negative long run impact on economic growth in South Africa. From the
regression results, it was noted that undervaluation of the currency significantly hampers growth
in the long run, whilst it significantly enhances economic growth in the short run. As such, the
policy of depreciating the exchange rates to achieve higher growth rates is only effective in the
short run and is not sustainable in the long run. Based on the findings of this study, the researcher
recommended that misalignment (overvaluation and undervaluation) of the currency should be
avoided at all costs. In addition, the results of the study showed that interest rates also have a
significant impact on growth and since interest rates have a bearing on the exchange rate, it was
recommended that the current monetary policy in South Africa should be maintained.
Keywords: Real exchange rates, Economic growth, South Africa.
ii | P a g e
DECLARATION
I, Kin Sibanda, the undersigned, hereby declare that this dissertation is my own original work
and that it has not been submitted, and will not be presented at any other university for a similar
or any other degree award.
Signature ……………………………………………....
Date ……../……. /……….
iii | P a g e
ACKNOWLEDGMENTS
This study would have been a fruitless exercise, if it was not for the aid, guidance and protection
of, first and foremost, The Omnipotent Lord of Hosts, God the Almighty-indeed you are good all
the time. And the steadfast contribution of the following people, to whom I feel grateful:
Firstly, I am greatly indebted to my mentor and supervisor Prof R Ncwadi for
conscientious and able research assistance. My appreciation also goes to the department
of Economics staff who provided helpful comments.
Secondly, all 2012 University of Fort Hare Economics masters students for valuable
support. Caroline, Mlambo, Mr. A and Mahali just to mention a few.
Thirdly, the government of Zimbabwe and the Govan Mbeki foundation for financial
support during my under graduate studies and post graduate studies, respectively.
Fourthly, my family and friends for their enormous support.
Last but not least I want to forward my humble and heartfelt gratitude towards the
Seventh-day Adventist Student movement for their prayers and encouragement.
iv | P a g e
DEDICATION
“Other things may change us, but we start and end with family” Anthony Brandt.
To my late father, Mr. Fredrick Sibanda, my loving mother Mrs. Madeline Sibanda, my two
wonderful brothers-Freeman, Innocent and my two ardent sisters-Winnie, Agifa and my lovely
grandmother Lillian.
v | P a g e
LIST OF ACRONYMS
ADF Augmented Dickey-Fuller
BS Balassa-Samuelson hypothesis
COSATU Congress of South African Trade Union
DTI Department of trade and industry
DW Durbin-Watson
ELGH Export-led growth hypothesis
EU European Union
FCF Fixed capital formation
GDE Gross domestic expenditure
GDP Gross Domestic Product
GEAR Growth, Employment and Redistribution Policy
GLS Generalized least squares
JB Jarque-Bera test
LDC Least Developed Countries
M Imports
MS Money supply
OECD Organisation for Economic Co-operation and Development
OLS Ordinary Least Squares
PP Phillips-Perron
REER Real effective exchange rates
vi | P a g e
RIR Real interest rates
SA South Africa
SACU Southern African Customs Union
SADC Southern African Development Community
SARB South African Reserve Bank
StatsSA Statistics South Africa
TFP Total Factor Productivity growth
OP Trade openness
VAR Vector Auto-regression
VECM Vector Error Correction Modeling
WTO World Trade Organisation
X Exports
vii | P a g e
LIST OF TABLES
Table 3.1 South Africa: Exchange rate regime changes ............................................................... 26
Table 3.2 Export share of products in the South African export basket to the rest of the world
(2000-2010)................................................................................................................................... 37
Table 3.3 Export share of products in the South African export basket expressed in percentages
(2000-2010)................................................................................................................................... 39
Table 3.4 Import share of selected products in the South African import basket 2006-2010 ...... 40
Table 5.1(a): Stationarity results of the Augmented Dickey-Fuller test. ...................................... 64
Table 5.1 (b): Stationarity results of the Phillips-Perron test........................................................ 65
Table 5.2: Pair-wise correlation results ........................................................................................ 67
Table 5.3: Lag order selection criteria .......................................................................................... 68
Table 5.4(a): Cointegration Rank Test (Trace) ............................................................................. 69
Table 5.4(b): Cointegration Rank Test (Maximum Eigenvalue) .................................................. 69
Table 5.5: Results of the long run cointegration equation ............................................................ 72
Table 5.6: Error correction model results ..................................................................................... 73
Table 5.7: Diagnostic checks results............................................................................................. 75
Table 5.8: Variance decomposition of GDP ................................................................................. 78
viii | P a g e
LIST OF FIGURES
Figure 2.1: Real exchange rates and inflation in South Africa 2005-2010................................... 12
Figure 3.1: Trends in real effective exchange rates: Average for the period - 15 trading partners
1994-2009 ..................................................................................................................................... 28
Figure 3.2: Economic growth trends in SA: 1994 2010 ............................................................... 31
Figure 3.3: Trends in real effective exchange rates and real GDP in percentage changes 1994-
2009............................................................................................................................................... 32
Figure 3.4. Trends in trade openness in South Africa 1994-2010 ............................................... 35
Figure 3.5 Investment and Gross Domestic Product 1994-2010 ................................................... 41
Figure 3.6 Real interest rates and Gross Domestic Product 1994-2010 ........................................ 43
Figure 3.7 Money supply (M3) and Gross Domestic Product 1994-2010 ..................................... 45
Figure 5.1(a) Plots of variables in levels for 1994 – 2010............................................................ 62
Figure 5.1 (b) Plots of first differenced variables for 1994-2010 ................................................. 63
Figure 5.2: Cointegration vector ................................................................................................... 71
Figure 5.4: Impulse response of GDP ........................................................................................... 75
ix | P a g e
Contents ABSTRACT ................................................................................................................................................... i
DECLARATION .......................................................................................................................................... ii
ACKNOWLEDGMENTS ........................................................................................................................... iii
DEDICATION ............................................................................................................................................. iv
LIST OF ACRONYMS ................................................................................................................................ v
LIST OF TABLES ...................................................................................................................................... vii
LIST OF FIGURES ................................................................................................................................... viii
CHAPTER ONE ........................................................................................................................................... 1
1.1 BACKGROUND OF THE STUDY ....................................................................................................... 1
1.2 STATEMENT OF THE PROBLEM ...................................................................................................... 3
1.3 OBJECTIVES OF THE STUDY ............................................................................................................ 4
1.4 HYPOTHESIS ........................................................................................................................................ 5
1.5 METHODOLOGY ................................................................................................................................. 5
1.6 SIGNIFICANCE OF THE STUDY ........................................................................................................ 5
1.7 ORGANIZATION OF THE STUDY ..................................................................................................... 5
1.8 SUMMARY ............................................................................................................................................ 6
CHAPER TWO: LITERATURE REVIEW ................................................................................................. 7
2.1 INTRODUCTION .................................................................................................................................. 7
2.2 THEORETICAL LITERATURE ........................................................................................................... 7
2.2.1 The Traditional Approach to Exchange Rates ............................................................................. 7
2.2.2 The Structuralist Approach ........................................................................................................ 11
x | P a g e
2.2.3 Balassa-Samuelson Hypothesis .................................................................................................. 15
2.2.4 Export-led Growth Hypothesis .................................................................................................. 16
2.3 EMPIRICAL LITERATURE ............................................................................................................... 18
2.3.1 Empirical Literature from Developed Countries ........................................................................ 18
2.3.2 Empirical Literature from Developing Countries ...................................................................... 20
2.4 ASSESSMENT OF LITERATURE ..................................................................................................... 23
2.5 SUMMARY .......................................................................................................................................... 24
CHAPTER THREE: TRENDS IN MACROECONOMIC VARIABLES IN SOUTH AFRICA ............. 25
3.1 INTRODUCTION ................................................................................................................................ 25
3.2 EXCHANGE RATE MANAGEMENT ............................................................................................... 25
3.2.1 Exchange Rate Policy in South Africa 1994-2010 .................................................................... 26
3.2.2 Trends in Real Exchange Rates in South Africa from 1994-2010 ............................................. 28
3.3 ECONOMIC GROWTH IN SOUTH AFRICA .................................................................................... 30
3.3.1 Exchange Rates and Economic Growth in South Africa ........................................................... 32
3.4 TRADE OPENNESS IN SOUTH AFRICA ......................................................................................... 33
3.5 GROSS CAPITAL FORMATION AND GROSS DOMESTIC PRODUCT (1994-2010) .................. 41
3.6 REAL INTEREST RATES AND GROWTH (1994-2010) .................................................................. 43
3.7 MONEY SUPPLY AND GROWTH (1994-2010) ............................................................................... 44
3.8 SUMMARY .......................................................................................................................................... 46
CHAPTER FOUR: RESEARCH METHODOLOGY ................................................................................ 48
4.1 INTRODUCTION ................................................................................................................................ 48
4.2 MODEL SPECIFICATION .................................................................................................................. 48
xi | P a g e
4.3 DEFINITION OF VARIABLES .......................................................................................................... 49
4.3.1 Priori Expectations ..................................................................................................................... 51
4.4 DATA SOURCES ................................................................................................................................ 51
4.5 RESEARCH TECHNIQUES ................................................................................................................ 51
4.5.1 Testing for Stationarity .............................................................................................................. 51
4.5.2 Augmented Dickey-Fuller (ADF) test ...................................................................................... 52
4.5.3 Phillips-Perron (PP) Tests .......................................................................................................... 54
4.5.4 Cointegration and Vector Error Correction Modeling (VECM) ................................................ 54
4.5.5 Johansen Technique Based on VARS ........................................................................................ 56
4.6 IMPULSE RESPONSE ANALYSIS .................................................................................................... 57
4.7 VARIANCE DECOMPOSITION ANALYSIS .................................................................................... 57
4.8 DIAGNOSTIC CHECKS ..................................................................................................................... 57
4.8.1 Heteroscedasticity ...................................................................................................................... 58
4.8.2 Residual Normality Test ............................................................................................................ 58
4.8.3 Autocorrelation LM Tests .......................................................................................................... 58
4.8.4 Misspecification Tests................................................................................................................ 59
4.9 SUMMARY .......................................................................................................................................... 59
CHAPTER FIVE: PRESENTATION AND ANALYSIS OF EMPIRICAL FINDINGS .......................... 60
5.1 INTRODUCTION ................................................................................................................................ 60
5.2 UNIT ROOT/STATIONARITY TEST RESULTS .............................................................................. 60
5.3 TESTS FOR COINTEGRATION ........................................................................................................ 66
5.4 VECTOR ERROR CORRECTION MODEL (VECM) ....................................................................... 71
5.5 DIAGNOSTIC CHECKS ..................................................................................................................... 74
xii | P a g e
5.6 IMPULSE RESPONSE ANALYSIS .................................................................................................... 75
5.7 VARIANCE DECOMPOSITION ANALYSIS .................................................................................... 77
5.8 SUMMARY .......................................................................................................................................... 78
CHAPTER SIX: CONCLUSIONS, POLICY RECOMMENDATIONS AND LIMITATIONS ............... 80
6.1 SUMMARY OF THE STUDY AND CONCLUSIONS ...................................................................... 80
6.2 POLICY IMPLICATIONS AND RECOMMENDATIONS ................................................................ 81
6.2.1 Exchange Rate Policy ................................................................................................................ 82
6.2.2 Investment policy ....................................................................................................................... 82
6.2.3 Monetary Policy ......................................................................................................................... 83
6.2.4 Trade Policy ............................................................................................................................... 83
6.3. LIMITATIONS OF THE STUDY AND AREAS FOR FURTHER RESEARCH ............................. 84
6.4 SUMMARY .......................................................................................................................................... 84
REFERENCES ........................................................................................................................................... 85
APPENDICES ............................................................................................................................................ 94
APPENDIX I: South African data used in regression ................................................................................ 94
APPENDIX 2: Editor’s Declaration ........................................................................................................... 94
1 | P a g e
CHAPTER ONE
INTRODUCTION AND BACKGROUND OF THE STUDY
1.1 BACKGROUND OF THE STUDY
The relationship between the real effective exchange rate and economic growth is fast becoming
an important area of study in both the developing and developed countries (Akpan, 2008).
Edwards and Garlick (2007) assert that the exchange rate plays a central role in public debate
around trade and trade policy in South Africa, with widespread calls for appreciation,
depreciation or simple stabilization. Rodrick (2007) concurs that economists have long known
that poorly managed exchange rates can be disastrous for economic growth. The real exchange
rate thus, serves as an international price for determining the competitiveness of a country.
Takaendesa (2006) explains that the real exchange rate plays a crucial role in guiding the broad
allocation of production and spending in the domestic economy between foreign and domestic
goods.
From 1994, South Africa has been suffering from high inflation levels, declining output and
unemployment. In an attempt to ensure price stability, the monetary authorities adopted the
inflation targeting regime to keep inflation parity with its major trading partners. This therefore,
meant that the country abandoned the fixed exchange rate regime for a market-based, free
floating exchange rate (Van der Merwe, 1996). The adoption of a free float exchange rate system
exposed the economy to the dangers of exchange rate misalignment. Misalignment according to
Montiel and Serven (2008) is the deviation of the actual or observed real exchange rate from the
equilibrium real exchange rate. Misalignment can either be an overvaluation or undervaluation of
the currency. Developing countries are often guilty of the latter, which negatively affects the
export sector and ultimately economic growth. With regard to this, Takaendesa (2006) argues
that emerging economies in particular, are encouraged to conduct their policies so as to get this
macroeconomic relative price right; that is, an exchange rate that does not stray too far from its
equilibrium value.
South Africa, like many emerging market economies, is an open economy which participates
heavily in international trade. As such it depends on imported capital goods and specialises in
2 | P a g e
commodity exports. In order to gain from this trade, it is important for the country to maintain a
very competitive exchange rate, one that is neither too weak nor too strong. An overvalued
currency can be dangerous to the economy of South Africa. Old mutual (2009) argues that a
strong rand negatively affects exports. The rationale being that a strong rand makes exports
expensive and imports cheap which contributes to an import boom which in turn deteriorates the
current account of the balance of payments. The exchange rate is also linked with the
manufacturing activities in South Africa. The manufacturing circle chairperson Stewart Jennings
was quoted as saying “the manufacturing sector has declined from contributing 25% to South
Africa’s gross domestic product during its heydays in the 1960s, to only about 15% in 2011”
(Prinsloo, 2011). This was blamed on what was called an overvalued rand and hence the calls
from the sector to devalue the currency.
Proponents of a weaker currency believe that real depreciation is beneficial to the economy.
Edwards and Garlick (2007) explains that real depreciation enhances export competitiveness,
encourages export diversification, protects domestic industries from imports and ultimately
improves the trade balance; this in turn promotes economic growth. Accordingly Berry (2006)
states that:
“The policy prescription that encourages countries to devalue their currencies in order to
stimulate aggregate demand remains enshrined in the orthodox tradition of stabilisation
policies since the 1960s. In the 80s and 90s, the policy of devaluation in stabilization
programmes gained more impetus against the backdrop of the first-tier South East Asian
countries, whose export-oriented strategies were supported by maintaining competitive
exchange rates through frequent devaluations” (Berry, 2006 in Ngandu and Gebreselasie,
2006)
Most countries are now recognising the fact that competitive exchange rates are an important
macroeconomic instrument in ensuring low inflation levels, promoting exports and enhancing
economic growth. The South African government through Growth, Employment and
Redistribution Policy (GEAR) undertook the relaxation of exchange controls in order to aid the
lowering of inflation. In the new growth path, the policy consensus focuses on the need for a
more competitive and stable exchange rate. According to Treasury (2012) the government in the
new growth path seeks to promote more active monetary policy interventions to achieve growth
3 | P a g e
and jobs targets through a more competitive exchange rate and a lower cost of capital. As such,
Tarawalie (2010) explains that in recent years, policy discussions sought to ensure exchange rate
stability and correct exchange rate-alignment as essential elements in the improvement of
economic performance.
Real exchange rates play a vital role in foreign trade and economic development. It is apparent
that changes in real exchange rates (either depreciation or appreciation) have wider and far
reaching economic effects. It is therefore very important to understand how exchange rates affect
economic growth in South Africa. One of the macro economic objectives of South Africa is to
maintain sustainable economic growth. High economic growth helps to maintain an adequate
level of foreign reserves and to create and maintain a sustainable, internationally competitive
exporting sector that will contribute to job creation and high incomes.
1.2 STATEMENT OF THE PROBLEM
South Africa’s fiscal discipline and relatively conservative economic policy since 1994 have
failed to produce the expected results for growth and employment (Rodrik, 2008). In view of the
policy framework adopted by the South African government, it is appropriate to ask how this has
transformed the wellbeing of the economy. High levels of unemployment and sluggish economic
growth can be recognised as vexing problems that have challenged South Africa even sixteen
years into democracy1. In light of the government’s exchange rate policy over the years, the
questions that basically come to mind concern the performance of the economy in relation to
economic growth and unemployment. Does the exchange rate matter in South Africa when it
comes to improving economic growth and alleviating unemployment? What are the
consequences of a stronger or weaker rand on the economy? This study aims to provide answers
to these questions by examining the impact of real exchange rate on economic growth in South
Africa.
Theories and empirical evidence present different conclusions regarding the relationship between
economic growth and exchange rates. The existing discourse about the impact of real exchange
rate strength in South Africa has its core in the contest between the views of the Washington
1 “South Africa’s unemployment rates have increased rapidly since the move to democracy and are now higher than
they were in 1993”( Nchimunya , 2011).
4 | P a g e
Consensus versus Rodrik’s (2007) conclusions2 (Mishi, 2011). These contrasting views have
their roots in theoretical literature. The traditional approach to exchange rate holds that
devaluation has expansionary effects on the economy (Salvatore, 2005). The Structuralist
approach to exchange rates, on the other hand, is equally convincing that devaluation is
contractionary to expansion in the economy (Acar, 2000). Empirical evidence from different
countries including South Africa using different methodologies also have conflicting conclusions
about the relationship (Akpan, 2008, Acar, 2000).
The lack of consensus in literature (both theoretical and empirical) justifies the contradictory
calls by different stakeholders in South African economy on how to deal with the buoyant Rand:
the manufacturing sector (tradable goods sector), the monetary authorities and labour bodies like
Congress of South African Trade Union- COSATU (concerned by high unemployment levels).
The call to devalue the currency comes with some costs related to interference with the market
forces. Since South Africa is an open economy, real exchange rates are a very crucial variable in
the growth process. In the light of this information, the study seeks to find out if the real
exchange rates directly and strongly influence the level of exports and economic growth. Failure
to find the true relationship might result in wrong policy formulations and resources wastage.
1.3 OBJECTIVES OF THE STUDY
The broad objective of this study is to assess impact of real exchange rates on economic growth
in South Africa between the period 1994 and 2010. The specific objectives of the study are as
follows:
To provide a review of the trends in exchange rates and economic growth in South Africa
over the period 1994-2010.
To evaluate the impact of exchange rates on economic growth in South Africa during the
period 1994-2010.
To make policy recommendations based on the findings.
2The Washington Consensus view holds that real exchange misalignment causes imbalances which are bad for
growth, (Williamson, 2004). In contrast, Rodrik (2007) argues that “undervaluation relative to purchasing power
parity is good for growth since it promotes the tradable sector and overvaluation destroys exports and hence
economic growth”.
5 | P a g e
1.4 HYPOTHESIS
H0: Real exchange rates have a significant impact on economic growth.
H1: Real exchange rates do not have a significant impact on economic growth.
1.5 METHODOLOGY
By applying the Vector Error correction modeling (VECM) approach, this study empirically
investigates the impact of real exchange rates on economic growth in South Africa. In testing for
the unit root properties of the time series data, the variables were subjected to the Augmented
Dickey-Fuller (ADF) and Philips-Peron unit root test. The cointegration and vector error
correction modeling (VECM) by Johansen (1991, 1995) was employed.
The study made use of diagnostic tests such as the residual normality test, heteroscedacity,
autocorrelation tests and Ramsey test for misspecification in order to validate the parameter
estimation outcomes achieved by the estimated model.
1.6 SIGNIFICANCE OF THE STUDY
This study explores how real exchange rates have impacted on economic growth in South Africa
during the period between 1994 and 2010. In this regard, the study serves to shed light on the
link that exists between exchange rates and economic growth. Neither economic theory nor
empirical evidence provides clear-cut answers to the question of how real exchange rates affect
economic growth. The link between exchange rates and economic growth has stimulated a lot of
debate in South Africa and abroad. Many researchers are investigating this relationship, and the
results of the empirical literature on the relationship between real exchange rates and economic
growth are inconclusive. It is on this basis that this study adds value to the ongoing discourse on
exchange rates and economic growth. The results of this study will hopefully make a
contribution towards policy planning and formulation at all levels of government, in the banking
sector, the business community, as well as labour unions.
1.7 ORGANIZATION OF THE STUDY
Chapter 1 provides the introduction and background to the study. Chapter 2 reviews both the
theoretical and empirical literature pertaining to the relationship between the real exchange rates
6 | P a g e
and economic growth in South Africa. Chapter 3 gives an overview of the trends in the link
between exchange rates and economic growth in South Africa between 1994 and 2010. Chapter
4 presents a discussion on methodology. Chapter 5 presents the estimation techniques and
interpretation of the results. Chapter 6 presents a summary, conclusions and policy
recommendations.
1.8 SUMMARY
This chapter presented the introduction and background to the study. The research problem was
outlined together with the aims and objectives of the study. The method of research to be applied
in this study was provided and the deployment of the study was outlined.
