II. Financial Reintermediation and New Challenges to Monetary Policy11
Over the past year, liquidity and credit in the Philippines have surged to exceptional high levels. On balance, this is a welcome rebound from the mid-1980s, when the debt-servicing crisis and a loss of confidence caused the banking sector to shrink. Indeed, financial reintermediation now is making possible the levels of investment needed to put the Philippines irrevocably on the path of sustained growth. At the same time, however, the surge in liquidity growth has created formidable new challenges for monetary policy, especially one anchored by monetary targets.
This paper looks at three issues. First, why has financial intermediation expanded so rapidly? Second, what does this expansion imply for a monetary policy anchored by an intermediate base money target? Third, should the current very high rates of money demand growth be expected to continue?12
Formulation of the intermediate targets that have traditionally guided policy is based on the historical relationships between liquidity and macroeconomic indicators. However, econometric evidence indicates that these relationships have broken down. While this breakdown likely due to the influence of financial reforms and improvements in confidence on money demand, these factors cannot, in practice, be used as guides for targeting monetary growth.
Nevertheless, a judgement is still necessary regarding the sustainability of the recent surge in liquidity. For this reason, episodes of very high monetary expansion in other Asian countries are reviewed to garner evidence on the size and sustainability of the shift in liquidity in the Philippines. These episodes have been similar in size but short-lived—around two years, suggesting that the recent surge in money demand in the Philippines may soon begin to level off.
The main policy implications of this paper are two. First, during this period of financial intermediation, the effectiveness of monetary policy would likely be enhanced by putting less weight on intermediate monetary targets and more on the final objective of price stability. The authorities are already taking steps in this direction. Second, policy should not be based on the assumption that this period will continue; therefore, the authorities should remain alert to signals that money demand is beginning to falter.
B. Financial Intermediation: a Review of Recent History
The recent surge in liquidity growth can be best understood against the background of the sharp contraction of the banking sector during the decade after the debt crisis (Chart II.1). For this reason, financial sector developments since 1980 are reviewed. Four phases can be identified.
CHART II.1PHILIPPINES: Indicators of Financial Intermediation, 1981-95
Source: Data provided by the Philippine authorities.
1. The early 1980s: hollow growth 13
In 1980, the Government introduced two important financial sector reforms to develop the financial sector. First, interest rate ceilings were eliminated. Second, a “universal” banking system was established, allowing banks to expand their activities into new areas and broaden their ownership base. These reforms achieved some success. Interest rates became market-determined, high fees and charges were reduced, and nine commercial banks converted to a universal bank charter.
However, banking sector growth was achieved at the expense of nonbanks, indicating that the roots of expansion were not deep. A string of trading-firm bankruptcies triggered the failure of a large number of investment houses and finance companies, and even created liquidity shortages for nonfinancial firms not directly related to the crisis. As a result, banks took on a larger share of private savings, partly by absorbing many of the surviving nonbank institutions in government-managed mergers.
2. The mid-1980s: financial disintermediation
During the mid-1980s, a series of severe shocks to the already weakened financial system resulted in a calamitous decline in financial activity, from which the Philippines is only today recovering. The ratio of broad money to GNP fell by 8 percent of GNP during 1984 alone, while the ratio of bank credit to GNP plummeted from 42 percent in 1983 to 16 percent just three years later. This rapid pace of financial disintermediation compressed liquidity and credit to levels that are significantly below those prevailing in other Asian countries (Chart II.2).
CHART II.2PHILIPPINES: Financial Intermediation in Selected Asian Countries, 1980-94
Source: Data provided by the Philippine authorities.
The contraction in financial activity was triggered by the external debt crisis of 1983. The prolonged and painful period of adjustment needed to restore external stability drained external and domestic sources of bank liquidity. The exclusion of the Philippines from international capital markets, together with payments to retire outstanding debt, forced commercial banks to shift from a substantial negative net foreign asset position to a surplus. At the same time, banks’ access to domestic financing sources was cut back. Heavy issuance of government domestic debt instruments to finance the fiscal deficit and repay foreign obligations attracted private savings away from bank deposits, and squeezed bank balance sheets, reducing the supply of bank lending to the private sector (Chart II.3).
CHART II.3PHILIPPINES: Broad Financial Balances, 1981-95
Source: Data provided by the Philippine authorities.
The impact of the external credit cutoff was exacerbated by domestic political instability, which undermined confidence in the financial sector and the peso.
