Market Fragility and Interconnectedness in the Asset Management Industry
Scott W. Bauguess, Acting Director and Acting Chief Economist, DERA
New York, New York
June 20, 2017
Thank you, Greg [Strassberg], for that kind introduction.
Thanks also to Joanna Rejman, and all of the other conference organizers, for the invitation to speak here today, at this year’s Buy-Side Risk USA 2017, and for all of the preparation required to make an event like this happen. I’m glad to be back and, over the course of the next 20 minutes, I plan to provide an update to some of the thoughts I last shared in April 2016. Before I begin, let me give the standard disclaimer that the views I express today are my own and do not necessarily reflect the views of the Commission or its staff.
My prepared remarks this morning center on the topic of “Market Fragility and Interconnectedness in the Asset Management Industry.” The relationship between asset management activities and financial stability risks has been a frequent topic of discussion among financial market participants and regulators in recent years. This discussion has arisen in light of the significant growth in the asset management industry and the increased focus on financial stability risks in the aftermath of the financial crisis. My remarks today are intended to highlight the underlying economics of these issues to help provide a better understanding of their nature and significance. They are informed by a continuously emerging body of academic research, some of which is being performed by staff at the SEC, and all of which demonstrates the inherent complexities in thinking about potential contributions of asset management practices to market fragility concerns.
Theories of Financial Stability Risk in the Asset Management Industry
Following the financial crisis, an area of immediate focus for those concerned with financial stability risk in the asset management industry was on run risk. This was likely attributed to the market’s reaction to the Reserve Primary Fund breaking the buck at the height of the financial crisis and the related run on institutional prime money market funds. The SEC addressed these specific risks through a series of regulatory money market reforms in 2010 and 2014. However, the money market discussions – and the ways that mutual funds differ from money market funds – brought to the forefront how the asset management industry and its risks compared to those of the banking system.
Run risk is a generally well-understood concept in the context of a traditional bank. For example, when depositors are uncertain about the quality of their bank’s assets, and concerned about whether their savings are safe, they line up outside the front doors and wait their turn to take their money out. The sooner they get in line, the more likely that money will be waiting for them, creating the well-known, first-mover advantage that can lead to bank runs and cause a bank to fail, even when its underlying assets are of good quality.
Run risk in the asset management industry is more nuanced. Overly simplistic attempts to apply run-risk concerns to traditional funds generally ignore two key differences when compared to the banking industry. First, assets are managed for investors who intentionally take on investment risk; it is not a balance sheet business and generally operates without significant economic leverage. Second, investment funds offer more transparency about their asset holdings compared to banks, allowing investors to assess, first hand, their quality and their associated risk.
These differences do not mean that run risk doesn’t exist at funds. And while the money market fund reforms addressed run risk at funds widely believed to have contributed to market fragility as revealed during the financial crisis, other types of funds remain in the spotlight. For example, the Financial Stability Board (FSB) – a G20-initiated organization comprised of the global financial market and banking regulators – has since 2011 engaged in an annual monitoring exercise of certain non-bank financial intermediaries, including many investment funds. In their most recent report, the FSB stated that the global asset management industry accounts for 65% of market-based financial activity that could give rise to financial stability risks. Put differently, this amounts to $22 trillion out of the $34 trillion in assets that the FSB classifies under the shadow banking moniker.
So, the scope of the potential concern is large. And the pejorative nature of the term shadow banking underscores some of the potential misperceptions of the value and risks created through the market-based intermediation activity it is intended to describe. From one perspective, the tremendous growth in fund activity has given households around the world economic exposure to valuable investment opportunities that were previously out of their reach. And it comes to them with a transparent risk and reward profile that the opaque banking model, by its design, is unable to replicate. For these reasons, we should never lose sight of the role that the asset management industry plays in ensuring the vibrancy of our markets. It is to ensure the stability of this crucial financial sector that regulators – and indeed, all market participants – should remain aware of potential fragilities.
And it is important to understand how these risks are shifting. While some market observers may be weary of hearing the phrase “this time is different,” it is undeniable that market function has changed fundamentally since the financial crisis. I believe that merits a fresh view of market fragility risks. This includes identifying not only new risks, but also monitoring and being aware of the new contexts in which the old risks can emerge.
