ICI Submits Comment Letter in Response to SEC's Latest Target Date Funds Release (pdf)

Via e-mail to rule-comments@sec.gov June 9, 2014 Mr. Kevin M. O’Neill Deputy Secretary Securities and Exchange Commission 100 F Street, NE Washington, D.C. 20549-1090 Re: Investment Company Advertising: Target Date Retirement Fund Names and Marketing (File No. S7- 12-10) Dear Mr. O’Neill: The Investment Company Institute (“ICI,” or the “Institute”)1 appreciates the opportunity to comment on the recent release2 of the Securities and Exchange Commission (the “Commission,” or the “SEC”) requesting additional comment related to its 2010 proposal3 to require certain disclosures in target date fund marketing materials.4 On behalf of our members, which are entrusted with the 1 The Investment Company Institute is the national association of U.S. investment companies, including mutual funds, closed-end funds, exchange-traded funds (ETFs), and unit investment trusts (UITs). ICI seeks to encourage adherence to high ethical standards, promote public understanding, and otherwise advance the interests of funds, their shareholders, directors, and advisers. Members of ICI manage total assets of $16.8 trillion and serve more than 90 million shareholders. 2 See Investment Company Advertising: Target Date Retirement Fund Names and Marketing, SEC Release Nos. 33-9570; 34- 71861; IC-31004 (April 3, 2014), 79 FR 19564 (April 9, 2014) (the “Release”). 3 See Investment Company Advertising: Target Date Retirement Fund Names and Marketing, SEC Release Nos. 33-9126; 34- 62300; IC-29301 (June 16, 2010), 75 FR 35920 (June 23, 2010)(the “2010 Proposal”). 4 We are using the term “marketing materials” in this letter to refer both to advertisements under Rule 482 of the Securities Act of 1933 (the “Securities Act”), and sales literature under Rule 34b-1 of the Investment Company Act of 1940 (the “Investment Company Act”). Securities Act Rule 156 applies to both advertisements and sales literature. Throughout this letter, we are referring to these three rules as the “Rules.” The 2010 Proposal and the Release also solicit comment on whether certain disclosures should be included in fund prospectuses, statements of additional information (“SAIs”), or shareholder reports. Mr. Kevin M. O’Neill June 9, 2014 Page 2 retirement savings of more than 50 million U.S. households,5 the Institute has been, and continues to be, committed to working with the Commission, the Department of Labor (“DOL”) and others to protect the interests of target date fund investors and enhance their understanding of these useful investment products6 In 2010, we strongly voiced our support for the spirit and core of the Commission’s 2010 Proposal. In fact, we strongly supported its most critical element—the asset allocation glide path illustration. This illustration effectively communicates key pieces of information about target date funds to investors through a prominent table, chart, or graph conveying the changing asset allocation of the target date fund over the entire life of the fund, including at the target date and the point where the fund arrives at its final asset allocation.7 The Commission requested additional comment after the Commission’s Investor Advisory Committee (the “Committee”) recommended in April 2013 that the Commission develop a glide path illustration for target date funds that is based on a standardized measure of fund risk as a replacement for, or supplement to, the proposed asset allocation glide path illustration.8 While we too are committed to enhancing investor understanding of target date funds, we believe that mandating a risk- based glide path illustration would require the adoption of a single-dimensional risk standard that simply cannot adequately convey the multi-dimensional aspects of risk. All risk measures have certain inherent limitations; therefore, requiring the use of any such measure may confuse and even possibly mislead target date fund investors. For these reasons, we have concluded that the adoption of a risk- based glide path is counter to the Commission’s goal in proposing the Rule amendments of “provid[ing] enhanced information to investors concerning target date retirement funds and reduc[ing] 5 See Figure A7 in Burham, Bogdan, and Schrass, Ownership of Mutual Funds, Shareholder Sentiment, and Use of the Internet, 2013, ICI Research Perspective 19, no. 9 (October 2013), available at www.ici.org/pdf/per19-09.pdf. 6 See, e.g., Letter from Karrie McMillan, General Counsel, the Investment Company Institute, to Elizabeth M. Murphy, Secretary, Securities and Exchange Commission (Aug. 23, 2010) (commenting on the 2010 Proposal, available at www.sec.gov/comments/s7-12-10/s71210-34.pdf) (the “2010 ICI Comment Letter”). In addition, in June 2009 the Institute published Principles to Enhance Understanding of Target Date Funds (the “Principles”), which reflected the results of the ICI Target Date Fund Disclosure Working Group project that began earlier that year. The ICI Target Date Fund Disclosure Working Group included representatives from a broad range of member firms, representing more than 90 percent of target date mutual fund assets. The Working Group reviewed the then existing disclosures applicable to target date funds, determined that the public’s understanding of target date funds could be enhanced by identifying key pieces of information that should be prominently conveyed by target date funds, and developed the Principles that spell out this key information, which includes an asset allocation glide path illustration accompanied by narrative disclosure. The Principles are available at www.ici.org/pdf/ppr_09_principles.pdf. 7 See the 2010 ICI Comment Letter. 8 The Recommendation of the Investor Advisory Committee, Target Date Mutual Funds (adopted April 11, 2013)(the “Recommendation”) is available at www.sec.gov/spotlight/investor-advisory-committee-2012/iac-recommendation-target- date-fund.pdf. On November 20, 2013, SEC Chair White sent a letter to the Committee, notifying it that the Commission would be seeking additional comment on standardized risk-based glide path illustrations for target date funds in connection with the Commission’s 2010 Proposal (available at www.sec.gov/spotlight/investor-advisory-committee-2012/chair-white- letter-target-date-funds.pdf). Mr. Kevin M. O’Neill June 9, 2014 Page 3 the potential for investors to be confused or misled…”9 We urge that the Commission not adopt Rule amendments that stipulate the use of a risk-based glide path illustration for target date funds.10 Rather, we urge the Commission to continue with its approach to the glide path set forth in the 2010 Proposal. More specifically, our comments in response to the Release include the following points, all of which are discussed in greater detail below. • There is no single measure of risk on which the industry has settled. This is due to the complexity and multi-faceted nature of risk and the inherent limitations of any single measure. • Managers of target date funds attempt to address a variety of risks faced by individuals investing for retirement. In addition to return volatility risk, target date fund managers also consider longevity or shortfall risk (i.e., the risk of outliving one’s assets), inflation risk (i.e., the risk that the purchasing power of one’s assets will erode over time), and income replacement risk (i.e., the risk that the income provided for in retirement will not be sufficient). • Risk metrics can be helpful tools for assessing one of these risks, i.e., the potential return volatility risk of an investment portfolio, but a simplistic use of such measures, alone or in the construction of a risk-based glide path, will harm, rather than help, investors. Widely identified weaknesses include the limitations of historical data on which they are based, the challenges associated with adapting such data to a particular portfolio, and the inability of data to account sufficiently for market events with no historical precedent. • While a few foreign jurisdictions require or permit funds to disclose “risk rating” information, the use of such ratings and their underlying methodologies have been criticized for these and other limitations that affect risk-based metrics generally. • Adoption of a mandatory risk measure, either alone or in connection with the construction of a risk-based glide path illustration, presents several unique challenges when applied to target date funds, including the following: 9 See 2010 Proposal, 75 FR at 35920. 10 While the vast majority of ICI members agree with the positions expressed in this letter, we understand that at least one ICI member is generally supportive of the Committee’s recommendations with respect to the use of a risk-based metric as a supplement to the asset allocation glide path illustration. Mr. Kevin M. O’Neill June 9, 2014 Page 4 o A risk-based glide path focused solely on return volatility risk cannot capture all forms of risk to which target date funds are subject, such as inflation risk and longevity risk; o As a fund’s asset allocation changes, its risk profile also changes, and therefore it is not possible to take a risk statistic such as standard deviation and extrapolate the same risk metric out over the life of a target date fund; and o The limitations associated with the use of historical data are more pronounced for target date funds, which are composed of multiple asset and sub-asset classes whose weightings and representative investments change over the funds’ investment horizons; interpreting from data how and at what level of precision these varying assets correlate presents difficult challenges. • Investors may be confused, and may be misled, by a risk-based glide path illustration, because such an illustration: o may not be comprehensible to investors not familiar with the statistical concepts underpinning the glide path’s construction (e.g., standard deviation or beta); o would suggest that future levels of risk in a fund are reasonably predictable, but return volatility risk measures are probabilistic in nature, not exact, and investors likely will view the illustration as predictive of future performance, or even regard it as promissory in nature; o would not accurately reflect how most target date funds are managed; and o would cause investors to de-value other important investment considerations, such as longevity and inflation risks and return potential, which will make it more difficult for them to realize their retirement goals. • An asset allocation glide path is not only an effective proxy for return volatility risk, it also shows actual intended asset allocations, allowing for better glide path comparability among target date funds. • The asset allocation glide path illustration and other disclosure requirements contemplated in the 2010 Proposal, together with all of the information that target date fund investors and plan administrators currently have available, provide an effective and comprehensive picture of fund risk. • The SEC has requested comment on a complex topic with no new specific Rule amendments and very limited analysis. In the event that the SEC chooses to pursue Mr. Kevin M. O’Neill June 9, 2014 Page 5 some type of risk-based glide path, the SEC formally should propose specific new Rule amendments accompanied by careful analysis (including a detailed regulatory impact analysis). • In the event that the SEC chooses to pursue some type of risk-based glide path, the U.S. Department of Labor should impose similar rules on non-mutual fund target date funds and arrangements to assure that all retirement investors receive the same basic information about these important retirement savings vehicles. I. Introduction Asset allocation is one of the most important factors in long-term portfolio performance. Target date funds provide an efficient way for an investor to invest in a mix of asset classes through a single fund that rebalances its portfolio over time to become less focused on growth of principal and more focused on current income and lessening principal fluctuation. For this reason, target date funds have proven to be a valuable innovation for defined contribution plans and other individual retirement account savings vehicles. At year-end 2013, target date mutual funds had $618 billion in assets, including $430 billion held in defined contribution plans and another $124 billion held in IRAs.11 At year-end 2012, 72 percent of 401(k) plans in the EBRI/ICI 401(k) database included target date funds in their investment lineups, 41 percent of 401(k) plan participants had at least some portion of their account in these funds, and 