Elections and Financial Stability
Click to Print This Page
When Hillary Clinton unveiled her proposals to assure financial stability earlier this month, few people bothered to notice. Nonetheless, one central piece of legislature Clinton proposed to increase trust of the everyday investor in our financial markets, a “tax on the high-frequency trading that makes our markets less stable and less fair,” has potentially far-reaching consequences. The Clinton proposals elicited 550-something comments on Bloomberg, with criticism coming from the usual suspects. We are all acquainted with a contemporary practice whereby political operatives stuff “comments” sections with their talking points and completely flood the exchange between the readers.
Heretofore, I took notice only when Larry Harris, a highly respected expert on market microstructure and the author of a seminal textbook, Trading and Exchanges: Market Microstructure for Practitioners, suggested that “any tax on high frequency trading firms’ trading activity will increase the cost of trading for retail investors and for pensions that serve retirees.” In the preface to my 2009 book, Microstructure and Noise in Financial Markets: Rigorous and not-so rigorous results in market microstructure, I quoted his testimony to the US Congress as a model explanation of market microstructure to the laypeople.
For accuracy, I have to inform the reader that Larry Harris is a partner in a number of trading firms. Yet, given his stature, his prognosis cannot be simply dismissed as a right-wing rant.
Presumably, the currency of his argument is that HFT, along with dark pools, prevents unusual movements of asset prices caused by the purchase or unwinding of large numbers of shares. Assuming that this is true, one would conclude that the agents making block trades unrelated to the fundamental movement of prices—to accommodate exceptionally large contributions, or redemptions, or statutory requirements—are pension funds, mutual funds, and insurance companies. Why Larry Harris put “retail investors” into that category is a mystery to me—unless he implied that retail investors highly depend on ETFs—which also might suffer from the necessity to buy and sell securities in large blocks.
Yet, if inordinate dependence of the retail investors on ETFs is what Mr. Harris had in mind, this by itself can be a problem. Reliance on sophisticated structured products in search of short-term liquidity can hardly be prescribed to everyone after our recent experience with subprime mortgages.
Indeed, the timely accuracy with which asset prices reflect the expectations of investors concerning asset value is the reason for existence of exchanges in the first place. Yet, I observed, in my recent research that, in the short term, only 1% or so of stock price changes can be attributed to fundamental events: decisions of the Fed, manufacturing reports and such. While my methodology can be subject to criticism, I believe that this or a similar figure is quite robust with respect to the methodology because I have seen a small share of “fundamentally defined” trades/price corrections reproduced in a number of unrelated contexts. Domination of “technical trading” in formation of short-term volatility suggests that the role of exchanges as keepers of fundamental value is highly exaggerated.
The role of HFT as a provider of liquidity has recently been questioned as well. For instance, research by Fricke and Gerig demonstrated that the optimal speed of execution required for maximum liquidity is not microscopic but, by the estimates of the authors, amounts to 0.2-0.9 seconds. I also discuss optimum speed of execution in different contexts in my book (Lerner, 2009, quoted above, see sections 2.4, 7.3 and 7.4). There are credible practitioners stating that, while extra high speed of executions can be beneficial for equities, it may be detrimental for fixed income securities, especially U.S. Treasuries.
However, if the proliferation of HFT and dark pools is the problem for financial stability—I am not so sure though I have a much lower threshold for conviction that Prof. Harris—there are other possible ways to address this problem besides a financial transactions tax. One solution would be to require exchanges and trading platforms to batch all orders for a given time (say, one second) and to execute them simultaneously at the clearing price. Another would be to introduce random delay to orders, though this rule can be easily defeated by “order stuffing”, i.e. submitting artificially large number of orders with expectation that only a fraction of them will be executed. Alternatively, my solution would be to require exchanges to auction off the bandwidth beyond some reasonable threshold—for instance, 1 Hz/trade, and require the order submitters to pay a market-determined price to play fast and loose.
All in all, I think that a tax on HFT would not be as pernicious as its critics suggest, however, there could be better ways to address abuses, and the significance of this problem for financial stability is not clear. It is unfortunate, however, that these problems will be decided (or undecided) not by deliberation of experts but by political rhetoric on evening networks.
-Peter Lerner, PhD, MBA is a semi-retired financial researcher who lives in Ithaca, NY.