An analysis of demand for money in the Lao people’s democratic republic

Journal of Economics and Development Vol. 14, No.3, December 2012, pp. 47 - 62 ISSN 1859 0020 An Analysis of Demand for Money in the Lao People’s Democratic Republic Tran Tho Dat National Economics University, Vietnam Email: tranthodat@gmail.com Ha Quynh Hoa National Economics University, Vietnam Somphao Phaysith Bank of the Lao PDR, Laos Abstract This paper is aimed at exploring the dynamic relationship between money bal - ance and four other macroeconomic variables: real GDP, expected inflation, exchange rates, domestic and foreign interest rates by modeling and testing for sta - bility of money demand functions in the Lao People’s Democratic Republic (PDR) during the period of 1993:Q1-2010:Q2. Demands for narrow money, broad money and board money in foreign currencies were estimated. The estimated results sug - gested that all demand functions are stable. They can be intermediate targets of the Bank of the Lao PDR. The substantial results point out: (i) there is an evidence of ample influence of exchange rates and interest rate on money balances in the Lao PDR; (ii) expected inflation indicates the effect of high inflation episodes on money balances, especially in terms of foreign currency, and (iii) the local currency, the Kip, is used predominantly for transaction purposes rather than foreign currencies. Keywords : Demand for money, long-run relationship, narrow money, broad money, error correction model. Journal of Economics and Development 47 Vol. 14, No.3, December 2012 1. Introduction opened for more than one year . Demand for money plays a major role in The financial market is developing within a macroeconomic analysis, especially in select - limited scope. Credit is limited and meets only ing appropriate monetary policy actions. 15 percent of the requirements with high non- Consequently, a steady stream of theoretical performing loans. The Lao economy is also and empirical research has been carried out partially dollarized. The total amount of for - worldwide over the past several decades. eign currency deposits to broad money was Money demand function was first conduct - 59.3 percent in 1992 and 55 percent to the end ed in developed countries where financial sys - 2011. tems developed and the central banks realized Therefore, in order to control the banking the role of money demand in conducting mon - system efficiently, BOL should consider the etary policy. However, lately there has been demand side when conducting monetary poli - considerable interest among several other cy. Up to now, there is no empirical study industrial and developing countries. about money demand for the Lao economy. Thus, this is the first study about demand for The Lao PDR is in the process of a transi - money for the Lao PDR. tion towards a market economy. The Lao econ - omy has experienced high fluctuations of This paper aims to explore the dynamic inflation rates. Monetary growth rates have not relationship between money balance and four been calculated by considering the demand other macroeconomic variables: real GDP, side. The implementation of financial sector expected inflation, exchange rates , and domes - policies has been slow in solving several tic and foreign interest rates by modeling and issues. The monetary policy framework is lim - testing for stability of money demand func - tions in the Lao PDR during the period of ited and incomplete. It is mainly based on the 1993:Q1-2010:Q2 . The paper is structured as obligation and issuance of bonds while BOL follows : Section 2 gives theoretical and empir - credit and marketing officers may have not yet ical overviews about demand for money. used them. It is for such reasons that the Section 3 presents the empirical results and sources of money and credit are restricted. The analysis of the results. Section 4 includes the exchange rate mechanism is not yet fully con - conclusion and provides policy implications of sistent with the actual conditions, thereby lim - the findings. iting the efficiency of its implementation. The main tools of BOL are interest rates, reserve 2. Overview of theoretical and empirical requirements , and discount window lending. studies on money demand The BOL has only used open market opera - 2.1. A brief theoretical overview tions since the Laos stock market has been There is a stream of theories about demand Journal of Economics and Development 48 Vol. 14, No.3, December 2012 for money. Theoretical developments on teenth and early twentieth centuries. Since the money demands began from the classical tra - classical economists believed that wages and dition. All theories try to explain two motives prices were completely flexible, they posited for holding money, namely transaction motive that the level of aggregate output produced in and asset motive. a normal economic period (Y) would remain at 2.1.1. Quantity theory of demand for money the full employment level, so Y by definition is The quantity theory of demand for money a nation’s total potential level of output. Fisher proposes a direct and proportional relationship assumed that the ratio between the level of between the quantity of money and the prevail - transactions, T, and output, Y, is reasonably ing price level. This relationship emerges with - stable (Y = t xT) and hence T can be treated as in the classical equilibrium framework using a constant in the short-run. two separate, but equivalent expressions. The Fisher