More on High Frequency Trading and Liquidity

by on October 17, 2011 at 7:44 am in Data Source, Economics | Permalink

Tyler is more optimistic about financial innovation than I am. Strange, but true. I recommend Andrew Haldane’s speech, The Race to Zero, on high frequency trading (HFT). Haldane is Executive Director for Financial Stability at the Bank of England and his speech is eminently quotable. First, some background from Haldane:

  • As recently as 2005, HFT accounted for less than a fifth of US equity market turnover by volume. Today, it accounts for between two-thirds and three-quarters.
  • HFT algorithms have to be highly adaptive, not least to keep pace with the evolution of new algorithms. The half-life of an HFT algorithm can often be measured in weeks.
  • As recently as a few years ago, trade execution times reached “blink speed” – as fast as the blink of an eye….As of today, the lower limit for trade execution appears to be around 10 micro-seconds. This means it would in principle be possible to execute around 40,000 back-to-back trades in the blink of an eye. If supermarkets ran HFT programmes, the average household could complete its shopping for a lifetime in under a second.
  • HFT has had three key effects on markets. First, it has meant ever-larger volumes of trading have been compressed into ever-smaller chunks of time. Second, it has meant strategic behaviour among traders is occurring at ever-higher frequencies. Third, it is not just that the speed of strategic interaction has changed but also its nature. Yesterday, interaction was human-to-human. Today, it is machine-to-machine, algorithm-to-algorithm. For algorithms with the lifespan of a ladybird, this makes for rapid evolutionary adaptation.

Consistent with the research cited by Tyler, Haldane notes that bid-ask spreads have fallen dramatically.

Bid-ask spreads have fallen by an order of magnitude since 2004, from around 0.023 to 0.002 percentage points. On this metric, market liquidity and efficiency appear to have improved. HFT has greased the wheels of modern finance.

But at the same time that bid-ask spread have decreased on average, volatility has sharply increased, as illustrated most clearly with the flash crash

Taken together, this evidence suggests something important. Far from solving the liquidity problem in situations of stress, HFT firms appear to have added to it. And far from mitigating market stress, HFT appears to have amplified it. HFT liquidity, evident in sharply lower peacetime bid-ask spreads, may be illusory. In wartime, it disappears.

In particular, what has happened is that stock prices have become less normal (Gaussian), more fat-tailed, over shorter periods of time.

Cramming ever-larger volumes of strategic, adaptive trading into ever-smaller time intervals would, following Mandelbrot, tend to increase abnormalities in prices when measured in clock time. It will make for fatter, more persistent tails at ever-higher frequencies. That is what we appear, increasingly, to find in financial market prices in practice, whether in volatility and correlation or in fat tails and persistence.

HFT strategies work across markets (e.g. derivatives), exchanges, and stocks and can have negative externality effects on low frequency traders. As a result, micro fat-tails can become macro fat-tails.

Taken together, these contagion channels suggest that fat-tailed persistence in individual stocks could quickly be magnified to wider classes of asset, exchange and market. The micro would transmute to the macro. This is very much in the spirit of Mandelbrot’s fractal story. Structures exhibiting self-similarity magnify micro behaviour to the macro level. Micro-level abnormalities manifest as system-wide instabilities.

For these reasons I am not enthusiastic about innovations in HFT. Earlier I compared high-tech swimming suits and high-frequency trading:

High-tech swimming suits and trading systems are primarily about distribution not efficiency.  A small increase in speed over one’s rivals has a large effect on who wins the race but no effect on whether the race is won and only a small effect on how quickly the race is won.  We get too much investment in innovations with big influences on distribution and small, or even negative, improvements in efficiency and not enough investment in innovations that improve efficiency without much influencing distribution, i.e. innovations in goods with big positive externalities.

Chris October 17, 2011 at 8:08 am

I am skeptical. Shouldn’t it be possible to design an algorithm that takes advantage of higher volatility, if that higher volatility is entirely HFT-generated?

