Slow Running

How mutual funds can limit damage when a run appears imminent.
March 9, 2009
Print this page

In 2004, the Securities and Exchange Commission launched an investigation of the mutual fund Putnam Investments amid allegations of market-timing fraud. The estimated losses incurred by investors as a result of Putnam’s market-timing practices were about $4.4 million over seven years.

But it was the aftermath that really hurt: in the first three months of the SEC’s investigation, many of Putnam’s investors panicked, prompting massive redemptions of their holdings. One prominent expert’s report to the SEC estimated that investors who stuck with the fund ended up losing about $48 million.

Since then, Putnam has rebuilt its reputation and its holdings, but the 2004 run illustrates a classic coordination failure, one to which open-end funds, such as mutual funds and hedge funds, are particularly sensitive.

Unlike closed-end funds, open-ended funds allow investors to redeem their shares at the end of any trading day. That easy access to redemption can pose a problem, says Professor Wei Jiang. “To pay out to redeemers, the fund has to liquidate assets, which is costly. The remaining investors bear the cost, in the form of trading commissions, price impacts and deviation from the fund manager’s desired portfolio allocation.” Market observers had given little attention to the implication of this effect because, until recently, runs on open-ended funds had been rare.

Mutual fund performance isn’t persistent on its own: poor performance today by a fund doesn’t doom a fund to poor performance in the future, nor does good performance today guarantee the same tomorrow. “However,” says Jiang, “we do see that people tend to redeem from funds that have performed poorly, especially from funds that invest primarily in illiquid assets such as small-cap stocks.”

Too many such redemptions can create fragility in a fund. “Even if an investor thinks a fund’s poor performance is just bad luck,” Jiang says, “if that investor thinks the poor performance will prompt others to redeem, he is more likely to rush to redeem. And this self-fulfilling, amplifying effect can be much more prominent during a crisis.”

Jiang, working with Qi Chen of Duke University and Itay Goldstein of the University of Pennsylvania, examined Morningstar data on flows to mutual funds and the trading liquidity of underlying assets to learn more about the dynamics of runs on mutual funds. Their analysis reveals two key factors that influence the severity of self-fulfilling redemptions.

The first factor is how much it costs the fund, and remaining investors, when some investors choose to redeem their shares. For funds invested in highly liquid stocks, like S&P 500 stocks, the cost of liquidating assets to cover redemptions is likely to be only a few basis points, a price few investors remaining in a fund are likely to be concerned about. But when a fund’s underlying assets are highly illiquid, the transaction cost of covering the redemptions is likely to be much higher, and the risk of a “run” on the fund much higher. Remaining investors who anxiously watch assets become diluted may consider redeeming as a way to avoid future losses.

The second factor the analysis considers is investors’ beliefs about their fellow investors. Whether investors view their fellow investors as stable and likely to stay in a fund, or likely to leave a fund, can also contribute to the likelihood and intensity of a run. The researchers looked at who holds funds and found that institutional investors are less likely to chase performance, while retail investors tend to move money around far more frequently. As a result, the same institutional investor is more likely to redeem from a losing fund if the fund is open primarily to retail investors and is likely to stay if the fellow investors are also institutional.

Can mutual funds — and similar open-end funds — preempt such runs? The researchers’ analysis suggests that funds can adjust their cash holdings and impose some redemption restrictions — a less extreme version of the practice associated with closed-end funds. But both practices have some inherent limitations.

Cash holdings, while protecting against potentially costly redemptions on illiquid assets, don’t earn a return. Redemption restrictions, while they may postpone or slow down a potential run, are not usually welcome, particularly by unsophisticated investors, who may not understand the benefits this method offers. A fund must proceed carefully when instituting such restrictions, because it risks driving unhappy investors to seek out less restrictive investment funds.

In 2006, retail fund managers began to express interest in entering alternative investing fields (like real estate and private equity, which have traditionally been the exclusive territory of private funds open to institutions and high-net-worth individuals), and mutual fund families started talking about establishing mutual fund shares for their investors to pool money to invest in alternative funds.

But all this talk came at a time when funds were less subject to runs than they are now, and it’s likely fund managers weren’t considering the risk that coordination failures can carry for alternative investments. “The assets in these alternative funds are very illiquid, so a fund might have to sell a dollar’s worth of asset in order to raise 90 cents,” Jiang says. “Given the financial crisis, the entry of retail investors into those fields is unlikely to move forward any time soon.”

Wei Jiang is the Sidney Taurel Associate Professor of Business in the Finance and Economics Division at Columbia Business School.

Wei Jiang

Wei Jiang is Arthur F. Burns Professor of Free and Competitive Enterprise in the Finance and Economics Division, and the Vice Dean (for Curriculum and Instruction) at Columbia Business School.  She is also a Scholar-in-Residence at Columbia Law School, a Senior Fellow at the Program on Corporate Governance at Harvard Law School, and a Research Associate of the NBER—Law...

View full profilePersonal Website