Financial Liberalization and the Information Content of Money in Indonesia, Korea, and the Philippines

A. J Hamann

doi:10.5089/9781451851083.001.A001

Jump to Content Jump to Main Navigation

Financial Liberalization and the Information Content of Money in Indonesia, Korea, and the Philippines

Author: Mr. A. J Hamann https://orcid.org/0000-0002-6363-0398

Publication Date:: 01 Nov 1993

eISBN:: 9781451851083

Language:: English

Keywords:: WP; M2; output equation; price; equation; price behavior; error correction terms; money growth; price development; managed float; money equation

Download PDF (1.8 MB)

Using a VAR approach, this paper studies the relationship between money, output, and prices in a group of Pacific Basin countries that underwent financial sector reform during the 1980s: Indonesia, Korea, and the Philippines. Special attention is paid to assessing the information content of money. Money was found to contain valuable advance information on output and prices in Korea, on prices only in the Philippines, and did not contain any advance information in Indonesia. The introduction of financial sector reform was not found to lead to a structural break in the price and output equations; however, the information content of money was affected. Further tests show that exchange and interest rates—variables that gained flexibility with the reforms—contain valuable information about future developments in prices in Korea and the Philippines.

Abstract

I. Introduction

The use of monetary policy in order to achieve a moderate path for the growth of nominal variables is common practice in developed countries. However, in much of the developing world, especially in high- inflation countries, the conduct of monetary policy has been driven by other concerns. Typically, the lack of well-functioning tax administrations, or even the reluctance of the authorities to raise the needed tax revenues in a relatively nondistortionary way, have led to the use of seigniorage as an important source of revenue, and turned monetary policy into an auxiliary fiscal policy tool. In these countries, monetary policy is clearly not guided by the objective of minimizing the departures of nominal variables from a stable path.

Low-inflation semi-industrialized economies such as those of fast growing Pacific Basin countries, constitute an interesting group in which the exercise of monetary policy has typically followed the objective of price stability to an important extent, but in a context in which financial markets were not fully developed; consequently, the monetary authorities have had to rely on a number of direct instruments of monetary control in order to achieve their policy goals. The experiences of many such countries, however, show that the prevalence of controls in the financial sector and the lack of money and capital markets provide only a short-lived framework for monetary policy. In fast-growing financially repressed economies the need for more sophisticated financial markets and the gradual erosion of the effectiveness of the controls have led to the adoption of financial sector reforms.

During a period of financial sector reform, the framework of monetary policy can be drastically altered as the transmission channels of monetary policy change and the economy’s financial structure is transformed. Moreover, the availability of new financial instruments may introduce instability in narrow definitions of money and reduce the authorities’ ability to control broader aggregates. Thus, financial sector reform warrants a reassessment by the monetary authorities of their basic monetary policy framework, and of the adequacy of their intermediate targets or information variables. ²/

One way of assessing the value of a variable to be monitored (or targeted) in the monetary policy process is by examining the robustness of the empirical relationship between that variable and the authorities’ ultimate policy goals, usually price stability and sustainable growth. Traditionally, empirical studies of the relationship between money, output, and prices have relied on the VAR methodology and have concentrated on the issue of causality (in the Granger sense) to determine whether money “matters.” ³/ Nevertheless, the approach has been criticized on the grounds that statistical evidence indicating Granger-causality from money to output may be capturing other phenomena such as simultaneous endogeneity of those variables to a common exogenous factor, or policy anticipations in the context of reverse causation (from output to money). Recently, however, Friedman and Kuttner (1992) have argued that independently of what the results of causality tests mean, if lagged values of money help predict future movements in prices or real output, then money contains valuable information that can be used by policymakers to guide their policy decisions.

This paper follows Friedman and Kuttner’s approach and examines the relationship between money, output, and prices using the VAR methodology in a sample of Pacific Basin countries that underwent financial sector reform during the 1980s: Indonesia, Korea, and the Philippines. The paper investigates first whether movements in money or credit are useful for predicting subsequent movements in output or prices. The focus is on the marginal predictive content of the monetary and credit aggregates, that is, whether these variables help predict fluctuations in output or prices when conditioned on other information. The paper then goes on to examine whether the observed empirical relationships in the different countries have changed over time in response to financial sector reforms.

The paper is organized as follows: Section II discusses the possible effects of financial sector reforms on the relationship between money, output, and prices; Section III provides a brief description of the experiences of Indonesia, Korea, and the Philippines with financial sector reform; Section IV looks at the empirical evidence on the money-output-price relationship in those countries and examines the effects of financial liberalization on that relationship; Section V assesses the information content of exchange and interest rates after the reforms; and Section VI summarizes the main findings of the paper and presents the conclusions.

II. Financial Sector Reform and the Relationship Between Money, Output, and Prices

As mentioned earlier, the empirical relationship between money, output, and prices may be affected by the introduction of financial sector reforms because such reforms are likely to prompt a change in the framework of monetary policy. This is the focus of Khan and Sundararajan’s (1991) study on the relationship between financial sector reform and monetary policy. Their analysis stresses the role of three factors: (i) instability in the money demand in the form of one-time variations in the effects of the variables determining money holdings; (ii) changes in the money supply process as liberalization alters the behavior of the public and banks; and (iii) changes in the propagation of the effects of monetary policy as liberalization induces a shift in monetary transmission channels from quantity rationing to market price rationing.

Instability in the money demand function is likely to arise mainly from the development of new financial assets, the liberalization of interest rates, and a reduction in transaction costs involved in portfolio reallocations. The money supply process is affected by financial sector reform in several ways, including: the substitution of indirect instruments of monetary policy for administrative ceilings in interest rates and credit aggregates, and less predictable fluctuations in the money multiplier as a result of financial deregulation. The transmission mechanism of monetary policy is affected as the roles of the exchange and interest rates may be enhanced at the expense of a reduced role for the credit rationing mechanism.

The effects described above will change the framework of monetary policy and, as a result, policymakers will need to respond by reassessing the usefulness of alternative financial variables used in the monetary policy process. This is the focus of the present paper. As a first step, the robustness of the money-output-price relationship is examined in the light of the financial reform measures. Then, the information content of several financial aggregates is examined before and after the reform, and the issue of which asset contains the most information on future movements of output and prices is examined.

Several empirical studies of the effects of financial liberalization on the money market have typically concentrated on the issue of stability of the money demand. For example, Tseng and Corker (1991) examine the effects of financial liberalization on money demand in a sample of Asian countries. They find that although the data confirm the existence of a stable long run demand for money, the introduction of financial sector reform created substantial instability in the short run. Goldsbrough and Zaidi (1989) also concentrate on the issue of the stability of the money demand in their study of the effects of financial sector reform and changes in monetary policy in the Philippines, and find similar results.

On the other hand, some studies have looked at the effects of financial innovations or deregulation on the relationship between money, output, and prices in developed economies. Friedman and Kuttner (1992), for example, look at the relationship between money, output, prices and interest rates in the face of deregulation and financial innovation, using U.S. data, and find that the relationship weakens when data from the 1980s are included in the regression. They also found that, at the same time, the role of interest rates becomes more important. A few empirical studies are available on the relationship between money, output, and prices in developing economies, but have not addressed explicitly the issue of the effects of financial sector reform on the information content of money. ⁴/

The following sections apply the VAR methodology to the analysis of the effects of financial sector reform on the monetary policy process. Using a simple time series approach in order to study the joint dynamic behavior of a set of macroeconomic variables, the paper focuses on the relationship between money, output, and prices and, in particular, assesses the information content of money. As in the case of the studies mentioned above the methodology used here does not allow for a detailed disaggregation of the effects of financial sector reform. Financial sector reforms are usually implemented over a period of time, using a particular sequencing, and the short run behavior of economic agents is likely to be dominated by intertemporal substitution effects caused by the anticipation of certain reform measures. Moreover, financial sector reforms are often implemented as a part of an overall reform program and, thus, it becomes practically impossible to discern between the effects of the financial reform measures and the other reform policies.

III. Financial Liberalization in Indonesia. Korea, and the Philippines

During the 1980s a number of Asian countries undertook financial sector reform programs. In most cases, these programs were part of a broader strategy aimed at both stabilizing the economy in the face of external shocks, and improving overall resource allocation in the economy. It was widely accepted that more sophisticated financial sectors were needed in order to cope with the increasing demands of their rapidly growing economies. Several studies have analyzed the reform programs implemented in those countries; thus, only a brief description of the issues most relevant for the empirical analysis of the money-output-price relationship examined in this paper is provided here. ⁵/

Prior to the adoption of the reforms, all three countries shared some characteristics of financial repression such as interest rate controls, domestic credit controls, high reserve requirements, segmented financial markets, and controls on international capital flows. As in other LDCs, these measures led to disintermediation in domestic banks, escalation of unregulated financial markets and nonbank financial institutions and, ultimately, to a significant reduction in the effectiveness of direct instruments of monetary control. Thus, financial sector reform was needed not only to improve overall resource allocation, but also to improve the effectiveness of monetary policy.

In most cases, the reforms implied the liberalization of interest rates, a reduction in the use of credit controls, the adoption of measures to enhance competition in the financial system, and the strengthening of the supervisory framework (Tseng and Corker (1991)). The reforms, however, were implemented at different speeds. In Indonesia and the Philippines significant action was taken early in the programs (first part of the 1980s), whereas in the case of Korea measures were adopted in a more gradual way in the first part of the decade but were intensified in the second half.

All three countries moved toward more flexible exchange rate regimes during the 1980s. Indonesia moved from a fixed exchange rate arrangement during the late 1970s when the rupiah was pegged to the U.S. dollar, to a more flexible regime in 1986. Korea, which had maintained a fixed exchange rate system since 1960, switched to a managed float in January of 1980. ⁶/ The Philippines adopted a managed float regime in October 1984.

