Inferring Rate Expectations from Fed Funds Futures and OIS
Summary
The document outlines how to estimate market-implied policy-rate changes using fed funds futures or overnight indexed swap rates. Fed funds futures settlement reflects an arithmetic average of overnight fixings, weighted by the number of days each fixing applies; this calendar weighting matters when interpreting monthly contracts. An OIS curve can also be read by assuming the overnight fixing stays constant between effective policy-change dates and accounting for any spread between the fixing and the benchmark rate.
Under assumptions that policy changes occur only at scheduled meetings and are in fixed increments, the difference between successive meeting-to-meeting forward rates can be translated into an implied probability of a change. In several currencies, liquid forward OIS contracts spanning adjacent meetings can isolate that interval. OIS rates more broadly provide risk-neutral expectations of cumulative overnight rates, but extending them into policy-rate probabilities requires assumptions about the link to the target rate, expectations, and rate variance. These are market-implied measures, not certain forecasts.
Key ideas
- Fed funds futures settle on a day-weighted arithmetic average of overnight fixings.
- OIS forwards can be interpreted by holding the overnight fixing constant between effective policy dates.
- The fixing may differ from the policy benchmark, so that spread should be tracked.
- Meeting-to-meeting forward rate changes can imply policy-change probabilities if meeting timing and step sizes are assumed.
- OIS rates express risk-neutral overnight-rate expectations, and converting them into policy expectations requires additional assumptions.
Tags
Full text
# How to Calculate (month by month) what hikes are priced into OIS? # How to Calculate (month by month) what hikes are priced into OIS? I am looking to see on a monthly basis, what the market is pricing in in terms of rate hikes in the OIS. Would it be sufficient to look at fed funds futures at those monthly dates? My gut says not. It would be great to understand better how the OIS (and OIS futures curve?) could be used to see how many hikes are priced in at any frequency out to any horizon ## Answer by FinanceGuyThatCantCode (score 1) https://quant.stackexchange.com/a/33195 Fed funds futures are nearly sufficient. You need to know the precise way in which Fed Fund futures are calculated for settlement purposes - e.g. that it is an arithmetic average of the fed funds fixings where a fixing is weighted by the number of days between the fix and the next fix (i.e. a Friday fix will be weighted 3 times if there is no holiday on Monday). OIS curves are great too. Easiest thing to do would be to assume that on average, the fix will be constant between CB meeting dates (or the effective dates of rate changes really - In Europe, the effective date can be 6 or 7 days after the meeting, but otherwise effective date is usually the next day). Also, good to keep track of where the fixes are relative to the benchmark rate as there can be a few bps difference at times. Then if you make the assumption that there will be no inter-meeting rate changes and that rate changes can only be 25 bps (for the US) - then if the forward rate between a given central bank meeting date and the next meeting date is $n$ bps higher than previous such forward rate, you might say that the market is implying an $n/25%$ chance of a rate change during that meeting. In EUR, GBP, AUD, NZD, and CAD, there are reasonably liquid forward OIS contracts that do exactly this - they start on one meeting date and end at the next meeting date so they purely capture the central bank's implied prob without you having to do any interpolation or handling of the evil interest rate conventions, ## Answer by Igor Pozdeev (score 0) https://quant.stackexchange.com/a/39315 We have an elaborate discussion of different OIS in our paper (link). In short, the OIS can give you the risk-neutral expectation of what the cumulative overnight rate is going to be in, say, 6 or 9 months. Bad news is, you cannot go further without assumptions: that the overnight rate is the same as the policy target rate, that the risk-neutral and natural expectations coincide in this case, and that the variance of the overnight rate is known. Good news is, many of these assumptions are not as restrictive as they seem! You can work off the formula in the paper, modifying it to the multi-period setup.
Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.