Long–and Short–Run Determinants of Demand for Money and its Stability in the Gambia: An Empirical Investigation Kebba Jammeh Research Paper submitted in fulfillment of the Honours Requirements for the Bachelor of Science in Economics 1 Department of Economics and Management Sciences University of the Gambia March, 2012 2 Approval I certify that I have supervised this research project and I am of the opinion that it meets the acceptable standards of scholarly work and is adequate in quantity and quality as a research project for the fulfillment of Honors requirements for the Bachelor of Science in Economics. ………………………………………………. Prof. Jan Sture Gunnarsson (Supervisor) This research project is submitted to the Department of Economics and Management Sciences and is accepted as the fulfillment of the Honours Requirements for the Bachelor of Science in Economics. …………………………………………. Mohammed Jammeh Ag. Head of Department of Economics and Management Sciences 3 Dedication This thesis is dedicated to my sister, brothers and our parents. 4 Acknowledgments I would like to register my sincere appreciation to my supervisor, Prof. Jan Sture Gunnarsson, for his wholehearted support and time during the course of writing this Thesis. I am also greatly indebted to Mr. Yaya S. Jallow whose encouragement and expert advice helped pilot this paper. Under his tutelage, I had also been inspired towards contributing to the frontiers of knowledge. I would also like to record my sincere thanks to Dr. Momodou Mustapha Fanneh for his intellectual and material contributions which supplemented the necessary ingredients to making this thesis a product of extensive research. I would specially like to express my profound gratitude to Mr. Bai Madi Ceesay, Mr. Lamin Camara, Mr. Alhagie Fadera, Mr. Abdou Kadrie Ceesay and Madam Amie Kolleh Jeng for their understanding and constructive suggestions after reading the first draft of this paper. My thanks go to Mr. Hassoum Muhammed Loum whose brotherly advice and words of encouragement made me realized that ‘when the going gets tough, the tough get going’. Very importantly, my special acknowledgement goes to Mrs. Amie Tunkara, Ms. Saffiatou Sanneh and Mr. Samba Bah for their support. I would also convey my heartfelt gratitude to Ms. Fatou Sillah who has helped me a lot in the search for up-to-date and relevant information. I also gratefully acknowledge the authors of all the articles that provided me a wealth of useful information and the reliable scientific basis for this paper. Above all, I am grateful to the Almighty God for His grace, wisdom and knowledge He granted onto me and mostly importantly for making me know these great people. Words are indeed inadequate for me to express the enormousity of my oceanic gratitude to all those who have contributed towards the successful completion of this thesis. May God bless and strengthen you all. 5 Abstract The objective of this paper is to empirically examine the long–and short-run determinants of demand for money and its stability in the Gambia using quarterly time series data from 1993:I to 2008:IV. The Johansen cointegration test shows the existence of a long-run equilibrium relationship between money demand, income, interest rate, inflation rate and exchange rate. However, the short-run dynamic model did not confirm the existence of this relationship in the short-run. The estimated long-run money demand function shows that money demand in the long-run has a positive relationship with income and exchange rate, and a negative relationship with interest rate and inflation rate. The estimated long-run income elasticity of money demand is greater than one, which implies that money can be considered a luxury in the Gambia (Valadkhani, 2008). Furthermore, the ECM model shows that although these variables are important determinants of money demand in the Gambia in the long-run, they are not significant drivers of money demand in the short-run. Hence, the transactional, precautionary and speculative motives of money demand hold in the Gambia only in the long-run. Moreover, the asset-substitution phenomenon also holds only in the long-run while the currencysubstitution phenomenon does not exist in the Gambia. The Impulse Response Function (IRF) shows that a percentage increase in real income increases money demand by 2.4 percent after one year and die down to 2.1 percent in the second year, while a percentage shock in interest rate reduces money demand by 0.023 percent in the first year and 0.043 percent in the second year. The error correction factor shows that about 2.9 percent of the disequilibrium in real money demand is corrected by short-run adjustment within a quarter. Furthermore, the ECM stability test shows that inflation is not purely a monetary phenomenon in the Gambia because money demand is not stable during the period under study. Thus, monetary authorities in the Gambia should be flexible in the use of broad money supply as an intermediate target because money supply targeting may not translate to changes in interest rates due to the instability of money demand. An alternative policy option of targeting interest rates rather than money supply may improve the outcome of monetary policy in the Gambia. 6 Table of contents Approval.................................................................................................................................................. 3 Dedication ............................................................................................................................................... 4 Acknowledgments ................................................................................................................................... 5 Abstract .................................................................................................................................................. 6 List of Figures…………………………………………………………………………………….………………………….…………..9 List of Tables……………………………………………………………………………………….……………………………………10 CHAPTER 1 ............................................................................................................................................ 11 Introduction .......................................................................................................................................... 11 1.1 Background of the study .................................................................................................................. 11 1.2 Problem Statement.......................................................................................................................... 12 1.3 Research Objective .......................................................................................................................... 13 1.4 Significance of the Study .................................................................................................................. 14 1.5 Research Hypotheses....................................................................................................................... 15 1.6 Scope of the Study ........................................................................................................................... 15 CHAPTER 2 ............................................................................................................................................ 16 Literature Review .................................................................................................................................. 16 2.0 Introduction .................................................................................................................................... 16 2.1 Theoretical Framework .................................................................................................................... 16 2.1.1 Quantity Theory of Money Demand........................................................................................ 17 2.1.2 Liquidity Preference Theory of Money Demand ...................................................................... 18 2.1.3 Portfolio Theory of Money Demand........................................................................................ 20 2.1.4 Inventory Theory of Money Demand…….…………………………………………………………………………….….21 2.2 Empirical Evidence ........................................................................................................................... 23 CHAPTER 3 ............................................................................................................................................ 26 Financial Sector Development and Trend Analysis of Board Money Demand in the Gambia ................... 26 3.0 Introduction .................................................................................................................................... 26 3.1 Financial Sector Development.......................................................................................................... 26 3.2 Trend Analysis of Broad Money Demand in the Gambia…………………………………………………….28 7 CHAPTER 4 ............................................................................................................................................ 30 Methodology......................................................................................................................................... 30 4.0 Introduction .................................................................................................................................... 30 4.1 Model specification and Definition of Variables ............................................................................... 30 4.2 Dataset and its Source ..................................................................................................................... 31 4.3 Pre-estimation tests: Unit Root and Cointegration Tests .................................................................. 32 4.4 Estimation Technique: Error Correction Model………………………………………………………………..………………34 4.5 Post-estimation techniques: Impulse Response, Variance Decomposition and Stability Tests ........... 35 4.6 Expected Results…………………………………………………..……………………………………………………………36 CHAPTER 5 ............................................................................................................................................ 38 Analysis and Interpretation of Results ................................................................................................... 38 5.0 Introduction .................................................................................................................................... 38 5.1 Unit Root and Johansen Cointegration Tests Results ........................................................................ 