Having outlined the conceptual framework of the study in this introductory chapter, the scene is
set to present the theoretical framework of the study. The theoretical and empirical literature is
provided in the next chapter.
7 | P a g e
CHAPER TWO
LITERATURE REVIEW
2.1 INTRODUCTION
In order to provide a conceptual framework and appropriate policy recommendations in this
study, it is important to present a theoretical framework which underpins this study. In addition
to the various theories that will be discussed in this chapter, empirical literature is also presented.
The aim of presenting empirical literature is to explore work done by others and the various
methods of research applied in this field in order to identify any existing gaps in literature. This
chapter is divided into two sections. The first section deals with theories that explain the
relationship between exchange rates and economic growth and the second section deals with
empirical literature. The assessment of literature and concluding remarks are provided towards
the end of the chapter.
2.2 THEORETICAL LITERATURE
This section examines theories that deal with exchange rates. The theories discussed in this
section include the traditional exchange rate approach, the Structuralist approach, the Balassa-
Samuelson Hypothesis and export led hypothesis.
2.2.1 The Traditional Approach to Exchange Rates
This view holds that depreciation have expansionary effects on economic growth. This is
popularly known as the traditional view. This approach holds that devaluation of a currency will
cause local goods to be cheaper abroad and this will increase their demand, leading to an
increase in exports (Salvatore, 2005). The view that devaluation has expansionary effects on
output is evident in that devaluation of the currency improves trade balance, alleviates balance of
payments difficulties and accordingly expands output and employment (Acar, 2000). The case
for depreciation is that when a country depreciates its currency, it enhances the cost
competitiveness of its exports which are a component gross domestic product.
8 | P a g e
The traditional approach to exchange rates assumes that exchange rates affect economic growth
through two main channels; the Total Factor Productivity growth channel and the Capital
accumulation channel. An analysis of these two follows below.
2.2.1.1 The total factor productivity growth channel
The Total Factor Productivity growth channel holds that currency depreciation shifts the output
composition of a country from the production of non-traded goods to traded goods. The link
from output composition to growth is through economy-wide productivity improvements,
generated by the production of some types of traded goods (exported manufactured goods)
through mechanisms such as technology and skill transfers associated with learning by doing that
is external to the firm (Montiel and Serven, 2008). This shift to the production of traded goods
and improvements technology results in an increase in investments locally, exports and
ultimately economic growth.
The Total Factor Productivity growth channel places the structure of domestic production at the
core of the analysis (Eichengreen, 2008). A depreciated real exchange rate, equivalent to an
increase in the price of tradables relative to non tradables, improves the profitability of the
tradable sector. As production moves from the non tradables to the tradable sector characterized
by higher (marginal social) productivity levels the overall productivity in the economy increases.
Such economy-wide productivity improvement ultimately fosters growth (Mbaye, 2012). Acar
(2000) explains that currency devaluation switches demand from imports to domestically
produced goods by increasing the relative prices of imports. Export industries on the other hand
become more competitive in international markets by stimulating domestic production of
tradable goods and inducing domestic industries to use more domestic inputs.
Dabla-Norris and Floerkemeier (2005) are of the view that changes in exchange rates have an
effect on aggregate demand through improvements in international competitiveness as well as
net exports. In other words, when the rand weakens against other currencies, local people opt for
domestic goods thus improving economic performance of the manufacturing sector. At the same
time the country exports more and imports less resulting in favourable net exports. Ngandu and
Gebreselasie (2006) further explain that increased exports, through the multiplier effect, are
expected to increase aggregate demand, and ultimately domestic production and employment.
9 | P a g e
Given that depreciation tends to be inflationary, it is argued that the increase in the overall price
level leads to the lowering of the real wage, which leads to more hiring and increased production,
assuming that there is unemployment in the economy.
The criticism leveled against the Total Factor Productivity growth channel is that it does not give
clear-cut ways in which it operates except for the learning by doing assumption. Also it lacks
empirical support; Mbaye (2012) for instance argues that there is no empirical investigation on
the TFP transmission channel of the effect of real exchange rate undervaluation on growth. The
second channel under the traditional approach which seems to be increasingly common in policy
circles links a depreciated real exchange rate to growth through effects on the domestic saving
rate (Montiel and Serven, 2008). Under this channel a depreciated real exchange rate increases
the domestic saving rate which in turn stimulates growth by increasing the rate of capital. This
channel is discussed below.
2.2.1.2 Capital accumulation growth channel
The capital accumulation approach is a view that holds that exchange rates affect economic
growth through their effect on savings. This approach claims that real exchange rate
undervaluation enhances growth through an increase in the capital stock of the economy as a
whole (Mbaye, 2012). A backdrop for this view is that a depreciated real exchange rate has a
tendency to increase the domestic saving rate. Higher saving rates induced by depreciation
stimulate growth by increasing the rate of capital accumulation.
According to Montiel and Serven (2008) while there is no consensus on the precise channels
through which this effect is generated, an increasingly common view in policy circles points to
saving as the channel of transmission, with the claim that a depreciated real exchange rate raises
the domestic saving rate, which in turn stimulates growth by increasing the rate of capital
accumulation. The capital accumulation channel holds that there are two sources of capital
accumulation. Mbaye (2012) explains that in the first mechanism, the capital accumulation
operates exclusively in the tradable goods sector whose share in GDP increases while in the
second, the stock of capital in the economy increases through the expansion of overall savings
and investment.
10 | P a g e
The relationship between a depreciated real exchange rate and the saving rate arises because a
depreciated real exchange rate tends to shift aggregate demand away from traded to non-traded
goods, requiring an increase in the real interest rate to maintain internal balance (Montiel and
Serven, 2008). The increase in interest rates constrains aggregate demand in part by raising the
domestic saving rate. Thus from this perspective, causation runs from the real exchange rate
through the real interest rate to the saving rate which in turn increases economic growth. Dooley,
et al. (2004) argue that high saving rates in many Asian countries, including China, are at least in
part the result of the pursuit of such an exchange rate policy (depreciated currency) in their
export driven strategies.
In understanding the capital accumulation channel of exchange rates, it is important to note that
there are two conceptual links in the causal chain that underlies the capital accumulation channel.
The first being the real exchange rate to the saving rate, and second, from the saving rate to
growth (Montiel and Serven, 2008). The second link of saving and economic growth is familiar
and has long been an underpinning of mainstream growth process. The first link of exchange
rates and savings, however, is controversial both theoretically and empirically. Montiel and
Serven (2008) further argue that if the real exchange rate is adopted as a policy target, an
improvement in a country’s current account balance resulting from depreciation, increases the
country’s saving rate relative to its rate of investment.
A more depreciated real exchange rate can also result in higher saving through a different
channel (Levy-Yeyati and Sturzenegger, 2007). Thus, a more depreciated real exchange rate will
result in firms paying lower wages in real terms. Lower wages in turn reduce costs of production,
inducing firms to invest more. Consequently, firms increase their savings to finance their
additional investment, which ultimately raises the aggregate savings. The validity of the capital
accumulation channel is subject to debate, for instance, Bernanke (2005) argues that causation
runs from a high saving rate to a depreciated exchange rate not the other way round. The
rationale behind this reasoning is that, a high saving rate tends to depress domestic demand. In
order to sustain internal balance, countries maintain a depreciated real exchange rate. If this view
is correct, an empirical correlation between exchange rates and savings cannot be interpreted as
the capital accumulation channel at work. Even if the causation runs in the correct direction from
exchange rates to savings, the second link of savings to growth might not result because the same
11 | P a g e
high real interest rate that induces a higher domestic saving rate would also tend to discourage
domestic investment. Montiel and Serven (2008) believe that the existence of a link between the
real exchange rate and the saving rate, as well as the interpretation of that link if it exists is
questionable.
The traditional approach poses a challenge when it comes mostly to less economically developed
countries which rely mostly on imported capital goods and infrastructure. In as much as
devaluation of the currency makes exports cheaper for external buyers, imports in contrast
become dearer for local buyers. Local firms which rely on imported capital goods pass the extra
cost to customers in terms of high prices thus leading to inflation. Failure to consider this
challenge in part resulted in the formulation of the Structuralist approach to exchange rates. Acar
(2000) explains that consensus on the view that devaluation leads to output expansion was
broken at the end of the 1970s. An alternative line of approach emerged, which raised the
possibility that depreciation could be contractionary, especially in developing countries. This
approach is referred to as the Structuralist approach.
2.2.2 The Structuralist Approach
Contrary to the traditional approach, this view argues that currency depreciation might have a
contractionary effect on output and employment, especially for less economically developed
countries. This approach shows how currency depreciation might cause a reduction in output.
Depreciation of the currency can cause a contractionary effect on output in many ways, but the
increase in the price of imports it causes is an important issue and requires much attention.
Depreciation increases the cost of imports in particular, and the cost of domestic production in
general, through imported inputs (Acar, 2000). If the costs of inputs rise, it is possible that they
may be a rise in the cost of production and firms will pass this on to their prices. This is because,
firms can only get rid of an increase in the cost of production by increasing their prices. This
increases the general price level. Acar (2000) notes that “decreasing imports in this context
imply insufficient inputs necessary for production. Eventually, because of the lack of enough
inputs and increasing costs, production will slow down, leading to a contraction in total supply”.
In this case, depreciation would be contractionary in that it causes a slowdown or decrease in the
growth rate of output in the economy.
12 | P a g e
This view seems to be supported by South African data. The inflation rate3 followed the rand
exchange rate4 from 2005 to 2010. This is shown by the figure below.
Figure 2.1: Real exchange rates and inflation in South Africa 2005-2010
Source: data compiled from SARB online bulletins (2012)
Figure 2.1 above confirms the Structuralist point of view that currency depreciation or
depreciation is inflationary. From year 2005 to year 2008 the rand was weakening and that was
followed by year on year increase in inflation for the same period. The exchange rate of the rand
strengthened from 7.8 to 12.3 in 2009 and 2010, respectively. This was followed by decrease in
inflation from 7.8 in 2009 to 4.6 in 2010. The rand exchange in the year 2008 was at the lowest
and the inflation rate was at the peak. From 2009 onwards the rand started to appreciate inducing
a decrease in inflation. Generally, as confirmed by Figure 2.1 above depreciation is inflationary.
This increase in inflation according to the Structuralist approach is mainly imported inflation
whereby expensive imported capital goods increase the cost of production which will later be
passed on to customers as high prices. Acar (2000) explains that in the case of price rigidities,
prices of non-traded goods will adjust slowly. In this case, total output will decrease if demand
for non-traded goods decreases by more than the rise in net foreign demand for traded goods.
3 Consumer price index
4 Real effective exchange rate of the rand: Average for the period - 15 trading partners - Trade in manufactured
goods 1-Term % change
0
5
10
15
-15 -10 -5 0 5
10 15
2005 2006 2007 2008 2009 2010
CP
I
RE
ER
YEAR
REER and CPI 2005-2010
RER CPI
13 | P a g e
2.2.2.1 Channels through which depreciation may create negative effects on aggregate
demand
For the Structuralist approach to hold, there are channels through which it operates. These
channels show how the exchange rate depreciation causes a reduction in aggregate demand and
ultimately national output. These are discussed below:
2.2.2.1.1 Reduction in real wealth or real balances
As a result of depreciation, prices of traded goods increase relative to non-traded goods. This
leads to an increase in the general price level. As prices rise, real money balances (M/P) decline
(Acar, 2000). The larger the share of traded goods in the consumption, the more severe the
increase in general price level and decrease in real money supply. As real money balances go
down, expenditures also fall. Ngandu and Gebreselasie (2006) state that in order to restore real
balances a fall in expenditure is needed thereby lowering consumption demand and providing an
offsetting contractionary effect on output. The lowering of consumption also implies that the
increase in import prices of final consumer goods will diffuse to the consumer price index.
2.2.2.1.2 New investment constrained by rising prices of imported machinery
A depreciating currency more often than not limits investment of capital and equipment which
are usually imported in the case of less economically developed countries. As discussed above,
after a depreciation of the rand, imported goods become expensive and exported goods cheaper.
Acar (2000) asserts that this channel seems to be applicable to most developing countries.
Developing countries during the growth phase import much of the capital goods and
infrastructure abroad and the expenditure is too high when the currency is weaker. Depreciation
on the contrary, may stimulate domestic capital and equipment industries if supported by right
policies.
2.2.2.1.3 The time lag in inducing non-traditional tradables may be too long
Traditional exports of developing countries frequently lie in sectors that offer unattractive
demand prospects and limited inter-sectoral linkages, such as agriculture and minerals. This
means that there may be limited potential for expansion in existing industries (Schweicker,
14 | P a g e
Thiele and Wiebelt, 2006). Currency depreciation might not stimulate investment into new non-
traditional exports, except it is generated through foreign direct investment.
2.2.2.1.4 Increased debt and debt service payments in local currency
One of the critical factors that play a role in macro-economic difficulties of less economically
developed countries is the existence of a large amount of accumulated external debt stock and
the interest burden on it (Caves et al. 1996). For a country that has accumulated external loans
denominated in foreign currency, a devaluated local currency means that in real terms the
country and businesses pay more than the real worth of the debt, thus influencing the country’s
debt servicing capacity. When the currency is depreciating, it is hard for South Africans and the
government to pay their debts denominated in external strong currencies.
2.2.2.1.5 Wage indexation based on foreign and domestic price levels
Increased prices for tradables caused by depreciation may lead labour to demand higher wages,
which could produce adverse supply effects. Ngandu and Gebreselasie, (2006) agree that the
possibility of an inflationary impact of depreciation might lead labour to demand higher wages.
If wages are flexible they will adjust to the new prices following the depreciation. Similarly, if
there exists a wage indexation mechanism, which automatically increases nominal wages in
proportion to price changes, then production costs will increase through higher wages (Acar,
2000). An increase in wages contributes to higher input costs causing a reduction in production,
causing output to contract. This might not apply to the South African context because of high
unemployment levels and wage rigidities.
2.2.2.1.6 Import cost channel
Acar (2000) contends that currency depreciation usually takes place when countries have a
foreign trade deficit and related external balance difficulties. The opinion holds mostly for
developing countries that usually run unfavorable trade balances. The effect of currency
depreciation on aggregate demand for a country that has a trade deficit is negative. Following the
depreciation, given that imports exceed exports, price increases of traded goods reduce the home
country’s real income and raise the real income of the outside world, since foreign exchange
payments (import costs) exceed foreign exchange receipts (export revenues). Schweicker et al.
15 | P a g e
(2006) explain that within the home country the value of foreign savings goes up, aggregate
demand goes down, and imports fall along with it. The larger the initial deficit, the greater the
contractionary outcome.
The Structuralist approach to exchange rates in a nutshell is a view that holds that exchange rates
depreciations can be contractionary. This approach explains that the contractionary effect is due
mainly to an increase in price levels through a number of channels. Given this approach, a policy
to depreciate the currency can end up contradicting macro policies that seek to stabilise the
macro-economy through a reduction in inflation.
2.2.3 Balassa-Samuelson Hypothesis
The Balassa-Samuelson Hypothesis follows the work of Balassa and Samuelson of 1964 who
gave a theoretical explanation of the long run trends in the real exchange rates (RER). According
to Solanes and Flores (2009) the basis of the Balassa-Samuelson Hypothesis is that there is a
positive correlation between the relative economic growth and real exchange rate. It is therefore
anticipated that fast-growing countries usually experience real exchange rate appreciations as
opposed to slow-growing countries. This theory is in contrast with the commonly held view that
depreciation enhances economic growth. It has been subjected to numerous empirical tests5 and
found to be relevant in some counties, for example Japan and Korea.
The Balassa-Samuelson Hypothesis is based on the following assumptions: a) there are two
sectors in the economy that produce tradable and non-tradable goods, respectively, with the
same production function; b) the prices of tradable goods and the interest rate are determined in
the world market; c) Purchasing Power Parity holds in the tradable sector; d) labour is perfectly
mobile across sectors inside the country, but less mobile between countries; e) wages are led by
developments in the tradable sectors, and then translated to the non-tradable sector (wage
equalisation across sectors). Given the above assumptions the Balassa-Samuelson Hypothesis is
such that:
1. The differentials in productivity growth rates between the tradable goods and non-
tradable goods sectors cause relative price changes.
5 Hsieh (1982), Marston (1987) and Drine and Rault (2003)
16 | P a g e
2. The ratio of non-tradable prices to tradable prices is higher in a fast growing country.
3. The ratio of tradable prices across countries remains constant.
4. A combination of 2 and 3 causes a real exchange rate appreciation.
The productivity differential between the tradable and non-tradable sectors is the main
determinant of real exchange rates (MacDonald, 2000). Since improvements in the tradable
sector productivity are normally linked to economic growth, a correlation between relative
economic development and the real exchange rate is also postulated. Thus, it is expected that
countries growing faster will tend to experience real exchange rate appreciations with respect to
other, slowly growing economies (Solanes and Flores, 2009). This postulation challenges the
notion that fast growing countries run depreciated currencies.
The Balassa-Samuelson Hypothesis though applicable in some instances is not in others. It
usually suits fast growing countries than slow growing countries. According to Ito et al. (1999),
the Balassa-Samuelson Hypothesis may not be applicable when the economy has just come out
of being the primary goods exporter or planned economy even if the economy is growing fast.
2.2.4 Export-led Growth Hypothesis
The export-led growth hypothesis postulates that export expansion is a key factor in promoting
long-run economic growth. According to Medina-Smith (2001) the export-led growth hypothesis
(ELGH) postulates that export expansion is one of the main determinants of growth. This view
holds that countries do not only grow by increasing the amounts of labour and capital within the
economy, but also by expanding exports. Advocates of export led growth hypothesis argue that
exports can perform as an “engine of growth” (Schweicker et al., 2006). Araujo and Soares
(2011) contend that stronger exposure to international competition by higher exports is
considered to increase the pressure on the export industries to keep costs low and provide
incentives for the introduction of technological change. In this vein the growth of exports is seen
to have a stimulating influence on productivity of the economy as a whole via externalities of
exports on other sectors.
Many industrial economies according to Garnaut et al. (1995) have come to recognise that
reliance on world markets, a strategy known as outward looking, gives much greater scope for
17 | P a g e
economic growth than reliance on domestic markets. Outward looking countries, instead of only
relying on local markets, can enjoy extended markets abroad which help improve the balance of
trade and ultimately exports. James et al. (1989) reckons that the success of the East Asian
countries in the 1970s is a result of their adoption of export led growth. James, et al. (1989)
further argues that the remarkable growth and industrialisation first of Japan, and then of Hong
Kong, Taiwan and the Asian tigers (Singapore and Korea) challenged the negativity regarding
the applicability of export led growth strategy to other less developed countries.
Agosin (1999) is of the opinion that exports can be a catalyst for income growth, as a component
of aggregate demand. The rationale behind this reasoning is that in a small open economy,
demand for products is not sustainable in domestic markets to stimulate economic growth.
Export markets, in contrast, are almost limitless and hence do not involve growth restrictions on
the demand side. Kubo (2011) argued that exporting is beneficial as it enables related firms to
avail of certain benefits, such as the enhancement of efficient resource allocation, exploitation of
economies of scale, foreign technological knowledge through learning-by-doing and
technological innovation stimulated by exposing foreign-market competition.
Palley (2011) believes that export-led growth generates a win–win outcome for developing and
industrialized economies. Both the exporter and importer benefit from the global application of
the principle of comparative advantage. This is so because export oriented strategies discourage
the use of trade barriers and encourage free trade that will at the end benefit the countries
involved. Free trade, on the other hand, comes with its own evils, for instance, it can cause an
import boom and defeat the same purpose of export growth that countries need to achieve. If
unregulated, free trade can lead to dumping whereby countries export poor quality products at
cheap prices.
The export-led growth hypothesis is not without its own critiques. One such critique according to
Palley (2011) is that in a Keynesian world of demand shortage, trade can reduce domestic
demand, leading to reduced output, employment, and national welfare. In the Keynesian world,
export subsidies are not a gift but may instead steal demand and employment.
The other criticism leveled against the export-led growth is that it disturbs free trade. Usually
exporting countries devalue their currencies to make exports cheaper and reduce imports. This
18 | P a g e
strategy invokes retaliation by other countries which may lead to exchange rate wars, were
countries involved devalue their currencies. Also classical economists believe that any
employment induced by depreciation is at worst temporary because monetary effects are neutral.
Classical economists’ standpoint is that money is neutral hence inflation caused by depreciation
will not affect the real side of the economy (employment, economic growth and output).
2.3 EMPIRICAL LITERATURE
A lot of researchers (Schweicker et al., 2006; Acar, 2000; Chen, 2012) examined the impact of
real exchange rate on economic growth using different methods and countries. They came to
different conclusions depending on the country, method and time of study. This section presents
the various studies done, the methods used, the countries of research and the results obtained.
The section is divided according to literature from developed countries, literature from
developing countries and South Africa.
2.3.1 Empirical Literature from Developed Countries
Chen (2012) investigated the role of the real exchange rate in economic growth and in the
convergence of growth rates among provinces in China. The study employed the generalized
method of moments (GMM). Data from 28 Chinese provinces for the period 1992–2008 together
with dynamic panel data estimation was used. The outcome of the study was that there is a
positive effect of real exchange rate appreciation on economic growth in the provinces. The
GMM method used in this study demonstrates a superior performance on finite samples. Bond et
a. (2001) claims that the strength of this GMM lies in that it has the potential for obtaining
consistent parameter estimators even in the presence of measurement error and endogenous
variables. The results of the study showed that an appreciation of the real exchange rate boosts
economic growth and supports the Balassa-Samuelson Hypothesis that there is a positive
correlation between real exchange rate and economic growth. However, his results are not
consistent with Rodrick (2008). The latter found that depreciations are expansionary to growth
using data for developing countries that consists of a maximum of 184 countries and eleven 5-
year time periods from 1950-1954 through 2000-2004 for developing countries.