15 percent of total assets in 401(k) plans in the EBRI/ICI 401(k) database was in target date funds.12 Target date funds invest in multiple asset classes, ranging from domestic and international equities to corporate and government bonds and cash. To achieve the same benefits with a self- managed portfolio, an investor would have to select and monitor a number of individual funds and regularly transfer money among them. The design of target date funds avoids the extreme asset allocations observed in some retirement accounts. For example, some young workers invest very conservatively by allocating all, or almost all, of their accounts to fixed income investments, while some participants nearing retirement invest very aggressively, allocating all, or almost all, of their accounts to 11 See Table 20 in Investment Company Institute, The U.S. Retirement Market, Fourth Quarter 2013 (March 2014), available at www.ici.org/info/ret_13_q4_data.xls. 12 In an ongoing collaborative effort, the Employee Benefit Research Institute and the Investment Company Institute collect annual data on millions of 401(k) plan participants as a means to accurately portray how these participants manage their accounts. The EBRI/ICI database includes data on target date funds offered as mutual funds, collective investment trusts, and other investment vehicles. The EBRI/ICI 401(k) database is the largest, most representative repository of information about individual 401(k) plan participant accounts. The EBRI/ICI 401(k) database includes statistical information on 24.0 million 401(k) participants, in 64,619 employer-sponsored 401(k) plans, holding $1.536 trillion in assets at year-end 2012. See Holden, VanDerhei, Alonso, and Bass, 401(k) Plan Asset Allocation, Account Balances, and Loan Activity in 2012, ICI Research Perspective 19, no. 12, and EBRI Issue Brief, no. 394 (December 2013), available at www.ici.org/pdf/per19- 12.pdf. Mr. Kevin M. O’Neill June 9, 2014 Page 6 equity investments.13 Target date funds follow professionally designed asset allocation models to eliminate such extremes. The Commission issued the 2010 Proposal in response to the market events of 2008. Most asset classes (with the notable exception of U.S. Treasury securities) lost value in 2008. Consequently, many funds (including target date funds) and investment accounts generally experienced negative returns. It has been observed, however, that target date funds’ use of a consistent asset allocation strategy positioned their shareholders to benefit from the subsequent market recovery.14 The Committee’s Recommendation stems from its concern about the effectiveness of the asset allocation glide path as a risk measure.15 The Recommendation states that asset allocation “is a particularly unreliable proxy for risk where the asset classes are defined quite broadly, as they are in the Commission proposal” and therefore requests that the Commission “develop an alternative glide path illustration based on the target risk level over the life of the fund.” The Committee does not offer a specific proposal in this regard, but suggests that the Commission “focus on factors such as maximum exposure to loss or volatility of returns that are directly relevant to the primary concerns of those approaching retirement.”16 For the reasons set forth below, we strongly object to moving away from the asset allocation glide path illustration in favor of a risk-based glide path illustration. The discussion below responds to a number of the questions posed in the Release. II. Assessing a Risk-Based Glide Path The Release asks whether there are quantitative measures of risk that would be useful to and understandable by investors as the basis for a fund’s risk-based glide path illustration. Of course, this is not the first time the Commission has explored the possible adoption of risk-based measures for investment companies. In its concept release issued in 1995, the SEC requested comment on how to 13 See 401(k) Plan Asset Allocation, Account Balances, and Loan Activity in 2012, ICI Research Perspective 19, no. 12, and EBRI Issue Brief, no. 394, supra. 14 See Morningstar, Target-Date Series Research Paper: 2013 Industry Survey (“Morningstar Survey”), at 20 (“And although the more aggressive tendencies of some shorter-dated funds produced unexpectedly severe losses in 2008, results since then have worked in favor of investors who were patient. From the pre-crisis peak in 2007 through the end of 2012, all but one of the currently extant 2015 funds have recouped their losses, and most have produced double-digit positive gains.”), available at corporate.morningstar.com/us/documents/ResearchPapers/2013TargetDate.pdf. 15 The Committee’s stated rationale in the Recommendation regarding glide paths is similar in many respects to that offered in the Comment Letter of Foliofn Investments Inc. to the 2010 Proposal (Jan. 24, 2011) (the “Foliofn Comment Letter”). 16 See the Recommendation, at 3-4. Mr. Kevin M. O’Neill June 9, 2014 Page 7 improve risk disclosure for investment companies,17 and noted in 1998 that “[c]omments submitted in response to the Commission’s Risk Concept Release asserted that investors have too wide a range of investment goals and ideas of what ‘risk’ means to be well served by a single quantitative risk measure. In addition, commenters argued that, if the Commission mandated a risk measure, investors might rely on it as a definitive standard despite the lack of general agreement on how to measure risk.”18 We believe that these important points are equally valid today, particularly with respect to target date funds. In our view, the absence of industry agreement on a single risk measure is not due to a lack of effort or ingenuity. Rather, it is due to the complexity and multi-faceted nature of risk, and the inherent limitations of any single measure—challenges that are exacerbated with target date funds, given their multiple asset classes and changing asset allocations. A. Defining “Risk” As an initial matter, a proper evaluation of the relative merits of an asset allocation glide path illustration and a risk-based glide path illustration requires an understanding of “risk.” Risk, broadly and properly understood, is a multi-faceted concept. The variance in investment returns for an investment (e.g., standard deviation, a measure of variability around an investment’s average returns) provides one potential measure of risk.19 The sensitivity of an investment’s returns to those of the market generally (i.e., beta) would provide another potential measure of risk. In layman’s terms, these measures focus on the likelihood that an investment will go up and down in value relative to a base line. Throughout this comment letter, we will refer to the concept that these measures attempt to capture as “return volatility risk.” Neither of these measures of return volatility risk, however, whether alone or together, fully captures “risk.” Investors generally, and those saving for retirement in particular, must consider and plan for risks beyond return volatility risk. Target date funds generally are used as retirement savings vehicles, and therefore their investors also consider longevity or shortfall risk (i.e., the risk of outliving one’s assets), inflation risk (i.e., the risk that the purchasing power of one’s assets will erode over time), and income replacement risk (i.e., the risk that the income provided for in retirement will not be 17 See Improving Descriptions of Risk by Mutual Funds and Other Investment Companies, SEC Release No. IC-20974 (Mar. 29, 1995), 60 FR 17172 (Apr. 4, 1995) (the “Risk Concept Release”). 18 See Registration Form Used by Open-End Management Investment Companies, SEC Release No. IC-23064 (Mar. 13, 1998), 63 FR 13916, 13929 (Mar. 23, 1998)(the “1998 Release”). 19 Another risk measurement referred to in the Release is semi-variance, which can be used to measure the variability of returns below the average return. Mr. Kevin M. O’Neill June 9, 2014 Page 8 sufficient).20 Investing for retirement requires trade-offs between different forms of risk, and the manner in which these trade-offs are and should be made will depend in large part on the particular and varied financial circumstances of investors. As previously acknowledged by the Commission,21 managers of target date funds attempt to address a variety of risks faced by individuals investing for retirement, including investment risk (i.e., return volatility risk), inflation risk, and longevity risk. Target date fund managers take different approaches to balancing these risks, and typically hold varying levels of equities, bonds, cash, and other types of investments in a portfolio that becomes more conservative as the fund approaches (and in some cases, passes) the targeted retirement date. Moreover, target date funds continue to evolve with respect to their portfolio and risk management strategies and techniques. For instance, target date funds increasingly are diversifying beyond equities and bonds into new asset and sub-asset classes such as emerging markets, real estate, Treasury inflation-protected bonds, natural resources, and commodities.22 B. Evaluating Potential Risk Measures That Could Be Used in the Construction of a Risk- Based Glide Path We discuss below the risk measures identified in the Release.23 We also include a summary of criticisms expressed regarding the various “risk-based” ratings adopted in other jurisdictions as well as a survey of those approaches. The clear lesson learned from this review is that, while risk metrics can serve a critical function when used by investors who are knowledgeable of their limitations,24 a simplistic presentation of these measures alone or in the construction of a risk-based glide path illustration raises significant concerns. 20 Of course, a number of other risks may be subsumed by return volatility risk for a particular target date fund, including stock market risk, small- and mid-cap stock risk, foreign risk, currency risk, interest rate risk, credit risk, commodity risk, and asset allocation risk, among others. 21 See 2010 Proposal, 75 FR at 35921. 22 See Morningstar Survey, at 37. 23 For a discussion of other potential measures of risk (e.g., Sharpe ratio, Treynor ratio, Jenson’s alpha, etc.) and their related limitations, see Letter from Paul Schott Stevens, General Counsel, the Investment Company Institute, to Jonathan G. Katz, Secretary, Securities and Exchange Commission (July 28, 1995), available at www.ici.org/pdf/7147.pdf. We believe that these measures of risk are susceptible to the same fundamental limitations as standard deviation, as discussed below. For a discussion of value at risk (VaR) and its limitations, see Andreas Krause, Exploring the Limitations of Value at Risk: How Good Is It in Practice?, The Journal of Risk Finance (Winter 2003), available at www.gravitascapital.com/Research/Risk/ValueAtRisk/VAR%20Limitations%20in%20Practice%20feb03.pdf. 24 See infra, note 53 and accompanying text. Mr. Kevin M. O’Neill June 9, 2014 Page 9 1. Standard Deviation and Semi-Variance The Release specifically mentions and briefly describes standard deviation and semi-variance as risk measures before asking whether there is a particular quantitative risk measure, or group of risk measures, that are helpful in evaluating the risks of target date funds. It also asks whether fund investors would be likely to understand these risk measures and be able to use them effectively in making investment decisions. Standard deviation is a measure of variability around an investment’s average returns. While some investors no doubt find this measure useful, when applied to target date funds its relevance and utility for all investors is unclear. Investors may have difficulty deciphering the raw standard deviation figures upon which the risk-based glide path would be constructed. Specifically, standard deviation is used to calculate the probability that under certain circumstances a fund’s returns will fall within a range of its average returns approximately two-thirds of the time.25 Many investors, however, are not likely to understand the measure or appreciate its limitations with respect to predictability and precision.26 For instance, a fund that consistently loses 10 percent annually would have a standard deviation of zero, and investors may mistakenly view such a fund as having no risk, even though the negative return would leave the investor exposed to inflation, shortfall, and income replacement risks. Second, like most risk metrics, standard deviation is based on historical data. Investors, however, may place too much confidence in such a measure as being able to predict future performance, or even regard it as promissory in nature, despite accompanying disclosure to the contrary. If the Commission required the use of a risk-based glide path that “projected” standard deviations for future years, this danger would be magnified. As a result, investors may not understand that a fund in practice could behave very differently as a result of market conditions, changes in underlying portfolio holdings and the correlations in returns of those holdings. Third, given a target date fund’s changing asset allocations, measures of a target date fund’s standard deviation likely would be limited to relatively short time intervals (e.g., annualized standard deviation of returns). Such a relatively short-term volatility measure would not be particularly relevant to retirement investors with long-term time horizons. We are concerned that a lack of investor understanding about this potential mismatch in risk measure and time horizon could hurt their chances of adequately saving for retirement. 