Market Makers and Flash Crashes

"Blue Horseshoe loves Anacott Steel."
--Bud Fox (Wall Street)

Previously we discussed the role of increasingly fragmented exchange structure in precipitating flash crashes. Fragmentation alone does not adequately explain the phenomenon because modern arbitragers can quickly eliminate gross mispricings that might occur on a particular exchange.

Other factors, such as the evolution of market makers alongside changes in exchange profit status and order flow offer more explanatory power. Market makers are employed by exchanges to take the other side of trades within a reasonable, often pre-defined spread. The smaller the spread between the bid and ask price for a security, and the more volume that can be done in that spread, the more 'liquid' that security is.

Exchanges pay market makers to provide liquidity in order to attract customers who may be hesitant to do business on the exchange unless they are confident that trades can be executed near prevailing quoted prices. In return for providing liquidity, market makers are compensated by the exchange in the form of information and trade execution privileges as well as rebates for shares traded at quoted bid/ask prices.

In their more concentrated days, exchanges were run as non-profit mutual organizations. The exchanges were funded by the institutions they served. They were not treated as profit centers. Because retail and institutional investors dominated the order flow, exchanges were motivated to employ market makers who kept spreads narrow for the retail and institutional flow.

Today's fragmented exchanges are run as for-profit entities. No longer funded by institutions on a non-profit basis, modern exchanges must seek sources of revenue. Primarily, these revenues take the form of fees from trading customers.

But today's customers are no longer retail and institutional investors. Instead, order flow is dominated by firms engaging in high frequency trading (HFT). HFT, also known as algorithmic or 'algo' trading, is primarily the domain of hedge funds and proprietary trading desks (a.k.a. 'prop desks').

HFT is short term and momentum-based. Liquidity is not necessarily a friend of the algos. In fact, algos may seek to create temporarily illiquid situations in order to scalp profits created by wider spreads.

Because today's exchanges depend on HFT customers for profits, and those HFT customers may not value liquidity, the role of market makers in this arrangement becomes unclear. Seemingly, exchanges have less need for 'old school' market makers since primary order flow is no longer liquidity dependent. Absent incentive to keep spreads tight, market makers are less likely to take the other side of trades that they aren't being compensated for.

Moreover, market makers often run their own prop desks and engage in HFT themselves. Because they are granted privileged views of order flow, market makers can insert this info into their algorithms and perhaps gain advantage over non-market-maker HFTs. Accusations of 'front running' are increasingly common, where market makers are seen as executing trades ahead of other exchange customers. Last week's AP tweet flash crash, for example, has been said to have been caused by market maker prop desks getting a head start on other algos.

The sheer volume done by front running market makers, who both a) are no longer governed by strict liquidity mandates by the exchanges and b) might operate on multiple exchanges, make it difficult for arbitragers to quickly narrow gross mispricings--thereby leaving the system more vulnerable to the flash crash phenomenon.

Some people, such as the Santelli Exchange guest in the second video here, appear to want a return to the old days of concentrated, mutually owned market structure, perhaps with added government oversight. However, such structure carries moral hazard that invites excessive risk-taking under the pretense that a 'liquidity mandate' provides a bail out mechanism for bad trades. Moreover, regulation would create entry barriers to entrepreneurs with prospective game-changing innovations, thereby shielding incumbent marketplaces from competitive pressure.

Rick Santelli, on the other hand, proposes a free market solution. If the current market structure is not attractive to particular customers, then those customers don't have to participate. This is already occurring. Retail and institutional investors comprise smaller and smaller fractions of order flow.

If (big if) regulations and other entry barriers do not impair the development of alternative exchanges that could serve these customers better, then supply will follow demand.