Introduction

Project Hypothesis

This project is intended to develop a top-down allocation model for Latin American countries that will out-perform a Latin American benchmark. Latin America was chosen specifically because it is an emerging market with a history of being quite volatile. The particular countries that were used within our model were Mexico, Chile, Venezuela, Brazil, and Argentina. Our intention is to capitalize on the volatility of these countries by using our model to identify money-making opportunities as well as hedge against any negative market movements. Our interest in this project arose from the recent Brazilian currency crisis and the possible application of a top-down model to protect portfolios against possible future devaluation.

Latin America Benchmarks

The benchmarks used for Latin America were the:

  • International Finance Corporation Latin America U.S. Dollar Price Index, and
  • MSCI Latin America U.S. Dollar Price Index

These files provided the best data of Latin American returns in which our top-down allocation model could be tested.

Out of Sample Testing

We purposely did not include the monthly data from 1999 so that we could use this data to perform an out-of-sample test of our model.

Project Details

Country Information

The countries of Mexico, Venezuela, Chile, Brazil, and Argentina were selected to represent Latin America in the top-down allocation model. The following is a snapshot of each country:

INSERT COUNTRY SPECIFIC INFO HERE

Country Correlation Matrix


The correlation matrix shows a significant relationship of each country to the other. As a result, each country can be used to hedge against the other by exploiting the high correlation to reduce the portfolio risk and receive a return premium.