The reforms were also accompanied by changes in the conduct of monetary policy. As explained above, since the reforms changed the basic framework of monetary policy, the monetary authorities were faced with the need to change their basic policy rules. As noted by Tseng and Corker, in most cases the changes took the form of moves from credit targeting to money or interest rate targets, due especially to the greater exchange rate flexibility adopted during the reform process. Indonesia adopted interest rate targeting following the reforms of 1983, reflecting concerns that money targeting in the face of financial reform may lead to unacceptable levels of real interest rates. ⁷/ However, when the rupiah was subject to a speculative attack in 1986, policy shifted gradually to targeting a certain level of international reserves. In Korea there was no change in the intermediate target, as M2 kept being used by the monetary authorities. In the Philippines, after the exchange rate was allowed to float in 1984, the authorities switched from targeting the net domestic assets of the central bank to base money. ⁸/

During the reform period all three countries moved toward greater reliance on indirect instruments of monetary control, especially open market operations, and less reliance on direct instruments such as reserve requirements or selective credit controls. Korea and Indonesia lowered reserve requirements substantially, while the Philippines began paying interests on reserves held at the central bank.

Clearly, then, the monetary policy framework changed significantly during the 1980s in the three countries. Some changes point toward a strengthening of the money-output-prices relationship: more flexible exchange rate regimes improved the authorities’ ability to control monetary aggregates; also, freer interest rates and more flexible exchange rates enhanced the role of these variables in the transmission mechanism. On the other hand, relaxation of several credit controls with rudimentary money and capital markets may have been a source of instability in monetary aggregates, and the appearance of new financial assets may have rendered the conventional monetary aggregates obsolete. The overall effects of these measures on the empirical relationship between money, income, and prices are examined in the next two sections.

IV. Money, Output, and Prices in Indonesia, Korea, and the Philippines

1. Times series properties of the data

This section presents the results of unit root tests for the time series of money, output, and prices for Indonesia, Korea, and the Philippines. ⁹/ As it is well known, such analysis is necessary in order to avoid “spurious correlation” problems in regression analysis. Table 1 summarizes the results of augmented Dickey-Fuller tests (ADF) performed for each of the series available under the null hypothesis of one (or two) unit roots. ¹⁰/ For all variables marked with an asterisk, the null hypothesis was rejected at the 5 percent level, except for currency in Korea which was rejected only at the 10 percent level.

Table 1.

Time Series Properties of the Macroeconomic Variables

^1/In parenthesis are shown the number of lags of the dependent variable that minimized the Schwartz Information Criterion.

Table 1.

Time Series Properties of the Macroeconomic Variables

Variable		Indonesia	Korea	Philippines
Tests for one unit root
Currency		-0.71 (2)	-0.50 (11)	-2.72 (3)
MB		-1.53 (3)	-1.86 (7)	-2.09 (4)
M1		-1.72 (3)	-1.15 (4)	-3.07 (3)
M2		-3.00 (1)	-2.51 (3)	-2.62 (4)
Credit		-1.63 (2)	-0.87 (8)	-1.54 (6)
Real GDP		-3.84 (1)*	-2.32 (5)	-1.57 (2)
CPI		-1.11 CD	-0.78 (1)	-3.15 (5)
Tests for two unit roots
Currency		-4.78 (2)*	-3.27 (11)*	-5.84 (3)*
MB		-4.17 (2)*	-2.98 (6)	-6.74 (3)*
M1		-3.45 (2)*	-4.72 (2)*	-6.12 (3)*
M2		-5.25 (0)*	-4.19 (5)*	-4.88 (3)*
Credit		-5.86 (1)*	-3.67 (7)*	-2.67 (5)
Real GDP			-7.44 (3)*	-8.54 (3)*
CPI		-4.17 (6)*	-5.24 (0)*	-5.79 (4)*
*Critical Values*
	*5 percent*		*-3.45*
	*10 percent*		*-3.15*

^1/In parenthesis are shown the number of lags of the dependent variable that minimized the Schwartz Information Criterion.

Table 1.

Time Series Properties of the Macroeconomic Variables

Variable		Indonesia	Korea	Philippines
Tests for one unit root
Currency		-0.71 (2)	-0.50 (11)	-2.72 (3)
MB		-1.53 (3)	-1.86 (7)	-2.09 (4)
M1		-1.72 (3)	-1.15 (4)	-3.07 (3)
M2		-3.00 (1)	-2.51 (3)	-2.62 (4)
Credit		-1.63 (2)	-0.87 (8)	-1.54 (6)
Real GDP		-3.84 (1)*	-2.32 (5)	-1.57 (2)
CPI		-1.11 CD	-0.78 (1)	-3.15 (5)
Tests for two unit roots
Currency		-4.78 (2)*	-3.27 (11)*	-5.84 (3)*
MB		-4.17 (2)*	-2.98 (6)	-6.74 (3)*
M1		-3.45 (2)*	-4.72 (2)*	-6.12 (3)*
M2		-5.25 (0)*	-4.19 (5)*	-4.88 (3)*
Credit		-5.86 (1)*	-3.67 (7)*	-2.67 (5)
Real GDP			-7.44 (3)*	-8.54 (3)*
CPI		-4.17 (6)*	-5.24 (0)*	-5.79 (4)*
*Critical Values*
	*5 percent*		*-3.45*
	*10 percent*		*-3.15*

^1/In parenthesis are shown the number of lags of the dependent variable that minimized the Schwartz Information Criterion.

The financial variables analyzed here, currency outside banks, base money, M1, M2 and domestic credit by the banking system, were all found to be difference-stationary, with the exception of base money in Korea and domestic credit in the Philippines, for which the null hypothesis of two unit roots could not be rejected. ¹¹/ The price levels, as defined by each country’s CPI, were also found to be difference-stationary. Real GDP, on the other hand, was stationary in levels in Indonesia but difference-stationary in Korea and the Philippines.

Given the results of the unit-root tests, the short term relationship between money, output, and prices will be modeled as a 3-variable unrestricted vector autoregression (VAR) in first differences for all variables except for Indonesia’s GDP, for which deviations from a linear trend will be used.

2. The money-output-price relationship and the information content of money

This subsection assesses the information content of money by testing the marginal predictive content of money in the output and price equations of an unrestricted vector autoregression system (VAR). For each of the three countries in the sample, the estimated reduced-form equations were:

\begin{matrix} Δ p_{t} = Σ_{i = 1}^{4} α_{i}^{p} Δ p_{t - i} + Σ_{i = 1}^{4} β_{i}^{p} Δ m_{t - i} + Σ_{i = 1}^{4} γ_{i}^{p} {Δ y}_{t - i} + ε_{p t} & (\begin{matrix} 1 \end{matrix}) \end{matrix}

\begin{matrix} Δ y_{t} = Σ_{i = 1}^{4} α_{i}^{y} Δ p_{t - i} + Σ_{i = 1}^{4} β_{i}^{y} Δ m_{t - i} + Σ_{i = 1}^{4} γ_{i}^{y} {Δ y}_{t - i} + ε_{y t} & (2) \end{matrix}

\begin{matrix} Δ m_{t} = Σ_{i = 1}^{4} α_{i}^{m} Δ p_{t - i} + Σ_{i = 1}^{4} β_{i}^{m} Δ m_{t - i} + Σ_{i = 1}^{4} γ_{i}^{m} {Δ y}_{t - i} + ε_{m t} & (3) \end{matrix}

The system (l)-(3) provides a general description of the relationship between money, output, and prices. ¹²/ Within that structure, the information content of money is commonly assessed by testing the significance of the βs in equations (1) and (2). Notice that what is being tested is the marginal predictive content of money; that is, not just whether some ‘m’ can predict the future variations in output (prices), but whether it can predict the part of the variation in output (prices) not already predictable from the observed past movements in output (prices) itself (themselves). If prices and output (or some pre-established path for their rates or growth) constitute the policy goals of the monetary authorities, the money variables that prove to be significant in equations (1) and (2) contain important information that could be used by policymakers in the formulation of their short run policies.

Despite Friedman and Kuctner’s disregard for equation (3) in the system described above, it is estimated in this paper mainly in order to see if the monetary policy changes implemented along with the reforms could be captured in this simple formulation of a monetary policy rule equation. Following the work of Sims (1972), the VAR methodology has focused on the issue of causality (in the Granger sense) between money and prices, or money and output. In the system (l)-(3), money is said to Granger-cause output if the βs in equations (1) or (2) are significantly different from zero. If both the βs in equation (1) and as in equation (3) are significantly different from zero, then there is two-way causality between money and prices, with the implication that money is not exogenous to movements in prices. This result presumably weakens the evidence on the information content of money based on one-equation regressions, and questions the validity of tests based exclusively on the βs in equation (1) to assess the usefulness of a given monetary aggregate in the monetary policy process. ¹³/

The concept of causality defined in this way has been the subject of much debate, as many authors question the significance of the term “causality” based simply on time series tests. As mentioned earlier, the main focus of this paper is the information content of money and not necessarily causality, i.e., whether the βs in equations (1) and (2) are different from zero, irrespective of whether that implies causality or not.

Equations (l)-(3) were estimated five times for each country, using alternatively the five financial aggregates analyzed in Table 1 as the relevant “money” variable. ¹⁴/ For each regression, F tests were performed in order to test the significance of the other endogenous variables, Including tests of block exogeneity of money, output, and prices. The results of these tests are reported in Table 2; an asterisk denotes that the null hypothesis of zero coefficients is rejected at the 5 percent level of significance.

Table 2.

F-Te s ts in Reduced-Form Equations

Table 2.

F-Te s ts in Reduced-Form Equations

	I. PRICE EQUATION			II. OUTPUT EQUATION			III. MONEY EQUATION
Financial Aggregate	Indonesia	Korea	Philippines	Indonesia	Korea	Philippines	Indonesia	Korea	Philippines
	(A) Ho: Coefficients on money = 0			(A) Ho: Coefficients on inflation = 0			(A) Ho: Coefficients on inflation = 0
Currency	1.92	3.73*	7.05*	0.55	2.33	1.62	1.06	0.93	0.29
Monetary Base	1.11	2.10	3.34*	0.96	2.11	2.09	0.63	0.04	1.02
M1	0.91	1.31	2.01	1.16	2.50	3.36*	0.91	0.17	1.01
M2	1.13	3.46*	2.78*	1.33	2.88*	1.53	0.95	0.44	0.49
Domestic credit	1.12	3.42*	1.27	1.49	1.84	2.28	0.50	3.03*	0.24
51 critical value	2.51	2.5-2	2.48	2.51	2.52	2.48	2.51	2.52	2.48
	(B) Ho: Coefficients on output = 0			(B) Ho: Coefficients on money = 0			(B) Ho: Coefficients on output = 0
Currency	0.23	2.02	0.35	0.32	1.12	0.35	0.44	0.48	0.53
Monetary Base	0.24	1.07	0.35	0.94	2.74*	1.05	1.02	1.45	0.32
M1	0.19	1.42	0.32	0.66	2.87*	2.09	1.14	1.40	0.37
M2	0.17	1.66	0.24	0.98	2.61*	0.71	0.48	1.39	1.23
Domestic credit	0.29	2.01	0.41	1.02	0.92	0.70	2.90*	1.34	0.42
52 critical value	2.51	2.52	2.48	2.51	2.52	2.48	2.51	2.52	2.48
	(C) Ho: Block exogeneity			(C) Ho: Block exogeneity			(C) Ho: Block exogeneity
Currency	1.01	2.66*	3.77*	0.72	1.54	1.27	0.69	0.S4	0.41
Monet »ry Base	0.60	1.83	1.89	1.04	2.44*	1.65	0.79	0.78	0.70
M1	0.50	1.36	1.21	0.90	2.51*	2.22*	0.98	0.83	0.64
M2	0.61	2.52*	1.60	1.07	2.36*	1.46	0.63	1.11	0.92
Domestic credit	0.60	2.50*	0.83	1.08	1.43	1.50	1.53	1.84	0.30
52 critical value	2.08	2.09	2.05	2.08	2.09	2.05	2.08	2.09	2.05

Table 2.

