What David Swensen Didn’t Write About Stock Trading Spreads

Bad Eggs My initial take on Michael Lewis's 2014 best-seller, "Flash Boys," was respectful but skeptical.

Lewis, of course, is the leading chronicler of financial shenanigans, possessing deep subject knowledge along with his ability to spin a yarn. Clearly, Lewis had done his homework; he understood that book's arcane topic, high-frequency traders. Using data that other investors do not receive, and technology that they do not possess, HFTs flit like mosquitoes, alighting to sip from other parties' trades, then darting away before being seen.

(Star Trek fans may picture a different simile: the invaders of "Wink of an Eye". Worry not for Deela—she will evade that phaser blast.)

Lewis told that story well, and I believed him. My doubts came when he assessed the damage. That HFTs are parasitic is without question. Also unquestioned is the culpability of the stock exchanges, which, by accepting payments from HFTs in exchange for exclusive data feeds, sell out the interests of their other customers. The picture is not pretty.

Less-Bad Outcomes However, aside from principles, it didn't appear that HFTs harmed retail investors. HFTs extract much less than a penny per share when making their trades, meaning that they depend on very high volume to generate their profits. This makes them a danger to other giant traders, but not to the rest of us. An investor with $1,000,000 held directly in stocks and portfolio turnover of 50% will shed about $10 per year to HFTs. As scandals go, that's not much.

(The math: $500,000 of stock sold per year, another $500,000 purchased, for a total of $1 million. Assuming a share price of $50—as good as any assumption—that means 20,000 shares traded. With Deutsche Bank estimating that HFT profits have dropped in recent years to one-twentieth of a cent per share, those 20,000 shares would yield $10 for the HFT.)

Indeed, the individual investor may well have ended up ahead on the deal. Although it's not possible to measure HFTs' effect on stock-market spreads directly, we do know that since HFTs began to surface, the average trading spread for the U.S. stock market has dropped significantly. Many if not most institutional investors agree with Vanguard CEO Bill McNabb's perspective from a few years back, which is that HFTs have probably played their part in that success.

Thus, while I shared Lewis’s qualms, I couldn’t muster any real dislike for HFTs. Yes, they are are unprincipled. In that, they can join a very large crowd. If I became outraged each time people acted in self-interest, without regard to others, I would have a very long enemies’ list. The key point for me was that Lewis failed to make the case as to how investors were harmed, and therefore I shrugged off his book as being entertaining and informative, but ultimately not terribly important.

The New Case Two weeks ago, Yale chief investment officer David Swensen challenged this belief. In "Wall Street Profits by Putting Investors in the Slow Lane," published in The New York Times, Swensen (and co-author Jonathan Macey) argued—well, appeared to argue, more on that shortly—that banning HFT activities from stock exchanges leads to reduced trading spreads.

In other words, retail investors do pay more because HFTs exist, which makes HFTs a financial menace as well as morally deficient.

Swensen’s argument was as follows:

1. Every stock exchange except one accepts "kickbacks" from favored traders, presumably HFTs. (Strangely, the article does not directly mention HFTs in this section, but they are surely the intended target.)
2. The sole exception is a new exchange called IEX, which opened last summer, and which features a "speed bump that prevents high-frequency traders from front-running ordinary investors."
3. Since its launch, IEX has "crushed" rival stock exchanges by offering by far the lowest spreads on stock trades. This May, for example, IEX offered the lowest effective spreads on 497 of the S&P 500's 505 stocks. (Hey, the Wilshire 5000 has but 3,600 stocks; by comparison, the S&P index is a model of accuracy.)

That rocked me. Lewis had profiled IEX in “Flash Boys,” portraying its founder as the angel among stock-exchange devils, but his morality play didn’t emphasize lower trading spreads. It’s one thing to oppose HFTs because they don’t seem to be good people; it’s quite another to oppose them so that everyday investors get significantly better prices. If that is the train, then I’m hopping aboard.

Correlation, but Not Causation However, that train does not exist.

True, IEX is the only stock exchange that bans HFT activities. And yes, IEX has the lowest effective spreads, as Swensen states. However, the first item does not lead to the second. The reason that IEX’s reported spreads beat those of its competitors is not because of what occurs within IEX’s conventional trading facility. It is instead because IEX offers the best prices to institutions conducting “dark-pool” trades. Those dark-pool transactions count toward the overall statistics that Swensen cites, but they don’t directly help you or me.

That is, IEX does two things differently than other stock exchanges. One, it bans latency arbitrage trades. Two, it serves dark-pool traders very well. The first involves HFTs but does not improve the exchange's trading spreads. The second does not involve HFTs but does improve IEX's spreads. That, of course, is not the impression given by the Times article.

For an example of how conventional, “lit-pool” trading results may differ from dark-pool figures, let’s return to May 2017. In that month, cited by Swensen as demonstrating IEX’s across-the-board excellence, IEX posted the sole best bid or offer price for

The upshot, writes one observer: "So while it is true that 'effective spreads' are lower at IEX, that is the direct result of [institutional investors] sending orders to them 'blind', and, in particular, from orders sent to them probing for midpoint fills (as compared with orders that are willing to cross the spread.) The data makes this extremely clear."

(That writer, to be sure, may well be an interested party in this debate, as may another who offers a similar take. Yale's Swensen is also an interested party; a topic this abstruse has nothing but interested parties. However, for background, I spoke with the chief equities trader at one of the world's largest money managers, and he fully supported that claim: IEX's results look good because of how it serves dark-pool investors, not for how it treats retail shareholders.)

Half Measures Swensen (and Lewis for that matter) has done right for investors. Many years back, before the merits of low costs were fully understood, he instructed fund buyers to pinch their pennies, avoid chasing performance, and consider the merits of indexing. Those are praiseworthy things. As is IEX, which is indeed attempting to run a clean business, in a less-than-clean industry.

However, half a truth in service of a greater good remains only half a truth—and half a lie. There is a good case to make for IEX. But allowing readers to draw the wrong conclusion would not seem to be the way to make it.

John Rekenthaler has been researching the fund industry since 1988. He is now a columnist for Morningstar.com and a member of Morningstar's investment research department. John is quick to point out that while Morningstar typically agrees with the views of the Rekenthaler Report, his views are his own.