believed that the velocity of money, first expression is associated with the V, is determined by the institutions in an econ - American economist , Irving Fisher and is omy, because these directly affect the way in called the “equation of exchange”. The second which individuals conduct transactions. For expression is associated with Cambridge example, if consumers use charge accounts University’s Arthur C.Pigou and is called the and credit cards to conduct their transactions, “Cambridge approach” or the “cash balance and consequently use money less often when approach”. making purchases, less money is required to a) Fisher’s “equation of exchange” conduct the transactions generated by nominal Fisher’s equation of exchange provides an income ( M decreases relative to PT ). Hence, important relation between four macroeco - velocity, defined as (PT)/M , will increase. On nomic variables to determine the nominal the other hand, if consumers find it more con - value of aggregate income . The four variables venient to purchase items with cash or checks in the equation of exchange are : the total (both of which are counted as money), more amount of money in circulation (M), an index money is used to conduct the transactions gen - of the total value of aggregate transactions (T), erated by the same levels of nominal income, the price level of articles traded ( P), and a pro - hence velocity will fall. Fisher theorized that portionality factor (V) denoting the “transac - institutional and technological features of the tion velocity of money”. The equation is given economy that affect velocity change only below: slowly over time, so velocity can safely be MV = PT (1) considered constant in the short-run. By divid - The classical economists (including Fisher ing both sides of the equation of exchange by himself) built on this relationship in the nine - V, the money demand function is obtained : Journal of Economics and Development 49 Vol. 14, No.3, December 2012 Md = ( 1/V)PT (2a) als desire money because money is a medium of exchange and a store of wealth. Cambridge Or equivalently, economists concluded that money demand Md =kPT (2b) would be proportional to nominal income and Equation (2b) states that because k is a con - expressed the demand for money function as: stant in the short-run (because V and T are con - Md =kPY (3) stant in the short-run), PT pins down the quan - In the short–run, k is the constant of propor - tity of money that people demand, Md. Fisher tionality and money demand does not depend believed that people hold money only to con - on the interest rate. However, money demand duct transactions and have no freedom of can depend on the interest rate when velocity action in terms of the amount they want to is not constant over time. hold. The demand for money is determined by the level of transactions generated by the level From the above discussion, the quantity the - of nominal income, PY , and by the institutions ory of money emerges as the theory with a in the economy that affect the way people con - simpler approach to estimating money duct transactions that determine velocity, V, demand. The estimating equation is : and hence k. Therefore, Fisher’s quantity theo - MV = PY (4) ry of money suggests that the demand for where M denotes nominal money stock, V money is purely a function of income. Interest denotes the income velocity of circulation, P rates have no effect on the demand for money. denotes the prevailing price level and Y b) Cambridge approach to money demand denotes real income. A group of classical economists, including Note that the elegant expression for money Alfred Marshall and Arthur C. Pigou in demand given by the quantity theory of money Cambridge studied the demand for money by relies on the assumption of constant velocity. considering how much individuals want to In reality, however, the velocity is not constant hold, given a set of circumstances. Pigou held especially during periods of financial liberal - the central assumption that individual demand ization. In these cases, equation (4) cannot for money is driven by the institutional envi - capture the complex relationship between the ronment, as this is the main factor that affects money demand and other macroeconomic whether individuals use money (i .e., cash and variables. Hence, we will turn to two other check) to conduct transactions. In the approaches to the theory of money demand: Cambridge model, individual demand for the Keynesian approach and Friedman’s mod - money is completely bound by institutional ern quantity theory approach. Both approaches constraints , such as whether one can use cred - consider the demand for money as part of the it cards to make purchases. Instead, individu - general issues of wealth allocation, but place Journal of Economics and Development 50 Vol. 14, No.3, December 2012 emphasis on different aspects of the problems. However, this is not an important weakness of 2.1.2. Keynesian approach these models because all three motives togeth - In 1936, Keynes offered a theory of demand er influence an individual’s optimal level of for money that emphasized the importance of money holding. interest rates. Keynes’ theory of money 2.1.3. Friedman’s model of the demand for demand (referred to as liquidity preference money theory), focuses on factors that influence indi - In 1956, Friedman developed the modern vidual decision-making. He postulated that quantity theory of demand in a famous article, there are three