Nancy Lebovitz October 17, 2011 at 8:10 am

Presumably, there are people investing to take advantage of increased volatility. Would you expect this to dampen the volatility? Amplify it? Something more complicated?

NAME REDACTED October 17, 2011 at 8:17 am

Its harder to go short volatility than to go long volatility. This is a serious problem right now.

I am K October 17, 2011 at 9:02 am

It’s not. For example, if you buy stocks you are short a certain kind of volatility, precisely the kind we are talking about here, while being long the right kind of volatility ( the option on getting better than expected future cashflows )

NAME REDACTED October 17, 2011 at 10:24 am

“For example, if you buy stocks you are short a certain kind of volatility, precisely the kind we are talking about here”

And what kind is that?

I am K October 17, 2011 at 10:58 am

Short-term “markets are freaking out” volatility

Right Wing-nut October 17, 2011 at 12:10 pm

Right up until the point that the exchange decides to unwind all the trades. >:E

byomtov October 17, 2011 at 10:28 am

“A serious problem?” For who?

corporate_serf October 17, 2011 at 4:20 pm

You can go short realized variances to within replication error.

NAME REDACTED October 18, 2011 at 9:26 am

You /can/ its just more costly overall to short variance than to go long variance.

Tyler Cowen October 17, 2011 at 8:19 am

What’s wrong with higher volatility? Aren’t computer vs. computer chess games more tactical and more volatile? In that case, we *know* they are playing better moves.

I am K October 17, 2011 at 9:00 am

The downside of higher volatility at short run frequencies is that it can induce people to be more fearful on longer term frequencies as well and make business investment more volatile, which is not good In theory, if that happens, traders should be able to arbitrage it away by shorting “economically unjustified” volatility.

mjw149 October 17, 2011 at 9:00 am

That’s the stock market fallacy. We should not encourage ‘better moves’. We need stock prices to fluctuate based on actual value of the companies in the real world. That’s the whole point of capitalism. Greater volatility is necessarily a negative, as it introduces uncertainty artificially and uncertainty decreases investment and demand.

So the question is, can machines achieve better understandings of listed companies’ value than humans? Probably, but that’s not what these computers do.

I’m quite concerned at the pattern for a market crash now. I have a reasonable expectation that anytime a real crash occurs, the human owners will pull the plug to analyze the machine’s moves, making it worse.

To be fair, it’s possible that they’ve kept the market afloat in the last couple months when all the uncertainty over Europe and the US simultaneously cause less confident human investors to pull out. As long as you’re operating one of these machines, after all, you can reasonably expect to make money even as the market goes down. So perhaps there’s a benign reason that their volume has increased. Perhaps the fact that the models can’t read the news for macro conditions and none of the programmers considered the US credit rating as a variable rather than a constant is what saved us.

But that’s a guess of course.

Andrew' October 17, 2011 at 9:30 am

Whatever happened to the story that Taleb lost his shirt in the flash crash? Losing large sums of money is probably something that the users will want to be patched in the next release of HFT 2.0.

Why wouldn’t companies put in orders to buy if their shares drop to pennies? Short squeezes on insolvent banking and shadow banking institutions is a separate, known issue and needs a separate consideration as long as the gov insists on protecting fractional reserve.

Your last paragraph is outstanding, but the more the computers can pick up the pennies, why doesn’t this free up the alleged brain drain to finance?

derek October 17, 2011 at 11:48 am

>possible that they’ve kept the market afloat in the last couple months when all the uncertainty over Europe and the US simultaneously cause less confident human investors to pull out.

And when they pull the plug, there are no buyers. And there will be no bottom.

NAME REDACTED October 17, 2011 at 10:25 am

Volatility is another name for riskiness.

mark October 17, 2011 at 11:56 am

Agree with other comments. Volatility tends to drive prices and multiples down. The wealth impact far outweighs the 20 bps of compression of the bid ask spread.

