Securities lending programs became increasingly popular during the last decade among many investment companies (mutual funds, ETF's, pensions, endowments). Long-only funds, pensions, and endowments were attracted to the incremental income from lending securities which otherwise remained invested? The demand from certain hedge funds and other speculators with shorting selling strategies made such programs all the more worthwhile. In addition, custodians of the securities also got a chance to earn additional fees for providing the service.
As with many other investment trends in the last decade, the music for securities lending abruptly stopped playing around the bankruptcy of Lehman Brothers. Interestingly, the reason for losses was not specifically related to the securities which were lent. Except in the case of securities held in accounts at Lehman Brothers, securities generally could be recalled by the lender. Instead, investment companies which lent their securities suffered losses from supposedly safe short-term securities using the cash collateral from borrowers. Regrettably, the securities lending transaction added counterparty risk and collateral price risk in exchange for a fee which probably did not fully compensate the lender for these risks.
The web is filled with ample explanations of how these lending programs worked, and the media diligently covered the losses on reinvested collateral (and also popularized the SIV acronym). The immediate purpose of this posting is to provide a simple graphical perspective on securities lending activity among registered investment companies.
Using SEC filings from primarily mutual funds and ETF's, we can get a 50,000-foot view of securities lending activity. Starting in 2007, funds pulled back their level of lending activity in order to re-assess the risks and protect against further unforeseen losses on reinvested collateral. The interesting observation from the following graphs is the apparent recovery in lending activity.
Please note two important caveats concerning the data behind these graphs.
1. Investment companies disclosed their policy to lend securities, specifically whether lending was allowed and whether it occurred during each reporting period. However, the data does not reflect exactly how much of the securities were actually loaned out. Of course, if an investment company had a policy to not allow securities lending or did not engage in the practice, then we assume no securities were lent.
2. The data has a few months lag. The graphs cover a period through June 2009, but the sample during the second quarter of 2009 gets thin and leaves open the possibility of material restatement as more data arrives. To gauge the sample size, refer to the dollar amount of sampled portfolios is displayed underneath the months along the horizontal axis.
Starting from December 2003, the following graphs (separately for equity-dominated portfolios and debt-dominated portfolios) show the overall mix of securities across the following three categories:
1. Fund policy did not permit lending out securities during the reporting period.
2. Fund policy did permit lending out securities but was not engaged in lending any during the reporting period.
3. Fund policy did permit lending out securities and was engaged in lending some amount of them during the reporting period.
Here are three graphs for equity funds, looking at the trend according to total assets (AUM), average AUM per fund, and number of funds.
The green bars are intentionally placed underneath the yellow bars. The height of the combined green and yellow bars corresponds roughly to the capacity for lending activity, so hence the green bars are a gauge for how much this capacity was utilized.
The return of securities lending does not appear to be consistent across the population. While lending utilization based on total AUM has been trending upwards since January 2009 (first graph), the number of funds participating in this rebound has been trending downwards (second graph), albeit the pace of decline in 2009 is slower than in 2008. This divergent trends implies that funds lending securities are larger on average than previously. Furthermore, the third graphs shows that smaller funds which can lend securities do not exercise the privilege (yellow), when compared to larger funds (green).
Here are the same three graphs for bond funds.
For bond funds, the trends are consistent across the board. In terms of total AUM and number of funds, securities lending utilization has trended upwards during 2009 to historic highs. When looking at the population based on the average AUM, those funds lending securities appear to be increasingly smaller than those which do not authorize the practice. Given the challenges in managing and monitoring a securities lending program, one might have expected larger bond funds to be more engaged in the practice. However, smaller bond funds may be relying on their custodian banks or third parties to manage the programs, hence not impacted by internal resource constraints.
Finally, one obvious and expected outcome from this analysis: equity funds have consistently engaged in securities lending much more than bond funds. Hedge funds which borrow securities probably engage in long/short trading strategies, but if trading in bonds becomes more electronic or
Check back in the future for updates of these graphs and possibly other breakdowns, such as passive vs active strategy, custodian specific, and fund complex. Feel free to post or email other suggestions for assessing these trends.
Further reading:
Wall Street Journal: Is Your Fund Pawning Shares at Your Expense?
Wall Street Journal: Getting a Fair Share from ETFs
The Motley Fool: The Risky Business of Securities Lending