38 5.2 Estimated Long-run Money Demand Function ................................................................................. 40 5.3 Estimated Short-run Money Demand Function…………..…………………………….……………………………….......42 5.4 Impulse Response and Variance Decomposition Tests Results………………………………………………..……….43 5.5 ECM Stability Test Result…………………………………………………………………….……………………………………….….45 CHAPTER 6 ............................................................................................................................................ 47 Conclusion and Recommendations ........................................................................................................ 47 6.0 Introduction……………………………………………………………………………………………………………………….……………47 6.1 Conclusion…………………………………….……………………………………………………………………………………47 6.2 Recommendations……………………………………………………………………………………………………………..49 References ............................................................................................................................................ 51 Appendix ............................................................................................................................................... 55 Declaration............................................................................................................................................ 64 8 LIST OF FIGURES Figure A1: Annual Growth Rate and Share in GDP of the Financial Sector in the Gambia…………27 Figure A2: Trend of Broad Money Demand In the Gambia from 1990 to 2010…………………………28 Figure A3: The Graph of Log Real GDP at Level…………………………………………………………………………60 Figure A4: Graph of Interest Rate at Level………………………………………………………………………………..60 Figure A5: Graph of Log Real Broad Money Balance at Level……………………………………………………61 Figure A6: Graph of Inflation Rate at Level……………………………………………………………………………….61 Figure A7: Graph of Log Exchange Rate at level……………………………………………………………………….62 Figure B1: Graph of Cumulative IRF for ECM model…………………………………………………………………63 * Impulse factor= Real income and Response factor= Real money demand Figure B2: Graph of Cumulative IRF for ECM model…………………………………………………………………63 * Impulse factor= Interest rate and Response factor= Real money demand 9 LIST OF TABLES Table B1: Summary Statistics of the variables in the model……………………………………………………..55 Table B2: Correlation Matrix of the variables in the model………………………………………………………55 Table B3: Augmented Dickey Fuller Test Results………………………………………………………………………55 Table B4: Results from Johansen Cointegration Test………………………………………………………………..56 Table B5: Results from the estimated Long-run money demand function and the Ramsey RESET Test ……………………………………………………………..…………………………………………………………………..……..56 Table B6: Results of the Lag Selection-order criteria………………………………………………………………..56 Table B7: Regression results of the error-correction model of money demand equation..……….57 Table B8: Result of the cumulative Impulse Response Function and FEVD……….………………………57 *Impulse factor= Real income and Response factor= Real money demand Table B9: Result of the Cumulative Impulse Response Function and FEVD……………………………….58 *Impulse factor= Interest rate and Response factor= Real money demand Table B10: Result of Lagrange-multiplier test……………………………….……………………………….…………58 Table B11: Result of Jarque-Bera test……………………………….……………………………….…………………….58 Table B12: Result of Eigenvalue stability condition……………………………….………………………………….59 10 Chapter 1 INTRODUCTION 1.1 Background of the study Studies on the demand for money and its stability has long remained in the domain of rigorous investigation dating back to the period when it helped in eliminating the problems associated with the barter system. The demand for money plays a central and very crucial role in macroeconomic analysis as it serves as the core link between the monetary and the real sector of an economy. Goldfeld (1994) noted that the relationship between the demand for money and its main drivers is an important building block in macroeconomic theory and is a crucial component in the conduct of monetary policy. Theories on the determinants of money demand and its stability has attracted debates in the academic circle dating back to the period when Irving Fisher implicitly put forward the earliest theory of money demand in his quantity theory of money. The stability of the money demand function also has crucial implications on the way the central bank carries out its monetary policy under a monetary targeting framework. A stable money demand function is prerequisite for any policy-driven change in monetary variables to have predictable effect on output, interest rate and ultimately prices through the transmission mechanism of monetary policy. A stable money demand function also provides empirical evidence that money supply targeting is an effective monetary policy option to controlling inflation and effective demand. 11 During the last decade, the Central Bank of the Gambia (CBG) has been using various monetary policy instruments to control inflation and effective demand. The CBG had been following an interest rate targeting regime by using its Bank Rate as a monetary instrument to control inflation and effective demand. The CBG reduced the Bank Rate from 29% in 2003 to 9% in 2007, but was raised to 10% at the end of 2007 to check effective demand and inflationary pressures on the economy. The Bank Rate has since then remained at 10%. However, with the introduction of the Monetary Policy Committee (MPC) Policy Rate, the Bank Rate has become ineffective and non-operational. The MPC then started using the CBG rediscount Rate as its policy instrument. The CBG raised its rediscount rate by one percentage point from 14% to 15% in June 2007, and further to 16% in October 2008 to counter emerging inflationary pressures on the economy. The rediscount rate since then remain unchanged until December 2009 when the MPC reduced it by two percentage points to 14% due to the declining trend of inflation rates. Currently, the CBG follows a monetary regime using broad money – defined as M21– as the intermediate target and reserve money as the operating target in order to achieve its policy objectives which are price stability and real economic growth. The CBG uses control instruments such as the statutory minimum reserve requirement and open market operations to regulate the quantity of money in the economy. 1.2 Problem Statement Controlling inflation to a single digit has long been a policy objective for the Gambian government. The Government of the Gambia through the Central Bank of the Gambia does 1 M2 is the sum of M1 and Quasi money. While M1 is the sum of currency in circulation outside banks and demand deposits, Quasi money is the sum of saving deposits and time deposits. 12 alter monetary aggregates to achieve this policy objective. The specification of an appropriate money demand function is vital in determining the optimum quantity of money to be supplied in an economy. Effective monetary policy implicitly assumes a stable money demand function. However, if the relationship is not stable, money supply targeting might not be an effective policy option for controlling inflation. Thus, the problem of this thesis is to explain the long–and short-run determinants of money demand and its stability in the Gambia because understanding the main drivers of money demand is crucial in understanding the transmission mechanism of monetary policy in the Gambia. 1.3 Research Objectives The main objective of this thesis is to examine the long–and short-run determinants of broad money demand and its stability in the Gambia from 1993 to 2008. This is because so far there is little that is known on the stability and determinants of demand for broad money in The Gambia. Thus, the fundamental objectives of this paper, among other things, are: A. To provide an empirical investigation on both the long–and short-run determinants of broad money demand in the Gambia. This will help to provide answers to vital macroeconomic questions such as; do changes in real income affect the demand for money in the Gambia? If yes, what is its magnitude? How does money demand in the Gambia response to changes in the opportunity cost of holding it? Does the assetsubstitution phenomenon holds in the Gambia? Moreover, how does domestic money demand adjust when the Dalasi depreciates? Does the currency-substitution phenomenon or the wealth-effect hypothesis hold in the Gambia? 13 B. To estimate how money demand responds to changes in interest rate and income, since the sensitivity of money demand in the Gambia to these two variables determines the slope of the Gambian LM Curve C. To evaluate the stability of broad money demand in the Gambia during the period under study. This is because the recession in 2003 was a potential source of money demand instability as inflation was as high as 20 percent accompanied with a sharp depreciation of the Dalasi by 100 percent within a year. D. To provide recommendations that would help improve monetary management in the country. 1.4 Significance of the Study A correctly specified money demand function is very important in the determination of the optimal way in which the Central Bank formulates and conducts its monetary policy to ensure that it supplies the optimum quantity of money for the stabilization of the economy. Moreover, the stability of the specified money demand function is also fundamental if monetary targeting is to have any predictable effect on the ultimate purposes of the policy which are price stability and real economic growth. Among the importance of this thesis is that it evaluates the stability of the money demand function which could help to determine if inflation is purely a monetary phenomenon in the Gambia and to ascertain the effectiveness of monetary policy in the Gambia as a policy tool. Moreover, the paper will also provide vital information that could help in tracking both interest rates and money stock. This is essential because it will help in assessing the impact of monetary policy in the Gambia. Furthermore, this research is significant because 14 it will add up to the limited available literature on the Gambian economy and hence will provide information for those students or researchers who may want to further explore this relationship in the Gambia. 1.5 Research Hypotheses The working hypotheses of this paper are presented below and they are differentiated by the number of asterisks associated with each hypothesis. H0* Real income has no effect on the demand for money in the Gambia H0** Interest rate has no effect on money demand in the Gambia H0*** Inflation has no effect on money demand in the Gambia H0**** Exchange rate has no effect on demand for money in the Gambia H0*****Money Demand is not Stable in the Gambia 1.6 Scope of the Paper Following this Introduction as Chapter 1, Chapter 2 presents a review of literature on both the theoretical and empirical evidence. Chapter 3 presents the financial sector and a trend analysis of broad money demand in the Gambia. While Chapter 4 discusses the methodology used in this study, Chapter 5 presents the analysis and interpretation of the results obtained using the methodologies mentioned in Chapter 4. Finally, Chapter 6 highlights the conclusion and recommendations on monetary management in the Gambia. 