25 For example, if the average annual return of a fund were 15 percent and the standard deviation were 12 percent, there would be about a two-thirds chance that the annual return would be between 3 percent and 27 percent. There would be about a one-third chance that the annual return would be outside of this range. 26 Indeed, the validity of standard deviation rests upon various conditions being satisfied. These conditions include constancy of the standard deviation, which history has shown to be highly variable, and a normal distribution of the fund’s returns, which is generally not the case. Even assuming that these conditions are satisfied, there would be a one-third chance that the return would be outside of the range. Mr. Kevin M. O’Neill June 9, 2014 Page 10 The Commission appears to grasp the inherent difficulties in using standard deviation as a basis for a risk measure for target date funds. In this respect, the Release asks how to address the concerns of investors at different points in the cycle of accumulating and distributing retirement assets This is an important concern, because views of and responses to risk by target date fund investors will not be uniform. Unfortunately, we do not believe it is possible for the Commission to select a risk measure that would have a high degree of relevance for differently situated investors. For example, it may be entirely appropriate for a relatively young investor decades away from retirement to conclude that the potential for increased long-term rates of return is worth taking on increased short-term volatility. The danger is that such an investor could be discouraged from investing in a fund with a relatively higher standard deviation, concluding based on a review of this single metric that the fund is “too risky.” Thus, if the Commission adopted standard deviation as the basis for a risk-based glide path, it implicitly would be encouraging many investors unwisely to favor funds with less short-term volatility, and thus forego potentially superior long-term fund performance. Semi-variance can be used to measure the variability of returns below the average return.27 A fund with a larger semi-variance has returns that are more spread out below the average return. Insofar as this measure presents only the downside of volatility, i.e. the extent to which returns may fall short of average returns, the possibility of discouraging investors from selecting funds with a larger semi-variance (and increased potential for long-term gains) is perhaps greater than that associated with funds with larger standard deviations, which at least acknowledge the upside return potential relative to average returns. Otherwise, the limitations and difficulties of using standard deviation described above (e.g., the potential to confuse investors, and the limitations of using historical data in a risk measure) equally apply to the use of semi-variance in constructing a risk-based glide path. 2. Beta The Release also mentions and briefly describes beta, and specifically asks what issues, if any, are associated with the selection of an appropriate benchmark for calculating beta. Beta, i.e. the sensitivity of an investment’s returns to those of the market generally,28 would provide another potential measure of return volatility risk upon which a risk-based glide path could be constructed. Presumably, use of beta in this way would involve a correlation between a fund’s historical returns (or perhaps, for future years, those of its future expected asset classes) and those of a benchmark index. Unlike standard deviation, it provides a relative, rather than absolute, measure of volatility. 27 By contrast, standard deviation shows variability of returns above and below the average return. 28 The relevant market’s beta is by definition 1. A fund with a beta greater than 1 tends to move more than the market’s return. For example, if a fund has a beta of 2, a 10 percent market gain could be expected to result in a 20 percent fund gain, and a 10 percent market loss could be expected to result in a 20 percent fund loss. A fund with a beta less than 1 would move less than the market’s return. For example, if a fund has a beta of 0.5, a 10 percent market gain could be expected to result in a 5 percent fund gain, and a 10 percent market loss could be expected to result in a 5 percent fund loss. Mr. Kevin M. O’Neill June 9, 2014 Page 11 Beta is meaningful only in relation to an index. This presents a unique challenge for target date funds, which invest in a number of asset and sub-asset classes whose weightings shift over time, and for which there is unlikely to be an obviously appropriate index to use for this purpose. If a broad-based equity index were used, we would expect the risk-based glide path to show increasing divergence from the index, and a declining beta relative to that index, over time.29 The practical benefits of such a presentation for investors are questionable. Because beta is a measure of relative risk, it would not convey the absolute return volatility risk of either the index or the fund. This glide path would be useful only to those investors who already have a good sense of the absolute volatility of the index, and even then, they would have to, in effect, extrapolate the fund’s comparative absolute volatility from the illustration. In addition to these unique limitations and difficulties in using beta, we believe that the limitations and difficulties of using standard deviation described above (e.g., the potential to confuse investors, and the limitations of using historical data in a risk measure) equally apply to the use of beta in constructing a risk-based glide path. C. Other Jurisdictions’ Use of Risk Measures for Funds As discussed above, a few foreign jurisdictions require or permit funds to disclose risk rating information. Appendix A includes a summary of these approaches. The risk ratings for funds and their underlying methodologies currently in use in such jurisdictions have many of the same flaws that affect risk-based metrics generally, particularly when they are presented in a simplistic way. In fact, a review of these risk rating systems and the criticisms they have drawn readily illuminates the pitfalls of attempting to identify and measure risk by using a single standardized numerical score or label (e.g., “Very High [Risk]”), either in conjunction with the glide path or on a stand-alone basis. In our view, the risk rating systems described in Appendix A and below, or any variations thereof, are not appropriate for U.S.- registered target date mutual funds (or indeed, for U.S.-registered funds generally). For instance, the European Securities and Markets Authority (“ESMA”) has required that all UCITS IV funds include the Synthetic Risk and Reward Indicator (“SRRI”) number (which ranges from 1 (lowest risk/reward) to 7 (highest risk/reward)) in their Key Investor Information Documents (“KIIDs”).30 Commentators have expressed numerous concerns with the SRRI, including, for example, that different sources of risk are not captured by the measure;31 that risk ratings may cause investors to 29 The Foliofn Comment Letter provides examples of how such a measure s may look in application. Interestingly, these graphical representations do not look very different from a graphical depiction of a fund’s equity allocation over time, i.e., an asset allocation glide path. 30 See infra, note 75, for a description of KIIDs. 31 Paul D. Kaplan, PhD, Quantitative Research Director, Morningstar Europe, Ltd., What the Synthetic Risk Reward Indicator Doesn’t Indicate About Risk or Reward (2011), available at www.vff.no/filestore/PaulKaplan- WhattheSyntheticRiskRewardIndicatorDoesntIndicateAboutRiskorReward.pdf (stating, among other things, that SRRI Mr. Kevin M. O’Neill June 9, 2014 Page 12 place too much emphasis on volatility and forego the opportunity to benefit from higher returns;32 and that, because similar types of funds tend to fall within just a couple of the risk categories (i.e., “crowding”), the benefits of these comparative scores are limited.33 Similarly, Australia’s use of the Standard Risk Measure (“SRM”) (which utilizes a 7-level classification system with risk labels ranging from “Very Low” to “Very High”), which Australian authorities “strongly recommended” in 2011 for Australian super funds,34 has also been criticized. Commentators have expressed concern that, in the determination of SRM, the size of an adverse event is not considered in the calculation and, rather, any negative return regardless of magnitude is treated the same; that the calculation methodology is subjective, creating an incentive to adopt a methodology resulting in a lower risk score;35 that the methodology misleadingly applies a single level of risk to any investment option, given that the level of risk that is undertaken by investors relates to their investment timeframes and objectives, which will vary; that annual volatility, which is what SRM measures, is not the only risk and is much less of a risk for younger investors than inflation/adequacy; and that SRM reinforces short-term thinking and investing through its focus on volatility in annual returns.36 Finally, Canadian fund managers are required to provide risk ratings (based on a 5-level scale ranging from “Low” to “High”) for mutual funds in their Fund Facts.37 As with Australia’s SRM, Canada’s current approach has been criticized for, among other things, the subjectivity of the methodology used.38 does not adequately capture certain sources of risk (e.g., market risk, active risk, and currency risk), and that the possibility of extreme events is not revealed). 32 Alex Hoctor-Duncan, Head of Retail in Europe, the Middle East and Africa for BlackRock, Crouching Tiger Hidden Danger, Strategic Solutions Magazine Q1-2 2012 issue, available at www.blackrockinternational.com/public/en- zz/literature/brochure/strategic_solutions_crouching_tiger.pdf. 33 Ed Moisson, Head of UK & Cross-Border Research, Lipper, and Kevin Pollard, Research Analyst, Methodology, Lipper, SRRI European Overview (May 2012), available at share.thomsonreuters.com/PR/Lipper/Reports/Lipper_SRRI_European_May2012.pdf. For the period studied, 94.5% of equity funds had a risk rating of 6 or 7, and 74.4% of bond funds had a risk rating of 3 or 4. 34 See infra, note 78, for a description of superannuation and Australian super funds. 35 David Bell, Is APRA’s Standard Risk Measure Helpful? (Feb. 20, 2013), available at www.cuffelinks.com.au/is-apras- standard-risk-measure-helpful/?COLLCC=2680257826. 36 Australian Institute of Superannuation Trustees, APRA Discussion paper, Reporting Standards for Superannuation (November 2012), available at www.aist.asn.au/media/11489/2012.11.16_APRA_DP_Reporting_Standards.pdf. 37 See infra, note 81, for a description of the Canadian Fund Facts. 38 See Canadian Securities Administrators (“CSA”) Notice 81-324 and Request for Comment—Proposed CSA Mutual Fund Risk Classification Methodology for Use in Fund Facts, in which the CSA noted that it had received feedback from stakeholders indicating that a standardized risk classification methodology would be more useful to investors as it would provide a consistent and comparable basis for measuring the risk of different mutual funds. Mr. Kevin M. O’Neill June 9, 2014 Page 13 D. Limitations of Risk-Based Measures Are Magnified When Applied to Target Date Funds In light of their limitations and difficulties discussed above, we oppose mandatory risk metrics being used either alone or in connection with the construction of a risk-based glide path. For the reasons set forth below, mandatory risk metrics are particularly unsuitable and problematic to implement (in the form of a risk-based glide path or otherwise) for target date funds. 