As the article points out, these firms provide liquidity which makes trading faster and cheaper. Theoretically, more liquidity should also make it easier to move capital to the most profitable enterprises.
Actually, the article suggests that the firms pulled out of the market, and that their actions amplified the crash. HFT is a zero sum game, and these firms just demonstrated that they exist to extract value from the market, not provide it liquidity.
[edit]:Though there are obvious differences, I feel like HFT in general is just another form of side betting, alarmingly similar, in principle to CDOs and other derivatives, in at least a few ways. They're exotic, and theirfore unregulated, they dwarf and therefore distort "real" market activity, and they appear to provide value, but the geniuses behind them underestimate the risk because they can never account for all the possible black swan events.
When "real" markets depend on these educated guesses, and the guesses turn out to be wrong, the real markets suffer. Sure we know how to create fancy new financial products to back up these guesses, and exploit the leverage provided by developments in math and computer science, but the fact remains that we still can't even identify, let alone account for what we don't know.
The good news is that because these particular bets are shorter term, in this case the immediate damage is reversible, though their were definitely consequences that were irreversible, and if we keep allowing our equity markets get pushed around by HFT, the next time the system hiccups, the irreversible consequences may spiral out of control.
But I don't have a very deep finance background. I'd love to hear thoughts from some people who do.
Without high frequency traders, the market became far less liquid. This was considered a bad thing by almost all involved. It's almost as if high frequency traders are actually providing a valuable service for the rest of the world!
Also, HFT is not very vulnerable to "black swan" events. HFT doesn't usually use leverage, there are very few counterparties, no leverage, and almost no contagion.
I totally agree.
>This was considered a bad thing by almost all involved.
What I'm trying to propose is that the regulators involved in overseeing markets are becoming more and more out of touch with the techniques being used to improve their efficiency. What's important for long term economic health does not align with what is important for short term market efficiency. And as a result of this, the long term economic policy cannot keep up with the increasing degree to which short term regulation can affect markets.
Yet long term economic health has an impact on other metrics of "utility" beyond itself, and we're in trouble if the power of the people in charge of making the rules can't address the needs of the people who depend on them.
In other words, our regulators actions are predicated on the assumption that we knew how to improve the efficiency of the equity market. Lately, the influence they have on the markets has been increasing, but the amount of oversight they have been subjected to is decreasing. The skills that are relevant to getting politician into office are diverging from the skills and attention that are required to selecting these regulator in the first place. The politicians wind up screwing the people who both elected and depend on them.
--
Let's consider the current, using arbitrage to improve the market's liquidity case, as an example. The most effective arbitrage techniques are made possible in part by simultaneously exploiting developments in computer science theory with those financial theory, and combining our improved ability to process this newly discovered information with our improved ability to implement our improved knowledge. In other words, in this case our increasingly sophisticated understanding of finance and computer science is what allowed us to more efficiently implement and test our hypotheses in the first place.
Previously we used CDO's to improve our ability to assess risk. We did this because we assumed that assessing risk was important to the efficiency of the markets. However, CDOs were made possible by developments in financial theory and implementation, as well as statistics, and the politicians in charge of influencing our dependance on ability to improve this sophistication were out of touch with the long term effects that their decisions have.
Why are crashes bad?
I think the point that you may be missing is that there will always be booms, busts, panics, euphorias, etc. Those are byproducts of human psychology.
The real problem is when people blindly expect the market to always be stable, to always increase in price, etc. Most of the people who lost money in the recent financial crisis were people who couldn't afford to lose on the bets they made or who had essentially let it all ride without understanding that sometimes share prices go down too.
It's a self-serving argument at best. I sometimes wonder why the whole thing couldn't be replaced with a giant daily auction.
If I put in an order to sell one share of GOOG for $1, would you start to believe that the market price of GOOG was $1 per share?
A market price is only as valid as the degree of consensus. Increased liquidity means that there is a high degree of consensus about the current price. This means that one can feel additional confidence that one is getting a fair price.
The main concept to think about here is an order book. Think about the psychology behind orders that are a ways away from the best price (both bid and ask). Those parties are willing to transact, but at a premium. In other words, they prefer to hold unless a sweet enough offer is made.
Thus, for a very large order one has to pay that premium.
The investors get a daily price, and they have to notify us they day before.
There seem to be two obvious problems here. First, what actually happened was that at the point where that service would have been most valuable they stopped providing it. There seems to be a difference in value between idealized theoretical HFTs that provide liquidity at all times, and real HFTs that provide liquidity when there's plenty of it around anyway.
Second, "they abruptly took X away and the results were bad, therefore X was providing something valuable" is an unsound argument -- e.g., coming off a drug cold turkey can be very bad even if the drug was doing you no good at all. It doesn't seem implausible that a sudden drop in liquidity during a market crash might be a very bad thing quite regardless of the optimal level of liquidity in general.
That said, there's also traditional leverage used in many cases because you may have a fully hedged position across different exchanges that don't recognize each other's instruments in margin calculation. For instance, you can be long Jan 11 Brent Crude Oil on CME and be short just as much on ICE. The risk is practically zero, so firms borrow as much money as they can to put on as much position as they can. If you've got more margin than others in these sorts of trades, you win huge when everyone else is backed up to the wall.
ALL companies exist to create wealth for their shareholders. Google does't provide search out of pure altruism, but providing their service and making money go hand in hand.
Incidentally this sort of trading is only zero-sum if you ignore time. Taking a position based on events that have not happened yet is in fact buying or selling risk.
We do this because by facilitating companies we get employment, goods, services and wealth for owners. Of these the utility from wealth for owners is the least significant.
From the point of view of a retirement saver who invests at a single price per day, high frequency traders do hold positions for close to zero time. So in this zero sum game hft traders are just reducing our income.