Skip to content
All library documents

Currency Choice and Regional Factors in Fama-French Regressions

Article Quant Q&A · Author: vonjd
Read the original

Summary

The document considers how to run Fama-French regressions for euro-denominated investments when the available factors are in US dollars and the investments may also carry international currency exposure. It contrasts converting returns to dollars, using euro returns directly, and including exchange-rate changes as an additional factor. The answer says the appropriate choice depends on assumptions about purchasing power parity and whether the assets can hedge currency risk. If those assumptions are unsuitable, the researcher should use empirical tests that account for exchange-rate risk.

A separate issue is factor geography: US portfolio factors may not describe European investments well. The responses recommend constructing factors for Europe, individual countries, or sectors, and suggest world factors when analyzing internationally invested firms. The document offers conceptual guidance and points to relevant research, but does not provide regression specifications, empirical comparisons, or a universal best practice. Results depend on the assets, currency treatment, and factor portfolio chosen.

Key ideas

  • Currency conversion choices depend on assumptions about purchasing power parity and the assets’ ability to hedge exchange risk.
  • When those assumptions are not appropriate, the analysis should empirically account for currency risk.
  • US-constructed Fama-French factors may not be suitable for explaining European investment returns.
  • Researchers can build regional, country, sector, or world factors to better match the investments being studied.

Tags

EquitiesForexFactor investingStatistics
Full text
# Best practice for international Fama-French analysis


# Best practice for international Fama-French analysis












First I have to admit that I have never been really good at thinking about the implications of investments in different currencies. I don't know why but it makes my head spin, this is why I am no FX guy... anyway: Bear with me if this is a silly question!

I have the following situation: I want to thoroughly analyse different equity investments in the Eurozone. Fama-French regression is of course part of that. Now all Fama-French factors (there is a whole bunch of them now) are in USD. Another complicating factor is that - while all investments are in Euros - some are heavily investing internationally themselves and are therefore exposed to currency risk too.

My question How should I conduct this analysis? Should I

- leave the investments as is and regress the changes of the Euro denominated investments against the respective USD factors,

- convert the investments into USD and regress against the respective USD factors or even

- leave the investments as is and regress the changes of the Euro denominated investments against the respective USD factors + add another currency factor (EUR/USD exchange rate changes)?

What is best practice here? What are the implications? How to interpret the results? Sorry again if this is a silly question.

## Answer by phdstudent (score 2, accepted)

https://quant.stackexchange.com/a/25146

It depends on the assumptions you are willing to make.

If you either assume:

- (i) Complete purchasing power parity (relative prices of goods are the same everywhere and an exchange rate is just the ratio of the nominal prices of any good in two countries) or

- (ii) the assets you consider cannot be used to hedge exchange risk

then you are fine by using your the solution outlined in your method 2.

- If you don' want to assume neither of those, then you need empirical tests that allow for exchange rate risk. Look into Fama and Farber (1979) or Zhang (2006).

One caveat though. The Fama-French factors (built on US porfolios) are not the correct factors to analyze Europe returns. Either you build the same factors for Europe, or if the firms you are considering are international you should build world factors (look into Asness, Moskowitz and Pendersen 2013).

## Answer by Tim  (score 2)

https://quant.stackexchange.com/a/25175

This is a much discussesd topic in the litrature, I recommend you to read FF2012: "Size, value, and momentum in international stock returns". Best practice is to create Europe, country or even sector specific factors.

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.