1. Conventional Risk Measures Do Not Work for Target Date Funds As acknowledged by the SEC, a target date fund’s asset allocation changes throughout the fund’s investment horizon (i.e., its glide path) until it reaches a “landing point” at which the asset allocation generally becomes static.39 As a fund’s asset allocation changes, its risk profile also changes. It therefore is not possible to take a risk statistic such as standard deviation, calculated over a set period based on a given asset allocation at a point in time, and extrapolate the same risk metric out over the life of a target date fund. In addition, the plotting of changing risk indicators over specific points on the glide path (e.g., the starting point, retirement date and landing point) would be problematic. In particular, it would not directly address the return volatility risk of the target date fund over its entire investment horizon or, for example, the impact that taking on greater return volatility risk in the early years of the glide path would have on the fund’s ability to increase its returns over its entire investment horizon. Another difficulty in applying risk-based measures to target dates funds relates to the limitation of historical data and challenges in adapting such data to a specific target date fund portfolio. As noted above, all risk metrics use historical data to make assumptions about future risk, and history may not be a reliable predictor of future markets nor sufficiently account for major, unforeseen market events with no historical precedents (e.g., 2007–2008 market events). Beyond these obvious limitations, target date funds are composed of multiple asset and sub-asset classes whose weightings and representative investments will change over time. Interpreting from data how and at what level of precision these varying assets correlate presents potentially unknowable challenges—particularly as target date funds diversify beyond equities and bonds and seek exposure to new asset and sub-asset classes such as emerging markets, real estate, natural resources and commodities. And complicating matters further, aside from these contemplated changes in asset allocation (i.e., expected movements along the glide path over time), a target date fund’s asset allocation glide path itself will not necessarily remain fixed. In this respect, consistent with disclosure in their prospectuses, funds often reserve the right to deviate from the future asset allocation percentages identified in seeking to meet the funds’ investment objectives.40 39 See 2010 Proposal, 75 FR at 35921. 40 For example, by policy a target date fund may be permitted to deviate +/- 5 percent from the asset allocation percentages reflected in its asset allocation glide path. Mr. Kevin M. O’Neill June 9, 2014 Page 14 Given the inherent limitations in applying a conventional risk measure to target date funds with changing asset allocations, we expect that some commenters might recommend the use of stochastic modeling, such as Monte Carlo simulation techniques,41 in connection with the construction of a risk-based glide path. While fund sponsors can and do use such techniques to help them gauge the probability of achieving a desired return and construct asset allocation glide paths, there is no industry wide standard for doing so. Minor changes to the inputs and assumptions used can cause wide differences in the results.42 The potential for inconsistent inputs and assumptions is a significant impediment to systematic, rigorous analysis across target date funds. Additionally, the use of such techniques necessitates value judgments regarding, among other things, the periods along the glide path during which investors are best able to respond to return volatility risk and at what levels. The SEC could as a theoretical matter consider the adoption of a standardized approach to the use of such techniques, but, to actually implement such an approach in practice, it would need to make a series of technical decisions and value judgments regarding, among other things, the level of specificity required for inputs associated with asset classes, sub-asset classes and statistical correlations among such assets, and the number of simulations to be used.43 Finally, the challenges in applying risk-based metrics to target date funds identified above apply regardless of whether the metric is backward-looking or forward-looking.44 While one could argue that a backward- looking metric is not inherently speculative in the way that forward-looking metrics are,45 a backward-looking metric would be particularly ill-suited for use with a target date fund in light of such fund’s changing asset allocation. For instance, suppose a fund with a target retirement date of 2050 launched in 2010. It would have just over four years of actual performance—or backward-looking 41 For a discussion of Monte Carlo techniques, see Roger Eckhardt, Stan Ulam, John von Neumann, and the Monte Carlo Method, Los Alamos Science, Special Issue (15), 131-137 (1987). See also Risk Analysis in Capital Investment, Harvard Business Review (Sept. 1, 1979). 42 See Allen Roth, Is Financial Monte Carlo Simulation Dead?, Money Watch (Feb. 22, 2010). 43 In this respect, the Release requests comment on whether required risk measures, if adopted in final rules, should be based on a standardized methodology or methodologies developed by the Commission. If the SEC were to pursue a risk-based glide path, it would either: (i) have to set forth, in a very detailed and prescriptive way, how funds would construct their glide paths, which would require the SEC to make a number of technical and value-based policy choices, or (ii) allow funds discretion in creating a methodology for constructing their glide paths, which would introduce a subjective “black box” element to their construction and make comparability between target date funds impossible (indeed, the Canadian Securities Administrators have recognized similar concerns with its current risk rating system). We oppose either of these general means of construction, and the attendant flaws of each further call into question the benefits and value of the end result, i.e., the risk-based glide path. 44 The Release asks whether a risk-based glide path illustration should be backward-looking (showing past actual risk measures of a target date fund or group of target date funds) or forward-looking (showing projected risk targets for a target date fund or family of target date funds). 79 FR at 19567. 45 See Appendix A and compare the risk rating approaches taken in Europe and Australia. Mr. Kevin M. O’Neill June 9, 2014 Page 15 data—on which to gauge the potential risk of the target date fund with an investment horizon of several more decades and with asset allocations that will change over time. 2. Target Date Fund Investors May Be Confused and Potentially Misled by Risk Metrics The Release asks whether investors in target date funds would be likely to understand risk measures, or any related illustrations based on those measures. In addition to the difficulties with applying conventional risk measures to target date funds described above, we believe that the investors who are the focus of the Commission’s rulemaking would have difficulty understanding and may be misled by a risk-based glide path illustration. The market for target date funds consists largely of investors saving through retirement plans or retail investors investing through IRAs.46 The decision to make a particular target date fund available through a plan is generally made by plan fiduciaries who may be presumed to be more sophisticated and knowledgeable than investors generally. But, like many investment professionals who were said to have relied too heavily on risk metrics in failing to gauge the severity of the financial crisis in 2007–2008,47 even sophisticated investors can be susceptible to overreliance on risk metrics. Target date fund investors are not necessarily knowledgeable enough to recognize the limitations of risk metrics and likely would place too much confidence in such metrics. In this respect, presented on its own or as an indicator on a glide path, a risk metric would imply a level of precision and reliability that is unwarranted. Simply providing, say, a backward-looking measure of a fund’s return volatility risk such as standard deviation could lead investors to assume that the numbers presented are intended to represent a true indication of future volatility. Insofar as a risk-based glide path would also project risk in future years, the dangers of overreliance and overconfidence are compounded. The Release asks whether target date fund strategies are based on a changing target risk level or a changing target asset allocation over time, or some combination of these approaches. Based on our review of our members’ practices, most target date funds do not construct a risk-based glide path based on a risk metric and manage to it over time. Consequently, the imposition of a risk-based glide path would be something of a forced construct for most funds, and would give the impression, misleading in most cases, that the funds were being managed to a single, targeted risk metric and not an optimized risk/return goal, as is more likely the case, and for which the asset allocation glide path is better suited. Thus, the inclusion of a risk-based glide path may confuse and potentially mislead investors about how most target date funds are managed.48 46 See supra, note 11 and accompanying text. 47 See Joe Nocera, Risk Management, The New York Times Magazine (Jan. 2, 2009). 48 The Release also asks whether a risk-based glide path illustration should be required only for target date funds with an investment objective or strategy of managing to a target risk level. We believe that it should not, because even for those funds that do manage to a target risk level, a mandated risk-based glide path illustration may not reflect the way any Mr. Kevin M. O’Neill June 9, 2014 Page 16 The Release asks whether disclosure of risk measures could influence investors to choose lower or higher risk investments than would be consistent with their goals for accumulating retirement assets. We are concerned that investors reviewing risk-based glide path illustrations may reach wrong conclusions. For example, in side-by-side comparisons, lower-volatility funds (e.g., those with relatively heavy allocations to bonds and/or cash) may “look better” than higher-volatility funds, but the former could be inappropriate for certain investors and subject them to increased longevity risk and inflation risk and cause them to forego potentially superior returns. The Release asks whether the inclusion of two glide path illustrations (i.e., both asset allocation and risk-based glide paths) in the same document would tend to confuse investors. We believe that it would. Placing a risk-based glide path next to the asset allocation glide path would add little to investors’ understanding of a fund’s overall risk profile and, for the reasons discussed above, could confuse and/or mislead investors. Investors may also be left wondering how, if at all, the two illustrations are meant to complement one another. Presumably aware of the potential for confusion,49 the Commission asks whether a required explanatory statement preceding or accompanying the risk-based glide path illustration would be particular fund is managed, and in any event such disclosure would still be subject to the same weaknesses discussed elsewhere in this letter. However, we distinguish a risk-based glide path of the type recommended by the Committee from those that are not intended to project future fund performance or risk. For example, we understand that some members use “volatility targets” as part of their investment strategies and may therefore wish to disclose a volatility target glide path (e.g., a fund may seek to target a specific level of volatility, with such target levels declining over time). Such a glide path is distinguishable from the type recommended by the Committee in that it identifies a risk target to which the fund seeks to be managed, but does not attempt to project future fund performance or risk. See infra note 51, explaining that our concerns do not extend to such type of disclosure. At the same time, consistent with our view that funds should not be required to disclose a risk-based glide path, we believe that funds that use non-predictive measures or targets in managing their funds should not be required to provide glide path illustrations of such measures or targets. 