SMSG Opinion on ESMA's Report on performance and cost of retail investment products

Securities and Markets Stakeholder Group ESMA SMSG • CS 60747 – 103 rue de Grenelle • 75345 Paris Cedex 07 • Tel. +33 (0) 1 58 36 43 21 • www.esma.europa.eu/smsg Executive summary In January 2019 ESMA published a report on “Performance and costs of retail investment products in the EU”. The report follows a mandate from the European Commission to the ESAs and provides a compre￾hensive overview of the EU markets and a country-by-country analysis for UCITS from 2008 to 2017, highlighting insightful findings about the returns generated by mutual funds in different asset classes, the costs borne by retail investors, and the active vs. passive management debate. The SMSG believes that the report is a very valuable tool for retail investors as well as for the industry. Moreover, clear and comprehensive information on performance and costs of products offered across the EU can be beneficial to regulators and policymakers too. The SMSG also appreciates the discussion in the report about methodological issues and data limitations. The SMSG observes that it would be useful to provide retail investors with a common perspective of the cost and performance of all substitute retail investment products. The SMSG therefore advises ESMA to consider whether the Joint Committee of the ESAs could take this up. Even if one Joint Committee report would prove hard to achieve in the near future, the SMSG advises ESMA to develop, together with the other ESAs, a common format for the three ESA reports on this matter, in order to facilitate comparability of the results and to identify differences in data availability across product categories, in line with the European Commission mandate. In this opinion the SMSG further observes that important data - such as a part of the distribution costs, transaction costs and performance fees – are not available, as explicitly acknowledged by ESMA. This lack of data may limit the significance of the evidence provided in the report. Additionally, country-by￾country comparability appears to be partial, due to data limitations. The SMSG observes that it is not yet possible to estimate how much an investor pays to invest via mutual funds both on an all-inclusive basis and to the different professionals involved (asset managers, brokers, salespersons). This opinion also dis￾cusses technical matters that might affect certain messages delivered by the report. The SMSG believes that the ESMA report provides very useful data, information and insights. The SMSG advises ESMA to consider the revisions suggested in this opinion, especially since the publication of this report appears to be a long- term project, the aim of which is to provide on a yearly basis a reliable sta￾tistical reference to understand performance and costs of mutual funds in the EU. I. Background Date: 3 October 2019 ESMA22-106-1990 SMSG OPINION SMSG opinion on the ESMA Report on performance and cost of retail invest￾ment products

2

1. The EC request to the ESAs. On October 13, 2017 the European Commission (henceforth “EC”) invited, with a formal request, the European Supervisory Authorities (“ESAs”) to issue recurrent reports on the cost and past performance of the main categories of retail investment, insurance and pension products (“EC Request”).
2. The rationale of the EC Request to the ESAs. Within the actions planned by the EC in the Capital Mar￾kets Union project, the EC Request is meant to contribute to the objective of fostering the participation of retail investors in capital markets through the assessment of net returns and impact of fees and charges. In our understanding the EC Request to the ESAs has three rationales. First, it is a way to make easily accessible to retail investors (and to the general public) key information on performance and cost of investment products in a compact and aggregated form. This is made possible thanks to the disclo￾sures and reporting requirement imposed by different legislative measures in the European Union (e.g. UCITS, MiFID/MiFIR, IDD, IORPII and PRIIPs). Second, such recurrent reports are intended to sup￾port fair competition among industry players and drive pricing discipline in these markets. Third, they may also form a basis for public authorities to identify market segments and regions where investors are in a sub-optimal situation and identify those issues which call for careful analysis.
3. The principles of the reporting stated in the EC Request. The reporting of cost and past performance of different categories of products, subject to the availability of data, should be based on the following prin￾ciples: Member State by Member State analysis; comparability of indicators; appropriate level of granu￾larity with categories comprising products with broadly homogenous characteristics; net returns and im￾pact of costs; all fees impacting the net performance of retail investment products should be reported; the reporting on cost and performance should present the net return for various periods of time (i.e., time horizons); and the impact of inflation should be taken into account. Qualitative and quantitative comparisons of the costs and performance indicators of different categories of products should be pro￾vided both at the levels of each Member State and the EU and should include a comparison between net returns of different categories of products as well as the proportion of impact of costs on the net return.
4. Steps. The EC Request outlines a pragmatic and staggered approach. Due to the expected difficulties in accessing the full set of data, it was envisaged that the first iteration of the report could be based on already available but potentially incomplete databases. The EC Request states that the first iteration should establish a baseline for the future editions of the report and serve also as a basis for the further development of the reporting (e.g., extension of the scope, modification of the methodology). The time￾table for the consecutive iterations will be communicated by the Commission based on the experience with the first iteration. II. ESMA Report on performance and costs
5. The publication of the Report. In January 2019 ESMA published the annual statistical report “Perfor￾mance and costs of retail investment products in the EU” (“ESMA Report” or simply “Report”). The Report provides a comprehensive overview of the EU markets and a country-by-country analysis for UCITS from 2008 to 2017: it shows a differentiated range of performance indicators; it provides details on past performance and costs over a multi-year period; it distinguishes different time horizons, asset classes, retail and institutional investors, actively and passively managed funds, and the impact of in￾flation. For retail alternative investment funds and other structured retail products, evidence is severely limited, and the Report only provides an overview of the EU market.
6. The findings of the Report. The ESMA Report provides a full performance and cost analysis for UCITS. At just under EUR 10tn net asset value (NAV), UCITS represents the largest retail investment fund

