Back to Portfolio for the Future™

When Exchanges are Complements, not Competitors

One regular theme of coverage and analysis that involves the listed equities, including analysis of the pursuit of alpha, is that many of the users of these markets claim, often quite obstreperously, that the exchanges charge more than makes sense and some restructuring of the industry is in order. The exchanges are doing well, so they push back against such critics.

The “Streetwise Professor,” also known as Craig Pirrong of the Bauer College of Business, University of Houston, added a new perspective to this debate recently when he argued on his blog that the exchanges actually can charge more even than a monopolist could, and indeed that as the market for markets fragments, their take increases rather than decreases.

This happens, he contends, because in an important respect exchanges aren’t competitors with one another. They complement one another.

Squeezing Brass Manufacturers

The problem of complementary pricing is an old one in economic literature. In a manufacturing context, it goes back to an 1838 book by Augustin Cournot. Cournot’s example was the manufacturing of brass, a sought-after alloy of copper and zinc.

Assuming the demand for brass is sufficient to keep the manufacturers in business, and assuming the amount of competition in the markets involved is imperfect, how much can the suppliers of zinc and copper squeeze out of the deal? A lot. After all, if the zinc miner increases its price, the copper miner is not incentivized to decrease its price so that the brass company’s total outlay for raw materials stays even. The copper miner doesn’t care that the brass company is prosperous, just as long as it stays in business. Indeed, the copper miner’s incentive may be in the opposite direction, to match the zinc miner’s increases with its own. That’s the problem of complementary pricing.

Pirrong suggests, then, that exchanges are complements to one another, not competitors or substitutes. Consider the high-frequency trading hedge fund that arbitrages on the basis of small and fleeting difference in prices in NYSE and BATS, when both exchanges list the same company. Our HFT firm’s profit is brass. NYSE is the copper. BATS is the zinc. The HFT firm needs “near immediate and simultaneous access” to both in order to make its alloy.

Although execution services among exchanges are in competition, the data services are complements. The HFT guys need to have available to them a complete set of exchange data.

This isn’t a problem for the HFTs exclusively. Pirrong also mentions the situation of a buyside institution that has developed a “clever order routing strategy” to reduce its execution costs. It may lose the value of that efficiency because the exchanges through whom it routes are now complements.

Blame Reg NMS

In general, the execution services on the US exchanges are in fact in competition, but the data services are complements to one another. The pricing that this allows, which can “border on the extortionate” in Pirrong’s words, is an illustration of the law of unintended consequences. Virtually every government law, regulation, or policy has unintended consequences that will outweigh and/or threaten the intended benefits.

In this case, the regulation at issue was Reg NMS, the SEC’s rule established in 2005 to foster competition and decentralization among exchanges. Reg NMS included an order protection rule, which requires traders to transact on the venue offering the best price. The revised national market system introduced in this way helped fragment the market, enhancing the efforts to find arbitrage plays and clever routing through them, and enhancing the Pokemon-like desire to “catch them all.”

This worked. It achieved its intended purpose. It made execution services more competitive. But it also backfired. It had the unintended effect of creating or exacerbating the complements problem regarding the data.

What is the solution? Is there a better way to structure markets? Pirrong is not definitive about this, but he suggests that property rights in data may be malleable. Exchanges don’t necessarily have to be regarded as the exclusive owners of their share price data. They do own it in some sense, but these rights may be attenuated in some manner in order to “split the baby” and solve the problem of exchanges as complementary suppliers.

The attenuation of rights need not itself be a regulation of policy. It could be achieved through a “Coasean bargain” between the users and the exchanges.