motives driving the demand for “The quantity theory of money: A restate - money: transaction motive, precautionary ment”. He simply stated that the demand for motive, and speculative motive. With this money must be influenced by the same factors view, money demand is a function of real that influence the demand for any other asset. income ( Y) and interest rate ( r). An individual’s demand for money should be a M / P = f (r,Y) (5) function of his wealth and his expected relative Equation (5) has the key implication that (to money) return on alternative investments. velocity is not constant and is positively corre - Friedman developed his theory on the lated with the interest rate, which fluctuates demand for money within the context of the substantially. Initially, Keynes suggested a liq - traditional microeconomic theories of con - uidity-preference schedule as in the following sumer behavior and of the producer demand equation: for input. Consumers hold money because it Md = M1 + M2 = M1(Y) + M2(r) (6) yields a direct utility stemming from the con - where: Md is the total demand for money, venience of holding an immediate form of pay - M1 is the sum of transaction and precautionary ment. Producers hold money because it is a demands , and M2 is speculative demand. In productive asset that smooths the payment and this schedule, transaction and precautionary expenditure streams over time. Therefore, the demand depends only on the level of income, sum of demand for money by both consumers Y, where dM1/dY > 0 . The speculative demand and producers is the demand for real balances. depends only on the level of interest rate, r, Intuitively, this demand should depend on the where dM2/dr < 0 . level of real income (or real output) as well as on the returns of alternative assets such as Although the Keynesian approach to ana - bonds or durable goods (for consumers). lyzing the demand for money focuses on the Therefore, the equation below gives us the three motives for holding money, the models demand function for real balances: do not allow us to uniquely identify an individ - ual’s particular motive for holding money. rm = M/P = f(Y, r 1, r 2,..., r n) (7) Journal of Economics and Development 51 Vol. 14, No.3, December 2012 where rm is the demand for real balances variables , and financial development and the sequence r1, r 2,, r n represent the real 2.2.1. Definition of money rates of return on alternative (i.e., non-money) Empirical studies have focused on three assets. monetary aggregates M1, M2, and M3. The In particular, Friedman considers durable component of monetary aggregate differ from goods as an important category of alternative country to country and depends on many fac - assets to money for consumers. With this view, tors, e.g., a country’s level of financial market the demand for consumers’ durable goods development. Economists have shown that depends on the expected inflation rate, πe. studies that interchange the use of M1, M2, or Then, the demand function for real balances M3 to estimate the demand for money face the also depends on the expected rate of inflation. problem of estimating heterogeneous assets. rm = f(Y, r, πe) (8) For example, cash and demand deposits may where drm/dY>0 , drm/dr<0 and drm/d πe< 0 differ significantly in terms of transaction In conclusion, all money demand models costs, risks of loss, and ease of concealment of can be broadly lumped into three separate illegal or tax-evading activities. One solution frameworks namely , transactions, asset and is to separately estimate the demand functions consumer demand theories of money. The for cash and demand deposits. This approach optimal stock of real money balances is has yielded more robust empirical results, but inversely related to the rate of return on earn - it does not resolve the underlying empirical ings of alternative assets and is positively difficulties. Any analysis in the Lao PDR will related to real income. This is the starting point face similar issues regarding the definitions of of all empirical studies. money and should leverage the advances made 2.2. Some empirical problems in estimating by economists to deal with these empirical money demand functions problems . All empirical studies are based on a conven - 2.2.2. Scale variable tional textbook formulation of a simple theo - Recently, scale variables were typically cre - retical demand for money function , , relating ated by using data on a country’s GNP , perma - demand for real money balances ( rm ) to a nent income or wealth, and cash measured in measure of transactions or scale variable ( Y) real terms. A number of other related variables and the opportunity cost of holding money ( r). that move together with GNP , such as net However, the demand for money functions national product ( NNP ) and GDP have also estimated for different countries are not the been heavily utilized in creating scale vari - same because of differences in the definition ables without any significant differences of dependent variables, availability of scale induced by the substitution. Traditionally, Journal of Economics and Development 52 Vol. 14, No.3, December 2012 GNP has been used for transaction-oriented Asia and South Asia. Various central bank offi - models, while modern-quantity theories relied cials realize that understanding money demand on permanent income. function is the cornerstone of monetary policy. Whichever measure of transactions is ulti - In this section, the set studies