F-Te s ts in Reduced-Form Equations

	I. PRICE EQUATION			II. OUTPUT EQUATION			III. MONEY EQUATION
Financial Aggregate	Indonesia	Korea	Philippines	Indonesia	Korea	Philippines	Indonesia	Korea	Philippines
	(A) Ho: Coefficients on money = 0			(A) Ho: Coefficients on inflation = 0			(A) Ho: Coefficients on inflation = 0
Currency	1.92	3.73*	7.05*	0.55	2.33	1.62	1.06	0.93	0.29
Monetary Base	1.11	2.10	3.34*	0.96	2.11	2.09	0.63	0.04	1.02
M1	0.91	1.31	2.01	1.16	2.50	3.36*	0.91	0.17	1.01
M2	1.13	3.46*	2.78*	1.33	2.88*	1.53	0.95	0.44	0.49
Domestic credit	1.12	3.42*	1.27	1.49	1.84	2.28	0.50	3.03*	0.24
51 critical value	2.51	2.5-2	2.48	2.51	2.52	2.48	2.51	2.52	2.48
	(B) Ho: Coefficients on output = 0			(B) Ho: Coefficients on money = 0			(B) Ho: Coefficients on output = 0
Currency	0.23	2.02	0.35	0.32	1.12	0.35	0.44	0.48	0.53
Monetary Base	0.24	1.07	0.35	0.94	2.74*	1.05	1.02	1.45	0.32
M1	0.19	1.42	0.32	0.66	2.87*	2.09	1.14	1.40	0.37
M2	0.17	1.66	0.24	0.98	2.61*	0.71	0.48	1.39	1.23
Domestic credit	0.29	2.01	0.41	1.02	0.92	0.70	2.90*	1.34	0.42
52 critical value	2.51	2.52	2.48	2.51	2.52	2.48	2.51	2.52	2.48
	(C) Ho: Block exogeneity			(C) Ho: Block exogeneity			(C) Ho: Block exogeneity
Currency	1.01	2.66*	3.77*	0.72	1.54	1.27	0.69	0.S4	0.41
Monet »ry Base	0.60	1.83	1.89	1.04	2.44*	1.65	0.79	0.78	0.70
M1	0.50	1.36	1.21	0.90	2.51*	2.22*	0.98	0.83	0.64
M2	0.61	2.52*	1.60	1.07	2.36*	1.46	0.63	1.11	0.92
Domestic credit	0.60	2.50*	0.83	1.08	1.43	1.50	1.53	1.84	0.30
52 critical value	2.08	2.09	2.05	2.08	2.09	2.05	2.08	2.09	2.05

Quite surprisingly, the hypothesis that lagged values of some financial aggregate do not help predict the behavior of prices is not rejected in 9 out of 15 cases, although this is largely due to the results obtained for Indonesia. In fact, in Indonesia, no financial aggregate was found to be significant in any price equation, while in both Korea and the Philippines, 3 of the 5 financial aggregates used in the tests proved to contain Information on the future behavior of prices.

The money variables appear to be less significant in the output equations. Lagged values of some measure of money (MB, M1, and M2) were found to be useful predictors of output only in Korea; in Indonesia or the Philippines none of the money variables was significant. Only in two cases (the Philippines when M1 is used and Korea when M2 is used), past values of inflation were found to be useful for predicting fluctuations in output. In all other cases (11 out of 15), output can be treated as block-exogenous (i.e., the hypothesis that the coefficients of the lagged values of the other two variables are jointly equal to zero could not be rejected).

The money equations, on the other hand, seem to indicate that all financial aggregates could be treated as block exogenous in the three variable system. Tests on the significance of lagged values of a single variable, however, show that credit in Korea depends on past values of inflation, and the same variable in Indonesia depends on lagged values of output.

Table 3 summarizes the dynamics of the money-output-price relationship implied by the results of Table 2. For each pair of variables, the table shows whether they help predict future movements of each other, and summarizes the results in terms of non-existent, unidirectional or two-way causality (in the Granger sense). ¹⁵/ For example, in Indonesia no money variable helps predict output fluctuations, although lagged values of output growth do help predict future movements in credit. Thus, the table shows one-way Granger-causality going from output to credit. This is the only clear relation obtained for Indonesia; in all other cases money lacks information content and the variables are best regarded as block-exogenous.

Table 3.

The Money-Output-Price Relationship in Indonesia, Korea, and the Philippines

Table 3.

The Money-Output-Price Relationship in Indonesia, Korea, and the Philippines

	Indonesia	Korea	Philippines
(A) money/output
lags of m → y	…	(MB, M1, M2)	…
lags of y → m	(CR)	…	…
	y → CR	MB → y
		M1 → y
		M2 → y
(B) money/prices
lags of m → p	…	(CU, M2, CR)	(CU, MB, M2)
lags of p → m	…	(CR)	…
		CU → p	CU → p
		M2 → p	MB → p
		CR ↔ p	M2 → p
(C) output/prices
lags of y → p	…	…	…
lags of p → y	…	(M2)	(M1)
		P → y (M2)	P → y (M1)

Table 3.

The Money-Output-Price Relationship in Indonesia, Korea, and the Philippines

	Indonesia	Korea	Philippines
(A) money/output
lags of m → y	…	(MB, M1, M2)	…
lags of y → m	(CR)	…	…
	y → CR	MB → y
		M1 → y
		M2 → y
(B) money/prices
lags of m → p	…	(CU, M2, CR)	(CU, MB, M2)
lags of p → m	…	(CR)	…
		CU → p	CU → p
		M2 → p	MB → p
		CR ↔ p	M2 → p
(C) output/prices
lags of y → p	…	…	…
lags of p → y	…	(M2)	(M1)
		P → y (M2)	P → y (M1)

The lack of information content in money in Indonesia, especially in the inflation equations, is not inconsistent with previous studies of dynamics of inflation in that country. Parikh (1990) found that the authorities’ control over the money supply is limited by the influence of other factors, such as commodity prices, which affect monetary developments through their effect on the government’s budget. As a result, money is likely to lag rather than lead developments in the general price index.

In Korea, the dynamics of the system are more complex. Three financial aggregates (MB, M1, and M2) Granger-cause output growth and two financial aggregates (currency and M2 and) Granger-cause inflation; two-way causality between credit and inflation exists. However, M2 is the only aggregate significant in all the equations (notice that when M2 is the relevant financial aggregate inflation Granger-causes output). In that scenario, money is exogenous to the system, and changes in it have direct effects on both inflation and real output, as well as an indirect effects on output, through changes in inflation.

Other studies of the dynamics of money, output, and prices in Korea have yielded mixed results. For example, using annual data for the period 1951-1986, Ghatak and Deadman (1989) find no causality in either direction between M1 or M2 and prices. Similarly, Dhakal and Kandil (1993) find that changes in M1 are not significant in their inflation equation once an index for import prices is included as a regressor. They also find that neither prices nor output or interest rates Granger-cause money growth, a result consistent with our finding of block exogeneity of the money equations. This result, however, does not support the notion that monetary policy has been pro-cyclical in that country, as explained by Fry (1988, p. 373). The argument is that the use of automatic rediscounts at subsidized rates in connection with export credits led to temporary losses of control over monetary aggregates, as they accommodated positive shocks to the export sector.

In the Philippines there is no causality in either direction between any financial aggregate and output. Three financial aggregates (currency, base money, and M2), however, Granger-cause inflation, while inflation Granger-causes output when M1 is the relevant financial aggregate. Unlike Korea, however, no financial aggregate plays a significant role in more than one equation. In other words, three monetary aggregates contain information valuable for predicting fluctuations in prices, but no monetary aggregate helps predict output growth.

The results on the information content of money in the price equation for the Philippines contrast with findings of related studies. Using annual data for the period 1950-1986, Ghatak and Deadman (1989) found one way Granger-causality going from M1 to prices, but no relationship in either direction between M2 and prices (exactly opposite to the results of this paper regarding the information content of M1 and M2). Moreover, Lim (1987) found that the short run dynamics of inflation were closely related to cost-push elements, but not to demand developments induced by monetary policy.

It is interesting to note that in Korea and the Philippines both a narrow monetary aggregate, such as currency outside banks, as well as a broad monetary aggregate, such as M2, contain information about the future behavior of inflation, yet M1 does not contain any advance information in either country (and neither does the base in Korea). This result may seem somewhat puzzling, given the general belief that monetary aggregates move together. The explanation, however, lies in the high degree of substitutability between checking and saving deposits in both countries, as shown in Figures 1 and 2. ¹⁶/ in both cases, total deposits represent a relatively stable share of M2 (which contains significant information), but the shares of each type of deposits are highly volatile and move in opposite directions. The simple correlation coefficient between the shares of checking and saving deposits in M2 is -0.96 in Korea and -0.94 in the Philippines. Also, the relative stability of the share of total deposits in M2 implies also a stable relationship between currency and M2.

As for the lack of information content of base money in Korea, the explanation seems to lie in the very erratic reserve requirement policy followed by the Bank of Korea, especially during the 1980s, which led to significant variability in the money multiplier. ¹⁷/ Having targeted M2 before and after the reforms, and achieved a relatively stable path for it, this policy-induced volatility in the multiplier introduced excessive noise in the nominal base money series (Figures 3 and 4).

