Econometrics concepts and applications

¶ … exchange rates and Inflation

Can manipulating Currency Rates effectively reduce inflation?

Recent trend have shown that developing countries have been prone to higher inflation than industrialized countries. This trend has formed the basis of monetary policy in these countries since the early 1980s (Bleaney and Fielding, 2005). Bleaney and Fielding surmised that inflation could be controlled in these countries by adjusting currency exchange rates. Bleaney and Fielding suggest that the method of pegging the developing nation's currency exchange rate to the currency of a more advanced nation, or perhaps a basket of such currencies was the most effective means to control inflation.

This was the method that dominated the monetary policy of developing nations before the 1980s. However, in the mid 1980s things began to change as developing nations began to move towards capital market systems and enter into the global marketplace. Advances in communications and technology made it possible for developing nations to compete on a global basis and they began to adjust their monetary policy to resemble more closely those of industrialized nations.

We have seen this in Asia and South America in recent years. When developing nations changed monetary policy, they soon found that there was a right way and a wrong way to do this. Countries such as Brazil and Mexico experienced dramatic rises in inflation as a result. They had to adopt drastic policies to control runaway inflation and regain some order in their economy. Other countries, such as Morocco, took these lessons to heart and adopted floating exchange rates at a much slower pace. This helped to ease the transition.

This research user regression analysis to support the hypothesis that inflation rates will be directly related to changes in its currency exchange rate among developing nations. In order to accomplish this task the study will examine data from developing nations and industrialized nations using a regression analysis to determine if any correlation exists.

Theoretical Basis of Analysis

Governments have several mechanisms available to them to assist in the regulation of their fiscal health. Manipulating the exchange rate and the adoption of various monetary policies are only two of the solutions available to manage the economy. These are the two most commonly used mechanisms to assure that their currency retains value and that they do not experience run away inflation and all of the ills that come with it.

The goal of every government is stability in the economy. They seek to avoid scenarios such as high inflation, currency devaluation and other situations that would harm their ability to maintain peace within their boundaries. Low inflation and the ability to maintain their competitive position on the global marketplace have become major goals of policy makers around the globe. One of the key characteristics of an industrialized nation is fiscal and political stability. Developing nations are associated with monetary volatility and political upheaval. In order to ease pressures, many developing nations have decided to adopt policies similar to those used by industrialized nations in hopes of achieving similar stability and peace. Let us examine how monetary policy is used to achieve this goal.

Monetary Policy and Exchange Rate

The real exchange rate is used as a measure to assess the countries competitiveness as compared to other countries around the globe. Inflation tells us less about the fiscal health of the country than real exchange rate. Inflation is derived from monetary expansion, the devaluation of currency and other factors. One of the key difficulties is that different measures are used to report these indices. When one wishes to compare inflationary indices of different countries, one has to be certain that they are comparing like factors. This is one the difficulties in using a regression analysis to assess the relationship between two factors. Differences in the way inflationary indices are reported may affect the ability to draw conclusions about the relationship between exchange rate and inflation in developing countries.

Countries can choose several different structures for their exchange rate policy. They can use a fixed rate policy where their exchange rate is directly tied to that of another country, such as the United States or France. They can also choose a variable rate that that allows them much more flexibility to respond to shocks within their economy. This type of exchange rate is referred to as a floating exchange rate. A fixed rate policy ensures more stability but does not allow them to respond to situations as they arise.

The type of policy chosen depends on the specific needs of a particular country. The type of exchange rate policy directly affects the ability to draw a correlation between exchange rates and inflation. Countries that use a floating exchange rate would be more responsive to monetary shocks, such as rising interest rates or changes in GDP. Many developing nations use a fixed exchange rate. However, as they evolve toward becoming an industrialized nation, they will often adjust their exchange rate so that it is more responsive to the market.

In several cases, during the time period examined by this study, countries changed their fiscal policies. These examples are the most interesting as far as examining the correlation between exchange rates and inflation are concerned. These countries provide an example where one can examine the independent variable (inflation) before and after the adoption of new monetary policies involving exchange rates.

A prime example of a country that changed its exchange rate policy from a fixed rate to a floating rate is Kenya. Until 1974 Kenya's exchange rate was pegged to the U.S. dollar. However, after several devaluations they dropped the peg and decided to tie the currency directly to the markets. Kenya now has a floating exchange rate (Calvo and Reinhart, 2001). Many countries that were previously tied to the U.S. dollar have chosen to peg to a basket of currencies as a result of devaluation of the Dollar (Calvo and Reinhart, 2001). All of these changes have an effect on the analysis of exchange rate and inflation.

