"The Fed fund futures predict an eighty percent chance of a quarter-point rate hike at the hike at the next FOMC meeting," someone on CNBC intones. How is that calculation made?
The calculation utilizes the thirty-day Fed funds futures. The CBOT introduced its Fed funds futures in late 1988. Fed fund futures are listed for the current month and the 24 succeeding months. QCharts employs FF as the symbol for these futures, and then appropriate symbology for the year and futures-expiration month would also be used. Therefore, on QCharts, the June 2006 Fed fund futures would be FF06M.
This June contract expires the last business day of the delivery month, on June 30, with trading closing at 2:00, Chicago time. On the CBOT site (www.cbot.com), a sidebar on the right-hand side allows you to follow a link to the thirty-day Fed fund futures and find more information such as tick size, price quote, settlement and trading hours. However, be aware that these futures are subject to volatile moves in jittery markets, especially times when any FOMC member might be slated to speak. This article isn't meant to encourage you to trade them, but rather to understand all those predictions you hear.
The futures are associated or aligned with the FOMC's targeted rate, but market forces also impact them. Other sites offer information on how the Fed uses its open market operations to supply the reserves that keep the Fed funds aligned with the targeted rate, but that's not the purpose of this article, either. We're interested in how the futures are used to make that prediction. On June 12, when this article was originally prepared, the June Fed fund futures closed at 94.9650. To calculate the suggested interest rate at the delivery of that contract at the end of June, the Fed funds futures value would be subtracted from 100, resulting in a suggested rate of 5.035 (100 - 94.9650).
On the same date, the July Fed fund futures closed at 94.79, resulting in a suggested rate of 5.21 at July expiration on the last business day of July. That's a spread of 0.175 or 17.5 basis points between the June and July contracts, as of June 12 at least. According to the CBOT site, it's normal to see a spread of 4-8 basis points (or 0.04-0.08) from one month's contract to the next. If the spread in one month's Fed futures rate and the next month's is less than 9 basis points (or 0.09), the market is not anticipating a change in the Fed's policy. If that spread widens to a range of 12-15 basis points, the market does anticipate that the FOMC will change its target rate. Clearly, on June 12, markets were anticipating a hike in interest rates, but we all knew that, didn't we?
How much of a hike were the futures predicting, and what was the probability assigned to that hike? First, it's important to know that there is no July FOMC meeting. Checking http://www.federalreserve.gov/fomc/#calendars for a schedule of meetings turns up that information. That means that those July futures must have been pricing in a chance that the FOMC will raise rates in late June since they were predicting a higher interest rate by the expiration of those July futures. Since the Fed changes rates in increments of 0.25 or 25 basis points, those futures were pricing in a chance of a 0.25 hike in rates, not a 0.21 hike.
How can we determine the probability of that hike? Dividing 0.21, the spread between the current Fed funds target rate set by the FOMC and the future rate predicted by the July contract, sets that probability. Making that calculation determined that on June 12, the market considered the likelihood that the FOMC will raise rates by 0.25 or 25 basis points to be 0.84 or 84 percent. John B. Carlson, William R. Melick, and Erkin Y. Sahinoz, writing "An Option for Anticipating Fed Action," (http://clevelandfed.org/research/Com2003/0901.pdf) also suggested using futures other than front-month ones (June, in this case) to make predictions of the Fed funds target rate and suggested a further out futures contract, one two months out.
If the consensus on June 12 was that the FOMC would either pause or raise by a quarter basis point, as it seemed to have been at the time, then the probability that the FOMC will pause would be 16 percent (100 percent minus the 84 percent the market has assigned to the probability that a 25 basis-point hike will occur). If the consensus is that the FOMC will either raise 25 basis points or 50 basis points, then that 50-basis-point hike is also assigned a 16 percent probability.
What if there had been no market consensus on June 12, and some had believed that there would be a pause; some, a 25-basis-point hike and others, a 50-basis-point hike? Those would have formed three possibilities to be divided among a 100 percent probability.
Then it becomes more difficult to make those predictions. Some experts suggest that even when only two possibilities are suggested, it's perhaps simplistic to use the Fed funds futures in the manner described above to predict the possibility of a certain hike. If those futures were to suggest a 30 percent chance of a hike, for example, do you know whether most were pricing in only a 20 percent chance of a hike, but then a few went out to a 80 percent chance, so that the average would work out to 30 percent although fewer futures buyers believed in a hike?
Since the introduction of options on the Fed funds, many have begun suggesting that options might be better used to make these predictions. Those options allow for an examination of how the spread in prices across several strikes differs from the actual differences in prices. For example, if there's a relatively big difference in the prices of a lower-strike call option and a higher-strike one when compared to actual difference in the strikes, options buyers are betting that the Fed funds futures will settle above the lower-priced call but below the lower-priced one, some suggest. If the spread in prices for the two call strikes is near the difference in the two strikes, they're betting that the Fed fund futures will settle above both. The use of options does not restrict calculations to two possible outcomes since several strike pairs could be examined. A bell-curve of probabilities can be obtained.
Others suggest another problem with using the simple calculation explained above to predict the Fed funds target rate the FOMC will set at the next meeting. In "How Well Does the Federal Funds Futures Rate Predict the Future Federal Funds Rate?" (http://clevelandfed.org/research/Com2003/0901.pdf), Ed Nosal sets forth a conclusion that Fed funds futures rates may overestimate the future target Fed funds rate when those rates are falling and underestimate the future target Fed funds rate when rates are rising. Nosal moved four months out from the current month to remove bias, further than Carlson, Melick and Sahinoz's two months to accomplish the same purpose.
Despite the potential problems with using the Fed funds futures to predict FOMC moves, it's helpful to see what those futures suggest and understand all those probabilities being discussed. Want to know more or read it described in different words? Check out the Federal Reserve Bank of Cleveland's site at http://www.clevelandfed.org/research/policy/fedfunds/faq.cfm for more details and examples of how these calculations are made. Then you, too, will know what the Fed fund futures predict.