3 segment in the Union. UCITS gross annual past performance (i.e. before fund fees) largely follows the performance of the underlying asset classes and can – due to differences in national market structures – vary significantly across Member States. Actively managed equity funds provide a slightly better gross performance than passively managed funds, even though the margin is small. Key findings related to the cost impact are: (i.) costs fluctuate less than gross performance; (ii.) the largest cost impact comes from ongoing costs, while subscription and redemption fees have a significantly lower impact; (iii.) across asset classes, costs are highest for equity and alternative UCITS, followed by mixed, bond and money market UCITS; (iv.) costs are higher for retail compared to institutional investors; (v.) costs are higher for actively managed equity funds compared to passively managed equity funds, which leads to lower performance net of costs for active compared to passive funds (i.e., after cost, passive equity funds outperform active equity funds); (vi.) high heterogeneity in costs across Member States. 7. The limitations of the Report. The ESMA Report also transparently highlights some limitations. In this respect ESMA clarifies that it is currently not possible to estimate exactly the overall costs charged to investors. The ESMA Report emphasizes the unavailability of important cost data elements such as a part of distribution costs, transaction costs, performance fees, a lack of data granularity as well as the heterogeneity of data availability and content across Member States. III. General Comments 8. Objective of this opinion. The EC Request explicitly states that the first edition of the Report should establish a baseline for the future editions. Accordingly the Report also states that ESMA will work on further enhancing the coverage, methods, and assessments in the future editions. The objective of this report is therefore to offer ESMA the SMSG opinion on specific points of the first edition of the Report that ESMA could consider when designing the content of future editions. 9. Content of the Report. The ESMA Report contains a data driven analysis of performance and costs of EU UCITS and includes a statistical annex with a number of tables and graphs as well as a detailed description of data, limitations and statistical methods. The SMSG notes that the publication of this Report is a costly task for ESMA, taking into account data availability issues and technical complexities. 10. Value of the Report. The SMSG believes that the ESMA Report is a very valuable tool for retail investors as well as for the industry. Retail investors are expected to benefit from the availability of detailed in￾formation on costs and past performance of UCITS as they have the possibility to enhance their under￾standing of the key benefits and risks to be considered when investing in mutual funds. Information on past performance and costs of products offered across the EU can also help asset managers (as well as regulators and policy makers) to understand and assess the dynamics of the industry. 11. Comparability with the other ESA reports. As mentioned above, the EC Request was addressed to the three ESAs and each ESA has, therefore, produced a report in response to the mandate received by the EC. One of the rationales of this initiative is to provide investors with information that could be used to take informed decisions when investing. In this respect it would be useful to provide retail investors with a common perspective of the cost and performance of all substitute retail investment products. The SMSG therefore advises ESMA to consider whether the Joint Committee of the ESAs could take this up. The SMSG understands that comparison across products belonging to different sectors of the financial industry (mutual funds, banking, insurance) might be difficult, because of differences in in￾formation available and requested by the regulations currently in place in the different industries. Nev￾ertheless, even if one Joint Committee report would prove hard to achieve in the near future, the SMSG advises ESMA to develop, together with the other ESAs, a common format for the three ESA reports on

