markets and prices
the connectivity revolution
HFT and dark pools
What is a market? It’s a place where individuals congregate in order to exploit the gains available from exchange.
The benefits of bringing a multiplicity of agents together are: to generate diversity in what is available for exchange, to facilitate competition, and to increase choice, and thus to optimise the gains from trade.
Is it necessary that the rates of exchange between different goods and money are the same for everyone, at any quantity — i.e. that there is at each moment a single price for a given good? Not at all. It’s known, for example, that allowing price discrimination can generate better (Pareto-superior) outcomes under certain circumstances. 
Indeed, in some cases it may be better if there is no price mechanism at all. The outcome of bargaining between pairs of agents, where the amount of money exchanged for a particular quantity of a good is agreed deal-by-deal, may be Pareto-superior to the equilibrium involving prices — if the market doesn’t comply with the textbook model of perfect competition, as most markets don’t. This becomes more relevant as search costs and other transaction costs approach zero.
What does always make a market more efficient is the presence of a high-quality infrastructure. The tenant of a shopping arcade (and his customers) will benefit from such things as a high degree of reliability of access to his premises at set times, the presence of crime-deterring features such as video cameras or security personnel, and so on.
Conversely, when the Nasdaq stock exchange suffers a computer glitch, so that trading in stocks such as Microsoft becomes impossible for several hours — as has happened several times over the past couple of years — this severely reduces the reliability, and hence the efficiency, of that market.
The benefits of trade are often confused with the quite separate notion of fairness. Whether a particular kind of contract or trade between two parties should be regarded as ‘unfair’ is a separate question from whether it is efficient (i.e. maximally improving the joint position of the two parties).
The issue of fairness usually revolves around how the gains from trade are split. For example, there has for some years been concern over whether powerful retail chains are able to extract most of the gains from trade with producers, generating profits for the chains’ shareholders while leaving their suppliers making barely more than the minimum.
In addition to such concerns, there may be qualms over whether different agents transacting in identical situations face identical opportunities. The principle that all investors wishing to (say) buy shares in Apple at any given moment should be able to do so at the same price falls into this category. People are used to thinking of the price of something as having a single value, but it doesn’t follow that price homogeneity is necessarily more conducive to generating an overall ‘fair’ outcome, in the sense usually meant. Nor is price homogeneity essential for efficiency, as we noted above.
Until recently, most quoted shares were in the position of having almost all their trade carried out on a single stock exchange and at a uniform price, visible and available to everyone. What are the benefits of this? There are no doubt efficiency gains, but on the other hand (as in the case of prohibiting price discrimination) there may conceivably be costs as well. While the established system seems a sensible way to go on, one shouldn’t blithely assume that deviations from convention are necessarily to be abhorred.
To reiterate: a single, exactly identical price faced by everyone at any given moment is a possible, but not necessary, aspect of an efficient market.
The revolution in information technology is having a major impact on the way trading works, particularly in homogeneous goods such as financial instruments. This is leading to the disruption and recreation of certain kinds of market. Markets are increasingly globalised, interconnected and astonishingly quick to respond to one another.
A significant positive change in technology, such as the connectivity revolution, means there is a new, Pareto-superior equilibrium available to an economy. One of the interesting things about the way that an economy moves to a new equilibrium is that it may proceed via a series of steps, not all of which may appear to be (and some of which may not be) efficiency-improving.
Once an existing setup is out of kilter with a new optimal equilibrium, distortions are liable to exist which are capable of generating profit opportunities. Exploitation of those opportunities may not appear to be beneficial to anyone other than the exploiters, but such exploitation is nevertheless part of the process by which one gets to the desirable new equilibrium. If there weren’t such opportunities or the desire to exploit them, there wouldn’t necessarily be any forces that would take one, relatively quickly, to the new equilibrium.
So, for example, the creation of a new stock exchange in London (Chi-X) may, as has been argued, have depended partly on the desire to exploit the kind of timing advantages often complained about in relation to high-speed trading. Nevertheless, the existence of a multiplicity of exchanges, in place of former monopolies, may represent either an aspect of the new equilibrium, implied by the change in available technology, or a transitional stage en route to the new equilibrium.
In the case of equity markets, there is now widespread use of automated trading strategies which are capable of generating rapid feedback loops in a way that was not possible a few decades ago.
Moreover, the information that is available sufficiently quickly to be useful seems to have reached a timescale where a new kind of strategy is possible. Because it reflects, in a more timely way, the dynamics of intraday decisions, the information accessible by those who have the right technology seems to allow for a greater level of prediction — and potentially manipulation — of investor psychology than was previously feasible.