The uneven policy response during the early phase of the adjustment period further undermined the stability needed to restore confidence in the financial system. Monetary growth and real interest rates fluctuated wildly from 1983 through early 1985, before stability was achieved at the cost of very high rates during 1986 and 1987. 14
These gyrations in monetary policy were closely related to heavy central bank losses. During the mid-1980s, an increasing share of the assets of the old Central Bank of the Philippines (CBP) earned no income, resulting in losses averaging 2 percent of GNP from 1983 to 1992. The decline in CBP asset quality was a result of foreign exchange swap and forward cover losses, and the bailout of several large banks. The heavy operating losses reduced the ability of the central bank to undertake measures, such as open market sales, that would further reduce its income. Consequently, the National Government financed CBP operations by making large noninterest paying deposits, and aided liquidity management by varying the amount of auctioned treasury bills to influence market conditions.
Finally, financial intermediation was further compressed by high intermediation costs. Wide spreads between bank deposit rates and lending rates reflected three factors. First, two large direct taxes were—and are today—exacted on bank-related financial transactions. Depositors pay a 20 percent withholding tax, and banks pay a 5 percent gross receipts tax on loans. Second, the raising of reserve requirements to 25 percent during the mid-1980s, necessitated by the need to narrow central bank losses, imposed a costly tax on bank intermediation. Third, in addition, bank operating margins were wide, reflecting the lack of competition in the banking sector.
3. The late 1980s and early 1990s: nascent recovery
The financial sector began to rebound around 1987. Private saving accelerated, allowing a sharp buildup of bank deposits, even as the stock of government securities continued to swell. Banks also began to tap new sources of liquidity. Owing to high reserve requirements on regular deposits, bank turned to off-balance sheet common trust funds (CTFs), which were used mostly to buy government securities. Banks also obtained large amounts of foreign exchange from foreign currency deposit units (FCDUs), especially after the foreign exchange system was deregulated in 1992. The buildup in liabilities was matched by a rebuilding of foreign assets, and, as a result, bank credit didn’t begin to recover until 1992.
Meanwhile, the factors that crippled the financial sector began to turn around. The domestic political situation stabilized, a comprehensive reform program increased the role of the market, and the December 1992 debt restructuring agreement with commercial banks renewed access to the international capital markets.
4. 1994-95: financial reintermediation
Liquidity and credit have accelerated during the current recovery, especially during the past year. Annual liquidity growth rose from 15 percent during 1993 to 36 percent in the first quarter of 1995, a historical high, while bank credit to the private sector swelled by 40 percent during the year ending March 1995. 15
This expansion has been driven by sharp upward shifts in both supply and demand. Beginning in mid-1994 policy was loosened, and the impact of base money expansion on overall liquidity was amplified by unexpected increases in the money multiplier, owing to an acceleration in the shift from currency to deposits (Chart II.4). Strong liquidity growth allowed bank credit to the private sector to expand rapidly.
CHART II.4PHILIPPINES: Currency-Deposit Ratio, 1986-95 1/
Source: Data provided by the Philippines authorities.
1/ Not adjusted for upward shifts around the time of the May 1992 and May 1995 elections.
At the same time, there is evidence that increases in money supply were matched to a large extent by increases in money demand. First, the composition of savings shifted from currency to deposits as confidence improved and financial sector reforms took hold. Second, notwithstanding the surge in liquidity, inflation fell to 7 percent by end-1994 (the relationship between inflation and liquidity is considered below). Third, regression analysis, presented in the following section, indicates that the standard cyclical determinants of money demand do not explain liquidity growth during the past year.
C. Analysis of the Recent Surge in Monetary Growth
The sharp growth of liquidity during the last year poses two challenges to monetary policy. First, it has complicated the formulation of the monetary targets, which have traditionally anchored monetary policy. The derivation of monetary targets could be modified in three ways to account for the shifts in liquidity growth: (i) if it can be determined that the liquidity surge is one-off or reversible, then the methodology used to derive monetary targets should not be changed; (ii) if the increase in liquidity shift is judged to be lasting and predictable, then the methodology should be altered; and, (iii) if money demand is no longer predictable, then the choice of monetary targets should be reconsidered.
Second, shifts in money demand have created considerable uncertainty regarding the appropriate stance of policy, which requires a judgement as to the sustainability of the very high rates of liquidity growth.