are carefully mately chosen, the question of whether it can chosen on the basis of potential relevance to be disaggregated into several scale variables the Lao PDR context. remains an open question. Economic aggre - Some Asia-specific studies (Fan and Liu gate proxies for scale variables in estimating (1970); Aghevli et.al (1979); Khan (1980); demand for money function depend much on Tseng and Corker (1991); Watanabe S. and development of statistic systems and available Pham T. B. (2005); Nguyen, D. H., and W. D. data. Pfau, (2010) ; Hoa, H.Q. (2008); Dat, T.T. and 2.2.3. Opportunity cost of holding money Hoa, H.Q. (2010)) show that demand for Interest rates in money demand function money is a proportion of income level, and this includes two groups: the own-rate of money is constrained by a measure of the wealth that and the rate of return on alternative assets. can be proxied by either income or permanent Tobin (1958) and Klein (1974) argue that both income. The demand for money fluctuates of these rates are important and should be with changes in the opportunity cost of holding included in any model for the demand for money. This opportunity cost depends on the money. This may be the interest rates of gov - relative return on non-money assets such as ernment securities, commercial paper, or sav - other financial investments and real goods. In ing deposits. In countries where the financial addition, expectations are important. The sector is not well developed and that also suf - demand for money depends not only on the fers from hyperinflation, the expected rate of prevailing level of factors such as the interest inflation is also a useful variable to calculate rate and inflation, but also on the future the opportunity cost of holding money. expected values of each of these factors. In the 2.3. Some Asia-specific studies on the case of dollarization, the interest rate of the money demand function dollar and the exchange rate are also an inter - A large body of literature is available to esti - esting explanation for demand for money bal - mate money demand functions. The initial ances. work in this area was confined primarily to In developed countries, the nominal interest industrial countries, especially the U.S. and the rate considers an appropriate proxy for the U.K. However , there has also been consider - opportunity cost of holding money, whereas able attention paid to studying the money the weak financial markets and administrative demand function in developing countries in interest rates are the overriding feature in most Journal of Economics and Development 53 Vol. 14, No.3, December 2012 developing countries. In most developing of money demand function and vary among countries the nominal interest rate is institu- empirical studies. tionally determined and it doesn’t fully capture Following the empirical literature on money the opportunity cost of holding money. demand in developing countries (Goldfeld and Furthermore, the administrative nominal inter- Sichel, 1990), the long-run money demand can est rates are not often adjusted for changes in be specified in the following (natural) logarith- inflation and consequently real interest rates mic form: become negative. Therefore, to overcome this d =+b bbbpe + +e + problem, researchers often use the consumer lnrmyit 01 lnttt 2 ln 3 t (10 ) price index as the proxy for the interest rate In most empirical studies, the interest rate variable. In fact, asset substitution in develop- term is used in non-logarithmic form, which ing countries usually takes place between leads to the following: money and real assets as inflation hedges and d e not between money and other financial assets. lnrmyit =+b0 bbbpe123 lnt +tt + +t (11 ) Thus the expected rate of inflation rather than rmd the nominal interest rate can be regarded as a where t is the desired demand for real better proxy for the opportunity cost of holding money balances, defined as the demand for money in developing countries. money supply deflated by the price level p, yt is a scale variable (for example, real measured 3. Estimating money demand function for income), i is the nominal interest rate on the Lao PRD t financial assets, which represents alternatives 3.1. Estimation Model e to holding money, π t is expected inflation The theory-based money demand function which measures the rate of expected return on for the Lao PRD is assumed to take the follow- physical assets, and ε is an error term. The d t ing form: rmt function is increasing in yt, and decreas- d π e M /P = α0 + α1Scale Variable (Y) ing in both it and t . When physical assets rep- resent the major alternative to holding money + α2Opportunity Cost Variable(r) (9) in high or hyperinflationary countries, the where Md is money demand balance, P is money demand may be specified as a function the price level, is therefore the demand for real of expected inflation alone Md/P=f(πe) (Peter money, Y is the real income that represents the Bofinger, 2001). scale variable and r is the interest rate on the alternative assets which represents the oppor- In developed countries, the nominal interest tunity cost variable. The selections of the scale rate is considered as an appropriate proxy for variable and the opportunity cost of holding the opportunity cost of holding money, where- money depend on the theoretical background as in most developing countries, the nominal Journal of Economics and Development 54 Vol. 14, No.3, December 2012 interest rate is