Traders who react sufficiently quickly to the information they get may be able to predict what other players are going to do in the next few seconds or even milliseconds. If they can predict how investors would react to certain small changes, they may be able to trigger behaviour in others by making small catalytic transactions of their own  or by placing orders that are quickly removed again and not intended to be executed.
Possibly this is only a temporary phenomenon, generated by the leap in technology. Once the investment community in general becomes aware of the strategies being used to predict and influence, counter-strategies will develop that will render many of the current techniques of high-speed trading obsolete. This is consistent with data pointing to a decline in the high frequency trading sector. According to a Business Week article, such trading generated US$5 billion in revenue during 2009, but only 1 billion in 2012.
High frequency trading (HFT) — the name given to some of the strategies made possible by the new information speeds — is technologically dazzling but a bit mysterious, which may be why it has lent itself to becoming a scapegoat for various market complaints. For example, the flash crash of 2010, in which the aggregate value of the US stockmarket dropped about 6 per cent in the space of five minutes, is often blamed on HFT.
The fact that the people making money from HFT are mostly individualistic outfits — rather than large, respectable institutions — may also contribute to the hatred the business seems to inspire in certain critics, including several celebrity economists. But there is more than a whiff of hysteria about the hostility towards HFT. The intuition seems to be: the business doesn’t “add value” to the operation of the stockmarket, therefore it must be doing harm. However, the opinions one reads often seem more based on prejudice than on data.
A PricewaterhouseCoopers report on the issue, for example, manages to stick to the facts — until it suddenly starts presenting theory as if it were generally accepted wisdom. Noting that the so-called Twitter Flash Crash was probably caused by the use of algorithms, the report goes on to assert, apparently as an indubitable proposition, that the presence of HFT
means that when such sell-offs are initiated, high-speed firms are driven into action, exacerbating and accelerating declines.
In fact, it is still not clear what the key factors were that led to the Twitter Crash or the more disturbing 2010 crash, nor to what extent HFT (as opposed to automated trading in general) exacerbated what happened. I believe the system-weakness explanation has not received nearly enough attention.
An analysis of trading on the exchanges during the moments immediately prior to the flash crash reveals technical glitches in the reporting of prices on the New York Stock Exchange [...] that might have contributed to the drying up of liquidity.
According to this theory, technical problems at the NYSE led to delays as long as five minutes in NYSE quotes being reported [...]
At the same time, there were errors in the prices of some stocks (Apple, Sothebys, and some ETFs). 
Extreme levels of volatility, it should be noted, appear to be possible even in the absence of such intensive automation as we have at present. The October 1987 crash happened in the very early days of computerised trading, yet was in some ways more extreme (a 20% drop in stock indices in the space of a few hours) than anything seen since, including the 2008 meltdown.
Have the financial markets changed in some way that makes them more prone to instability? Anecdotally, it’s tempting to believe that is the case.
Apart from computerisation, globalisation, and increasing mathematical analysis, the main noticeable difference between the financial sector now and thirty years ago is that there’s less of a spirit of individual responsibility and diligence, though paradoxically this is paralleled by a greater superficial obsession with risk avoidance. Collectively, there may be more willingness to throw new regulations at a problem — typically after the event — but individually, there is more tendency to assume someone else (whether human being, or computer) will deal with any issues that may arise. One contributory factor may be that the market is less driven by individuals, and more by institutions, who are playing with other people’s money, not their own.
The preferred attitude seems to be “no worries”, let automation take care of it. This can easily create scope for glitches, whether on a macroscopic scale (major shutdowns) or a microscopic one (general day-to-day volatility).
One of the lessons of the sub-prime crisis is surely that the ethos in financial institutions has fundamentally changed. The new casual, gung-ho approach expresses itself in various ways. To implement a new model that is prima facie successful, without getting too hung up about possible defects or the risk of exceptional circumstances generating unforeseen problems, is now often regarded as a “no brainer”. If an automation or other technological improvement suggests itself, there is a greater tendency to say, “let’s go ahead and not think about it too much; anything which saves human effort has to be a good thing”.
Nicholas Carr, writing in The Atlantic about automation and plane crashes, says that
Pilots today work inside what they call “glass cockpits”. The old analog dials and gauges are mostly gone. They’ve been replaced by banks of digital displays. Automation has become so sophisticated that on a typical passenger flight, a human pilot holds the controls for a grand total of just three minutes. [...]
And that, many aviation and automation experts have concluded, is a problem. Overuse of automation erodes pilots’ expertise and dulls their reflexes, leading to [...] “a de-skilling of the crew”.
The problem may go beyond “de-skilling”. There may be “de-motivation” which, together with a more relaxed attitude towards detail, can lead to “de-responsibility”. The same factors may, a fortiori, apply to the financial services industry.