Also (and this is a small point by comparison), the bid ask spread compression is a bit misleading. I’ve learned to put in limit orders outside the bid-ask spread because most of the time HFT’s intraday variability will within a few minutes drive the clearing price outside the bid ask spread at any given moment. So it’s actually not increasing efficiency – that is, it’s not reducing the spread between bid ask spreads that would otherwise exist, it’s creating more, and a wider range of, bid ask spreads. I don’t see that is intrinsically as more efficient.as it sounds.

Last, it doesn’t seem to have anything to do with improving capital allocation. To the extent secondary trading is valuable because the results show who are good capital allocators, this subverts that. It’s as if one set out to select a sniper by running a contest for blindfolded contestants with machine guns.

figleaf October 17, 2011 at 12:35 pm

Possibly the only meaningful contribution George Lucas ever made was the tossed-in sequence in THX-1138 where one technician is training another to use a piece of equipment that happens to cause involuntary muscle contractions in prisoners, with the prisoner in question being THX himself. As the trainee struggles to master the technology the trainer says something like “careful not to go over 4.6″ while THX’s agonized spasms become progressively more violent. A moment later the THX collapses and the trainer says with disappointment in his voice something really bland like “oh, you let it go to 4.7.”

Volatility that results from modeling chess as a finite-state automata is interesting in the sense that it leads to increasingly optimal algorithms for winning — some of which might be of interest or even useful for subsets of live players. The difficulty with volatility in markets is that, like THX, you’re operating on actual corporeal… well… corporations. For the trader it’s no doubt frustrating when a trainee induces “dissolution at 4.7,” as I’m sure it is for the abstractly-inclined economics professor. For the people running the actual businesses who’s stock is affected “frustration” probably isn’t the word that comes to mind.

On paper there should be nothing wrong with volatility in your core body temperature either, especially since over time your average temperature remains 98.6. To argue therefore that increasing volatility should have no adverse health effects would be missing the point.

As my old abstract math prof occasionally said, optimum body temperature isn’t the same thing as maximum body temperature.

figleaf

Crenellations October 17, 2011 at 3:52 pm

Volatility definitely reduces M&A activity. I dunno if that’s a “wrong” or a “better move”.

Lou October 17, 2011 at 5:44 pm

Investors will require higher returns for more volatile assets, all else equal (logically, because volatility is a form of risk). This means they will pay lower prices. This makes financing more difficult, and financing is the reason financial markets exist. Whether or not volatility is evidence of superior investing tactics, it hinders the market’s primary function and its limitation is a worthy goal.

D October 17, 2011 at 6:00 pm

Chess has a final outcome that is zero sum. Should the market be like that? It’s a real question. I don’t know enough to have an opinion.

mpkomara October 17, 2011 at 8:22 am

“We get too much investment in innovations with big influences on distribution and small, or even negative, improvements in efficiency”

Who gets too much investment? You? Me? The world? Shouldn’t the investors be able to determine the returns on their investment? What’s the right amount of market efficiency, and how do you measure it? Is there an “H” in HFT that you think optimizes bid/ask liquidity versus volatility? Market participants were slightly less high-frequency last year and less the year before that; maybe we should limit technology to 2009 levels? From your comments should we only be able to trade as fast as we blink?

Andrew Montgomery October 17, 2011 at 8:25 am

“HFT strategies …. can have negative externality effects on low frequency traders”.

Most of us are low-frequency traders, either directly or through our pension savings. If HFT has negative externalities which affect the majority of voters, then a democratic government should ban it. Or have I missed something?

I am K October 17, 2011 at 8:57 am

Yes you have missed a lot. Let’s ban frozen pizza too. It has negative externalities on a majority of voters by making people fat. Let’s ban soft drinks, pornography, etc., etc. But more importantly, HFT is impossible to ban. How would you go about it without destroying the market place?

mjw149 October 17, 2011 at 9:04 am

Of course you can and should regulate HFTs. You think frozen pizza isn’t regulated? When was the last time someone you know died from packaged food?