15 Chapter 2 LITERATURE REVIEW 2.0 Introduction This chapter presents a review of literatures on the theoretical and empirical fronts for both developed and developing economies as well as the Gambia. 2.1 Theoretical Framework The theory of money demand has been in the forefront of academic debates for many years. The earliest theory of money demand was implicitly put forward by Irving Fisher when he laid the quantity theory of money demand. Fisher (1911) argued that the demand for money is solely a function of income. Keynes (1936) in his liquidity preference theory of money demand argued that people demand money for transactional, precautionary and speculative motives. He argued that money demand depends on both income and interest rate. Portfolio theories of money demand treat money like any other asset and used the assets’ demand theory to derive the money demand theory. They argued that people hold money as part of their portfolio of assets because money offers different combination of risk and return than other assets. Inventory theories of money demand postulate that money demand for transactional motives have a positive relationship with income and a negative relationship with nominal interest rate earned on alternative assets. The Baumol-Tobin model is the most well-known inventory theoretic approach model. A cursory look at these theories is presented below. 16 2.1.1 Quantity Theory of Money Demand The earliest theory of money demand was implicitly put forward by Irving Fisher when he laid the foundation of the Quantity Theory of Money. The quantity theory of money demand is explained using the equation of exchange. According to the old Fisher equation of exchange, the demand for money in an economy is solely a function of the volume of transaction in an economy. In other words, people demand money solely for transaction purpose and the more money people need for transactional purpose, the more money they will demand. This relationship between money demand and the level of transaction is expressed in the equation below: Where, is the quantity of money balances; price level and is the transactional velocity of money; is the is the volume of transactions. Fisher (1911) argued that people demand money only for transactional purpose and the demand for money is inelastic to interest rate changes. This equation was later modified by the Cambridge School and they presented a slightly different version of the old equation by replacing with . The modification is due to the fact that there is a problem inherent with the original Fisher equation because the number of transactions in an economy is difficult to calculate. Hence, output transaction is used as a proxy for because the more an economy produces, the more goods and services are bought and sold. With this modification by the Cambridge economists, the equation of exchange becomes: 17 This equation is transformed into the Quantity Theory of Money Demand by solving for the real money balance and thus rewriting the equation as: Equilibrium in the money market is where the quantity of real money supplied to the demand for real money balance and is also equal to is equals which is constant reflecting institutional and technological features of the economy which are stable in the short run. This now gives us the quantity theory of money demand as: Thus, the quantity theory of money demand function shows that the demand for real money balance is solely a function of real income and this relationship is stable over time. 2.1.2 Liquidity Preference Theory of Money Demand Keynes in his famous book “The General Theory of Employment, Interest and Money” identified three motives why people demand money: the transactional motive, the precautionary motive and the speculative motive. Keynes developed a more general and realistic theory of money demand than Irving Fisher in his liquidity preference theory. Contrary to Fisher, Keynes believed that the demand for real money balances depends on both interest rate and income. According to Keynes (1936), the volume of transactions is positively related with income and if income increases, the demand for real money balances also increases for transactional and precautionary motives. Moreover, Keynes argued that money demand for speculative motives 18 is interest rate elastic because interest rate is the opportunity cost of holding money. Hence, the Keynesian money demand function is expressed as: According to the model above, the demand for real money balance is a function of income and nominal interest rate. Money demand is positively related with income and inversely related with interest rates. Keynes further argued that the velocity of money is not constant but instead it is positively related with interest rates which fluctuate considerably. Moreover, the liquidity preference function implicitly captures the effect of inflation on the money demand. This can be demonstrated by introducing the Fisher equation, which states that nominal interest rate is the sum of the real interest rate and expected inflation, into liquidity preference function. The Fisher effect is written as: The Fisher equation states that there is a one-to-one relationship between expected inflation and nominal interest rates. This now gives us the Keynesian money demand function that captures the effect of inflation on real money demand as: This equation implies that the demand for real money balance also depends negatively on the expected rate of inflation. Inflation rate positively influence nominal interest rate and nominal interest rate is the cost of holding money. This explanation gives us a more sophisticated explanation of money demand than the quantity demand of money theory because it shows 19 that money demand is an increasing function of income and a decreasing function of both interest rate and expected rate of inflation. 2.1.3 Portfolio Theories of Money Demand Portfolio theories of money demand provide a microeconomic explanation of money demand which emphasize the function of money as a store of value. Portfolio theories of demand for money emphasize that people hold money as part of their portfolio of assets because money is one asset among several and it offers a different combination of risks and return than other assets. Friedman (1956) and Tobin (1958) formulated the most well-Known portfolio theoretic approach models. Portfolio theories of money demand treat money like any other asset and used the assets’ demand theory to derive the money demand theory. According to these theories, the demand for money should be a function of the risk and return offered by money and by the alternative assets that households can hold instead of money. Moreover, it should also be a function of wealth, since the size of wealth determines the amount of the portfolio to be allotted between money and the alternative assets. This version of money demand can be expressed as: Where is the expected return on stock/equity, expected return on money, is the expected return on bonds, is the expected inflation rate and is the is the permanent income which is used as a proxy for wealth. According to portfolio theories, since the demand for assets increases as wealth increases, the demand for money is also positively related with permanent 20 income because higher wealth means larger portfolio. The three main assets identified by proponents of portfolio theories of money demand were bonds, stocks and goods; and they argued that the incentive to hold money depends on the attractiveness of these assets comparing to holding money. Expected returns on these assets are negatively related with the demand for money. As or increases, money demand declines because it becomes less attractive to hold money comparing to stock or bond holding. Furthermore, an increase in expected inflation also reduces money demand because money became less attractive as its real value depreciates over time. Mankiw (1997) highlighted that from the viewpoint of portfolio theories of money demand, the liquidity preference function of money demand is just a useful simplification of the general theory of money demand because firstly it uses real income as a proxy for real wealth and secondly only the nominal interest rate on money is included while ignoring the returns on other alternative assets. 2.1.4 Inventory Theories of Money Demand Another microeconomic theory of money demand is the inventory or transaction theories of money demand because they emphasize the role of money as a medium of exchange. Inventory theories of money demand consider the demand for money for transactional motives. The Baumol-Tobin model is the most well-known inventory theoretic approach model. Baumol (1952) and Tobin (1956) explicitly formulated a transactions demand for money in an inventory theoretic approach that provides a microeconomic explanation of money demand by analyzing the costs and benefits of money holding. The cost of holding money is the forgone interest and the benefit of money holding is liquidity. According to this theory while agents receive income 21 periodically for instance monthly, they make transactions at a constant rate over the period. The agent can decide to hold his entire income to make his daily transactions or save his entire income in an interest-bearing savings account or other interest-earning assets. However, there is a trade-off between the costs of holding his income – interest cost cost – and the transaction of converting interest-bearing assets into money. Thus, the optimal strategy is to hold a portion of his income as money and another portion in interest-bearing assets. The BaumolTobin model postulated that the optimal average money demand is given by: Where; rate, is the cost of converting interest-bearing assets into money, is the price level and is the nominal interest is income. Therefore, the Baumol-Tobin model postulates transactional money demand to have a positive relationship with income and a negative relationship with nominal interest rate earned on alternative assets. Moreover, the transaction cost of converting wealth between interest-bearing assets and money also has a positive relationship with money demand. If transaction cost declines, for example the introduction of Automatic Teller Machines, more wealth is held in the form of interest-bearing assets and less in the form of money. Thus, according to the Baumol-Tobin model the income elasticity of money demand and the interest rate elasticity of money demand are and – , respectively. This implies that average money demand should increase by 5 percent when income increases by 10 percent and average money demand should decline by 5 percent when interest rate increases by 10 percent. Inventory theories of money demand provide a microeconomic 22 explanation of the liquidity theory of money because they show that money demand is positively related with income and negatively with interest rates. 