My global fund portfolio is not yours: the effect of home bias on European- and US-managed convertible bond fund exposures

Abstract

This paper shows that global convertible bond funds (CBFs) and their resulting equity-bond exposures are regionally biased. Global bond fund managers display home bias, resulting in CBFs that are not only tilted towards the home market but also reflect the different bond-equity exposures of European and US convertibles. More specifically we find that global funds managed by a European asset management firm are more bond-like than global funds managed by a US-based asset manager. Hence, investors have to account for the asset management company's origin to avoid that the performance of the fund and its correlation with other assets is not in line with investor's ex ante expectations about globally managed portfolios. Our results also indicate that for investors of European-based CBFs this home bias has resulted in an ex post opportunity cost up to 1.38% per year, depending on the sample period.

Keywords:

• convertible bonds
• home asset bias
• mutual funds
• portfolio allocation

JEL Codes:

• G11
• G15

Acknowledgement

We are very grateful for the collaboration with Morningstar on this mutual fund project. We thank the Hercules Foundation (Project No. AKUL/11/02) for the financing provided to purchase Datastream and acknowledge the financial support provided by the National Bank of Belgium (NBB) special fund that helped financing this project. Also, we thank UBS for giving us permission to using their index data. We are grateful to Jan Dierick for excellent research assistance. We take full responsibility for all remaining errors.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Geert Van Campenhout http://orcid.org/0000-0001-5821-6750

Rosanne Vanpée http://orcid.org/0000-0002-1673-2535

Notes

1. The latter must be included in in annual reports, while portfolio holding must be included in semiannual and annual reports.

2. Our sample differs greatly from Ammann, Kind, and Seiz (2010) who use a total sample of 114 US CBFs. The reason for this is that they include convertible arbitrage hedge funds and their definition of a CBF is less strict: they include funds that have a proportion of zero to 98% of their total assets invested in convertible bonds, with an average of 38%. The minimum proportion of convertible bonds in our funds is 50%, with an average of 77%.

3. Other countries include Australia, New Zealand, Latin America, Africa and the Middle East.

4. Ammann, Kind, and Seiz (2010) include a fourth factor, notably the return on a convertible arbitrage index, to capture variations in returns arising from a convertible arbitrage strategy. As we have excluded the convertible arbitrage hedge funds from our sample, there is no need to consider this variable. In addition, to our knowledge, no region-specific convertible arbitrage indices are available.

5. So superscripts refer to the domicile of the asset management firm, while subscripts refer to the region of equity, bond and option market exposures.

6. Alternative estimation methods notably pooled OLS, fixed effects with robust standard errors, a model with clustered regional effects, lead to slightly lower standard errors of the coefficient estimates, but the majority of the significance levels remain unchanged. These alternative estimation results are available upon request.

7. Note that the documented European home bias should be considered as an average home bias for European-based funds. As the home bias of specific EU countries might differ, an analysis of such differences might be a valuable avenue for further research. We would like to thank one of the anonymous referees for raising this point.

8. Levy (2013) assumes that all markets have the same mean return, standard deviation and the same pair-wise correlation. Looking at the ex post characteristics of the convertible bond markets, these assumptions seem too strict in our case.

9. The correlation between the returns of the US convertibles and the European convertibles drops from 70% to 61%, the correlation between US and Asian convertible bond returns drops from 62% to 54% and the correlation between European and Asian convertible bond returns decreases from 61% to 48%.

10. For funds managed in the USA, portfolios with a high (low) home bias have an average Sharpe ratio of 0.19 (0.16). These Sharpe ratios are indicative only.