The market needs to be controlled, yes it would be difficult (hell we haven’t even fixed the derivatives issue have we?) but given that we’re only human, I have a feeling we’ll just wait for a real crash. There’s no way to properly assess risk and coordinate action otherwise.

I am K October 17, 2011 at 9:07 am

What makes you think HFT is not regulated now? All the firms in the industry are regulated by CFTC and SEC as well as exchanges they trade on. All these firms have compliance staff that monitor everything, probably to a much better degree than your average bank. The issue is not whether or not HFT should be regulated, because it already is, the issue is, should it be regulated out of existence or not.

gappy October 17, 2011 at 10:16 am

Sorry, you’re missing the point in so many ways. First, people die from packaged food quite often, as these happen to be contaminated by Salmonella or other viruses. Second, packaged food regulation doesn’t affect the quantity or type of pizza anyone wants to buy or sell. Third, it’s far from obvious that HFT imposes negative externalities on *low-frequency* traders. There is a very compelling case that they have positive externalities for these investors, by reducing spreads and improving price discovery. On the other side, if you are a day trader and think you have an edge, think again.

I look forward to hearing how do you “fix the derivative issue”.

Andrew Montgomery October 17, 2011 at 9:20 am

This article is about HFT; let’s stick to the subject.

How would you ban it? Simply by imposing a rule that all stocks must be held for at least five seconds / five minutes / five hours if you want to push out all short-term speculation (probably a bad idea). We didn’t have HFT twenty years ago, the market can clearly function without it.

I am K October 17, 2011 at 9:32 am

Ok, let’s do that. Then a bunch of orders will be sitting there for 5 seconds, will be occasionally mispriced relative to factors in the world, HFT algorithms will hit those orders instead of providing liquidity. Result? HFTs are making as much money as before, or perhaps even more. but there is less overall liquidity in the world.

We had HFT 20 years ago in the form of specialists and runners on the exchange. Only it was less fair than now because it’s harder to get a runner than to get a computer.

Andrew Montgomery October 17, 2011 at 10:42 am

Another way to limit HFT strategies would be to convert the entire stockmarket into a Dutch Auction market. Buyers and sellers place bids, and the market clears at one price every few minutes. Illiquid stocks in smaller companies would clear every few hours. Not only are HFT strategies thus rendered useless, but it also curtails other fast-mover advantages such as first access to breaking news.

Without more research I don’t know if this would be better than the current system, or even practical. I offer it merely as a starting point for other ideas.

Tom West October 17, 2011 at 10:45 am

I don’t think that anyone doubts that HFT has some benefits. It’s simply a matter of whether those benefits outweigh the perceived risk of extreme damage to the entire financial system. The advantage of systems operating in a human time frame is that when the unanticipated occurs (which, by definition is not anticipated, so we can’t prepare for it), we have time to react to the emergency. With HFT as it currently stands, if there’s an unanticipated crisis, it’s already too late.

K I Am October 17, 2011 at 11:09 am

The point is precisely this: nobody knows if it would be better than the current system or even practical, it’s very hard to predict what would happen. But if you did this, there is clearly a huge risk of destroying the market place.

byomtov October 17, 2011 at 10:30 am

Those are not negative externalities. Your pizza consumption doesn’t make me fat.

I am K October 17, 2011 at 11:03 am

it makes me fat which makes me sicker which makes medicare payout more which raises your taxes. QED baby

Andrew Montgomery October 17, 2011 at 11:30 am

In Denmark, they banned trans fats entirely because of this negative externality. The idea of a soda tax has been around for quite a while, although practical measures such as banning soda machines from schools are easier to implement. Negative externalities are fine if the cost is recaptured in tax – that’s partly why we pay gas tax. I don’t see a mechanism to recapture the /alleged/ negative externality that HFTers impose on low-frequency market participants.

I K AM October 17, 2011 at 12:08 pm

The key here is alleged. The externality from trans fats is not alleged, but very real. They did not ban frozen pizza, did they ? Secondly, yes they do that sort of thing in Denmark. Do we want to turn U.S. into Denmark? but that is a whole other conversation.