2.2 Empirical Evidence On the empirical front, there is a large body of literature documenting the determinants and stability of money demand in both developed and developing economies. This large body of literature on the topic could be divided into two generations. The first generation of this literature is the pre-1974 literature and the second generation is the post-1974 literature. This is because the earliest researches on the demand for money that were conducted before 1974 work reasonably well in estimating money demand and they concluded that the money demand function is stable over time (Tomori, 1972 and Goldfeld, 1973). However, during the post-1974 era the consensus that the demand for money is stable started to fall apart. In the United States, the estimated money demand functions over-predicted actual money demand, M1, which according to Goldfeld (1976) led to ‘the case of the missing money’ because actual money demand were lower than the estimated money demand. Whereas, the estimated money demand functions over-predicted actual money demand in the US, money demand equations under-predicted actual money demand in the United Kingdom (Artis and Lewis, 1974). The break-down of these money demand functions was partly explained by great financial innovations such as electronic funds transfer which altered the working definition of money even though the official definition did not change. The 1973 Arab oil embargo resulting to oil prices hike, higher inflation rates and sharply higher interest rates also contributed to the instability of such money demand functions. 23 However, currently the stability of money demand remains controversial because different researchers reached at different conclusions on the stability of money demand in different economies. Sriram (2009) evaluated broad money demand in the Gambia using an ECM model and monthly observations from 1988:1 to 2007:2. He found that there appeared to be a longrun relationship between real money balance, real GDP, interest rates on deposits at the commercial banks, yields on Treasury bill, and expected inflation; but the relationship was not stable. He argued that foreign-influence variables such as foreign interest rates and expected depreciation were not significant determinants of real money demand in The Gambia. Tomori (1972), pioneering empirical estimation of money demand in Nigeria, used OLS and annual data from 1960 to 1970 and concluded that income is a significant variable that explained variation in money demand regardless of the definition of money adopted. He further argued that this relationship is stable by running a separate regression for the period 1960 to 1966, and comparing the coefficients with the coefficients obtained from the full sample. Akinlo (2006) found that there exist a cointegrating relationship between broad money, income, interest rates and exchange rates in Nigeria. He further tested the stability of the function using an Autoregressive Distributed Lags Model (ARDLM) which revealed that the relationship is somewhat stable. Moreover, Kallon (1992) estimated the money demand function in Ghana using quarterly data from 1966:1 to 1986:4, and found that the money demand function in Ghana was stable. On the other hand, Andoh and Chappell (2002) estimated money demand in Ghana and tested if there was a structural break using annual data from 1960 to 1996. Their study revealed that there was a structural break of the Ghana’s money demand function in 1983. 24 Furthermore, Hamori (2008) empirically analyzed the money demand function in the SubSaharan African region using a non-stationary panel data analysis for 35 countries based on annual data from 1980 to 2005. His findings revealed that there exists a cointegrating relationship of the money demand function in the Sub-Saharan African region over the period under study regardless of whether M1 or M2 is used as a measure of money supply. Similarly, Narayan and Seema (2009) studied the demand for money function from a panel of 5 South Asian countries using data spanning 1980 to 2000. They found that there exist an equilibrium relationship between money demand and its determinants both for individual countries and for the panel. Using diagnostics testing tools, they found that the money demand functions of all these economies are stable except for Nepal. Valadkhani (2008) estimated the long- and short-run determinants of money demand in six countries in the Asian-Pacific region as a function of real income, interest rate spread 2, inflation rate, real effective exchange rate and the US real interest rate using panel data (1975-2002). He found that the long-run income elasticity is greater than unity and both the currency substitution and capital mobility hypotheses hold only in the long run. From this empirical evidence, it could be seen that variables such as income and inflation rate are important determinants of demand for money in developing countries while foreign exchange and foreign interest rates are less important determinants of money demand. Moreover, the mixed conclusions on the stability of money demand in different economies might be attributed to the fact that the factors that affects money demand vary in accordance to the realities of different economies. 2 The difference between the deposit and lending rates 25 Chapter Three FINANCIAL SECTOR DEVELOPMENT AND TREND ANALYSIS OF BROAD MONEY DEMAND IN THE GAMBIA 3.0 Introduction This chapter presents the Gambia’s financial Industry and a trend analysis of broad money demand in the Gambia between 2000 and 2010. 3.1 Financial Sector Development The Gambian financial system is small and under-developed. There are no merchant banks; no stock markets and the capital markets are underdeveloped. Like most nations, the Central Bank of the Gambia (CBG) is at the apex of the financial system of the country which is the institution charged with the responsibility of regulating the financial and monetary system in the Gambia. The sector is dominated by commercial banks, accounting for 97 percent of the sector’s total assets. The industry continues to show signs of resilience as it recorded an annual growth rate of 6.4 percent between 2001 and 2010, and contribute on average about 3.2 percent of GDP. The growth trend of the financial sector and its contribution to GDP from 2001 to 2010 is shown in Figure A1. From Figure A1, it can be seen that the contribution of the financial sector to GDP has not been changing much. However, growth in the industry has been volatile with a sharp growth from 2007 to 2008 which was mainly driven by the influx of commercial banks into the country during that period. 26 -10 0 10 20 30 Figure A1: Annual Growth Rate and Share in GDP of the Financial Sector in the Gambia 2000 2002 2004 2006 2008 2010 TIME Annual Growth Rate (%) Share in GDP (%) Data Source: Ministry of Finance and Economic Affairs, Republic of the Gambia. Commercial banking in the country is highly concentrated with the three largest banks accounting for more than 60% of total assets, although their share has declined over the years. Despite the large number of commercial banks in the country, financial intermediation is still low with a loans-to-deposits ration as low as 0.4 which also influences demand for money. This is because government securities which have low default risk and high yield are very attractive to commercial banks. In January-June 2010 and 2011, Treasury Bills accounted for 92.3% and 93.0% of Commercial Banks investment respectively (The Gambia Macroeconomic Bulletin, 2011). 27 3.2 Trend of Broad Money Demand in the Gambia Monetary policy in the Gambia continues to focus on containing inflation at a single digit, maintaining exchange rate stability and a viable external position. To sustain this single digit inflation rate target, the CBG target broad money growth at about 12½ percent during 2010 by limiting reserve money growth to 9.5 percent during the year (The Gambia Macroeconomic Bulletin, 2011). Figure A2 shows that the growth trend in broad money (M2) has generally been increasing between 1990 and 2010. On the supply side, broad money supply recorded a growth of 21 percent in June 2010, compared to 21.2 percent a year ago. This growth was supported by 9.6 percent growth in currency in circulation outside banks, 14.1 percent growth in demand deposits, 15.9 percent growth in savings deposits and significant growth of 45.3 percent in time deposits. 0 5000 m2 10000 15000 Figure A2: Trend of Broad Money Demand in the Gambia from 1990 to 2010 1990 1995 2000 time 2005 2010 Data Source: Central Bank of the Gambia 28 On the demand side, this 21 percent increase in broad money supply was substantially driven by a 5.9 percent growth in net foreign assets and 27.1 percent growth in net domestic assets over a year. Domestic credits increased by 23.7 percent from D6.9 billion in June 2009 to D8.6 billion in June 2010, sustained by 28.4 percent growth in government borrowing and 29.6 percent growth in credits to the private sector, while credits to public entities declined by 15.8 percent over one year. The increasing trend of broad money demand shown in Figure A2 illustrates that people want to hold more money for their transactional, precautionary and speculative purposes. 29 Chapter 4 METHODOLOGY 4.0 Introduction This chapter presents a bird’s-eye view of the model and estimation techniques used in this paper as well as the dataset and expected results. 