How suitable is the Fama-French ve-factor model for describing German and Norwegian stock returns?

2016

In this thesis we investigate the suitability of the Fama and French (2015a) ve-factor model

for describing German and Norwegian average monthly stock returns in the period from

1991 to 2015. We do this by estimating factor exposures and risk premiums of test portfolios

formed on three di erent double sorts on rm characteristics. Estimation is done by means of

Fama and MacBeth (1973) regressions. To secure unbiased standard errors, we apply a GMM

approach when estimating the risk premiums. We evaluate both absolute and relative model

performance on the basis of the test statistic developed by Gibbons, Ross, and Shanken (1989).

The three-factor model serves as benchmark when describing German stock returns and a

two-factor model comprising the market and the size factor is the basis for comparison in the

case of Norway.

With our thesis we make three contributions to existing literature. Firstly, we construct

the Fama-French pro tability and investment factors for the German and Norwegian stock

markets. Although the market, size and value factors are provided by others, we rebuild

them based on our model assumptions to secure internal model consistency. Secondly, by

estimating risk premiums with a GMM approach, we introduce robust standard errors to the

original estimation done by Fama and French. Thirdly, we show that the ve-factor model does

not outperform the more parsimonious benchmark models neither in describing German nor

Norwegian stock returns within our sample period and that this result is una ected by several

changes in underlying assumptions. By this, we extend others’ ndings about the German

and the Norwegian stock markets with the conclusion that the inclusion of pro tability and

investment factors, at least in our setting, does not add value to already existing models.