NAME REDACTED October 18, 2011 at 9:29 am

Socialization of /anything/ means creating externalities. This is what government is, its an externality factory.

Marwan October 17, 2011 at 8:25 am

Capitalist inefficiency in a nutshell: “We get too much investment in innovations with big influences on distribution and small, or even negative, improvements in efficiency and not enough investment in innovations that improve efficiency without much influencing distribution, i.e. innovations in goods with big positive externalities.”

Gunnar Tveiten October 17, 2011 at 8:28 am

Lower spreads are good, as such – if we could get them for free. But a large increase in volatility is too high a price to pay for it. Especially since spreads already, by their nature, matter more to the speculant and less to the investor.

If you hold a stock for a decade, a difference in spread from say 0.5% to 0.2% matters essentially not at all – 0.3% for a decade translates to less than 0.03% difference in the yearly return on the investment.

If, on the other hand, you hold stock for a week, then the same difference subtracts 0.3% of your value once a week, 52 times a year, or in short, it makes 15% pro year worth of difference, which is a enormously huge difference.

On the flipside, if you *are* a long-time investors (as opposed to a short-term speculant) then the antics of the HFT-crowd end up hurting you less. If their quicker turn-around ends up with them ahead of you by 0.1% on every trade, then that matters hugely for the short-time-speculant, but not really at all for the decade-perspective buy-and-hold investor.

Anon October 17, 2011 at 8:41 am

Chris nailed it. The traders and their algos will adapt to profit from the tails/volatility.

The big flick is that the spread has dropped to nothing. That means that I, a regular old investor, get a much better deal when I finally get around to buying or selling. Thank you HFT!

NAME REDACTED October 17, 2011 at 10:28 am

+1

Tom West October 17, 2011 at 10:54 am

That means that I, a regular old investor, get a much better deal when I finally get around to buying or selling.

Much better deal? I don’t think the spreads materially impacted long-term buyers.

However, if because of HFT, there’s a long term drift away by mainstream investors from stocks as a whole, or if the nightmare scenario occurs and the markets get completely hashed in some unanticipated HFT-related crash, I’m *much* worse off.

Frankly, saving 0.2% every few years against a (number chosen out of the air) 1% chance of losing big seems a pretty bad deal for most investors.

mark October 17, 2011 at 12:00 pm

You don’t. The impact of volatility has reduced multiples. The lost wealth outweighs the transactional savings.

msgkings October 17, 2011 at 1:38 pm

This. The markets have never been more perceived as casinos (correctly I’d say), and this does in fact affect their only meaningful function, as the method for allocating capital efficiently from savers to companies that need it.

To most of the world, ‘spread compression’ and ‘price discovery’ are useless abstractions. To most of the world, Goldman’s computers are rigging the game and they don’t want to play. Anyone notice how few companies even bother with public markets if they can avoid it?

Anon October 17, 2011 at 8:42 am

Gunnar, don’t pay for volatility. Use limit orders that don’t expire for 30 days.

NOZE October 17, 2011 at 3:14 pm

The risk of HFT is that standing, visible GTC or day limit orders get dimed. With little advantage to placing them (other than convenience), the decks of standing bids below and offers above thin out and apparent sweep prices widen (that’s the expected average fill price on a market order of X shares).

NAME REDACTED October 18, 2011 at 9:30 am

If anything, that volatility will help you get even cheaper prices, if you use limit orders. :-)

Anon October 18, 2011 at 12:10 pm

I get the feeling we’re surrounded by people who don’t understand what we’re talking about.

gappy October 17, 2011 at 9:12 am

I think the paper deserves a closer reading, so I would not feel comfortable summoning it to shore up my own thesis, whatever it may be. However, relatively early on, Haldane says:

“So is there any evidence of increasing abnormality in market prices over the past few years? Measures of
market volatility and correlation are two plausible metrics. 9 Chart 9 plots the volatility of, and correlation
between, components of the S&P 500 since 1990. In general, the relationship between volatility and
correlation is positive. Higher volatility increases the degree of co-movement between stocks.”