Jaussaud and Rey (2009) investigated the long-run determinants of Japanese Exports to China
and the United States during the period of 1971–2007. The study adopted the ARCH
19 | P a g e
(Autoregressive Conditional Heteroscedasticity) and the GARCH (Generalized ARCH) model.
The results indicated that Japanese sectoral exports to China and the United States have
depended on real exchange rate fluctuations and external demand (Gross Domestic Product of
the country of destination). Generally, the real exchange rate fluctuations and GDP have had the
expected negative effects. In particular, a real appreciation of the yen and a bigger uncertainty
has reduced Japanese exports. The results of this study are supported by the traditional approach
which holds that currency depreciation improves exports and hence growth and vice versa. Also
the results of the study are consistent with Edwards and Garlick’s (2007) findings that trade
volumes are sensitive to real exchange rate movements. Ito and Krueger (1999), however, tested
the Balassa-Samuelson hypothesis, by examining the relationship between the growth rate and
changes in the real exchange rate in Asia- Pacific Economic Cooperation Council (APEC)
countries and economies. The positive relationship between economic growth and real
appreciation that is a hallmark of the Balassa-Samuelson Hypothesis is found in Japan and, to a
much lesser extent, in Chile. Thailand and Malaysia experienced high growth with real
depreciation, although the magnitude of depreciation was small. The results of Jaussaud and Rey
(2009) and Ito and Krueger (1999) differ in Japan and Chile but are consistent for Thailand and
Malaysia.
Razin and Collins (1997) studied the real exchange rate misalignments and growth for a large
sample of developed and developing countries. The paper used regression analysis to explore
whether real exchange rate misalignments are related to country growth rates. The findings were
that over-valuations lower economic growth. Moderate to high (but not very high) under-
valuations are associated with more rapid economic growth. The findings of the study are
supported by the traditional theory of exchange rates, in that depreciations are associated with
rapid growth. To a greater extent, the results are not consistent with the results found for
developed countries, for instance, Kalyoncu et al. (2008) who assessed the long and short-run
effects of real exchange rate depreciation on output level in OECD countries. The results unlike
Razin and Collins (1997), were mixed as in the long run, a currency depreciation exerted a
negative impact on output expansion in Austria, Hungary, Poland, Portugal, Switzerland and
Turkey while in three countries, namely, Finland, Germany and Sweden, this relationship was
found to be positive. In the short run, a depreciation exerted a negative impact on output for
20 | P a g e
Finland, Germany and Turkey while depreciation exerted a positive impact on output growth for
Hungary and Switzerland.
Vieira and MacDonald (2010) empirically investigated the relationship between real exchange
rate misalignment and long-run economic growth in almost one hundred countries using panel
data techniques, including fixed and random effects, panel cointegration and system GMM
(generalized method of moments).The results for the two-step system GMM panel growth
models indicated that the coefficients for real exchange rate misalignment are positive for
different model specification and samples, which means that amore depreciated (appreciated)
real exchange rate helps (harms) long-run growth. The estimated coefficients are higher for
developing and emerging countries.
In conclusion, the reviewed studies in the developed countries are inconclusive about the impact
of exchange rate on economic growth. The empirical literature from developed nations show
mixed results about the impact of real exchange rates on economic growth. Chen (2012) for
example concurred with the Balassa-Samuelson hypothesis that there is a positive effect of
exchange rates on economic growth. Razin and Collins (1997) on the other hand found that
currency overvaluations are associated with lower growth which is consistent with the traditional
approach to exchange rates. Jaussaud and Rey (2009) also confirmed the traditional approach in
that Japanese exports to China and United States were found to decline if currency appreciated.
This in turn affected growth negatively. Vieira and MacDonald (2010) findings were that a more
depreciated (appreciated) real exchange rate helps (harms) long-run growth. It should be noted
however that though there is no consensus, many studies confirmed that currency overvaluations
are associated with lower growth.
2.3.2 Empirical Literature from Developing Countries
Tarawalie (2010) employed econometric techniques using quarterly data to find the relationship
between real effective exchange rates and economic growth in Sierra Leone. He also used a
bivariate Granger causality test as part of the methodology to examine the causal relationship
between the real exchange rate and economic growth. The empirical results suggest that the real
effective exchange rate correlates positively with economic growth, with a statistically
significant coefficient. These results which show a positive correlation between real effective
21 | P a g e
exchange rate and economic growth are supported by the Balassa-Samuelson hypothesis.
Empirically the study matches that of Ito and Krueger (1999) who found a positive correlation
between exchange rate and growth in Japan and Chile.
Ndlela (2011) investigated implications of Real Exchange Rate Misalignment in developing
countries with particular reference to growth performance in Zimbabwe. The study followed
ARDL (autoregressive distributed lag) approach to the cointegration method. The main
advantage of the ARDL method is that it can be applied irrespective of whether the variables are
I(0), I(1) or fractionally integrated. The main findings show that exchange rate misalignment
exerts a negative and highly statistically significant impact on growth. Also the findings support
the notion that real exchange rate overvaluation was a key fundamental in the post-2000
economic growth contraction in Zimbabwe. In turn, Masunda (2011) investigated the impact of
real exchange rate misalignment on sectoral output in Zimbabwe. To achieve this, the feasible
generalized least squares panel data techniques using data for the period between 1980 and 2003
from a sample of Zimbabwean sectors that include agriculture, manufacturing and mining sectors
was employed. The study indicated that real exchange rate misalignment is harmful to sectoral
output. The findings of the study were that undervaluation negatively affects the sectoral output
while exchange rate overvaluation negatively and significantly affects sectoral output.
McPherson et al. (2000) studied the direct and indirect relationship between the real and
nominal exchange rates and GDP growth in Kenya for the period 1970 to 1996. A number of
approaches that include the following: single equation regressions, a system of simultaneous
equations, and a VAR model test and cointegration techniques were used. The results of the
study showed no evidence of a statistically significant strong direct relationship between changes
in the exchange rates and GDP growth, rather growth responds to fiscal, monetary policy and
foreign aid. The results of this study are not supported by the theories reviewed in this chapter.
Acar (2000) studied the effects of depreciation on output growth in Less Developed Countries
(LDCs). Data from 18 sample countries in a fixed-effect procedure were employed. LDCs were
divided into two categories and two different regression analyses were conducted. Data from a
group of 10 countries for both manufacturing product exporters as well as agricultural and
primary product exporters to estimate a model of real output behaviour for a period of 25 years
were used. Data (for a 20 year period) from two different groups of countries (8 manufacturing
22 | P a g e
exporters, 8 agricultural and primary exporters) were analysed to investigate if there exists a
qualitative difference between different countries in terms of the effect of devaluation on
economic growth. The results indicate that depreciation creates a contractionary effect on output
in the first year, whereas it has an expansionary effect in the following year.
Aguirre and Calderon (2005) evaluated the growth effects of real exchange rates (RER)
misalignments and their volatility. Real exchange rates misalignments are calculated as
deviations of actual real exchange rates from their equilibrium. The study covered the period
1965-2003 for 60 countries using panel and time series cointegration methods. The study
employed the dynamic data techniques and the findings of the study were that real exchange
misalignments negatively affect growth though the effects are non-linear, for instance, growth
declines are larger in direct proportion with the size misalignments. The study’s other findings
were that although larger undervaluations hinder growth, small to moderate currency
undervaluation enhance growth. The other finding is that growth is hampered by highly volatile
real exchange rate misalignments.
Munthali et al. (2010) used a simple univariate model of GARCH to investigate the impact of the
real exchange rate on economic growth in Malawi. The results show that real effective exchange
rate (REER) volatility has adverse effects on economic performance. Also, an appreciated REER
is significantly and positively correlated with economic growth, reflecting Malawi’s net-importer
position. In contrast, REER volatility is significantly and negatively correlated with growth,
reflecting investors’ preference for a stable exchange rate. The positive relationship between real
exchange rate and economic growth was also found in Nigeria. Akpan (2008) used the ordinary
least squares (OLS) technique to investigate the link between the foreign exchange market and
economic growth in an emerging petroleum-based economy (Nigeria) for the period between
1970 and 2003. The paper maintains that there is a positive relationship between the exchange
rate, volatility and economic growth in Nigeria. Both these studies concurred with the Balassa-
Samuelson Hypothesis.
Domac and Shabsigh (1999) examined the effect of real exchange rate misalignment on the
collective economic growth of Egypt, Jordan, Morocco and Tunisia. The study constructed three
measures of exchange rate misalignment based on the purchasing power parity, a black market
23 | P a g e
exchange rate and a structural model. The results of the study were that real exchange rate
misalignments, especially overvaluation, have adverse effects on growth.
The review of literature from developing countries also presents mixed results, for instance,
Tarawalie (2010) and Akpan (2008) found a positive correlation between exchange rates and
economic growth. These results support the Balassa-Samuelson hypothesis. Acar (2000)
confirmed the Structuralist approach that devaluations may be contractionary to growth. Aguirre
and Calderon (2005), Domac and Shabsigh (1999) are aligned to the traditional approach that
devaluations enhance economic growth. McPherson, Rakovski and Kennedy (2000) however
found no direct relationship between exchange rates and economic growth. Exchange rate
volatility also has negative effects on growth as is the case in Munthali et al. (2010).
In conclusion, the reviewed studies from developing countries predicted mixed results
concerning the impact of exchange rate on economic growth. Results demonstrated that
economic growth reacts differently to an overvaluation or undervaluation of the currency in
different countries. In some instances real exchange rate devaluations had a negative impact on
economic growth, while in others a positive impact existed. Although a general conclusion on
the impact of real exchange rate on economic growth in developing countries cannot be easily
determined, the researcher based on cited researches (Domac and Shabsigh, 1999; Akpan, 2008;
Tarawali, 2010) is of the view that to a greater extent currency undervaluations are expansionary
to economic growth in developing countries. To a lesser extent however, economic growth
responds positively to currency overvaluations.
Empirical literature on the impact of real exchange rate on economic growth in South Africa
focused mainly on the impact of exchange rate volatility on exports, for instance, Todani and
Munyama, 2005; Raddatz, 2008; Edwards and Garlick, 2007. The findings of these researches
suggest that exchange rate volatility negatively affects exports in South Africa.
2.4 ASSESSMENT OF LITERATURE
The hypothesis of this study is drawn from the traditional approach to exchange rates. This
theory was chosen because most empirical evidence seems to support its postulations. The study
modifies the model by Acar (2000). Acar (2000) used economic growth (GDP) as a dependent
variable explained by movements in the explanatory variables namely real exchange rate, terms
24 | P a g e
of trade, government expenditure and money supply. In this study terms of trade are replaced by
trade openness, and government expenditure is replaced by fixed capital formation (Investment).
Real interest rates are added as one of the explanatory variables. Money supply and real
exchange rate remains in the model as explanatory variables.
2.5 SUMMARY
This chapter reviewed theoretical literature that shows the link between exchange rates and
economic growth. This helped to identify potential variables to be included in the model. The
first part of the chapter looked at the theoretical literature. Theoretical literature included the
traditional exchange rate approach which holds that devaluation of a country’s currency is
expansionary on the economy; the Structuralist approach on the other hand holds that currency
devaluation is contractionary. The Balassa-Samuelson Hypothesis holds that there is a positive
correlation between exchange rates and growth. Lastly, the export-led growth hypothesis was
reviewed which holds that exports plays a major role in the growth of a country.
The second part of this chapter explored empirical studies conducted by previous researchers on
impacts of real exchange rate on economic growth in developed and developing countries as well
as in South Africa. Studies reviewed employed several quantitative models to test the impact of
real exchange rate on economic growth. Most of the studies concluded that real exchange rate
devaluations/depreciations are expansionary to growth in developed and developing countries as
well as in South Africa. However, it is important to note that in South Africa, a large gap exists
in literature regarding the impact of real exchange rate on economic growth.
Having outlined the theoretical foundations of this study, the next chapter presents an overview
of the trends of exchange rates and economic growth in South Africa.
25 | P a g e
CHAPTER THREE
TRENDS IN MACROECONOMIC VARIABLES IN SOUTH AFRICA
3.1 INTRODUCTION
The aim of this chapter is to present an overview of macro-economic variables in South Africa
over the period 1994 to 2010. This chapter is divided into two sections. The first section deals
with exchange rates management. The second part of the chapter presents trends in economic
growth using real GDP in South Africa. This is followed by a discussion on the trends of the real
interest rates, trade openness, money supply and fixed capital formation investment. Concluding
remarks are provided towards the end of the chapter.
3.2 EXCHANGE RATE MANAGEMENT
Poorly managed exchange rates can be disastrous for economic growth (Rodrick, 2007). With
the increasing global integration of developing countries into the global trading system and
participation in international production networks, exchange rate management has taken on an
added importance. Exchange rate policy can affect macro-economic factors such as the GDP,
aggregate demand, inflation, economic growth, employment creation and income distribution.
Flassbeck (2004) points out that the fact that exchange rate movements directly influence the
overall competitiveness of a country, exchange rates have the potential to directly improve the
overall trade performance in a country. South Africa in particular is an open economy and is
involved in the exportation and importation of goods and services. This necessitates the need to
properly manage the exchange rate.
There has been a longstanding debate in many countries on the desirable degree of foreign
exchange rate flexibility. According to Engel (2009) one side of the debate holds that exchange
rates should be freely determined by market forces independently of any foreign exchange
intervention or targeting by central bank monetary policy. The other extreme holds that the
central bank should have control over the exchange rate market. The former view is based on the
notion that markets work better than the government to determine the appropriate level of the
exchange rate while the latter holds that the central bank can be handy in dealing with
undesirable aspects such as currency volatility and exchange rate misalignment.
26 | P a g e
3.2.1 Exchange Rate Policy in South Africa 1994-2010
South Africa’s management of the exchange rate is characterised by numerous regime changes,
that is, exchange rate regimes have evolved from being fixed, to managed floating and finally to
free floating in the year 2000. These regime changes are indicative of the importance attached to
the exchange rate, possibly as one of the stable instruments for the monetary authority in its
desire to achieve macro-economic stability. The following section is dedicated to the exchange
rate regimes that were adopted by South Africa from 1994 to 2010. According to Van der Merwe
(1996) “in order to cope with economic and political crises the country faced after 1994, South
Africa devoted significant attention to stabilisation measures in the domestic foreign exchange
market”. This was done through numerous changes to the exchange rate regime.
Since 1994 the country adopted three main regimes namely dual exchange rate regime under a
managed float commercial and free float financial rand which was in effect from September
1985 to February 1995, unitary exchange rate (managed float rand) from march 1995 to January
2000 and from February 2000 to present the country adopted a free floating rand, with inflation
targeting framework of monetary policy. The three are summarised in Table 3.1 below.
Table 3.1: South Africa: Exchange rate regime changes
Episode Date Exchange rate regime
I Sept 1985-Feb 1995 Dual exchange rate regime: managed float commercial
and free float financial rand
II Mar 1995 – Jan 2000 Unitary exchange rate : Managed float rand
III Feb 2000- present Unitary exchange rate: free floating rand, with
inflation targeting framework of monetary policy
Source: adapted from Mtonga (2011:3)
The choice of an exchange rate regime was mainly influenced by socio-political events that
hindered the development of the foreign exchange market in South Africa from the late 1984 to
1994 (Aron et al., 2000). These problems forced the authorities to opt for more direct control
measures to manage exchange rates. Van der Merwe (1996) explains that as a result of the
financial sanctions imposed on the country, the Reserve Bank was forced to re-enter the foreign
27 | P a g e
exchange market as an active participant under conditions of direct control measures to regulate
the influence of capital flows on monetary reserves. Van der Merwe (1996) further states that
during the first two years of the new Government of National Unity (1994-1995) South Africa’s
international financial relations normalised and steps were taken in the development of a forward
market with less Reserve Bank involvement and progressive relaxation of exchange control.
The period from March 1995 to January 2000 saw the country adopting a unitary exchange rate
under the managed float rand. Under this regime the spot exchange rate was determined by
market forces under conditions where exchange rate control is exercised only over residents in
respect of capital movements. According to Aron et al. (2000) changing to this regime was a
great step towards the gradual liberalisation of the financial markets and repositioning South
Africa into the global economy. Mtonga (2011) explains that financial liberalisation resulted in
the gradual removal of exchange control regulations. On eliminating the financial rand the
exchange control was abolished on transactions of non-residents. Van der Merwe (1996) asserts
that under this regime, the Reserve Bank did not prescribe fixed buying and selling rates for
dollars to be quoted by the banks in their transactions with the public, neither did it quotes its
own predetermined buying or selling rates for spot dollars. The managed float allows the
currency to fluctuate under market conditions but allows the reserve bank to intervene in the
market to minimise short run variability by adjusting the stock level of gold and foreign
exchange reserves (Nattrass et al., 2000).
The year 2000 witnessed another shift in South Africa’s monetary policy framework. The
country adopted inflation targeting as a framework for monetary policy. This change was
followed by the adoption of the free floating exchange rate which complements the fundamentals
of inflation targeting regime, for instance, Masson et al. (1998) believes that for inflation
targeting to be effective there has to be no pre-commitment to an exchange rate target. This
implies that the rand exchange rate is basically determined by the forces of demand and supply in
the foreign exchange market. Mtonga (2011) argues that “the year 2000 demarcates the previous
years of controls from the present regime in which market conditions are allowed to influence
the domestic foreign exchange market”. The move to a free floating exchange rate regime was
mainly due to the fact that for inflation targeting to work well, there was need for an independent
28 | P a g e
monetary policy. The independence of monetary policy is limited if the exchange rate is targeted
because the primary goal of the monetary policy will be that of defending the exchange rate.
The current policy of the central bank is generally to stay out of the market and to allow market
forces to determine the exchange rate. In recent years, however, the Bank has been building up
foreign exchange reserves and this involves the purchase foreign exchange from the market
(SARB, 2012). Thus the central bank influences the equilibrium exchange rate since it interferes
with the demand for foreign exchange. The bank though it ceased the direct control on the
foreign exchange still influences the exchange rate by participating in the market by buying or
selling other currencies. SARB (2012) also contends that the exchange rate, however, is not the
objective or the target of the Bank. The decisions by the Bank regarding reserve accumulation
should rather be seen as management of international liquidity, not exchange rate policy.
3.2.2 Trends in Real Exchange Rates in South Africa from 1994-2010
Figure 3.1: Trends in real effective exchange rates: Average for the period - 15 trading
partners6 1994-2009
Source: Data compiled from SARB (2012)
6Annual average figures were used to analyse the trends in the real effective exchange rates. Real effective
exchange rate of 15 major trade partners instead of frequently used bilateral (usually against USD) real exchange rate was chosen because it is richer measure of competitiveness. Also South African trade is not only against the United States of America but against more countries hence an average for these trading partners is a more realistic measure
0 20 40 60 80
100 120 140
RE
R
YEAR
Real effective exchange rates 1994-2010
29 | P a g e
The rand exchange as is shown in Figure 3.1 was fairly stronger during the period 1994-1995.
The inception of a democratic government in 1994 accompanied by the removal of sanctions
attracted a lot of foreign direct investment. These developments resulted in increased capital
inflows and hence demand for the local currency. Jordaan and Harmse, (2001) explain that
increased capital inflows result in appreciation of the rand. After the elections in April 1994, the
economy experienced a relative large inflow of investments into the economy leading to a stable
and relatively strong rand between 1994-1997 (Figure 3.1). Takaendesa (2006) argues that the
period (1995-1997) resembled a return to a unified exchange rate system and the implementation
of gradual relaxation of exchange controls. This meant that the rand was now determined in a
competitive foreign exchange market and all exchange controls were abolished for the non-
residents.
Figure 3.1 above also shows that the real effective exchange rate index (on a 2000 base year)
depreciated significantly from 119.22 in 1997, to 108.05 in 1998, where after it depreciated to
100 by the end of 2000 (SARB, 2012).The depreciation was driven mainly by a significant
relaxation of exchange controls over residents, the effects of the Asian financial crisis and a
decline in the price of precious metals (Takaendesa, 2006). In 2001 the real effective exchange
rate of the rand depreciated to 91.39 from 100 in 2000 and further depreciated to 82.55 in 2002
(Figure 3.1). According to Mnyande (2010) this depreciation is mainly attributed to the attacks
on America on 11 September 2001 and a volatile political situation in Zimbabwe which led to
substantial capital withdrawals from South Africa.
Todani and Munyama (2005) state that positive investor sentiment, a rise in international
commodity prices and South Africa’s relatively healthy balance of payments contributed to the
recovery of the rand during the period 2002-2005. The real effective exchange rate of the rand
appreciated from 82.55 in 2002 to 103.23 in 2003. The recovery continued with the real effective
exchange rate of the rand being 110.13 and 112.5 in 2004 and 2005 respectively. The rand
depreciated between 2006 and 2008 from 112.5 in 2004 to 94.09 in 2008. Mnyande (2010)
explains that this period was largely comprised of oil shocks, world price crisis and the further
deterioration of the Zimbabwean economy which led many Zimbabweans resorting to the
informal use of the Rand. The rand appreciated in the year 2009 to 101.41 from 94.09 in 2008
30 | P a g e
and this appreciation persisted into the year 2010 were the real effective exchange of the rand
was 113.85.