4.1 Model Specification and Description of Variables The general specification of money demand in most macroeconomic literature postulates money demand as a function of income and interest rates. However, the conventional money demand function is extended by introducing exchange rate to take account of the currencysubstitution hypothesis and inflation rate to take account of the asset-subsistution hypothesis. The inclusion of inflation also has its theoretical foundation from the portfolio theoretic approach models discussed in . Mundell (1963) also suggested the exchange rates to be included in the standard money demand function to take account of the currencysubstitution phenomenon. A similar approach was used by Sriram (2009) and Bahmani-Oskooee and Chi Wing Ng (2002). However, both models were modified by excluding interest rates on deposits and foreign interest rates from the models of Sriram (2009) and Bahmani-Oskooee and Chi Wing Ng (2002) respectively. Against this background, this paper postulates an open economy money demand function which relates real money balance as a function of real 30 income, interest rates, inflation rate and exchange rate, and can be econometrically presented as: Where is real broad money demand, transactions and precautionary demand for money, to capture speculative demand for money, asset-substitution hypothesis and is the real GDP as a proxy to capture is the interest rate on T-bills as a proxy is the inflation rate as a proxy to capture the is the nominal dollar-to-dalasi exchange rate as a proxy to capture the currency-substitution phenomenon. 4.2 Dataset and its Source The dataset used in this paper was obtained from the IMF Statistical Appendix of The Gambia 1990 and 2008 reports and the CBG database. The variables used are nominal broad money demand (M2), Consumer price index (2004 base period), real GDP at 1974/75 price, 91-Day Tbills rates, CPI inflation rates, and the dalasi-to-dollar exchange rate. The real money demand was computed by dividing the nominal money demand (M2) with average prices (CPI). M2 was used as a measure of money demand because according to De Brouwer and Subbaraman (1993, p.10) a broader measure of money is more appropriate for modeling purposes because it: a) is less distorted by financial deregulation and innovations; and b) has a more reliable relationship with income. M2 and GDP are measured in millions of dalasi, interest rates and inflation rates are in percentage. All variables are at their end period rates. The variables are all 31 in quarterly frequencies except GDP. Hence, the Denton method of disaggregation was used to convert the annual figures of GDP into quarterly frequencies. The Denton was applied because it has the ability to disintegrate the annual data into quarterly frequencies in such a way that the sum of the quarterly frequencies will also be equal to the annual figure. Moreover, the dataset stretches from 1993 first quarter to 2008 fourth quarter, giving a total of 64 observations. 4.3 Pre-estimation Tests: Unit Roots and Cointegration Tests Several pre-estimation tests were conducted before undertaking any regression analysis. The unit root and cointegration tests are very important pre-estimation tests that are often used to circumvent the inherent limitations of traditional models as well as avoid spurious regression result (Hendry, 1986). The objective of these tests is to examine the properties of the time series used in this paper in order to avoid a nonsensical regression. Nonsense or spurious regression arises when an OLS regression is estimated with non-stationary variables and residuals. The Augmented Dickey-Fuller (ADF) test is employed to test whether the variables used in the estimation are stationarity or not. The test is performed by augmenting each variable with its lags. The number of lags to be used in all the pre-estimation and estimation models in this paper is determined by using the Alkaike Information Criterion (AIC). The maximum lag length is chosen based on the minimum AIC criterion. The ADF test in this study is conducted by including a constant only and a constant and with a time trend. The test could be estimated with the following regression: 32 Where term, is the individual variable at time , is the constant, , is a pure white noise error is the number of lags which should be large enough to ensure that the error terms are white noise and small enough to save degree of freedom, variable in quarters and is the trend . The equations above are the ADF with a constant and time trend, and ADF with only a constant. In each case, the null hypothesis is that ; which means that there is unit root or the time series is nonstationary. The alternative hypothesis is that ; this means that the variable is stationary, using a Tau ) statistics. At ninety-five (95) percent confidence level, if the p-value is less than or equals to 0.05, we reject the null hypothesis otherwise we do not reject the null hypothesis that the variable is nonstationary. Among the rich menu of pre-estimation techniques, unit root test is the first pre-estimation test employed in this study because it is prerequisite for cointegration test. Cointegration test is used to determine if there exists an equilibrium or long-run relationship between two or more variables. If two or more variables are non-stationary but a linear combination of them is stationary, then the variables are said to be cointegrated. In this paper, the Johansen cointegration technique is used to examine the existence of an equilibrium relationship 33 between money demand and its determinants. The Johansen cointegration model is a VARbased test and it is presented as: is an (n x 1) vector of endogenous variables and is also an (n x 1) vector of white noise error term, where n is the number of all the variables used in this paper. The rank of the matrix coefficient indicates the long run relationship among the variables. Full rank r = n means that the variables are cointegrated. Rank r = 0 means that the variables are not cointegrated and reduced rank where r lies between zero and n means that there are r cointegrating vector among the variables. According to (Bashier and Dahlan, 2011) this technique is preferred to the two-step Engle-Granger procedure because it can test for multiple cointegrating vectors. Johansen developed two test statistics – trace eigenvalue statistics and maximum eigenvalue statistics. In the trace test, the null hypothesis is that there are r cointegrating vectors and the alternative hypothesis is there are n cointegrating vectors. On the other hand, the maximum eigenvalue test tests the null hypothesis of r cointegrating vectors against the alternative hypothesis of r +1 cointegrating vectors. 4.4 Estimation Techniques: Error Correction Mechanism (ECM) An ECM model is employed to estimate money demand in the Gambia. The ECM technique has been the dominant estimation technique used in estimating money demand in most literatures. Therefore, we estimate money demand and evaluate its stability in the Gambia using a similar 34 technique. An ECM model helps us to study the short-run dynamics in the relationship between money demand and its determinants. After testing for the existence of a long-run relationship between money demand and its determinants using the Johansen cointegration test, the error correction model enables us to reconcile the short-run behavior of money demand with its long-run behavior. The ECM model is specified as: Where: Where is the error correction term, it is the residual from the cointegrating equation, is the error correction coefficient and are the estimated short-run coefficients. The error correction coefficient works to push short-run money demand disequilibrium back towards its long-run equilibrium and its shows the speed of this adjustment. Our interest in the ECM model is the error correction coefficient, the Impulse Response Function (IRF), the Forecast Error Variance Decomposition (FEVD) and the stability test. 4.5 Post-estimation Techniques: Impulse Response Function, Forecast Error Variance Decomposition and Stability Test The post-estimation techniques employed in this paper are the IRF, the FEVD and the stability test. The IRF measures the percentage change in the response variable due to a percentage change in the impulse factor. In this paper, the IRF test is conducted using money demand as the response variable, and income and interest rates as the impulse factors. Hence, the IRF in this paper measures the percentage change in money demand due to a percentage in either 35 interest rates or income. The FEVD measures the percentage of the variation in the response factor that is explained by a percentage shock in the impulse. In other words, the FEVD in this paper measures the percentage of the variation in real money demand that is explained by a percentage shock in either interest rate or income. The ECM stability test is also very important in this study. This test was used to determine if the relationship between money demand and its determinants in the Gambia were stable during the period under study. Rapid expansion of and innovations in the financial sector, high inflation rate, sharp depreciation of the Dalasi, a recession or the combination of these factors are factors that might cause money demand instability. Interestingly, these macroeconomic disequilibria where experienced in the country during the period under study. For instance during the recession in 2003, inflation was as high as 20 percent accompanied with a sharp depreciations of the Dalasi by 100% within a year. 4.6 Expected Results The expected sign and magnitude of the coefficient of the real income (income elasticity of money demand) has a very interesting meaning. If applies; if , then the quantity theory of money , the Baumol-Tobin inventory theoretical approach is applicable; and if , money can be considered as a luxury (Valadkhani, 2008). A prior, we expect the sign of real income to be positive. This is because as real income increases, people demand more money for their transactional and precautionary motives. We also expect the signs of all the other explanatory variables except exchange rates to be negative. The coefficient on interest rate is expected to be negative because it measures the real cost of holding money. If the interest rates on savings accounts, time-deposits accounts or T-Bills increase, people would be 36 willing to hold less money at their disposal. The magnitude of the coefficient of interest rates which is the interest elasticity of money demand is also very central in the debate over whether fiscal or monetary policy is a more powerful policy option in an economy. A low coefficient implies that monetary policy has a greater effect on output than fiscal policy while a high coefficient value imply that fiscal policy has a larger effect on output than monetary policy, ceteris paribus. The coefficient on inflation rate is also expected to be negative since an increase in inflation reduces the incentive to hold money. The rate of inflation serves as a proxy to measure the return on real assets as an alternative to holding domestic currency. Nachega (2001) argued that in developing countries where interest rate ceiling and capital controls prevail, assets substitution is likely to be between money and physical assets rather than between money and financial assets. The sign of the coefficient on exchange rates is unspecified because exchange rate depreciation may increase or reduce domestic money demand. According to Bahmani-Oskooee and Chi Wing Ng (2002) variation in the foreign exchange rate may have two effects on domestic money demand, which are the wealth effect and the currency-substitution effect. Exchange rate depreciation may be perceived as an increase in wealth by wealth holders in foreign economies which may leads to an increase in money demand. However, exchange rate depreciation may cause expectation of further depreciation. This will lead to a reduction in domestic demand of money because people are given enough reasons to hold less domestic money in order to avoid the associated capital loss. The sign of the coefficient will eventually be determined by the predominant effect. 