4 this matter, in order to facilitate comparability of the results and to identify differences in data availa￾bility across product categories, that might be addressed in future iterations of the reports. IV. Specific Comments 12. Data limitations I (unavailability of specific categories of costs). The SMSG notes that important data – such as a part of the distribution costs, transaction costs and performance fees – are not available. Additional issues in this respect are: a lack of data granularity as well as the heterogeneity of data avail￾ability and content across Member States. The ESMA Report recognizes some of those limitations. For example, on p. 13 the ESMA Report states that “Distribution costs are not included as a specific cost as we are not able to identify such fees. However, it should be noted that distribution costs may be part of the analysis to the extent they are included in ongoing costs and/or the entry charges presented in the KIID. This is a significant limitation in this first analysis given the role of distribution channels and related costs across the EU”. 13. Data limitations II (limited comparability). The SMSG also notes that comparability of country-by￾country results is limited. Data are based on the domicile of the fund, not the domicile of the investor. Therefore, the analysis cannot provide a full picture of the national markets where domestic and foreign funds are competing. For example, the clear majority of funds are sold cross-border in Luxembourg (95%) and Ireland (97%): consequently, funds domiciled in these countries are also present and mar￾keted in other countries. Heterogeneity of cost data availability across Member States is also relevant. There are differences in the definitions of costs used by commercial databases and current EU Regula￾tions and Directives. Such limitations are also acknowledged in the ESMA Report. 14. Data limitations III (unavailability of data on SRPs). The ESMA Report notes that evidence is severely limited for Structured Retail Products (SRPs) and merely provides an overview of the EU market in this first edition. The ESMA Report also notes that in the future it might be possible to make use of infor￾mation published in key information documents (KIDs) for packaged retail and insurance-based in￾vestment products (PRIIPs) to assess costs of SRPs, even though doing so could be very resource-in￾tensive in many cases. In this respect the SMSG observes that the data collection work might be facili￾tated as a result of the FinDatex pan-European industry project. In respect of performance of SRPs, the ESMA Report notes that performance data are not generally available at present for SRPs. In this re￾gard, the SMSG observes that the ESMA Report might usefully review existing national studies, where available. 15. Risk dimension and value for money. Apart from a distinction between asset classes, the analysis in the Report does not distinguish between different risk levels of the UCITS funds. The ESMA Report recog￾nizes such limitations on p. 12: “Asset classes are self-reported in Thomson Reuters Lipper and cover several different strategies within equity, bond, mixed, MMF UCITS as well as within alternative UCITS and UCITS ETFs. This implies different risk levels within an asset class, which we are not able to take into account at this stage”. A meaningful limitation is also the lack of comparison with benchmark (or market) returns. 16. Interpretation of the findings I (retail costs vs institutional costs). The Report shows strikingly clear, and expected, evidence for retail vs. institutional costs: retail investors incur higher costs in both equity funds (2% for retail vs. 1% for institutional, see graph ASR-PC.19) and bond funds (1,5% for retail vs. 0,8% for institutional, see graph ASR-PC.20). In this respect a comparison with non-EU markets and other products might also be useful to provide the reader with reference values to assess the overall cost structure of EU funds.