3.3 ECONOMIC GROWTH IN SOUTH AFRICA
The performance of the South African economy fluctuated considerably in the post-War era, but
particularly since the early 1980s. Two trends stand out: firstly, the decline in real per capita
GDP since 1981 and, secondly, the growth revival since 1995 (Du Plessis and Smit, 2007). South
Africa experienced sluggish economic growth between the years of 1984 and 1993 owing to
internal political instability, trade and financial isolation. South Africa however experienced
considerable improvement in economic growth since the advent of democracy in 1994; this
increase in growth can be attributed to improved macro-economic performance as a result of
improved policies and removal of financial sanctions in 1994. A calm political environment
coupled with increased foreign direct investment ensured improvement in the performance of the
economy.
The inception of democracy in 1994 created the possibility of a peaceful and a more stable future
which led to the restoration of investor confidence. According to Faulkner and Loewald (2008)
the decade prior to the year 1994, South Africa was under economic sanctions and investor
confidence was low which made it hard for the economy to attract investment. Faulkner and
Loewald (2008) further argue that another political economy factor in South Africa’s improved
growth performance since 1994 has been the policy response to reintegration with the world
economy and globalisation. Figure 3.2 below shows an overview of the trends in economic
growth for the period between 1994 and 2010.
31 | P a g e
Figure 3.2: Economic growth trends in SA: 1994 2010
Source: Data compiled from DTI (2012)
The economic performance of the post-apartheid economy has been quite strong, averaging
growth in real gross domestic product (GDP) of 3.3 percent and 1.35 percent in per capita terms
for the period 1995 to 2005. This growth trend was an improvement, if compared with the rates
of the 1985 to 1994 period, where the respective average rates were 0.8 and –1.3 percent. Du
Plessis and Smit (2007) argue that the improved growth performance is largely attributable to
strong domestic demand and a large foreign capital inflow in the face of low inflation and
interest rates. The recovery continued between 1995 and 1997 where growth rates of 3.1 per
cent, 4.3 per cent and 2.6 per cent were recorded, respectively.
In 1997, growth in GDP dropped sharply from 2.6 per cent from 4.3 per cent in 1996. Du Plessis
and Smit (2007) attribute this to the effects of the global financial crisis. Also the Asian financial
crisis and the world recession contributed to this dismal performance. In 1999 there was
considerable improvement to 2.4 per cent growth rate from 0.5 per cent recorded in 1998. The
strides towards high growth rate continued in 2000 with South Africa registering a 4.2 per cent
growth rate. The progress slowed down with a slight decrease in growth rate to 2.7 per cent in
2001. Since 2002 when the inflation targeting framework came into effect, South Africa enjoyed
positive and sustained growth rates with 2006 recording a growth rate of 5.6 per cent, the highest
since 1981. In 2007 the rate dropped to 3.6 per cent. GDP growth declined in 2008 and even
-2
0
2
4
6
% c
han
ge
GD
P
YEAR
Percentage change in GDP 1994-2010
GDP
32 | P a g e
turned negative in 2009, at the height of the global financial crisis. According to Mnyande
(2010), the year 2009 was disappointing for most parts of the world especially advanced
economies. South Africa for the first time since independence recorded a negative growth rate of
negative 1.5 per cent for the year as a whole. 2010 was a relief as seen by an improved growth
rate to 2.9 per cent.
3.3.1 Exchange Rates and Economic Growth in South Africa
The level of real exchange rate is important on economic growth as it determines the value of
imports and exports of a country. Walters and De Beer (1999) explain that a country’s exchange
rate is an important determinant of the growth of its cross-border trading and it serves as a
measure of its international competitiveness. Figure 3.3 below shows the trends of real effective
exchange rate and economic growth over the period 1994-2010.
Figure 3.3: Trends in real effective exchange rates and Real GDP in percentage changes
1994-2009
Source: - Data compiled from SARB and DTI (2012)
The movements in both economic growth as represented by gross domestic product and real
effective exchange rate have been used to analyse the relationship between the two. From Figure
3.3 it can be seen that over the period 1994-2010, economic growth trends were fairly stable as
compared to volatile real exchange rates. Between 1994 and 1995 the rand exchange was fairly
0 20 40 60 80 100 120 140
0 5000
10000 15000 20000 25000 30000 35000 40000
19
94
19
95
19
96
19
97
19
98
19
99
20
00
20
01
20
02
20
03
20
04
20
05
20
06
20
07
20
08
20
09
20
10
RE
ER
GD
P
YEARS
GDP and real effective exchange rates in 1994-2012
GDP REER
33 | P a g e
stronger followed by a sharp depreciation notably in 1996 where the real effective exchange rate
of the rand was at a low of 111.86. During this period, economic growth in million rands was
stable but increasing from 28536 in 1994 to 29431 in 1996. In the year 1997 there was a sharp
appreciation as shown by a change in the REER from 111.86 in 1996 to 119.22 in 1997. The
resulting change was an increase in growth to R29582 million but there was a drop to 2.6 per
cent in the growth rate. There was a sharp depreciation in 1998 as shown by the change in the
REER to 108.05 and economic growth responded by dropping to R29116 million. A weaker rand
was persistent from 1998 to 2002, and during this time economic growth was fairly stable and
rising save for the year 2001.
A massive appreciation in 2003 was accompanied by a decrease in economic growth, that is, in
2002 the REER was at 82.55 and 103.23 in 2003 and economic growth was R30581 and R30992
respectively. From 2004 to 2008 the rand exchange took a downward trend whilst economic
growth increased slightly year on year save for 2008. In 2009 and 2010 the rand was stronger
and economic growth was low especially in 2009 where GDP growth took a negative swing from
R36942 billion in 2008 to R35936 billion in year 2009. Year 2010 experienced a further
appreciation of REER to 113.89 from 101.41 in 2009. This was accompanied by a slight increase
in growth to R36594.
3.4 TRADE OPENNESS IN SOUTH AFRICA
When South Africa was reintegrated into the global economy in 1994, the contribution of
imports and exports rose strongly as a percentage of GDP. The economy became more open,
more productive and more outward orientated (Flatters and Stern, 2007). The entire economy
became more outward-oriented, with export orientation and import penetration increasing across
both primary sectors and manufacturing. Trade openness is popularly measured as exports plus
imports divided by gross domestic product (X+M)/GDP and can be seen as the extent to which a
country is engaged in international trade. It then forms part of the current account section of the
balance of payments. According to Hausmann (2008), the results of a study conducted by the
Harvard University Center for International Development showed that for South Africa to
achieve the desired annual growth target of 6 percent, exports growth needs to rise significantly.
This then suggest that to improve economic growth prospects of the country policies to stimulate
export growth and diversify the composition of exports should be high on the government’s
34 | P a g e
agenda. Apart from improving economic growth and current account balance boosting, exports
can be handy in addressing the unemployment problems which are high in South Africa.
Trade in South Africa endured a lot of restrictions during the period prior to 1994 and some of
these constraints persisted after 1994. As a result, South Africa experienced mediocre trade
volumes in the early 1990s compared to other middle-income economies (Hausmann, 2008;
Edwards et al. 2000). Before 1994 trade was mainly affected by the introduction of sanctions
during the 1980s which crippled many sectors of the economy including the manufacturing
sector. After 1994, conditions did not quickly change with regard to international trade.
According to Edwards and Lawrence (2006), in 1994 when the democratically elected
government took office, protection remained high in many sectors and the tariff structure was
complex compared to a range of other developing economies. This continued protection resulted
in stagnant trade growth during the early 1990s.
After 1994, trade liberalisation became a central part of South Africa’s post-Apartheid
development strategy. Since 1994 trade policy has shifted towards achieving greater openness
through removal of restrictive trade policies and promotion of some previously underperforming
export sectors. Trade liberalisation arose in response to the decline in the contribution of import
substitution policies (adopted prior 1970s) towards growth, a continued dependence on gold as a
source of foreign exchange and increased trust on exports brought about by rapid export-led
growth in some of the newly industrialised countries of South East Asia (Jenkins et al.,
1997).Since 1994 the government has done a lot with regards to trade liberalisation reflecting the
government’s strong commitment to outward-oriented industrialisation.
Trade liberalisation began in earnest when the new government came into power. Some of the
reforms introduced included the removal of import surcharges on capital goods in 1994 and
consumer goods in 1995.In 1995 an offer was made to the World Trade Organization consisting
of a five-year tariff reduction and rationalisation program (Edwards et al.,2000). The export
incentive scheme was abolished by 1997 and the number of tariff lines declined by 40 percent by
1999. Average tariff rates were halved and the country moved towards its proposed
rationalization targets (Jenkins et al., 1997). The reforms saw South Africa participating heavily
in international trade and has since joined international trade institutions formally by means of
treaties and trade agreements. South Africa in 1996 signed to be a member SADC and a trade
35 | P a g e
protocol enabled the creation of a free-trade zone over eight years. South Africa also has good
trade relations with other major trading blocs such as SACU and EU. Trends in trade openness
are presented in Table 3.4 below.
Figure 3.4 Trends in trade openness in South Africa (1994 to 2010)
Source: - Data compiled from SARB website (2012)
In general the level of trade openness was on average rising since 1994, though it slightly fell in
1999 and rose again in 2000. Another major decrease in trade openness was in 2003 but the trade
activities of the economy rapidly recovered and increased again from 2004 up to 2008 when it
significantly fell again in 2009 and 2010. From the figure 3.4 trade openness can be divided into
three categories namely the period 1994-1999 characterised by slow and steady increase in
openness, followed by the period 2000 to 2008 characterised by rapid increase in openness and
finally 2009 and 2010 characterised by a sharp decrease in openness. The year 2010, however,
showed some signs of improvement.
Trade openness reduced South Africa’s overdependence on primary products to more
sophisticated manufactured goods and service industry. South Africa is naturally endowed with
lot of rich mineral resources such as gold, platinum and diamonds. Edwards and Lawrence
(2006) argues that this factor endowment was the major reason that led to South Africa’s
reluctance to develop an internationally competitive manufacturing industry. Exports have been
dominated by resources-based and relatively low value-added commodities while imports are
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Tra
de
oppen
nes
s
Year
trade oppenness
Trade openness
36 | P a g e
primarily dominated by higher value-added goods. Mining sector contributed more than 50% of
total exports before independence. A lot has changed because of concerted efforts by the
government to diversify trade which later opened many markets for South African products
regionally and overseas. Below is Table 3.2 which shows export share of products in the South
African basket export basket.
37 | P a g e
Table 3.2 Export share of products in the South African export basket to the rest of the
world (2000-2010)
Product 2000 2002 2004 2006 2008 2010
Agricultural products 22678
35009
30062
33531
58067
57596
Fuels & mining products 37894
56537
80721
135155
236007
223415
Manufactures 97073
148609
146439
185036
314373
240453
Iron and steel 19128
25355
36435
38221
73110
56637
Chemicals 14252
22653
20362
26460
47233
37175
Pharmaceuticals 749
1009
780
914
1617
1230
Machinery& transport Equip 31694
55285
51193
76422
133917
98491
Office & telecom equip 2837
4186
3857
5102
7352
5404
Electronic data processing & office
equip
978
1294
813
1807
2228
1647
Telecommunications equip 1685
2240
2354
2402
3482
2387
Automotive products 11845
25261
23878
33180
63843
49073
Textiles 1644
2587
1941
2044
2484
1691
Total exports R (millions)
244142
383380
401189
542128
946337
777308
Source: Data compiled from WTO (2012)
38 | P a g e
Table 3.2 shows that the South African economy is now hugely diversified when it comes to
exports. This has improved trade with the rest of the world. From the year 2000 the exports of
South Africa included the previously not so important products like Electronic data processing &
office equipment, Telecommunications equipment and Pharmaceuticals. By 2000 the
manufacturing sector was the most dominant contributor to the export basket with exports worth
R97073 million compared to mining and agriculture which contributed R37894 and R22678
million respectively. In order to show the progress of South African Exports to the rest of the
world, Table 3.3 below shows the share of South African export basket expressed in percentages.
From the Table 3.3 below it can be seen that South Africa has taken too many strides towards
opening up its borders to international trade. A quick look at the export structure of South Africa
shows that improvements were made in this sector. Fuels and mining products improved
significantly from contributing only 15.5 % in the year 2000 to contributing 28.8 % of exports by
the year 2010. The manufacturing sector contributed 39.8% of South African exports in 2000
which was the greatest, but in 2010 it contributed only 31%. Agriculture contributed 7.4% of
exports in 2010. South Africa also made important inroads in the previously not so dominant
sectors like Automobile products, electronic data equipment and chemicals which contributed
6.3%, 0.2% and 4.8% to the South African exports respectively.
39 | P a g e
Table 3.3 Export share of products in the South African export basket expressed in
percentages (2000-2010).
Product 2000 01 02 03 04 05 06 07 08 09 2010
Agricultural products 9.3 8.2 9.1 8.7 7.5 7.3 6.2 5.8 6.1 8.3 7.4
Fuels & mining
products
15.5 23.1 14.7 18.6 20.1 21.2 25.0 26.5 25.0 27.0 28.8
Manufactures 39.8 35.3 39.0 36.9 36.5 36.1 34.1 33.6 33.2 31.2 31.0
Iron and steel 7.8 5.6 6.6 8.0 9.1 8.1 7.1 7.7 7.6 6.3 7.3
Chemicals 5.8 5.2 6.0 4.9 5.1 5.5 5.0 4.5 5.0 5.1 4.8
Pharmaceuticals 0.3 0.2 0.3 0.2 0.2 0.2 0.2 0.2 0.2 0.2 0.2
Machinery&
transport Equip
13.0 13.4 14.4 13.4 12.8 13.3 14.1 13.9 14.2 13.4 12.7
Office & telecom
equip
1.2 1.7 1.1 1.0 1.0 0.8 0.9 0.9 0.8 0.9 0.7
Electronic data
processing & office
equip
0.4 0.4 0.3 0.3 0.2 0.2 0.3 0.3 0.2 0.3 0.2
Telecommunications
equip
0.7 0.7 1.6 0.5 0.4 0.4 0.4 0.5 0.4 0.4 0.2
Circuits & electric
components
0.1 0.1 0.2 0.2 0.2 0.2 0.2 0.1 0.2 0.2 0.3
Automotive products 4.9 5.4 6.6 6.3 6.0 6.0 6.1 6.1 6.7 6.3 6.3
Textiles 0.7 1.0 0.7 0.6 0.5 0.4 0.4 0.3 0.3 0.3 0.2
Clothing 0.6 1.6 0.7 0.6 0.4 0.2 0.2 0.1 0.1 0.1 0.1
Source: - Data compiled from WTO (2012)
South Africa did not only open up to exports, as is shown by Table 3.4 the country also imports
quite significantly. South Africa mainly imports manufactured goods, chemicals, fuel, mining
and agricultural products.
40 | P a g e
Table 3.4 Import share of selected products in the South African import basket 2006-2010
Product 2006 2007 2008 2009 2010
Agricultural products 24950.67
34086.86
44262.12
39907.96
39708.29
Fuels & mining products 100743.3
123303.9
185605.5
125300.9
127948.1
Manufactures 331802.6
397042.8
487139.1
367651
412108.1
Iron and steel 7734.91
11096.57
12781.88
8690.316
9797.104
Chemicals 41746.86
50946.88
70552.04
56309.87
63432.22
Machinery& transport
Equip
202522
241973
295509.9
211815.9
239355.4
Automotive products 71840.6
86381.13
93079.18
61119.08
81510.73
Textiles 6598.02
7160.216
8408.482
7644.103
8383.919
Clothing 7599.566
7012.074
8193.938
8892.809
9980.159
Food 20355.74
28725.52
37867.05
35124.06
34326.47
Office & telecom Equip
48128.33
51842.79
60823.28
50378.52
59141.41
Total Imports R
(millions)
999907.9
1200102
1531953
1143823
1265811
Source: - Data compiled from WTO (2012)
41 | P a g e
3.5 GROSS CAPITAL FORMATION AND GROSS DOMESTIC PRODUCT (1994-2010)
The link between the growth rate of real output and the share of gross fixed capital formation
(investment) in GDP has been an important subject of analysis and debate in both developed and
developing nations (Ghali and Al-Mutawa, 1999). The analysis addresses the question of the role
of fixed investment in promoting economic growth. Investment in physical capital is an
important factor in explaining economic growth since changes in the share of fixed investment in
GDP affects the output growth rate. Levine and Renelt (1992) explain that the gross capital
formation contributes to sustainable economic growth not only on the demand-side, but also on
the supply-side, because an important part of these expenditures are dedicated to the renewal of
the firms’ fixed capital. Fixed capital is one of the main production factors and hence changes in
this variable have a huge impact on growth. Figure 3.5 shows trends in Gross fixed capital
formation in South Africa from 1994 to 2010.
Figure 3.5. Investment and Gross domestic product 1994-2010
Source: - Data compiled from SARB and DTI website (2012).
Figure 3.5 gives an overview of fixed capital formation (investment) and economic growth in
South Africa. Prior to 1994 Gross domestic fixed investment in South Africa took significant
slump due to economic sanctions, resulting in huge disinvestment and outflow of foreign capital.
-10
-5
0
5
10
15
20
-2
-1
0
1
2
3
4
5
6
FC
F
GD
P
Year
Fixed capital formation (investment) and GDP
GDP FCF
42 | P a g e
However, the situation improved with the abolition of sanctions and disinvestment ended in 1994
(Du Toit and Moolman, 2007). Gross fixed capital formation followed an upward trend from
1994 to 1996 as depicted by Figure 3.5. The increase in investment in this period was mainly due
to the end of apartheid and abolishing of sanctions imposed on the country. Mnyande (2010)
argues that this saw South Africa being economically independent and able to open up its
borders to the international community.
Investment dropped considerably in 1999. Du Toit and Moolman (2007) stress that high interest
rates in 1999 of 18.69% led to a decrease in investment by 7.60%. Other factors such as the
Southeast Asian Crisis might have contributed to the decrease in investment, but interest rate
increases and volatility were the main reasons for decreases in investment. The economy
managed to recover in 2000 and investment increased by 1.59%, as investors built confidence
about the country due to the adoption of the formal inflation targeting policy, which implied a
more stable economy (Merwe, 2004). The upward trend was however short-lived, in 2001 due to
emerging market concerns and uncertainties about recession in the US, Europe and Japan which
led to sharp reversals in capital flows and considerable currency volatility in many emerging
markets including South Africa (Du Toit and Moolman, 2007). This led to failure to increase
investment in South Africa significantly, leading to investment only increasing with smaller
margin of 2.07% in 2001.
South Africa was on an upward swing investment wise from 2001 to 2008. The increase in
investment was attributed to the stableness of the economy, evidenced in the low inflation rate
which was below 10% for many years, besides 2008 were inflation was 11.5% (Mnyande, 2010).
Year 2009 was characterized by a decline in investment, the fact was attributed to the global
financial crises that affected many economies and the effects spilled over to 2010. Generally
Figure 3.5 portrays a positive correlation between investment and economic growth. Since the
advent of democracy in South Africa in 1994, gross fixed capital formation took an upward
swing. From Figure 3.5 it can be seen that between 1994 and 1997 gross fixed capital formations
was high and at the same time GDP was following the trends in investment.
43 | P a g e
3.6 REAL INTEREST RATES AND GROWTH (1994-2010)
There is a growing debate in the emerging market on the choice of an appropriate monetary
policy that could lead to sustainable economic growth (Davidson, 2007). Inflation targeting has
become one of the policy alternatives and has since year 2000 been implemented in South Africa
and some of the emerging markets in Asia and Latin America. As suggested by economic theory,
lower and positive levels of inflation lead to positive economic growth. The South African
Reserve Bank uses interest rates in its inflation targeting framework. The interaction therefore
between real interest rates and economic growth became very important in this study.
The real interest rate is the nominal interest rate adjusted for expected inflation and is usually
measured as the difference between the nominal interest rate and the expected or actual inflation.
Davidson (2007) argues that real interest rates lie at the heart of the transmission mechanism of
monetary policy. Investment spending is affected by the cost of capital, and this provides a link
between the financial sector and the macro-economy. In agreement Kahn and Farrell (2002)
contend that a traditional indicator of monetary policy stance is the interest rates. Interest rates
are an important link by which changes in the money supply are transmitted to the real economy.
This provides a logical reasoning for focusing on the interest rates and economic growth link.
Figure 3.6 below gives an overview of real interest rates and economic growth in South Africa
from 1994 to 2010.
Figure 3.6 Real interest rates and Gross domestic product 1994-2010
Source: - Data compiled from SARB and DTI website (2012).
0
5
10
15
20
-2
0
2
4
6
RIR
GD
P
YEAR
Real interest rates and GDP 1994-2010
GDP RIR
44 | P a g e
As depicted in Figure 3.6, real interest rates were on a clear upward trend after 1994 until 1998
and mostly above 7%. Shelile (2006) explains that the increase was a result of the monetary
policy instrument adopted by the SARB during this period. During this period, in order to
achieve its objective of protecting the rand through low inflation, the bank used money supply
rather than interest rates as a way of fighting inflation. Frederick and Fouri (2009) contend that
low growth rate in money supply led to increases in interest rates for the same period. The
interest rates started to trend downwards from 1999 to 2000 and continued to decrease in 2001.
The reduction in interest rates was related to the inflation targeting policy adopted by the SARB
in 2000. According Shelile (2006), the temporary increase in 2002 was due to supply shocks in
the country resulting from the global financial crises pressure.