37 Chapter 5 Analysis and Interpretation of Regression Results 5.0 Introduction This chapter deals with the analysis and interpretation of the findings of this thesis. The chapter is divided into five sections. Section 5.1 reports the results obtained from the pre-estimation tests. Section 5.2 presents the results of the estimated long-run money demand function. The results of the short-run money demand function are presented in Section 5.3. Section 5.4 shed light on the IRF and FEVD results, and Section 5.5 presents the ECM Stability test. 5.1 Unit Root and Johansen Cointegration Tests Results Before any formal unit root test was conducted, we plotted all the variables used in this paper in order to have initial glue about the properties of the variables. Figure A3 to A7 [in the Appendix] show a visual plot of all the variables. The Augmented Dickey-Fuller (ADF) test of unit root shows that all the variables in the model are stationary at first difference with the exception of interest rate and inflation rate which were stationary at their levels. The econometrics result of the ADF test is presented in Table A1. These results show that the null hypothesis of unit root at level cannot be rejected for money demand; income and exchange rate at 95 percent confidence level while the null hypothesis of unit root at level for interest and inflation rates can be rejected. The numbers of lags are included in the estimation in order to eliminate the possibility of autocorrelation in the error terms. 38 Table A1: Augmented Dickey Fuller Test Results No. of Lags Variables Intercept Only Levels First Difference Intercept and Trend Levels First Difference Log Real M2 5 0.9666 0.0160 0.1726 0.0831 Log GDP 2 0.9981 0.0078 0.4397 0.0114 91-Day T-Bills rate 4 0.0171 0.0000 0.0679 0.0000 CPI inflation rate 2 0.0000 0.0000 0.0596 0.0000 Log Forex 6 0.8268 0.043 0.5446 0.0521 Given that interest and inflation rates are stationary at their levels, and real money demand, real income and exchange rate are stationary at their first differences, we proceeded to determine whether there is a long-run cointegrating relationship among these variables. The trace statistics and the maximum eigenvalue statistics revealed the existence of at least four cointegrating vectors between real money balance and its determinants. The result is shown in Table A2. Table A2: Results from Johansen Cointegration Test Eigenvalues (lambda) .923607 Ho: rank<= (r) r 0 H1: Maximum lambda statistics (rank<=(r+1)) 159.45558 Trace Statistics (rank<=(p=5)) 232.12859 .59072856 1 55.389354 72.673008 .23113812 2 16.296324 17.283655 .01276239 3 .79636058 .98733066 .00307542 4 .19097007 .19097007 39 This signifies that there is a long-run relationship between money demand, income, interest rate, inflation rate and exchange rate in the Gambia. 5.2 Estimated long-run money demand function Given the evidence that the variables in the specified money demand model have a long-run relationship as shown by the Johansen cointegration technique, our next step is to estimate how the demand for money responds in the long-run to changes in its determinants. The econometrics result of the estimated long-run money demand function is shown in Table A3. Table A3: Result of the estimated Long-run money demand function Number of obs = 63; R-squared = 0.9900; Adj R-squared = 0.9893; Prob > F = 0.0000 The coefficients of the estimated long-run money demand model are consistent with a prior expectation and theoretical postulations regarding signs, and they are all statistically significant. The result shows that a percentage increase in real income will in the long-run increase real money demand by about 1.823%. Real income has the expected sign; hence it shows that the transactional and precautionary motives of money demand hold in the Gambia in the long-run. Its magnitude also has an interesting implication because it indicates that money can be considered as a luxury in the Gambia because the income elasticity of money demand is greater than one (Valadkhani, 2008). Interest rates has its expected sign (-0.005) which means that a 40 one percent increase in interest rate will in the long-run lead to a decline in real money demand by about 0.005%. This effect is statistically significant at 95 percent confidence level and it is very important because it shows that in the long-run people also hold money in the Gambia for speculative reasons. Inflation rate also has its expected sign (-0.013) which shows that a percentage increase in inflation rate will in the long-run leads to a decline in money demand in the Gambia by about 0.013%. This finding is in conformity with Nachega (2001) who argued that in developing countries where interest rate ceiling and capital controls prevail, assets substitution is likely to be between money and physical assets rather than between money and financial assets. This means that physical assets are substitutes of money in the Gambia and thus, inflation leads to a shift from money holding to asset holding. Suggesting that demand for money also has implications for portfolio decisions in the Gambia. The coefficient of exchange rate (0.579) shows that the currency-substitution phenomenon does not hold in the Gambia. The positive estimated coefficient on exchange rate is consistent with the fact that depreciation of domestic currency raises the domestic currency value of an individual’s foreign assets, and if this is perceived as an increase in wealth, then money demand would increase (Arango and Nadiri, 1981). This means that the wealth-effect of currency depreciation dominates the currency substitution-effect of currency depreciation in the Gambia in the long-run which postulates that money demand increases in the Gambia when the dalasi depreciates. The wealth-effect could be attributed to the fact that remittances play a very crucial role in the Gambian economy. A depreciation of the dalasi may be perceived as an increase in wealth because it raises the domestic value of dollar inflows which would increase money demand in the Gambia. 41 An Adjusted R-squared of 0.9893, used to measure the goodness-of-fit of the estimated model, indicates that the model explain about 99 per cent of the long-run behavior of money demand in the Gambia. This is similar to the findings of Valadkhani (2008) who also found that real income, interest rate spread, inflation rate, real effective exchange rate and the US real interest rate explained about 99 percent of the long-run variation of demand for money in six Asian-Pacific countries. The Ramsey RESET test was also employed to determine if the model has omitted a relevant variable and it indicates that the model is correctly specified [Result shown in the Appendix]. 5.3 Estimated Short-Run Money Demand Function Another experiment in this paper is to examine the short-run dynamics of the variables within an error correction model (ECM) given the evidence that the variables in the money demand function model are cointegrated. The ECM model shows how the model adjusts to the long-run equilibrium implied by the cointegrating equation. The ECM model shows that although the independent variables used in the model are important determinant of money demand in the Gambia in the long-run, they are not necessarily important determinants of money demand in the short-run. The result from the ECM also shows that the error correction mechanism (ECM) which is the residual from the cointegrating equations has its expected negative sign (.0290569) but statistically insignificant (0.787). The absolute value of the coefficient of the error-correction term indicates that about 2.9 percent of the disequilibrium in the money demand function is corrected by short-run adjustment within a quarter. This suggests that following short deviations, 2.9 per cent of adjustment to the long-run are corrected within one- 42 quarter period either by market mechanism, government intervention or combinations of both. The coefficient indicates that the speed of adjustment in the money demand function is relatively low in the Gambia. The adjusted R-squared of 0.5142, used to measure the goodness-of-fit of the estimated short-run money demand function, indicates that the model fairly explain about 51 percent of the short-run behavior of demand for money in the Gambia. The result is shown in Table B7 [in the Appendix]. 5.4 Impulse Response Function and Variance Decomposition Tests Results The IRF was also implemented to single-out the effect of a percentage change in one of the impulse factors on the response variable. The IRF test is conducted in this paper using real income and interest rate as the impulse variables, and real money demand as the response variable. This is because we want to estimate how money demand responds to changes in both interest rate and income. These two variables were selected as the impulse factors due to their strong theoretical backup. From the econometrics result shown in Table A4, the IRF reveals a positive relationship between money demand and real income, and a negative relationship between money demand and interest rates. The result shows that a percentage increase in real income increases money demand by 2.4 percent after one year and die down to 2.1 percent in the second year. Furthermore, the IRF also reveals that a percentage increase in interest rate reduces money demand by 0.023 percent after one year and 0.043 percent in the second year. The relative small change in real money demand due to a percentage change in interest rates give evidence that domestic assets holders in the Gambia still prefer to hold money to 43 themselves rather than investing in interest bearing monetary assets which is attributed to the cash-based nature of the Gambian economy. Table A4: Result of the Cumulative Impulse Response Function Step Real Income Interest Rate 0 0 0 1 .881385 -.000622 2 .375368 -.006525 3 .563695 -.016104 4 2.4492 -.022739 5 3.07514 -.026437 6 2.7019 -.03063 7 2.18506 -.036375 8 2.10806 -.042618 From the reported results shown in Table A5, the FEVD was also used to measures the percentage of the variation in money demand that is explained by a percentage shock in either income or interest rate. The FEVD result shows that 8.1 percent of changes in real money demand are explained by a percent change in real income after one year and this diminishes to 6.7 percent after two years. The FEVD result also illustrates that 0.8 percent of changes in real money demand are explained due to a percent change in interest rate in the first year and this lessen to 0.5 percent in the second year. The relatively small proportion of the variation in money demand that is explained by a percentage shock in interest rates is attributed to the cash-based nature of the Gambian economy. 