Securities Lending as an indicator of fund underperformance

From the desk of Roy Zimmerhansl,

Practice Lead, Pierpoint Financial

Could it be that securities lending activity is a tip-off for anticipated future fund performance? That is the premise of a forthcoming study from Derek Horstmeyer, Professor of Finance at George Mason University. An old friend (although he might just say ‘friend’) sent through a link to a Wall Street Journal article a few days ago with the following title:

“Your Fund Manager Is Lending Out Your Holdings. Should You Be Worried?”

Guaranteed to get my attention, and of course, that is the objective for all headlines. A few lines later, Dr Horstmeyer suggests the answer is “a resounding yes”.

I have been publicly asking for academic evidence that refutes the various studies that demonstrate markets with short-selling bans have wider spreads and reduced liquidity, and by definition, covered short selling requires securities lending. The first time I asked for such evidence was about 10 years ago in my original blog series “Stocklending Today – an industry under siege” in the wake of several academic studies concluding that markets with short-selling bans have wider spreads and reduced liquidity.

Could this analysis be what I have been searching for? A rigorously researched assessment that demonstrates securities lending is a negative for markets. Will this challenge my long-held belief that properly conducted securities lending has a beneficial impact driving my career-long near-obsession with communicating that message to as broad an audience as possible?

The article was silent as to whether the study had been published or still being finalised, so I reached out to Dr Horstmeyer, who responded just a few hours later. The study will be released through normal channels in the coming weeks, and I look forward to reading it.

Most people will not have had the experience of writing an article for a journal – you try to get your thoughts and story across in relatively few words, concisely covering the full set of issues. The typical article will be half the size of a blog post such as this one. That’s hard to do. Mark Twain is often cited as having said:

“I didn’t have time to write a short letter, so I wrote a long one instead.”

Indeed.

So be clear that the following questions, comments, and criticisms may in part be due to the limited article space. Nevertheless, these are the key points I anticipate exploring with the author and/or having addressed on the release of the document.

Dr Horstmeyer and his research assistant cite the full list of actively managed equity mutual funds, separated into their specialisation: US growth, US value, US large-cap, international and emerging markets. For each fund, they looked at the average dollar value of the portfolio on loan during the year and split the funds into two groups: those with greater than or less than 1% of the assets on loan. It’s not clear whether they included funds that don’t lend at all, and it doesn’t’ disclose the sample size nor give any context of how much of the US mutual fund industry lends.

Disclosure. The article opens with:

“Unannounced to their investors, mutual-fund managers will often lend the shares they hold to short-sellers …”.

As I understand it, there is a requirement for US mutual funds to disclose securities lending activity either in the prospectus or in the Statement of Additional Information. Of course, securities lending income can be found in the financial statements. And the most obvious point is that information must be available as to what individual funds do, otherwise, how would the study have been possible?

Active Management. The study found that active fund managers who lend out more than 1% of their holdings on average during the year underperform other mutual-fund managers by an average of 0.62% a year. It is important to note that the article only refers to active managers rather than index-trackers/passive funds. Active managers are those that select securities in line with the fund mandate and whose sole objective is to outperform the benchmark index. This group of managers searches for stocks to buy that will outperform the rest of the market, so is tasked with seeking unconventional choices – after all, if most fund managers also select the same portfolio, there is no hope of outperformance. The study doesn’t isolate the impact of securities lending itself, for example, by comparing the performance of two passive funds tracking the same index, one that lends, one that doesn’t. On a like for like basis, securities lending must contribute to outperformance.

Timeline. It is also unclear how underperformance and lending was determined and over which periods the assessment was conducted. The WSJ story carries graphs of the 10-year annualized returns by specialisation, the 5-year and the volatility. Broadly speaking, these show outperformance for those funds with under 1% loaned out and lower volatility. However, there is no detail on how the analysis was compiled – was the 1% dividing line a one-off snapshot or consistently applied over several years with the performance tracked accordingly? If more than a snapshot, how did they categorise funds that had >1% one year followed by <1% the next? This will no doubt be revealed on publication and may provide insights.

The active manager with the greater lending activity on the portfolio may be a contrarian looking for oversold stocks or be a value manager searching for unloved stocks – after all what is less loved than a stock with a lot of shorts? In either case, it may take some time for the eventual ‘winner’ to be determined. Think of Tesla shares, the poster child for short selling debate. Ask long investors who have held for many years while also lending out their shares – did they mind the volatility, the fight against the short-sellers and the collection at times of enormous fees? Absent the timeline information, it is hard to say whether the poorly performing managers have eventually been successful with their selections.

Performance. The study’s findings appear to suggest that lending is “bad” because it leads to poorer performance. If that were the case, the two samples would be ‘funds that lend’ vs ‘funds that don’t lend’, and unfortunately, that isn’t how I interpret the study’s construction. Instead, the conclusions are drawn based on the funds that have ‘a lot on loan’ compared to those with fewer securities on loan as outlined in the following article extract:

“…if the manager is lending out a good amount of the fund’s holdings, this means there is a lot of demand by other investors to bet against the exact holdings the fund manager has in the mutual fund.”

All this statement says is that if investors want to borrow and short more of Fund A’s portfolio rather than Fund B, then Fund A is more likely to underperform in the ensuing period. Short sellers expect/hope the stocks they have shorted drop in price, and if a lot of short-sellers agree about the same stock, then the likelihood of that stock underperforming is greater than if fewer short sellers believe that. I’m not certain, but there may also be an element of self-selection in the results. Managers that are poor stock pickers will underperform, partly because they hold stocks that short sellers expect to drop or drop further, and I think people would anticipate such an outcome over the short term.

If Fund A contains many stocks with a lot of shorting activity (aka borrowing demand), then the active manager of the mutual fund is facing a lot of short-selling headwinds. Whether Fund A lends is immaterial - these headwinds are the issue – the disagreement between the long investors and the short-sellers on the stocks in the portfolio. And with over $25 trillion US of securities available for loan on any given day, the impact of a lending investor withdrawing from lending is essentially zero. Inevitably there will be a handful of specific securities where a large shareholder withdrawing from lending would impact trading activity but these are exceedingly few and far between.

Maximum Limits. There is also a reference that 2% of funds that focus on US equities have an average of more than 20% of the assets on loan and that this is “close to SEC guidelines” maximum. For the record, the SEC figure is 33 1/3%.

Volatility. The article also describes increased volatility for funds that have more than 1% on loan, but that is not due to securities lending, rather it is the effect of the dispute between long and short investors – the greater the range of market views on a given stock, the greater the price movement is likely to be. I am not clear how the article can close with the suggestion that investors may wish to avoid funds that lend as the focus of the study is the difference between funds that lend a lot or lend a little, not whether funds lend or don’t lend. Perhaps there is more on this in the research itself.

Summary

I appreciate the contact from Dr Hosrtmeyer and I hope that I can access the full study at some point so that I can correct any errors in this blog, fill in the gaps that I have identified, and gain valuable insights into the study’s findings.

Unfortunately, based on the WSJ article alone, I don’t think this study will satisfy my hunt for new evidence challenging the value of short selling and securities lending. As that Irish band says so eloquently, I still haven't found what I'm looking for. .