Intermediaries’ Incentives across Share Classes in the Same Fund

Abstract

We provide supporting evidence that intermediaries’ incentives vary across retail share classes in the same fund. We find that when a fund has multiple share classes with different distribution fees, flow is less sensitive to poor performance for share classes with higher distribution fees. These results are more pronounced for funds when intermediaries are more inclined to favor one share class over another—specifically, for funds serving only retail investors, having a large dispersion in distribution fees across share classes, or having a share class that charges the maximum allowed distribution fee. Our results hold for funds with small spread in investors’ performance sensitivities and disappear in a placebo test. These findings cannot be explained by differences in share-class load fees or investor clientele.

Keywords:

• 12b-1 fee
• conflicts of interest
• flow
• intermediaries
• mutual funds
• performance
• share classes

PL Credits:

• 2.0

The views expressed herein are those of the authors and do not necessarily reflect the position of the Federal Reserve Bank of Richmond or the Federal Reserve System. All errors are our own.

Acknowledgments

We thank the executive editor William Goetzmann, co-editor Daniel Giamouridis, managing editor Luis Garcia-Feijoo, and two anonymous referees for their helpful comments. We thank Anna Agapova, Kevin Ahlgrim, Vadim Balashov, Eliana Balla, Yashar Barardehi, Alexander Barinov, Onur Bayar, Hank Bessembinder, Mehdi Beyhaghi, Karan Bhanot, Natasha Burns, Scott Cederburg, Ji-Woong Chung, Tony Ciochetti, Douglas Cumming, Mark DeSantis, Mike Dong, Kristopher Gerardi, Ralph Giraud, Nikolay Gospodinov, Olivia Huseman, Mark Jensen, George Jiang, Yawen Jiao, Domingo Joaquin, Palani-Rajan Kadapakkam, Gary Koppenhaver, Shahdad Naghshpou, David Pedersen, Dave Porter, Nan Qin, Michael Reher, Mohammad Robbani, Zhi Li, Atanas Mihov, Efdal Ulas Misirli, Indrajit Mitra, Eli Sherrill, Neal Stoughton, Thomas Thomson, Paula Tkac, Hai Tran, Tim Trombley, Chun Kai Tseng, Abhishek Varma, Marc Weidenmier, John Wald, Larry Wall, Zijun Wang, Bin Wei, and Clas Wihlborg for valuable feedback. We thank the seminar participants at Chapman University, University of California-Riverside, the Federal Reserve Bank of Richmond, the Federal Reserve Bank of Atlanta, the participants at the 14th Annual California Corporate Finance Conference-Loyola Marymount University, the 2019 Financial Management Association Meeting-New Orleans, and the 3rd Annual Boca Corporate Finance and Governance Conference.

Disclosure

No potential conflict of interest was reported by the author(s).

Notes

1 See https://www.sec.gov/news/press-release/2019-28 and https://www.sec.gov/news/press-release/2019-200).

2 A one-time kickback has different terminologies across countries, such as trailer fee, finders’ fee, referral fee, or acquisition commission. For example, in Canada the trailer fees are paid to the fund dealer (Cumming, Johan, and Zhang 2019).

3 See https://www.cfainstitute.org/-/media/documents/article/position-paper/sales-inducements-in-asia-pacific.ashx.

4 See, for example, Christoffersen, Evans, and Musto (2013); Bergstresser, Chalmers, and Tufano (2009); Wahal and Wang (2011); Del Guercio and Reuter (2014); Chalmers and Reuter (2015); Linnainmaa, Melzer, and Previtero (2018); Egan, Matvos, and Seru (2019); Egan (2019).

5 See, for example, O’Neal (1999); Nenninger and Rakowski (2014); Navone and Pagani (2015); Handy, Ricketts, and Smythe (2020); Gómez, Prado, and Zambrana (2022).

6 We provide an overview of 12b-1 fees and share classes in Appendix B.

7 See, for example, Bergstresser, Chalmers, and Tufano (2009); Christoffersen, Evans, and Musto (2013); Del Guercio and Reuter (2014); Chalmers and Reuter (2015); Navone and Pagani (2015); Lemeunier (2017); Linnainmaa, Melzer, and Previtero (2018); Egan (2019); Egan, Matvos, and Seru (2019); Gómez, Prado, and Zambrana (2022).

8 See, for example, Gruber (1996); Wermers (2000); Chevalier and Ellison (1997); Sirri and Tufano (1998); Huang, Wei, and Yan (2007); Wahal and Wang (2011); Nenninger and Rakowski (2014); Christoffersen and Xu (2017).

9 See, for example, Dellva and Olson (1998); Sirri and Tufano (1998); O’Neal (1999); Lesseig, Long, and Smythe (2002); Barber, Odean, and Zheng (2005); Jones, Lesseig, and Smythe (2005); Bergstresser, Chalmers, and Tufano (2009); Christoffersen, Evans, and Musto (2013); Chalmers and Reuter (2015); Oh, Parwada, and Tan (2017); Roussanov, Ruan, and Wei (2021).

10 To comply with the new United Kingdom rule, asset managers have created new share classes called “clean” share classes, that is, share classes clean of retrocessions fees. Although there is no ban on 12b-1 fees in the United States, to comply with the 2017 Department of Labor fiduciary rule that made the group of advisors subject to fiduciary standards wider, some investment managers in the United States have created new T shares and “clean shares.” The SEC’s 2019 Regulation Best Interest also intends to address the issues of potential conflicts but relying more on disclosure rules. Section IA.3.2. in the Online Supplemental Material Appendix provides more details on Dodd Frank and Regulation Best Interest.