institutionally determined and it The inclusion of foreign interest rates in the does not fully capture the opportunity cost of money demand function is to capture the effect holding money. Furthermore, the administra - of capital mobility and the expected exchange tive nominal interest rates are not often adjust - rate captures the substitution between domes - ed for changes in inflation and consequently tic and foreign currencies. Its impact on the the real interest rate becomes negative. demand for money can be either positive or Therefore, to overcome this problem, econo - negative. mists often use inflation rates as a measure of The error correction model (ECM) is used the opportunity cost of holding money to determine money demand and explain the (Bahmani-Oskooee and Tanku, 2006). dynamics of the economic model equation d (15 ) if observed variables are non-stationary lnrmycpit =+b01bb lnttt + 2 ln + e (12 ) and they are co-integrated (Engle and Granger , In fact, asset substitution in developing 1987). If the obtained results from unit root countries usually occurs between money and tests and the co-integration test of Johansen real assets as inflation hedges and not between approach are provided as in the Engle and money and other financial assets. Thus the Granger representation theorem, then the short expected rate of inflation , rather than the nom - run dynamics of money demand can be inal interest rate , can be regarded as a better described by ECM. The model in general form proxy for the opportunity cost of holding presents as: n n money in developing countries. Furthermore, DDD=+bb + bcg + + e lnlnrmrmECtiti01åå- ji t- i 1 given the fact of currency substitution in some i=10i= developing countries, many studies suggest to t = 0 - ++ge11ECtt- (15) include nominal exchange rate as an explana - tory variable in the estimated equation ECtt--1=--ln rm 1011bbc t - (Samreth and Sovannroeun, 2008). where EC is error-correction term , which d t-1 r lnrmt =+b01bb ln ytttt + 2 ln cpi + b 3 ln er + e () (1313 ) is derived from the long-run relationship and γ , is speed of adjustment to long run equilib - To capture the effects of foreign factors, 1 rium. χ is a set of explanatory variables. many studies on the demand for money in t Equation ( 15 ) will be estimated by OLS developing countries have included the impact method . of foreign interest rates and the expected rate depreciation of the domestic currency The ECM has proved to be the most suc - (Oluwole and Olugbenga, 2007). cessful tool in researching money demand. This type of formulation is a dynamic error- d * * lnrmt =+b 01bb ln yttttt + 2 ln cpi + b 3 ln er + be 4 i + () (1414 ) correction representation in which the long- Journal of Economics and Development 55 Vol. 14, No.3, December 2012 run equilibrium relationship between money Expected rate of inflation, exchange rates and its determinants is embedded in an equa - and interest rates are used as proxies of oppor - tion that captures short-run variation and tunity costs of holding money in Lao PDR. dynamics. The ECM is shown to contain infor - The past value of the actual inflation is used as mation on both the short- and long-run proper - a proxy of expected inflation rate. The quarter - ties of the model with disequilibrium as a ly series of saving USD interest rate is used as process of adjustment to the long-run model. a proxy of foreign currency interest rate due to In addition, the long-run equilibrium is speci - USD deposits taking the highest proportion. fied by economic theory while short-run Average exchange rates Kip/Dollar and dynamics are defined from the data. When co- Kip/Baht are used as proxies of exchange rate. integrated holds and if there is any shock that 3.3. The empirical results causes disequilibrium, there exists a well- As a result of the non-stationary I(1) process defined short-run dynamics adjustment in each series and co-integrating relations, the process such as error-correction mechanisms ECM is estimated to capture the long run rela - that will put back the system toward long-run tionship of money demand. On account of the equilibrium. VARs method and Johansen tests, it considers 3.2. Data description and issues the effects of all series in the whole system and The data used in this analysis is taken from verifies the co-integration of the multivariate the BOL. The estimated sample uses quarterly non-stationary which is helpful to avoid mis - data in the period from Q1/1993 to Q2/2010. specification. As a result, the ECM is estimat - ed in the first differencing form with up to six The study will apply both narrow money lags. The short-run dynamics presents in the M1 and broad money M2 as dependent vari - specific form as: ables. In addition, given the fact that there is the multi-currencies use phenomenon in the Lao PDR, hence, monetary aggregate will be classified by currency as local currency (Kip) and foreign currencies. M1 is narrow money including cash in circulation and current The error-correction term can be derived account. M2 is broad money consisting of M1, from the long-run equation as: savings and time deposits. EC t-1 = ln rm t - β0 - β 1ln rgdp t-1 - β2ln cpi t-1 According to the data availability, the scale - β3ln er t-1 - β4iusd t-1 - β5ikip t-1 (17) variable used in this study will be gross OLS estimation is applied for this two-step domestic product ( GDP ) as an income meas - error correction model in order to draw a rela - urement. tionship between money