49 In this same vein, the Release asks whether forward-looking disclosures such as projected future volatility (or other risk measures) or expected returns give rise to potential liability concerns. We believe they do and, if the SEC were to pursue this course, it would have to be mindful of the forms of liability to which funds and certain of their related persons could be subject by complying with any new Rule or Form requirements. For instance, under Sections 11(a) and 12(a)(2) of the Securities Act, investors who purchase shares of a mutual fund while the fund’s registration statement contains a material misleading statement have broad legal recourse (because Rule 482 advertisements are “prospectuses,” they would also be subject to liability under Section 12(a)(2)). From a disclosure and liability perspective, requiring a target date fund to project future levels of risk is far more problematic than requiring such a fund to project future asset allocations, because the former is much harder to predict and control. Consequently, we strongly recommend that the SEC implement measures to shield from liability funds that make such future risk projections in good faith that turn out to understate actual, realized fund risk. The Commission has precedent for using its authority to shield from potential liability certain forward-looking information that registrants are required to provide. See, e.g., Item 303(c) of Regulation S-K (applying the safe harbor provided in Section 27A of the Securities Act and Section 21E of the Securities Exchange Act of 1934 (the “Exchange Act”) to forward-looking information related to off-balance sheet arrangements and certain contractual obligations contained in the “Management's Discussion and Analysis” section of a registrant's disclosure documents) and Disclosure in Management's Mr. Kevin M. O’Neill June 9, 2014 Page 17 helpful to investors (which, among other things, could explain that certain risks are not captured by the illustration). Prophylactic disclosure about these additional risks would be difficult to accommodate in advertising materials, and would not adequately temper the message from the illustration in any event. In this way, a risk-based glide path could detract from, rather than improve, decision-making and jeopardize the achievement of retirement goals. We also do not think that the difficulties associated with a risk-based glide path could be cured if the illustration were placed in the prospectus, SAI, or shareholder report instead of marketing materials.50 The only potential benefit to doing so is that there would likely be more space to describe the illustration and its limitations. But as mentioned above, we do not believe that any prophylactic disclosure will sufficiently counterbalance the message delivered by the illustration. Otherwise, the arguments made above against the risk-based glide path’s mandated inclusion in marketing materials equally apply to its inclusion in the prospectus, SAI, or shareholder report. The Release notes that some target date funds already provide quantitative risk measures in certain materials on a historical basis. Such funds are doing so on a voluntary basis. We believe there is an important distinction between risk information that is permitted and provided by certain funds voluntarily (and that is still subject to meeting a general standard of not being misleading) and information that the SEC requires.51 The latter would have the SEC’s de facto imprimatur and would be elevated in importance (especially given that it would be presented in graphic form), and therefore runs the risk of taking on a heightened level of importance in the minds of investors. While the limitations associated with risk-based metrics make them inappropriate for mandated use in marketing materials and other required disclosure materials, we recognize that risk-based metrics can be useful in explaining how a target date fund’s glide path or portfolio holdings are intended to help the fund achieve its investment objectives. As the Commission is aware, most target date funds are selected as investment options in employer-provided retirement plans. In connection with the selection of a particular target date fund for a plan, plan fiduciaries are required under ERISA to obtain information that will enable them to evaluate the prudence of any target date funds they are considering, including information regarding investment strategies and glide paths which may impact the way in which a target date fund performs.52 In many cases, target date fund providers provide this information to, and review it with, plan fiduciaries and their investment consultants. Notably, the SEC staff has recognized Discussion and Analysis about Off-Balance Sheet Arrangements and Aggregate Contractual Obligations, SEC Release Nos. 33-8182, 34-47264 (Jan. 28, 2003), 68 FR 5982 (Feb. 5, 2003). 50 The Release asks about appropriate potential locations for the risk-based glide path illustration. 79 FR at 19576. 51 It is important to note that our concerns with the mandated use of risk metrics do not extend more generally to funds that use or disclosure of volatility targets as part of their investment strategies. See supra note 48, explaining that funds that identify volatility targets as part of their investment strategies consider the targeted glide path as an express part of the funds’ strategies, but not a measure or projection of future fund risk. 52 See the DOL’s Target Date Retirement Funds—Tips for ERISA Plan Fiduciaries, available at www.dol.gov/ebsa/pdf/fsTDF.pdf. Mr. Kevin M. O’Neill June 9, 2014 Page 18 in certain circumstances a distinction between customized information, which is intended for sophisticated investors (e.g., wealthy individuals, pension funds, universities, and other institutions that have sufficient assets to justify the cost of a one-on-one presentation), and information intended for investors generally.53 Here too, while risk metrics can serve as useful tools in target date fund selection and oversight, the potential for overreliance on their predictive qualities or for making investment decisions makes their mandated use in disclosure materials problematic. Finally, risk-based glide paths would require projections of risk in future years, which is potentially inconsistent with FINRA’s public communications rules54 and their underlying policy rationale. To fully explain the glide path, its underlying assumptions and limitations, FINRA could require lengthy explanations and/or footnotes, which would pose an additional obstacle to investor comprehension and would certainly detract from the clarity of the presentation. In the 2010 Proposal, the SEC stated that its main objective in proposing the Rule amendments was to “reduce the potential for investors to be confused or misled regarding [target date funds] and other investment companies.”55 In this regard, given the difficulty that investors would have in understanding the information conveyed and the possibility that they would be misled, a risk-based glide path would represent a major step back from the asset allocation glide path contemplated in the 2010 Proposal.56 III. Assessment of the 2010 Proposal In addressing the Committee’s criticisms of the asset allocation glide path, it is important to assess how the disclosures described in the 2010 Proposal in their entirety, together with other 53 See Investment Company Institute, SEC No-Action Letter (pub. avail. Sept. 23, 1988) (stating that the SEC staff would not recommend enforcement action to the Commission if an investment adviser provides prospective clients with performance results for advisory accounts on a “gross basis” (i.e. without deducting advisory fees or other expenses paid by clients) in a one-on-one presentation as described in the letter, provided certain conditions are satisfied). 54 FINRA Rule 2210(d)(1)(F) generally prohibits communications that make unwarranted forecasts. In addition, Securities Act Rule 156(b)(2) provides that representations about past or future investment performance could be misleading because of statements or omissions made involving a material fact, including situations where representations, whether express or implied, are made about future investment performance. 55 See 2010 Proposal, 75 FR at 35920. 56 The Release asks whether radio and television advertisements should be required to include information about a target date fund’s risk-based glide path. In the 2010 ICI Comment Letter, we supported the Commission's decision not to apply the asset allocation glide path illustration requirement to radio or television advertisements, noting that it would be difficult, if not impossible, to convey this information effectively in those media and could result in substantially increased costs for additional advertising time. Our views on this are unchanged, and would apply with even greater force to a risk- based glide path, which is much more difficult to understand. Mr. Kevin M. O’Neill June 9, 2014 Page 19 information that investors and plan administrators currently have at their disposal, will enhance investor understanding of target date funds. A. The Committee’s Recommendation The Committee questions the use of an asset allocation glide path illustration as a proxy for risk. The Committee’s Recommendation begins by acknowledging that “[m]uch of those differences in risk [i.e., differences in returns in 2008 and 2009 among funds with a 2010 target date] can be explained by differences in the asset allocation models and glide paths used by different funds, as different target-date funds with the same target dates may pursue vastly divergent investment strategies.” The Recommendation notes that “choices of assets within the various asset classes, inclusion of assets from outside the traditional asset classes, and other risk management practices can also have a significant impact on fund risk levels.” The Recommendation then concludes on this issue: “A glide- path illustration based solely on asset allocation is therefore unlikely to reliably capture potentially significant differences in fund risk levels. Asset allocation is a particularly unreliable proxy for risk where the asset classes are defined quite broadly, as they are in the Commission proposal.”57 In response to this critique, the Release asks whether the proposed asset allocation glide path illustration, without a risk-based glide path illustration, would adequately convey risk information to investors. Evidence indicates that, in 2008, a broad asset allocation glide path (i.e., one showing a fund’s investments in stocks and bonds only) would have provided effective information to investors in funds near their target dates. Appendix B presents results of a statistical analysis that seeks to explain the 2008 returns of individual target date funds. It is based on what investors knew from data that were publicly available in 2007 about the percentages of these funds’ assets that were held in stocks and in bonds. The results indicate that very basic asset allocation information explains most of the wide variability in the 2008 returns of target date funds had investors known in 2007 how stock and bond markets would behave the following year. From this analysis, two conclusions can be drawn. First, to the extent investors in 2010 target date funds were surprised in 2008 by the performance of their funds, it most likely reflects an unawareness that funds near to their target retirement dates may still hold a significant portion of their assets in stocks or other securities that are more volatile than, say, U.S. government bonds. Second, any measure intended to highlight to investors the range of returns they might experience will heavily depend on actual market outcomes. In years when the stock market declines, those target date funds with heavier allocations to stocks may experience unfavorable returns. Alternatively, in years when the bond market declines but the stock market advances, such as in 2013, these same funds would experience favorable returns. Thus, the perceptions that investors form of their target date funds will depend in large part on specific stock and bond market outcomes. This illustrates 57 See the Recommendation, at 3. Mr. Kevin M. O’Neill June 9, 2014 Page 20 a difficulty that could arise from a risk-based glide path: presumably, a fund with a heavy allocation to fixed income securities would be seen as “less risky,” but in a year such as 2013 such a fund might nevertheless surprise its investors because of muted or