So-called dark liquidity is a contemporary feature of stock trading that has to risen to prominence about the same time as HFT, though it’s a separate phenomenon. A dark pool is essentially a small, private stock exchange, usually run by a financial institution such as Barclays or Goldman Sachs, via which clients can trade shares without the details of their transactions being made public.
The existence of dark pools may reflect the greater ease with which the infrastructure for such exchanges can be set up these days. What is less clear is why their share of total trading volumes has been increasing.
The most common explanation is that dark pools allow buyers and sellers to escape the scalping and manipulation techniques of HFT to which those dealing via conventional exchanges are now prone. On the other hand, it seems some financial institutions have allowed HFT traders access to their dark pools, so this may not be the whole story. It’s certainly not safe to assume that the explanation given by the institutions for their pools is necessarily the only, or principal, reason they are running them.
A more worrying possibility is that dark pools reflect a loss of confidence (or anticipation of problems to come) with regard to conventional exchanges. If you no longer feel the established systems can be trusted, you are more likely to want to do it for yourself.
Once a good thing — such as a market — has been generated by individuals, there is a tendency for it to get appropriated by the collective, and for commentators to start boggling over whether it should be permitted to change.
This can creep into discussions almost unnoticed. Thus a recent article asserts that whether dark pools are “needed” is “open to debate”. But clearly they are needed, by at least some participants, or they wouldn’t have arisen. It may of course be claimed that they have sufficiently harmful effects to consider intervening. Asking whether something that’s happening is needed, however — even if this is used as shorthand for something else — is tendentious. It implies the existence of objective outside observers who are somehow more capable of assessing participators’ interests than the participators themselves.
Michael Lewis’s book Flash Boys, published earlier this year, is
the explosive story of how one group of ingenious oddballs and misfits set out to expose market abuses on Wall Street.The supposed market abuses concern the new types of high-speed activity such as HFT which were made possible by computerisation, and particularly by the use of fibre-optic links that have shortened communication times between traders and exchanges to milliseconds.
According to Lewis, the result of such activities is that the prices investors see on their screens are no longer a meaningful representation of the prices they can actually trade at, or of the ‘true’ prices of the stocks in question; and that different investors face different prices depending on the way they access the market. We are mostly talking about very small differences — Apple at $98.05 versus Apple at $98.10, for instance. Nevertheless, such discrepancies are supposedly undermining the fundamental principles of stockmarket efficiency and equality.
Flash Boys is not merely a popular bestseller but appears to have become influential in determining government policy. It has provided fuel for the EU’s proposed transaction tax, and seems to be affecting the SEC’s plans for new US legislation. Perhaps it was also the inspiration for Attorney General Eric Schneiderman’s recent suit against Barclays. It contains some interesting background facts, particularly about the history of high-speed connections, but does it deserve to be treated as a serious guide to what’s wrong with equity markets?
The book tries hard to be a rollicking read in spite of the technicality of its subject, coming across like a movie script co-written by John Grisham and Quentin Tarantino, though the sweary truculence of the characters gets a little tiresome.
Ronan brought in oversized maps of New Jersey showing the fiber-optic networks built by telecom companies. [...]
When he unrolled his first map, a guy who worked in RBC’s network support team burst out, “How the fuck did you get those? They’re telecom property! They’re proprietary!”
Ronan explained, “When they said they wouldn’t give them to me because they were proprietary, I said, ‘Well, then, proprietarily fuck off.’” 
While it’s informational in parts, the main purpose of the book seems to be the message: The markets are rigged, man. You (Honest Joe) are getting fleeced — a theme reinforced by the racy style and conspiracy-theory tone.
The story told makes satisfying sense, but only if one presumes what Lewis ostensibly sets out to prove: that HFT is inherently dodgy, and that it’s responsible for increased market instability such as flash crashes. Also, that various financial institutions are in league to rip off investors by means of devices such as HFT and dark pools.
In making its case, the book mixes up various issues without much attempt to disentangle them:
- whether HFT makes use of arbitrage opportunities that oughtn’t to exist;
- whether HFT creates opportunities for types of trading that are already considered illegal, e.g. front running;
- whether HFT technology provides timing advantages that are somehow morally wrong, e.g. by being as close as possible physically to the source of the information, including inside the same premises (“co-location”);
- whether dark pools aren’t sufficiently transparent, and therefore shouldn’t exist;
- whether financial institutions don’t care if their clients are getting a raw deal;
- whether financial institutions try to exploit their clients, via HFT or otherwise.
Lewis lists three principal HFT activities that are said to be “grotesquely unfair”. By far the most widespread one, he notes, is what is referred to as “slow market arbitrage”:
when a high-frequency trader was able to see the price of a stock change on one exchange, and pick off orders sitting on other exchanges, before the exchanges were able to react. 