An understanding of the different types of exchange rate policies and how their effects on volatility are imperative in the development of a model for the conduct of this study. Fixed exchange rates are not reflective of the actions and reactions that are taking place within the country itself, but rather reflect the movements of the country to which the exchange rate is pegged. Fixed exchange rates are not as reactionary to changes within the country as floating exchange rates. Therefore only countries that use floating exchange rates will be examined.

International trade has caused many formerly third world countries to adopt an exchange that more closely follows the markets. Many are choosing a floating exchange rate rather than a pegged one. Morocco and Tunisia recently made this move. Until 1984 their markets were pegged to the French Franc, but in 1984 they decided to peg their exchange rates to the international markets (Grand and Dropsy, 2004). This has the effect of making their country more attractive to investors because there is a considerable chance for growth. However, when Latin countries made this move it led to currency overvaluation. To avoid this same situation Morocco and Tunisia developed a plan to manage the currency rate while the markets were being developed. The than from a pegged currency to a floating one requires carefully planning and is usually accomplished in stages.

There is another type of exchange rate that may affect this analysis. As countries are making the transition from a fixed exchange rate to a floating one, they may decide to use what is called a managed floating exchange rate. The terms of managed exchange rates differ from country to country. Managed exchanged rates allow the exchange rate to float with the market, but have mechanisms in place to prevent unwanted effects such as runaway inflation and a drop in the GDP.

The purpose of this research is to measure interaction between exchange rates and inflation, in order to effectively accomplish this task necessitates the most highly reactionary variable possible. Using countries that have a fixed or managed floating exchange rate do not offer a true measure of market reaction,. Therefore only countries that use a floating exchange rate will be used for this regression analysis. Special cases such as countries that changed their exchange rate policy will be examined both before and after the monetary policy change to examine the effects of the policy more closely.

There are many factors that cannot be controlled for in the course of this research, such as the methods used to measure inflation, GDP, and other market indices. These factors may affect the outcome of the results of this research. However, at the current time there is no standard method for reporting these factors. However, many of the countries conform to the standards set by industrialized nations and it can be assumed that the information is accurate. Data will be obtained from credible sources such as the OECD, WMF, and the U.S. Bureau of Labor Statistics.

Model Development

The purpose of this study is to determine the macroeconomic factors that contribute to changes in inflation such as economic fundamentals and policies. The second part of the research uses a Markov switching model with time-varying transition probabilities to capture the changes in inflation and their determining factors. This model was developed through the evolution of several previous studies and is considered to be relevant to the research at hand.

The Markov Switching model used is the result of similar studies by Bleaney (1997) and Blix (1999) that used a switching VAR model to obtain time-varying probabilities of inflation processes. A suitable model was examined in Dropsy and Grand (2000) that asked a similar question to the one being explored in this research using a similar data set. The Markov switching model used by them sufficiently describes the data set being used in this research model.

The data for this research is derived from country data among 26 countries, both developing and industrialized. All of the countries examined have changed their exchange rate policy several times to one of those discussed previously. Floating policies are the most interesting as far as this research is concerned. The period being examined in this research is from 1982 to 1998, which is the period when many of the changes took place.

At some time during this period the countries chosen had a floating policy. This period was isolated and selected for the regression analysis. The period before the after the changes were closely examined. Several countries changed back to a pegged exchange rate after several years of floating rates. One such example is Argentina, which changed back to a pegged rate after runaway inflation sent the economy spiraling (U.S. BLS, 2001).

The Markov model used for the analysis accounts for these changes using shifts from high inflation to low inflation. This model was preferred because it allowed the shifts to move in either direction, therefore they could sufficiently explain the data set being examined in this study with little modification. The results of this model help to measure the effectiveness of inflation target policy using currency rates as a means of control. Countries were able to apply various exchange rate models in an attempt to control inflation. Therefore, exchange rate policies act as the independent variable in this study. Inflation is reactionary to exchange rates, therefore will serve as the dependent variable.

The results of the Markov Switching model used for this analysis serve as a measurement of the degree of effectiveness of inflation target policy (Dropsy and Grand, 2004). As reported by Dropsy and Grand, this model has certain limitations that may hinder its effectiveness as an assessment tool. For instance, the model uses fixed transition probabilities from one policy form to another over the entire period of time being considered. Using a fixed approach to exchange rate has an averaging effect over time. The market responds to shocks and these may be lost due to this averaging effect.

