Demonizing the financial industry has proven to be a political winner. Exposure of the evil banks was followed by Michael Lewis's exposé of the markets in Flash Boys. The markets are rigged, we are told, byhigh-frequency traders (HFTs) who are a bunch of tech-savvy parasites.1 The banks have been thoroughly chastened by Washington politicians and now it is the turn of the markets with more to come in the way of fines, regulations, and punitive taxation in the form of a financial transaction tax.
Financial transaction taxes would be politically popular. They would disrupt the HFT gravy train while imposing a minimal burden on "legitimate" users of the markets and raising bundles of cash.
Unfortunately, there is a lot wrong with this tidy assessment. HFT is badly misunderstood even by prominent economists, and it is attacked with about a dozen wrong but convincing-sounding accusations. The fact is that the world of finance going forward will be ever more dependent on a HFT infrastructure, and a new tax on the industry isn’t likely to cure what ails us.
The Role of High-Frequency Trading in the Marketplace
The early days of HFT involved a lot of very simple relationships, such as buying IBM on one exchange and selling it on another. But now much of that low hanging fruit has been picked; the simple pure speed plays are being replaced by a variety of firms employing HFT software to monitor and trade relationships central to their businesses.
It is important to refute the idea that the blinding speeds serve little purpose except to beat out other HFTs, and for nefarious ends. Speed makes it possible for ETFs and other derivatives to be accurately priced against their underlying securities, raw materials against their end products, and interest rates relative to each other across the globe. Most exchange traded instruments relate in some way to others, and through computer algorithms and extreme speed those relationships can be made consistent, making for efficient risk transfer and shared liquidity.
Businesses which previously had to take unwanted risks on materials they needed or sold can find growing numbers of liquid risk transfer vehicles. The various exchanges have been launching a steady stream of new instruments to satisfy these needs. In the past, such new contracts stood little chance of becoming viable, but with today's increased linking they can flourish. Eventually, instruments all around the world (at least in those countries permitting HFTs to operate profitably) will be linked. Different individual securities might be monitored as part of a dozen potential combinations by a dozen different firms. The result will be a slew of bids and offers for even thinly traded instruments.
Speed also makes for much more efficient market making, which means a considerable saving to market users. A market maker uses various cues to try to center his quotes on where the market is. High speed algorithms let a market maker monitor a dozen or more parameters that help him determine where to post his bid/ask and how wide to make it. It also enables him to resume his valuable activities sooner in cases where he has had to pull back due to chaotic conditions.
The need for speed becomes especially clear when a single completed transaction requires multiple purchases and sales. What happens if one or two pieces are missed? I call this the "busted play risk." Take for example an oil "crack spread," buying three crude oil contracts against selling two gasolines and one heating oil. What if your machine failed to sell the two gasolines? Now you have a busted play with dozens of possible actions to consider. Busted plays took years off my life when I was in the business — the alarm goes off, the machine tries a few pre-programmed emergency actions and then dumps the mess on the human trader. A tiny increment in speed, just a few milliseconds in some cases, can make a huge difference in the number of costly busted plays per day.
The Attack on HFTs — and Potential Consequences
Much criticism of HFTs has centered around the huge number of orders entered and canceled. This is supposedly clear evidence of waste and chicanery. But, anyone policing a relationship by bidding a certain number of points above or below something else will cancel and enter a new order every time the lead item changes—– possibly 100,000 times in a day.
The HFT vigilantes also have their sights on taxing order entries, cancelations, and orders not left in for a minimum time. This would be devastating to any HFT, but particularly to those HFTs losing out to faster or more sophisticated rivals. These "losers" provide two absolutely invaluable services: 1) they force their faster rivals to outbid them, which is a huge benefit to other market participants, and 2) they are primary drivers of the expanding infrastructure of linked markets since they are always desperately searching for new relationships to trade. These marginal firms would be gone in a flash with order taxes. That would leave the winners to back off on their bids and make dramatically bigger profits at the expense of market users (if the "winners" weren't also driven out).
Any HFTs that did stick around would probably switch to what I call "invisible following mode." Instead of actively entering bids and offers, they simply watch until someone else shows a bid or offer the HFT likes, and then they grab it. This would turn HFTs into liquidity takers rather than liquidity providers. It would doom many thinly traded markets and be a big setback to others. The markets would lose the liquidity and valuable information they derive from the steady flow of visible HFT orders.
Financial transaction taxes belong to a very destructive class of taxes, ones that reach deep into the guts of economic value creation, producing major distortions and inefficiencies.
Our markets would be severely handicapped without their increasingly important HFT infrastructure, and that might lead to the markets themselves exiting. Modern markets are quite mobile. Twenty years ago a move by the NY Stock Exchange to another country would have been unthinkable. Thousands of people would have to relocate, and there would be huge communication, legal system, and real estate headaches. Today the NYSE is mostly a bunch of computers and software owned by Intercontinental Exchange (ICE,) which also owns exchanges in Canada, Singapore, and London, among other places. ICE already has the necessary local contacts in those places, knows the regulatory environments, the legal systems, and the communications and facilities particulars.
HFT will become central to the entire world financial industry, and a healthy HFT infrastructure in our markets is an absolute necessity if the US is to maintain its preeminence in world finance. Transaction taxes could destroy HFT here, a desirable outcome to many of the proponents, but one with major unintended consequences. Right now we are at the center of HFT development and deployment, and it would be a shame to sacrifice this important edge for cheap political gain, economic shortsightedness, and a windfall tax mirage.
John Chapman is a retired commodity trader, having headed his own small trading firm for over thirty years.