Market Participant Behaviors That Keep Regulators Up at Night
Among the most prominent areas of market change that have placed increased attention on potential risks in the asset management industry is the very rapid growth in open-end funds that invest in fixed income securities and other alternative investments. As of December 2016, there were $2.5 trillion of fund assets in these categories. This represents a 673 percent increase in AUM [assets under management] since December 2000. And it corresponds to an annual growth rate of approximately 14 percent, approximately three times greater than the growth in domestic equity fund assets over the same period.
This growth is notable because portfolios of fixed income securities are generally less liquid than those comprised of equity. And when subject to an adverse financial shock, they may be more susceptible to the risk of fire sales – the forced selling of assets at discounted prices. This is because open-end funds, regardless of the liquidity of their underlying assets, must offer investors daily redemption rights at end-of-day NAV [net asset value]. In times of market or fund stress, if redemptions arrive at a rate faster than fund managers can reasonably liquidate the underlying assets, they may be required to sell assets at prices below their intrinsic value. And the resulting losses from discounted selling can itself engender further selling pressure, thereby providing greater opportunity for prices to deviate from fundamentals.
On a large scale, fire sales could pose financial stability concerns. Of course, it is important to remember that, outside of money market funds, the U.S. asset management industry weathered the financial crisis well. There have not been any major disruptions to financial markets from investor redemption requests at funds. Nevertheless, under current market conditions, with strong investor demand for funds with less liquid assets, many researchers are now focused on evaluating this likelihood. Some studies are reassuring. Others are less so. In all cases, they provide important evidence and motivation for both fund managers and regulators to remain vigilant in assessing risks and potential impact of future market stability events.
One area of financial stability research focuses on the prevalence of investor herding and whether smart capital will step in when such behavior results in an imbalance in the supply and demand for securities. Investor herding occurs when the actions of a few are mimicked by many, but without a separate evaluation of the decision that triggered the initial actions. The concern during financial market stress is that the time required to perform a separate evaluation could exceed the amount of time available—if you see mass exit in a market, it could be individually rational for each investor to join the exit without additional information or evaluation, even if the collective action is irrational.
In the context of the asset management industry, for herding behavior to contribute to run risk in this way, two fundamental questions need to be answered. First: “Are there countercyclical investors who will step in with smart capital to arbitrage, and benefit from, the deviations in asset prices away from their fundamentals?” As Warren Buffett is famous for saying, “You want to be greedy when others are fearful.” But this smart capital frequently relies on funding from the same investors who may be redeeming fund shares. So, the second important question to answer is: “Where does the capital from redemptions go?” If the answer is: “under the mattress,” then there is reason to be fearful. But if investors put the money back into the financial market or in a bank, this provides for opportunistic arbitrageurs to profitably buy assets, potentially before they reach fire sale prices.
Many large investors in public capital markets have long-term investment horizons and a structural ability to weather short-term financial market pressures. In particular, insurance companies and pension funds face redemption pressures on the basis of events largely orthogonal to [independent of] macroeconomic events: natural disasters, accidents, retirement, and death. As a result, their customers do not necessarily have liquidity or consumption needs requiring a return of capital during times of financial market stress, and even if they do, their contractual agreements may not allow it. So they are not subject to the same funding constraints that other financial institutions face, particularly those that are levered. This institutional design affords flexibility for insurance companies and pension funds to make smart money decisions that rely on a long-term strategy or focus.
There is empirical evidence in support of this view. Recent research shows institutional herding is substantially higher in the corporate bond market relative to equity markets and is more pronounced on the selling side.  But this effect is less pronounced among insurance companies. And when it is evident, for example, as a result of forced sales on account of rating downgrades, there is some evidence that mutual funds and pension funds step in to take advantage of this market friction. This provides mixed, but encouraging news—while there is observed herding behavior by institutions in less liquid markets, there are also countercyclical effects that offset such behavior.
Whether offsetting capital is available in sufficient quantities during times of financial market stress is equally important to whether there are individual members or segments of the market willing to provide it. There is well-documented evidence that capital is slow moving, even when value enhancing arbitrage opportunities are available. One cited cause is that the requisite knowledge to profit is often separated from the capital required to do so. And when it is not, market participants who face losses and the prospect of investor redemptions may have only limited ability to place smart money against the prevailing winds out of fear that their capital will disappear. 
