CHAPTER 1

 INTRODUCTION

1.1. Background

Irving Fisher (1930) points out that, with a well-functioning capital market, the one-period nominal rate of interest is the equilibrium real return plus the fully anticipated rate of inflation. If the statement holds true, it implies that real interest rates, which are important variables affecting investment and output decisions, are determined only by real factors and cannot be influenced by monetary policy.

The past decade has witnessed great success in Vietnam in macro management in areas such as: inflation control, interest liberalization and an improvement in exchange rate system. The relationships between macro variables have become of great interest for numbers of researchers. However, the relationship between nominal interest rate and inflation rates, whose management have been considered the outstanding achievements in the past decade, seem not to be paid a due attention. There has never been before a study of the Fisher effect in Vietnam.

This thesis aims to contribute a better understanding of the relationship between the nominal interest rate and inflation in particular, and the efficiency of interest policy in general.

1.2. Focus and Scope

 The thesis focuses on building an economic foundation for the Fisher statement through an asset - pricing approach in order to examine the relationship between nominal interest

rates and the inflation rate in Vietnam over the past 10 years from 1991 to 2001.

1.3. Research Questions

The major objective of this thesis is to answer the three following questions:

1.      Is the Fisher statement about the relationship between nominal interest rates and inflation rate reasonable under the asset - pricing approach?

2.      Can the Fisher equation be applied in Vietnam over the past 10 year period?

3.      If it can, what is the relationship between the nominal interest rate and the inflation rate and is it a one- to- one relation?

1.4. Methodology

The method of analysis includes the asset - pricing approach in building a model and an adoption of econometric regression for quantitative analysis. The monthly data are sourced from the State Bank of Vietnam (SBV), the International Monetary Fund (IMF), the World Bank (WB), the General Statistic Office (GSO) and related published works.

1.5. Structure

The thesis consists of five chapters. Chapter 1 reveals the introduction. Chapter 2 discusses various studies of the Fisher hypothesis in term of both theoretical concepts and empirical evidence. Chapter 3 provides an overview of interest rate liberalization in Vietnam as well as inflation scenario during the period of study. Chapter 4 outlines the model using the asset - pricing approach; presents econometric methodology and discusses the empirical evaluation. Chapter 5 summarizes the main findings of the thesis, offers policy recommendations and discusses the research limitations for further studies.

        CHAPTER 2

RELATED LITERATURE REVIEW

2.1. Interest Rate and Inflation: Theoretical Review

Since the early 1970s the effect of inflation on nominal interest rates, which was discussed extensively by Fisher (1930), again has gained an attention in economic literatures. Because there was a coincidence of rising inflation rates, rising nominal interest rates, and accelerating money growth that characterized much of phenomenon in 1970s. Up to now, the Fisher hypothesis has still widely been criticized and extended in empirical analysis.

2.1.1. Fisher- styled Equations

Strong form of Fisher Equation

Fisher (1930) hypothesized that if an unit change in inflationary expectations resulted in an equal unit change in nominal interest rates in the same direction.

For convenience, the Fisher equation is organized as follows:

                                                               (2.1)

where                the nominal (or market) interest rate at time t

                       - the expected (or ex ante) real rate at time t-1 forming at time t

                      - the expected inflation rate at time t-1 forming at time t.

Equation 2.1 is cited as the “strong form” of the Fisher hypothesis

Weak form of Fisher Equation

 In reality the Fisher- styled- relationship between nominal interest rates and inflation is not 1-for-1 relationship but 1- for- b relationship with b less then 1. Such is called as the “weak form” of Fisher- styled- relationship which is presented as

                        where  £ 1                     (2.2)

The above said equation is “weak form” of Fisher equation. There are several reasons that  would not be equal to unity which will be introduced in the next section.