Finally, notice that the credit variable is relevant only in Korea, where 2-way causality between it and inflation exists. Surprisingly, credit is not significant in the output equations of any country. Several authors have been argued that credit plays an important role in the monetary transmission process (most recently Bernanke and Blinder (1988), Blinder (1987), and Brunner and Meltzer (1988)), and that a broad measure of credit should be used as an intermediate target for monetary policy (Davis (1979), Kopcke (1983)). The results of our tests are especially interesting given that one of the often cited characteristics of developing countries is the presence of some form of credit rationing, which makes this aggregate a direct link between the monetary and real sectors. Clearly, this issue deserves more attention, but such work is beyond the scope of this paper.

3. The “best” financial aggregate

Having identified cases where some money variable contains information on the future behavior of output or prices, the next step is to determine which of the financial aggregates that have proven significant in the inflation or output equations contains the most valuable information. With this purpose, price and output equations for Korea, and price equations for the Philippines are examined in Tables 4 and 5. The equations were run using as the relevant money variable only those that proved to be significant in Table 2. In each case, we looked at 3 “goodness-of fit” indicators, the adjusted R², the log of the likelihood function, and the root of the mean square error (RMSE) over the estimation period; we also computed the root of the mean square forecasting error (RMSFE) over different forecasting horizons.

Table 4.

Korea: Best Financial Aggregate in Price and Output Equations

Table 5.

Philippines: Best Financial Aggregate in the Price Equation

Table 4 shows the results for the inflation and output equations for Korea. The “goodness-of-fit” indicators for the inflation equation are very similar for the three aggregates used as the relevant money variable (currency, M2, and credit), with a small edge in favor of currency as the most informational variable. The RMSFE for different forecasting horizons, however, shows that M2 outperforms the other two aggregates. Since the focus of the present analysis is on the predictive content of money, the tests suggest that M2 contains the most information about the future behavior of prices. The best results for Korea’s output equation are obtained when M1 is used as the relevant monetary aggregate, according to the “goodness-of-fit” statistics although, as in the case of the price equation, the differences are minimal. Nevertheless, based on the RMSFE, M2 is again the best financial aggregate.

In the case of the Philippines a similar exercise was performed, but only for the price equation, since no monetary aggregate was found to have significant power for predicting the behavior of output in that country. The results are shown in Table 5. Of the three, aggregates with significant coefficients in the inflation equation, currency outperforms the other two in terms of the three “goodness-of-fit” statistics; however, using the RMSFE criterion the results are less clear. The best results are obtained either using currency or the money base, depending on the forecasting period.

These results provide some evidence that in the cases of Korea and the Philippines the monetary aggregates chosen as intermediate targets by the monetary authorities (over the latter part of the estimation period) turn out to be the ones containing the most advance information on the behavior of prices (Korea and the Philippines) and output (Korea). Moreover, in the case of Indonesia, where no “quantity” variable was used as an intermediate target, all financial aggregates examined here lacked advance information on either prices or output. The question arises, then, of whether this result reflects that advance information was an important criterion used by these countries to select an intermediate target, or whether the use of these aggregates as intermediate targets has in any way “contaminated” the data. A partial answer to this question is provided in the next subsection.

4. Financial sector reform and the money-output-price relationship

This section examines the effects of financial sector reform on the money-output-price relationships identified in the previous sub - sections. As a first step, simple Chow tests for the stability of the set of parameters of the reduced form equations (1)-(3) are performed, taking the first quarter of 1980 as the cutoff point. Although the reforms started in different years and were implemented at different speeds, 1980 was selected as the common cutoff point in order to avoid degrees of freedom problems.

The results of the Chow tests are reported in Table 6. In almost all cases the null hypothesis of no structural change in the parameters could not be rejected; the exceptions were the price equations with currency and M1 as the financial aggregates, and the M2 equation for Korea; and the M1 equation for the Philippines. All of the equations for Indonesia passed the stability tests. Notice that the M2 equation for Korea and the M1 equation for Indonesia were both found to be independent of price and output developments; thus, the results point in the direction of a change in the univariate process generating those aggregates (i.e., independent changes in monetary policy rules) which, as explained in Section III, did take place during the reform period. Notice also that the unstable price equations in Korea are those using currency and M1 as the pertinent monetary aggregate, but the price equation with M2 (the aggregate with the most advance information) is stable. Overall, the results of the Chow tests provide little evidence in favor of any structural breaks in the price and output equations associated with financial sector reform.

Table 6.

Chow Tests for Structural Break in the 1980s

Table 6.

Chow Tests for Structural Break in the 1980s

		Indonesia	Korea	Philippines
(A)	Price eauation
	Currency	0.70	2.94*	0.80
	MB	0.75	1.85	0.57
	M1	1.04	2.72*	1.66
	M2	0.80	1.64	1.05
	Credit	0.81	1.03	1.15
	*5 percent critical value*	*1.83*	*1.86*	*1.79*
(B)	Output equation
	Currency	1.01	0.74	0.94
	MB	1.48	0.67	1.25
	M1	1.23	0.62	1.29
	M2	1.13	0.84	1.38
	Credit	0.62	0.84	0.63
	*5 percent critical value*	*1.83*	*1.86*	*1.79*
(C)	Honey equation
	Currency	0.77	1.68	1.56
	MB	0.96	0.77	1.50
	M1	1.42	1.03	2.00*
	M2	0.82	1.91*	1.29
	Credit	1.09	1.44	0.35
	*5 percent critical value*	*1.83*	*1.86*	*1.79*

Table 6.

Chow Tests for Structural Break in the 1980s

		Indonesia	Korea	Philippines
(A)	Price eauation
	Currency	0.70	2.94*	0.80
	MB	0.75	1.85	0.57
	M1	1.04	2.72*	1.66
	M2	0.80	1.64	1.05
	Credit	0.81	1.03	1.15
	*5 percent critical value*	*1.83*	*1.86*	*1.79*
(B)	Output equation
	Currency	1.01	0.74	0.94
	MB	1.48	0.67	1.25
	M1	1.23	0.62	1.29
	M2	1.13	0.84	1.38
	Credit	0.62	0.84	0.63
	*5 percent critical value*	*1.83*	*1.86*	*1.79*
(C)	Honey equation
	Currency	0.77	1.68	1.56
	MB	0.96	0.77	1.50
	M1	1.42	1.03	2.00*
	M2	0.82	1.91*	1.29
	Credit	1.09	1.44	0.35
	*5 percent critical value*	*1.83*	*1.86*	*1.79*

The possibility of structural changes in any quarter throughout the estimation period was investigated by applying CUSUM tests to the residuals of the equations with the “best” financial aggregates: inflation and output equations with M2 for Korea (Figure II.1), and the inflation equation with base money for the Philippines (Figure II.2). ¹⁸/ Notice that in no case the cumulative sum of errors crosses the significance lines, extending our previous findings of no structural breaks to the entire estimation period for these equations.

A final look at the relationship between financial sector reform and the information content of money is provided in Tables 8-10 in Appendix III, where the F significance tests shown in Table 2 are recalculated for two sub-periods taking, again, the first quarter of 1980 as the cutoff point. No changes are found in Indonesia, where the basic result of no information content of any monetary aggregate for the purpose of predicting the future behavior of either prices or output holds also for each sub-period. The only changes are present in the money equations. While lagged values of output are found statistically significant in the currency and money base equations for the entire period, they are not significant in either sub-period individually considered. On the other hand, lagged values of output were found significant in the domestic credit equation during the sub-period 69:3-79:4 but not in the other sub-period or over the entire sample.

In Korea, a clear pattern emerges from Table 9, regarding the information content of money. The significance of money (or any other variable) disappears completely during the 1980s. In no equation ran for that period is the null hypothesis of zero coefficients rejected. In the particular case of the price equations, notice that the significance of currency over the entire sample holds during the 1970s but not during the 1980s. M2 and credit, which are significant for the entire period, are not significant in either sub-sample, while M1, which is not significant during the entire period is significant during the 1970s. Similar results are obtained for the output equation. M2 is significant throughout the entire period and also during the 1970s, but not during the 1980s. ¹⁹/ The money base and M1 are significant for the entire period but not for any sub-period.

A different result is obtained for the price equations in the Philippines, where no financial aggregate is significant in any equation during the sub-period 64:3-79:4 but four aggregates are significant during the 1980s. Lagged values of M1 or credit, which are not significant in either the entire period or the 1970s, become significant during the 1980s. Currency and M2, which are significant for the entire period are also significant during the 1980s but not during the period 64:3-79:4. The money base, which is not significant in either sub-period was not only significant for the entire period but had also proved to contain the most advance information.

In sum, the results show that despite the lack of clear evidence in favor of structural break in the 1980s, there are changes in the statistical importance of the regressors between the two sub-periods. The information content of money collapses in Korea but seems to increase for some aggregates in the Philippines. This result seems to suggest that the net effect of the financial reforms on the information content of money may differ across countries depending on the relative importance of the factors tending to strengthen or weaken the relationship between money, income and prices described in Section III.

On the other hand, the findings of this sub-section seem to suggest that the selection of a particular intermediate target has not biased the results of the previous sub -sections. In Korea, the information content of M2 disappears during the 1980s, precisely after M2 is selected as an intermediate target. ²⁰/ In the Philippines, base money was targeted starting in 1984, yet the results of Table 10 show that during the 1980s all financial aggregates except base money contained advance information on inflation.

V. The Information Content of the Exchange and Interest Rates

As mentioned earlier, financial sector reform in Indonesia, Korea, and the Philippines entailed greater exchange rate flexibility, liberalization of interest rates, and the emergence of new financial instruments in all three countries. As a result, both exchange and interest rates could be expected to play an increasingly important role in the transmission process. Moreover, greater flexibility in the foreign exchange market, and the development of money and capital market may render both rates important leading indicators of output and price developments. This section focuses on the information content of exchange and interest rates. As mentioned earlier, the methodology used here focuses on the marginal predictive content of the variables; that is, on whether these variables contain Information not already provided by the other variables In the regression equations. Thus, if found to contain valuable information, both rates could also be used in the exercise of monetary policy, especially because they are usually available with a shorter lag than is the case for monetary or credit aggregates.