The monetary market is dynamic and constantly changing. However, incorporating each and every change over a period of time becomes cumbersome. In addition, such detailed analysis makes it difficult to sort out trends from single events. This has been reported as a limitation of the Markov switching model. However, for the purposes of this research it may be an advantage rather than a hindrance. If a spike is found, then that particular set of data can be re-examined to determine if the change fits the category being examined in this analysis. Although the model does not typically account for shocks, it can still be used to examine them more closely. The Markov switching model as used in Dropsy and Grand (2004) is an excellent tool for examining the data set used in this analysis.

Dropsy and Grand overcame the obstacles by using an estimation of the changes from high inflation policies to low inflation policies using their transition probabilities. This model developed by Dropsy and Grand is a result of the expansion of several earlier models. It is based on the Markov-switching model with fixed transition probabilities developed by Hamilton (1989). This model, as used in Dropsy and Grand (2004) was expanded by allowing the transition probabilities to change over time using fluctuations of an information variable (Filardo, 1994, 1998).

An examination of these works further explains how these changes help to account for shocks. The changes introduced by Filardo are the key to making the model an effective instrument in the current study. The model used as the basis for this regression was originally used to measure the effects of policy changes in Morocco and Tunisia (Dropsy and Grand, 2004). The changes that took place in these countries were similar to the changes in other developing nations that were examined as part of this study. The economic profile of Morocco and Tunisia make them excellent testing grounds for the Markov-switching model using a similar data set to that being used in this study. Dropsy and Grand received reliable results using this model and it can be expected that it will perform as effectively on data from the countries used in this research due to the similarities in data.

The Model

An explanation of the Markov-switching model can be found in Dropsy and Grand (2004). However, it is important to explain how the various components of the model relate to the current research model and the interpretation of the data. The following explains how the model was applied to the data set in this research.

The first portion the model assumes that inflation can be derived using the following equation:

It the equation used by Dropsy and Grand ? is the inflation rate. The equation states that the statistical mean (?) and the variance (?) of inflation depend on the state of the economy (ts). In this modes a high inflation period is indicated as ts=0 and a low one is represented by ts=1. According to the authors of the model a high inflation period is characterized by much more volatility than the low inflation one. This agrees with the information obtained in the theoretical development of the model. This means that inflation variance can vary with the state of the economy.

Hamilton (1989) defined the following transition probability matrix:

Hamilton defined the transition probabilities as follows:

According to Hamilton's model, the probability of being in either one of the two states depends solely on the state in period t-1. A key limitation of Hamilton's model, as noted in Dropsy and Grand (2004) is that this matrix makes the probability of being in a particular state constant over time. This is where Filardo (1994, 1998) improved on the model. He allows transition probabilities to change over time using an indicator variable zt as follows:

"In this revision, p is the probability of staying in a high inflation regime, q is the probability of remaining in a low inflation regime and Zt = {zt, zt-1,...} is the set of exogenous variables considered to predict the future course of inflation," (Dropsy and Grand, 2004).

Filardo's revision to Hamilton's model estimated the transition probabilities (p) and the parameters of the equation concurrently. If 0-0, 1, = = I i, then the transition probabilities are fixed as in Hamilton's model (Dropsy and Grand, 2004).

This is the revision that allows model to be used on the current data set. This revision allows us to consider the factors that influence the high inflation period. As one will recall, one of the key limitations of the regression study is that even though one event closely follows another it does not mean that causality can be assumed. This revision of Hamilton's model allows us to better determine causality and the effects of influences other than the dependent variable.

Empirical Results

Using data from Calvo and Reihart (2001) 25 period of floating monetary policy were found among the 26 countries considered for the study. Inflation data was obtained from a compilation of the International Monetary Fund (IMF, 2002).

The primary goal of the research was to determine if a relationship exists between exchange rates and inflation. In order to do this we had to isolate the type of exchange rate policy that is considered to be the least volatile and the most reflective of the country being examined.

Not all of the countries considered in the final analysis were found to have floating policies, or at least during the period in question. For instance, France, Greece, Germany, Egypt, Columbia and Chile have never used a floating exchange rate policy (Calvo and Reinhart, 2001). For those that did have a floating policy for some time they later switched to a managed or limited floating policy. A small number switched back to a pegged policy after a period of floating policy, but most adopted a managed floating policy after the floating policy period ended (Calvo and Reinhart, 2001).