44 Table A5: Result of the Forecast Error Variance Decomposition Step Real Income Interest Rate 0 0 0 1 0 0 2 .001439 .000737 3 .030066 .000966 4 .081423 .008145 5 .080092 .008221 6 .068097 .006539 7 .06435 .005601 8 .06765 .00512 The Lagrange-multiplier (LM) test was also employed to determine the presents of serial autocorrelation between the residual. As shown in Table B10 [in the appendix], there is no serial autocorrelation between the residual at any lag, up to lag two. Based on the Jarque-Bera (JB) normality test, multivariate normality of all the variables in the model cannot be rejected. 5.5 ECM Stability Test Result The ECM Stability test was also used to find out if demand for money is stable in the Gambia during the period under study. As shown in Table B12 [in the Appendix], the stability test shows that the relationship between demand for money and its determinants was not stable in the Gambia during the period under study which is consistent with the conclusion of Sriram (2009). The instability of the money demand function implies that inflation is not purely a monetary phenomenon in the Gambia. EMPU (2009) found that inflation in the Gambia is not purely a 45 monetary phenomenon because money supply is not an important determinant of inflation. Thus, the instability of the money demand function found in this paper shows that money supply targeting is not a reliable policy option to controlling inflation since shocks in money demand also influence the LM curve which destabilizes money supply targeting. The instability of the money demand function could not also be associated with macroeconomic disequilibria in 2003 when inflation was as high as 20 percent accompanied with a sharp depreciation of the Dalasi by 100 percent within a year. This is because a dummy variable was used to test whether there was a structural break in 2003 but we found out that demand for money was instable prior to and after 2003. 46 Chapter 6 Conclusion and Recommendations 6.0 Introduction This chapter is divided into two sections. Section 6.1 presents the conclusion and Section 6.2 highlights the policy recommendations based on the findings highlighted on this Thesis. 6.1 Conclusion A correctly specified money demand function is very important in the determination of the optimal way in which the central bank formulates and conducts its monetary policy, whether it follows a monetary regime of money supply targeting or interest rates targeting. The stability of money demand is prerequisite for any policy-driven change in monetary variables to have predictable effect on output, interest rate and ultimately prices through the transmission mechanism of monetary policy. This paper examines the long–and short-run determinants of money demand and its stability in the Gambia using quarterly time series data from 1993:I to 2008:IV. The Augmented Dickey-Fuller test clearly indicated that all the variables in the specified money demand function are stationary at first difference except interest rate and inflation rate which were stationary at their levels. The Johansen cointegration test shows that there is a long-run equilibrium relationship between real money demand, real income, interest rates, inflation rate and exchange rate in the Gambia. The estimated long-run money demand function shows that money demand in the long-run have a positive relationship with income and exchange rate, and a negative relationship with interest rate and inflation rate. The 47 estimated long-run income elasticity of money demand is greater than one, which implies that money can be considered a luxury in the Gambia. Moreover, the error correction model which captures short-run dynamics of money demand shows that although these variables are important determinant of money demand in the Gambia in the long-run, they are not significant determinants of money demand in the short-run. Hence, the transactional, precautionary and speculative demand of money holds in the Gambia only in the long-run. Likewise, the asset-substitution phenomenon also holds only in the long-run, which implies that domestic assets holders in the short-run do not view holding of real assets as an attractive alternative to monetary assets during periods of high inflation but as time passes by they tend to modify this view and start to view real assets holding as an attractive alternative to monetary holding because of the inflationary psychology developed. The insignificant effect of interest rate on money demand is evidence that domestic assets holders in the Gambia still prefer to hold money to themselves rather than investing in interest bearing monetary assets which is attributed to the cash-based nature of the Gambian economy. In addition, the currencysubstitution phenomenon does not hold in the Gambia. The wealth-effect of currency depreciation dominates the currency substitution-effect of currency depreciation in the Gambia in the long-run which might have arisen due to the crucial role remittances play in the Gambian economy. Thus, to answer the working hypotheses of the thesis presented in Section 1.5 of the Introduction, we conclude that real income, interest rate, inflation and exchange rate have effect on demand for money in the Gambia. Moreover, demand for money is not stable in the Gambia. 48 Furthermore, the error correction factor shows that about 2.9 percent of the disequilibrium in money demand is corrected by short-run adjustment within a quarter. This suggests that following short deviations, 2.9 per cent of adjustment to the long-run are corrected within onequarter period either by market mechanism, government intervention or combinations of both. The Impulse Response Function shows that a percentage increase in real income increases money demand by 2.4 percent after two years while a percentage increase in interest rate reduces money demand by 0.043 percent after two years. Moreover, the ECM stability test shows that money demand is not stable in the Gambia during the period under study. Hence, the Central Bank of the Gambia should be flexible in the use of broad money supply as an intermediate target because money supply targeting may not translate into changes in interest rates due to the instability of money demand. 6.2 Recommendations Based on the specified money demand function, in conducting monetary policy in the Gambia, monetary policy makers in the Central Bank of the Gambia (CBG) should consider real income, interest rate, inflation rates and exchange rates as key policy factors. Monetary authorities should also be cognizant of the fact that inflation is not purely a monetary phenomenon in the Gambia because the money demand function is not stable. Thus, a monetary policy which only targets money supply is not a reliable policy option to controlling inflation because money supply targeting may not translate into changes in interest rates due to adjustments or the instability of money demand a phenomenon which could be associated with Keynes’s liquidity traps. The CBG should therefore adopt an interest rate targeting regime because interest rate 49 targeting is an alternative policy option that could improve the outcome of monetary policy in the Gambia. Paul Samuelson in support of this approach noted that “God gave us two eyes so we can keep one on the money supply and the other on interest rate”. In interest rate targeting, the CBG should use instruments of monetary policy to set interest rates at levels that it deems would be consistent with low inflation and continued economic growth. However, the CBG should be aware that there is no straight - Jacket policy instrument to control inflation and effective demand and should use a mixture of policy options in conducting monetary policies. 50 REFERENCES Akinlo, A. E. (2006) “The Stability of Money Demand in Nigeria: An Autoregressive Distributed Lag Approach”, Journal of Policy Modelling Andoh, Samuel K. and Chappell D. (2002) “Stability of Money Demand Function: Evidence from Ghana”, Applied Economics Letters, Vol. 9, Issue 13 Arango, S., and M. Nadiri (1981) “Demand for Money in Open Economics,” Journal of Monetary Economics, 7, pp. 69-83. Arize, Augustine C., Darrat, Ali F. and Meyer, Donald J. (1990) “Capital Mobility, Monetization, and Money Demand: Evidence from Africa”, Center for Economic Research on Africa, School of Business, Montclair State University, Upper Montclair, New Jersey 07043 Artis M. J and Lewis M. K (1974) “How Unstable is the Demand for Money in the United Kingdom?” Economica, Blackwell Publishing, New Series, Vol. 51, No. 204 (Nov. 1984), pp.473476 Bahmani-Oskooee, Mohsen and Chi Wing Ng, Raymond (2002) “Long-Run Demand for Money in Hong Kong: An Application of the ARDL Model” International Journal of Business and Economics, Vol. 1, No. 2, 147-155 Ball, L. (2001) “Another Look at Long-Run Money Demand”, Journal of Applied Econometrics, 13, 57-76 Bashier, Al-Abdulrazag and Dahlan, Abdullah (2011) “The Money Demand Function for Jordan: An Empirical Investigation”, International Journal of Business and Social Science, Vol. 2, No. 5; [Special Issue- March 2011] Baumol, W. (1952) “The Transactions Demand for Cash: An Inventory Theoretic Approach”, Journal of Econometrics, pp. 545-56 Ceesay, Alieu (2000) "Demand for Money and Interest Rate Liberalisation in the Gambia" M.A. Thesis (Dakar, Senegal: United Nations Institute for Economic Planning and Development). 