This is a non-sequitur. First of all, asset-asset correlation has notthing to do with “abnormality”. Second, increased correlation and volatility have a parsimonious explanation: they are the consequence of higher factor volatility *while keeping everything else constant* (factor correlations, idiosyncratic volatilities). These factors argually have little or nothing to do with high frequency trading, and have been increased in volatility well before the advent of HFT. In fact, there is a rich literature trying to explain higher volatility.

I haven’t read Haldane’s presentation in detail, but if the HFT -> vol causal relationship doesn’t hold, much of the theses in the paper fall as well. A way to test this relationship would be to relate HFT participation rate (which, from some papers I read, can be inferred, at least in some Europan exchanges), and a volatility factor.

I would expect a paper by the Executive Director for Financial Stability at the Bank of England to be almost unimpeachable, but this doesn’t seem to be the case.

I am K October 17, 2011 at 9:18 am

If I could upvote this a million times, I would

gappy October 17, 2011 at 9:43 am

Ahah! Thanks!

Jacob October 17, 2011 at 1:08 pm

me too!

David Zetland October 17, 2011 at 9:14 am

When is Tyler NOT more optimistic than you? Top ten request!

eccdogg October 17, 2011 at 9:25 am

How do we know HFT is to blame for all the extra volatiltiy? During most of the time being discussed for higher volatility we have also had huge financial crisis with news coming out every few months that make people think the world is going to end.

And as others have said if you think volatility is too high then sell it. Set your price target for the stocks you follow and put bids and offers out 10% below and above. You would have made at least 10% in one day during the flash crash.

The other thing is the flash crash which everyone points to as being the huge example of how bad HFT is had zero impact on me as long term investor. I view the battle to regulate HFT as a battle between high frequency computer traders and high frequency human traders not long term investors.

Right Wing-nut October 17, 2011 at 12:22 pm

NOT TRUE! The exchange rolled back the trades. Your 10% limit orders hit, but you don’t get the money.

eccdogg October 17, 2011 at 12:31 pm

To the extent that that happend and it did not happen to everyone. That is a problem with the exchange not HFT.

Dredd October 17, 2011 at 10:15 am

The cops finally showed up on Wall Street along with the free press in what some are calling the first shots in the germ wars which may spread around the globe.

It was a very liquid high frequency investment.

Ansel October 17, 2011 at 11:05 am

Why doesn’t someone do a trading experiment and settle the question? Experimentally set maximum trading frequencies or minimum order lifetime and see if volatility has a causal relationship.

Scott Sumner October 17, 2011 at 11:27 am

Volatility of equity prices increases sharply when low AD is a big problem and there is lots of policy uncertainty. That was true in the 1930s, and it’s been true since 2008. It’s not obvious to me that HST played a role in this.

s jay October 17, 2011 at 11:16 pm

Amen. HFT is a huge distraction from the larger, more important financial regulatory issues facing the world. When Merkel and Sarkozy met that weekend when Europe was falling apart, and their sole announcement was that Europe needed a financial transactions tax, everyone should have been very, very afraid. European banks are on the verge of imploding, taking the western world with it, and the only diversion they can come up with is anti-HFT rhetoric. Sometime, when you have some spare time, compare the HFT profits made in the transparent, hugely competitive US equity markets, with the take in the opaque, uncompetitive OTC fixed income markets. Interest in HFT amounts to an uninformed fad, or fiddling while Rome burns. It’s being driven by those whose ox is being gored by HFT, along with a desperate need to keep attention off the truly expensive parts of the financial markets, to the huge detriment of the common investor. The quotable, speculative drivel from Mr. Haldane exposes his level of competence (or lack thereof) at a time when we desperately need effective leadership on the real issues.