This section brings empirical evidence to bear on these two issues. First, econometrics estimates of a standard money demand function are presented. Next, since the standard model leaves a large portion of money demand unexplained, an alternative determinant of money demand is introduced. As even these results are not conclusive, the experience of other countries that underwent episodes of very high liquidity growth is reviewed.
1. Econometric estimates
a. The model
where real liquidity is the log of the sum of broad money and CTFs, deflated by the CPI; real GNP is in log form; and the interest rate is the benchmark 91-day Treasury bill auction rate. The long-run properties of the model can be gauged by testing whether the et time series is mean-reverting, i.e., stationary. If statistical tests indicate that real is stationary, then the data support the existence of a stable long-run relationship between real liquidity, and interest rates. If this is the case, a value of et near zero is evidence that the money market is in equilibrium—based on the historical relationships between liquidity and macroeconomic aggregates. Positive values of et—actual liquidity exceeds predicted liquidity—indicate that a correction toward long-run equilibrium is in order and, therefore, real liquidity growth should decelerate in the next several periods. Since et captures the need for a subsequent money market correction, it is called the error correction term.
With reliable estimates of the long-run relationship parameters in hand, the second step involves unrestricted estimation of the short-run dynamics of money demand:
where ê is the estimated error correction term from the levels equation. 18 The signs of b1 and b2 are expected to be negative, since, as noted previously, an overshooting of the long-run level of liquidity should subsequently push down liquidity growth. Near-zero values of ft would indicate that monetary growth is proceeding as projected, taking into account the adjustment of, balances toward long-run equilibrium captured in the error correction term.
The feedthrough of yesterday’s departures from long-run equilibrium into liquidity growth today is based on the notion that money holdings do not adjust instantaneously to the desired levels consistent with growth and interest rates. One reason for this partial adjustment of money holdings is the cost of portfolio reallocation, for example, from money to bonds, which may be substantial in a country such as the Philippines, owing to the thinness of secondary markets. Slow adjustment of expectations would be a second reason. If money demand is, in practice, based on expected future values of growth and inflation and if expected values are only partially revised each time new information on current values become available, then actual money demand will react to current values with a lag.
b. Standard determinants of money demand
Presentation of the empirical results begins with estimates of the long-run liquidity relationship specified in equation (1). The first regression is based on data from the first quarter of 1981 (the first period for which quarterly real GNP data are available) through the second quarter of 1995. These results provide strong evidence for important structural shifts in the relationship between liquidity, income, and interest rates (Table II.1, top panel, column A). The R-squared is lower than the norm for long-run money demand estimates (Boughton, 1992 and Tseng and Corker 1991), and the augmented Dickey-Fuller (ADF) statistic indicates that this regression is not capturing a long-run relationship. 19
|(Long-run equilibrium relationship)|
|Real GNP elasticity||1.72|
|Interest rate semi-elasticity||-0.51|
|1 percent critical value||-2.60||-2.62||-2.62|
|Error correction term, lagged||-0.05|
|Error correction term, lagged twice||0.29|
|Real GNP elasticity||0.41|
|Real GNP elasticity, lagged||0.30|
|Interest rate semi-elasticity||-0.29|
|Interest rate semi-elasticity, lagged||-0.18|
|Bank branches, lagged||…||0.53|
T-statistics in parentheses.
T-statistics in parentheses.
The instability of the relationship over a data sample inclusive of the turbulent early and mid-1980s is no surprise. Instability was addressed by dropping observations one by one until the ADF statistic indicated that the residual was stationary at the one percent level of significance. A stationary residual was obtained by beginning the sample in the first quarter of 1984 (Table II.1, top panel, column B).
As would be expected, omitting the period of financial disintermediation resulted in a larger real GNP elasticity estimate. Also, shortening the sample period halved the estimated interest rate parameter and generated a larger standard error.
The most striking result of this regression is the strong evidence that standard money demand relationships leave unexplained a large portion of the level of liquidity since early 1994. This is demonstrated by the behavior of the residual, i.e., the error correction term (Chart II.5). The error correction term fluctuated widely during the volatile mid-1980s, was generally negative during the late 1980s period of policy consolidation, and turned positive during the early 1990s. During the past year, the error correction term has been positive and consistently large relative to history.
CHART II.5PHILIPPINES: Error Correction Term from Long Run Level of money Demand Equation, 1984-95
Source: IMF, staff estimates.