In late 2006 interest rates started to rise again. The increase was a result of financial turmoil due
to the crisis that originated from the United States of America in 2007 and 2008 (Havrylchyk,
2010). The SARB managed to reduce interest rates in 2009 to achieve inflation targeting. In the
year 2010 the interest rates started to pick up again. From Figure 3.6 it is evident that changes in
real interest rates triggered some changes in economic growth. From 1994 to 1998 interest rates
were increasing and this was accompanied by positive changes in economic growth. High
interest rates in 1999 led to a decrease in economic growth. Though there were many factors that
contributed to a decline in economic growth such as the South East Asian Crises, interest rates
and its volatility were the main causes for the decreases in investment and economic growth (Du
Toit and Moolman, 2007). From 2003 to 2007 South Africa faced low interest rates and this was
coupled by high and positive growth rates. From 2008 to 2010 there was a steady increase in real
interest rates and growth also improved from the sharp decrease of 2009.
3.7 MONEY SUPPLY AND GROWTH (1994-2010)
Money supply is the amount of money within a specific economy available for purchasing goods
or services. For the purposes of this study, the broad definition money supply (M3) is adopted
which includes all bank notes and coins in circulation plus all deposits of the domestic private
sector with banking institutions. The rationale behind choosing M3 as proxy for money supply is
that it has a stable relationship with the domestic demand compared to M1 and M2.
45 | P a g e
M3 which is the broad money demand in South Africa directly depends on income (GDP) in the
sense that when income increases, the transactions in the economy increase. The increase in the
economy’s transactions triggers an increase in demand for money to facilitate increased
transactions. Money supply affects economic growth in many ways and this is the reason why it
has been included as one of the variables in the growth model.
There is controversy amongst economists about the effect of money supply on economic growth,
and Rad (2012) explains that economists differ on the effect of money supply on economic
growth. Ogunmuyiwa and Ekone (2010) add that there has always been a persistent concern
among monetary economists about the relationship between money supply and output. While
some agree that variation in the quantity of money is the most important determinant of
economic growth, others are skeptical about the role of money on gross national income.
According to Rad (2012) increases in money stock increases investment in an economy and the
capacity of internal production also increases. In addition, the extra liquidity from the demand
side can invoke economic growth. Ogunmuyiwa and Ekone (2010) on the other side argue that if
the increase of the money stock leads to the inflation and rent seeking activities, most probably it
will have a negative impact on the economic growth. Figure 3.7 below presents the movements
in both money supply and economic growth in South Africa from 1994 to 2010.
Figure 3.7 Money supply (M3) and Gross domestic product 1994-2010
Source: - Data compiled from SARB and DTI website (2012).
0
5
10
15
20
25
-2
0
2
4
6
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
MS
GD
P
YEAR
Money supply and GDP 1994-2010
GDP MS
46 | P a g e
As already explained money supply exerts considerable influence on economic activity in both
developed and developing economies, South Africa included. This section attempts to analyse
the relationship between output and money supply growth in South Africa using the graphical
method. Figure 3.7 above shows the trends in both economic growth and money supply from 994
to 2010.
Figure 3.7 indicates that the correlation between money supply and output is generally
considerable. A positive but moderate correlation is observed between the two variables between
1994 and 1996. Money supply grew from 15.71 % in 1994 to 17.53 % in 1996 and during the
same period GDP rose from 3.2 % to 4.3 % respectively. As depicted in Figure 3.7 there was
negative link between money supply and growth between 1997 and 2004. The degree of negative
correlation is strong between these years and it is most notable between 1999 and 2000 were
money supply dropped from 10.45% to 7.26 % and GDP rose from 2.4% to 4.2%.However,
positive correlations are depicted in the years 2005 and 2010 were the trends between the two
variables were following each other. The year 2009 was generally poor by many standards, GDP
growth rate in this year was 1.5%. Generally from 1994 to 2010 it can be deduced that there is a
positive correlation between money supply and growth though there are some instances showing
a negative trend. This relationship provides an indication that to a greater extent an increase in
money supply is likely to lead to an increase in real GDP.
3.8 SUMMARY
This chapter gave an overview of real exchange rates and economic growth trends in South
Africa over the period of 1994 to 2010. Firstly, the chapter discussed the exchange rate system
and exchange rate regimes in South Africa. The exchange rate system evolved over time in South
Africa and there were important changes which led to the adoption of the free float exchange rate
system. The second part discussed trends in the real exchange rates, and exchange rates were
found to be unstable with many factors affecting them which include government policies,
economic and political developments inside and outside South Africa. Thirdly, trends in
economic growth were also reviewed followed by the concurrent discussion of exchange rates
and economic growth. Lastly, the chapter looked at other variables that affect economic growth
in South Africa.
47 | P a g e
From the discussion in this chapter, the main highlights are that the exchange rate regimes
changed over time in South Africa from periods of more exchange controls to a free float rand.
Real exchange rates from 1994 were generally stronger except for the years 2002 to 2003 were
there was a sharp decrease. Economic growth fairly improved averaging around 3.3 percent with
the exception of the year 2009, where the growth rate was negative owing to the global financial
crisis. Investment was generally improving from 1994 to 2010 except notably the year 1999
where investment was at the lowest owing to the high interest rates experienced that year. The
country opened up to trade as seen by the removal of trade restrictions and trade diversification.
Money supply and economic growth followed each other, suggesting a positive relationship and
interest rates were negatively related to growth.
Having outlined the trends in various macroeconomic variables the next chapter presents the
methodology used in the study.
48 | P a g e
CHAPTER FOUR
RESEARCH METHODOLOGY
4.1 INTRODUCTION
This chapter sets the analytical framework used in this study by providing the model used to
examine the impact of exchange rates on economic growth in South Africa from 1994-2010.
This chapter also includes confirmation of data sources, research techniques and diagnostic tests
employed in this study.
4.2 MODEL SPECIFICATION
In examining the impact of exchange rates on economic growth in South Africa the explanatory
variables in this study are fixed capital formation (FCF), real interest rates (RIR) trade
openness(OPEN), broad money supply ( ) and real exchange rates (RER).
This study modifies (Acar, 2000) model. The reduced form equation for output in this model is
formally specified below.
……………………………………..4.1
Where
: Real output
: Constant term
: Parameter that captures the trend rate of growth
: Time period
: Relative size of government (the ratio of government expenditures to nominal output)
: Money supply term (the difference between actual and expected rate of growth of nominal
money supply)
: Terms of trade
: Real exchange rate
: Error term with mean zero and constant variance.
49 | P a g e
The main objective of this study is to investigate the impact of the real exchange rate on
economic growth (GDP). An output growth model is thus specified by adding the real exchange
rates (RER) to the set of explanatory variables. In this study, the dependent variable is economic
growth(y) as explained by the movements in other variables which are real interest rate (RIR),
fixed capital formation (FCF), broad money supply ( ), trade openness (OP) and real effective
exchange rates (REER) which is the main explanatory variable. The model is specified as
follows:
………………………………….4.2
Where:
: The intercept
: Coefficients of the explanatory variables
: Error term which represents omitted variables in the specification of the model
: Real output
: The real effective exchange rates
: is the real interest rates
: represents trade openness
: is broad money supply
: represents fixed capital formation investment.
To obtain elasticity coefficients and remove the effect of outliers, the variables have to be
transformed to logarithms. In log linear form the function becomes:
Log
…………4.3.
4.3 DEFINITION OF VARIABLES
Gross Domestic Product
This shows the Gross Domestic Product in South Africa measured in billion Rands. The growth
rate is the percentage increase or decrease of GDP from the previous measurement cycle.
50 | P a g e
Real exchange rates
The real exchange rate is the nominal exchange rate that takes the inflation differentials among
the countries into account. It is used as an indicator of competitiveness in the foreign trade of a
country.
Trade openness
This measures a country’s willingness to trade with other countries. According to Squalli and
Wilson (2006) the various measures of trade openness provide a method for determining how
open an economy is to world trade and the income growth benefits that flow from trade
irrespective of the trade openness measure used. The most basic measure of trade openness
which is used in this study is .
Investment
This refers to Gross fixed capital formation in R billion. This is the government investment in
infrastructure. This includes both public and private investments.
Real interest rate
The real interest rate is the nominal interest rate adjusted for expected inflation and is usually
measured as the difference between the nominal interest rate and the expected or actual inflation.
Broad money supply
is the broadest definition of money supply as it includes all bank notes and coins in
circulation plus all deposits of the domestic private sector with banking institutions. It is the
money supply aggregate that has the most stable relationship with domestic demand and is
unaffected by deposit shifts between different maturities.
51 | P a g e
4.3.1 Priori Expectations
The parameter captures the effect of real exchange rate on output growth. For the purpose of
this study, the sign of this parameter is critical. As long as the parameter is statistically
significant, a positive sign will indicate an expansionary, while a negative sign will indicate a
contractionary effect. An increase in real interest rates increases the cost of capital and can be
contractionary and hence is expected to be below zero.
is expected to be greater than 0, for
instance, trade openness positively affect output. is also expected to be positive because an
increase in money supply is expected to affect output positively. is expected to be positive as a
rise in fixed capital formation is expansionary to economic growth.
4.4 DATA SOURCES
This study uses quarterly data covering the period 1994-2010. GDP and trade figures are
obtained from the South African Department of Trade and Industry, data on investment and
money supply is obtained from the South African Reserve Bank (SARB) publications, and
Interest rate and exchange figures are obtained from the South African Department of Statistics
(StatsSA).
4.5 RESEARCH TECHNIQUES
The study employs the Johansen (1995) and Johansen and Juselius (1990) cointegration
technique. The technique is well known for establishing the long run relationship between
variables. This approach applies maximum likelihood estimation to a vector error correction
(VEC) model to simultaneously determine the long run and short run determinants of the
dependant variable in a model. Firstly data has to be integrated of the same order. To achieve
this, unit root tests to examine stationarity of data sets are carried out. In investigating the unit
root properties of the time series data, the variables shall be subjected to the Augmented Dickey-
Fuller (ADF) and Phillips-Perron unit root test.
4.5.1 Testing for Stationarity
Stationarity is defined as a quality of a process in which the statistical parameters (mean and
standard deviation) of the process do not change with time (Challis and Kitney, 1991).The
52 | P a g e
assumption of the classical regression model necessitate that both the dependent and independent
variables be stationary and the errors have a zero mean and finite variance. According to
Newbold and Granger (1974) the effects of non stationarity includes spurious regression, high R2
and low Durbin-Watson (dw) statistic. Below are basic reasons why data must be tested for non
stationarity.
First, the stationarity or otherwise of a series can strongly influence its behaviour and properties,
for instance, persistence of shocks will be infinite for non-stationary series. Secondly, if two
variables are trending over time, a regression of one, on the other hand, could have a high
even if the two are totally unrelated and this is known as spurious regressions. Thirdly, if the
variables in the regression model are not stationary, then it can be proved that the standard
assumptions for asymptotic analysis will be invalid. In other words, the usual “t-ratios” will not
follow a t-distribution, so it is impossible to validly undertake hypothesis tests about the
regression parameters (Bowerman and O'connell, 1979).
4.5.2 Augmented Dickey-Fuller (ADF) test
The augmented dickey fuller test modifies the work done by Dickey and Fuller (1979 and 1976
respectively).The aim of the Dickey Fuller theory was to test the hypothesis that in:
………………………………………………………………………………4.3.
Thus, the hypotheses are formulated:
: Series contains a unit root
: Series is stationary.
The rejection of the null hypothesis under these tests means that the series does not have a unit
root problem.
The standard Dickey Fuller test estimates following equation:
………….………………………………………………………..4.4
Where is the relevant time series, Δ is a first difference operator, t is a linear trend and is
the error term. The error term should satisfy the assumptions of normality, constant error
variance and independent error terms. According to Gujarati (2004) if the error terms are not
independent in equation (4.4), results based on the Dickey-Fuller tests will be biased.
53 | P a g e
The weakness of the DF test is that it does not take account of possible autocorrelation in the
error process or term ( ). Clemente, et al (1998) noted that a well–known weakness of the
Dickey–Fuller style unit root test with I(1) as a null hypothesis is its potential confusion of
structural breaks in the series as evidence of non-stationarity.
Blungmart, (2000) stated that the weakness of the Dickey-Fuller test is that it does not take
account of possible autocorrelation in error process, . If is auto-correlated, then the OLS
estimates of coefficients will not be efficient and t-ratios will be biased. In view of the above
mentioned weaknesses the Augmented Dickey-Fuller test was postulated and is preferred to the
Dickey-Fuller test.
The presence of serial correlation in the residuals of the Dickey-Fuller test biases the results
(Mahadeva and Robinson, 2004). When using the Dickey-Fuller test the assumption is that the
error terms are uncorrelated. But in case the
are correlated, Dickey and Fuller developed a
test, known as the Augmented Dickey-Fuller test to cater for the above mentioned problem.
The Dickey-Fuller test is only valid where there is no correlation of the error terms. If the time
series is correlated at higher lags, the augmented Dickey-Fuller test constructs a parameter
correction for higher order correlation, by adding lag differences of the time series. The
Augmented Dickey-Fuller test estimates the following equation:
………………………………………………….4.5.
Where is a pure white noise error term and where = ( , ),
etc. According to Gujarati (2004) the number of lagged difference terms to include is often
determined empirically, the idea being to include enough terms so that the error term in (4.5) is
serially uncorrelated. In ADF as in DF the test is whether = 0 and the ADF test follows the
same asymptotic distribution as the DF statistic, so the same critical values can be used.
The calculated value of ADF is then compared with the critical value. If the calculated value is
greater that the critical, we reject the null hypothesis that the series have unit root, thus
confirming that the series are stationary.
54 | P a g e
In a nutshell Gujarati (2004) states that an important assumption of the DF test is that the error
terms are independently and identically distributed. The ADF test adjusts the DF test to take
care of possible serial correlation in the error terms by adding the lagged difference terms of the
regressand.
4.5.3 Phillips-Perron (PP) Tests
The Phillips-Perron tests are a more comprehensive theory of unit root non-stationarity. Gujarati
(2004) states that the Phillips-Perron use non-parametric statistical methods to take care of the
serial correlation in the error terms without adding lagged difference terms. According to Brooks
(2008) the tests are similar to ADF tests, but they incorporate an automatic correction to the DF
procedure to allow for auto correlated residuals. The PP test and the ADF test have the same
asymptotic distribution. Brooks (2008) explains that the PP tests often give the same conclusions
as, and suffer from most of the same important limitations as, the ADF tests.
4.5.3.1 Criticisms of Dickey-Fuller and Phillips-Perron type tests
Brooks (2008) argues that the most important criticism of the unit root tests is that their power is
low if the process is stationary but with a root close to the non-stationary boundary. For instance,
an AR (1) data generating process with coefficient 0.95. If the true data generating process is
the null hypothesis of a unit root should be rejected. It has been thus
argued that the tests are poor at deciding, for example, whether , especially
with small sample sizes.
Brooks (2008) further argues that the source of this problem is that, under the classical
hypothesis-testing framework, the null hypothesis is never accepted, it is simply stated that it is
either rejected or not rejected. This means that a failure to reject the null hypothesis could occur
either because the null was correct, or because there is insufficient information in the sample to
enable rejection.
4.5.4 Cointegration and Vector Error Correction Modeling (VECM)
When dealing with time series data, there is need to check if the individual time series are either
stationary or that they are co-integrated. If that is not the case, there is great chance of engaging
in spurious (or nonsense) regression analysis (Gujarati, 2010). If two series appear to move
55 | P a g e
together over time, it suggests that there exist an equilibrium relationship. This therefore shows
that even though the variables are non-stationary in the short run if they are co-integrated, they
will move closely together over time and their difference will be stationary.
The vector autoregressive (VAR) model is a general framework used to describe the dynamic
interrelationship among stationary variables. Dolado et al. (1999) states that if the time series are
not stationary then the VAR framework needs to be modified to allow consistent estimation of
the relationships among the series. The vector error correction (VEC) model is just a special case
of the VAR for variables that are stationary in their differences (for instance, I (1)). The VEC can
also take into account any cointegration relationships among the variables.
In order to justify the use of vector error correction model (VECM) there is need to test for
cointegration. A VECM is intended to be used with non-stationary series that are known to be
cointegrated. Brooks (2008) contends that the VECM has cointegration relations built into the
specification so that it restricts the long-run behaviour of the endogenous variables to converge
to their co-integrating relationships while allowing for short-run adjustment dynamics. Brooks
(2008) also states that the cointegration term is known as the correction term since the deviation
from long-run equilibrium is corrected gradually through a series of partial short-run adjustments
estimated. Thus, the presence of a cointegration relation(s) forms the basis of the vector error
correction model (VECM) specification.
There are several methods of testing for cointegration, but two often stand above the rest namely
the Engle-Granger approach which is residual based and the Johansen and Julius (1990)
technique which is based on maximum likelihood estimation on a VAR system. Brooks (2008)
argues that the problems of the Engle-Granger approach include lack of power in unit root tests,
simultaneous equation bias and the impossibility of performing hypothesis tests about the actual
cointegration relationships.
In light of the above mentioned shortfalls of the Engle-Granger approach this study applies the
vector error correction modeling (VECM) by Johansen (1991; 1995). The rationale behind being
that this approach applies maximum likelihood estimation to a vector error correction (VEC)
model to simultaneously determine the long run and short run determinants of the dependant
variable in a model. This approach also provides the speed of adjustment coefficient, which
56 | P a g e
measures the speed at which Gross Domestic Product reverts to its equilibrium following a short
term shock to the system (Greene, 2000).
4.5.5 Johansen Technique Based on VARS
According to Greene (2000) the following steps are used when implementing the Johansen
procedure:
(1) Step 1: Testing for the order of integration of the variables under examination. All the
variables should be integrated of the same order before proceeding with the cointegration test.
(2) Step 2: This step involves setting the appropriate lag length of the model. Also in the step is
the estimation of the model and the determination of the rank of П.
(3) Step 3: With regards to the deterministic components in the multivariate system the choice of
the appropriate model is made. An analysis of the normalised co-integrating vector(s) and speed
of adjustment coefficients is made.
(4) Step 4: Step 4 includes the determination of the number of co-integrating vectors. Causality
tests on the error correction model to identify a structural model and determine whether the
estimated model is reasonable is done in this last step.
After ascertaining the existence of co-integrating relationships, the vector error correction model
(VECM) is estimated to test for the short-run dynamics. We consider the following VAR of
order P:
………………………………………………………..4.6.
where is is a -vector of non-stationary I(1) variables, is a d-vector of deterministic
variables, and is a vector of innovations. In order to use the Johansen test, the VAR (4.6)
above needs to be turned into a VECM specification (Brooks, 2008). We may rewrite this VAR
as:
..............................................................................4.7.
Where:
57 | P a g e
…………………………………………………………4.8.
Granger’s representation theorem asserts that if the coefficient matrix has reduced rank ,
then there exist matrices and each with rank such that and
is I(0). is
the number of cointegration relations (the cointegration rank) and each column of is the
cointegrating vector. The elements of are known as the adjustment parameters in the VEC
model. Johansen’s method is to estimate the matrix from an unrestricted VAR and to test
whether we can reject the restrictions implied by the reduced rank of (Green, 2007).
4.6 IMPULSE RESPONSE ANALYSIS
Brooks (2008) stress that impulse response analysis traces out the responsiveness of the
dependent variable in the VAR to shocks to each of the other variables. In this study therefore, it
shows the sign, magnitude and persistence of real and nominal shocks to the real growth. Brooks
(2008) further states that impulse response analysis is applied on the VECM and, provided that
the system is stable, the shock should gradually die away. This study applies the generalised
impulse response analysis. Lutkepohl (1993) cited in Rusike (2007) explains that this approach
fully takes into account historical patterns of correlations amongst the different shocks.
4.7 VARIANCE DECOMPOSITION ANALYSIS
After performing the impulse response analysis further information on the link between
economic growth and exchange rates is found using the variance decomposition analysis. Brooks
(2008) explains that variance decomposition analysis provides the proportion of movements in
the dependent variables that are due to its own shocks, against shocks to other variables.
4.8 DIAGNOSTIC CHECKS
The diagnostic tests are very important in the analysis of the impact of real exchange rates on
economic growth in South Africa because it validates the parameter estimation outcomes
achieved by the estimated model. Diagnostic checks test the stochastic properties of the model
such as residual autocorrelation, heteroscedasticity and normality, and many more. These
mentioned tests are applied in this study and, hence, they are briefly discussed below.
58 | P a g e
4.8.1 Heteroscedasticity
The OLS makes the assumption that for all j. That is, the variance of the error term is
constant a condition termed homoscedasticity. If the error terms do not have constant variance,
they are said to be heteroscedastic. The study employs the White heteroscedasticity test.
According to Greene (2000) white test computes the White (1980) general test for
heteroscedasticity in the error distribution by regressing the squared residuals on all distinct
regressors, cross-products, and squares of regressors. The test statistic, a Lagrange multiplier
measure, is distributed Chi-squared (p) under the null hypothesis of homoscedasticity. The null
hypothesis for the White test is homoscedasticity and if we fail to reject the null hypothesis then
we have homoscedasticity. If we reject the null hypothesis, then we have heteroscedasticity.
4.8.2 Residual Normality Test
The assumption of normality is . The null is that the skewness ( ) and kurtosis ( )
coefficients of the conditional distribution of (or, equivalently, of the distribution of ) are 0
and 3, respectively:
: (if then f ( ) is skewed to the left)
(if then f ( ) is leptokurtic)
The above assumptions can be tested using the Jarque-Bera test (JB). The JB test follows the null
hypothesis that the distribution of the series is symmetric. The null hypothesis of normality
would be rejected if the residuals from the model are either significantly skewed or leptokurtic
(or both).
