“A key structural vulnerability from asset management activities is the potential mismatch in open-ended funds between the liquidity of fund investments and daily redemption of fund units.” Financial Stability Board, January 2017
In the aftermath of Britain’s July 2016 vote to exit the European Union, six U.K. open-end property funds with nearly £15 billion in assets suspended redemptions. These funds had routinely engaged in an extreme version of liquidity transformation: offering investors the ability to convert their shares into cash daily on demand, while holding highly illiquid commercial properties. Fortunately, the overall sector was small, and its post-referendum disruption neither spilled over broadly to funds holding other assets, nor prompted a wave of fire sales that might have undermined the balance sheets of leveraged intermediaries. Nevertheless, the episode was of sufficient concern that the U.K. Financial Conduct Authority (FCA) is now reviewing its “regulatory approach to open-ended funds that invest in illiquid assets” (see here).
The FCA is not alone in its concerns. Other regulators have been looking closely at risks associated with the liquidity transformation performed by open-end funds. And, interest in the official sector has been accompanied by a wave of academic research on liquidity management in open-end funds that generally buttresses the regulators’ concerns. In this piece, we briefly highlight the work of the regulators, summarize the research, and finally reprise our proposal to convert open-end funds into exchange-traded funds (ETFs).
Starting with the United States, last October, after an extended period of study, proposal and industry comment, the U.S. Securities and Exchange Commission (SEC) imposed a limit on illiquid holdings of mutual funds and required fund managers to adopt “liquidity risk management programs” that include periodic assessments of liquidity risk and a manager-determined minimum for highly-liquid holdings. It also permitted “swing pricing”—an adjustment to net asset value (in this case, of up to 2 percent) designed to impose on shareholders engaging in purchases or redemptions some of the associated trading costs.
Similarly, last month, the Financial Stability Board (FSB) issued a set of recommendations aimed at containing stability risks arising from asset management activities, including those associated with “liquidity mismatch and redemption terms” (see here). In addition to greater transparency, strengthened liquidity risk management, and guidance on stress testing, the FSB also calls on regulators to provide fund managers tools like redemption fees or swing pricing that limit “first-mover advantage” and hence investor incentives to run in a period of market stress.
Before turning to the academic research, it is worth taking a quick look at the role that open-end funds play in the U.S. financial system. As recently as 1980, open-end funds represented only about 1 percent of domestic financial intermediation, compared to more than 40 percent by depository institutions. Today, those shares are about 15 percent and 20 percent, respectively. If anything, heightened regulation of leveraged intermediaries will encourage a further shift toward mutual funds and away from depositories.
Measured relative to the size of the economy, the rise of mutual fund intermediation is especially notable. Today, the value of all open-end funds taken together―including all bond, stock, and hybrid funds―exceeds 73 percent of GDP, up from a bit more than 2 percent in 1980. More recently, ETFs have grown even faster, up from only ½ percent in 2000 to nearly 13 percent today. Importantly, over the past 25 years, the mix of open-end funds has shifted notably toward less liquid instruments. In the chart below, we show the path of assets held in various intermediaries—traditional depository institutions (banks), open-end funds holding liquid or illiquid assets, and ETFs—since 1990. Relative to GDP, the holdings of “illiquid funds” have climbed from ½ percent in 1980 to nearly 30 percent today, while the “liquid” funds have risen from nearly 2 percent to about 43 percent.
Domestic Financial Intermediation by Type of Intermediary (Ratio of Assets to GDP), 1990-3Q 2016
This broad shift away from depositories and toward mutual funds has contributed to financial stability. The Global Financial Crisis of 2007-2009 arose primarily in the world of large, complex, opaque, and highly interconnected leveraged intermediaries. These included, for the most part, depositories, securities firms and, in a few cases, insurers and government-sponsored enterprises. Some of these institutions had on- and off-balance sheet leverage exceeding 30 to 1.
In contrast, U.S. mutual funds operating under the Investment Company Act of 1940 have long functioned with strict limits on leverage. They also typically disclose a great deal of information regarding their holdings. By limiting use of derivatives, SEC rules proposed at the end of 2015 would further limit effective leverage. Moreover, unlike insured depositories and government-sponsored enterprises, mutual funds (excluding MMMFs) have operated without the explicit or implicit government guarantees that prompted many intermediaries to take greater risk at taxpayers’ expense.