even negative returns. The Institute’s analysis suggests that even a simple asset allocation glide path illustration (e.g., one showing only equities, bonds, and cash) can effectively explain fund performance (after the fact), and conveys useful information about a target date fund’s overall risk profile and return potential. While the Committee is correct that a more granular asset allocation glide path illustration (e.g., one showing sub-asset classes such as domestic and international equities and bonds, or alternative asset classes) would be a better proxy for return volatility risk, our research suggests a diminishing predictive return when additional variables are considered in the form of sub-asset classes or alternative asset classes. We are concerned that a more elaborate asset allocation glide path could reduce the clarity and understandability of the presentation, thereby outweighing the benefits of increased granularity. In other words, there are trade-offs between the amount of detail that can be captured in this type of presentation and its clarity and understandability, especially for those investors who are the focus of the Commission’s concern. B. Evaluating the Asset Allocation Glide Path Illustration Beyond conveying useful information about a target date fund’s overall risk profile and return potential, the asset allocation glide path illustration offers a number of other benefits, especially in comparison with a risk-based glide path. First, asset allocation glide path illustrations, combined with narrative disclosures that help investors to infer levels of risk from an asset allocation glide path, convey critical information about both the funds’ future investments and their investment risk profiles in an easy-to-understand format. An asset allocation glide path illustration conveys important information about a fund’s overall risk profile and return potential without implying that it is possible to actually predict it. For instance, an investor could look at an asset allocation glide path illustration of a fund with a relatively heavy allocation to equities and infer that the fund places more emphasis on return generation and managing longevity and inflation risk, and less emphasis on minimizing short-term volatility. By contrast, a risk-based glide path may not be comprehensible to investors not familiar with statistical concepts upon which it would be constructed, such as standard deviation, beta, or whatever other risk-based alternative may be chosen. We believe that those investors who are the focus of the Commission’s concern are far more likely to understand the basic risk characteristics of stocks, bonds and cash, and what a reduction of the stock allocation over time means for a fund’s overall risk profile. Second, target date fund managers more commonly (when looking at overall assets of target date funds or the overall number of funds) construct a targeted asset allocation and manage to it over time, rather than construct a risk-based glide path based on a risk metric and manage to it over time. Consequently, an asset allocation glide path illustration better reflects how most portfolio managers think about and manage their funds. We would caution the Commission against placing too much Mr. Kevin M. O’Neill June 9, 2014 Page 21 emphasis on the distinction between the two management approaches. Even for funds that are managed to an asset allocation glide path, risk considerations (broadly understood to include, e.g., return volatility risk, inflation risk, longevity risk and income replacement risk) are a critical component in the construction and continuing evaluation of the asset allocation glide paths and the selection of investments in building portfolios. Third, an asset allocation glide path illustration shows actual intended asset allocations, and therefore allows for better glide path comparability among target date funds than would a risk-based glide path. A risk-based glide path can be based only on predictive measures of volatility in future years, over which portfolio managers have far less control than a fund’s asset allocation. C. Other Information Complements the Asset Allocation Glide Path In addition to the asset allocation glide path, the 2010 Proposal requires a statement summarizing key features of the fund, including the importance of considering the investor’s risk tolerance, personal circumstances, and complete financial situation, as well as a statement that an investment in the fund “is not guaranteed and that it is possible to lose money by investing in the [fund].”58 This information, together with all of the information that target date fund investors and plan administrators currently have at their disposal, provides a comprehensive and multi-faceted picture of fund risk. Investors and plan administrators receive, or have ready access to, risk information in prospectuses,59 SAIs,60 shareholder reports,61 and marketing materials. Prospectuses and SAIs contain detailed narrative risk disclosure, and shareholder reports (through, e.g., management’s discussion of fund performance) include (often more topical) risk information as well. In addition, the performance bar charts in fund prospectuses62 and summary prospectuses and the performance “line graphs” in annual reports present useful and intuitive graphic measures of risk. The bar chart requirement is intended to illustrate graphically the variability of a fund’s returns and thus provide investors with an 58 See 2010 Proposal, Proposed Rule 482(b)(5)(ii). 59 See Items 4(b) and 9(c) of Form N-1A. 60 See Items 16(b) and 17(b) of Form N-1A. 61 See Item 27(b)(7)(i) of Form N-1A. 62 A target date fund (or any other mutual fund) that has annual returns for at least one calendar year is required to include in its prospectus (i) a bar chart showing the fund’s annual total returns for each of the last 10 calendar years (or for the life of the fund if less than 10 years), (ii) the fund’s highest and lowest returns for a quarter during the 10 years or other period of the bar chart, and (iii) a brief explanation of how the information illustrates the variability of the fund’s returns. See Item 4(b)(2) of Form N-1A. Mr. Kevin M. O’Neill June 9, 2014 Page 22 idea of its risk.63 Accordingly, target date fund investors currently receive a graphical illustration of return volatility risk. The bar chart’s virtues are that it (i) includes information that investors already understand (calendar year returns), (ii) shows sequential year-by-year returns (positive or negative) in a fashion that also demonstrates intuitively how a fund’s returns have varied historically (e.g., dramatic changes in the year-to-year returns of a fund would lead investors to conclude that the fund has highly variable returns and high short-term return volatility risk), and (iii) consists entirely of historical, fund- specific, verifiable data. A risk-based glide path cannot claim likewise. The asset allocation glide path illustration contemplated in 2010 Proposal for marketing materials would effectively complement this currently available information. With respect to the asset allocation glide path illustration, the Release asks whether the Commission should specify the particular asset categories required to be shown, and if so, how narrow those asset categories should be. Many target date fund managers use an asset allocation glide path that presents the fund’s asset allocation in a more granular fashion. Such granularity is typically presented to differentiate a particular fund from others in the market generally, and is usually accompanied by a narrative description of the relevance of the additional asset classes depicted to the investment strategy of the fund. For example, depicting allocations to alternative asset classes, including real estate and commodities, might be relevant to an investment strategy focused on minimizing return volatility risk. Similarly, a more granular depiction within asset classes, e.g., equities, might be used together with a narrative description to explain a manager’s position that the types of equities held by the fund are relevant in ascertaining a fund’s expected volatility.64 In such circumstances, we agree that a more granular glide path illustration can be helpful in conveying additional useful information. As we noted in our 2010 comment letter, we support the Commission’s flexible approach in the 2010 Proposal in not prescribing asset classes (in part to accommodate the strategies of funds seeking exposure to alternative asset classes such as commodities and real estate)65 and continue to believe that the antifraud rules under the federal securities laws, FINRA public communications rules, and FINRA review of 63 See the 1998 Release. The SEC adopted the bar chart after determining that investors needed improved disclosure about the risks of investing in a fund. Interestingly, the SEC was careful even when adopting this new disclosure requirement to place it in proper context: “In adopting the bar chart requirement, the Commission does not mean to suggest that all, or even a significant portion of all, fund investors equate variability in a fund’s returns with the risks of investing in the fund. As discussed below, the Commission acknowledges that investors have a wide range of ideas of what ‘risk’ means.” See n.52 of the 1998 Release. 64 For example, a fund’s allocations to relatively low volatility equities like blue chip stocks might be highlighted. 65 In the 2010 Proposal, the Commission proposed requiring disclosure of asset allocation among “types of investments” and clarified that “types of investments” would mean the underlying asset classes in which the target date fund invests, whether directly or through other funds (i.e., target date funds would have to “look through” any underlying funds). The 2010 ICI Comment Letter urged the Commission to reexamine this proposed approach, because requiring funds to use the look- through approach would add little to enhancing understanding of the fund for most investors, and could be difficult to accomplish in marketing materials (e.g., underlying funds could be unaffiliated, making this information harder to obtain). Mr. Kevin M. O’Neill June 9, 2014 Page 23 marketing materials provide ample protection for investors from misleading presentations of asset allocations.66 In cases where a fund opts to use a simpler asset allocation glide path illustration—perhaps due to concerns that the increased granularity would complicate the illustration—investors and plan fiduciaries would still be able to obtain a more detailed view of the fund’s asset allocation by reviewing portfolio holdings information. This information must be provided to shareholders and/or filed with the SEC at least quarterly (many funds voluntarily disclose portfolio holdings information more frequently on their websites).67 Funds may also be reluctant to be more granular in their asset allocation glide paths because they may not be able to project with specificity and confidence how much they intend to invest in sub-asset classes (e.g., international equities or corporate bonds) twenty or thirty years into the future. This makes the construction of a granular glide path in future years a more speculative exercise. IV. Process Considerations in Connection with the Release The 2010 Proposal contained specific proposed Rule amendments. The specificity of the amendments and the accompanying detailed discussion in the 2010 Proposal provided the public with a specific proposal to consider and upon which to provide comment. The Release, on the other hand, takes a much different approach. It provides background concerning target date funds and summarizes the 2010 Proposal and the Committee’s Recommendation, and then raises a series of questions, designed to assist the SEC in determining whether it should develop a risk-based glide path illustration either as a replacement for, or supplement to, the asset allocation glide path proposed in 2010. On the subject of the costs and benefits associated with the adoption of a standardized risk measure, the Release is similarly open-ended, and consists almost entirely of a series of questions. This approach taken in the Release would be highly insufficient as a prelude to final rulemaking given that (i) the subject of investment risk is exceedingly complex, (ii) there is likely to be strong disagreement, as indicated by our comments here, that substantial changes to the glide path contemplated in 2010 are warranted, and (iii) the Committee’s Recommendation itself was not specific with respect to the manner in which the risk-based glide path should be constructed. Indeed, the Release is reminiscent in many ways of the SEC’s Risk Concept Release. In addition to the substantial overlap in subject matter, each of the Risk Concept Release and the Release provided some background and raised a number of important questions, but did not propose amendments to rules or forms. As noted in the Release,68 the SEC received over 3,700 comment letters in response to the Risk Concept 66 FINRA Rule 2210 requires, among other things, that mutual fund retail communications be based on principles of fair dealing and good faith, be fair and balanced, and provide a sound basis for evaluating the facts of any particular security. 