11 We use variable open_to_inv to identify whether a fund is open to investors.

12 We use variable retail_fund to identify whether a share class is a retail share class.

13 The Rule 18f-3 was adopted in 1995. There is no fund that satisfies our dataset construction criteria before 1999.

14 See, for example, Rosenberg and Czepiel (1984); Sterne (2002); Gallo (2014); Armstrong and Kotler (2020); Belch and Belch (2021).

15 The section “Robustness Tests” shows that our results are robust to different estimation procedures and measures of performance (see Sections IA.2 and IA.3 in the Online Supplemental Material Appendix).

16 Subscript i,q is omitted for brevity.

17 The results hold if we use the 12b-1 fee itself instead of the indicator variable High-IntermFee Dummy. For ease of interpretation, we report results using the indicator variable.

18 Note that the minimum 12b-1 fee in our sample is nonzero, 0.03 (Panel A). This fact helps with the concern that our baseline results are driven by investor heterogeneity across different share classes. Investors in share classes that do not levy distribution fees are likely more sophisticated than investors in classes that charge 12b-1 fees. However, investors in share classes that charge a 1% fee and those in share classes that charge a 0.75% fee are more likely to have similar characteristics such as sophistication, behavioral biases, or other constraints.

19 The mean gross expense ratio (12b-1 fee + net expense ratio) of our sample funds is higher than other studies because all funds in our sample charge a nonzero 12b-1 fee.

20 Results are robust if we drop style fixed effects and include Category Flow instead.

21 These results are not at odds with the results reported in Huang, Wei, and Yan (2007). They provide fund-level evidence that flow is more sensitive to performance when a fund experiences poor performance and charges a high distribution fee. We show that within each fund, where all share classes have the same investment portfolio, when fund performance is poor, flow is more sensitive for the share classes with a lower 12b-1 fee, all else equal. Huang, Wei, and Yan (2007) also show in their Table VIII on page 1305 that the coefficient on Low × C Class Dummy is insignificant, indicating that the flow between A and C shares is indistinguishable when the fund performs poorly. Notably, both A and C shares can be low 12b-1 fee share classes relative to the rest of the shares that a fund offers.

22 The performance rank is calculated based on the gross risk-adjusted return of a share class in the paper, since all share classes in the same fund have almost identical gross returns. Our results hold when performance rank is calculated based on the net risk-adjusted return of a share class. For example, the parameter estimate for Low Carhart Alpha × High-IntermFee Dummy (β1) in column (4) of Table 3 is −4.49% and significant at the 1% level when performance rank is calculated based on the net Carhart alpha. Using a sample of only institutional investors, we find that β₁ is positive and insignificant. Institutional investors are sophisticated and do not rely as much (or at all) on financial intermediaries; thus, in this case we do not expect that 12b-1 fees affect the flow sensitivity to bad performance. In this sense, this test could serve as a placebo test to our main results.

23 Huang, Wei, and Yan (2007) show that fund flows are more (less) responsive to the medium (high) fund performance range. Oh, Parwada, and Tan (2017) document that these results are restricted to before 2000 and that in the post-2000 period, flow–performance sensitivity disappears from the medium-performance range and is largely concentrated in the high-performance range.

24 A survey conducted by the Investment Company Institute also shows that intermediaries collect the majority of the 12b-1 fees. Considering that our test sample includes only funds charge positive 12b-1 fees, these funds are most likely broker-sold funds (https://www.ici.org/pdf/fm-v14n2.pdf).

25 Bergstresser, Chalmers, and Tufano (2009) and Christoffersen, Evans, and Musto (2013) obtain the fund-level distribution channel information from the Financial Research Corporation data and the N-SAR filing data, respectively.

26 This approach is in line with the prohibition under the 1993 NASD rules of calling funds no-load funds if they charge front-end, rear-end, or 12b-1 fees in excess of 0.25%.

27 Distribution costs include fees paid to brokers who sell fund shares as well as costs such as advertising, printing, and mailing prospectus to new investors. Service costs are expenses related to serving current fund shareholders, such as operating and staffing information hotlines. The SEC does not limit the size of 12b-1 fees, but the rules of the Financial Industry Regulatory Authority, Inc. (FINRA) limit the 12b-1 fees to 1% annually, with distribution/marketing fees and service fees limited to 0.75% and 0.25%, respectively.

28 Source: https://www.ici.org/pdf/fm-v14n2.pdf (last accessed January 2022).

29 Detailed descriptions of estimating alphas can be found in Section “Methodology and Variables” and Appendix A.

30 Specifically, Section 913(g) of the Dodd-Frank Act articulated a uniform fiduciary standard which essentially gives an “Authority to establish a fiduciary duty for brokers and dealers” to the SEC.

31 Retrieved from https://crsreports.congress.gov/product/details?prodcode=R46115.

Additional information

Notes on contributors

Ivalina Kalcheva

Ivalina Kalcheva is an Associate Professor from the Finance Department, Carlos Alvarez College of Business, University of Texas at San Antonio, Texas.

Ping McLemore

Ping McLemore is a Senior Financial Economist from the Federal Reserve Bank of Richmond, Baltimore, Maryland.