As before, the information content of interest rates and exchange rates is examined on the basis of F exclusion tests performed on an extended version of equations (l)-(3) that includes as explanatory variables four lags of interest rates and exchange rates. The interest rates used were the money market rate for Indonesia and Korea, and the Treasury bill rate for the Philippines. Data for the interest rate series were available for a shorter period of time (roughly the 1980s), which prevented a comparison of the effects of introducing interest rate variables in the reduced-form system described by equations (l)-(3) before and after the reforms. ²¹/

The results of the tests are shown in Table 7.²²/ In general, the introduction of lagged values of interest rates and exchange rates did not have much of an effect on the output equations. The main results for the output equations can be summarized as follows: (i) the information content of money was not enhanced by the introduction of exchange and interest rates; and (ii) neither exchange rates nor interest rates were found significant in any of the output equations.

Table 7.

The Information Content of Honey, Interest Rates, and Exchange Rates in the Price and Output Equations

Table 7.

The Information Content of Honey, Interest Rates, and Exchange Rates in the Price and Output Equations

I. Price Equation				II. Output Equation
	Indonesia	Korea	Philippines		Indonesia	Korea	Philippines
(A) Ho: Coefficients on money = 0				(A) Ho: Coefficients on money = 0
Currency	0.52	0.91	1.75	Currency	1.12	0.59	0.36
Monetary base	0.79	1.07	1.58	Monetary base	1.48	0.55	0.68
M1	0.62	1.21	1.32	M1	0.55	0.78	0.41
M2	1.97	1.82	0.59	M2	1.04	1.41	0.14
Domestic credit	1.24	1.08	0.79	Domestic credit	0.51	1.54	0.55
52 critical value	*2.72*	*2.67*	*2.64*	52 critical value	*2.72*	*2.67*	*2.64*
(B) Ho: Coefficients on interest rates = 0				(B) Ho: Coefficients on interest rates = 0
Currency	0.26	0.28	2.75*	Currency	1.77	0.23	1.07
Monetary base	0.34	0.62	3.19 *	Monetary base	0.63	0.04	1.13
M1	0.20	0.21	1.96	M1	0.91	0.11	0.87
M2	0.46	0.23	1.48	M2	1.55	0.28	1.20
Domestic credit	0.75	0.27	1.13	Domestic credit	1.70	0.13	1.45
52 critical value	*2.72*	*2.67*	*2.64*	52 critical value	*2.72*	*2.67*	*2.64*
(C) Ho: Coefficients on exchange rates = 0				(C) Ho: Coefficients on exchange rates = 0
Currency	0.97	2.65	8.75*	Currency	2.26	1.32	0.86
Monetary base	0.70	2.84*	12.26 *	Monetary base	2.34	1.04	0.56
M1	1.32	2.97*	13.14 *	M1	1.34	1.56	0.64
M2	1.03	3.07*	11.87*	M 2	1.87	2.41	0.82
Domestic credit	0.59	3.31*	11.64*	Domestic credit	1.86	2.37	0.77
*52 critical value*	*2.72*	*2.67*	*2.64*	52 critical value	*2.72*	*2.67*	*2.64*

Table 7.

The Information Content of Honey, Interest Rates, and Exchange Rates in the Price and Output Equations

I. Price Equation				II. Output Equation
	Indonesia	Korea	Philippines		Indonesia	Korea	Philippines
(A) Ho: Coefficients on money = 0				(A) Ho: Coefficients on money = 0
Currency	0.52	0.91	1.75	Currency	1.12	0.59	0.36
Monetary base	0.79	1.07	1.58	Monetary base	1.48	0.55	0.68
M1	0.62	1.21	1.32	M1	0.55	0.78	0.41
M2	1.97	1.82	0.59	M2	1.04	1.41	0.14
Domestic credit	1.24	1.08	0.79	Domestic credit	0.51	1.54	0.55
52 critical value	*2.72*	*2.67*	*2.64*	52 critical value	*2.72*	*2.67*	*2.64*
(B) Ho: Coefficients on interest rates = 0				(B) Ho: Coefficients on interest rates = 0
Currency	0.26	0.28	2.75*	Currency	1.77	0.23	1.07
Monetary base	0.34	0.62	3.19 *	Monetary base	0.63	0.04	1.13
M1	0.20	0.21	1.96	M1	0.91	0.11	0.87
M2	0.46	0.23	1.48	M2	1.55	0.28	1.20
Domestic credit	0.75	0.27	1.13	Domestic credit	1.70	0.13	1.45
52 critical value	*2.72*	*2.67*	*2.64*	52 critical value	*2.72*	*2.67*	*2.64*
(C) Ho: Coefficients on exchange rates = 0				(C) Ho: Coefficients on exchange rates = 0
Currency	0.97	2.65	8.75*	Currency	2.26	1.32	0.86
Monetary base	0.70	2.84*	12.26 *	Monetary base	2.34	1.04	0.56
M1	1.32	2.97*	13.14 *	M1	1.34	1.56	0.64
M2	1.03	3.07*	11.87*	M 2	1.87	2.41	0.82
Domestic credit	0.59	3.31*	11.64*	Domestic credit	1.86	2.37	0.77
*52 critical value*	*2.72*	*2.67*	*2.64*	52 critical value	*2.72*	*2.67*	*2.64*

The introduction of lagged values of interest rates and exchange rates did change the picture in the price equations, however. Exchange rates were found significant independently of the monetary aggregate being used in Korea and the Philippines, especially in the latter were exchange rates appear to contain substantial information on the future behavior of prices. Interest rates, on the other hand, were significant only in the Philippines and only when narrow monetary aggregates were used (currency or base money). Nevertheless, when interest rates were found significant the information content of money diminished noticeably, a result found by many studies for developed countries.

VI. Summary and Conclusions

In a recent paper Friedman and Kuttner have argued that, independently of how monetary policy is conducted (that is, either committing to an intermediate target, monitoring monetary developments as part of a set of information variables used in the formulation of feedback rules, or a combination of the two), money can be a helpful variable to policymakers if it contains information on future developments in output and prices, the usual variables associated with the policymaker’s ultimate objectives. This paper follows their approach and focuses on the information content of money in a group of Pacific Basin countries that underwent financial sector reform during the 1980s: Indonesia, Korea, and the Philippines. The uncertainty about the effects on financial sector reform on the relationship between money, income and prices and, more specifically on the information content of money is an Issue commonly faced by policymakers in countries undergoing financial sector reform.

The results obtained in this paper are varied and warn against the use of generalizations. In Korea money was found to contain valuable advance information on output and prices; in the Philippines money was found to contain information only on the future behavior of prices; and in Indonesia money was found not to have any advance information. A common result obtained from the tests is that in those cases where some financial aggregate was found to be a significant leading indicator of output or price behavior, a measure of money, narrow or broad, outperformed credit in terms of information content.

Different results were also obtained from the various tests on the effects of financial sector reform on the information content of money. The information content of financial aggregates practically disappeared after the reform in Korea, but was enhanced in the Philippines; money consistently lacked information content in Indonesia. Further tests were conducted in order to assess the information content of exchange and interest rates, variables that gained flexibility with the reforms. Exchange rates were found to contain valuable information about future developments in prices in Korea and the Philippines. Interest rates, on the other hand, were found to be significant only in the Philippines; however, in the presence of interest rates no monetary aggregate was significant. These results suggest strongly that exchange rates could be valuable indicators or guides of monetary policy in those countries. Since daily information on exchange rates is usually available almost immediately, a more detailed analysis of the information content of exchange rates with lower frequency data should be carried out, in order to assess the usefulness of exchange rates in the daily or weekly operations of monetary policy.

Finally, a few words of caution on the VAR methodology used here. Recently, several studies have proven the relative importance of including the long-term relationship between the relevant variables via error correction terms (for example, Caramazza and Slawner (1991)). We did not proceed along those lines for a practical reason: three variable systems using five variables each time for money would have required a very large set of cointegration tests for two or three variables and for the corresponding error correction representations. Implicitly we have assumed that those costs outweigh the possible benefits.

APPENDIX I Data Sources

All series used in this paper were taken from the International Monetary Fund’s International Financial Statistics. The series and their corresponding line numbers are:

Except for the real GDP series, all series were available quarterly for the periods 1968:1 to 1991:4 (Indonesia), 1970:1 to 1991:4 (Korea) and 1963:1 to 1991:4 (Philippines). For the monetary variables the beginning and end of period figures were averaged. Price figures are period averages. A series of quarterly real GDP was obtained for Indonesia and the Philippines by interpolating the annual figures using the following quarterly related series: volume of exports and real imports (Indonesia), and volumes of exports and imports (the Philippines).

The series were not seasonally adjusted, but all regressions included seasonal dummies. The estimations were carried out using TSP, version 4.2. Since the program does not generate automatically significance lines for the CUSUM tests, these were computed using the formulae suggested by Harvey (1989), p. 257.

APPENDIX II

APPENDIX III

Table 8.

Indonesia: Tests of Parameter Stability

Table 8.

