THE BOND MARKETS: Estimation

The FOMC meets 8 times each year. There are two types of meetings of the FOMC: one-day meetings and two-day meetings. There are usually 2 two-day meetings and 6 one-day meetings every year. Our sample over the ten-year period consists of 58 one-day meetings and 22 two-day meetings. In total there were 80 meetings, i.e. one every 6-7 weeks. The first and fourth meetings every calendar year are two-day events. They begin at 2:30 pm on the first day, continuing at 9:00 am the following day. The one-day meetings always begin at 9:00 am. All meetings begin on Tuesdays.

At these meetings, the FOMC examines information about the economy and decides on whether to undertake open market operations in the dollar or other currencies. They also determine the level of short-term rates. The usua^ issues relating to the economic outlook are considered: consumer spending, non-farm payroll, industrial production, retail sales, real business fixed investment, nominal deficit, consumer price inflation, currency rates, money supply (М2,М3), and housing activity. At the two-day meetings additional policy directives are issued. In particular, these relate to domestic open market operations, authorization of foreign bank limits for foreign currency operations, foreign currency directives, and procedural instructions with reference to foreign currency operations. We find that the two-day meetings appear to have a greater information impact than one-day meetings.

In addition to foreign currency directives, the Fed also undertakes other distinct activity at the two-day meetings. By (the Humphrey-Hawkins) law, the Fed must report to Congress twice a year on monetary policy, i.e. in February and July. The two-day meetings are the setting for the discussions on monetary policy as well. The FOMC thus votes on the range of growth rates of М2,М3 and the debt levels it expects to see.

Thus, two-day meetings tend to evidence more forward-looking discussions than usually occur at one-day meetings. However, these votes are not announced immediately, and only get reported in minutes two weeks after the meeting. Thus, it is not clear that this activity of the Fed in any way forms an information event. However, we do find that the two-day meetings seem to impact parameter estimation, in contrast to the one-day meetings.
w6631-15
To begin, we carry out a few simple regressions to ascertain if the volatility of interest rates is in any way related to information released at FOMC meetings. This is done by regressing the squared change in interest rates on interest rate level and a dummy variable for the FOMC meeting. The regression equation is as follows:

[7’г+1 —rt]2=a-\-brt-\- cft+i + Ct+1

where ft is the dummy variable indicating the FOMC meeting. It may take four different forms as described in Table 6 below. Since some of the meetings last 2 days, combinations are possible. First, we assign a dummy variable which is the first day of all meetings. As can be seen, this has little impact on volatility, and hence provides evidence of no unexpected information, A similar result holds when we examine only 1 day meetings. However, when we set the dummy variable to be the first day of a 2-day meeting, the coefficient comes in strongly positive. This indicates that there may be a significant information release on the first day of the 2-day FOMC meetings. We also examined whether the information impact occured on the second day of the 2-day meeting and found little effect. Thus, if there is an information effect, it occurs on the first day of the two-day meeting. Table 6 summarizes the regression results.