Interconnectedness and Liquidity Management
Another potential contributor to market fragility is interconnectedness of the asset management industry that comes from the tendency of fund managers to respond similarly to a market shock. The decision of one asset manager can appear to be in lockstep with another, even when there is no intention to coordinate these actions or herd. This may be most acute in how asset managers manage fund liquidity, particularly during times of stress. Research shows that fund managers, during the financial crisis, when faced with increasing illiquid portfolios, engaged in trading activity that shifted holdings toward more liquid assets. And for funds that faced greater outflow pressure, managers disproportionally sold the most liquid assets.
Both of these behaviors are consistent with other research that documents cash hoarding by mutual funds in anticipation of redemptions, particularly during times of stress. The concern, like with herding, is that a common stress event could trigger simultaneous sales by many fund managers. While such sales are prudent from an individual fund’s perspective, the funds’ common pursuit for increased liquidity could, in turn, exacerbate downward selling pressure. Anticipating the precise impact of this pressure can be difficult for individual fund managers whose liquidity management experience during normal times may not apply to stress times.
The SEC has been cognizant of this concern. Rules adopted in 2016 require that registered mutual funds and ETFs establish a liquidity risk management program and allow flexibility for fund managers and their boards to establish these programs in ways that best suit their circumstances. But even sophisticated models designed to account for the negative externalities from other fund manager actions can be difficult to estimate. This is because significant stress events are generally infrequent, and some potential stress scenarios are so rare that they have yet to occur. And, as a result, there is little or no data with which to calibrate risk models to account for extreme events.
Among potentially emerging issues in the asset management industry is the rise in rules-based investing executed by fixed algorithms. Robo-advising is a potential example of this. Another example is smart beta ETFs. As most of you already know, these [purportedly passive] funds weigh investments on the basis of certain factors. In particular, rather than track an index by value- or equal- weighting the constituents, a smart beta portfolio may weigh constituents by their earnings, volatility, dividends, or similar metrics.
These rules are designed to capture specific, well-defined, and repeatable strategies aimed to earn better risk-adjusted returns. The concept is appealing. Sophisticated strategies can be implemented at relatively low costs. The financial stability concern, which at this point is largely theoretical, is that these rules could unintentionally engender mechanical herding. In particular, if decision rules are derived from changes in market conditions, and the resulting portfolio reallocations then affect market conditions, then a feedback loop of self-fulfilling behaviors without human judgment is generated.
At what point could this contribute to market fragility? That is a difficult question to answer. But looking at history, the 1987 market crash offers an extreme example of how this type of behavior could manifest with deleterious consequences. In particular, retrospective analyses show that program trading designed to limit losses in a declining market – offered to investors as portfolio insurance – exacerbated selling pressure and provided a disincentive for institutional investors to step in as buyers. These predetermined actions occurred without incorporating the situational awareness of human judgement.
Whether smart beta could generate similar risks and consequences depends on the answers to two questions. First: “Are the underlying rules designed in a way that could generate negative feedback actions?” For example, a rule that replaces high volatility stocks with low volatility stocks could, during a period of financial market stress, exacerbate the volatility of certain securities. Second: “If, indeed, actions facilitate a negative feedback, are they in sufficient quantity to amplify the market effects?”
By one estimate, the current footprint of smart beta funds includes a $347 billion global ETF market. While that is a large number, it remains a relatively small portion of global ETF assets. But if the popularity and growth of smart beta funds continues at a rapid rate, this could change. So this is an area that warrants continued monitoring for potential market risks.
Monitoring Markets for Emerging Risks
Turning to what these and other potentially persistent risks mean for regulators, I want to first reiterate that the same financial market practices that may give rise to market fragility concerns are generally rooted in financial intermediation activities that provide tremendous benefits to consumers, households, and the public in general. So, any proposals to address these concerns by “de-risking” markets are likely to come at a very high cost if unintended consequences are not carefully considered.
Second, it is important to note that human judgement is irreplaceable in responding swiftly to market developments and risks. Regulations are necessarily rooted in historical experiences and are not always designed to address even well-known risks that merge in an unexpected context. And while I’ve alluded previously to SEC rules that are designed to ameliorate certain persistent risks – including those that emerge in novel ways – regulators must remain ready to respond nimbly.
As a result, diligent market monitoring remains an essential function of regulators. In particular, risks become systemic only if we allow them to emerge and grow without abatement. In the asset management industry, this means continuous assessment of the known risks just discussed. And when market disruptions occur, it is important to quickly assess the affected market participants and potential spillover risks into other areas of the markets.
