Skip to content
All library documents

Choosing Fama–French Three-Factor Data for European Returns

Article Quant Q&A · Author: Saad Al
Read the original

Summary

The discussion addresses which daily Fama–French three-factor series to use when explaining returns for European-country stocks, whether one factor file can be applied across multiple countries, and which risk-free rate to use when constructing a market excess return. Its main guidance is to choose global or European factors according to the research question: global factors address global exposure, while European factors focus on the regional market. The currency denomination of the stock price alone does not determine which factor set is appropriate.

For a study using the same selected regional or global dataset, the answer says the factor time series can be shared across the country samples while the individual return series differ. If constructing the market premium independently, it suggests matching the risk-free rate to the factor region and purpose. The response is qualified: it says the choice depends on the research objective and is uncertain about how the published European factors define their risk-free rate. It does not offer empirical comparisons or detailed data-construction rules.

Key ideas

  • Choose global or European factors based on the market exposure the study aims to explain.
  • Daily stock data can be analyzed with factors at another frequency if returns are converted consistently.
  • A shared factor dataset can be used across country samples, with each country's returns remaining distinct.
  • When constructing a market excess return, align the risk-free rate with the intended regional analysis.
  • The response flags uncertainty about the published European series' risk-free rate convention.

Tags

EquitiesFactor investingStatisticsMulti-asset
Full text
# Estimating French and Fama 3 - factors for global markets


# Estimating French and Fama 3 - factors for global markets












I am working on this project where I am estimating FF three factors for some European countries. So I collected daily prices in US dollars for these countries since I will be using FF three factors from Kenneth website. I made sure that countries in my sample are included in listed of countries in Fama and French's list of countries. Now I have some questions:

- If my data is daily, which Fama and French three factors should I use from Kenneth's website: Fama/French Global 3 Factors [Daily], Fama/French Global ex US 3 Factors [Daily], or Fama/French European 3 Factors [Daily]? and what is the difference between them. I am assuming Fama/French European 3 Factors [Daily] since it only includes European countries but keep in mind that although companies are European but I collected prices in U.S. dollar, not in Euros.

- Do I use the same factors file for all the European countries? In other words, If my sample were Spain and Finland, would I use the same Mkt-RF SMB HML and RF for both countries and the only difference would be on the daily return?

- If I am not using FF market risk premium, Mkt-RF, and would like to construct my own, do I use U.S. risk-free rate or the European country's risk-free rate? It makes no sense to use U.S. risk free rate but if not mistaking, FF use one month T-bill rate.

Thank you in advance for your help

## Answer by phdstudent (score 1, accepted)

https://quant.stackexchange.com/a/41724

It really depends on your goal.

- Regarding your first question. If you want to measure how well global factors of value, size explain the cross-section of returns you should use the global factors. If you want to explain the cross-section of returns using only European factors then use the European dataset. It really depends on what you are trying to achieve. Also even if your data is daily you can always convert it to monthly or annual and do the analysis with another time scale. Again depends on your goal.

- That is correct if you want to use the FF global or european factors. Only the daily return changes. The factor time-series are the same.

- Again it depends on the application. If you want to estimate Mkt-Rf on your own and you are using european factors, then use the european Rf. In any case shouldn't matter much for the results. FF I think are not very clear which Rf they use on the European factors. They might have a paper where they explain it, but I am not sure.

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.