I strongly recommend to read the whole article, but the main points are:
- HFT trading creates lots of volume, but more volume is not prima facie better (unless you are collecting per-trade commissions on that volume, which is why Wall Street likes HFT).
- HFT trade systems “win” (aka profit) – but who loses (who pays)? Wall Street lobbyist like to argue that HFT is victimless, but if there was no profit in HFT why would anyone do it? It sounds too good to be true – and it is. We all pay for HFT trading – via exchange fees and false information about liquidity.
- Why don’t the exchanges lower their transaction fees for everyone, instead of marking up each trade and rebating a fraction of the mark-up to just the popular clique? HFT firms and stock exchanges essentially gouge the public on every trade, and split the spoils between themselves.
- "HFT adds no meaningful value to the public. It is a poorly veiled way for a privileged few to benefit from exchange transaction fees".
High frequency trading (HFT) has been in the news a lot lately, as the SEC tries to determine what additional regulations (if any) are required – and many in the public view, rightly or wrongly, that HFT is just another way Wall Street firms “screw over the little guy”.
I would argue that HFT is bad, not because it screws over the little guy (it may or may not) – but because HFT undermines the whole economic purpose of having markets in the first place: it obfuscates value (prices) rather than aiding value (price) discovery.
Lets step back for a minute and remember the reason why financial markets exist at all in capitalist economies. Providing liquidity is only one small part of the reason – if business owners want to sell their company (or part of it), they can do so without selling shares to the public. They don’t need an actively traded stock, or even a listed stock. Plenty of transactions are done between private entities, and most companies are actually sold (privately) to new owners – no financial markets involved at all. Many countries around the world have private companies, but they don’t have stock markets. In many others (including G7 countries), private company valuations are in aggregate many times the total market capitalization traded on exchanges.
The theoretical reason for having financial markets is to assign prices to large companies that (at least on average) equal the value of those companies. By assigning the “correct” price to each company, the markets direct finite capital to the places it can best serve society. Capital allocation is the reason why financial markets exist. Price discovery (as an approximation of value) is why financial markets exist.
The problem is that individual traders range from the very astute to the village idiot. Even the so-called experts can, and often do, make mistakes. Financial markets overcome this using “crowd wisdom”. Numerous studies have shown that a large crowd of people, each making estimates of a value, will usually beat even the so-called experts. Not just some of the time, but most of the time.
One frequently cited example is having a large group guess the number of marbles in a jar – the crowd average guess is usually very accurate even when experts guess wrong. There are still people who make crazy guesses (outliers), but on average the absurdly high and absurdly low guesses tend to cancel out. On average, the crowd gets it right – and usually beats even the experts who in theory should have an edge.
On average, a large crowd will estimate the correct value of a company as well, if not better, than the experts.
But crowd wisdom relies on several assumptions, perhaps the most crucial being that the bets are independent. If there is large scale collusion between crowd members, then essentially you no longer have a crowd, you merely have a handful of estimates that are just echoed by sub-groups. Outliers no longer cancel out, and the estimate becomes biased by a few guesses. Instead of having a true student government election, you have a handful of high school cliques. The popular clique “wins”, everyone else grumbles.
How does HFT fit into this? HFT is done by a computer sitting at the exchange data center. The computers are “co-located” because HFT strategies must be executed as fast as possible, they don’t think carefully what the company is actually worth – that is not part of their programming; it would take too long. For all the sales pitches and PR spin, most HFT algorithms are not all that different. They all use a moving average or two, an RSI indicator, and a list of pending trades (limit orders / market orders). They really aren’t all that different from one another — except for execution speed.
Essentially, HFT is the popular clique. For much of 2009-2011, HFT trading represented as much as half of all market trades. Half the trades were not independent bets, and worse, they weren’t even guesses about the true value of the underlying companies. They were ultra-short term mean reversion trades (selling gamma) without regard to whether the short term mean was valid.
HFT trading creates lots of volume, but more volume is not prima facie better (unless you are collecting per-trade commissions on that volume, which is why Wall Street likes HFT). Volume has both a quantity and a quality component – and HFT has zero quality. HFT algorithms do not attempt to value the underlying company, they don’t have the time. One might argue the quality of HFT trade volume is negative – because it tends to dry up when markets are extra volatile, exactly when liquidity is most needed.
Why do exchanges pay for HFT trade flow? Where does that money come from? HFT trade systems “win” (aka profit) – but who loses (who pays)? Wall Street lobbyist like to argue that HFT is victimless, but if there was no profit in HFT why would anyone do it? It sounds too good to be true – and it is.
. . . .We all pay for HFT trading – via exchange fees and false information about liquidity. What do we get in return? HFT systems are not required to balance order flow, as the old stock specialists on the NYSE used to do in return for their privileged market access. Getting 0.000001 higher price is legally a better execution – but practically speaking it is not. One has to trade 10,000 shares for the price difference to matter by a penny. Almost no one benefits from HFT in practical terms.
Why don’t the exchanges lower their transaction fees for everyone, instead of marking up each trade and rebating a fraction of the mark-up to just the popular clique? HFT firms and stock exchanges essentially gouge the public on every trade, and split the spoils between themselves. The HFT firm that gives the highest rebate back to the exchange gets to do the gouging.
. . . .HFT adds no meaningful value to the public. It is a poorly veiled way for a privileged few to benefit from exchange transaction fees.