51 De Brouwer, G., I. Ng, and R. Subbaraman (1993) “The Demand for Money in Australia: New Tests on an Old Topic”, Research Discussion Paper, No. 9314, Sydney: Reserve Bank of Australia EMPU (2009) “The Interrelations among inflation and Major Fiscal, Monetary and Real Variables- A Case Study for the Gambia”, Institutional Support Project for Economic and Financial Governance (ISPEFG), Economic Management and Planning Unit (EMPU), Department of State for Finance and Economic Affairs (DOSFEA), The Republic of the Gambia, November 2009. Fisher, I. (1911) “The Purchaasing Power of Money”, New York, Macmillian Friedman, M. (1956) “The Quantity Theory of Money: A Restatement”. In M. Friedman, ed., Studies in the Quantity Theory of Money, Chicago: University of Chicago Press, 3-21 Friedman, M. and Schwartz, Anna J. (1982) “Monetary Trends in the United States and the United Kingdom: Their Relation to Income, Prices, and Interest Rates” 1867-1975, Chicago: University of Chicago Press Gerlach, S. and L. Svensson (2004) “Money and Inflation in the Euro Area: A case for Monetary Indictors”, Journal of Monetary Economics 50, 1649-1672 Goldfeld, Stephen M. (1994) “Demand for Money: Empirical Studies”, In the New Palgrave Dictionary of Money and Finance. Edited by P. Newman, M. Milgrate and J. Eatwell. London: Macmillan Press. Goldfeld, Stephen M. (1973) "The Demand for Money Revisited," Brookings Papers on Economic Activity, 3:1973, pp. 577-638 Goldfeld, Stephen M. (1976) “The Case of the Missing Money” Princeton University Haffer, R. W and Jansen, Dennis W. (1991) “The Demand for Money in the United States: Evidence from Cointegration Tests”, Journal of Money, Credit and Banking, Vol. 23, No. 2 (May, 1991), pp. 155-168 52 Hamori, Shigeyuki (2008) "Empirical Analysis of the Money Demand Function in Sub-Saharan Africa." Economics Bulletin, Vol. 15, No. 4 pp. 1-15 Hendry, D. F. (1986) “Econometric Modeling with Co-integrated Variables: An Overview. Oxford Bulletin of Economics and Statistics, Vol. 48, No. 3, pp. 201-12. Hendry, David F. and Ericsson, Neil R. (1991) “An Econometric Analysis of U.K. Money Demand in Monetary Trends in the United States and the United Kingdom by Milton Friedman and Anna J. Schwartz”, The American Economic Review, Vol.81, No.1, pp. 8-38 Hoffman, Dennis and Rasche, Robert H. (1989) “Long-Run Income and Interest Elasticities of Money Demand in the United States” NBER Working Paper Series, National Bureau of Economic Research, Cambridge, MA 02138, Working Paper No. 2949 Judd, John P., and John L. Scadding (1982) “The Search for a Stable Money Demand Function: A Survey of the Post-1973 Literature,” Journal of Economic Literature, Vol. 20 (September), pp. 993–1023. Kallon, Kelfala M. (1992) “An Econometric Analysis of Demand for Money in Ghana”, The Journal of Developing Area 26 Keynes, J. M. (1936) “The General Theory of Employment, Interest and Money”, Macmillan Mankiw, Gregory N.(1997) “Macroeconomics,”Third Edition, Havard University, Worth Publishers, New York, New York 10003, USA Ministry of Finance and Economic Affairs (2001) Economic Watch, First edition, December 2001 Ministry of Finance and Economic Affairs (2011) The Gambia Macroeconomic Bulletin, First Half 2011, Macroeconomic Policy Analysis Unit (MPAU), Ministry of Finance and Economic Affairs, The Quadrangle, Banjul, The Republic of The Gambia Mundel, R. (1963) “Capital Mobility and Stabilization Policy under Fixed and Flexible Exchange Rates”, McGill University 53 Narayan, Peresh and Seema, Mishra (2009) “Estimating Money Demand Functions for SouthAsian Countries”, Empirical Economics, Vol. 36, No. 3 Pedroni, P. (2001) “Purchasing power parity tests in Cointegrated panels”, Review of Economics and Statistics, 83, 727–731. Skrabic, B. and Tomic-Plazibat, N. (2009) “Evidence of the Long-run Equilibrium between Money Demand Determinants in Croatia”, World Academy of Science, Engineering and Technology 49 Sriram, Subramanian S. (2009) “The Gambia: Demand for Broad Money and Implications for Monetary Policy Conduct”, IMF Working Paper, African Department, International Monetary Fund Taylor, John B. (1994) “The Inflation/Output Variability Tradeoff Revisited” in Goals, Guidelines, and Constraints Facing Monetary Policy Makers (Federal Reserve Bank of Boston). Tobin, J. (1956) “The Interest Elasticity of Transactions Demand for Cash”, Review of Economics and Statistics 38 (3), pp. 241-47 Tobin, J. (1958) “Liquidity Preference as Behavior Toward Risk,” Review of Economic Studies, pp. 65-86 Tomori, S. (1972) “The Demand for Money in the Nigerian Economy” Nigerian Journal of Economics and Social Studies, 14(3): 337-345 Valadkhani, Abbas (2008) “Long- and Short-Run Determinants of the Demand for Money in the Asian-Pacific Countries: An Empirical Panel Investigation”, Annals of Economics and Finance, 9(1), 47-60 54 APPENDIX PART A1 Table B1: Summary Statistics of the variables in the model Variable Log Real M2 Log GDP Interest rate Inflation Log Forex Obs. 64 64 64 63 64 Mean 36.49258 -.4606271 16.95516 1.314513 2.76691 Std. Dev. 21.6073 .2025934 5.383732 2.022264 .4665733 Min 13.21817 -.727833 12 -4.84386 2.174752 Max 81.68407 -.0575475 31 5.910915 3.509454 Table B2: Correlation Matrix of the variables in the model Log Real M2 Log GDP Interest rate Inflation rate Log Forex Log Real M2 1.0000 0.9777 0.1693 0.1420 0.8837 Log GDP Interest rate Inflation Exchange rate 1.0000 0.1126 0.1685 0.8698 1.0000 0.3031 0.4907 1.0000 0.2984 1.0000 Table B3: Augmented Dickey Fuller Test Results No. of Lags Variables Log Real M2 Log GDP 91-Day T-Bills rate CPI inflation rate Log Forex 5 2 4 2 6 Intercept Only First Levels Difference 0.9666 0.0160 Intercept and Trend First Levels Difference 0.1726 0.0831 0.9981 0.0171 0.0000 0.8268 0.4397 0.0679 0.0596 0.5446 0.0078 0.0000 0.0000 0.043 0.0114 0.0000 0.0000 0.0521 55 Table B4: Results from Johansen Cointegration Test Eigenvalues (lambda) Ho: rank<= (r) r 0 1 2 3 4 .923607 .59072856 .23113812 .01276239 .00307542 H1: Maximum lambda statistics (rank<=(r+1)) 159.45558 55.389354 16.296324 .79636058 .19097007 Trace Statistics (rank<=(p=5)) 232.12859 72.673008 17.283655 .98733066 .19097007 Table B5: Results from the estimated Long-run money demand function and the Ramsey RESET Test Real M2 Coef. Std. Err. T P>t Log GDP Interest rate Inflation rate Log Exchange rate _cons F( 4, 58) = 1439.08 1.822544 -.0049575 -.0127306 .5794709 .1177146 .0024716 .0042127 .0585384 15.48 -2.01 -3.02 9.90 0.000 0.050 0.004 0.000 2.754004 .1825855 Prob > F R-squared = 0.0000 0.9900 [95% Interval] 1.586913 -.0099049 -.0211632 .4622935 15.08 0.000 2.38852 = Adj R- Number of obs = 63 squared = 0.9893 Conf. 2.058176 -.0000102 -.0042979 .6966483 3.119489 * significant at 5 percent Results for Ramsey RESET Test Ramsey RESET test using powers of the fitted values of lpm2 Ho: model has no omitted variables F(3, 55) = 31.61 Prob > F = 0.0000 Table B6: Results of the Lag Selection-order criteria Lag 0 1 2 3 LL -252.265 62.0496 121.914 158.714 LR Df P 0.000 0.000 0.000 FPE .069631 2.8e-06 6.4e-07 3.2e-07 AIC 8.68695 -1.42541 -2.91233 -3.61744 HQIC 8.74194 -1.1505 -2.41749 -2.90267 628.63 119.73 73.602 16 16 16 4 182.431 47.433* 16 0.000 2.6e-07* -3.87902* -2.94432* SBIC 8.8278 -.721161 -1.64468 1.78639* -1.48457 56 Table B7: Regression results of the error-correction model of money demand equation Coef -.0290569 -.3553059 -.4951452 -.2673922 .8125127 -.9739894 1.433838 -.000591 -.0031625 -.0019006 -.0142546 -.0001515 -.0026822 .2924242 .4008054 .0395493 cons .0378087 Log likelihood = P>chi2= 273.0351 0.0003 Std. Err. .1077249 .194202 .1674684 .1683449 2.368355 4.118918 2.570943 .0052543 .0052938 .0052416 .0216771 .0242572 .0049296 .1579699 .1618347 .1710921 .0148168 Rsq=0.5142 Z -0.27 -1.83 -2. 69 -1.59 0.34 -0.24 0.56 -0.11 -0.60 -0.36 -0.66 -0.01 -0.54 1.85 2.48 0.23 2.55 chi2= 44.45388 P>|z| [95% Conf. Interval] 0.787 -.2401939 .1820802 0.067 -.7359348 .025323 0.003 -.8233772 -.1669133 0.112 -.5973423 .0625578 0.732 -3.829379 5.454404 0.813 -9.04692 7.098942 0.577 -3.605118 6.472794 0.910 -.0108891 .0097072 0.550 -.0135381 .0072131 0.717 -.0121739 .0083727 0.511 -.056741 .0282318 0.995 -.0476948 .0473918 0.586 -.0123441 .0069797 0.064 -.0171911 .6020395 0.013 .0836152 .7179956 0.817 -.2957851 .3748837 0.011 .0087683 .0668491 AIC= - SBIC HQIC 6.238479 = - = 3.104567 5.015127 Table B8: Result of the cumulative Impulse Response Function and FEVD Step (1) (1) Cirf Fevd 0 0 0 1 .881385 0 2 .375368 .001439 3 .563695 .030066 4 2.4492 .081423 5 3.07514 .080092 6 2.7019 .068097 7 2.18506 .06435 8 2.10806 .06765 *Impulse factor= Real income and Response factor= Real money demand 57 Table B9: Result of the Cumulative Impulse Response Function and FEVD Step (1) (1) Cirf Fevd 0 0 0 1 -.000622 0 2 -.006525 .000737 3 -.016104 .000966 4 -.022739 .008145 5 -.026437 .008221 6 -.03063 .006539 7 -.036375 .005601 8 -.042618 .00512 *Impulse factor= Interest rate and Response factor= Real money demand Table B10: Result of Lagrange-multiplier test Lag chi2 1 22.6533 2 24.4559 H0: no autocorrelation at lag order Df 25 25 Prob > chi2 0.59780 0.49317 Table B11: Result of Jarque-Bera test Equation chi2 Df Prob > chi2 D_lpm2 0.335 2 0.84590 D_lgdp 2.280 2 0.31978 D_irate 173.994 2 0.00000 D_infl 5.651 2 0.05927 D_lforex 39.619 2 0.00000 ALL 221.879 10 0.00000 58 Table B12: Result of Eigenvalue stability condition Eigenvalue Modulus 1 1 1 1 1 1 1 1 7053378 + .3536431i .789028 .7053378 - .3536431i .789028 .7477387 + .1018127i .754638 .7477387 - .1018127i .754638 .3533548 + .6576239i .746545 .3533548 - .6576239i .746545 -.07562946 + .7384904i .742353 -.07562946 - .7384904i .742353 -.6861313 + .129883i .698316 -.6861313 - .129883i .698316 -.3672004 + .4458319i .577583 -.3672004 - .4458319i .577583 -.4065421 + .1694841i .440456 -.4065421 - .1694841i .440456 .1563731 + .3582525i .390893 .1563731 - .3582525i .390893 The VECM specification imposes 4 unit moduli. 59 PART A2 -.4 -.6 -.8 1993q1 1997q1 2001q1 year and quarter 2005q1 2009q1 15 20 25 30 Figure A4: Graph of Interest Rate at level 10 lrgdp -.2 0 Figure A3: The Graph of Log Real GDP at level 1993q1 1997q1 2001q1 year and quarter 2005q1 2009q1 60 3.5 2.5 3 lrm2 4 4.5 Figure A5: Graph of Log Real Broad Money Balance at level 1993q1 1997q1 2001q1 year and quarter 2005q1 2009q1 2005q1 2009q1 -4 -2 0 infl 2 4 6 Figure A6: Graph of Inflation Rate at level 1993q1 1997q1 2001q1 year and quarter 61 2 2.5 lforex 3 3.5 Figure A7: Graph of Log Exchange Rate at level 1993q1 1997q1 2001q1 year and quarter 2005q1 2009q1 62 PART A3 Figure B1: Graph of Cumulative IRF for ECM model orderb, lgdp, lpm2 3 2 1 0 0 2 4 6 8 step Graphs by irfname, impulse variable, and response variable * Impulse factor= Real income and Response factor= Real money demand Figure B2: Graph of Cumulative IRF for ECM model orderb, irate, lrm2 0 -.01 -.02 -.03 -.04 0 2 4 6 8 step Graphs by irfname, impulse variable, and response variable * Impulse factor= Interest rate and Response factor= Real money demand 63 Declaration I declare that this research project is the outcome of my own investigations, except where otherwise stated. Other sources are acknowledged by footnotes giving explicit references. Kebba Jammeh Signature: ………………… Date: ……………………. 64