A return-based approach to identify home bias of European equity funds

ABSTRACT

This paper introduces a return-based approach to studying a possible home bias of European equity funds by estimating their exposures to their domestic markets. We first confirm the robustness of our approach using simulated portfolios with different proportions of domestic and foreign stocks. The empirical analysis examines equity funds domiciled in 15 European countries that invest in European stocks. We examine individual funds as well as portfolios comprising funds that are all domiciled in a particular country. Our findings reveal that the portfolios of four domiciles show a significant home bias. Moreover, we observe that in seven domiciles more than a quarter of the individual funds are home-biased. These results are robust when controlling for fund-specific benchmarks or for the average country exposures of all funds in our final sample. Finally, a home bias of individual funds is not related to superior performance, but actually results in higher investment risk consistent with underdiversification.

KEYWORDS:

• home bias
• equity mutual funds
• market conditions
• fund performance

JEL CLASSIFICATIONS:

• G11
• G12
• G14
• G23

Acknowledgements

We are grateful for helpful comments and suggestions from John C. Adams, Rainer Baule, Christine Crozier, Oliver Entrop, Hannah Lea Hühn, Olaf Korn, Andrew Lynch, Martin Rohleder, Marco Wilkens, Gulnara R. Zaynutdinova, and the participants of the Southwestern Finance Association Annual Conference 2016 in Oklahoma City, the Workshop of the German Operations Research Society (GOR AG FIFI) 2016 in Augsburg, the Workshop Finance 2016 in Göttingen, the International Ph.D. Seminar 2016 in Nürnberg and the Southern Finance Association Annual Meetings 2017 in Key West. We are responsible for any remaining errors. Financial support from the Dr. Michael Munkert-Stiftung is gratefully acknowledged.

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 For an excellent overview and potential enhancements, see Mishra (2015).

2 For details see e.g., Securities and Exchange Commission (2004).

3 Note that our peer group-adjusted returns are different from the active peer benchmarks of Hunter et al. (2014). They account for peer group effects in estimated fund alphas by including an additional explaining factor to capture peer group effects. However, our focus is to analyze whether the fund’s exposure deviates from that of the peer group which can be achieved by using fund returns in excess of the peer group returns.

4 The stock classification into sectors or industry groups is based on the ‘Global Industry Classification Standard’ (GICS) industry classification of stocks from MSCI and S&P.

5 The country market indices represent the equity markets of Austria, Belgium, Denmark, Finland, France, Germany, Ireland, Italy, the Netherlands, Norway, Portugal, Spain, Sweden, Switzerland and the United Kingdom.

6 Before the orthogonalization, the VIFs ranged from 1.56 (MOM) to 456.50 (MKT). Most notably, the VIFs of France, Germany, the Netherlands and the United Kingdom exceed 10.

7 We chose the STOXX Europe 600 rather than the MSCI AC Europe IMI as the stock universe because we assume that fund managers tend to invest in stocks with a larger market value. From this point of view, it would be preferable to select an index that contains larger stocks, e.g., STOXX Europe 200. However, because we simulate home subportfolios and foreign subportfolios for all domiciles, we need a minimum number of available stocks from each country. Therefore, the STOXX Europe 600 can be seen as a compromise between average stock size and the number of available stocks in each country.

8 .

9 We chose these 15 countries with respect to the 15 domiciles investigated in our empirical analysis of equity funds.

10 The Pearson correlation of the STOXX Europe 600 with the MSCI AC Europe IMI is 0.997. However, the estimated regression of the STOXX Europe 600 on Equation (2) yields negative coefficients for the country-specific factors of France, Germany and the United Kingdom.

11 These domiciles are Andorra, Greece, Guernsey, Jersey, Latvia, Malta, Poland and Slovenia.

12 The non-European benchmarks are ‘Euronext AEX All Share TR EUR’, ‘MSCI EMU NR EUR’, ‘MSCI Germany Small Cap NR EUR’ and ‘MSCI Spain’.

13 Specifically, we interpret funds with ‘Europe’, ‘Europa’, ‘Europ’, ‘Eurp’ and ‘Eurooppa’ in fund names as European-focused funds. These funds are included in the final sample.

14 Moreover, we calculate equal-weighted portfolios and redo all analysis. The results tend to be stronger in supporting the presence of home bias in European equity funds which indicates that smaller funds tend to be more home-biased. These additional results are available upon request. We report results here for value-weighted portfolios, to be conservative with respect to home bias.

15 The benchmark indices are ‘MSCI Europe Growth’ (39 funds), ‘MSCI Europe High Div Yld’ (34), ‘MSCI Europe’ (499), ‘MSCI Europe Small Cap’ (57), ‘MSCI Europe Value’ (128) and ‘STOXX Europe Mid 200’ (28). We use Morningstar fund-specific benchmarks instead of self-declared fund benchmarks, because the latter could lead to distorted findings if funds choose inappropriate benchmarks, as discussed, e.g., in Sensoy (2009), Cremers and Petajisto (2009) or Cremers et al. (2016).

16 Respective t-statistics for the alpha and the Carhart (1997) factors are unreported for the sake of brevity. These values are available upon request.

17 Results for an equal-weighted peer group are excluded for the sake of brevity and are available upon request.

18 For a detailed description see https://fred.stlouisfed.org/series/4BIGEURORECDM.

19 The time series of the CLI can be downloaded from the OECD homepage http://www.oecd.org/std/leading-indicators/oecdcompositeleadingindicatorsreferenceturningpointsandcomponentseries.htm.

20 In additional robustness tests, we analyze different market states or different phases of investor sentiment. For market states, we follow the methodology from Cooper, Gutierrez, and Hameed (2004) and divide market states in up and down states. We observe a slight tendency towards a more pronounced bias (positive and negative) during insecure times (down market states). With regard to investor sentiment, we focus on times of optimistic and times of pessimistic sentiment. We find a slight tendency towards more pronounced home bias during insecure times (pessimistic sentiment). Detailed results for these tests are available upon request.

21 In additional robustness checks, we analyze fund performance and risk by using the one-factor model and the Fama and French (2015) five-factor model. We thank Kenneth French for providing these factors on his website (http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html). We cannot observe a statistically higher or lower average alpha for decile D10 than for one of the other nine deciles with respect to t-statistics. However, the pairwise differences between the average alpha of decile D10 and those of the other nine deciles are mostly significant in the non-parametric tests. The detailed results are available upon request.

22 In addition, we calculate funds’ idiosyncratic risk by applying a one-factor and the Fama and French (2015) five-factor model. In doing so, our main findings, that more highly concentrated funds (D10) yield higher risk, remain unchanged. The differences between the average idiosyncratic risk of decile D10 and that of one of the other nine deciles are mostly statistically significant. These results are available upon request.

Additional information

Notes on contributors

Moritz Maier

Moritz Maier is a Research Assistant and PhD Student at the Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany. He studied business administration at the Baden-Wuerttemberg Cooperative State University Ravensburg, Germany and at the Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany. His research focuses on home bias, performance analysis and empirical finance.

Hendrik Scholz

Hendrik Scholz is a Professor of finance and banking at the Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany. He received his PhD in finance from the University of Göttingen and his habilitation from the Catholic University of Eichstätt-Ingolstadt, Germany. His research focuses on performance analysis, asset management, structured financial products and bank management. He has authored numerous articles in finance journals. He is a member of the Southern Finance Association, the German Finance Association and the German Academic Association for Business Research.