4.8.3 Autocorrelation LM Tests
Serial correlation happens when the error terms from different time periods (or cross-section
observations) are correlated. In time series studies it occurs when the errors associated with
observations in a given time period carry over into future time periods. Serial correlation (also
called autocorrelation) in the residuals means that they contain information, which should itself
be modeled. The Durbin-Watson statistic is used in the study to test for the presence of first order
serial correlation in the residuals. The null hypothesis is no serial correlation . The
59 | P a g e
DW statistic lies in the 0 to 4 range, with a value near 2 indicating no first order serial
correlation. The Lagrangian Multiplier was used to test for serial correlation.
4.8.4 Misspecification Tests
Misspecification errors happens when some important variables are omitted from the model. The
Ramsey Reset tests are used in this study.
4.9 SUMMARY
This chapter which was informed by economic theory and empirical evidence specified the
model of the impact of exchange rates on economic growth in South Africa. The dependent
variable is economic growth and the main explanatory variable being real effective exchange
rates. Other explanatory variables as suggested by the theory include fixed capital formation as
proxy for investment, exports, trade openness, real interest rates and broad money supply m3. In
order to make sure that data is stationary the Dickey-Fuller and Augmented Dickey-Fuller unit
root tests were employed. The cointegration and vector error correction modeling (VECM) by
Johansen (1991, 1995) technique was also proposed. The study makes use of a number of
diagnostic tests such as the residual normality test, heteroscedacity and the autocorrelation tests
in order to validate the parameter estimation outcomes achieved by the estimated model.
Having outlined the methodology that is used in this study, the next chapter presents the
estimation, presentation and analysis of the research findings.
60 | P a g e
CHAPTER FIVE
PRESENTATION AND ANALYSIS OF EMPIRICAL FINDINGS
5.1 INTRODUCTION
The main aim of this chapter was to address some of the questions raised in the first chapter.
Results from this chapter explains the impact of real exchange rates on economic growth in
South Africa using quarterly data for the period between 1994 and 2010. The analysis follows
the analytical framework presented in Chapter Four. The chapter is divided into six subsections.
The unit root test is presented first, followed by cointegration tests. This leads to the formulation
of the vector error correction model (VECM) which is followed by diagnostic checks, impulsive
response and variance decomposition. A summary for the chapter is finally provided.
5.2 UNIT ROOT/STATIONARITY TEST RESULTS
The opening stage of the Johansen procedure is to test for stationarity in time series. There are
two main methods to test whether time series are stationary or not, namely; the graphical method
which is informal and then the formal test. This study first presents the visual plot of graphs
before the formal tests. The formal tests conducted are the Augmented Dickey-Fuller and the
Phillips-Peron tests. These tests are very important as they give insight into the structural breaks,
trends and stationarity of the data set (Brooks, 2008).The graphical results from the test for
stationarity are presented in Figure 5.1 on the next pages: (a) which shows data in level form and
Figure 5.1(b) for first differenced data. The Augmented Dickey-Fuller and the Phillips-Peron test
results are shown in Tables 5.1(a) and 5.1(b).
Figure 5.1(a) shows that Gross Domestic Product (GDP), Gross fixed capital formation
investment (FCF) and Money supply (MS) show trendy behaviour. These three variables have a
growth trend; real interest rate (RIR) and real effective exchange rate, however, with a
downward trend. REER shows a downward trend until 2002 followed by a growth trend in the
years after. RIR generally shows a downward trend except for the years between 1994 and 1998,
which show a growth trend. The series in levels is clearly non-stationary. Figure 5.1(b) shows
that all the differenced variables fluctuate around the zero mean hence the variables are
integrated of order one I(1).
61 | P a g e
Using the graphical method, data that fluctuate around the zero mean indicate stationarity. It can
be noted that Figure 5.1(a) shows series before differencing and hence are non-stationary as the
mean is not zero. Figure 5.1(b), however, shows stationary series after differencing and the
means are fluctuating around zero. It can, therefore, be concluded that Figure 5.1(b) shows
stationary data after differencing. This implies that the data is stationary if integrated of order
one. The first order integrated series ensure that economic data is stationary for the purpose of
avoiding spurious regressions. The informal method, however, is not enough to conclude that
data is stationary as it is informal, hence the need for a more formal method to complement it.
Consequently, other formal tests were conducted to support findings from the graphical findings.
In this regard, the Augmented Dickey-Fuller and the Phillips-Peron tests were adopted and the
results are presented in Table 5.1(a) and 5.1(b) below.
62 | P a g e
Figure 5.1(a) Plots of variables in levels for 1994 – 2010
10.2
10.4
10.6
10.8
11.0
11.2
11.4
11.6
94 96 98 00 02 04 06 08 10
FCF
12.4
12.5
12.6
12.7
12.8
12.9
13.0
13.1
94 96 98 00 02 04 06 08 10
GDP
12.0
12.5
13.0
13.5
14.0
14.5
15.0
94 96 98 00 02 04 06 08 10
MS
4.3
4.4
4.5
4.6
4.7
4.8
4.9
94 96 98 00 02 04 06 08 10
R E E R
0.5
1.0
1.5
2.0
2.5
3.0
94 96 98 00 02 04 06 08 10
RIR
-.9
-.8
-.7
-.6
-.5
-.4
94 96 98 00 02 04 06 08 10
TO
63 | P a g e
Figure 5.1 (b) Plots of first differenced variables for 1994-2010
-30,000
-20,000
-10,000
0
10,000
20,000
1994 1996 1998 2000 2002 2004 2006 2008 2010
Differenced GDP
-6,000
-4,000
-2,000
0
2,000
4,000
6,000
1994 1996 1998 2000 2002 2004 2006 2008 2010
Differenced FCF
-3,000,000
-2,000,000
-1,000,000
0
1,000,000
2,000,000
3,000,000
1994 1996 1998 2000 2002 2004 2006 2008 2010
Differenced MS
-20
-15
-10
-5
0
5
10
15
1994 1996 1998 2000 2002 2004 2006 2008 2010
Differenced REER
-.08
-.04
.00
.04
.08
1994 1996 1998 2000 2002 2004 2006 2008 2010
Differenced TO
-4
-2
0
2
4
6
1994 1996 1998 2000 2002 2004 2006 2008 2010
Differenced RIR
64 | P a g e
Table 5.1(a): Stationarity results of the Augmented Dickey-Fuller test
Augmented Dickey-Fuller
Order of integration Variable Intercept Trend and
intercept
None
Level LGDP 0.213 -2.529 3.300
1st difference DGDP -8.607*** -8.641*** -6.849***
Level LREER -2.301 -2.090 -0.265
1st difference DREER -6.308*** -6.411*** -6.359***
Level LRIR -2.056 -3.138 -0.936
1st difference DRIR -5.550*** -5.573*** -5.594***
Level LFCF 0.082 -1.543 2.374
1st difference DFCF -5.048*** -5.065*** -4.206***
Level LMS -0.077 -9.839 0.361
1st difference DMS -36.943*** -41.179*** -23.471***
Level LTO -2.776 -2.518 0.401
1st difference DTO -5.711*** -5.743*** -5.735***
1% Critical values -3.535 -4.403 -2.601
5% -2.907 -3.480 -1.946
10% -2.591 -3.167 -1.614
Values marked with a *** represent stationary variables at 1% significance level, and **
represent stationary at 5% and * represent stationary variables at 10%.
65 | P a g e
Table 5.1 (b): Stationarity results of the Phillips-Perron test
Phillips-Perron
Order of integration Variable Intercept Trend and
intercept
None
Level LGDP -0.120 -3.073 3.136
1st difference DGDP -11.502*** -11.489*** -10.092***
Level LREER -2.432 -2.143 -0.427
1st difference DREER -7.591*** -7.631*** -7.651***
Level LRIR -1.966 -3.239 -0.677
1st difference DRIR -6.700*** -6.657*** -6.750***
Level LFCF 0.149 -1.327 3.135
1st difference DFCF -6.198*** -6.213*** -5.419***
Level LMS -1.707 -4.762 -0.041
1st difference DMS -19.590*** -21.247*** -19.700***
Level LTO -2.630 -2.660 0.200
1st difference DTO -10.244*** -10.327*** -10.276***
1% Critical values -3.532 -4.101 -2.600
5% -2.906 -3.478 -1.946
10% -2.590 -3.167 -1.614
Values marked with a *** represent stationary variables at 1% significance level, and **
represent stationary at 5% and * represent stationary variables at 10%.
Table 5.1(a) shows the Augmented Dickey-Fuller results. The test has a null hypothesis of unit
root. The calculated value of ADF was compared with the critical value. If the calculated value is
greater than the critical, we then reject the null hypothesis that the series have unit root, thus
confirming that the series are stationary. The ADF tests variables in (a) intercepts, (b) trends and
intercepts and (c) no trend and no intercept. For variables in levels, the test in intercepts revealed
that all variables were not stationary. For the intercept, all the data in levels was not stationary as
66 | P a g e
reflected by the non-rejection of the null hypothesis at both 1% and 5 % significance levels. All
the differenced variables were stationary at 1% significant level; hence the null hypothesis of
unit root is rejected. For the test under trend and intercept and trend and no intercept data series
were all non-stationary in levels but became stationary at 1% significant level when first
differenced.
Table 5.1(b) shows the Phillips-Peron results. According to Brooks (2008) the tests are similar to
ADF tests, but they incorporate an automatic correction to the DF procedure to allow for auto
correlated residuals. For variables in levels, the test in intercepts revealed that none of the
variables were stationary. All differenced variables on intercept were stationary at 1%
significance level. On trend and intercept all variables were non-stationary in levels but all
variables on trend and intercept were stationary at 1% significance level when first differenced.
For the test under no trend and no intercept, all variables in levels were non-stationary. When
first differenced, all the variables were stationary at 1% significance.
Both methods used to test for stationarity significantly revealed that the data series were non-
stationary in levels and stationary when first differenced. Therefore, the series are integrated of
the same order I(1).
5.3 TESTS FOR COINTEGRATION
If the variables are integrated of the same order, it is very important to determine whether there
exists a long-run equilibrium relationship amongst them. Cointegration describes the existence of
an equilibrium or stationarity relationship between two or more times series each of which is
individually non stationary. For the purposes of this study cointegration examines the long run
relationship between the gross domestic product and its determinants. It is very important to
assess whether there exists long run relationships between gross domestic product and the chosen
determinants, in order for a viable economic conclusion to be reached from the results obtained.
The cointegration approach allows researchers to integrate the long run and short run relationship
between variables within a unified framework (Andren, 2007). The Johansen cointegration
approach is preferred over the Engle and Granger residual-based methodology to test for
cointegration because of the obvious reasons mentioned in Chapter 4.
67 | P a g e
Since all variables are non-stationary in level, the next procedure is to test for the existence of
long run relationships among the variables in the model. The cointegration test using Johansen
test requires the estimation of a VAR equation. The variables i.e. LREER, LRIR, LMS, LOP,
and LFCF are entered as endogenous variables. Table 5.2 below presents the pair-wise
correlation test results.
Table 5.2: Pair-wise Correlation results
GDP REER RIR OP MS FCF
GDP 1.00 -0.200179 -0.308773 0.481542 0.249385 0.104350
REER -0.200179 1.00 -0.042382 -0.171443 -0.012654 0.173781
RIR -0.308773 -0.042382 1.00 -0.171443 0.132809 -0.378257
OP 0.481542 -0.171443 0.021417 1.00 0.439242 0.084065
MS 0.249385 -0.012654 0.132809 0.439242 1.00 -0.016419
FCF 0.104350 0.173781 -0.378257 0.084065 -0.016419 1.00
From the pair-wise correlation results shown in Table 5.2, it is observed that OP is highly
correlated with GDP, followed by RIR. Variables (REER and RIR) are negatively correlated
with the dependent variable (GDP). This negative correlation is in line with theoretical
underpinnings which suggest that increases in interest rates and real exchange rates will
discourage investment and exports respectively thus reducing economic growth. FCF and OP
have positive correlation with GDP which validates the theoretical underpinnings which hold
that increases in investment and trade activities results in increases in economic growth.
The information criteria approach is applied in this study as a direction to choose the lag order. It
is a requirement of the Johansen technique to show an indication of the lag order and the
deterministic trend assumption of the VAR. Table 5.3 confirms the lag lengths selected by
different information criteria.
68 | P a g e
Table 5.3: Lag order selection criteria
Lag
LogL
LR FPE AIC SC
HQ
0 321.7273 NA 1.79e-12 -10.02309 -9.818982 -9.942813
1 734.9081 734.5437 1.13e-17 -21.99708 -20.56833* -21.43515
2 788.9004 85.70206 6.60e-18 -22.56827 -19.91486 -21.52467
3 820.3785 43.96931 8.25e-18 -22.42471 -18.54666 -20.89946
4 892.3342 86.80379 3.09e-18 -23.56617 -18.46346 -21.55925
5 951.8420 60.45229* 1.95e-18* -24.31244* -17.98509 -21.82387*
Notes
* indicates lag order selected by the criterion
LR: sequential modified LR test statistic (each test at 5% level)
FPE: Final prediction error
AIC: Akaike information criterion
SC: Schwarz information criterion
HQ: Hannan-Quinn information criterion
Table 5.3 confirms that the criteria selected 5 lag. Consequently, using the information criteria
approach, the Johansen cointegration test was conducted using 5 lag for the VAR.
The trace test results based on the Johansen cointegration are shown in Table 5.4(a). The null
hypothesis of the trace test is that the number of co-integrating equations is greater than the
number of variables involved. If the test statistic is smaller than critical values of the trace tests
we do not reject the null hypothesis. Table 5.4(b) presents the results of the Johansen
cointegration test based on the maximum eigenvalue. The maximum eigenvalue test was
conducted on a null hypothesis of the number of cointegration equations (r) against the
alternative hypothesis of number of cointegration equations plus one (r +1). We do not reject the
null hypothesis if the test statistic is smaller than the maximum eigenvalue test’s critical values.
69 | P a g e
Table 5.4(a): Co-integration Rank Test (Trace)
Hypothesised
No. Of CE(s)
Eigenvalue Trace Statistic 0.05 Critical
Value
Prob**
None*
0.615084 62.05753 40.07757 0.0001
At most 1
0.544707 33.87687 51.14299 0.0002
At most 2
0.330947 26.12297 27.58434 0.0760
At most 3
0.169760 12.09264 21.13162 0.5385
At most 4
0.078585 5.319883 14.26460 0.7010
At most 5
0.003438 0.223852 3.841466 0.6361
Trace test indicates 1cointegratingeqn(s) at the 0.05 level
* denotes rejection of the hypothesis at the 0.05 level
**MacKinnon-Haug-Michelis (1999) p-values
Table 5.4(b): Co-integration Rank Test (Maximum Eigenvalue)
Hypothesised
No. Of CE(s)
Eigenvalue Trace Statistic 0.05 Critical
Value
Prob**
None*
0.615084 156.9599 95.75366 0.0000
At most 1*
0.544707 94.90234 69.81889 0.0002
At most 2
0.330947 43.75935 47.85613 0.1151
At most 3
0.169760 17.63638 29.79707 0.5930
At most 4
0.078585 5.543734 15.49471 0.7487
At most 5
0.003438 0.223852 3.841466 0.6361
Max-Eigen value test indicates 1 cointegration Eqn(s) at the 0.05 level
* denotes rejection of the hypothesis at the 0.05 level
**MacKinnon-Haug-Michelis (1999) p-values
70 | P a g e
Figure 5.4 shows the results of the trace test which reflect that at least one co-integrating
equation exists at 5% significance level. The null hypothesis of no co-integrating vectors is
rejected since the trace (test) statistic of 62.057 is greater than the 5% critical value of
approximately 40.077. Using a similar explanation, the null hypothesis that there is at most 1 co-
integrating vector cannot be rejected since the test statistic of approximately 33.877 is less than
the 5% critical value of about 51.143. For that reason, the trace statistics specified 1 co-
integrating relationship at 5% significance level. The maximum Eigen value test in Table 5.4(b)
put forward that there is only 1 co-integrating relationship in the gross domestic product model.
The maximum Eigen value test also rejected the null hypothesis of no cointegration, but failed to
reject that at most 1 co-integrating vectors exist, since the test statistic of about 43.759 is less
than the 5% critical value of about 47.856. Therefore, it can be concluded that there is one
significant long run relationship between the given variables (using the trace test). Since
variables can either have short or long run effects, a vector error correction model (VECM) is
used to disaggregate these effects.
Table 5.4(a) indicates the existence of one cointegration vector. The cointegration vector
represents the deviations of the endogenous variable from its long run equilibrium level. Figure
5.2 shows that over the period 1994 to 2010, the deviations of economic growth from
equilibrium were stationary and this is critical in its use as an error correction model.
71 | P a g e
Figure 5.2: Cointegration vector
5.4 VECTOR ERROR CORRECTION MODEL (VECM)
The detection of a cointegration equation in the previous section means that a VECM can be
used. This has led to a distinction between the long and short run impacts of variables so as to
establish the extent of influence that real exchange rates has on economic growth. Using the
results from the cointegration test the VECM was specified. The VECM results are presented in
Tables 5.5 and 5.6.
-.10
-.08
-.06
-.04
-.02
.00
.02
.04
.06
95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10
Cointegrating relation 1
72 | P a g e
Table 5.5: Results of the long run cointegration equation
Variable
Coefficient Standard error t-statistic
Constant
-10.23693 - -
LOG_GDP 1.000000 - -
LOG_REER -0.163071 0.03974 -4.10355
LOG_RIR 0.091851 0.01127 8.15153
LOG_OP -0.199206 0.05632 -3.53710
LOG_MS -0.257361 0.01814 -14.1872
LOG_FCF 0.124363 0.04866 2.55592
The long run impact of real exchange rates on economic growth as presented in Table 5.5 is
illustrated using Equation 5.1:
Equation 5.1 shows that REER, RIR and FCF have a positive long run relationship with GDP.
On the other hand, MS and OP show a negative long run relationship with GDP. All the
variables are statistically significant in explaining economic growth since they have absolute t-
values greater than 2. The results suggest that a unit increase in REER which is a depreciation of
the South African rand against its trading partners, reduces economic growth in the long run by
approximately 0.163. This shows that despite the fact that a negative relationship exists in the
short run as depicted by Table 5.6 below; in the long run it is not sustainable. In the long run the
results follow the Structuralist view to exchange rates reviewed in Chapter Two which holds that
depreciation might have a contractionary effect on output and employment, especially for less
economically developed countries. Depreciation increases the cost of imports in particular, and
the cost of domestic production in general, through imported inputs
In the long run a unit increase in RIR increases economic growth by approximately 0.092. An
increase in RIR in the long run attracts foreign direct investment especially portfolio investments
which improves South Africa’s balance of payments accounts thus increasing economic growth.
A unit increase in FCF increases economic growth in the long run by approximately 0.124.
Investment in public and private infrastructure like roads, plant and equipment expands the
73 | P a g e
country’s production capacities. This in turn guarantees increases in the gross domestic product
of a country in the long run.
In this study, MS has a negative long run effect on economic growth. A unit increase in MS
reduces economic growth by approximately 0.257. An increase in MS in the long run can be
inflationary and this has negative effects on economic growth. Lastly, OP was also found to be
negatively related to economic growth. A unit increase in OP reduces GDP by approximately
0.199. Trade openness of South Africa to the whole world comes with its own disadvantages
such as capital flight, effects of international financial instability, for instance global financial
crises and the South East Asian crises of 1998 and 1999, the Zimbabwean crises and imported
inflation. These factors contributed a great deal to the negative impact of OP on growth during
the study period.
Table 5.6: Error correction model results
Variable
Coefficient
Standard error
t-statistic
D(LOG_GDP) -0.139042 (0.15133) -2.91882
D(LOG_REER) 3.046736 1.10049 2.76853
D(LOG_RIR) -12.02454 3.57355 -3.36487
D(LOG_OP) 0.414884 0.58940 0.70390
D(LOG_MS) -0.120453 0.28205 -0.42707
D(LOG_FCF) -0.486794 0.33937 -1.43441
In Table 5.6, the coefficient of D (LOG_GDP) of -0.139 shows that the speed of adjustment is
approximately 13.9 percent. This means that if there is a deviation from equilibrium, only 13.9
per cent is corrected in one quarter as the variable moves towards restoring equilibrium. This
means that there is no strong pressure on economic growth to restore long run equilibrium
whenever there is a disturbance. This speed of adjustment is statistically significant with an
absolute t-value of approximately 2.91882. The low speed of adjustment by economic growth
may reflect the existence of some other factors affecting economic growth in South Africa which
are not specified in the model.
In the short run it can be seen in Table 5.6 that the real exchange rate has a negative effect on
growth. A unit increase in real exchange rates which is a depreciation of the South African rand
74 | P a g e
increases economic growth by approximately 3.046. This is compatible with the traditional
approach to exchange rates which holds that devaluations have expansionary effects on
economic growth. Depreciation of a currency will cause local goods to be cheaper abroad and
this will increase their demand leading to an increase in exports thereby improving the trade
balance and accordingly expand output and employment. Depreciations may in some cases be a
quick fix in the short run, but not sustainable in the long run. As is the case in this study, in the
long run the relationship becomes positive.
Real interest rates in the short run have a negative impact on growth. An increase in RIR reduces
growth by approximately 12.024. An increase in the real interest rates in the short run increases
the cost of borrowing thereby dampening investment in the country, leading ultimately to a
decrease in gross domestic product in the short run. In the long run, it is different as high interest
rates helps in keeping price stability and low inflation which in turn improves growth prospects.
