THE BOND MARKETS: ANALYTICAL METHODOLOGY 7

w6631-14

Day of the Week Effects. In this section we examine whether jumps are more likely to occur on specific days of the week, by introducing a modification to make the arrival intensity of jumps depend on the day of the week. There are several reasons which make jumps more likely on some days of the week rather than others. For example, jumps would be more likely on Mondays since the release of pent up information over the weekend may drive up the possibility of a large change in interest rates. Likewise, option expiry may inject jumps into the behavior of interest rates, and this would be likely on Wednesdays and Thursdays. Jumps may also occur on Fridays when last minute trading may create excess volatility.

By using dummy variables for each day of the week, we assume a linear model for the arrival intensity of jumps in the short rate of interest, where Ao is the arrival probability of a jump if the day is Friday, and А*, г = 1. 2, 3. 4 is the incremental arrival intensity of jumps over Friday’s level when the day of the week is Monday, Tuesday, Wednesday and Thursday respectively, d,, i — 1,2,3,4 are dummy data variables indicating the day of the week for Monday, Tuesday, Wednesday and Thursday respectively. Estimation was conducted over the two models containing jumps, i.e. (i) the jump-diffusion model and (ii) the ARCH-jump-diffusion model.

The results of the estimation are presented in Table 5. Intuitive results emanate from this analysis. The likelihood of jumps is highest on Fridays, but jumps are also likely on Wednesdays and Thursdays, when information from options expiry is released. This lends credence to the proposition that jumps are caused by large bursts of information being released into the market. Once again, there is no evidence of skewness, but kurtosis exists. The jump tends to be of the order of 50 basis points.

Federal Reserve Activity. Jumps may arise from intervention by the Federal Reserve in the bond markets. The Federal Open Market Committee (FOMC) meets periodically, and informs their open market desk of the range they wish to establish for the Fed Funds rate. Short rates tend to track this rate rather closely. Existing models do not account for Federal Reserve activity. It is possible that these meetings form an important information event. If so, a model that accounts for this will prove to be superior for traders. In this section, we enhance our jump model by making the jump intensity depend on the FOMC meeting. By examining the impact of the meeting on the jump probability we can ascertain whether the meeting is a significant information event.