Further points on high-frequency trading

Here is a good survey of some of the debates, plus Paul Krugman mentions the topic in his column today.  Here's my earlier post but I'd like to add or reiterate a few points:

1. On one hand, critics wish to charge that there is little or no advantage to having prices move more quickly to reflect new information.  On the other hand, some of these same critics charge that short-run volatility of prices — assuming this is in fact the result of HFT (and that is not proven) — creates social costs.  That's not quite a contradiction but it is an odd mix of views about the relevance of the short run.

2. I haven't seen a good estimate, or for that matter a bad estimate, of the social loss involved from investing resources in HFT.  Even if the practice has no gain, I suspect the loss is small.  It's the symbolic nature of the issue which excites people — bailed-out elites doing fancy things with powerful computers in a non-egalitarian manner — rather than the belief that it is a policy priority.  Even if you think HFT is bad, on an actual list of bad policies or practices in our world, would it be in the top million?  Mostly it's a canvas on which to paint complaints about the continuing political and economic power of finance, but we shouldn't let that skew our judgment of the practice itself.

3. There is no argument to date, and probably no argument period, that HFT can lead to financial insolvency or collapse on a major scale.  The cost, if there is one, is that the associated trading strategies bring a temporary collapse of asset prices for some period of time (how long?) or perhaps greater ongoing price volatility, or uncertainty about order execution, in the short run.

When I read that HFT may give markets "a new, currently unknown set of emergent properties" I think buying opportunity.

4. Research by Hans Stoll indicated that program trading was not in fact an instrumental culprit behind Black Monday in 1987, yet media coverage of HFT seems to be indicating that it was.  Many of the HFT debates echo themes from the earlier program trading debates from the late 1980s but in fact program trading did not turn out to be a major problem.  We have been down this path before and it turned out there was much less there than the critics thought at the time.

5. The more I read these debates, the more nervous I get about the idea of a financial products safety commission.  Essentially on innovation we're seeing a flipping of the burden of proof and I don't think it is possible to easily fine-tune that flipping in a way to capture good innovations and rule out bad ones.  We should still follow the rule of regulating practices shown to be harmful or likely to be harmful.


On one hand, critics wish to charge that there is little or no advantage to having prices move more quickly to reflect new information.
Of course it's a good thing. The question is whether the costs outweigh the benefits.

Question 1: You call your broker and place a buy order. Someone taps the call and trades in front of you, raising the price. Is this a good thing? The price moves quickly to reflect the new information.

Question 2: Someone trades on material non-public information. Is this a good thing? The price moves quickly to reflect the new information.

The issue is that HFT moves money from the investors who are looking to invest in the production aparatus of society to people who are engaged in naked wealth extraction. Effectively, goldman is looting the entirety of the stockexchange, and thus also looting, the part of the economy listed there, which is a fairly big chunk. So, yes, its a serious social problem, in pretty much the exact same way bank robberies and fraud is. These people are thieves, and in need of stopping.

As the returns on high-frequency trading strategies are not correlated with returns on other asset classes, HFT was a space that performed very well in 2008, especially when compared to the rest of the market. So what do we see in 2009? Beefed up investments in high-frequency trading infrastructure, both with respect to technology and the quants that develop strategies in that space. Both established firms and financial "entrepreneurs" trying to raise capital and start new HF funds.

To make a comparison to another recent era in the markets from about, oh, 9 years ago, in the short term there will be a lot of hype and gnashing of teeth. Blood will spill, and more teeth will be gnashed about the damage done to the average 401k investor, although most of the carnage will take involve the HF participants themselves. In the long run, the markets will be faster and more efficient and at least a couple of gee-whiz innovations will come out of all the wreckage. The partners at a few firms, only some of which are already stinking rich, will become stinking rich.

Why isn't it objectionable that the trading floor isn't level? The call isn't *tapped*, but rather the information is *given* to the fast trader. Suppose you were at the horse track and your bet was first passed to the guy at booth #1, before your bet was processed.

This trading also can prevent yours. If they were mean enough, they could always intercept your trade, ensuring that the desired shares were always bought up before your trade was processed.

In effect, by letting others skip in line, armed with your trading information, they are collecting a tax on all other traders. This isn't a trading floor any longer but "the only game in town".

I haven't read Krugman's article. But I just stumbled on Ezra Klein's take because I follow his RSS feed (I think because Tyler recommended him).


Is there a more shallow commentator alive? So he's supposedly the "economics expert" at the Wa. Post.

The idea that consumer protection is going to lessen innovation is a myth, well-exposed and debunked today by Elizabeth Warren.


But it's no surprise that it keeps coming up: Scaring people with the costs of a loss of innovation was a common tactic leading up to the crisis. (Don't take my word for it: http://www.newdeal20.org/?p=3476) At some point, those who argue that a lack of innovation will cost us in the long-run will need to own up and admit the huge cost that a plethora of innovation has visited on us since last year.

If you clear the markets once per second, you can bet that high-frequency traders will adapt. Once per second? Is that all you got, Wellman?

Are we really at the point where Goldman et al replaces GM in the old syllogism of what's good for America? What have we learned from the last year? That anyone who claims to be able to accurately access financial risk is deluding themselves.

What possible social benefit is there in allowing a parasitic coterie to enrich themselves at the risk of destabilizing global financial markets. It will take out children's lifetimes to pay off our current fiasco, do we want to offer up our grandchildren too?