NAME REDACTED October 18, 2011 at 9:32 am

“Europe was falling apart, and their sole announcement was that Europe needed a financial transactions tax, everyone should have been very, very afraid.”

If they do this, there will be a massive sucking sound as European money moves to the US.

Eddie October 17, 2011 at 11:27 am

The implications of high volatility is that it will eventually scare away a lot of long-term investors, compress overall P/E ratios, and consequently hurt anyone who is currently invested long-term in the market.

As mentioned up-thread, lower spreads really only help short term traders. So, HFT creates more volatility and the only ones who benefit from HFT are HFT traders. Wait, what?

Becky Hargrove October 17, 2011 at 11:38 am

Alex, when I consider HFT I am reminded of the beauty of a fractal, and the way that it replicates itself smaller and smaller as it cycles through the pattern that was created. Gradually, you can’t even see the replications before the pattern changes.

Steve C. October 17, 2011 at 12:18 pm

I’m neutral on the features and benefits of HFT. I’m pleased that the cost of trading has fallen dramatically since 1990. But that cost compression had nothing to do with HFT which is has evolved over the past 10 years. Is there really that much of an opportunity left in squeezing trading costs?

My concern is that HFT is the latest financial market manifestation of giving guns to monkeys. Or, to paraphrase Jeff Goldblum in Jurrasic Park, “When you found out that you could do it, did you ever stop to ask yourself if you SHOULD do it?”

Innovation and creativity are important. I’m just not sure that developing the means to scalp $.00001 per share in 10 milliseconds counts as innovative or creative. And the benefits to the majority of the market participants seem awful puny compared to the risk forced on the rest of us.

Have any of these outfits ever gone out of business? Surely HFT isn’t like Lake Woebegone where everyone’s above average.

figleaf October 17, 2011 at 12:49 pm

Alex, I disagree with you often enough to be a bit anxious about agreeing with you here. But this seems like a very good post.

If the only measure of success in the stock market is narrowing ask-bid spreads then HFT is clearly a ringing success. But only if.

Pretty much by definition rounding errors occur only to the right of the last significant digit. That in theory HTF is basically about making money by pushing further and further to the right of that last significant digit then in theory there shouldn’t be a problem. Heh.

Again, nice post.

figleaf

Jason Ruspini October 17, 2011 at 2:08 pm

Could it be that volatility has sharply increased mainly because we are in a depression (that economists don’t seem to be able to explain)? Look, there are some abuses with HFT, and it could increase volatility, but the main problem for most investors is simply that stocks have not gone up for 10 years!

Stop posturing, and join Tyler in trying to explain why that might be, e.g. rate of technological change, rate of globalization, demographics.

And by the way professor, the Occupy Wall St bell tolls for thee. If there were some kind of prediction market for college tuitions, the price would be breaking down right now.

albatross October 18, 2011 at 10:09 am

Do the states that let you prepay in-state tuition release figures for how many parents have signed up each year? That’s at least somewhat related to a prediction market for college tuitions.

buddyglass October 17, 2011 at 8:41 pm

Two thoughts:

1. Supposing HFT does increase liquidity, at what cost? We’ve already mentioned volatility. What about the cost in terms of human capital? It’s like an arms race between investors. A ton of money and effort and intellectual capital is expended in order to out-gun the competition. I wonder if a sort of “mutual disarmament” (forced by the introduction of a tax on transactions) would pay dividends in the form of redirecting this effort somewhere more constructive.

2. While a tax on transactions might kill HFT and thereby reduce liquidity, the revenue it generates could then be used to replace other taxes. It seems like this might represent a net win, depending on what tax(s) were replaced.

3. If added volatility has the effect of creating both more “big winners” and “big losers” (which might not be the case) then this strikes me as a net loss, considering the goals of many small investors. For instance, retirement savings.

buddyglass October 17, 2011 at 8:42 pm

Okay, three.

jdm October 17, 2011 at 9:15 pm

Leaving aside that it is far from certain that the increase in volatility in equity markets since 2007 is caused primarily by the growth of HFT, what specific changes would you advocate making to the rules of equity trading or equity market structure? As a side note, there are a number of markets around the world where for one reason or another it is difficult or impossible to do HFT – markets like those in India, Taiwan, and China. Are these markets less volatile than markets in comparable countries (say Hong Kong) where HFT is possible?