Estimation of the short-run dynamic model indicates that the growth of liquidity, as well as its level, exceeded expectations during the current recovery (equation 2). Generally, the parameter estimates are in line with expectations, although the influence of the error correction term is concentrated on the second lag, rather than the first, contrary to the usual case (Table II.1, bottom panel, column B). The income elasticity and interest rate semi-elasticity parameter estimates enter with the expected signs, but the standard errors are large for the lagged estimates. The R-squared of 54 percent indicates a good fit for a developing country money demand growth regression. Finally, recent liquidity growth rates have generally exceeded rates predicted by the model; during the first quarter of 1995, the model’s prediction of growth fell 7 percentage points short of the actual rate (Chart II.6).
CHART II.6PHILIPPINES: Residual from Short-Run Dynamics of money Demand Equation, 1984-95
Source: IMF, staff estimates.
What can be learned from these results regarding the sustainability of very high liquidity growth rates? The excess of the actual over the fitted level of money demand through the first half of 1995 suggests that a decline in real money balances lies ahead, either in the form of lower nominal money or higher inflation. Of course, there remains the possibility that to some extent the upward shifts in money demand are permanent and, consequently, that the model would continue to underpredict. For this reason, a look at alternative determinants of money demand in the Philippines is warranted.
c. Alternative determinants of money demand
The recent shift in money demand reflects more than just a cyclical upturn. In fact, it may stem from the sharp improvement in the performance and prospects of the Philippine economy, which was made possible by a consolidation of financial policies and a comprehensive program of structural reform.
The program of structural reform included important measures that reduced the cost of holding bank deposits. First, reserve requirements were reduced from 25 percent during the late 1980s to 17 percent in May 1995. Second, regulations that had stymied the opening of new bank branches, including restrictions on location and requirements for new capital for each branch, were lifted beginning in 1993; as a result, the number of bank branches rose by 34 percent during 1993-94. Third, the number and use of automatic teller machines (ATMs) were accelerated by a series of central bank circulars and by the merger of the two ATM vendors. Fourth, restrictions on the establishment of new foreign banks were relaxed effective May 1994. By the end of 1995, the number of foreign banks will have increased from four to 14. Although foreign banks are not expected to enter retail banking, their presence is already enhancing competition and innovation.
Moreover, financial confidence improved sharply, partly reflecting the reduction in inflation and the more stable exchange rate. This may have enhanced the willingness of savers to maintain their balances in peso bank deposit accounts.
Gauging the impact of reforms and confidence on money demand is difficult, since these factors are hard to quantify. 20 Fortunately, in the case of the Philippines, one indicator of structural reform—albeit crude—is available. This is the number of bank branches. 21
The number of bank branches is correlated with the portion of liquidity unexplained by the standard determinants of money demand (Table II.1, top and bottom panels, column C). Not surprisingly, this direct measure of monetization reduces the estimated elasticity of real GNP, since the increases in the number of bank branches has moved closely with GNP growth. Also, the error correction term from the levels equation is more stable, suggesting that including this crude measure of financial reform provides a better fit of the long-run money demand relationship.
What are the implications of the result for monetary policy? As a practical matter, it would be exceedingly difficult to include quantitative measures of financial reforms in the formulation of monetary targets. The reason is that the impact of financial reform on the monetary transmission process is not well known conceptually, and would be even more difficult to capture empirically. For example, some of the deposits flowing into new bank branches will simply be transfers from existing offices rather than a portfolio shift toward money. Therefore, using the projected relationship between proxies of reform and money demand in the formulation of monetary targets would be risky.
The empirical results can be summarized as follows. First, use of the standard determinants of money demand to derive monetary targets is now problematic because they leave unexplained a large portion of liquidity growth during 1994-95. Moreover, while the leftover monetary growth is correlated with a crude proxy of financial reform, it cannot be used to formulate targets.
Second, regarding the stance of monetary policy, this framework predicts a slowdown in real liquidity growth during the second half of 1995. However, the inability of the standard money demand determinants to explain liquidity behavior leaves a gap in the understanding of money demand in the Philippines. In this light, it may be fruitful to examine the experience of other countries.
2. The experience of other countries
This section examines the magnitude and duration of episodes of very high liquidity growth in four other countries in east Asia with a view to ascertaining whether the levels of liquidity growth in the Philippines during the past year are sustainable. Inverse velocity, defined as the ratio of the average of broad money (including CTFs for the Philippines) to current GNP, is the yardstick used in these comparisons.
