Does the stock market under-react to the Federal Reserve bank's monetary policy actions?

Review of Business Research, March, 2008 by Levon Goukasian, L. Keith Whitney

4. THE DATA AND OUR ANALYSIS

First, we are grateful to Daniel Grombacher of the CBOT and Ray Sasaki of Sempra Energy for providing data and valuable information on Federal Funds Futures. This research was possible only because of their generosity.

In Table 1 we report the FFTR levels, the expected and unexpected changes in the rate for all the days for which we study the stock market's reaction. This covers 156 event-days from June, 1989 to January, 2006, the period of our study. We also comment on the calculation of the surprise changes on the days that are at the beginning, at the end of months, or in some other unusual circumstances. Table 2 reports the descriptive statistics of the changes in FFTR and shows the distribution of the FFTR for three different time periods. Out of all the 156 event-days, there were no changes made to the FFTR 86 times, once the FFTR was raised by 75 basis points (bps), and 12 times it was cut by 50 bps. The vast majority of FOMC decisions were announced on FOMC meeting days: 135 out of 155.

We also report the statistical average and the standard deviation of the FFTR. By comparing two time periods, it is clear that the surprise change in the FFTR is less volatile in post-1994 period than pre-1994 (when policy actions were less transparent and not generally made in the FOMC meeting). We also study the 86 days on which there were no changes in FFTR. In 46 of those days the market was expecting an action by the Fed, so no change was considered an unexpected change that triggered a reaction by the stock market.

5. THE MARKET REACTION AROUND THE FED DECISIONS

Finally, in order to study any deviations from EMH's expected "unpredictable" market reaction to the changes in the monetary policy, we follow the standard event-study approach. Our event-days are the days on which the decisions were made on FFTR. There are some exceptions: on some days the decision of the rate change was announced after the market close. In those cases we consider the next business day as the event day to capture the market reaction to announcements and subsequent behavior. We used the notes provided to us by Ken Kuttner, our own investigation, and Kuttner (2001) to resolve the timing issues of some of pre-1994 days, on which the news on the policy was out after the close of the futures market. We test whether abnormal returns immediately before and after the event-days are zero, for example. We compute then abnormal and cumulative abnormal returns to see if the prices adjust to the announcements quickly and appropriately. If prices adjust to new information in efficient ways, there should not exist any predictable ways to trade for profit in the market. As noted before, in our study we use the broad market index, the CRSP Value-Weighted Index. (We also studied the reaction of the CRSP Equally-weighted index and found similar results.) Under the no-predictability or efficient markets hypothesis (EMH), there is no predictable way of earning statistically significant returns. That is, we test the following hypothesis: