Stock Market Reaction to Monetary Policy: An Event Study Analysis of the Brazilian Case

ABSTRACT

This article examines the relationship between the monetary policy implemented by the Central Bank of Brazil and the stock market. We implement event study analysis and analyze the effect of the anticipated and unanticipated components of monetary policy decisions on the returns of the IBOVESPA index and 53 stocks. We find that monetary policy has a significant effect on the stock market, but is only responsible for a small proportion of market variation. The analysis at the sector level with expected returns identifies that the financial sector is the most affected by this policy, whereas with excess returns only industrial goods are significantly affected. Moreover, individual assets respond in a rather heterogeneous fashion to monetary policy; however, when we look at excess returns, we identify a reduction in the intensity and in the number of companies impacted by monetary policy. Finally, the monetary shock is explained by unanticipated variations in the unemployment rate, in the Industrial Production Index, in the General Market Price Index, and in the Broad Consumer Price Index.

KEY WORDS:

• Brazil
• monetary shocks
• stock market

Notes

¹. The Federal Open Market Committee is the US central bank committee in charge of establishing the base interest rate for the economy, among other affairs.

². Reversal indicates that a change occurred in the direction of the SELIC rate between one meeting of the COPOM and the previous one.

³. The IBrX is a price index that measures the return of a theoretical portfolio made up of 100 stocks selected from among the most negotiated ones on the IBOVESPA in terms of the number of deals and financial volume. These stocks are weighted in the index portfolio according to their respective number of stocks available for trading on the market. Source: BM&FBovespa.

⁴. Statistical influence is a method of identifying possible outliers. This kind of statistic seeks to measure the difference in the results of a regression because of a unique observation (i.e., how different is an observation from the others in the sample used as an estimate).

⁵. The RStudent measure is numerically identical to the t-statistic, which ends up including a dummy variable in the original equation, where 1 is for observation i and zero otherwise. This may be interpreted as a test of the significance of an observation.

⁶. The initial date for gathering the stocks aimed at having 6 months of history for returns before the first COPOM meeting to estimate the market model and excess returns from the stocks in January 2005.

⁷. This crisis practically froze the interbank markets and the repo and securitization markets of the developed world, causing greater damage than had been predicted by the stress tests of many central banks and major players in the international financial market. In a chain reaction, the liquidity crisis evolved into a systematic risk crisis, contaminating developed financial systems and leading them and even national economies to bankruptcy.

⁸. We performed a robustness exercise using normal returns subtracted from S&P500 index as control variable. Using the excess returns, we obtained six stocks with significant responses to the monetary shocks. With the S&P500 as a control variable, we also obtained six significant stocks, of which three (FIBR3, RAPT4, and UGPA3) with significant responses in both methodologies. An explanation for similar results is the relatively high correlation (average of 55%) between the returns of the S&P500 and the expected return for each stock, given by the market model, in each day of implementation of the new SELIC rate.