Real exchange rates and real interest rates are statistically significant in explaining economic
growth in South Africa in the short run as seen by absolute t-values of above 2. The other
variables FCF, OP and MS have insignificant t-values of below 2.
5.5 DIAGNOSTIC CHECKS
The economic growth model was subjected to thorough diagnostics tests. The model was tested
for normality, serial correlation, autoregressive conditional heteroscedasticity and stability.
Diagnostic checks are performed to the GDP modeling in order to validate the parameter
evaluation of the outcomes achieved by the model. Any problem in the residuals from the
estimated model makes the model to be not efficient and the estimated parameters will be biased.
For the purposes of this study, the VAR model was subjected to diagnostic checks. The
diagnostic test results are presented in Table 5.7 below and these assist in checking for serial
correlation, normality and heteroscedasticity. These diagnostic checks are based on the null
hypothesis that: there is no serial correlation for the LM test; there is normality for the Jarque-
Bera test and there is no heteroscedasticity for the White heteroscedasticity test.
75 | P a g e
Table 5.7: Diagnostic checks results
Test
Null Hypothesis
t-Statistic Probability
Langrage
Multiplier (LM)
No serial correlation
17.450 0.634
White (CH-sq)
No conditional
heteroscedasticity 0.875 0.489
Jarque-Bera (JB)
There is a normal
distribution 5.737 0.056
Results from Table 5.7 show that the test for serial correlation produced an LM statistic of
17.450 with a probability of 0.634. For the Histogram and Normality Test, Jarque-Bera is 5.737
and the probability is 0.056. Thus, the Jarque-Bera statistic is insignificant as it is above the 5
percent significance level. More so, the histogram is bell-shaped, thus, the residuals are normally
distributed. Therefore, the null hypothesis of a normal distribution was not rejected.
Heteroscedasticity tests showed the F-statistic of 0.875 and the probability of 0.489 which means
that the null hypothesis of no heteroscedasticity was accepted. The alternative hypothesis was
that there is heteroscedasticity. This means that the residuals are homoscedastic. The results for
the diagnostic checks for serial correlation and heteroscedasticity show that the data is fairly well
behaved. Results indicate the presence of non-normal residuals.
5.6 IMPULSE RESPONSE ANALYSIS
Impulse response analysis traces out the responsiveness of the dependent variables in a VAR to
shocks from each of the variables (Brooks 2008). Results of the impulse response analysis are
presented in Figure 5.4 on the next page.
76 | P a g e
Figure 5.4: Impulse response of GDP
Since this study focuses on the impact of real exchange rates on economic growth, only the
responses of economic growth to real exchange rates and the responses of economic growth to
economic growth are reported in Figure 5.4. These impulse response functions show the dynamic
response of economic growth to a one-period standard deviation shock to the innovations of the
system and also indicate the directions and persistence of the response to each of the shocks over
10 quarters. For the most part, the impulse response functions have the expected pattern and
-.010
-.005
.000
.005
.010
.015
1 2 3 4 5 6 7 8 9 10
Response of LOG_GDP to LOG_GDP
-.010
-.005
.000
.005
.010
.015
1 2 3 4 5 6 7 8 9 10
Response of LOG_GDP to LOG_FCF
-.010
-.005
.000
.005
.010
.015
1 2 3 4 5 6 7 8 9 10
Response of LOG_GDP to LOG_MS
-.010
-.005
.000
.005
.010
.015
1 2 3 4 5 6 7 8 9 10
Response of LOG_GDP to LOG_RIR
-.010
-.005
.000
.005
.010
.015
1 2 3 4 5 6 7 8 9 10
Response of LOG_GDP to LOG_TO
-.010
-.005
.000
.005
.010
.015
1 2 3 4 5 6 7 8 9 10
Response of LOG_GDP to LOG_REER
77 | P a g e
confirm the results from the short run relationship analysis. Shocks to all the variables are
significant although they are not persistent. Shock to the FCF, OP and MS have an enormous
dampening impact on economic growth. In the long run this effect seems to be reversed. A
REER shock will have a positive impact on economic growth except for the first two quarters
and the 8th
quarter onwards where there is a negative impact. Shocks to MS show a very
turbulent nature for example in the first two quarters, a shock to MS shows a dampening effect
on growth. From the second quarter, shocks to MS have positive effect to growth until 2.5
quarters were it reaches equilibrium and continues growing positive until the 5th
quarter then
turns negative again from the 6th
quarter to the 8th
quarter were it turns positive again. Shocks to
MS from the 8th
quarter to the 10th
quarter take a negative effect on growth.
5.7 VARIANCE DECOMPOSITION ANALYSIS
Variance decomposition analysis provides a means of determining the relative importance of
shocks in explaining variations in the variable of interest (Andren, 2007). Variance
decomposition provides a way of determining the relative importance of shocks to real exchange
rates in explaining variations in economic growth. The results of the variance decomposition
analysis are presented in Table 5.8 and these show the proportion of the forecast error variance in
economic growth as explained by its own innovations and innovations in real exchange rates.
Table 5.8: Variance decomposition of GDP
Period
S.E
LOGGDP LOGREER LOGRIR LOGOP LOGMS LOGFCF
1 0.022530 100.0000 0.000000 0.000000 0.000000 0.000000 0.000000
2 0.027007 88.42171 0.018915 5.021784 5.027464 0.028820 1.481311
3 0.028071 87.64746 0.092347 5.773033 4.663513 0.050707 1.772944
4 0.029311 87.87705 0.085513 5.395342 4.334555 0.067604 2.239933
5 0.031396 88.80387 0.170496 5.061382 3.778293 0.103042 2.082911
6 0.033063 88.49551 0.191862 5.489679 3.631355 0.131207 2.060387
7 0.034327 88.65880 0.181087 5.536713 3.378323 0.147839 2.097243
8 0.035546 88.89307 0.171688 5.422261 3.152349 0.162573 2.198058
9 0.036880 89.16606 0.170898 5.356722 2.934238 0.179910 2.192168
10 0.038141 89.26224 0.172740 5.409653 2.772700 0.195378 2.187284
78 | P a g e
Since this study focuses on the movements of economic growth following shocks to itself or real
exchange rates, the study reports only the variance decomposition in economic growth and it
analyses the relative importance of real exchange rates in influencing its movements.
The study allows the variance decompositions for 10 quarters in order to ascertain the effects
when the variables are allowed to affect economic growth for a relatively longer time. In the first
quarter, all of the variance in economic growth is explained by its own innovations (shocks), as
suggested in Brooks (2008). For the 5th quarter ahead forecast error variance, reported in column
2 of Table 5.8 under S.E., economic growth itself explains about 88 per cent of its variation,
while the other variables explain only the remaining 12 per cent. Of this REER explains 0.17 per
cent, RIR explains about 5.1 per cent, OP explains 3.8 per cent, MS explains 0.10 and FCF
explains about 2.08 per cent.
However, after a period of 10 quarters, economic growth explains about 89 per cent of its own
variation, while other variables explain the remaining 11per cent. The influence of REER was
still at 0.17 per cent, while RIR increased to about 5.4 per cent, OP decreased to 2.8 per cent, MS
increased to 0.19 and FCF increased to about 2.19 per cent. These results are similar to those
from the impulse response analysis in that all the variables have a significant impact on
economic growth in the short run.
Economic growth explains most of its variations, followed by real interest rates and then trade
openness and fixed capital formation. Money supply and real exchange rates though significant,
do not explain much of the variations in economic growth. Using the variance decomposition
analysis it can be seen that real interest rates are a very important variable in explaining
economic growth in South Africa over the study period.
5.8 SUMMARY
This chapter was divided into six sections. The first section presented the unit root test where the
Dickey-Fuller and the Philips-Peron tests were used to test for stationarity. Both methods
revealed that the data series are non-stationary in levels and stationary when first differenced.
Therefore, the series were integrated of the same order I (1).
79 | P a g e
Following the unity root tests was the cointegration tests in the second section. The cointegration
tests were done using the Johansen maximum likelihood approach. Also presented in this chapter
was the pair-wise correlation matrix and the lag order selection criteria. A maximum of 5 lags
were used to permit adjustments in the model and accomplish well behaved residuals. The trace
and maximum Eigen value cointegration tests were used to test for cointegration. The results
indicated that both the Trace and Maximum Eigen value tests reject zero in favour of at least one
cointegration vector. The results were significant at 5 percent level. These results prove that the
variables are tied together in a single way in the long run, that is, there is one unique long run
equilibrium relationship.
The third section presented the (VECM) model since variables can either have short or long run
effects. All the variables had a statistically significant effect on economic growth in South
Africa. The results showed that REER, RIR and FCF have a positive long run effect on economic
growth while on the other hand OP and MS have a negative long run effect on growth.
The last section presented the results of the diagnostic tests carried in the study. A number of
residual diagnostics tests were carried out and these revealed the fitness of the model. All of the
diagnostic tests supported the statistical appropriateness of the equation. Diagnostic checks were
performed to the GDP model in order to validate the parameter evaluation of the outcomes
achieved by the model. Any problem in the residuals from the estimated model makes the model
to be not efficient and the estimated parameters will be biased. Both the impulse response and
variance decomposition were found to be compatible with economic theory. Therefore, reliable
results and policy recommendations can be derived from this study.
Having outlined the estimation, presentation and analysis of the results in Chapter 5, the stage is
now set for the conclusions, policy recommendations and limitations of the study. These are
presented in the last and final chapter.
80 | P a g e
CHAPTER SIX
CONCLUSIONS, POLICY RECOMMENDATIONS AND LIMITATIONS
6.1 SUMMARY OF THE STUDY AND CONCLUSIONS
The purpose of this study was to econometrically evaluate the impact of real exchange rates on
economic growth in South Africa between the period 1994 and 2010. The first chapter was the
introduction and background of the study. This chapter laid all the ground work necessary to the
study including the objectives, hypothesis, statement of problem and the organisation of study.
Chapter two analysed the applicable theoretical and empirical literature. Following the chapter
on literature review was chapter three which gave an overview of the macroeconomic variables
over the study period. The methodology of the study was given in chapter four followed by
chapter five which presented the estimation, analysis and interpretation of the results. The last
chapter presents the conclusions, policy recommendations and limitations of the study.
Graphical analysis was used to perform a preliminary examination of the data and to explain the
behaviour and trends of the variables over the study period. An overview of South African trends
showed that real exchange rates were not stable as symbolised by fluctuations over the study
period. The rand exchange in South Africa is not fixed by the government; it is determined by
the forces and demand in the foreign exchange market.
After an extensive review of literature both theoretical and empirical, it became apparent that
real exchange rates were a significant variable in explaining changes in economic growth.
Theories considered in this study were: the Traditional approach to exchange rates, the
structuralist approach to exchange rates, the Balassa-Samuelson Hypothesis, and the export led
hypothesis. These theories had different foundations and assumptions, but the basic idea in them
all was that policy makers can use real exchange rates as a tool to achieve high levels of
economic growth. The other side (traditional approach) held that there is a positive correlation
between real exchange rats and economic growth whilst the other side (Structuralist approach)
held that exchange rates impacts growth negatively. Most of the empirical literature reviewed in
this study disclosed that real exchange rates affect economic growth in both developed and
developing countries as well as in South Africa.
81 | P a g e
In order to achieve the main objective of this study, it was necessary also to include other
variables in the growth model as suggested by both theoretical literature and empirical literature.
The choice of these variables was informed by an extensive review of literature on both
exchange rates and economic growth, and availability of data. This led to the specification of the
empirical growth model. The explanatory variables in this study included; real exchange rates,
real interest rates, fixed capital formation, trade openness and money supply.
After specifying the model the next step was to determine both the long and short run
relationships among the variables, which was done using the Johansen cointegration and error
correction methodology. This methodology was preferred over alternative methodologies like the
Engle-Granger approach because of its overwhelming benefits explained in Chapter 4. When
dealing with time series data, there is need to check if the individual time series are either
stationary or that they are cointegrated. To achieve this, a preliminary test for unit root was first
carried out using the graphical method. Due to the limitations of the graphical method, the
formal unit root tests were performed using the Augmented Dickey-Fuller and Phillips-Perron
tests. All the tests showed that all the time series were non stationary in levels but became
stationary after first differencing. The impulse response analysis was done to check the
responsiveness of the dependent variable in the VAR to shocks to each of the other variables. It
was noted that shocks to all the variables were significant although they are not persistent.
Variance decomposition analysis was done to check for the variables that explain most of the
variations in the dependent variable. It was observed that GDP explains much of its variations
followed by real interest rates. Diagnostic tests were performed on the residuals. This was to
ensure that the residuals were well behaved. If residuals are serially correlated and have non
constant error variance, it may indicate that the model is not efficient and as such the parameters
estimated could be biased. The long run results of the mode showed that FCF, REER and RIR
have a long run positive relationship with growth, while MS and OP have a long run negative
relationship with economic growth. All the explanatory variables proved to be statistically
significant in explaining economic growth in South Africa.
6.2 POLICY IMPLICATIONS AND RECOMMENDATIONS
Results in this study have a number of policy implications. This section divides them into
exchange rate policy, investment policy, trade policy and monetary policy.
82 | P a g e
6.2.1 Exchange Rate Policy
There are two main contrasting views regarding the impact of exchange rates on economic
growth. On one hand the traditional view holds that currency depreciation increases economic
growth and on the other hand, the Structuralist view holds that a currency depreciation dampens
growth prospects of a country (Salvatore, 2005). Empirical literature also presents mixed
conclusions regarding this matter. The long run equation in the previous chapter however
suggested that real exchange rates have a negative impact on economic growth. The coefficient
of REER from the long run equation suggested by VECM results implies that a unit rise in South
Africa’s real exchange rates leads to a 0.163 decrease in economic growth in the country. Results
from the short run equation however, shows that a unit rise in South Africa’s real exchange rates
leads to a 3.046 increase in economic growth in the country.
In this regard, for South Africa to increase economic growth, the policy of devaluating the
currency can only work in the short run. Based on the short run relationship, depreciation
increases growth but this can be seen only as a quick fix that has negative consequences in the
long run. In the long run a depreciation/ devaluation can only reduce economic growth; hence
depreciation/devaluation works well in the short run but has negative consequences in the long
run. Based on these finding the policy of depreciation to increase exports and employment in the
economy might not be the best policy for South Africa. In order to avoid misalignments
(overvaluation or undervaluation of the rand) the best policy is to leave the determination of
exchange rates to the forces of demand and supply, were the rand exchange reverts to its own
equilibrium.
6.2.2 Investment policy
Estimation results in this study revealed that investment expenditure has a positive impact on
economic growth in South Africa in the long run, that is, a unit increase in fixed capital
formation leads to a 0.124 increase in economic growth. The implication for policy is that the
government should spend more on investment in order to improve economic growth in the
country. The government should invest in infrastructure; roads, plant and equipment, hospitals.
The government should also invest in human capital through educating its people. The returns to
investment are not immediate as revealed by the short run equation in Chapter 5. In the short run
83 | P a g e
the impact of investment on growth is negative; however the impact is positive and significant in
the long run. The rationale being that in the short run it is not easy to recoup some of the costs
incurred in investments of both physical and human capital, but in the long run the returns to
investment will be high. In order to achieve high levels of economic growth in the long run, the
government has to invest heavily though the returns are not so high in the short run.
6.2.3 Monetary Policy
Monetary policy is the deliberate manipulation of money supply and its price (interest rates) to
achieve desired changes in the economy. Estimation results in this study revealed that money
supply has a negative impact on economic growth in South Africa both in the short run and long
run. From the long run equation, a unit increase in money supply reduces growth by
approximately 0.257. Long run results also suggest that a unit increase in real interest rates
increases growth by approximately 0.0912. In the short run increase in real interest rates reduce
growth by approximately 0.120. The implication for policy is that in the short run an
expansionary monetary policy is effective. This implies that a reduction on a repo rate can induce
investments thereby improving economic growth. However, the effects of a contractionary
monetary policy can be experienced only in the long run.
Furthermore, taking into account the impulse response analysis and variance decomposition
analysis, it can be seen that real interest rates explain much of the variations in economic growth.
Interest rates therefore are a very important tool in influencing economic growth in South Africa.
The policy framework currently being used by the central bank of inflation targeting is relevant
and effective in the current South African economic climate. The government uses repo rate to
control both money supply and inflation. Given the long run relationship and the variance
decomposition analysis, this dissertation recommends that the current monetary policy in South
Africa be maintained.
6.2.4 Trade Policy
Estimation results in this study revealed that in the short run trade openness increases economic
growth in South Africa by approximately 0.415. From the long run equation though, a unit
increase in money supply reduces growth by approximately 0.199. For trade openness to be
successful, there are many factors to be taken into consideration; terms of trade, trade
84 | P a g e
diversification, balance of trade and the nature of goods imported and exported. In order for trade
openness to be sustainable and profitable in the long run, South Africa has to diversify trade. The
over reliance on primary produce such as mining and agriculture products reduces the gains from
trade. There is need to diversify into value addition products which fetch high price at the world
market, for instance, the need to expand the already viable car manufacturing industry and other
value adding industry. South Africa should also improve on the service industry in order to
compete with developed countries.
South Africa should also increase the import reducing measures in order to protect and expand
local industry. This can be done by imposing import tariffs and export subsidies. This increases
domestic demand for local industry and making then stronger in order to compete with
international industries in the long run, thus improving trade balance and ultimately economic
growth.
6.3. LIMITATIONS OF THE STUDY AND AREAS FOR FURTHER RESEARCH
The first limitation was the unavailability of quarterly data for some variables suggested by the
theoretical model regarding the impact of real exchange rates and economic growth. Secondly,
some of the secondary data used in this study was obtained from diversified sources which are
also subject to error, hence absolute reliability of the data is not guaranteed. With regards to the
exchange rate, not only does misalignment (overvaluation or undervaluation) affect the economy
but also the volatility thereof. The rand has been volatile since the adoption of the free floating
system. A recommended area for further research is the impact of exchange rate volatility on
economic growth in South Africa.
6.4 SUMMARY
The main purpose of this study was to examine the impact of real exchange rates on economic
growth in South Africa for the period between 1994 and 2010. The hypothesis of this study was
that real exchange rates have a significant impact on economic growth in South Africa. Given the
regression results, the null hypothesis that the real exchange rates have a significant impact on
economic growth in South Africa was not rejected.
85 | P a g e
REFERENCES
ACAR, M. 2000. Devaluation in Developing Countries: Expansionary or Contractionary?
Journal of Economic and Social Research, 2 (1).
AGOSIN, M. 1999. Trade and Growth in Chile. Cepal Review, 68.
AGUIRRE, A. AND CALDERON, C. 2005. Real Exchange Rate Misalignments and
Economic Performance. Central Bank of Chile, Working Paper 315.
AKPAN, P.L. 2008. Foreign Exchange Market and Economic Growth in an Emerging
Petroleum Based Economy: Evidence from Nigeria (1970-2003). African Economic and
Business Review, 6 (2).
ANDREN, T. 2007. Econometrics. United Kingdom: Thomas Andren & Ventus publishing
ApS.
ARAUJO, R.A. AND SOARES, C. 2011. Export Led Growth' x `Growth Led Exports':
What Matters for the Brazilian Growth Experience after Trade Liberalization? MPRA
30562(2).
ARON, J., ELBADAWI, I. AND KAHN, B. 1997. “Determinants of the Real Exchange Rate
in South Africa”, Centre for the study of African Economics, WPS/97-16, Oxford: CSEA
publishing.
BERNANKE, B.S. 2005. The Global Saving Glut and the US Current Account Deficit.
Remarks at the Sandridge Lecture, Virginia Association of Economics, Richmond, Virginia
(March).
BERRY, A. 2006. Employment and Income Distribution Experiences of Minerals Exporters
and of Countries Achieving Growth Acceleration. Human Sciences Research Council.
BLUNGMART, M. 2000. The Augmented Dickey-Fuller test. Journal of Econometrics.
BOND, S., HOEFFLER, A. AND TEMPLE, J. 2001.GMM Estimation of Empirical Growth
Models. Economics Papers 2001-W21.
86 | P a g e
BOWERMAN, B.L. AND O'CONNELL, R.T. (1979). Time Series and Forecasting, Duxbury
Press, North Scituate, Massachusetts.
BROOKS, C. 2008. Introductory Econometrics for Finance. Cambridge: Cambridge
University Press.
CAVES, R.E., FRANKEL, J.A. AND JONES, R.W. 1996. World Trade and Payment.7th
edition, New York: Harper Collins College Publishers.
CHALLIS, R.E. AND KITNEY, R.I. (1991). Biomedical Signal Processing. Medical &
Biological Engineering & Computing, 28(6).
CHEN, J. 2012. Real Exchange Rate and Economic Growth: Evidence from Chinese
Provincial Data (1992-2008). Paris School of Economics, France, Working paper 2012-05.
CLEMENTE, J.A., MONTANES, A. AND REYES, M. 1998. Testing for a Unit Root in
Variables with a Double Change in the Mean. Economics Letters, 59.
DABLA-NORRIS, E. AND FLOERKEMEIER, H. 2006. Transmission Mechanisms of
Monetary Policy in Armenia: Evidence from VAR Analysis. IMF Working Paper.
DAVIDSON, R. (2007). Monetary Policy in South Africa. [Online]. Available: Http:
//www.bis. org /review/rg.pdf [Accessed 13 February 2011].
DOLADO, J.J., GONZALO, J. AND MARMOL, F. 1999. Cointegration. University of Carlos
III, Madrid, Working Paper.