demand and its fac - Journal of Economics and Development 56 Vol. 14, No.3, December 2012 tors. The short run dynamic models including result reflects through the inertia in holding narrow money demand, broad money demand money that a 100 percent change in real nar - functions both in Kip and foreign currency row money demand in previous 2.2 quarters have a sensible statistic test. All coefficients still influences the current change by around are significant and reasonable explaining the 45 percent, with regarding to the effects of the model by approximately 45-60 percent. other explanatory variables. Durbin-Watson statistic interprets the overall The coefficient of is positive (0.36). This model serially uncorrelated. However, not all means that the demand for real narrow money signs are intuitive and plausible. will increase by 0.36% if the previous differ - Therefore, the model is examined for its encing in real money in M1 increases by 1 %, adequacy. It also looks for remedies such as if given other factors are unchanged. an important variable has been omitted or the Even the coefficient sign of the real income wrong function form has been used. To deter - is negative which is different from expecta - mine whether model inadequacy results from tion, but it significantly affects the real money one or more of these problems, various meth - M1 after two quarters. This explanatory vari - ods to test residual performances are used. The able is included in order to ensure the model results of the diagnostic test suggest that the validity . error term fulfills the classical assumptions, The coefficient ∆ln rm 1,t-1 of the real except for model for M2. exchange rate of the Kip against the USD is On the basis of the diagnostic tests, the short 0.29 after one quarter and -0.26 after two quar - run dynamic model of money demand pro - ters. If the negative coefficient in the second vides the validity of outcomes, except the quarter reflects the higher opportunity costs of broad money M2 model with misspecification. holding money, then the real money demand Therefore , the following discussion does not for M1 will decrease. If the Kip loses value by cover M2 demand function. 1 percent in the last two quarters, then real 3.3.1. Short-run money demand functions money demand M1 will decline by 0.26 per - a) Narrow money demand function cent, given that other factors are constant. The first difference of the real exchange rate ∆ln rm 1,t =0.03+0.36 ∆ln rm 1,t-1 -0.5 ∆ln rgdp t- Kip/USD in the previous quarter with a posi - 1 -0.41 ∆ ln rgdp t-2 +0.29 ∆ ln rerks t-1 - tive coefficient is included in this model in 0.26 ∆ln rerks t-2 -0.45 EC t-1 (18) order to maintain model validity . Based on the short-run estimated results, the b) Broad money demand function in Kip adjustment coefficient of error correction for long -run equilibrium shows the intuitive sign ∆ ln rm 2k,t =0.042-0.42 ∆ ln rgdp t-2 - with speed of adjustment in 2.2 quarters. This 0.27 ∆ri usd t-1 +0.37 ∆ri usd t-2 +0.23 EC t-1 (19) Journal of Economics and Development 57 Vol. 14, No.3, December 2012 Referring to the short-run estimated results, Based on the short-run estimated results, the the adjustment coefficient of error correction adjustment coefficient of error correction for for long -run equilibrium shows the intuitive long -run equilibrium shows the intuitive sign sign with speed of adjustment in 4.3 quarters. with speed of adjustment in 4 quarters. This This result reflects that through the inertia in result reflects that through the inertia in hold - holding money that a 100 percent changes in ing money , that the 100 percent change in real real broad money demand in 4.3 quarters ago broad money demand in 4 quarters ago still still influences the current change by around influences the current change by around 25 23 percent, regarding the effects of other percent, regardless the effects of other explanatory variables. explanatory variables. Even the coefficient signs of real income are Even the coefficient sign of the real income negative , which are different from the expecta - is negative which is different from the expec - tions, but it significantly affects the real money tation, but it significantly affects the real M2 in the Kip after two quarters. This explana - money M2 in foreign currencies after two tory variable is included in order to ensure the quarters. These explanatory variables are model validity . maintained in order to ensure the model valid - The coefficient of real saving in the USD ity . interest rate is -0.27 , after one quarter and 0.37 The coefficient of expected inflation is -0.77 after two quarters. The negative coefficient after two quarters. The negative coefficient after one quarter reflects the higher opportuni - shows the substitution effect of holding money ty costs of holding money, and so the money by physical goods including gold, land, and demand M2 in Kip will decrease. If the real houses. Consequently, the money demand M2 saving USD interest rate increases by 1 percent in foreign currencies will decrease. after one quarter, real money demand M2 in Specifically, if people expect that inflation will the Kip will decline by 0. 27 percent, given that increase by 1 percent in the last two quarters, other factors remain constant. The second dif - real money demand M2 in foreign currencies ference of real saving in the USD interest rates will decline by 0. 