Skip to content
All library documents

Why Market Efficiency Tests Depend on the Expected Return Model

Article Quant Q&A · Author: Nga Tran
Read the original

Summary

The document explains the joint-test problem in empirical tests of market efficiency. A researcher does not observe market efficiency directly; instead, the test evaluates whether returns contain statistically significant abnormal performance relative to an expected-return model. The example given is CAPM, which supplies a benchmark for expected returns and therefore determines how alpha is measured.

A significant alpha can be read as evidence against the combination of market efficiency and the chosen return model. It does not, by itself, identify which part is wrong: the market may be inefficient, or the model may inadequately describe expected returns. The document offers a brief conceptual explanation rather than a testing procedure, dataset, or empirical result. Its main limitation is that the model dependence remains unresolved; assessing efficiency separately would require additional assumptions or a more credible benchmark for expected returns.

Key ideas

  • Market efficiency tests assess returns against a model of expected returns.
  • Alpha represents performance beyond the benchmark specified by that model.
  • A significant alpha challenges both the efficiency assumption and the expected-return model.
  • The test alone cannot determine which of those assumptions caused the result.

Tags

StatisticsFactor investingBacktesting
Full text
# Joint tests of market efficiency - Is it possible to test market efficiency with either one?


# Joint tests of market efficiency - Is it possible to test market efficiency with either one?












Tests of market efficiency are the joint tests of (1)the market is efficient and (2) expected return model.

Please help me (a) explain this and (b) is it possible to test market efficiency with either of these ones?

## Answer by user62342 (score 1)

https://quant.stackexchange.com/a/71058

When markets are said to be efficient then the expectation is that there is no excess returns (alpha).

The expected return is basically a model of forecasting returns such as CAPM.

So basically you want to test if alpha (excess returns) is significant.

It's a joint test in the sense that you are not only testing alpha but are also testing the validity of CAPM to determine returns in the first instance.

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.