2.1.2. Reason for The Existence of Weak Form of Fisher Equation.

Empirical investigations, for instance, Gibson (1972) and Lahiri (1976), both find the estimated impacts of expected inflation on nominal interest rate significantly below unity. There are various reasonable explanations relating to the negative relationship between real interest rate and expected inflation as follows:

Error in Proxy for Expected Inflation

The first reason cited measurement error in the proxy for expected inflation. Such measurement error could bias downward the estimated coefficient on anticipated

inflation.

Mundell Real Balance Effects

Mundell (1963) argued that higher inflation leads to a portfolio shift out of nominal assets (money) and into real assets, i.e. investment in real capitals and this makes the real rate which represents the rate of return on real capitals decline. Like Mundell, Tobin (1965) also explains the negative relationship between real interest rate and real inflation, with an emphasis on the cost of holding money balance as representing a liquidity premium for money. Since their ideas are the same as reasoning the shift in agents’ real balance in response to inflation changes, thus it is sometimes called Mundell-Tobin effect.

Output Disturbances

Fama (1981) argues that higher real interest rates represent greater productivity in the economy. Therefore higher real interest rates correlate with output disturbances. And higher real interest rates will correlate with lower inflation.

Liquidity Premium

A liquidity premium drives a distinction between real assets and nominal assets adjusted

 for inflation. Higher inflation raises the liquidity premium and therefore lowers real returns, generating a negative correlation between inflation and real rates.

In addition to the above four seasons, there are more credible arguments summarized by, IMF (1984) to rationalize why nominal interest rate does not adjust equally to the expected inflation changing. Those include:

Liquidity Effects

As additional money is injected into the economy, individuals may for a time experience excess liquidity, particularly if the increase in the money supply is not fully anticipated. Thus, before prices and inflationary expectations fully adjusted upward, the impact of excess money may, as Keynes argued long ago, lead to a lowering of the interest rate..

Economic Activity Effects

A slowdown in the economic activity is likely to pull the interest rate below the level that, certeris paribus, would exist if the economic activity remained at a “normal level”. This effect might be reflected in insufficient adjustment of the nominal interest rate for the expected inflation rate.

Institutional Constraints

If the movement of interest rate is constrained by legal or institutional limitations, then the observed rates may be lower than the rates that would be prevail in the absence of any limitations.

Money Illusions

There must be at least some individuals who, especially when the rate of inflation is low, confuse nominal interest rates with real interest rates. As long as some hold this illusion, the nominal rates may tend to increase less than expected inflation does.

Fiscal Deficits

This can influence the interest rate in various ways such as: the economic; changing the supply of money in the economy. However, a more direct effect is through the demand for loanable funds. As the government sells bonds to finance the deficit, the supply of bonds, certerris paribus, increases, thus making price of bonds falls, and the interest rate rises

Uncertainty

As investments are essentially commitments to a given set of future realistic prices, it is implied that a risk factor associated with investment rises. This rise in the risk factor induces a negative shift in the borrowing schedule which, per se, implies a lower real interest rate

Tax effects

When an income tax is levied on nominal interest earnings, with a rise in anticipated inflation to leave constant the after-tax expected real rate, the nominal interest rate must rise in anticipated inflation. Clearly, the rise in nominal interest rate must be [1/(1- )] time the rise in anticipated inflation, where  represents the marginal tax rate on interest income. We can write in form as follows:

                              (2.3)

2.2. Interest Rate and Inflation:  Empirical Review

There are various studies relating to the Fisher hypothesis. One group of researchers try to study the validity of the Fisher effect starting from the Fisher equation. The other groups adopts the rational expectation based on asset pricing model to attain the Fisher relationship.

2.2.1. Empirical Studies of Original Fisher Equation

Fama (1975) has proposed an influential paper on the empirical implications of the Fisher hypothesis under rational expectations investigating the relationship between nominal interest rates and expected inflation. Two major conclusions are drawn from the paper. First, during the 1953 –1971 period, the bond market seemed to be efficient in the sense that in setting one to six-month interest rates the market correctly used all the information about future  inflation rates that was in time-series of past inflation rates. Second, one could not reject the hypothesis that equilibrium expected real returns on one to six-month bill were constant during the period.