Selection of Economic Variables/ Comments on Coefficient Signs

Asset Class / Variables/Economic Reasoning Behind Selection
Chile Equities / One month lagged variation of International Reserves for Chile-
The International Reserves of Chile represent the money held by the government to support the economic transactions of the country including the financial obligations to other countries. The stock returns tend to fluctuate with the supply of international reserves; hence, this variable is positively correlated with the stock returns.
Two month lagged variations on Unemployment-
The variations on unemployment are an indicator of the expected economic activity within Chile. As the economy improves and the stock market rises, as it has been doing, unemployment falls. As a result, this variable is negatively correlated with the stock returns.
We use a 2 month lag structure for this variable because of the government’s timing for announcing the number. For instance, January oil exports are announced on February 29th, thus impacting March stock returns.
One month lagged variations on Mexico’s Returns-
The variations on Mexico’s returns act as an indicator of the economy of Chile. As the Mexican economy thrives, the economy in Chile booms as well. This is due to the fact that a free trade agreement exists between Mexico and Chile, and Mexico is considered the ‘gateway’ country into Latin America. In other words, as international activity is initiated in Latin America, Mexico is usually the first country entered. After Mexico, international activity reaches out to other Latin American countries, which includes Chile. Due to this relationship, this variable is positively correlated with the stock returns of Chile.
Venezuela Equities / One month lagged variations on Exchange Rates-
This exchange rate represents bolivars to dollars. When the percentage change in exchange rates is negative on a month to month basis, this indicates that the bolivar has appreciated against the dollar. As the bolivar appreciates, the stock returns grow positively. This variable is indicated by a negative coefficient, which represents the relationship between the change in exchange rates and the stock returns.
Two month lagged variations on International Lliquidity-
International liquidity is a measure relating to the country’s liability and also to their monetary reserves. As the economy experiences positive growth, the country becomes more liquid; therefore, this variable produces a positive coefficient.
We use a 2 month lag structure for this variable because of the government’s timing for announcing the number.
Two month lagged variations on Oil Exports-
The exportation of oil is a huge factor in Venezuela. As oil exportation increases, the growth in the economy of Venezuela also increases. For this reason, the country’s stock returns are positively tied to the exportation of oil.
We use a 2 month lag structure for this variable because of the government’s timing for announcing the number.
One month lagged variations on Price Earnings-
The alterations in P/E ratios can bring investors attention to the market showing over or under estimations of the current prices. As the P/E ratios increase, the country believes the economy is growing strong due in part to oil exports; therefore, they are more willing to put funds into Venezuela’s stock market.
One month lagged variations on U.S. Risk Free Rates-
The variations on US risk free rates are an indicator of expected economic activity and inflation on US. As the rates in the U.S. increase, Venezuela receives a monetary boost from their exports, which in turn helps the economy to grow. As the economy grows, the stock market benefits; hence, there is a positive correlation between Venezuela’s equity returns and the U.S. risk free rates.
Mexico Equities / One month lagged variations on International Investor Credit Rating-
The international investor credit rating is a measure of one country versus another in terms of riskiness. The higher the credit rating, the more willing investors are to add a particular country to their portfolio. In Mexico, as the credit rating improves, the investors pour more money into the country. As a result, the country’s stock returns grow positively. Therefore, this variable is represented by a positive coefficient.
One month lagged variations on Price Earnings-
The alterations in P/E ratios can bring investors attention to the market showing over or under estimations of the current prices.
As the P/E in Mexico increases, investors become wary and believe the stock market is overvalued; therefore, investors start to pull out of the market, and Mexico’s stock returns drop. This relationship is represented by a negative coefficient.
One month lagged variations on Exchange Rate-
This exchange rate represents pesos to dollars. When the percentage change in exchange rates is negative on a month to month basis, this indicates that the peso has appreciated against the dollar. As the peso appreciates, the stock returns grow positively. This variable is indicated by a negative coefficient, which represents the relationship between the change in exchange rates and the stock returns.
One month lagged variations on Mexico’s Interest Rates-
The variations of interest rate on last month contain expectations of future inflation, and are an element of purchase decisions. It is believed that the positive correlation of a rise in interest rates with a rise in the stock market returns is due to the fact that the Government has successfully contained inflation in the past. Therefore, as the government raises rates in an effort to curtail on-coming inflation, the market reacts positively and shows increased returns.
One month lagged variations on Price/Book Value-
This variable is used as a measure of the overall “value” or “expensiveness” of the Mexican market. We believe this variable is actively used by portfolio managers around the world in making asset allocation decisions amongst countries. Therefore, we expect the variable to have some predictive power in estimating future inflows and outflows of capital to the Mexican market.
As the price to book value ratio increases, investors believe that the market may be overvalued and not an attractive investment. Investors start to pull out of the market, and Mexico’s stock returns drop. This relationship is represented by a negative coefficient.
Brazil Equities / One month lagged variations on Brazil Price/Book Value divided by Latin America Price/Book Value-
This variable is used as a measure of the overall “value” or “expensiveness” of the Brazilian market. We believe this variable is actively used by portfolio managers around the world in making asset allocation decisions amongst countries. Therefore, we expect the variable to have some predictive power in estimating future inflows and outflows of capital to the Brazilian market.
As the price to book value ratio increases as a percentage of the Latin America price/book value, investors believe that the market may be overvalued and not an attractive investment. Investors start to pull out of the market, and Brazil’s stock returns drop. This relationship is represented by a negative coefficient.
One month lagged variations on Brazil Price/Earnings divided by Latin America Price/Earnings-
The alterations in P/E ratios can bring investors attention to the market showing over or under estimations of the current prices. Although a positive correlation is counter-intuitive to what we would expect given the negative correlation of the above price/book value ratio as a percentage of Latin America’s price/book value ratio, we believe that as the price/earning ratio of Brazil increases as a percentage of Latin America, investors see this as a positive sign of growth. Therefore, more funds enter into Brazil, and as a result the stock markets returns increase.
One month lagged variations on Exchange Rates-
This exchange rate represents reals to dollars. When the percentage change in exchange rates is negative on a month to month basis, this indicates that the real has appreciated against the dollar. As the real appreciates, the stock returns grow positively. This variable is indicated by a negative coefficient, which represents the relationship between the change in exchange rates and the stock returns.
One month lagged variations on Price Earnings-
The alterations in P/E ratios can bring investors attention to the market showing over or under estimations of the current prices.
As the P/E in Brazil increases, investors become wary and believe the stock market is overvalued; therefore, investors start to pull out of the market, and Brazil’s stock returns drop. This relationship is represented by a negative coefficient.
Argentina Equities / One month lagged variations on Bond Yield-
Bond yields movements are equivalent to the movement in interest rates. As the interest rates are falling, investors see this as a positive sign to enter the stock market. Hence, there is a negative correlation between the bond yields and stock market returns.
One month lagged variations on Inflation Rates-
Inflation is an indicator of the robustness of economic activity. Normally, we would expect a negative correlation between a rise in inflation rates and the returns on a stock market; however, Argentina is experiencing a positive correlation. We believe this is due to the fact that as inflation increases, the prices on the stock market follow suit by increasing as well. This is reflected by positive returns on the stock market. Over time, as Argentina becomes more developed, we would expect this trend to reverse and the coefficient to become negatively correlated with the stock market. This is due to the fact that the Argentina government will eventually gain credibility in its ability to control inflation.
One month lagged variations on Brazil Equity Returns-
The variations on Brazil’s returns act as an indicator of the economy of Argentina. As the Brazilian economy thrives, the economy in Argentina should boom as well, which should be reflected in with a positive coefficient. This is due to the fact that Brazil is considered the ‘gateway’ country (with Mexico) into Latin America. In other words, as international activity is initiated in Latin America, Brazil and Mexico are usually the first countries entered. After Brazil, international activity reaches out to other Latin American countries, which includes Argentina. However, our results indicate a negative correlation between the lagged returns in Brazil and the returns in Argentina. The economic reasons for this negative correlation is beyond our knowledge, and further investigation into this result should be conducted.

Asset Class Regressions


Summary of Predictive Regressions

Variable Correlation Matrices

Chile Correlation Matrix


Mexico Correlation Matrix


Brazil Correlation Matrix


Argentina Correlation Matrix


Venezuela Correlation Matrix