This sounds remarkably like conventional arbitrage — exploiting short-term price differences between two sources by buying at the lower price and selling at the higher — generally regarded as an essential component of a liquid market. Precisely why this particular sort should be viewed as grossly unfair is not explained. Yet the new stock exchange set up by the book’s protagonists, one purpose of which was to prevent such arbitrage from being possible, is presented as being somehow morally superior.
[IEX] did not ban but welcomed high-frequency traders who wished to trade on it. If high-frequency traders performed a valuable service in the financial markets, they should still do so, after their unfair advantages had been eliminated. 
To complain that there are people ‘unfairly’ taking advantage of small discrepancies in a market, and to argue they are therefore spoiling the market, is to misunderstand what a market does. A market is an equilibrating mechanism designed to match demand and supply in the most efficient way possible. The most liquid markets require players (e.g. market makers) who will risk their own money on predicting short-term price movements.
In order for a market to work in the way envisioned by traditional models, any discrepancies need to be quickly – ideally, instantaneously – eliminated. This is what arbitrage does, and why arbitrage is useful. If a new type of discrepancy crops up due to a change in technology, which was not previously being arbitraged away, then the appearance of a new bunch of arbitrageurs who trade away the new discrepancy (making money in the process – why else would they do it?) is to be welcomed. In that respect, HFT-ers are clearly a plus, not a minus, in the equation.
As to whether the market is “rigged”, the possible distortions that Flash Boys examines seem minute by comparison with the more serious manipulation that is known to have gone on in currency and interest rate markets, and which may yet turn out to be a feature of equity markets as well.
Manipulation by large financial institutions, and by government agencies, is likely to produce a more severe mismatch between price and value — and eventually a far more damaging effect on the position of ordinary investors — than anything discussed in Michael Lewis’s book, even if it would be harder to turn this into dramatic narrative for popular consumption.
Good old Simpsons. Next week will be their 25th anniversary. (The first full episode was aired on Fox in 1989.)
They are of course well past their peak. Most of the classic episodes occurred circa 1992 to 1998. Wry observation, served up with benign detachment, has been replaced by knowingness and a degree of sneer, giving a somewhat degraded flavour. In this, the programme is to some extent merely reflecting changes in the behavioural norms of America, and those of its cultural satellites e.g. Britain.
The original subtly witty and subversive flavour has more or less gone. The tone now often seems authoritative rather than speculative. And by the time former US Attorney Generals appear as guest stars, you know that a series has made the transition from left-field to establishment.
Still, The Simpsons is probably still the best available in animated humour. Futurama, Family Guy and American Dad are all amusing and diverting in their way (Roger the alien surely deserves a place in the cartoon character hall of fame) but they don’t have the same feeling for social reality, the sense that “yes, that’s how things are”.
Look at any small community and you will see most of the Simpsons characters, though they may require a bit of deciphering. The do-gooding preacher and his gossipy wife may be the couple running the local shop; the big-haired, rock-loving bus driver may be a small-pond media figure. Nevertheless, the ridiculousness of ‘normal’ everyday people is well caught in terms of archetypes whom you can probably identify, in some form or other, wherever you go on the planet.
Me, I tend to think of as somewhere between Ned, Monty and Homer — though when I took a “Which Simpsons character are you?” online test a few years ago, I was informed that I was like Lisa.
Congratulations, the character you most resemble is Lisa Simpson!
You are principled and loyal. For your friends, you would walk to the ends of the earth.
But woe to anyone who pushes you too far. They are in for a world of pain!
This site is a publicity device, not a sample. Web-based commentary is not a substitute for mainstream publishing from the position of an academic post. Oxford Forum is seeking financial backing to enable its members to become productive intellectuals.
Modern academia is not well suited to making major progress, which requires the capacity to depart from received wisdom. Those who can succeed on its terms may be clever by some definitions, but are unlikely to be capable of moving beyond the established paradigms.
Billionaires should regard it as a duty to support individual innovators, including those with unfashionable viewpoints forced to operate outside the system – irrespective of whether they agree with them.
1. Outcome B is Pareto-superior to outcome A if, under B, at least one person is better off and no one is worse off. See Wikipedia on Pareto efficiency.
2. Attempting to manipulate the price in these ways may fall foul of pre-existing legislation on market manipulation, e.g. with regard to bear raids. Whether on balance anything is gained by having such legislation is another matter.
3. From the Wikipedia article on the 2010 Flash Crash.
4. Flash Boys, chapter 3.
“Ronan” is Ronan Ryan, one of the co-founders of the IEX quasi stock exchange.
5. Flash Boys, chapter 6.
6. Flash Boys, chapter 6.
Strictly speaking, IEX is (at the time of writing) a dark pool rather than a stock exchange proper.