Barsky (1987) attributes the apparent change in the Fisher relation to differences in the stochastic process of inflation, rather than a change in any structural relationship between interest rates and expected inflation. Two countries selected to investigate include the US and the UK. The objective of his paper is to explain why data from the post-world war II period (particularly post-1960) look more “Fisherian” than do the pre-war data

Mishkin (1992) resolves the puzzle of why a strong Fisher effect occurs only during certain periods but not for others

Dutt and Ghosh (1995) study the validity of the Fisher hypothesis for Canada under both fixed and floating exchange rate regimes. They also give the definition of weak and strong forms of this hypothesis.

Choudhry (1996) investigates the long-run relationship between nominal interest rates and the inflation during the gold standard era (1879 - 1913). The study shows that if one-period inflation rate is non-stationary at levels , then expectation of future inflation will be dominated by current one-period inflation rate.

Thorton (1996) presents the existence of the Fisher effect between the treasury bill interest rates and inflation in Mexico during 1978 – 1994.

Engsted (1996) tests the long-term Fisher hypothesis under rational expectations and constant ex ante real rates using Danish data covering the post-world war II period (1948 - 1989). The analysis takes into account that interest rates and inflation rates are non-stationary processes, integrated of order one.

Yuhn (1996) investigates the Fisher effect of the short-and-long run interest rates over the modern (post-March 1973) flexible exchange rate experience in five industrialized countries: the United States, the United Kingdom, Japan, Germany and Canada..

Olekalns (1996) and Hawtrey (1997)  examines the Fisher relationship in Australia, using a procedure based on vector autoregressive innovations, which yields estimates of structural parameters in models featuring rational expectations.

2.2.2. Empirical Studies of Asset Pricing Model- based- Fisher Equation.

Next, we go through the work which adopts the rational expectation based on asset pricing model to form the Fisher relationship.

Shome, Smith and Pinkerton (1988) develop and estimated the model of the Fisher equation under uncertainty and risk aversion. The framework they used is one of the rational expectations models as developed by Lucas (1978). Their model is first tested by using variables based on the Livingston forecast data. The consideration period is run from June 1971 to December 1984. The empirical results generally support our interpretation of the Fisher effect and the risk aversion effect.

Rose (1988) attempts to provide robust evidence relevant to the stability of the ex ante short- term US real interest rate. Nevertheless, Rose^’s conclusions account for the small sample distributions for standard unit root tests in the presence of moving average errors that may characterize U.S inflation

Chan (1994) analyses the empirical importance of inflation uncertainty for the time-series behaviour of short-term real interest rates. It applies the consumption- bases asset pricing model to Treasury –bill rate. The resulting condition for an optimum portfolio provides an equation for real  interest  rates that  incorporates uncertainty about inflation.

Crowder and Hoffman (1996) extend the empirical literature on the Fisher evaluation in several new directions. The data they use are the three- month T- bill rate and annualized log changes in the price as a proxy for expected inflation

Furthermore, innovation analysis of both permanent and transitory shocks to the system was conducted as well. A permanent shock to the system causes the Fisher system to move to a new equilibrium. This shock may be traced to an inflation innovation. These dynamics allow the length of time for the equilibrium relationship to be restored, and a considerable room for a short - run relationship between real rates and inflation, as suggested by the Mundell-Tobin effect.

CHAPTER 3

OVERVIEW OF THE INTEREST REFORM AND INFLATION SCENARIO IN VIETNAM

By the late half of the 1980s, Vietnam' macro economic problem became all the more acute. Most of economic relations and activities were heavily distorted. Many reform measures were adopted. However, it was until the first half of 1989, interest reform - the core of financial reform - had been kicked off with the great adjustment of interest rate so that real interest rates are significantly positive. From that on, interest reform, year by year, has become more profound, making a contribution to creating a macro- economic stability in general and controlling inflation in particular.