5 17. Interpretation of the findings II (active management vs. passive management). The evidence provided in the Report does not seem to be clear-cut for the active vs. passive management debate. In terms of gross returns, active equity UCITS perform slightly better than passive equity UCITS over the 1-year and 3-year horizons (16% vs. 15% and 11% vs. 10% respectively, see graph ASR-PC.36). Cost levels are broadly stable over time and consistently higher for actively managed UCITS impacting annual gross return by 1 ppt (1%) compared to 0.6 ppt (0.6%) on average for passively managed UCITS. Conse￾quently, net annual returns are similar at 1-year and 3-year horizons, around 14% and 9% respectively, while being higher for passively managed UCITS compared to actively managed UCITS at longer time horizons: 9.7% against 8.4% at 7-year horizons and 5.8% against 5.5% at 10-year horizons (see graph ASR-PC.36). It is not evident that this difference in returns (0.3%) is economically and statistically significant. In this respect it is important to consider that only 10% of the funds in the Report are pas￾sive, consequently the two sub-samples (active and passive) are not balanced in terms of size. This feature might also affect the significance of the difference in returns between active and passive funds. Lastly, mutual fund returns appear to be dispersed, as shown by the fraction of funds with large differ￾ence between return and benchmark (see graph ASR-PC.22). This evidence is worth to be further in￾vestigated. 18. Technicalities I (survivorship bias and possible overestimation of returns). Statistics computed using databases including only existing funds (i.e., ‘survivors’) – and omitting funds no more existing (i.e., ‘dead’) – may overestimate performances as they do not consider unsuccessful funds that might have displayed low performances over time and might have been shut down1 . The Report could clarify how this issue was dealt with. 19. Technicalities II (compounding effect and misrepresentation of returns). Average returns in the Report are computed on the basis of arithmetic averages. However, arithmetic averages are not representative of actual returns, taking into account – as it should be – the effect of compounding2. A quick example explains this point: let’s just consider an asset the price of which starts at 100, after one year the price is 110 and goes back to 100 at the end of the second year. The return of the first year is 10%, the return of the second year is -9,1% and the simple average of the returns is 0,45% (i.e., (+10%-9.1%)/2) per annum. This positive average return is clearly at odds with the actual return, which is simply zero (as the initial price is 100 and the final price is 100 as well). Geometric averages can be used to avoid this issue. This issue is worth to be further investigated, as the Report seems to be using arithmetic averages. 20. Technicalities III (miscellaneous). The SMSG advises ESMA to include additional disclosure on the way in which passively managed funds are identified in the ESMA Report. Moreover, to provide further evidence about the management style debate (active vs. passive) as well as closet indexing3, additional indicators such as active share, R-squared and tracking error values, alone or – if needed – in combi￾nation, could be considered. The SMSG notes that the meaningfulness of such additional metrics

1 Further information on this point can be found in Elton, E. J., Gruber, M. J., Blake, C. R. (2001), “A first look at the accuracy of the CRSP mutual fund database and a comparison of the CRSP and Morningstar mutual fund databases” The Journal of Finance, 56(6), 2415-2430 and Carhart, M. M., Carpenter, J. N., Lynch, A. W., Musto, D. K. (2002), “Mutual fund survivorship”, The Review of Finan￾cial Studies, 15(5), 1439-1463. 2 Further information on this point can be found in Ilmanen, A. (2011), Expected returns, John Wiley & Sons (particularly Section 4.3 “Return Measurement Issues” and Box 28.1 “Arithmetic and geometric means”) and Hull, J. C. (2018), Options futures and other derivatives, Pearson (particularly Business Snapshot 15.1 “Mutual Fund Returns Can Be Misleading”). 3 On February 2, 2016 ESMA issued a statement on “Supervisory work on potential closet index tracking” and on June 27, 2018 held a supervisors’ workshop for National Competent Authorities (NCAs) to promote supervisory convergence in NCAs’ supervision of closet indexing.

6 depends on the quality and reliability of data, and that comparability is also an issue to be considered. The previous remarks imply that appropriate care should be taken in the interpretation of such metrics, especially when dealing with a sensitive topic like closet indexing. V. Concluding remarks 21. The SMSG believes that the ESMA Report provides very useful data, information and insights on EU UCITS. The SMSG advises ESMA to consider some revisions along the lines suggested in this opinion in future editions, especially since the publication of this Report appears to be a long term project, the aim of which is to provide on a yearly basis a reliable statistical reference to understand performance and costs of mutual funds in the EU. This opinion will be published on the Securities and Markets Stakeholder Group section of ESMA’s website. Adopted on 3 October 2019 [Signed] Veerle Colaert Chair Securities and Markets Stakeholder Group