In the capital asset pricing model (CAPM), estimating beta consistently
is important to obtain a consistent estimate of the price of
risk. However, it is often found that the estimate of beta is sensitive
to the choice of portfolios used in the estimation. This paper provides
a new test to evaluate whether the choice of portfolios in typical asset
price regressions is valid, in the sense that the portfolios satisfy
two conditions: (i) the way the portfolios are formed are exogenous;
and ...
In the capital asset pricing model (CAPM), estimating beta consistently
is important to obtain a consistent estimate of the price of
risk. However, it is often found that the estimate of beta is sensitive
to the choice of portfolios used in the estimation. This paper provides
a new test to evaluate whether the choice of portfolios in typical asset
price regressions is valid, in the sense that the portfolios satisfy
two conditions: (i) the way the portfolios are formed are exogenous;
and (ii) the choice of the group of assets to include in the portfolios
provides enough information to identify the parameters of interest.
Thus, checking the validity of the portfolio choice is an important
pre-requisite to ensure consistent estimates of the parameters of the
model.
We illustrate the performance of the test in small samples via
Monte Carlo simulations.The proposed test is also applicable to group
and pseudo panel data models.
+