3.1. Interest Rate Reform during 1991-2001

Before March 1989, despite some adjustment of interest rate undertaken by the SBV, the real interest rates were below than zero because of a continuous hyperinflation. Two striking feature of interest rate mechanism in this period were: (i) deposit rates of interest were lower than inflation rate, and (ii) lending rates of interest were lower then borrowing rates and inflation rates.

From March 1989, SBV had taken interest rates as a crucial tool to check the high inflation, lure idle money in circulation, get rid of government interest rate subsidies. Only within a very short period, had deposit interest rates been pushed to a rather high level Such action had brought about a sharp decline in inflation from 400% in 1988 to 34% in 1989.

In the 1^st October 1993, the adjustment had occurred in the direction of applying both "ceiling" interest rate and "negotiated" interest rate for loans and deposits (Decision No.184/QD-NH1). Essentially, under "negotiated" interest rate mechanism, interest rates were partly liberalized.

In 1^st January 1996, the SBV applied ceilings on lending interest rates and the spread of 0.35% per month between average mobilizing deposit and lending interest rates.

In August 1998, the Vietnam' interest rate liberalization achieved one more step ahead with two SBV important decisions: (i) the ceiling interest rate management mechanism is replaced by the basic one for VND lending and (ii) the interest rates applied for foreign currency lending are regulated by the market.  The basic rate as well as the band of fluctuation quoted monthly for VND lending and SIBOR rates plus relevant band serve as maximum level of the fluctuation of USD lending rates.

It was in June 2001, foreign currency lending rates were fully liberalized and decided by the demand for and supply of foreign currencies in the market.

It can be noticed that, in 1991-2001 period, interest rates could not reflect exactly the demand for and supply of capital in the market despite continuous change of interest rates undertaken by the SBV.

 3.2. Inflation Scenario during 1991-2001

After the first success in 1989 with inflation being pressed by 34.7%, inflation reemerged strongly in the two next consecutive years with the peak of 67.6% in 1991. In 1992, a law to ban money printing for financing budget deficit was passed, thus contributing to the sharp drop to 17.5% in 1992. From that on, the prices have become more stable.

Inflation continued to reduce by 5.2% in1993 due to the government's adoption of tight fiscal and monetary policies, but jumping to 14.4% in 1994 mainly because of loose monetary policy.

Over the 1994- 1997 period, Vietnamese government was very cautious to carry out tight monetary and fiscal policies. The price growth slow down to 12.7% in 1995 and to 4.5% and 3.6% in 1996 and 1997 respectively.

The 1999- 2000 saw a slump in inflation rate. The consumer prices declined in 8 consecutive months in 1999, bringing the inflation of this year to the unprecedented low of 0.1%. Both fiscal and monetary policies had been loosen in order to advocate the demand stimulation. For the first time since the beginning of the economic reform, the Vietnamese economy faced deflation. The stimulus package which was reflected through the swelling budget deficit and the rising money supply, had been intensified, thus lifting the consumer prices to some extent. However, the inflation still stood at a modest level of 0.8% in 2001.

To summarize, the performance of inflation in Vietnam since the economic reform can be divided in three phases: (i) the 1989- 1991 phase of high inflation; (ii) the 1992- 1998 phase of low inflation; (iii) the 1999- 2001 phase of deflation.

3.3. A “Weak Form” of Fisher Equation in Vietnam ?

This section will focus on briefing some main factors, available in Vietnam economy, in favor of a “weak form” of Fisher equation.

3.3.1. Interest Rate Control

Theoretically, the movement of constrained interest rates are often lower than the rates that would  be in the absence of any limitations, therefore a strong form of Fisher equation hardly appears in an economy with constraints.

