Could a small financial transactions tax improve the health of financial markets by rendering unprofitable a lot of the high-frequency trading strategies that have been creating phantom liquidity (and, many argue, real havoc) in so many of the trading markets of late?
Mark Buchanan, a physicist who clearly believes in blurring the lines among academic/scientific disciplines, has looked carefully at this question. Buchanan is the author of Forecast (2014), a look at “what physics, meteorology, and the natural sciences tell us about economics.” He has also been writing on HFT-related issues for Bull Market, a collective blog for innovative work on such matters.
By the way, this is as good a time as any to mention that “high-frequency trading” has become an unfortunate catch-all term. It is inapt in that it suggests speed as such is the problem, and dooms us to listen to the same story about Rothschilds and pigeons every time the subject comes up. The problems with HFT and algo trading as it is currently exercised are: that it incorporates gamesmanship that can make IPOs hazardous, and this effect alone has dried up the IPO market beneath a very large capitalization threshold; that it encourages de facto front-running; that it necessitates a needlessly expensive technological arms race. The other problems follow from those on this short list. Still, the term “HFT” has become generally accepted and will very likely stay with us.
Buchanan’s first suggestion, in a January 25th Bull Market post, was that markets might simply change the nature of the trading they host, moving from continuous to discrete transactions. Another way of putting this is that they might restructure themselves as batch auctions held every one-hundredth of a second. Markets, he said, would “serve real economic functions much better if they ran a bit slower.”
Comment and Reply
But in a comment on that article, James Kwak, an associate professor at the University of Connecticut, raised the related issue of a financial transaction tax. Surely, if FTT slows down speculation, it would be most likely to do so precisely at the expense of the sort of speculative effort that has to be repeated in millisecond cycles to produce its alpha … right? And that would make the issue of changing market structure moot. It’s better to use a single stroke of a hammer than to create a Rube Goldberg machine that culminates in a hammer stroke only if all the gears on the machine work properly … isn’t it?
So on January 31, Buchanan wrote again on the issue of HFT, this time as a full-dress response to Kwak’s point. His conclusion is … that he can’t reach a conclusion, beyond that there is “research going both ways.”
Frank Westerhoff, of the Department of Economics at the University of Bamberg, has done a lot of detailed work simulating market dynamics within various market structures, and Westerhoff has reached a butterfly’s-wing type of conclusion. Small details in the implementation of such a tax have big consequences for how the affected markets will play out thereafter. In a market in which orders are handled by a liquidity-provider, that is, a dealer, the tax seems to have beneficial consequences. If the market employs no dealer, rather using only simple automatic order-marching mechanisms, then the consequences of the same tax turn malign.
There is the danger that a central planner will learn of Westerhoff’s work and draw exactly the wrong conclusion. A planner might say, “Aha! Now that we know this, we can simply tax only the dealer-mediated markets. Problem solved.”
But that is point-missing: the issue of dealers-as-liquidity-providers is only one of many butterflies fluttering about in the atmosphere. Outcomes could also depend on the mix of participants in a particular exchange or pool, the depth of the markets, and perhaps factors no one has even thought of yet, much less tested for. The only real conclusion is that no one has any good reason to draw any conclusion about the consequences of an FTT for over-all market function.
My own inference here would go a good deal further than Buchanan’s does. I would submit that the best way to go about counter-acting the baleful consequences of HFT is to withdraw the components of the system that have made HFT dangerous in the first place. Decimalization, combined with the implementation of the NBBO mandate, threw a monkey wrench into modern markets. A range of systemic malfunctions, including those generally grouped under the terms “high-frequency trading,” followed. Let’s just remove the wrench, and see if the malfunctions reverse themselves.