Indonesia: Tests of Parameter Stability

Financial Aggregate	Price equation			Output equation			Muaey equat i
	(A) Ho: Coefficients on money = 0			(A) Ho : Coefficients on inflation = 0			(A) Ho: Coefficients on inflation = 0
	69:3-91:4	69:3-75:4	80:1-91:4	69:3-91:4	69:3-79:4	80:1-91:4	69:3-91:4	63:3-75:4	80:1-91:4
Cur rency	1.92	1.12	0.38	0.55	1.11	0.52	1.00	0.00	0.22
Monetary Dace	1.11	0.85	0.55	0.96	0.45	0.67	0.63	0.62	0.33
M1	0.91	1.50	0.10	1.16	1.06	0.88	0.91	0.99	1.93
M2	1.13	0.75	1.57	1.33	0.89	0.78	0.95	0.39	2.45
Domestic credit	1.12	0.86	1.50	1.49	0.69	0.75	0.50	1.31	0.68
52 critical value	2.51	2.75	2.67	2.51	2.75	2.67	2.51	2.75	2.67
	(B) Ho: Coefficients on output = 0			(B) Ho: Coefficients on money = 0			(B) Ho : Coefficients on output = 0
	69:3-91:4	69:3-79:4	80:1-91:4	69:3-91:4	69:3-79:4	80:1-91:4	69:3-91:4	69:3-79:4	80:1-91:4
Currency	0.2 3	0.28	0.94	0.32	1.47	0.42	0.44	0.75	0.78
Monetary Base	0.24	0.44	0.46	0.94	1.40	2.08	1.02	0.84	0.83
M1	0.19	0.52	0.81	0.66	1.04	1.61	1.14	0.37	1.48
M2	0.17	0.55	0.61	0.98	1.05	1.10	0.48	0.21	0.88
Domestic credit	0.29	0.24	1.10	1.02	0.52	0.21	2.90*	4.52*	0.31
52 critical value	*2.51*	*2.75*	*2.67*	*2.51*	*2.75*	*2.67*	*2.51*	*2.75*	*2.67*
	(C) Ho: Block exogeneity			(C) Ho: Block exogeneity			(C) Ho: Block exogentity
	69:3-91:4	69:3-79:4	80:1-91:4	69:3-91:4	69:3-79:4	80:1-91:4	69:3-91:4	69:3-79:4	80:1-91:4
Currency	1.01	0.62	0.62	0.72	0.98	0.65	0.69	0.82	0.51
Monetary Base	0.60	0.48	0.72	1.04	0.95	1.56	0.79	0.74	0.78
M1	0.50	0.81	0.47	0.90	0.75	1.30	0.98	0.67	1.45
M2	0.61	0.43	1.28	1.07	0.76	1.03	0.63	0.28	1.43
Domestic credit	0.60	0.48	1.24	1.08	0.48	0.53	1.53	2.47*	0.66
52 critical value	*2.08*	*2.33*	*2.25*	*2.08*	*2.33*	*2.25*	*2.08*	*2.33*	*2.25*

Table 8.

Indonesia: Tests of Parameter Stability

Financial Aggregate	Price equation			Output equation			Muaey equat i
	(A) Ho: Coefficients on money = 0			(A) Ho : Coefficients on inflation = 0			(A) Ho: Coefficients on inflation = 0
	69:3-91:4	69:3-75:4	80:1-91:4	69:3-91:4	69:3-79:4	80:1-91:4	69:3-91:4	63:3-75:4	80:1-91:4
Cur rency	1.92	1.12	0.38	0.55	1.11	0.52	1.00	0.00	0.22
Monetary Dace	1.11	0.85	0.55	0.96	0.45	0.67	0.63	0.62	0.33
M1	0.91	1.50	0.10	1.16	1.06	0.88	0.91	0.99	1.93
M2	1.13	0.75	1.57	1.33	0.89	0.78	0.95	0.39	2.45
Domestic credit	1.12	0.86	1.50	1.49	0.69	0.75	0.50	1.31	0.68
52 critical value	2.51	2.75	2.67	2.51	2.75	2.67	2.51	2.75	2.67
	(B) Ho: Coefficients on output = 0			(B) Ho: Coefficients on money = 0			(B) Ho : Coefficients on output = 0
	69:3-91:4	69:3-79:4	80:1-91:4	69:3-91:4	69:3-79:4	80:1-91:4	69:3-91:4	69:3-79:4	80:1-91:4
Currency	0.2 3	0.28	0.94	0.32	1.47	0.42	0.44	0.75	0.78
Monetary Base	0.24	0.44	0.46	0.94	1.40	2.08	1.02	0.84	0.83
M1	0.19	0.52	0.81	0.66	1.04	1.61	1.14	0.37	1.48
M2	0.17	0.55	0.61	0.98	1.05	1.10	0.48	0.21	0.88
Domestic credit	0.29	0.24	1.10	1.02	0.52	0.21	2.90*	4.52*	0.31
52 critical value	*2.51*	*2.75*	*2.67*	*2.51*	*2.75*	*2.67*	*2.51*	*2.75*	*2.67*
	(C) Ho: Block exogeneity			(C) Ho: Block exogeneity			(C) Ho: Block exogentity
	69:3-91:4	69:3-79:4	80:1-91:4	69:3-91:4	69:3-79:4	80:1-91:4	69:3-91:4	69:3-79:4	80:1-91:4
Currency	1.01	0.62	0.62	0.72	0.98	0.65	0.69	0.82	0.51
Monetary Base	0.60	0.48	0.72	1.04	0.95	1.56	0.79	0.74	0.78
M1	0.50	0.81	0.47	0.90	0.75	1.30	0.98	0.67	1.45
M2	0.61	0.43	1.28	1.07	0.76	1.03	0.63	0.28	1.43
Domestic credit	0.60	0.48	1.24	1.08	0.48	0.53	1.53	2.47*	0.66
52 critical value	*2.08*	*2.33*	*2.25*	*2.08*	*2.33*	*2.25*	*2.08*	*2.33*	*2.25*

Table 9.

Korea: Tests of Parameter Stability

Table 9.

Korea: Tests of Parameter Stability

Financial Aggregate	Price equation			Output equation			Money equation
	(A) Ho: Coefficients on money = 0			(A) Ho: Coefficients on inflation = 0			(A) Ho: Coefficients on inflation = 0
	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4
Currency	3.73*	6.64*	0.85	2.33	1.09	1.42	0.93	2.93*	0.44
Monetary Base	2.18	2.15	0.93	2.11	0.90	1.24	0.04	0.42	0.10
M1	1.31	3.68*	0.74	2.SO	0.49	1.33	0.17	1.47	0.52
M2	3.46*	2.13	1.54	2.88*	0.76	1.47	0.44	1.17	1.14
Dosestic credit	3.42*	1.01	0.99	1.84	1.04	0.91	3.03*	2.10	0.62
52 critical value	*2.52*	*2.90*	*2.67*	*2.52*	*2.90*	*2.67*	*2.52*	*2.90*	*2.67*
	(B) Ho: Coefficients on output = 0			(B) Ho: Coefficients on money = 0			(B) Ho: Coefficients on ontput = 0
	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4
Currency	2.02	1.30	0.82	1.12	0.84	0.54	0.48	0.22	0.10
Monetary Base	1.87	1.37	0.72	2.74*	2.03	0.58	1.45	0.74	0.19
M1	1.42	2.08	0.41	2.87*	1.39	0.94	1.40	1.59	0.48
M2	1.66	0.69	0.75	2.61*	3.20*	0.42	1.39	2.01	0.43
Domestic credit	2.01	0.61	0.27	0.92	0.50	1.08	1.34	0.60	0.67
52 critical value	2.52	2.90	2.67	2.52	2.90	2.67	2.52	2.90	2.67
	(C) Ho: Block exogeneity			(C) Ho: Block exogeneity			(C) Ho: Block exogeneity
	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4
Currency	2.66*	4.12*	0.74	1.54	0.99	0.99	0.54	1.87	0.32
Monetary Base	1.83	1.59	0.79	2.44*	1.71	1.02	0.78	0.53	0.33
M1	1.36	2.47	0.69	2.51*	1.32	1.23	0.83	1.42	0.50
M2	2.52*	1.58	1.12	2.36*	2.41	0.92	1.11	1.50	1.09
Domestic cr ed11	2.50*	0.94	0.82	1.43	0.79	1.31	1.84	1.43	0.90
52 critical value	*2.09*	*2.48*	*2.25*	*2.09*	*2.48*	*2.25*	*2.09*	*2.48*	*2.25*

Table 9.

Korea: Tests of Parameter Stability

Financial Aggregate	Price equation			Output equation			Money equation
	(A) Ho: Coefficients on money = 0			(A) Ho: Coefficients on inflation = 0			(A) Ho: Coefficients on inflation = 0
	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4
Currency	3.73*	6.64*	0.85	2.33	1.09	1.42	0.93	2.93*	0.44
Monetary Base	2.18	2.15	0.93	2.11	0.90	1.24	0.04	0.42	0.10
M1	1.31	3.68*	0.74	2.SO	0.49	1.33	0.17	1.47	0.52
M2	3.46*	2.13	1.54	2.88*	0.76	1.47	0.44	1.17	1.14
Dosestic credit	3.42*	1.01	0.99	1.84	1.04	0.91	3.03*	2.10	0.62
52 critical value	*2.52*	*2.90*	*2.67*	*2.52*	*2.90*	*2.67*	*2.52*	*2.90*	*2.67*
	(B) Ho: Coefficients on output = 0			(B) Ho: Coefficients on money = 0			(B) Ho: Coefficients on ontput = 0
	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4
Currency	2.02	1.30	0.82	1.12	0.84	0.54	0.48	0.22	0.10
Monetary Base	1.87	1.37	0.72	2.74*	2.03	0.58	1.45	0.74	0.19
M1	1.42	2.08	0.41	2.87*	1.39	0.94	1.40	1.59	0.48
M2	1.66	0.69	0.75	2.61*	3.20*	0.42	1.39	2.01	0.43
Domestic credit	2.01	0.61	0.27	0.92	0.50	1.08	1.34	0.60	0.67
52 critical value	2.52	2.90	2.67	2.52	2.90	2.67	2.52	2.90	2.67
	(C) Ho: Block exogeneity			(C) Ho: Block exogeneity			(C) Ho: Block exogeneity
	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4	71:2-91:4	71:2-79:4	80:1-91:4
Currency	2.66*	4.12*	0.74	1.54	0.99	0.99	0.54	1.87	0.32
Monetary Base	1.83	1.59	0.79	2.44*	1.71	1.02	0.78	0.53	0.33
M1	1.36	2.47	0.69	2.51*	1.32	1.23	0.83	1.42	0.50
M2	2.52*	1.58	1.12	2.36*	2.41	0.92	1.11	1.50	1.09
Domestic cr ed11	2.50*	0.94	0.82	1.43	0.79	1.31	1.84	1.43	0.90
52 critical value	*2.09*	*2.48*	*2.25*	*2.09*	*2.48*	*2.25*	*2.09*	*2.48*	*2.25*

Table 10.

Philippines: Tests of Parameter Stability

Table 10.