At the Commission, this occurs through constant collaboration across divisions and offices. Within DERA, there is an army of Ph.D. financial economists and other highly trained quants that work hand-in-glove with examiners in the SEC’s Office of Compliance, Inspections and Examinations [OCIE] and staff in the Division of Investment Management to develop models and metrics to help detect and respond to emerging market events.
When a stress event occurs at a fund, or in the fund industry, Commission staff respond by assessing the potential spillover of investor reaction to the news. The speed of the reaction is calibrated by the level of concern over potential run risk. One way to make this determination is to identify the closest peers by portfolio composition. The uniqueness and similarities of the stressed fund or funds, relative to the rest of the industry, inform staff on which other funds may also be at risk.
Another determinant of potential spillover risk is looking at whether markets underlying the stressed fund or funds are affected by the event. If so, they might provide a conduit of risk between funds, even if portfolios are otherwise dissimilar. Asset holdings that are thinly traded or traded over the counter pose greater concern. Quantitative analysis of dealer participation, frequency of trades, size of trades, and price impacts of trades are some of the metrics used to assess these spillover concerns.
All of these analyses, many of them relying on sophisticated analytical methods, can be performed in real time as an event unfolds. However, there are limitations to analyses using portfolio holdings; they often rely on stale information, as much as 3 months old. Thus, an assessment of potential spillover effects assumes that there has not been significant rebalancing during the intervening period. This can be a risky assumption, and it underscores the importance of the recently adopted rule requiring funds to report monthly portfolio holdings information on Form N-PORT, starting in June 2018.
In all of these activities, it is important to remember that in a global economy, conduits of risk know no geographical or institutional boundaries. Because of this, it is critical that regulators coordinate their knowledge and potential responses to emerging risks and market events. International organizations like the FSB and the International Organization of Securities Commissions (IOSCO) offer important communication channels to share information. The FSB’s annual shadow banking monitoring exercise is one example of a cross-jurisdictional collaboration to assess the scope of potential financial intermediation risks outside of the banking sector. It provides prudential and securities markets regulators an opportunity to actively discuss and report assessments on the potential impact of asset management on market fragility.
So where does all this discussion leave us?
Let me begin my concluding remarks by saying that it appears that we have already reached the point where our memory of the financial crisis is fading. The days are gone when someone asks me to recount what it was like at the SEC during 2008 and 2009. Nor does anyone ask me to provide a retrospective review of lessons learned. And when I speak to students studying finance and accounting, very few of them have adult experiences related to the events of those years. And when I explain the financial market risks from structured finance, poor use and preparation of mark-to-market accounting, the use of credit default swaps in the OTC market, and the efficacy of housing policy, it’s purely a history lesson. Those were the mistakes of others. Financial stability to them is already a textbook topic.
But traces of these past experiences provide excellent lessons for the future, and some of the risks we shouldn’t forget because they are timeless. I’m confident saying now, that leverage, and the use of derivatives that create synthetic leverage, will exacerbate the next significant financial market disruption, if it isn’t the cause of it.
Many of the market reforms that the SEC has put in place, some of which I’ve discussed today, address the potential contribution of asset management activities to market fragility concerns. Central to these reforms is the collection of new information put in place by our Division of Investment Management. That is because no rule we contemplate today can prevent future market stress events, but the timely collection of current market information and practices will enable both regulators and the market participants they monitor to more clearly assess and respond to emerging and ongoing risks in the industry using accurate and reliable data.
So, let me finish by reiterating a comment that I made earlier: It is important for the SEC and other market regulators, both domestic and foreign, to remain vigilant in assessing these risks and their potential impact. There is no substitute for diligent market monitoring. There is no substitute for good regulatory data in that pursuit. And there is no entity better positioned than a market regulator to assess how the collective decisions of market participants can manifest into significant market events.
Thank you so much for your time today.
 The Securities and Exchange Commission, as a matter of policy, disclaims responsibility for any private publication or statement by any of its employees. The views expressed herein are those of the author and do not necessarily reflect the views of the Commission or of the author’s colleagues on the staff of the Commission. I would like to thank Ralph Bien-Amie, Vanessa Countryman, Giulio Girardi, Christina McGlosson-Wilson, Sarah ten Siethoff, and Christof Stahel for their extraordinary help and comments.
 See Release No. IC-29132, 2010, Money Market Fund Reform, https://www.sec.gov/rules/final/2010/ic-29132.pdf; and Release No. 33-9616, 2014, Money Market Fund Reform; Amendments to Form PF, https://www.sec.gov/rules/final/2014/33-9616.pdf.