During the past interest reform, the VND lending interest rate were always in the control. The mechanism of  “ceiling” interest rate, which had existed since 1989, has gradually replaced by the mechanism of “basic” interest rate. Admittedly, the introduction of “basic” interest rate is a step forward in the process of interest rate liberalization. However, such a mechanism can not be as flexible as a mechanism of market-determined interest rate in spite of efforts paid by the SBV to adjust the interest rate to meet the market forces.

3.3.2.Liquidity Effects

Over the past 10 years from 1991 to 2001, the money supply of Vietnam had increased year by year with the annual average increase of around 30-35%. If in 1991, the M2 was 15704 billion VND, up to 2001 this figure reached to 258000 billion VND.

In the economy where the money supply rises, the price and inflation expectations, as argued by Keynes, are often adjusted more slowly than the interest rates because it must take time for individuals in that economy to fully anticipate the excess liquidity.

Therefore, in Vietnam whose M2 increased considerably year by year, the argument for a slow adjustment of inflation, in comparison to the change of interest rate, seem not to be groundless.

3.3.3 Dollarization and Portfolio Choices

Dollarization in Vietnam reached the peak of 41% in 1991, then fell to 23% within two later years and stabilized in the range of 20 to 24% in the mid 1990s. Starting in 1997, the ratio has risen to more than 34% by August 2001.

There are two main motives for the demand for foreign currency. First is currency substitution, which occurs when residents hold foreign currency as a means of payment, The second is asset substitution, when residents hold foreign currency denominated asset for investment purposes, rather than as a means of payment.

Dollarization has enlarged the portfolio choices and USD interest rate become a important factor in analyzing the movement of inflation. As said in the previous chapter, the poor proxy of interest rate and inflation can be likely to drive the Fisher equation, if any, to be under a “weak” form, rather than “strong” form.

CHAPTER 4

MODEL SPECIFICATION AND EMPIRICAL EVALUATION

4.1. Model Specification

The model starts with a representative household maximizing periodic utility subject to an intertemporal budget constraint. The assets traded in this economy are a one-period nominal default-free asset that pays in form of currency and a one-period risk-free asset that pays in a single consumption good. The Fisher relation is derived as follows:

Max

Subject to

Where           β = 1/(1+ p); p is the rate of time preference

                      U(C_t) - the utility function of a representative household

                      C_t - consumption in period t

                      A_t - value of nominal interest- bearing asset due in period t

                       a_t - amount of real interest- bearing asset due in period t

                      P_t - price level in period t

                      R_t - nominal rate of interest in period t

           r_t - real rate of interest in period t

           Y_t - household income in period t, and

E_t - expectation operation conditional on all the information available at time t.

The Lagrangian function is expressed as:

By solving the above said Lagrangian function (As referred in the Appendix A), we can get the following result:

R_t+1      =                                                             -------- (4.1)

where  and .

and       R_t+1 is the nominal short-term interest rate from period t to t+1

            E_t ΔlnP_t+1 is the ex post inflation rate from t to t+1,

            r is the mean of the short-term real interest rate from period t to t+1 minus the one-half the conditional variance of inflation,

            ς is the mean of risk premium.

The final equation reveals that the constant term in a simple relation linking nominal interest rates and inflation comprises the real interest rate minus factor that proxy the conditional variance of inflation less the mean of the risk premium.

4.4. Empirical Evaluation

Based on the suggestion of equation (4.1), we use data collected from the movement of inflation and nominal interest rate in Vietnam over the past 10 years to run the following regression:

DR3M = C + b₂ INF^e + b₃ D1 + b₄ D2 + b₅ D3               …… (4.4)

where  DR3M =  Nominal interest rate,%

INF^e =  Inflation rate,%

D1 = 0  for period before October 1993

      = 1  for period beginning in October 1993

D2 = 0  for period before January 1996

      = 1  for period beginning in January 1996

D3 = 0  for period before January 1998

      = 1  for period beginning in January 1998

Interpretation of the Result

Figure 4 presents regression result after elimination all the dummy variables from equation (4.4).