Philippines: Tests of Parameter Stability

Financial Aggregate	Price equation			Output equation			Money equation
	(A) Ho: Coefficients on money = 0			(A) Ho: Coefficients on inflation = 0			(A) Ho: Coefficients on inflation = 0
	64:3-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4
Currency	7.05*	2.14	4.97*	1.62	0.40	1.02	0.29	1.03	1.15
Monetary Base	3.34*	0.45	2.11	2.09	0.32	1.95	1.02	0.83	1.06
M1	2.01	2.53	2.90*	3.36*	0.74	2.23	1.01	0.79	3.01*
M2	2.78*	0.42	4.41*	1.53	0.21	0.97	0.49	0.18	0.75
Domestic credit	1.27	0.18	3.32*	2.28	0.51	0.53	0.24	0.49	0.20
52 critical value	2.48	2.63	2.67	2.48	2.63	2.67	2.48	2.63	2.67
	(B) Ho: Coefficients on output = 0			(B) Ho: Coefficients on money = 0			(B) Ho: Coefficients on ontput = 0
	71:2-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4
Currency	0.35	0.28	0.63	0.35	0.34	1.46	0.53	2.38	0.41
Monetary Base	0.35	0.10	0.68	1.05	1.40	2.13	0.32	0.85	1.50
M1	0.32	0.36	0.89	2.09	2.23	2.30	0.37	1.33	0.37
M2	0.24	0.23	0.97	0.71	2.39	0.91	1.23	1.22	1.40
Domestic credit	0.41	0.18	0.90	0.70	0.26	0.45	0.42	0.27	0.15
52 critical value	*2.48*	*2.63*	*2.67*	*2.48*	*2.63*	*2.67*	*2.48*	*2.63*	*2.67*
	(C) Ho: Block exogeneity			(C) Ho: Block exogeneity			(C) Ho: Block exogeneity
	64:3-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4
Currency	3.77*	1.19	3.02*	1.27	0.37	1.27	0.41	1.72	0.84
Monetary Base	1.89	0.33	1.47	1.65	0.92	1.65	0.70	0.78	1.58
M1	1.21	1.39	1.90	2.22*	1.35	1.74	0.64	1.20	1.87
M2	1.60	0.31	2.72*	1.46	1.43	0.97	0.92	0.72	1.30
Domestic credit	0.83	0.19	2.13	1.46	0.33	0.71	0.30	0.37	0.18
52 critical value	*2.05*	*2.16*	*2.25*	*2.05*	*2.16*	*2.25*	*2.05*	*2.16*	*2.25*

Table 10.

Philippines: Tests of Parameter Stability

Financial Aggregate	Price equation			Output equation			Money equation
	(A) Ho: Coefficients on money = 0			(A) Ho: Coefficients on inflation = 0			(A) Ho: Coefficients on inflation = 0
	64:3-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4
Currency	7.05*	2.14	4.97*	1.62	0.40	1.02	0.29	1.03	1.15
Monetary Base	3.34*	0.45	2.11	2.09	0.32	1.95	1.02	0.83	1.06
M1	2.01	2.53	2.90*	3.36*	0.74	2.23	1.01	0.79	3.01*
M2	2.78*	0.42	4.41*	1.53	0.21	0.97	0.49	0.18	0.75
Domestic credit	1.27	0.18	3.32*	2.28	0.51	0.53	0.24	0.49	0.20
52 critical value	2.48	2.63	2.67	2.48	2.63	2.67	2.48	2.63	2.67
	(B) Ho: Coefficients on output = 0			(B) Ho: Coefficients on money = 0			(B) Ho: Coefficients on ontput = 0
	71:2-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4
Currency	0.35	0.28	0.63	0.35	0.34	1.46	0.53	2.38	0.41
Monetary Base	0.35	0.10	0.68	1.05	1.40	2.13	0.32	0.85	1.50
M1	0.32	0.36	0.89	2.09	2.23	2.30	0.37	1.33	0.37
M2	0.24	0.23	0.97	0.71	2.39	0.91	1.23	1.22	1.40
Domestic credit	0.41	0.18	0.90	0.70	0.26	0.45	0.42	0.27	0.15
52 critical value	*2.48*	*2.63*	*2.67*	*2.48*	*2.63*	*2.67*	*2.48*	*2.63*	*2.67*
	(C) Ho: Block exogeneity			(C) Ho: Block exogeneity			(C) Ho: Block exogeneity
	64:3-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4	64:3-91:4	64:3-79:4	80:1-91:4
Currency	3.77*	1.19	3.02*	1.27	0.37	1.27	0.41	1.72	0.84
Monetary Base	1.89	0.33	1.47	1.65	0.92	1.65	0.70	0.78	1.58
M1	1.21	1.39	1.90	2.22*	1.35	1.74	0.64	1.20	1.87
M2	1.60	0.31	2.72*	1.46	1.43	0.97	0.92	0.72	1.30
Domestic credit	0.83	0.19	2.13	1.46	0.33	0.71	0.30	0.37	0.18
52 critical value	*2.05*	*2.16*	*2.25*	*2.05*	*2.16*	*2.25*	*2.05*	*2.16*	*2.25*

References

Beltas, A. and T. Jones, “Money, Inflation and Causality in a Financially Repressed Economy: Algeria, 1970-88,” Applied Economics, Vol. 25 (1993), pp. 473–480.
- Search Google Scholar
- Export Citation
Bernanke, B. and A, Blinder, “Credit, Money, and Aggregate Demand,” American Economic Review, Vol. 78, No. 2 (1988), pp. 435–439.
- Search Google Scholar
- Export Citation
Blinder, A., “Credit Rationing and Effective Supply Failures,” Economic Journal, Vol. 97 (1987), pp. 327–52.
- Search Google Scholar
- Export Citation
Bisat, A., R.B. Johnston and V. Sundararajan, “Issues in Managing and Sequencing Financial Sector Reforms, Lessons from Experiences in Five Developing Countries,” IMF Working Paper, WP/92/82 (Washington: International Monetary Fund, October 1992).
- Search Google Scholar
- Export Citation
Brunner, K, and A. Meltzer, “Money and Credit in the Monetary Transmission Process,” American Economic Review, Vol. 78, No. 2 (1988), pp. 446–451.
- Search Google Scholar
- Export Citation
Cagan, P., “Money-Income Causality: A Critical Review of Literature Since the Monetary History,” Columbia University Discussion Paper Series, No. 373, January 1988.
- Search Google Scholar
- Export Citation
Cagan, P., “The Choice among Monetary Aggregates as Targets and Guides for Monetary Policy,” Journal of Money, Credit and Banking, Vol. 14, No. 4, (Part 2) (1992), pp. 661–86.
- Search Google Scholar
- Export Citation
Caramazza, F. and C. Slawner, “The Relationship Between Money, Income and Prices,” Bank of Canada, IMF Working Paper, WP/91/4 (Washington: International Monetary Fund, 1991).
- Search Google Scholar
- Export Citation
Davis, R., “Broad Credit Measures as Targets for Monetary Policy,” Federal Reserve Bank of New York Quarterly Review, Summer 1979, pp. 13–22.
- Search Google Scholar
- Export Citation
Dhakal, D. and M. Kandil, “The Inflationary Experiences of Six Developing Countries in Asia: An Investigation of Underlying Determinants,” Applied Economics, Vol. 25 (1993), pp. 413–25.
- Search Google Scholar
- Export Citation
Fellner, W., “Criteria for Useful Targeting: Money versus the Base and Other Variables,” Journal of Money, Credit and Banking, Vol. 14, No. 4, (Part 2) (1982), pp. 641–660.
- Search Google Scholar
- Export Citation
Friedman, B. and K. Kuttner, “Money, Income, and Interest Rates,” American Economic Review, Vol. 82, No. 3 (1992), pp. 472–492.
- Search Google Scholar
- Export Citation
Fry, M., Money, Interest, and Banking in Economic Development (Baltimore: The Johns Hopkins University Press, 1988).
- Search Google Scholar
- Export Citation
Ghatak, S. and D. Deadman, “Money, Prices and Stabilization Policies in Some Developing Countries,” Applied Economics, Vol. 21 (1989) pp. 953–65.
- Search Google Scholar
- Export Citation
Goldsbrough, D. and I. Zaidi, “Monetary Policy in the Philippines During Periods of Financial Crisis and Changes in Exchange Rate Regime: Targets, Instruments, and the Stability of Money Demand,” IMF Working Paper, WP/89/98 (Washington: International Monetary Fund, December 1989).
- Search Google Scholar
- Export Citation
Harvey, A., Forecasting, Structural Time Series Models and the Kalman Filter, (Cambridge: University Press, 1989).
- Search Google Scholar
- Export Citation
Kareken, J., T. Muench and N. Wallace, “Optimal Open Market Strategy: The Use of Information Variables,” American Economic Review, Vol. 63, No. 1 (1973), pp. 156–72.
- Search Google Scholar
- Export Citation
Kamath, S., “Monetary Aggregates, Income and Causality in a Developing Economy,” Journal of Economic Studies, Vol. 12, No. 3 (1985), pp. 36–53.
- Search Google Scholar
- Export Citation
Khan, M. and V. Sundararajan, “Financial Sector Reforms and Monetary Policy,” IMF Working Paper, WP/91/127 (Washington: International Monetary Fund, December 1991).
- Search Google Scholar
- Export Citation
Kopcke, R., “Must the Ideal “Money Stock” Be Controllable?,” New England Economic Review (March/April 1983), pp. 10–23.
- Search Google Scholar
- Export Citation
Lim, J., “The New Structuralist Critique of the Monetarist Theory of Inflation: The Case of the Philippines,” Journal of Development Economics, Vol. 25 (1987), pp. 45–61.
- Search Google Scholar
- Export Citation
Parikh, A., “Money Supply and Prices in Indonesia 1969-1980: An Econometric Investigation,” Applied Economics, Vol. 22 (1990), pp. 1479–94.
- Search Google Scholar
- Export Citation
Sims, C., “Comparison of Interwar and Postwar Business Cycles: Monetarism Reconsidered,” American Economic Review, Vol. 70 (September 1980), pp. 250–257.
- Search Google Scholar
- Export Citation
Sims, C., “Money, Income, and Causality”, American Economic Review, Vol. 62, (September 1972), pp. 540–52.
- Search Google Scholar
- Export Citation
Stock, J. and M. Watson, “Interpreting the Evidence on Money-Income Causality,” Journal of Econometrics, Vol. 40 (1989), pp. 161–182.
- Search Google Scholar
- Export Citation
Sundararajan V. and L. Molho, “Financial Reform and Monetary Control in Indonesia.” IMF Working Paper, WP/88/4 (Washington: International Monetary Fund, January 1988).
- Search Google Scholar
- Export Citation
Togan, S., “The Influence of Money and the Rate of Interest on the Rate of Inflation in a Financially Repressed Economy: The Case of Turkey,” Applied Economics, Vol. 19 (1987), pp. 1585–1601.
- Search Google Scholar
- Export Citation
Tseng, W, and R. Corker, “Financial Liberalization, Money Demand, and Monetary Policy in Asian Countries,” IMF Occasional Paper No. 84 (Washington: International Monetary Fund, July 1991).
- Search Google Scholar
- Export Citation
Yang, Y., “Causality Between Money, Interest Rates and Prices in Taiwan: a Multivariate Time-Series Analysis,” Applied Economics, Vol. 22 (1990), pp. 1739–1749.
- Search Google Scholar
- Export Citation

I have benefitted from valuable discussions held with Francesco Caramazza at different stages of the preparation of this paper. I am also indebted to David Bigman, Carlo Cottarelli, Norman Fieleke, Jang-Yung Lee, Miguel Savastano, and participants in a MAE seminar for helpful comments. All remaining errors are my own responsibility.