Nevertheless, some open-end funds do engage in extensive liquidity transformation. It is this bank-like activity that makes them vulnerable to shareholder runs and panics, creating financial stability risks even in the absence of leverage and other common indicators of systemic vulnerability. These risks become more prominent as activity shifts away from leveraged intermediaries. Moreover, the typical macro-prudential toolkit (such as capital requirements, resolution plans, and various restrictions on credit supply) does not address risks arising from unleveraged intermediation activities.
Recent academic research examines at least three mechanisms that could lead to disruptive exits from relatively illiquid funds, increasing systemic risk: (1) the first-mover advantage arising from redemption externalities; (2) the potential for herding; and (3) the way in which cash is managed.
On the first-mover advantage, Chen, Goldstein and Jiang (2010) argue that greater illiquidity of assets raises portfolio adjustment costs created by redemptions, increasing run risks. More recently, Goldstein, Jiang, and Ng (2016) conclude that, when compared with those holding equities or Treasuries, flows out of and into U.S. corporate bond funds exhibit greater sensitivity to bad performance than to good performance. This outflow sensitivity to bad performance rises with illiquidity both across funds (higher in those invested in less liquid corporates) and over time (higher in periods of reduced market liquidity).
Second, Feroli et al (2014) argue that the frequent investor focus on relative performance by fund managers can lead to herd-like behavior of asset managers. In the case of illiquid assets, this leads to momentum in returns and fund flows, with self-amplifying feedback.
The third mechanism arises from the way in which cash is managed. Perhaps surprisingly, these funds hold a considerable volume of cash-like assets. In a large sample of open-end managed bond and U.S. equity funds, Chernenko and Sunderam (2016) find that an average of 7 to 8 percent of the assets is held in the form of cash. Furthermore, they conclude that the more illiquid the assets, the more cash the fund is likely to hold. (Similarly, Jiang, Li and Wang (2016) report average cash holdings of 5 percent for a sample of corporate bond funds.)
Against this background, Morris, Shim and Shin (2017) show that managers of open-end bond funds exercise considerable discretion in how they adjust cash balances in the face of investor flows. Contrary to conventional theory, where cash balances serve as a buffer to potential outflows, they find that the usual practice is to hoard cash when redemptions materialize. This is especially the case for those funds holding relatively illiquid emerging market sovereigns or corporate debt. Such procyclical behavior, analogous to what we see with bank leverage, tends to amplify fire sale risk.
To illustrate the mechanics and implications, the following chart shows a simple bond fund example where cash is used to buffer an outflow (on the left) and one where cash is hoarded (on the right). In both cases, we start with a customer redeeming $100 of shares from the fund (the blue bar). In the left panel, the manager sells $90 of the asset (the red bar) and allows the cash balance to decline by $10 (the shaded bar), partly absorbing the selling pressure. In the right panel, the manager sells $110 of the asset, first paying the redemption and then building up the cash balance by $10. In practice, this is what Morris et al find: for managers of open-end funds holding the most illiquid assets, cash hoarding is about $10 for every $100 of redemptions.
Bond Fund Cash Management: Two Scenarios of Investor Flows and Changes in Fund Holdings
We should note that some of these results are quite new, so a consensus about cash management practices has yet to fully develop among researchers. For example, while Morris, Shim and Shin emphasize the nature of cash hoarding, Chernenko and Sunderam’s attention is on how managers of U.S. equity and corporate bond funds engage in cash management practices designed to cushion fund flows, especially for those funds with illiquid assets.
The bottom line is that we still have a lot to learn about the behavior of open-end funds and the risks they pose to the broader financial system. But, we think there at least two lessons that regulators and market participants could draw today. The first one is that illiquid funds do not function as mere pass-through mechanisms that immediately execute investor purchases and sales. As a result, there is a need to ensure that liquidity risk management practices inside of open-end funds are up to task.
The second is to recall our earlier proposal to enhance the resilience of the financial system by encouraging open-end funds holding relatively illiquid assets to convert to ETFs. The point is that the source of instability is liquidity transformation and on-demand redemption, something that open-end funds typically promise regardless of the state of the world. In contrast, when the underlying assets become illiquid, ETFs function like closed-end funds: there is no promise of redemption.
In contrast to many who worry that investors overestimate the liquidity of ETFs, we view their design as intrinsically limiting the first-mover advantage associated with open-end funds invested in illiquid assets. To quote from the FSB’s recent report, “As a result of using in-kind redemptions, the transaction costs associated with redemptions from an ETF are imposed on redeeming shareholders rather than the fund and its remaining shareholders” (see Annex 3 here). That looks like a promising path to pursue for asset managers, investors, regulators, and society.