77 percent, given the other in the previous quarter with a positive coeffi - factors are unchanged. cient is included in this model in order to The coefficient of real saving USD interest maintain model validity . rates is 0.28 after one quarter. The positive c) Broad money demand function in foreign coefficient after one quarter reflects the incen - currencies tive for holding foreign currencies. ∆ln rm 2f,t =0.044-0.36 ∆ln rgdp t-1 -0.46 ∆ln rgdp t- Consequently, the money demand M2 in for - 2 - 0.77 ∆ln cpi t-2 + 0.28 ∆ri usd t-1 -0.25 EC t-1 (20) eign currencies will increase. If the real saving Journal of Economics and Development 58 Vol. 14, No.3, December 2012 USD interest rate increases by 1 percent after people tend to hold more foreign currencies one quarter, real money demand M2 in foreign such as USD or Baht when the local currency currencies will increase by 0. 28 percent, given loses value. Exchange rate elasticity of real that other factors remain constant. narrow money balances is -0.57 to Kip depre - 3.3.2. Long-run money demand functions ciation against the USD and -0.42 to Kip depreciation against the Baht. a) Narrow money demand function In the case of broad money demand function ln rm 1,t =2.34+0.81ln rgdp t-0.42ln rerkb t- in foreign currencies , the exchange rate coeffi - 0.57ln rerks t - 0.10 riusd t (21) cients of the Kip against the USD and against b) Broad money demand function in Kip the Baht are -0.77 and -2.33, respectively. This ln rm 2k,t =3.22+0.51ln rgdp t-1.96ln rerkb t- means that when the Kip depreciates against 0.4 riusd t (22) the USD or the Baht by 1%, people reduce for - c) Broad money demand function in foreign eign currencies from their portfolio by 0.77% currencies and 2.33%, respectively. This can be explained by the behavior of the Lao people, when the ln rm 2f,t =8.81+0.33ln rgdp t - 0.97ln cpi t-1 - value of the Kip decreases much, the citizens 2.33ln rerkb t-0.77 ln rerks t-2.59 rikip t+ expect inflation will be high. Hence, they 2.6 riusd t (23) hedge themselves by investing in real estate The long-run relationships for rm 1, rm 2k , and gold. As a result, M2 in foreign currencies and rm 2f have rational economic explana - tions. All signs are intuitive and plausible. decline. Demand for money in the Lao PDR has a pos - Capital mobility is sensible for the M1 and itive relation with income. If real GDP increas - M2 in Kip money demand models . The real saving USD interest rate elasticity has a nega - es by 1 %, demands for rm 1, rm 2k , and rm 2f will raise by 0.81%, 0.51% and 0.33% respec - tive sign, -0.1 in M1 function and -0.4 in M2 in tively . Thus, demand for money is most affect - Kip function. If the real saving USD interest ed by changes in output. The local currency, rate (opportunity cost of holding Kip) increas - the Kip, is used mostly for transaction purpos - es, people tend to hold less Kip balances . In es . the case of M2 in foreign currencies , the USD interest rate elasticity is 2.6 and the Kip inter - The Lao demand for M1, M2 in Kip money est rate elasticity is -2.95 . Laotian people will functions show the situation of a multi-curren - have more incentives for holding foreign cur - cy economy. Exchange rates, the Kip against rencies if the USD interest rate increases and the USD and against the Baht, have negative the Kip interest rates decrease. influences on money demand . This shows the effect of the currency substitution. Therefore , Expected inflation (cpi t-1 ) shows the impor - Journal of Economics and Development 59 Vol. 14, No.3, December 2012 tant implication to money demand in the Lao quarters. Both local and foreign interest rates PDR . Laotian people usually reduce their have significantly influenced real broad money balance holding when they expect high money demand in the local currency and in inflation . They hedge the risk of high inflation foreign currencies. Expected inflation rates by investing in physical assets such as gold, have a negative effect on real broad money housing and land. demands in foreign currencies after two quar - Outstanding results from money demand ters. These negative influences indicate a functions short-term substitution effect of foreign cur - Speed of adjustment from short-run dynam - rencies or real goods for real money demand. ics to long-run equilibrium: M1 model adjusts Opportunity cost proxies in the long-term to its long-run equilibrium relatively faster relations , M2 in foreign currencies model , than other models by using only 2.5 quarters include all plausible explanatory variables. while the others are 4.3 and 4 quarters for M2 This reflects the Lao situations appropriately in Kip and M2 in foreign currencies respec - with the phenomenon of multi-currencies use tively. This result reflects that the portfolio of which leads to currency substitution and the the M1 model is highly liquid with a cash- capital mobility effects. Expected inflation based component with relatively lower oppor - also presents a negative impact from the long tunity costs to reallocate its portfolio. period of high inflation, hence it sometimes Income is significant , taking contempora - requires people to adjust their rational attitude neous effects on the money demand adjust - by enhancing macroeconomic stability for a ment. Precisely, the elasticity of current money period of time until they are confident and feel demand differencing with respect to that of comparable. GDP is at least 0.4 percent for all types of 4. Conclusion and recommendations demand for money in the short-term. In