To address Jerome's horse betting example and others' questions about front running, the typical user of the flash order type is an investor who wants to make a large bet on a horse without "changing the odds", or in trader parlance, "information leakage". So a small group of other big bettors are offered the first opportunity to take the other side of the bet, at or at slightly better than the current prevailing odds, before the entire bet goes to the betting window. When such a match can be made, both the original bettor and the counterparty presumably benefit, even if the counterparty later uses its private information against the bettor in the future (yes, if I make it easy for Goldman Sachs to front-run me I still might be better off than if I let everybody front run me).

In one sense, the trader and the counterparty are screwing over the little guy, denying him access to this large source of liquidity. In another sense, two agents are engaging in a consensual, mutually beneficial transaction that improves efficiency (if the large trader's costs from information leakage are lower, they will trade more) and no other party has a cause to claim they were harmed.

I second mobile here. It is not that Goldman is paying for a special view of the market, it is that investors are paying Goldman (with information about their desires) in exchange for a better chance of getting filled at a better price. It is a legal practice and both parties know exactly what they are doing. If Goldman (or the other broker-dealers who see these flash orders) don't provide a useful service in return for that information, the big investors using those orders will stop using them. Those people are capable of figuring out if using flash orders is a losing proposition. We don't need to the government to come in and protect them from themselves.

Back @ Doc Merlin: All true. But my "local knowledge" is that there is a lot of froth and that a lot of the investments made in 2009 and 2010 will go badly (not that there's anything wrong with that). More than once I've come across someone who "has this friend" that "knows this guy" that wants to plow his Internet fortune/ arms dealing proceeds/trust fund into a "high frequency trading" fund. I think the people who know what they are doing will get very rich off the people who think they know what they're doing.

The idea that regulation which promotes fair playing fields somehow taxes society is false. Good frameworks which create fair playing fields encourage more participants making the social effect larger than unfair fields tilted for the benefit of a highly powerful few. It's interesting to watch traders (those who are currently benefiting from the system) resist this relatively basic concept that is absurdly valuable.

after reading more ive suddenly flip-flopped my moral position. hft is clearly just market making by another name.

Doc Merlin -- from 1850 to 1950 per capita real gdp growth averaged about 1.4%.
It is the same if you make the end data 1925 or 1930 so the depression does not distort the trend.

Since 1950 per capita real gdp growth has averaged about 2.1%.

So in essence growth post New Deal was about 50% better than pre-New Deal.

Is that what you wanted to point out?

If whether we regulate high frequency trading makes no difference one way or the other why are you exercised that we might regulate it?

Moreover, I would say that increase of high-speed trading changed arbitragers behavior from liquidity takers on both sides to liquidity movers (moving liquidity from hedge instruments to the instrument they quote) or to liquidity providers on both sides.

In old days when a lot of things were done manually and just few participants had high-speed systems guys who traded spreads (say stock versus it's ADR) would monitor price difference between two instruments and if mismatched, they would execute buy and sell orders simultaneously, essentially taking liquidity from both sides (stock and adr). Now this just doesn't work this way - you can't make money on it (if you're not trading something like Russian stock vs their ADRs), because such an imbalance exists only for a very short amounts of time, so even systems with couple of hundred microseconds latency can't pick it up). Now you have to maintain orders on market at least on quoted instrument, essentially providing liquidity in it. When you do hedge, yes, you can take liquidity in hedge instrument, but also you can choose to maintain order in hedge instrument too to limit your hedge host (to not pay spread and exchange commission there).

So, what I see (being in a trading systems vendor business for some time) is that speed brought liquidity onto markets.

Paul Krugman and I drink the same wine. We totally do. And we both have two arms and both of us graduated from fabulous schools!

"Essentially on innovation we're seeing a flipping of the burden of proof..."

which is entirely a good thing

First, let me compliment you on a well-written and thoughtful article. I was impressed to read this side of the argument so well put. However, I find I have to disagree with you.

I feel that you're overreacting to potential government intervention in regards to the financial products safety commission, to begin with. The idea that markets regulate themselves and that companies that don't work in the public interest will necessarily lose business to competition is a rather naive proposal that depends on actual competition being available. I'm not suggesting the government should be involved in running every business, but we made business practices like creating pools illegal in the first part of the last century. I don't see why so many people are opposed to upholding that law.

A lot of your arguments are based on the idea that there is no proof that HFT will have any effect on the markets as a whole. My two points on this are that there's no proof because we're still just starting the process, but I don't need to put my hand on a stove to know it'll hurt.

Secondly, it's not so much HFT itself that causes the problem, though the vast potential for illegal activities and immoral actions is clearly there (as with any new technology), but the irresponsible behavior this may encourage. I wrote an article on my blog about the NYSE's new fast trading hub and how it will promote activities like sculpting which undeniably affect stock prices and make the market more volatile. Being able to trade at faster speeds is a necessary advancement in the technology of investing, but with it comes dangers that are very hard to ignore.

As far as Stoll's arguments, I simply don't find them convincing in light of overwhelming analysis for the past 20 years that suggests otherwise. There will always be a few people that agree with any proposition, and with all due respect to Dr. Stoll, his paper argues a mathematical oddity, not a market trend.

Comments for this post are closed