Since 1987, inverse velocity growth in the four other Asian countries has grown by an average of 4 percent per year (Table II.2). Against this benchmark, the 23 to 35 percent annual inverse velocity growth rates in the Philippines during the four quarters through June 1995 is indeed high. Turning to individual countries, Indonesia during 1990-91 and Malaysia during 1992-93 are the only countries with inverse velocity growth rates that are comparable with that of the Philippines during the past year. For this reason, brief digressions on Indonesia and Malaysia are in order.
Ratio of money plus quasi-money to GNP
Ratio of money plus quasi-money to GNP
In Indonesia, inverse velocity surged by 18 percent and 24 percent during 1989 and 1990, respectively. Liquidity was driven by a jump in credit growth reflecting: (i) investment and financial reforms (including a sharp decline in reserve requirements); (ii) easy availability of subsidized credits; and (iii) low interest rates through early 1990. Subsequently, Bank Indonesia raised interest rates and introduced new regulations; with these measures in place, inverse liquidity was virtually unchanged during 1991.
Capital inflows drove liquidity growth in Malaysia during 1992-93, when inverse velocity rose by 13 percent and 14 percent, respectively. This episode was preceded by monetary tightening to resist inflationary pressures that had emerged during 1991-92. The subsequent increase in domestic interest rates, combined with a decline in international interest rates, attracted massive speculative capital inflows, which was followed by an acceleration of liquidity growth. The authorities responded by tightening monetary policy and introducing administrative measures during 1994. As a result, inverse velocity declined during 1994.
What are the lessons for the Philippines from the experience of other countries? First, the comparisons show that liquidity growth in the Philippines has been very high during the last year. Second, episodes of very high liquidity growth in other countries in the region have been short-lived—about two years. Finally, these episodes were followed by a sharp decline in liquidity growth. In brief, monetary policy in the Philippines needs to stay alert to a possible slowdown, perhaps a sharp slowdown, in money demand.
D. Implications for Monetary Policy
The inability of the standard money demand determinants to explain the recent surge in liquidity behavior suggests that the broad tenets of a monetary-based policy should be reviewed. First, is the relationship between money and inflation stable? Second, if not, should alternative intermediate targets be considered? These questions are considered in turn.
1. Base money targeting
In recent years, monetary policy has been anchored by monetary targets with a view to achieving lasting price stability and sustainable growth. This approach requires a steady and predictable relationship between money and inflation. However, the evidence shows that this relationship has broken down.
First, the sharp rise in liquidity and credit have not, so far, fed into higher inflation. Annual CPI inflation fell from about 10 percent in late 1993, to 5 percent in early 1995—the lowest rate since 1987, although in recent months inflation has been on the rise. These fluctuations partly reflect supply side developments, including the alleviation of power shortages and favorable harvests during 1994, and rising food prices from a drought during 1995. On balance, however, the absence of a pickup in inflation from the historically high rates of liquidity growth marks a break from the past.
Second, simple regression analysis confirms a breakdown in the relationship between money and inflation. This relationship was tested by regressing inflation on current and the first two lags of liquidity growth and the lagged exchange rate. During 1985 to 1993, these factors alone explained more than sixty percent of inflation, and the parameters are fairly stable over the this interval (Table II.3). However, extending the sample through 1995 produces sharply different parameter estimates, and there is evidence of a shift in the parameter values during 1993. 22
|Liquidity growth, first lag||0.17|
|Liquidity growth, second lag||0.29|
|Nominal effective exchange rate, percentage||-0.12|
Dependent variable is CPI inflation; data are quarterly and seasonally adjusted.
T-statistics are in parentheses.
Dependent variable is CPI inflation; data are quarterly and seasonally adjusted.
T-statistics are in parentheses.
If monetary balances no longer serve as a reliable signpost for prices, a reexamination of the targets of monetary policy is warranted.
2. Inflation targeting
Costly fluctuations in the final target of monetary policy may result when it is not linked with the intermediate policy target. In the Philippines during the past year, the evidence suggests that the relationship between inflation and the intermediate monetary target has broken down.
This leads to the main policy conclusion of this paper: the effectiveness of monetary policy in the Philippines would likely be enhanced during this transitional period by putting less weight on intermediate monetary targets and more on the final objective of price stability. This would reduce the risk of basing the policy stance on uncertain projections, and downplay judgements regarding whether liquidity shifts reflect swings in money demand. 23
Of course, this approach raises yet another challenge. Given that inflation reacts to liquidity growth with variable lags, there is considerable uncertainty about the correct path of monetary instruments consistent with a given inflation target. If lags are especially long, there is the danger that a prolonged and costly tightening would be needed to stem the impact of excess liquidity growth on inflation. According to the regression analysis, most of the feedthrough from liquidity growth to inflation occurs within two quarters. 24 While this may be enough time for policy to be adjusted to emerging inflationary pressures without high costs, there remains the possibility that a sudden drop in money demand would not be reflected in prices until substantial monetary overshooting had already occurred.