 The FSB report’s narrow measure of financial assets that could potentially contribute to market stability risks does not include those intermediated by entities in China. Chinese entities are otherwise estimated to account for 8% of the $92 trillion in non-bank assets.
 An update of statistics calculated in P. Hanouna, J. Novak, T. Riley, and C. Stahel, 2015, Liquidity and Flows of U.S. Mutual Funds, DERA White Paper: https://www.sec.gov/dera/staff-papers/white-papers/liquidity-white-paper-09-2015.pdf. The fixed income class is comprised of foreign, general, and corporate bonds. U.S. government bonds are excluded from that category.
 For a discussion of how collectively irrational decisions can occur in the bank run context, see, e.g., D. Diamond and P. Dybvig, 1983, Bank Runs, Deposit Insurance, and Liquidity, The Journal of Political Economy, Vol. 91, No. 3 pp. 401-419.
 See, also, remarks by Andrew Haldane, Bank of England, on the need for patient capital. http://www.bankofengland.co.uk/publications/Documents/speeches/2014/speech723.pdf.
 F. Cai, S. Han, D. Li, and Y. Li, 2016, Institutional Herding and Its Price Impact: Evidence from the Corporate Bond Market, Board of Governors of the Federal Reserve System, Working Paper.
 Ellul, Andrew, Chotibhak Jotikasthira, and Lundblad, Christian T., 2011, Regulatory Pressure and Fire Sales in the Corporate Bond Market, Journal of Financial Economics 101, 596 – 620; and Chiang, Chia-Chun, and Greg Niehaus, 2016, Investment Herding by Life Insurers and its Impact on Bond Prices, University of South Carolina, Working Paper.
 A. Sheleifer and R. Vishny, 1997, The Limits of Arbitrage, The Journal of Finance, Vol. 52, No. 1, pp. 35-55.
 See, e.g., M. Mitchell, L. Pedersen, and T. Pulvino, 2007, Slow Moving Capital, American Economic Review, Vol. 97, No. 2, pp. 215-220. http://pubs.aeaweb.org/doi/pdfplus/10.1257/aer.97.2.215; and D. Duffie, 2010, Presidential Address: Asset Price Dynamics with Slow-Moving Capital, The Journal of Finance, Vol. 65, No. 4. http://www.darrellduffie.com/uploads/pubs/DuffieAFAPresidentialAddress2010.pdf.
 The median fund moved its holding to more liquid assets during the financial crisis. See Figure 10, P. Hanouna, J. Novak, T. Riley, and C. Stahel, 2015, Liquidity and Flows of U.S. Mutual Funds, DERA White Paper. https://www.sec.gov/dera/staff-papers/white-papers/liquidity-white-paper-09-2015.pdf.
 Funds with outflows greater than the median outflow experience a decrease in portfolio liquidity, consistent with selling their most liquid assets in order to meet redemptions. See Table 18, P. Hanouna, J. Novak, T. Riley, and C. Stahel, 2015, Liquidity and Flows of U.S. Mutual Funds, DERA White Paper. https://www.sec.gov/dera/staff-papers/white-papers/liquidity-white-paper-09-2015.pdf.
 S. Morris, I. Shim, and H. Shin, 2017, Redemption Risk and Cash Hoarding by Asset Managers, Bank of International Settlement, Working Paper; and G. Girardi, C. Stahel, and Y. Wu, 2017, Cash Management and Extreme Liquidity Demand of Mutual Funds, SEC Working Paper. https://www.sec.gov/files/DERA_WP_Girardi-Stahel-Wu_Cash%20Management%20and%20Extreme%20Liquidity%20Demand.pdf.
 Securities and Exchange Commission Report, 1988, The October 1987 Market Break. See also M. Carlson, 2007, A Brief History of the 1987 Stock Market Crash with a Discussion of the Federal Reserve Response. https://www.federalreserve.gov/pubs/feds/2007/200713/200713pap.pdf.
 Sizing the Global ‘Smart Beta’ ETF Market, October 21, 2016, http://www.nasdaq.com/article/sizing-the-global-smart-beta-etf-market-cm696510.
 “2,000% rise in new money allocated to smart-beta funds,” Financial Times, May 14, 2017. https://www.ft.com/content/5c5960d0-3668-11e7-99bd-13beb0903fa3.