DOMAC, I. AND SHABSIGH, G. 1999. Real Exchange Rate Behaviour and Economic
Growth: Evidences from Egypt, Jordan, Morocco and Tunisia. IMF Working Paper,
WP/99/40.
DOOLEY, M.P., FOLKERTS-LANDAU D. AND GARBER, P. 2004.An Essay on the Revived
Bretton Woods System. NBER Working Paper 9971.
DRINE, I. AND RAULT, C. 2003. Do Panel Data Permit to Rescue the Balassa-Samuelson
Hypothesis for Latin American Countries? Applied Economics, 35 (3).
87 | P a g e
Department of Trade and Industry (DTI). 2012. Research Statistics. [Online], Available:
http//tradestats.thedti.gov.za/ReportFolders.aspx/sCS. [Accessed 14 August 2012].
DU PLESSIS, S. AND SMIT, B. 2007. South Africa’s Growth Revival after 1994.
Stellenbosch Economic Working Papers 01/06.
DU TOIT, C. AND MOOLMAN, E. 2007. Neoclassical Investments Function of the South
African Economy. Economic Modelling, 21(4).
EDWARDS, L. AND GARLICK, R. 2007. Trade Flows and the Exchange Rate in South
Africa. Trade and Policy Strategies. Working Paper.
EDWARDS, L., MLANGENI, T. AND VAN SEVENTER, D. 2000. Revealed Comparative
Advantage in SADC economies. Southern African Update, 5.
EDWARDS, L. AND LAWRENCE, R. 2006. South African Trade Policy Matters: Trade
Performance & Trade Policy. Centre for International Development at Harvard University.
Working Paper 135.
EICHENGREEN, B. 2008. The Real Exchange Rate and Economic Growth. World Bank
PREM Network, Commission on Growth and development. Working Paper 4.
ENGEL, C. 2009. Currency Misalignments and Optimal Monetary Policy: A
Reexamination. NBER Working Paper Series, 14829.
FAULKNER, D. AND LOEWALD, C. 2008. Policy Change and Economic Growth: A Case
Study of South Africa. Commission on Growth and Development. Working Paper 41.
FLASSBECK, H. 2004. Exchange Rate Management in Developing Countries: The Need for
a Multilateral Solution. Based on a speech delivered at Workshop on "New Issues in Regional
Monetary Coordination: Understanding North-South and South-South Arrangements “in
Hamburg, 7-9 July 2004.
FLATTERS, F. AND STERN, M. 2007. Trade and Trade Policy in South Africa: Recent
Trends and Prospects. Development Network Africa.
88 | P a g e
FREDRICK, C. AND FOURIE, N. 2009. How to Reason and Think in Macro Economics. 3rd
Edition. South Africa: Philips Burger and Juta & Co.
GARNAUT, R., GRILLI, R.E. AND RIEDEL, J. 1995. Sustaining Export Oriented
Development: Ideas from East Asia. Cambridge: Cambridge University Press.
GHALI, K.H. AND AL-MUTAWA, A. 1999. The Intertemporal Causal Dynamics between
Fixed Capital Formation and Economic Growth in the Group-of-Seven Countries.
International economic journal, 13(2).
GRANGER, C. AND NEWBOLD, P. 1974, Spurious Regression in Econometrics. Journal of
Econometrics, 2.
GREENE, W. 2000. Econometric Analysis, 4th
ed. Upper Saddle River, N.J: Prentice Hall.
GREENE, W. 2007. Econometric Analysis, 6th
ed. Upper Saddle River, N.J: Prentice Hall.
GUJARATI, D. N. 2004. Basic Econometrics. 4th
ed. New York: McGraw Hill.
GUJARATI, D.N. AND PORTER, C.D. 2010. Essentials of Econometrics 4th
ed. New York:
McGraw-Hill International Edition.
HAUSMANN, R. 2008. Final Recommendations of the International Panel on ASGISA.
Center for International Development. Working Paper 161.
HAVRYLCHYK, O. 2010. A Macroeconomic Credit Risk Model for Stress Testing the
South African banking Sector. South African Reserve Bank. Working Paper 3(10).
HSIEH, D.A. 1982. The Determination of the Real Exchange Rate: the Productivity
Approach. Journal of International Economics, 12.
ITO, T. AND KRUEGER, A.O. 1999. Economic Growth and Real Exchange Rate: An
Overview of the Balassa-Samuelson Hypothesis in Asia. Theory, Practice, and Policy Issues,
NBER-EASE7.
JAMES, E.W., NAYA, S. AND MEIER, M.G. 1989. Asian Development: Economic and
Policy Lessons. University of Wiscons in press.
89 | P a g e
JAUSSAUD, J. AND REY, S. 2009. Long-Run Determinants of Japanese Exports to China
and the United States: A Sectoral Analysis. University of Pau, France. Working Paper.
JENKINS, C., BLEANEY, M., HOLDEN, M. AND SIWISA, N. 1997. A Review of South
Africa’s Trade Policy. A paper presented at the Trade and Industrial Policy Annual Forum,
Muldersdrift, September 1997.
JOHANSEN, S. 1991. Estimation and Hypothesis testing of Cointegration Vectors in
Gaussianvector Autoregressive Models. Econometrica, 59.
JOHANSEN, S. 1995. Likelihood-based Inference in Cointegrated Vector Autoregressive
Models. Oxford: Oxford University Press.
JOHANSEN, S. AND JUSELIUS. K. 1990. The Full Information Maximum Likelihood
Procedure for Inference on Cointegration-with Applications to the Demand for Money.
Oxford Bulletin of Economics and Statistics, 52.
JORDAAN, J AND HARMSE, C. 2001. Risk and Speculation in the South African
Economy. University of Pretoria. Working Paper.
KAHN, B. AND FARRELL, G. 2002. South African Real Interest Rates in Comparative
Perspective: Theory and Evidence. South African Reserve Bank. Occasional Paper 17.
KALYONCU, H., ARTAN, S., TEZEKICI, S. AND OZTURK, I. 2008. Currency Devaluation
and Output Growth: An Empirical Evidence from OECD Countries. International Research
Journal of Finance and Economics, Issue 12.
KRUGMAN, P. AND TAYLOR, L. 1978. Contractionary Effects of Devaluation. Journal of
International Economics, 8.
KUBO, A. 2011. Trade and Economic Growth: Is Export-led Growth Passé? Economics
Bulletin, 31(2).
LEVINE, R. AND D. RENELT. 1992. A Sensitivity Analysis of Cross-Country Regressions.
The American Economic Review, 82(4).
90 | P a g e
LEVY-YEYATI, E. AND STURZENEGGER, F. 2007. The Fear of Appreciation. Kennedy
School of Governance. Working Paper 07-047.
LUTKEPOHL, H. 1993. Introduction to Multiple Time Series Analysis. Berlin: Springer.
MAHADEVA, L. AND ROBINSON, P. 2004. Unit Root Testing to Help Model Building.
Handbooks in Central banking studies, 22.
MARSTON, R. C. 1987. Real Exchange Rates and Productivity Growth in the United States
and Japan. NBER Working Paper 1922.
MASUNDA, S. 2011. Real Exchange Rate Misalignment and Sectoral Output in
Zimbabwe. Midlands State University, Zimbabwe. Working Paper.
MBAYE, S. 2012. Real Exchange Rate Undervaluation and Growth: Is there a Total Factor
Productivity Growth Channel? CERDI Clermont Ferrand. Working Paper E 2012.11.
McDonald, R. 2000. The Role of the Exchange Rate in Economic Growth: a Euro-zone
Perspective. National Bank of Belgium. Working Paper 09.
MCPHERSON, M.F. AND RAKOVSKI, T. 2000. Exchange Rates and Economic Growth in
Kenya: An Econometric Analysis. African Economic Policy: Discussion Paper 56.
MEDINA-SMITH, E. 2001. ‘Is the Export-led Growth Hypothesis Valid for Developing
Countries? A Case Study of Costa Rica. UNCTAD/ITCD/TAB/8, UNCTAD, Geneva. Paper
7.
MISHI, S. 2011. Real Exchange Rate Misalignment and Economic Growth: Empirical
Evidence from South Africa. Nedbank budget speech competition South Africa [Online],
Available: http://www.budgetspeechcompetition.co.za [Accessed 23 September 2011].
MNYANDE, M. 2010. Interest Rates in South Africa. South Africa Reserve: Working Paper
2010-06.
MONTIEL, P.J. AND SERVEN, L. 2008. Real Exchange Rates, Savings and Growth: Is
there a link? The World Bank, Development Research Group, Macroeconomic and Growth
Team. Policy Research Working Paper WPS4636.
91 | P a g e
MTONGA, E. 2011. Did it matter? Monetary Policy Regime Change and Exchange Rate
Dynamics in South Africa. Lusaka, Working Paper.
MUNTHALI, T., SIMWAKA, K. AND MWALE, M. 2010. The Real Exchange Rate and
Growth in Malawi: Exploring the Transmission Route. Journal of Development and
Agricultural Economics, 2(8).
NATTRASS, N., WAKEFORD, J. AND MURADZIKWA, S. 2002. Macroeconomics Theory
and Policy in South Africa. 3rd
Revised Edition. South Africa: David Philip Publishers.
NCHIMUNYA, H. 2011. Does a Weaker Exchange Rate Make Sense? Nedbank budget
speech competition South Africa [Online], Available:
http://www.budgetspeechcompetition.co.za [Accessed 23 September 2011].
NDLELA, T. 2011. Implications of Real Exchange Rate Misalignment in Developing
Countries: Theory, Empirical Evidence and Application to Growth Performance in
Zimbabwe. Monash University Department of Economics, Australia. MPRA Working Paper
32710.
NGANDU, S. AND GEBRESELASIE, T. 2006. When Might an Exchange Rate Depreciation
be Growth Inducing or Contractionary? Human Sciences Research Council. Working Paper.
OGUNMUYIWA, M.S. AND EKON, A.F. 2010. Money Supply-Economic Growth Nexus in
Nigeria. Journal of Social Science, 22(3).
OLD MUTUAL, 2000. Strong Rand Poses Risks Amid Weak Economy. Media release, South
Africa.
PALLEY, T.I. 2011. The Rise and Fall of Export-led Growth. Levy Economics Institute.
Working Paper 675.
PRINSLOO, L. 2011. Tough Decisions Could Revive SA’s Manufacturing Sector in Months.
South Africa. [Online], Available: http://www.polity.org.za/article/tough-decisions-could-revive-
sas-manufacturing-sector-in-months.[Accessed 22 May 2011].
92 | P a g e
RAD, F.M. 2012. The Effects of Money Supply on Economic Growth in Iran. Department of
Human Sciences, Islamic Azad University Mashhad Branch, Mashhad, Iran. Working Paper.
RADDATZ, C. 2008. Exchange Rate Volatility and Trade in South Africa. The World Bank.
Working Paper.
SCHWEICKER, R., THIELE, R AND WIEBELT, M. 2006. Macroeconomics and
Distributional Effects of Devaluation in a Dollarized Economy: A CGE Analysis for
Bolivia. Keiler Arbeitspapiere. Working Paper 1255.
RAZIN, O. AND COLLINS, M. 1997. Real Exchange Rate Misalignments and Growth.
NBER Working Paper 6174.
RODRIK, D. 2007. The Real Exchange Rate and Economic Growth: Theory and Evidence.
Kennedy School of Government Harvard University Cambridge. Working Paper.
RODRIK, D. 2008. The Real Exchange Rate and Economic Growth. Kennedy School of
Government Harvard University Cambridge. Working Paper MA 02138.
RUSIKE, G.T. 2007. Trends and determinants of inward foreign direct investment to South
Africa. Unpublished Master’s Thesis. Rhodes University.
SALVATORE, D. 2005. The Euro-dollar Rate Defies Prediction African. Journal of Policy
Modeling, 27(4).
SARB.2012.HistoricMacroeconomicInformation[Online],Available:http://www.resbank.co.za/
Research/Statistics/Pages/OnlineDownloadFacility.aspx. [Accessed 14 August 2012].
SARB.2012.ExchangeRatePolicy.[Onlineavailableathttp://www2.resbank.co.za/internet/Glossar
y.nsf/0/6e77f482c063ea5742256b430031f732?OpenDocument [accessed on 17-06-2012]
SHELILE, T. 2006. The Term Structure of Interest Rates and Economic Activity in South
Africa. Unpublished Master’s Thesis. Rhodes University.
SOLANES. G.J AND FLORES, F.T. 2009. The Balassa–Samuelson Hypothesis in Developed
Countries and Emerging Market Economies: Different Outcomes Explained. Economics:
The Open-Access, Open-Assessment E-Journal.
93 | P a g e
SQUALLI, J. AND WILSON, K. 2006. A New Approach to Measuring Trade Openness.
Zayed University, Economic & Policy Research Unit, Dubai. Working Paper.
TAKAENDESA, P. 2006. The Behaviour and Fundamental Determinants of the Real
Exchange Rate in South Africa. Unpublished Master’s Thesis. Rhodes University.
TARAWALIE, B, 2010. Real Exchange Rate Behaviour and Economic Growth: Evidence
from Sierra Leone. SAJEM 13(1).
TODANI, R.K. AND MUNYAMA, T. 2005. Exchange Rate Volatility and Exports in South
Africa. [Online]. Available http://www.tips.org.za/files/733.pdf.[Accessed on 25 January, 2007].
TREASURY. 2012. The New Growth Path: The Framework. [Online], Available:
http://www.info.gov.za/view/DownloadFileAction?id=135748. [Accessed 01-02-2012].
VAN DER MERWE, E.J. 1996. Exchange Rate Management Policies in South Africa:
Recent Experience and Prospects. South African Reserve Bank. Occasional Paper.
VIEIRA, F. AND MACDONALD, R. 2010. A panel Data Investigation of Real Exchange
Rate Misalignment and Growth. CESifo Working Paper 3061.
WALTERS, S. AND DE BEER, B. 1999. An Indicator of South Africa’s External
Competitiveness. South African Reserve Bank Quarterly Bulletin, September 1999.
WHITE, H. 1980. A Heteroscedasticity-consistent Covariance Matrix and a Direct Test for
Heteroscedasticity. Econometrica, 1.
WHO, 2012. Exports and imports of major commodity groups by region and selected
economy,19802011.[Online],Available:http://stat.wto.org/StatisticalProgram/WSDBStatProgra
mSeries.aspx?Language=E [Accessed 25-08-2012].
WILLIAMSON J., 2004, A Short History of the Washington Consensus. A Paper
commissioned by Foundation CIDOB for a conference “From the Washington Consensus
towards a new Global Governance,” Barcelona, September 24–25, 2004.
94 | P a g e
APPENDICES
APPENDIX 1
South African data used in regression
obs LOG_GDP LOG_REER LOG_MS LOG_FCF LOG_RIR LOG_OP
1994Q1 12.47520 4.832226 12.29480 10.40363 1.704748 -0.798508
1994Q2 12.52968 4.781641 14.62542 10.44987 2.098018 -0.867501
1994Q3 12.54093 4.778703 12.34555 10.45487 1.938742 -0.776529
1994Q4 12.55166 4.791484 12.38893 10.53550 1.864080 -0.820981
1995Q1 12.51567 4.784989 12.40652 10.49902 2.001480 -0.693147
1995Q2 12.53979 4.747798 12.46915 10.58489 1.916923 -0.755023
1995Q3 12.57814 4.786825 12.49194 10.56452 2.388763 -0.673345
1995Q4 12.58667 4.804922 12.52315 10.60321 2.484907 -0.713350
1996Q1 12.55339 4.802955 12.57841 10.59816 2.493205 -0.733969
1996Q2 12.59892 4.702478 12.62973 10.65561 2.639057 -0.713350
1996Q3 12.61407 4.683334 12.66674 10.65658 2.468100 -0.597837
1996Q4 12.62278 4.674790 12.69775 10.68730 2.356126 -0.693147
1997Q1 12.58733 4.769413 12.73937 10.66305 2.307573 -0.713350
1997Q2 12.63122 4.800819 12.77530 10.68410 2.336987 -0.693147
1997Q3 12.63723 4.790986 12.81435 10.70219 2.411439 -0.579818
1997Q4 12.63841 4.762088 12.86115 10.77046 2.505526 -0.616186
1998Q1 12.60007 4.772547 12.89706 10.73394 2.553344 -0.616186
95 | P a g e
1998Q2 12.63260 4.741797 12.92860 10.72457 2.781920 -0.634878
1998Q3 12.63873 4.577285 12.95107 10.75323 2.836150 -0.597837
1998Q4 12.64374 4.625659 12.97681 10.79602 2.685805 -0.673345
1999Q1 12.61048 4.589549 12.98145 10.70255 2.525729 -0.673345
1999Q2 12.65114 4.616209 13.00248 10.64833 2.415914 -0.733969
1999Q3 12.66615 4.627421 13.02571 10.65408 2.617396 -0.653926
1999Q4 12.67983 4.624384 13.07333 10.68732 2.610070 -0.673345
2000Q1 12.64538 4.635602 13.08273 10.67921 2.459589 -0.653926
2000Q2 12.68437 4.599454 13.08748 10.69229 2.261763 -0.693147
2000Q3 12.71686 4.611550 13.09920 10.71604 2.054124 -0.673345
2000Q4 12.72326 4.572957 13.13933 10.75577 2.014903 -0.653926
2001Q1 12.68239 4.542869 13.18472 10.73494 1.960095 -0.653926
2001Q2 12.72092 4.571407 13.21424 10.72709 1.994700 -0.653926
2001Q3 12.73219 4.543508 13.25837 10.73833 2.145931 -0.693147
2001Q4 12.74319 4.393090 13.29079 10.75677 2.202765 -0.693147
2002Q1 12.71682 4.339250 13.37015 10.73846 2.174752 -0.693147
2002Q2 12.75767 4.424966 13.40584 10.74188 2.116256 -0.673345
2002Q3 12.76682 4.400480 13.42669 10.77973 1.887070 -0.693147
2002Q4 12.78136 4.483341 13.46874 10.83177 1.435085 -0.673345
2003Q1 12.74838 4.567364 13.50126 10.83335 1.840550 -0.673345
96 | P a g e
2003Q2 12.78925 4.612642 13.54275 10.82919 2.041220 -0.693147
2003Q3 12.79642 4.661834 13.55996 10.88001 2.230014 -0.653926
2003Q4 12.80507 4.700935 13.59085 10.93809 2.393339 -0.653926
2004Q1 12.78517 4.651863 13.63323 10.95415 2.429218 -0.634878
2004Q2 12.82588 4.702115 13.65336 10.94803 2.402430 -0.634878
2004Q3 12.84541 4.723486 13.68317 11.00353 2.272126 -0.634878
2004Q4 12.86018 4.727211 13.72388 11.05930 2.066863 -0.597837
2005Q1 12.83821 4.744671 13.75193 11.05689 2.091864 -0.616186
2005Q2 12.87647 4.705739 13.80274 11.06617 1.987874 -0.579818
2005Q3 12.89852 4.693364 13.85476 11.10472 1.902108 -0.562119
2005Q4 12.90925 4.727565 13.88918 11.15512 1.931521 -0.597837
2006Q1 12.88778 4.775082 13.95928 11.16790 1.902108 -0.597837
2006Q2 12.92464 4.712139 14.01597 11.18115 1.945910 -0.544727
2006Q3 12.94966 4.634049 14.05093 11.22113 1.824549 -0.527633
2006Q4 12.97775 4.632299 14.10366 11.27097 1.909543 -0.462035
2007Q1 12.95225 4.638218 14.15536 11.31467 1.832581 -0.494296
2007Q2 12.97853 4.659564 14.22126 11.32100 1.791759 -0.510826
2007Q3 12.99951 4.649952 14.27464 11.34143 1.871802 -0.510826
2007Q4 13.02623 4.669646 14.31353 11.38848 1.766442 -0.510826
2008Q1 12.98925 4.544783 14.35226 11.42248 1.470176 -0.544727
97 | P a g e
2008Q2 13.02782 4.541591 14.40681 11.43901 1.266948 -0.510826
2008Q3 13.03822 4.606270 14.43337 11.48084 0.615186 -0.494296
2008Q4 13.04399 4.480627 14.46518 11.52071 1.360977 -0.562119
2009Q1 12.98045 4.498587 14.46905 11.48640 1.629241 -0.673345
2009Q2 13.00055 4.624973 14.47859 11.45006 1.308333 -0.755023
2009Q3 13.01807 4.667112 14.48272 11.40435 1.410987 -0.733969
2009Q4 13.03823 4.676374 14.48172 11.39141 1.504077 -0.713350
2010Q1 13.00234 4.691256 14.47796 11.41816 1.532557 -0.713350
2010Q2 13.03210 4.733827 14.49663 11.40654 1.704748 -0.693147
2010Q3 13.04874 4.752296 14.52569 11.41242 1.845300 -0.634878
2010Q4 13.06781 4.760891 14.54786 11.43286 1.735189 -0.653926
98 | P a g e
APPENDIX 2
EDITOR’S DECLARATION
I Dr Ketiwe Ndhlovu (Department of English) confirm that l edited Kin’s Sibanda’s
Masters thesis entitled The lmpact of Real Exchange Rates on Economic Growth: A
case study of South Africa. During the process of editing, the following changes were
recommended; grammatical, spelling, sentence construction and table/figure numbering
among other things. It is up to the candidate to effect these changes as he remains the
author of this document.
...................................................... ........................................................
Editor’s Signature Date
..................................................... .......................................................
Candidate’s Signature Date