Figure 4. Regression Results in Periods

The above figure shows that, over the past decade, the inflation and the nominal interest rate in Vietnam moved in the same direction. However, the relationship between nominal interest rate and inflation in Vietnam is not 1- for- 1 as in the strong form of Fisher equation, it is only 1- for- 0.421, thus belonging to the weak form of Fisher relation.

When the inflation increased by 1 percentage point, the nominal interest rate raised by 0.421 percentage point. And 1 percentage point decease in the inflation was accompanied by 0.421 percentage point decrease in the nominal interest rate.

The constant terms in three equations, expressed in the figure 4.4, indicate that, if the conditional variance of inflation and the mean of risk premium are unchanged, then the real interest rate seems to reduce in the past 10 years.

The existence of a Fisher equation between inflation and nominal interest rate under the “weak form” in Vietnam is not surprised much, because as shown in the Chapter 3, the Vietnam’ financial system has not yet fully liberalized and many constraints do still exist.

CHAPTER 5

CONCLUSION AND RECOMMENDATIONS

5.1. Summary of Main Findings

The model, based on an asset - pricing approach, shows the fact that the nominal interest rate has a close relation not only with the real interest rate and inflation but also with such  factors as the conditional variance of inflation and risk premium (presented by the conditional covariance between consumption growth and inflation rate). The findings provide a clearer explanation of consumers' behaviors as well as factors evolved in the relation between nominal interest rate and inflation rate.

The econometric analyses indicate the non-stability of the macro variables of the nominal interest rate and inflation in the Vietnam's economy over the past decade. Both variables follow random walk. However, they are all stationary at the order of one. The approval of cointergration test means the result can be accepted from a direct regression. The Granger causality confirms that the lagged inflation rates are not good predictors for the nominal interest rates while the lagged nominal interest rates have  certain predictive impacts on inflation.

Three dummy variables, corresponding to three great changes in Vietnam's interest policy, are all significant at the 99% and all are negative, thus showing that the changes of interest policy affected the economy and reduced the real interest rate if the conditional variance of inflation and the mean of risk premium are unchanged .

Lastly, the result from regression in the case of Vietnam also indicates a weak rather than  strong form of the Fisher equation as is the case in other countries. The relationship between the nominal interest rate and the inflation rate in Vietnam in the past 10 years is 1- for- 0.4 instead 1- for- 1, which proves that the financial market is far from fully liberalized. However, a positive relationship does exist which is counter to a negative one as in the Keynesian view.

5.2. Policy Recommendations

First, in the case of Vietnam, policy makers should pay attention to a positive relationship between the nominal interest rate and inflation in mapping out interest policy to intervene in the economy in general and in the fluctuation of inflation in particular.

Second, the weak form of the Fisher equation shows a non-neutrality of interest policy in the determination of inflation. The interest policy as well as monetary policy still does affect, to some extent, the degree of expected inflation and the changes in real interest rates. Therefore, the interest rate policy, in the case of Vietnam, is a tool for intervencing in the fluctuation of inflation.

Third, the weak relationship between the nominal interest rate and inflation rate in Vietnam shows that the financial market is far from full liberalization. Further liberalization of the financial market is necessary to make the market factors more able to perform efficiently.

5.3. Limitations and Suggestions for Further Studies

There still remain some limitations in the thesis, which could be further developed by deeper analysis, lying  in following areas:

The first relates to the proxy for a risk-free rate. The nominal interest rate used in the study is 3 month saving deposit rate and may contains some certain biases. In further study, together with the development of financial instruments, other kinds of nominal interest rate could be adopted.

The second is the proxy for the expected inflation rate. Here, only one way of estimating expected inflation can be applied. We cannot say for sure that such a proxy is the best.

The third is the omission of the effect of marginal tax rate on interest income. Because this effect is very difficult to estimate. Such an effect itself provides a good topic for other studies.

The last is the assumption of a constant risk-premium. This, once again, resulted from the difficulty in estimation. In the future, it may be better to include this variable in the regression to draw more exact results.