Interpreting Forward Rates Under Different Measures

Summary

The document asks how to interpret a forward rate for a future period when the rate is observed today. It points out that the current forward rate is generally not the real-world expected future rate, even though an expectation relationship holds under the corresponding forward measure. This distinction matters because model equations under a pricing measure do not directly provide forecasts under the real-world measure.

It also asks why the bond price implied by the forward rate can serve as a discount factor when it is not the real-world expected future bond price. The text raises these conceptual questions but supplies no answers, derivations, examples, or empirical evidence. It therefore identifies important issues in interest-rate modeling without establishing how discounting is justified or how the measure change affects practical forecasting. Readers will need additional material on risk-neutral pricing, numeraires, and forward measures to resolve them.

Key ideas

• A forward rate observed today need not equal the real-world expectation of the future rate.
• The expectation relationship for a forward rate depends on the choice of probability measure.
• The document questions why an implied zero-coupon bond price is a valid discount factor.
• The questions are posed without explanations or supporting examples.

Tags

# How should I interpret a forward rate?


# How should I interpret a forward rate?












Let $L(t, S, T)$ denote the forward rate from time S to T observed at time t, assuming t < S < T.

A lot of modelling work is centered around this rate, but how is this rate useful? How are we supposed to interpret it?

a. Under the real world measure, $E_t[L(S, S, T)] \neq L(t, S, T)$. So this rate is not a good predictor of L(S, S, T). In fact, the above-mentioned equation holds only under the T-forward measure, but how should we think about it if the equality is under the forward measure?

b. Let P(t,S, T) denote the zero coupon bond price implied from L(t, S, T). Again P(t, S, T) is not the expectation under the real work measure of P(S, S, T). What justifies using P(t, S, T) as a discount factor?

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.