MatC October 18, 2011 at 7:25 am

Obviously and rightly the systemic effects are what are concerning us here, but we shouldn’t leave out the possibility of the investment of resources in this type of financial invention being a fixed sum arms race with, from a societal point of view, negative returns.

Assume the people funding HFT outfits (funding the computing power, the labour costs etc.) are self-interestedly income-maximizing rational in that doing so pays off by their being able to better predict (or more quickly predict) price fluctuations. They make money, which at least covers costs. This comes, straightforwardly, from those who invested less in HFT or not at all. But overall, having competing groups invest more and more in HFT will mean that these resources are funded from those who do so less successfully or not at all. You can model this as an externality, but it’s simply a feature of any situation where the more successful prediction of future price changes carries financial rewards but requires spending resources in competition with others doing the same thing.

This observation is not an argument for banning HFT by the way.

MatC October 18, 2011 at 9:19 am

Sorry, didn’t see Buddy Gass above here on point 1. Same point I think.

albatross October 18, 2011 at 10:13 am

The big question to my mind isn’t whether there is an externality on other market participants because of increased volatility or whatever, but rather whether HFT introduces some new low-probability disasters that will in practice be very hard or impossible to fix. To the extent you can have a sudden crash and then unwind the trades and “make it didn’t happen,” the risk seems pretty manageable. Are there other ways the interacting network of HFT programs could wreck things, which would be very hard or impossible to undo?

Anon October 18, 2011 at 12:31 pm

I think you non-investors have forgotten just how high bid/ask spreads were for regular old stocks back in the day (before 2004). IIRC, 1/8 was not unheard of, 1/16 more likely.

This is a huge hit for a retirement account buying or selling 25,000 shares of a $10 stock — $3k or $1.5k. Sure, it looks small percentage-wise over the course of holding that stock for a few years, but that’s real money to me.

How much of the bid/ask collapse is due to HFT? I suppose that depends on how you define HFT. But I assure you, the basic idea of HFT goes back as long as markets have existed and those strategies have cost the market-makers money for just as long.

A skeptic October 18, 2011 at 4:21 pm

I am skeptical of any long term benefits of HFT. If anybody thinks Wall Street took a hit in commissions from HFT they are kidding themselves. More likely, via steering shares Wall Street is front running their clients and making money not from commissions but from the price of the security, which is not being efficiently determined for the average buy-and-hold investor. HFT is thus distorting the price of a security. Tax them and nobody save the HFTs will care much, as commissions are rarely that important to retail investors post the Big Bang and deregulation of commissions.

Paul Brassey October 18, 2011 at 6:06 pm

As an individual, Joe Shmoe investor in mutual funds for retirement, all of the investment advice I receive, whether in prospectuses or from financial advisers, has to do with short-to-medium term economic forecasts. None of it, and I mean zip zero nada, refers to management of millions of transactions in a second. It’s impossible for the lay investor to know whether his/her mutual fund manager has the best or the worst HFT algorithm. Whereas the advisers and pundits talk about economic forecasts, the actual market looks like a rigged casino.

Chris October 19, 2011 at 12:25 pm

So volatility has risen over the last years? Hardly a surprise since we have been (and are) going through the most severe financial crises in living memory. But of course a culprit must be found, and HFT it is.
I refer to the excellent work from Mr. Narang. He concluded that from 2000-2006 the average movement of the S&P500 was 0.83%, while from 2007 onwards it has been 1.06%. However, most of this rise – roughly two-thirds – is contributable to overnight volatility. I.e. from the close to the open the next day, when no trading – in particular HFT – is taking place.

Comments on this entry are closed.

Previous post:

Next post: