Expected Returns: An Investor's Guide to Harvesting Market Rewards

4

Road map to terminology

I briefly review many definitions, distinctions, relations, and pitfalls. (Readers looking for the “meat” may skip this chapter and later use it as a reference section, as needed. Readers only interested in empirical evidence may jump all the way to Part II.)

4.1. Constant or time-varying expected returns?

4.2. Rational or irrational expectations formation?

4.3. Return measurement issues.

4.4. Returns in which currency?

4.5. Risk-adjusted returns.

4.6. Biased returns.

The terminology used to describe expected returns can be confusing. For most assets, expected returns are uncertain ex ante but they are also unknowable ex post. This is in contrast to realized returns, which are knowable ex post but which do not reveal what investors expected [1]. Realized returns are only one of many needed clues for inferring what investors might have expected. Since we cannot directly observe market expectations of returns or rates, we must infer expected returns from empirical data, including market yields or valuation ratios, past returns, and investor surveys, and from models.

An exception is the expected return on a default-free, zero-coupon government bond. The yield on such a bond is the return that investors expect—and in fact receive—if the bond is held to maturity.

When reading about expected returns, it is important to distinguish between time series predictability—variation in the expected return for a given asset class over time—and cross-sectional predictability—variation ...