67 See Sections 30(b) and (e) of the Investment Company Act and Investment Company Act Rules 30b1-5 and 30e-1. 68 See the Release at n.20. Mr. Kevin M. O’Neill June 9, 2014 Page 24 Release, and the SEC concluded that “commenters did not broadly support any one risk measure, and the Commission acknowledged that investors have a wide range of ideas of what ‘risk’ means.”69 Following the issuance of the Risk Concept Release and after considering all comment letters received in response, the SEC proposed that a fund’s prospectus include a bar chart showing the fund’s annual returns for 10 calendar years, and noted that over 75 percent of individual investors responding to the Risk Concept Release favored a bar chart presentation of fund risks.70 The Commission subsequently adopted the bar chart requirement, which was intended to illustrate graphically the variability of a fund’s returns and thus provide investors with some idea of the risk of an investment in the fund.71 We believe that the SEC acted prudently in publishing a Risk Concept Release as a first step in the consideration of whether to adopt a risk metric: It carefully reviewed the comment letters it received; considered how mutual fund risk disclosure could be improved in light of those comments; subsequently proposed specific amendments to Form N-1A (including the bar chart proposal) and invited additional comment; and then adopted amendments to Form N-1A (adopting the bar chart, among other changes), with some modifications. In the event that the SEC concludes, as we have, that it should adopt a glide path much closer to that proposed in 2010 (i.e., an asset allocation glide path) rather than that outlined by the Committee (i.e., a risk-based glide path), then we would expect that its rulemaking could follow the normal course, and the SEC’s next release on the subject could contain finalized Rule amendments. If, however, the SEC decides to pursue a risk-based glide path requirement, either as a replacement for or supplement to the asset allocation glide path, then the SEC must first formally propose new specific Rule amendments and include in the proposing release a detailed regulatory impact analysis.72 Such an approach would be similar to the procedural approach taken following the 69 Id. See also 1998 Release. 70 See Registration Form Used by Open-End Management Investment Companies, SEC Release No. IC-22528 (Feb. 27, 1997), 62 FR 10898, 10904 (Mar. 10, 1997). 71 See 1998 Release, at 13922. 72 In the 2010 Proposal, the SEC asked whether target date fund marketing materials should be required to disclose a risk rating based on a scale or index that could be compared to other target date funds; if so, how the scale or index would be designed; and whether the scale or index should reflect only investment risk, or also longevity and/or inflation risk. Those inquiries were not a point of emphasis in the 2010 Proposal (they were not part of the proposed Rule amendments), and not surprisingly, the 2010 Proposal’s cost/benefit analysis section did not address the possibility of the final Rule amendments acting on them in some way. In the Release, the SEC noted that the comments received to its queries in the 2010 Proposal about risk ratings were “limited.” In response to these queries, the 2010 ICI Comment Letter stated that creating such a rating would be an enormous undertaking with questionable benefit significantly that was beyond the scope of the 2010 Proposal. We urged the Commission first to study comprehensively the utility of any such measure, and recommended that the Commission take into account the risk information already being provided to fund investors in the risk/return section of prospectuses. Mr. Kevin M. O’Neill June 9, 2014 Page 25 issuance of the Risk Concept Release and is crucial to ensuring that interested parties are provided with an adequate opportunity to fully address (and for the SEC staff to fully consider) any concerns associated with any risk-based indicator. The Recommendation and the Release are undeveloped on the actual risk-based measure that might ultimately be adopted, and therefore comments in response to the Release will necessarily be incomplete, broad, and somewhat speculative. Even informed investors and market participants that read the Recommendation and Release will have difficulty picturing exactly what a risk-based glide path illustration might look like, literally and figuratively—particularly as applied to a target date fund with a changing asset allocation. The SEC must issue a new proposal to address more thoroughly issues such as how liability concerns would be allayed if funds were required to provide future projections of volatility (e.g., we would expect such a release to address the extent to which the methodology for constructing the risk-based glide path would be standardized and, if any element of constructing the glide path involved some measure of discretion, how the SEC would ensure that actual fund performance that diverges from the mandatory projections made in good faith would not subject funds to lawsuits).73 Moreover, we believe that skipping a step in the rulemaking process by moving straight to finalized Rule amendments that incorporate a risk-based glide path and diverges from the requirements of federal law.74 The 2010 Proposal included a regulatory impact analysis and solicited comment on that analysis. The Commission’s decision to move forward with dramatic changes to the 2010 proposal contemplated in the Release would require a re-evaluation of that initial analysis, and would warrant allowing the public to reconsider and comment on the new analysis. V. Marketing Materials Used by Non-Mutual Fund Target Date Funds The financial crisis of 2008 prompted a number of policymakers, including the Commission and DOL, to focus on target date funds and express concern about whether retirement investors understood these products. As testimony at the joint Commission-DOL hearing showed, mutual fund 73 See supra, note 49. 74 See, e.g., Business Roundtable v. SEC, 647 F.3d 1144, 1146 (D.C. Cir. 2011) (vacating Exchange Act Rule 14a-11—the “proxy access rule”— because the SEC failed adequately to consider the rule’s effect upon efficiency, competition, and capital formation, as required by Section 3(f) of the Exchange Act and Section 2(c) of the Investment Company Act; the D.C. Circuit Court noted that the SEC “inconsistently and opportunistically framed the costs and benefits of the rule; failed adequately to quantify the certain costs or to explain why those costs could not be quantified; neglected to support its predictive judgments; contradicted itself; and failed to respond to substantial problems raised by commenters.”); Am. Equity Inv. Life Ins. Co. v. SEC, 613 F.3d 166, 177 (D.C. Cir. 2010)(vacating Securities Act Rule 151A—the “fixed indexed annuities” rule—because the SEC’s consideration of the effect of the rule on efficiency, competition, and capital formation, as required by Section 2(b) of the Securities Act, was arbitrary and capricious under the Administrative Procedure Act); and Chamber of Commerce of the United States v. SEC, 412 F.3d 133, 136 (D.C. Cir. 2005)(finding that the SEC had violated the Administrative Procedure Act by failing adequately to consider the costs that mutual funds would incur in order to comply with the conditions of certain new fund governance rules—namely, that funds wishing to rely on certain exemptive rules have fund boards with at least 75% independent directors and independent chairmen—and by failing adequately to consider a proposed alternative to the independent chairman condition). Mr. Kevin M. O’Neill June 9, 2014 Page 26 complexes are not the only entities offering target date funds to retirement plans. While many non- mutual fund target date funds used in plans are organized as bank collective trusts, they also can be offered in other forms, including as insurance company separate accounts and customized arrangements put together by consultants for plans. The Commission’s 2010 Proposal covers only one segment of the target date fund industry— mutual funds subject to the Commission’s jurisdiction under the Securities Act and the Investment Company Act. Bank collective trusts and insurance company separate accounts are exempt from securities law regulation if they meet the conditions of the exemptions in those laws. The result is that mutual funds used as investments in 401(k) plans or other defined contribution plans are subject to far more detailed substantive requirements and regulation of their disclosure and advertising than any other retirement investment providers. Plan participants who invest in mutual funds are well-served by this regulatory regime. To enhance the understanding of target date funds, non-mutual fund target date funds and arrangements must be subject to the same kinds of new disclosure rules as the Commission imposes on mutual fund target date funds. Only in this way will all retirement investors be assured of receiving the same basic information about these important retirement savings vehicles. Whenever these Rule amendments are finalized, we recommend that the DOL subject collective trust funds, insurance company separate accounts and other non-mutual fund arrangements to similar rules. We urge the Commission to continue working closely with the DOL on target date fund issues with the goal of coordinating the concurrent issuance of comparable target date fund advertising rules. ■ ■ ■ ■ ■ We appreciate your consideration of these comments. If you have any questions, please contact me at (202) 218-3563, David Abbey at (202) 326-5920, or Matthew Thornton at (202) 371-5406. Sincerely, /s/ Dorothy Donohue Dorothy Donohue Acting General Counsel cc: The Honorable Mary Jo White, Chair The Honorable Luis A. Aguilar, Commissioner The Honorable Daniel M. Gallagher, Commissioner The Honorable Kara M. Stein, Commissioner The Honorable Michael S. Piwowar, Commissioner Mr. Kevin M. O’Neill June 9, 2014 Page 27 Norm Champ, Director Susan Nash, Associate Director Division of Investment Management Securities and Exchange Commission Thomas M. Selman, Executive Vice President, Regulatory Policy and Legal Compliance Officer Thomas A. Pappas, Vice President, Advertising Regulation Joseph E. Price, Senior Vice President, Advertising Regulation/Corporate Financing Joseph P. Savage, Vice President and Counsel, Regulatory Policy Financial Industry Regulatory Authority, Inc. A-1 Appendix A Summary of Risk Ratings for Funds Used in Non-U.S. Jurisdictions 1. UCITS IV Funds and SRRI Since July 2012, the European Securities and Markets Authority (“ESMA”) has required that all UCITS IV funds include the Synthetic Risk and Reward Indicator (“SRRI”) number in their Key Investor Information Documents (“KIIDs”).75 Under the applicable guidelines, UCITS IV funds are divided into five main categories (market funds, life cycle funds, absolute return funds, total return funds and structured products) for this purpose, each with a different calculation methodology. A UCITS IV fund calculates its rating, which ranges from 1 (lowest risk/reward) to 7 (highest risk/reward), and this backward-looking rating is based on fluctuations (measured weekly) in its net asset value over the previous 5 years. If a fund’s performance history is less than 5 years, the methodology for the SRRI calculation is adjusted using the returns of a representative portfolio model, target asset mix or benchmark of the fund from the beginning of the same period, until the date of availability of the actual returns of the fund. A management company must compute a fund’s SRRI in compliance with ESMA requirements. The SRRI must be accompanied by a narrative section, briefly explaining how the indicator was calculated, and what other material risks are not captured by it (e.g., liquidity, interest rate, currency, etc.).76 2. Australian Super Funds and Their Standard Risk Measure The use of the Standard Risk Measure (“SRM”) was “strongly recommended” by the Australian Prudential Regulation Authority (“APRA”), Australian Securities and Investment Commission (“ASIC”), Financial Services Council (“FSC”), and Association of Superannuation Funds of Australia (“ASFA”) in 2011.77 Australian super funds78 are not required to provide investors with their SRM, 75 Since July 2012, KIIDs, the European analog to the summary prospectus used by many U.S. mutual funds, are a requirement of UCITS IV. The KIID provides a concise two-page overview of a UCITS fund and is written in plain language and in a standardized format. It is a pre-sales document replacing the simplified prospectus and is intended to explain to retail investors the key features of a UCITS fund and to enable easy comparisons among funds. The KIID must include: (1) investment objectives and policy; (2) SRRI; (3) costs and associated charges; (4) past performance; and (5) practical information about the UCITS. 