This risk can be addressed by utilizing indicators other than prices as early warning signals of pending inflationary pressures. For example, the authorities could monitor carefully the evolution of international reserves, the exchange rate, and domestic interest rates—all of which will provide clues as to the adequacy of the monetary stance.
Already, the authorities are moving toward inflation targeting. Their base money targets are now being adjusted upward by the amount that international reserves exceed expected levels, as long as inflation stays within a targeted range. This would appear to be a prudent response to the new challenges to monetary policy posed by financial reintermediation.
ArrauPatriciaJose deGregorioCarmenReinhart and PeterWickham “The Demand for Money in Developing Countries: Assessing the Role of Financial Innovation” Working Paper International Monetary Fund (Washington, D.C.) No. WP/91/45May1991.
BoughtonJames M. “International Comparisons of Money Demand: A Review Essay,” Working Paper International Monetary Fund (Washington, D.C.) No. WP/92/7January1992.
EngleRobert F.GrangerC. W. J. “Cointegration and Error Correction: Representation, Estimation, and Testing” Econometrica Vol. 55 No. 2 (March1987) pp. 251–76.
GoldfeldStephen M. and SichelDaniel E. “The Demand for Money” in Handbook of Monetary Economics Vol. 11990.
GoldsbroughDavid and IqbalZadiri “Monetary Policy in the Philippines during the Period of Financial Crisis and Changes in the Exchange Rate Regime” Working Paper International Monetary Fund (Washington, D.C.) No. WP/89/98December1989.
HendryDavid and NeilErickson “An Econometrics Analysis of U.K. Money Demand in Monetary Trends in the United States and the United Kingdom by MiltonFriedman and Anna J.Schwartz” American Economic Review 81 1 pp. 8–381991.
HesterD.D. “Innovations and Monetary Control” Brookings Papers on Economic Activity (1989) No. 1 pp. 141–189.
JohansenSoren “Statistical Analysis of Cointegration Vectors” Journal of Economic Dynamics and Control Vol. 12 (1988) pp. 231–254.
KhanMohsin and V.Sundararajan “Financial Sector Reforms and Monetary Policy” Working Paper International Monetary Fund (Washington, D.C.) No. WP/91/27December1991.
LeidermanLeonardo and Lars E.O.Svensson (eds) Inflation Targets London CEPR (forthcoming).
TsengWanda and CorkerRobert “Financial Liberalization, Money Demand, and Monetary Policy in Asian Countries” Occasional Paper International Monetary Fund (Washington, D.C.) No. 84July1991.
Prepared by Mark R. Stone.
The related issue of complications to monetary policy arising from capital inflows is considered in the accompanying background paper: Capital Inflows: Are They a “Problem?”.
Monetary and exchange rate policy during this period is reviewed in Goldsbrough and Zadiri (1989).
For long-term analysis, it is preferable to use liquidity—the sum of broad money and common trust funds—rather than broad money alone. Broad money figures for 1993 and 1994 are somewhat misleading because the imposition of reserve requirements on (off-balance sheet) common trust funds appears to have induced a shift into bank deposits.
Restriction of the analysis to money demand only does not control for the influence of money on output and prices, in contrast to the Johansen (1988) multiple equation procedure. However, the relatively small number of estimated parameters in the money demand approach used here allows a more direct analysis of structural changes.
Preliminary regression results indicated that adding more lagged explanatory variables did not markedly improve the explanatory power of the model.
Similarly, Tseng and Corker (1991) did not obtain stable money demand estimates using data from 1976 through 1988.
Arrau et. al (1991) include a deterministic trend variable in money demand regressions for countries undergoing prolonged periods of financial deepening.
The cumulative sum of squares test indicates a structural break at the third quarter of 1993, at the 5 percent level of significance.
The experience of other countries that have turned to inflation targeting is reviewed in Leiderman and Svensson (forthcoming).
This is unusually rapid by the standards of most countries. A possible explanation for the short lag is that the exchange rate in the Philippines appears very sensitive to changes in the money supply, and the passthrough from the exchange rate to prices is relatively rapid—perhaps even before import costs rise.