the Demand for narrow money, broad money in long-run relation, considering the other effects, Kip and board money in foreign currencies an increase in real GDP by 1 % raises demand were estimated. The estimated results suggest - for money by 0.81%, 0.51% and 0.33% per - ed that all demand functions are stable. cent for rm 1, rm 2k , and rm 2f , respectively. Coefficient signs are suitable with theory. The Other important results from the short-term study results indicate : (i) there is evidence of ECM dynamics is that all proxies of opportu - ample influence of the exchange rate and inter - nities cost have effects on the real money est rates on money balance dynamics in the demand such as exchange rate depreciations , Lao PDR and this outcome is associated with especially Kip against USD affects the real a high degree of multi-currency use in the Lao narrow money demand negatively after two PDR; (ii) expected inflation shows the effect Journal of Economics and Development 60 Vol. 14, No.3, December 2012 of high inflation episodes on money balances , money in foreign currencies as intermediate especially in terms of foreign currency , even targets of monetary policy. BOL should take though the country has improved its macro - into account the effects of currency substitu - economic stability over recent years; and (iii) tion and the capital mobility of the Lao PDR in the local currency is used mostly for transac - a multi-currency economy. BOL should use open market operations frequently in control - tion purposes compared to that of foreign cur - ling the money supply since it is the most rencies . effective tool in conducting monetary policy in The study of money demand for the Lao the world. BOL also needs to strengthen bank - PRD gives some suggestions for BOL and the ing supervision to make the banking sector Lao government in controlling the economy operate more efficiently . The Lao government and conducting monetary policy. BOL can use should stimulate the development of the finan - narrow money, broad money in Kip and board cial system and by step by step de -dollarizing . References Aghevli, B.B. , et.al (1979), ‘ Monetary Policy in Selected Asian Countries ’, IMF Staff Papers, Vol. 26, pp. 775-824. Asian Development Bank (ADB) Team (2012), ‘ Lao PDR Bond Market Guide ’, February 2012. Bahmani-Oskooee, M.. and Tanku, A. (2006), ‘Black market exchange rate, currency substitution and demand for money in LDCs’, Economic Systems , 30, p249-263 Dat T.T. and Hoa H.Q. , (2010), Cau tien va chinh sach tien te o Viet Nam (Money demand and monetary policy in Vietnam, National Economics University Publishing House. Engle, Robert E. and Granger, C., (1987), Cointegration and Error-Correction: Representation, Estimation, and Testing’, Econometrica . 55:251-76. Fan, L.S., and Liu, Z.R., (1970), ‘Demand for Money in Asian Countries: Empirical Evidence’, Indian Economic Journa l, pp. 475-481. Goldfeld, S M. & Daniel E. Sichel., (1990), ‘ The Demand for Money ’, In the Handbook of Monetary Economics, 1, B. Hoa H.Q., (2008), ‘ Cau tien va he qua doi voi chinh sach tien te (Demand for money in Vietnam and pol - icy implications ’, PhD Dissertation, National Economics University, Hanoi, Vietnam. Hossain, A., & Chowdhuey, A., (1996), ‘ Monetary and Financial Policies in Developing Countries Growth and Stabilization ’. Routledge Studies in Development Economics, London and New York. Jayant M., (2007), ‘ Dollarization and the Multiple Currency Phenomenon in Lao PDR: Costs, Benefits and Policy Options ’, ADB Institute Discussion Paper No. 58 Keynes, J. M. (1936), The General Theory on Employment, Interest and Money . Journal of Economics and Development 61 Vol. 14, No.3, December 2012 Khan, M.S., (1980), ‘ Monetary shocks and the dynamics of inflation ’, IMF Staff Papers, Vol. 27, pp. 250- 284. Klein, Benjamin., 1974, ‘Competitive Interest Payments on Bank Deposits and the Long-Run Demand for Money ’, American Economic Review , American Economic Association, vol. 64(6), pages 931-49, December. Milton Friedman., (1956), ‘ The Quantity Theory of Money: A Restatement ’, in Studies in the Quantity Theory of Money, edited by M. Friedman. Reprinted in M. Friedman The Optimum Quantity of Money (2005), pp. 51-67. Ministry of Planning and Investment.., (2011), ‘ The seventh five-year National Socio-Economic Development Plan (2011-2015) ’, Vientiane, October 7. Nguyen, D. H., and W. D. Pfau., (2010), . ‘ The Determinants and Stability of Real Money Demand in Vietnam, 1999-2009 ’, GRIPS Discussion Paper 10-14. Tokyo: GRIPS. Oluwole Owoye and Olugbenga A. Onafwora., (2007), ‘M2 targeting, money demand and real GDP growth in Nigeria’, Journal of Business and Public Affairs ; Volume 1, Issue 2. Peter Bofinger., (2001), Monetary Policy, Goals, Institutions, Strategies, and instruments , Oxford University. Samreth, Sovannroeun., (2008), ‘ Estimation Money Demand Function in Cambodia: ARDL Approach ’, Munich Personal RePEe Archive. Tobin, J. (1958), ‘Liquidity Preference as Behavior Towards Risk’, Review of Economic Studies , Vol. 25 (1), p.15-29. Tseng, W, and R. Corker., (1991), ‘ Financial liberalization, money demand, and monetary policy in Asian countries’ , IMF Occasional Paper 84, IMF. Watanabe, S. and Pham, T.B., (2005), ‘Demand for Money in Dollarized, Transitional Economy: The Case of Vietnam’, Paper presented at the 1 st VDF-Tokyo Conference on the Development of Vietnam. Journal of Economics and Development 62 Vol. 14, No.3, December 2012