It’s no surprise that short-selling carries heavy weight on Wall Street. But next to nothing has been known about the agent lenders from whom short-sellers borrow their shares—until now.
One major difference between professional equity traders and the rest of us is that the pros don’t view shares as merely an asset to buy or sell. They routinely borrow other people’s securities to short-sell them as part of arbitrage, hedging, and other complex trading activities.
Since short sellers’ activities may reflect new information about a stock, their activity could be interpreted as making the market more efficient—i.e. pushing material information into the light of day, where it can help markets set accurate prices. But that could also depend on whether the short-sellers’ counterparty—stock lenders—are greasing the wheels of the system or adding friction by placing high premiums on in-demand stocks.
Until recently, researchers were prevented from rigorously exploring the activities of stock lenders. “Aggregate levels of short interest per stock, which tells you how many shares have been lent out and shorted, have been available for a long time,” says Alexander Philipov, area chair of finance at Costello College of Business at George Mason University.
“However, we have historically not been able to observe how these shares were lent, and at what fees. The lending market has been quite opaque.”
That all changed with the introduction in 2006 of HIS Markit’s Securities Finance database, which contains detailed information on lending activity. Now, with almost 20 years of transaction data accessible for analysis, finance scholars can finally draw meaningful conclusions about this critical dimension of the equities market.
Philipov’s recently published paper in Management Science is an early entrant in this endeavor. Using econometric techniques, his research decomposes lending fees into an intrinsic fee and a premium or discount, which may reveal lenders’ strategic behavior. Lenders may choose to charge a premium when demand is inelastic, or offer a discount when demand is elastic, to increase lending revenues.
His co-authors were Gergana Jostova and Brian Henderson of George Washington University.
The researchers analyzed lending activity for the period July 2006-March 2013 (covering 4,558 firms in all), alongside relevant data about share price, available shares, and other stock fundamentals.
Thanks to the Markit database, they had a large enough basis for comparison to determine that lending fee discounts were associated with higher demand elasticity, while premiums were linked to inelastic demand. In addition, increases in the magnitude of both premiums and discounts were associated with revenue increases, which would imply strategic lender behavior.
“We saw that the big agent lenders seem to be very informed about the market, because the fees they charge seem to be getting them extra revenues,” Philipov says. “They seem to know for which stocks a decrease in fees would draw high additional demand.”
By the same token, lenders seemed skilled at identifying a select group of stocks for which borrowers would readily pay a high premium. Presumably, the high demand for these stocks stemmed from insider information available to short-sellers. Stocks lent out at a high premium fared the worst, in terms of future share price, among the entire sample, indicating that borrowers had negative knowledge about those companies not yet incorporated in their prices.
Consistently across the sample, the most discounted and highest-premium stocks were the biggest revenue generators for lenders—highlighting the importance of strategic fee-setting for their business model.
As with any other market, competition was a moderating factor. Lenders facing less competition, because they owned a large proportion of available shares of a security, could charge higher premiums on average. In a more crowded field, lenders successfully court demand through deeper discounts. Either way, their fee-setting patterns could best be explained with reference to revenue, rather than availability of shares per se.
Philipov cautions that because these findings are based on aggregated activity, it’s not possible to find out exactly what may have been going on between the players. “We are not able to look at the individual actions of individual lenders, and actually the lending fee database is trying to preserve that anonymity.”
But based on the general patterns revealed in his research, Philipov surmises that equity lending functions more as an efficient facilitator than a source of friction. “With lenders being strategic—and especially if they see that people want to short, and demand is elastic—they offer discounts, and that makes trading easier and more transmissible. Only for a very, very small corner of the market where they can extract big profits, they use high premiums that might prevent people from trading.”