Content area
Full text
When analyzing fund returns, investors apply various factor models to assess performance and investment styles. Based on the widely used one-factor model (Jensen 1968) and the three-, five- and six-factor Fama–French models (Fama and French 1993, 2015, 2018), researchers estimate alpha as a measure of risk-adjusted performance and factor loadings in assessing investment styles. Moreover, factor models are regularly applied to stock returns to determine abnormal returns in, for example, the context of event studies (MacKinlay 1997). To generate reliable results, it is crucial that the currency of stock or fund returns match the currency of the factor returns.
Researchers can obtain these factor returns in several ways. They can calculate the factors by themselves based on specific datasets. However, many researchers choose a more comfortable way and download factors from databases like Kenneth French’s data library, which provides factor returns for the United States and for various developed markets, including Europe, Japan, and Asian Pacific (Fama and French 2012). Since August 2019, the library has also included emerging market factors. Importantly, the factor returns are given in US dollars (USD), irrespective of the geographical region covered. Consequently, the currency of the factor returns needs to be adjusted when working with non-US samples from a non-USD perspective (e.g., European funds from a EUR perspective). To be more precise, the downloaded factors need to be converted into the respective non-USD currency; otherwise, exchange rate fluctuations affect factor returns and potentially distort the results when estimating factor models.
To gain initial insights into such distortions, we now look at the performance of a passive European market index from a EUR perspective. Specifically, we analyze the (excess) returns of the MSCI Europe IMI in EUR when applying the one-factor model. As the market factor, we use the returns of the European market in USD from Kenneth French’s data library.1 We apply two different specifications of this explanatory variable. First, we download the returns of the European market and convert them into EUR. Second, we use the European market factor in USD without applying a currency conversion. Exhibit 1 illustrates the findings of both regressions.
[Figure omitted: see PDF.]
Panel A shows the results using the converted European market factor in EUR. We find an intercept/alpha at near...