Using the regression model, it was found that floating policies resulted in high periods of high inflation that began shortly after the adoption of the floating monetary policy. Both exchange rate volatility and inflation were found to increase after adoption of a floating exchange rate policy. Average inflation rates for the high periods of ranged from in the hundreds to rises of over 3,000 in extreme cases. Low average rates remained under 5%.

The data revealed regional differences in characteristics of the regression results. For instance, South American countries were found to be more highly reactionary than established European countries. Currency exchange rate data was obtained from Economics Web Institute (2005). The most significant finding from this analysis was the there were vast differences in the reaction of inflation within 1 year following the change in exchange rate policy. However, a change in exchange rate policy was followed by volatility in inflation rates in a majority of the cases. It would appear from this initial analysis that inflation rates are affected by exchange rate policy and the resulting real exchange rates. However, the degree of the reaction cannot be determined and will every country will react differently to the policy changes.

These differences in reactions to exchange rate policy are perhaps the most interesting phenomenon in the study. One must ask, not if the country will react, but why the results are so different across seemingly similar developing nations. This brings us to look at the mechanisms that cause inflation in the post-policy change period. Our economic model allowed us to examine some of these features to gain a greater understanding of the mechanisms that drive these changes.

One of the factors that we have examined is if there are regional differences in the reactions to floating exchange rates. In order to examine this further we grouped the data together according to continent. An average of the data was used to re-calculate the Markov model. When this was completed, drastic regional differences were found.

The data was once again categorized according to countries with similar tax rates and average GDP growth, When this was completed the results demonstrated that countries with similar tax rates were more closely related in terms of reactions than those with similar GDP growth. Tax rates are another mechanism that can be used to influence the economic health of the country and encourage higher levels of foreign investment. Those countries with lower tax rates were found to experience lower levels of inflationary growth after the currency was no longer pegged to the U.S. dollar. The most dramatic case in point was Mexico, which at the time it decided to switch to the floating exchange rate two times during the study period. In both cases the country had excessively high tax rates.

There was another variable that further complicates the situation. Mexico was preparing to enter into the North American Free Trade Agreement. It was assumed that they needed the extra tax money for improvements to their infrastructure. However, according to the Worldbank, NAFTA did not give Mexico enough reason for the excessively high tax rates that they were experiencing (Chen, Martinez-Vasquez, Worldbank, 2001).

According to the analysis conducted as a part of this research Mexico experienced runaway inflation after the first switch to a floating exchange rate. Their currency devaluated almost 38% against the U.S. dollar (U.S. BLS, 2006). The second switch to a floating exchange rate only devalued their currency by approximately 7.5% (U.S. BLS, 2006). This case supports the hypothesis that changes in currency exchange rates are related to inflationary changes. However, the case with Mexico also indicates that although these two factors may appear to be closely linked there are other factors involved.

An exploration of the factors involved in the different reactions to changes in exchange rate in Mexico highlight the apparent relationship between exchange rate and inflation. However, it also illustrates the inability to draw a causal relationship between the two factors. There are other factors involved that can influence the degree of reaction in the country in question. If this is the case, then one must consider the possibility that these same factors could serve to mask or enhance the effects of exchange rate on inflation as well.

An exploration of the trade balance reveals an interesting phenomenon. When one examines periods of high inflation, it was associated with a trade deficit in a majority of cases. This analysis focused on Mexico because they had the most dramatic changes and had several identifiable factors at the same time. Mexico followed the pattern of changing from an exporter to a major importer after the adoption of a floating monetary policy (WTO, 2006). The patterns found in Mexico help to shed light on the circumstances that could have contributed to similar results in other countries. Many factors in Mexico were blamed on NAFTA, but in reality, they were probably going to happen with or without NAFTA.

Every country represents a unique situation. Although there does seem to be a propensity to experience a period of high inflation following changes currency rate, our analysis suggests that this single factor cannot fully explain the differences in reactions that were observed in different countries. Mexico provides a unique cast study in that hey have switched exchange rate systems frequently. In a majority of the countries identified, they only exhibited one period of true floating exchange rates. Most found that a managed system was needed to control runaway inflation. Mexico was one of very few examples where we could examine the switch twice within the study period.

Markov Switching methodology was chosen due to the respective patterns of switching between high and low inflationary periods that could be observed closely following periods of exchange rate policy. This methodology revealed that there are many factors that tend to follow similar patterns following a change in exchange rate policy. These reactions occurred a significant number of times to make them predictive in nature. This predictability only goes as far as to be able to predict that a period of high inflation will follow, but it cannot predict how large or small the impact will be.