ICI Response to Kauffman Foundation Paper, “Choking the Recovery: Why New Growth Companies Aren’t Going Public and Unrecognized Risks of Future Market Disruptions”

There’s much to think about in the recent paper from the Kauffman Foundation. The authors, who have vast knowledge of and experience with U.S. capital markets, have some good suggestions for strengthening our markets, as well as some ideas that merit further consideration. One problem that shines through the paper, however, is a fundamental misunderstanding of how exchange-traded funds (ETFs) operate. As a result, the paper levels several accusations against ETFs that are just not plausible. Let’s examine two.

One elementary misunderstanding that permeates the Kauffman paper rests in the concept of winding down—and up—an ETF. The paper suggests that an ETF may create systemic risk because, in the event of a selloff triggered by any type of market meltdown, the “rush to unwind” ETFs will aggravate a selloff, and that “some creators of ETFs may not be able to honor their obligations.” In the other direction, an ETF could be at risk of failure if it takes in a lot of cash to create shares, but its own demand for the underlying securities drives up the price of those securities, so the ETF does not have sufficient cash to track its index.

Both of these scenarios fail for one key reason: except in some very rare and unique circumstances, ETFs are created—and redeemed—in-kind. What does this mean? In short, an ETF itself will never participate in a “sell-off” or, on the other side, a bidding war, because ETFs generally do not buy and sell stocks. Rather, they exchange blocks of shares, usually 25,000 to 100,000 shares, to certain “authorized participants” (generally broker-dealers or other trading firms) for a proportional amount of the ETF’s underlying securities. Because of this process, an ETF would never need to sell securities in a falling market to provide a redeeming shareholder with cash. Rather, if an authorized participant wishes to redeem ETF shares, the fund takes in the shares and hands over a basket of securities. Because its obligations are limited to a proportional share of its securities, an ETF will always be able to meet them, even in a declining market.

Similarly, an ETF typically does not take in cash with an expectation of buying its component securities, but rather receives component securities in exchange for ETF shares. If the price of those securities is rising, an authorized participant may determine not to create ETF shares, but there is no risk that the value of existing shares will be diluted by an ETF being forced to buy in a rising market.

Another misunderstanding is evident in the Kauffman paper’s reprise of the theory—first raised in a widely excoriated report by Bogan Associates last September—that an ETF can collapse because of a high level of short interest. The theory was that if, due to short selling, there are more claims of shares outstanding than actual shares, and there were a high level of redemption requests, the ETF could only pay out some of the claims, leaving other shareholders holding the bag. The Kauffman paper takes this theory a step further, stating that in a fund with 100 percent short interest (that is, the same number of shares sold short as shares outstanding), 50 percent of the money invested in the fund is used to purchase the underlying securities, and the other 50 percent is lent to short sellers.

The flaw behind both of these assertions rests on the simple concept of possession, which is required to redeem an ETF share. An ETF share represents a fractional ownership of the assets in the ETF. The fraction is determined by the number of shares outstanding. The only person who has an actual claim on the fractional ownership is the one who is in possession of the shares. In the old days, this possession would have been represented by a stock certificate; now there are electronic analogs.

When an individual sells an ETF share short, he must locate shares available for borrowing and be prepared to borrow them to meet his short sale obligation. The individual who purchases the ETF shares from the borrower/short seller, and ultimately receives the electronic certificate of ownership, will have a claim on the ETF shares. The lender of the ETF shares may continue to view himself as owning the shares, since he continues to receive any economic benefits, such as dividends, from the shares. Until he retakes possession of the shares he lent, however, he does not have a claim on the fractional ownership of the ETF. When an ETF redeems shares (that is, pays out those claims), it requires a representation that the investor has possession of the shares (i.e., the electronic certificate of ownership). In funds with high levels of short interest and when there are high levels of redemptions, the ETF will typically require verification of that possession.

Going back to the theories: since the ETF should only honor those claims represented by an actual share (i.e., electronic certificate of ownership), and each share represents a proportional share of the fund, the full value of an investor’s money is indeed invested in the underlying securities. Further, the ETF should never be in a situation where it has more claims than assets to pay out, because during high levels of redemptions it should verify ownership before paying any claims. Bottom line: as long as funds are doing this—and it is in their best interest to do so—the level of short interest in an ETF is irrelevant.

Had the authors of the Kauffman paper better understood the basic mechanics, they would not have created these doomsday scenarios driven by ETFs. Much has been written about how ETFs work, but clearly more education is warranted. We look forward to working with regulators, academics, and the public to help explain how ETFs work. The recent market stresses have made everyone anxious, and we all have a vested interest in understanding them and strengthening our capital markets. Still, we urge the authors not to jump to conclusions about ETFs—or other instruments—without fully understanding them.