A simple test of the Fama and French model using daily data: Australian evidence

Abstract

The current study contributes to the empirical literature aimed at testing the Fama and French three-factor model, using daily Australian data. In general, the evidence found is quite favourable to the model based on formal asset pricing tests. However, when the estimated risk premia are taken into account, the support for the Fama-French model is less persuasive. In particular, a negative size premium is uncovered consistent with a wave of recent findings questioning its continued existence over recent years.

Notes

¹Black (1993) was particularly critical of the Fama and French approach on the grounds of ‘data mining’. At the most general level, Leamer (1983) ignited modern research interest in the issue of data snooping. More recently, Lo and MacKinlay (1990) have forced finance researchers to take this issue more seriously than they have in the past.

²This represents the small-sample adjusted version following MacKinlay and Richardson (1991). GMM is (asymptotically) distributed as a chi-square statistic with N degrees of freedom.

³The estimation technique employs heteroscedasticity and autocorrelation consistent covariance matrices and following Ferson and Foerster (1994) uses an iterated procedure.

⁴A good discussion and description of these indexes can be found on the Frank Russell Company website at: http://www.russell.com/indexes/australia/definition/icadefn.htm and the discussion which follows in the text borrows from the information provided there.

⁵See, for example, Gompers and Metrick (1998); Dimson and Marsh (1999); Gustafson and Miller (1999) and Horowitz et al. (2000).

⁶In unreported results it is also found that the correlation between the proxy SMB and HML factors is very close to zero. This is outcome is consistent with the mimicking portfolios of Fama and French (1993).

⁷This resources versus industrial sector split is motivated by the well documented apparent overpricing of mining and resources stocks when applying the CAPM and asset pricing models in Australia (Ball and Brown, 1980; Ball, 1986; Dolan, 1997 and Ord, 1998).