Links: HFT

There has been a fair amount of discussion of High Frequency Trading (HFT) as a result of Michael Lewis' new book. For example, see the discussion on Leo Kolivakis' blog on the HFT debate. I was never directly involved with HFT as part of my job, but I see the effects in the trading of my personal retail trading account. I think the views of "Kid Dynamite" (who used to work on Wall Street as an equities trader) is the closest to my position:  HFT may be causing problems for big existing institutions, but it's largely a non-issue for retail investors (but speculators may be unhappy). But the rise of HFT means that you should never use market orders, only limit orders.

In the post High Frequency Trading: The Little Guy is the Big Winner, Kid Dynamite argues:

I think it’s incredibly dangerous to lead retail investors to believe that they will ever be on a “level” playing field.   In reality, we should be telling them that there will always be someone smarter in the market.  There will always be someone faster than you in the market,  there will always be someone who gets the information before you get it.  There will always be someone who reacts before you can react.   But the great thing about current market structure is that those discrepancies have been reduced from tens of minutes to milliseconds:  the playing field is more level than it has ever been, and more importantly: those who want/need faster access can get it!   That was most certainly not the case in previous market structure – there was a closed monopoly system (NYSE specialists).   The current market structure has democratized the role that specialists filled formerly.  Despite all the HFT hysteria, I think that democratization is a phenomenal thing.

In one of his linked posts, he raises one key point: "Don't lose faith in markets, lose faith in market orders." With the rise of HFT, it is extremely dangerous to use "market orders" - you always need to set a limit to which you will buy and sell. With the advent of flash crashes, it is possible for a $40 share to go to $0.01 in a millisecond, which means you cannot be willing to sell at any price.

Once you make that adaptation, transaction costs for retail investors are smaller than they used to be, and there are better products available (ETF's).

On the other hand, retail speculators, e.g. day traders, are probably getting clobbered. But I failed to see how day trading made any economic sense in the first place. Any investment manager has to be able to answer the question: what is my edge? Even a buy-and-hold retail investor has an edge - they are taking exposure to risk premia, and they have the capacity to hold the position almost indefinitely. That is an important advantage. An inability to hold positions cost many sophisticated investors dearly during the financial crisis.

But it is hard to see how someone sitting at home watching financial television, using a retail trading engine and closing positions at the end of the day has any edge relative to other investors. The strategy sort-of worked during the Tech bubble, but that's not saying too much. There is no reason to believe that the financial markets are a magic money machine that spits out cash every day.

Social Uselessness

It is clear that the investment in ultra-fast HFT infrastructure is socially useless. But in its defence, it is one of the few industries actually willing to make capital investments, and pay good salaries (albeit to a select few). In any event, in a services economy, pretty well all economic activity is properly viewed as useless. How much has society invested into broadband communications so that people can watch videos of kittens riding Roombas(tm)? (Factoid: a Google search for "video kitten roomba" returned 190,000 results.)

And HFT investment is not the only culprit within finance. Gold mining is highly ecologically damaging, and we already have huge amounts of the stuff already refined and sitting uselessly in vaults (that need very expensive security). Banning retail investment in gold would probably be much better for society than banning HFT. But imagine the howling if some lawmaker proposed that!

A sales tax on equity transactions would probably end the HFT infrastructure investment madness. For example, a 1 cent per share tax would force market makers to widen bid/offer spreads to at least 3 cents, or possibly even 5 cents. This would limit transaction activity, and the need to build out infrastructure to support that activity. But it is predictable that equity trading amongst institutions would migrate to platforms that allow the tax to be avoided (e.g., equity over-the-counter derivatives). I am not an expert on the legal aspects of fiduciary duty, but my understanding is that buy side portfolio managers have a duty to find the lowest cost legal means of transacting. The end result is that sell side financial product structuring specialists will be the focus of expansion instead of HFT firms. Frankly, I think it is probably better to invest in technological infrastructure than the "regulatory arbitrage" infrastructure.

If the only objective is to "protect the little guy", a much simpler approach is to ban market orders, and force limit orders to not go through the best bid or offer by too large a percentage. This will help prevent unprepared retail investors from seriously blowing up their portfolios, but without taking sides in the turf struggles between institutional investors.

(c) Brian Romanchuk 2014