It is assumed throughout this paper, that monetary policy is conducted by targeting one or more variables and monitoring the development of a number of other relevant variables in order to formulate feedback rules based on new information as it becomes available. This strategy lies somewhere between the “pure intermediate target” approach proposed by Fellner (1982), and the no-target, “information variables” approach described by Kareken et al. (1973).

See, for example, Sims (1972), (1980) and Stock and Watson (1989). For a critical survey of the literature, see Cagan (1988).

See, for example, Beltas and Jones (1993) for an application to Algeria; Kamath (1985) for India; Parikh (1990) for Indonesia; Togan (1987) for Turkey; and Yang (1990) for Taiwan.

The experiences of the three countries studied here with financial sector reform are discussed in detail by Tseng and Corker (1991), and Bisat et al. (1992). This section is based on their work.

Notice, however, that during the 1960s the Korean won was devalued several times in discrete steps, through 1974. From then on, the rate was pegged to the US dollar until January of 1980.

Sundararajan and Molho (1988) provide a detailed description of the conduct of monetary policy in Indonesia during most of the 1980s.

See Goldsbrough and Zaidi (1989), section I, for a description of the evolution of monetary policy through 1988 in the Philippines.

A detailed description of the data set is provided in Appendix I.

10/

In all cases, the ADF focused on the t ratio of the coefficient 7 in the following regression:

Δ x_{t} = α + β T + γ x_{t - 1} + Σ_{i = 1}^{k} Δ x_{t - i} + μ_{t}

where T represents a linear trend variable and k was selected as the number of lags of the dependent variable that minimizes the Schwartz Information Criterion.

11/

Despite this result, we proceeded to include those variables in the regressions for illustrative purposes. In the case of MB in Korea, however, the t-ratio is very close to the critical value, and its value was very sensitive to the lag structure in the ADF equations.

12/

Given the lack of consensus on a rule for determining lag lengths in a multivariate VAR, we used 4 as a uniform lag length, following a common practice in this area.

13/

Although a useful “information” variable must contain information about the future behavior of prices and/or output, the same need not be true about the intermediate target variables. On this issue, see Cagan (1982).

14/

For convenience, henceforth the term “money” will be used in general to refer to any of the following aggregates: currency outside banks, base money, M1, M2, and domestic credit by the banking system. A more specific notation will be used when needed.

15/

This is done only as a way of summarizing the results of the tests. As mentioned earlier, the issue of causality has no bearing in deciding whether money contains or not useful information for policymakers.

16/

The terms “checking” and “saving” deposits are being used loosely to mean, respectively, all deposits included in M1 and all deposits included in M2 but not in M1.

17/

Reserve requirements were lowered from 20-27 percent in the late 1970s to 10-20 percent in 1980, 5.5 percent in 1981, 4.5 percent in 1984, and were raised later to 10 percent in 1988; marginal reserve requirements were introduced in 1989 (Tseng and Corker (1991), p. 39).

18/

The test consists on determining whether a plot of the cumulative sum of the prediction errors (CUSUM) remains within boundaries determined by the sample size and the level of significance. If the CUSUM plot crosses one of the boundaries the hypothesis of parameter constancy is rejected. The test is essentially descriptive as it does not determine accurately the timing of the structural break, A detailed description of the test can be found in Harvey (1989), pp. 256-8.

19/

The decline in the information content of M2 in Korea may be related to the fast increase in the relative importance of nonbank deposits during the 1980s.

20/

In Korea M2 was formally chosen as an intermediate target of monetary policy at the end of 1979.

21/

A detailed description of the exchange and Interest rate data, is provided in Appendix I.

22/

Interest rates were found to be stationary in levels in Indonesia and the Philippines but difference stationary in Korea; as a result they are included in the regressions in levels in the first two countries and in first differences in Korea. Exchange rates were found to be difference stationary in all three countries.

23/

Available from 1978:1.

24/

Available from 1976:4.

25/

Available from 1976:1.

Other Publishers

Save
Cite
Email this content

Share Link

Copy this link, or click below to email it to a friend
Email this content
or copy the link directly:

https://www.elibrary.imf.org/view/journals/001/1993/088/article-A001-en.xml
The link was not copied. Your current browser may not support copying via this button.

Link copied successfully

Financial Liberalization and the Information Content of Money in Indonesia, Korea, and the Philippines

Author: Mr. A. J Hamann

Volume/Issue:: Volume 1993: Issue 088

Publisher:: International Monetary Fund

ISBN:: 9781451851083

ISSN:: 1018-5941

Pages:: 44

DOI:: https://doi.org/10.5089/9781451851083.001

Keywords
I. Introduction
II. Financial Sector Reform and the Relationship Between Money, Output, and Prices
III. Financial Liberalization in Indonesia. Korea, and the Philippines
IV. Money, Output, and Prices in Indonesia, Korea, and the Philippines
V. The Information Content of the Exchange and Interest Rates
VI. Summary and Conclusions
APPENDIX I Data Sources
APPENDIX II
APPENDIX III
- References

			I. Price Equation
		(A) Estimation results, 1971:2-1991:4
Financial aegregate			${\bar{R}}^{2}$		Log LF	RMSE
Currency			0.55*		235.70*	0.0161*
M2			0.54		235.15	0.0170
Credit			0.54		235.06	0.0169
		(B) Ex-post forecasting RMSFE
				forecasting period
	1987:1=1991:4		1988:1=1991:4		1989:1-1991:4	1990:1=1991:4
Financial aggregate
Currency		0.0110		0,0091	0.0094	0.0106
M2		0.0101*		0.0086*	0.0083*	0.0086*
Credit		0.0123		0.0120	0.0112	0.0090
			II. Output Equation
		(A) Estimation results, 1971:2-1991:4
Financial aseregate			${\bar{R}}^{2}$		Log LF	RMSE
MB			0.98		164.16	0.0384
M1			0.98*		164.43*	0.0368*
M2			0.98		163.89	0.0393
		(B) Ex-post forecasting RMSFE
			forecasting period
	1987:1-1991:4		1988:1-1991:4		1989:1-1991:4	1990:1-1991:4
Financial aggregate
MB		0.0562		0.0633	0.0723	0.0532
M1		0.0472		0.0554*	0.0674	0.0551
M2		0.0442*		0.0593	0.0625*	0.0522*

			I. Price Equation
		(A) Estimation results, 1971:2-1991:4
Financial aegregate			${\bar{R}}^{2}$		Log LF	RMSE
Currency			0.55*		235.70*	0.0161*
M2			0.54		235.15	0.0170
Credit			0.54		235.06	0.0169
		(B) Ex-post forecasting RMSFE
				forecasting period
	1987:1=1991:4		1988:1=1991:4		1989:1-1991:4	1990:1=1991:4
Financial aggregate
Currency		0.0110		0,0091	0.0094	0.0106
M2		0.0101*		0.0086*	0.0083*	0.0086*
Credit		0.0123		0.0120	0.0112	0.0090
			II. Output Equation
		(A) Estimation results, 1971:2-1991:4
Financial aseregate			${\bar{R}}^{2}$		Log LF	RMSE
MB			0.98		164.16	0.0384
M1			0.98*		164.43*	0.0368*
M2			0.98		163.89	0.0393
		(B) Ex-post forecasting RMSFE
			forecasting period
	1987:1-1991:4		1988:1-1991:4		1989:1-1991:4	1990:1-1991:4
Financial aggregate
MB		0.0562		0.0633	0.0723	0.0532
M1		0.0472		0.0554*	0.0674	0.0551
M2		0.0442*		0.0593	0.0625*	0.0522*

	(A) Estimation results, 1964:3-1991:4
Financial aggregate		${\bar{R}}^{2}$	Log LF	RMSE
Currency		0.55*	280.26*	0.0217*
MB		0.49	273.14	0.0258
M2		0.48	271.97	0.0241
	(B) Ex-post forecasting RMSFE
		forecasting period
	1987:1-1991:4	1988:1-1991:4	1989:1-1991:4	1990:1-1991:4
Financial aggregate
Currency	0.0127*	0.0136	0.0140*	0.0171
MB	0.0148	0.0132*	0.0147	0.0169*
M2	0.0155	0.0169	0.0152	0.176

	(A) Estimation results, 1964:3-1991:4
Financial aggregate		${\bar{R}}^{2}$	Log LF	RMSE
Currency		0.55*	280.26*	0.0217*
MB		0.49	273.14	0.0258
M2		0.48	271.97	0.0241
	(B) Ex-post forecasting RMSFE
		forecasting period
	1987:1-1991:4	1988:1-1991:4	1989:1-1991:4	1990:1-1991:4
Financial aggregate
Currency	0.0127*	0.0136	0.0140*	0.0171
MB	0.0148	0.0132*	0.0147	0.0169*
M2	0.0155	0.0169	0.0152	0.176

Topics

Countries

Series

About

Resources

Abstract

I. Introduction

II. Financial Sector Reform and the Relationship Between Money, Output, and Prices

III. Financial Liberalization in Indonesia. Korea, and the Philippines

IV. Money, Output, and Prices in Indonesia, Korea, and the Philippines

1. Times series properties of the data

2. The money-output-price relationship and the information content of money

3. The “best” financial aggregate

4. Financial sector reform and the money-output-price relationship

V. The Information Content of the Exchange and Interest Rates

VI. Summary and Conclusions

APPENDIX I Data Sources

APPENDIX II

APPENDIX III

References

Same Series

Other IMF Content

Other Publishers

Asian Development Bank

Inter-American Development Bank

The World Bank

Share Link

Table of Contents

Headings

Figures