Yves here. Readers are likely to assume that the “broken market” of the headline is US housing related, say the private mortgage securitization market, but the subject is what once was the gold standard of trading markets, equities.
Index Universe has cited a study by the Tabb Group that finds that investor confidence in stock markets is even lower than in the period immediately following the flash crash of 2010. Back then, 53% of respondents had high or very high confidence in the markets, and only 15% weak or very weak confidence. As of its August 2012 survey, the number with positive views and negative views were equal, at 34%. This interview with Chris Sparrow, an expert on high frequency trading, describes why he thinks the market is now fundamentally flawed and what can be done to reform it.
By Paul Amery, editor of Index Universe. Cross posted from Index Universe
IndexUniverse.eu: Chris, what’s wrong with the current structure of equity markets?
Sparrow: The current market structure is fundamentally unfair, since different participants have unequal temporal access to information.
To put it simply, if other people can see that your order has been filled before you can, or that there’s been an update to a price quote before you are able to see it, you’ll lose confidence in the market.
This comes down to how price signals are propagated. Currently, no two participants receive price and quote information simultaneously. If you want to ensure fairness in the markets you need to level the playing field for everyone.
If we can fix this then we should be able to restore some confidence to the markets and trading volumes, which have declined dramatically in recent years, should recover.
IndexUniverse.eu: Is the fragmentation of equity market trading between different venues a concern?
Sparrow: I’m not concerned with fragmentation per se, as competition between trading venues is good. What’s missing is synchronisation and coordination.
Let’s take the air travel system as an analogy. Airlines want to minimise fuel use and so all have an incentive to land their planes first. If you allow that, you’re going to have lots of crashes. Regulators impose structure via an air traffic control system, reducing the systemic risk.
Such coordination simply doesn’t exist in the equity markets today and regulators tend to take a passive role, watching what happens and taking action after the event if something goes wrong.
I’m not for overregulation and in my opinion introducing competition between trading venues, which happened in the US under Regulation NMS in 2005 and in Europe following the introduction of MiFID in 2007, was a good thing, as I’ve said. But there have been some unintended consequences of these reforms that now need to be dealt with.
IndexUniverse.eu: Should exchanges be forced to go back to some kind of utility status from their current for-profit model?
Sparrow: I don’t think that’s necessary. I think the for-profit exchange model can continue to exist, but subject to a requirement for synchronisation.
Here’s another example. Let’s say I buy a solar panel and want to contribute electricity back to the grid. If I want to do that I have to supply the electricity at 60 Hz. I can’t unilaterally decide that I want to give the electricity back at 45 Hz.
There’s an infrastructure requirement that should be based on a policy of coordination. If I want to build an ATS (automated trading system) and plug it into the rest of the market grid I should have to do so in a standardised way. Otherwise the system will generate a huge amount of quote “noise” and introduce exploitable latencies that act to inhibit fair trading.
IndexUniverse.eu: What led you to identify the current trading system as a problem?
Sparrow: I’ve done a lot of work in transaction cost analysis (TCA). I was looking at a particular order and for purposes of comparison wanted a proxy for the Canadian equity market. For this I used the iShares S&P/TSX 60 ETF (TSE: XIU).
It turned out that there were a million quote updates in this ETF during a single trading day of 23,400 seconds. Why do we need so many updates? They impose significant storage requirements on everyone, take up significant network bandwidth and arguably do not contribute significantly to price discovery.
All the trading data that’s being produced is an externality on the whole market. We all have to buy bigger hard drives, bigger servers and network switches just to process the data, even though there’s little extra benefit in doing so.
IndexUniverse.eu: So what do you propose as an alternative?
Sparrow: My suggestion is the following. At a single point in time, which we call T0, you open up the order book and allow participants to enter orders, without anyone being allowed to see them. You and I can enter orders up to a second point in time, called T1, at any venue we choose, but we can’t see or react to each other’s activity.
At T1 (plus, in practice, a small delay to ensure the order entry session is closed) the trading venues try to match as many buy and sell orders as possible and to establish a single, market-wide clearing price. At that point all trades get printed and the residual state of the book, consisting of unmatched orders, gets published, as would the location of the unmatched orders. In other words, competition between trading venues could continue.
And by ensuring that matching orders get cleared, we do away with some of the problems that exist under the current system of continuous trading. For example, in the current structure it’s very easy for there to be “locked” markets, where the bid price at one trading venue equals the offer price at another venue, or “crossed” markets, where the bid at one venue exceeds the offer at another. These situations could be eliminated through co-ordination of the matching process.
IndexUniverse.eu: Who would coordinate the times at which this clearing process takes place, and how often would it occur?
Sparrow: I would suggest that regulators set the time standard to which exchanges and other trading systems must conform. And I think that the market clearing price could be set every second. Since the market would tick from clearing price to clearing price every second, from a larger time perspective trading would appear almost continuous, and that’s why I call this model the continuous call market.
IndexUniverse.eu: If the clearing price changes each second, how long does it take within each second for the price to be established and disseminated?
Sparrow: I think that this can happen within a couple of hundred milliseconds—that is, a fifth of a second—at most. In Canada, the geographic latencies between different data centres are in the order of single digit milliseconds.
IndexUniverse.eu: Would the system you’re proposing get rid of all abusive types of high-frequency trading?
Sparrow: No, not all. I think this system would get rid of the worst form of abuse—latency arbitrage—where individual participants can pay extra to exchanges to get a “first look” at price quotes. But you’d still have activities like “spoofing”, often called layering. The idea is to level the playing field with respect to the dissemination of information.
IndexUniverse.eu: What does “spoofing” mean?
Sparrow: Spoofing means that if I’m trying to buy a stock I enter a whole lot of offers I don’t really want to execute into the order book to make it look as though there’s selling pressure. This activity is designed to make other market participants nervous and to “cross the spread” and hit my bid for the stock.
IndexUniverse.eu: But that kind of thing could happen when traders were all physically located on an exchange floor and transacted face to face.
Sparrow: Exactly. You’re not going to get rid of all nefarious behaviour. But what I’d hope to do is remove is what I term “technological adverse selection”—where participants can pay up to be in a privileged position by comparison with everyone else. If you’re an HFT and I’m a traditional buy-side investor you can see and react to the market more quickly than I can and you have an advantage. Let’s remove that advantage.
Is the market about “the guy with the best technology wins”, or “here’s a mechanism to allow the transfer of capital”? I think it’s the latter and people should compete on a level playing field and to win based upon how smart they are, not because they’ve won the technological arms race.
Chris Sparrow’s proposal for reforms to market structure is outlined in more detail here