76 See Committee of European Securities Regulators, CESR’s guidelines on the methodology for the calculation of the synthetic risk and reward indicator in the Key Investor Information Document (July 2010)(Ref. CESR/10-673), available at www.esma.europa.eu/system/files/10_673.pdf. 77 See Financial Services Council and Association of Superannuation Funds of Australia, Standard Risk Measure Guidance Paper For Trustees (July 2011), available at www.superannuation.asn.au/ArticleDocuments/116/FSC- ASFA_StandardRiskMeasures_July2011.pdf.aspx. 78 Superannuation is a means by which Australians accumulate funds for retirement. It is government supported and encouraged. Employers are required to pay a proportion of an employee’s salaries and wages into a super fund. Individuals are encouraged to invest additional monies into super funds to supplement their employer-sponsored retirement savings. A-2 although the vast majority of super funds have adopted the SRM as a means of disclosing risks to the investing public.79 The SRM, if adopted by a fund, must be disclosed in the fund’s Product Disclosure Statement (“PDS”).80 The SRM measures risk simply as the estimated number of negative annual returns that could be expected over any 20-year period. This model is forward-looking (but based on capital market assumptions that depend on historical data), and looks forward one year to calculate the probability of a negative return and then extrapolates this figure for the 20-year calculation. The SRM predicts the likelihood, not magnitude, of negative annual returns. Establishment of the methodology used to calculate the SRM has been left to the individual trustees of the funds, and trustees are required to develop their own set of capital market assumptions that take into consideration return, volatility, and correlations for the relevant asset classes. The investment product provider, based on internally generated assumptions, must calculate the fund’s SRM. SRM utilizes a 7-level classification system, with risk labels ranging from “Very Low” to “Very High.” In addition, the PDS must disclose that the SRM is not a complete assessment of risk. 3. Canadian Risk Disclosure Requirements On June 18, 2010, the Canadian Securities Administrators (“CSA”) published CSA Staff Notice 81-319, which outlined the point of sale disclosure framework for mutual funds that includes, among other things, the creation and distribution of a Fund Facts.81 Fund managers are required to provide risk ratings for mutual funds in the Fund Facts. The required risk rating is based on a 5-level scale ranging from “Low” to “High.” Currently,82 the risk rating is based on the methodology adopted Super funds operate as trusts with trustees responsible for operation of the fund as well as formulating an investment strategy. 79 See SuperRatings, Standard Risk Measures, An Overview of Standard Risk Measures in Practice (June 2013), available at www.superratings.com.au/media/mediarelease/standard. 80 The PDS is a summary of significant information about the super fund. The PDS includes or addresses: (1) general information about the super fund; (2) how the super fund works; (3) the benefits of investing in the super fund; (4) the risks of the super fund; (5) use of proceeds; (6) fees and costs; (7) tax information; (8) insurance options; and (9) how to open an account. This document is similar to the summary prospectus that many U.S. mutual funds use. 81 The Fund Facts, similar to the summary prospectus that many U.S mutual funds use, is the document provided to investors in connection with their purchases of fund shares. The Fund Facts discloses, in plain language and in not more than two double-sided pages, the key information important to investors, including past performance, risks and the costs of investing in a mutual fund. 82 On December 12, 2013, the CSA issued Notice 81-324 and Request for Comment—Proposed CSA Mutual Fund Risk Classification Methodology for Use in Fund Facts (“CSA Notice 81-324”). The proposed risk classification methodology outlined in this Notice would replace the current requirement that fund managers determine the fund’s risk rating based on a methodology of their choosing. The proposal, conceptually similar to the SRRI in Europe, would require fund managers to use a standardized methodology based on standard deviation to measure the fund’s volatility by looking backwards at the fund’s performance history on a monthly basis over a 10-year period. The proposed methodology contemplates moving from the current 5-category scale to 6 risk categories that range from “Low” to “Very High.” Assignment of a rating would depend on the fixed “standard deviation band” into which a fund falls. The comment period for this proposal closed on March 12, 2014, and to date, the CSA has not adopted final amendments. A-3 by the fund manager in its discretion. The majority of managers use volatility of past returns in determining their funds’ risk ratings.83 83 See CSA Notice 81-324. B-1 Appendix B Target Date Fund Glide Paths and Returns As noted in the body of our comment letter, the SEC’s Investor Advisory Committee (the “Committee”) argues that a glide path illustration based solely on asset allocation is unlikely to reliably capture potentially significant differences in fund risk levels, especially “where the asset classes are defined quite broadly, as they are in the Commission proposal.” The Committee notes that in 2008 returns ranged widely for funds nearing their target dates, notably 2010 target date funds.84 Evidence indicates that a broad asset allocation glide path (i.e., one showing a fund’s investments in stocks and bonds only) would have been quite informative in 2008 for investors in funds near their target dates. Figure A.1 presents results of a statistical analysis that seeks to explain the returns of individual target date funds based on what investors knew from data that were publicly available the prior year about the percentages of these funds’ assets that were held in stocks and in bonds.85 For example, the upper-left panel shows for 2010 target date funds the returns that investors could have expected in 2008 based on those funds’ 2007 broad asset allocations and had investors known the returns to be gained in the stock and bond markets in 2008.86 The red line in each figure is the “line of perfect fit.” In other words, if the broad asset allocation in 2007 is a good indicator of a fund’s return in 2008, the blue dotted point (representing the expected return for that fund) will lie exactly on top of the red line. The blue dots are widely dispersed, ranging from roughly -4 percent to -42 percent for 2010 target date funds, which indicates, as the Committee noted, that the returns on individual 2010 target date funds varied widely in 2008. Also, however, the blue dots are scattered in a fairly narrow band around the red line, which indicates that knowledge of a fund’s broad asset allocation would have allowed investors in 2007 to gauge rather accurately the 2008 returns of these funds, had they known in 2007 how stock and bond markets would behave the following year. Statistically speaking, the analysis explains most (72 percent) of the wide variation in the 2008 returns of these funds (as measured by the statistical concept of “R-squared”). 84 The Recommendation of the Investment Advisory Committee on Target Date Mutual Funds states that “among [target date] funds with a 2010 target date, losses in 2008 ranged from 9 percent to 41 percent … Much of those differences in risk can be explained by differences in the asset allocation models and glide paths used by different funds … But choices of assets within the various assets classes, inclusion of assets from outside the traditional asset classes, and other risk management practices can also have a significant impact on fund risk levels. A glide-path illustration based solely on asset allocation is therefore unlikely to reliably capture potentially significant differences in fund risk levels.” 85 The statistical analysis is a “regression” of a fund’s return on the portions of its assets in stocks and bonds the previous year. The estimated regression is used to explain fund returns that would have been expected for the various target date funds based on their broad asset allocations in the previous year. 86 In effect, the regression coefficients are proxies for the overall returns in the stock and bond markets. B-2 The upper-right panel shows that similar results hold for the 2008 returns of 2020 target date funds. Although information about each fund’s broad asset allocation is not quite as informative as for 2010 target date funds, the broad asset allocation is still quite informative. Returns of 2020 target date funds were not as dispersed as were those of 2010 target date funds, but variation in returns was substantial, ranging from about -20 percent to almost -45 percent. Again, however, the blue dots are generally scattered closely around the red line, indicating that a fund’s broad asset allocation as of 2007 would have been quite informative about the fund’s returns the subsequent year, had investors known how stock and bond markets would fare the following year. These findings hold for later years, as well. For example, the lower-left and lower-right panels in Figure A.1 present the same kind of analysis for the 2012 returns of funds with 2010 and 2020 target dates, respectively. Returns for these funds are considerably less dispersed than in 2008. As before, the blue dots remain tightly clustered around the red line, indicating that knowledge of a fund’s broad asset allocation would have allowed investors in 2011 to gauge rather accurately the 2012 returns of these funds, had they known in 2011 how stock and bond markets would behave the following year. From this analysis, two conclusions can be drawn. First, the figure highlights that to the extent investors in 2010 target date funds were surprised in 2008 by the performance of their funds, it most likely reflects an unawareness that funds near to their target retirement dates may still hold a significant portion of their assets in stocks or other securities that are more volatile than, say, U.S. government bonds. Differences in asset allocations among funds sharing the same target retirement date largely reflect the different views fund sponsors may have about how best to structure target date funds in order to meet investors’ financial goals. This, in turn, suggests the appropriateness and value of highlighting for investors differences in broad asset allocations of target date funds through a presentation in marketing materials of an asset allocation glide path. Second, the analysis highlights that any measure intended to highlight to investors the range of returns they might experience will be heavily dependent on actual market outcomes. In years when the stock market declines, those who have invested in target date funds with heavier allocations to stocks may experience unfavorable returns on their target date funds. Alternatively, in years when the bond market declines but the stock market advances, such as in 2013, investors in these same funds would experience favorable returns on their target dates funds. Thus, the perceptions that investors form of their target date funds will depend in large part on specific stock and bond market outcomes. This illustrates a difficulty that could arise from a risk-based glide path: presumably, a fund with a heavy allocation to fixed income securities would be seen as “less risky,” but in a year such as 2013 such a fund might nevertheless surprise its investors because of muted or even negative returns. B-3