Last December, when we alerted readers to the shift in the “laser” tower at the “New York” Stock Exchange located in Mahwah, NJ…
… we wondered, rhetorically, “what would happen if one flies, say, a drone in front of one or all of those lasers during, say, peak market hours or, heaven forbid, just a few milliseconds before the Fed announces its next “most important ever” policy decision.”
We now have an answer, and it confirms what we have said all along since 2009 when we warned that the confluence of HFT and central bankers nationalizing markets, would destroy not only capital markets, but price discovery (and lead to such amusing Bloomberg articles as “Hedge Fund Managers Struggle to Master Their Miserable New World“).
In a fascinating study carried out by Andriy Shkilko and Konstantin Sokolov of Wilfrid Laurier University, the duo tried to answer one question:
“In modern markets, trading firms spend generously to gain a speed advantage over their rivals. The marketplace that results from this rivalry is characterized by speed differentials, whereby some traders are faster than others. Is such a marketplace optimal?”
Said otherwise, is an HFT-dominated market the best possible outcome, or – as the WSJ put it – “can a rainstorm in Cleveland make markets more liquid?” and if not, what are the alternatives?
The thesis is simple: since high frequency traders use ultrafast links to relay prices and other information between Chicago and New York, using either microwave or laser signals, one theory that can be empirically tested is what happens when the ultrafast signal is impaired due to weather.
To be sure, we live in an age when market tiering has never been as vast as it is now, with some HFT traders enjoying access to the fastest, laser-based, trading technologies; others – less fortunate ones – still use microwaves; and virtually everyone else (including the poorest HFTs) resorting to fiber optics, which while slowest, is also the most dependable and is not impacted by such “outlier” events as heavy rain, thunderstorms or snow.
So, since microwave transmissions are disrupted by water droplets and snowflakes, during heavy storms traders using the networks switch to fiber. This is what Shkilko and Sokolov used to test their hypothesis: using weather-station data from along the microwaves’ paths, they determined when storms occurred and then looked at what happened to bid-ask spreads in a variety of securities during those periods.
In other words, the two academics set out to find: whether, or rather weather HFTs loss of trading superiority has an impact on the overall market, and if so, what is it. As they write “to answer this question, we study a series of exogenous weather-related episodes that temporarily remove speed advantages of the fastest traders by disrupting their microwave networks.”
Their finding is stunning:
“During episodes [of heavy rain and snow], adverse selection declines accompanied by improved liquidity and reduced volatility. Liquidity improvement is larger than the decline in adverse selection consistent with the emergence of latent liquidity and enhanced competition among liquidity suppliers. The results are confirmed in an event-study setting, whereby a new business model adopted by one of the technology providers reduces speed differentials among traders, resulting in liquidity improvements.”
As for the conclusion, it is a blow to HFT advocates everywhere: the slowing down of the fastest high-frequency traders improved market liquidity.
This is how they put it:
This study examines the effects of speed differentials on liquidity. During our sample period, microwave networks stretched from Chicago to New York allow for the fastest information transmission and are only available to select trading firms. When it rains or snows in the area between the two cities, the networks are disrupted because rain droplets and snowflakes block the microwave paths. With the networks temporarily down, information transmission falls back onto the fiber-optic cable – a more reliable, yet slower transmission medium – effectively eliminating the speed advantages of the fastest traders. We show that when this happens, adverse selection and trading costs decline. This result is consistent with predictions of theory models that show that speed differentials among traders may be associated with lower liquidity.
That, in itself, is an amazing observation: that the more tiered and fragmented the market, the less liquidity, as not only do HFTs not provide liquidity, but soak it up, but also that those who traditionally provide liquidity, such as a conventional market makers, refuse to do so in a time when microwave (and laser)-based HFT scalpers and parasites are present, resulting in a market with no depth.
The conclusion continues:
The results shed light on latent liquidity. We show that when speed differentials among traders decline due to precipitation, the emergence of latent liquidity narrows spreads more than one would expect based only on the decline in adverse selection. We also find that in assets where spread reductions are not possible due to the binding tick size, latent liquidity improves quoted depths.
Our results are confirmed in an event-study setting. In winter of 2012-2013, one of the technology providers democratized microwave transmissions by introducing a new business model. Instead of selling bandwidth on its network, the firm began selling information on both sides of the Chicago-New York corridor. This one-time event had positive consequences for market quality similar to precipitation-related network disruptions. This result further supports the claim that the technological race that leads to a market with speed differentials may be suboptimal for market quality.
The technological race continues to drive spending in the trading industry. A recent example is a new data transmission tower proposed by the telecommunications company Vigilant Global to connect the U.K. and European markets. The tower will be among the tallest structures in the U.K. and will rival the height of the Eiffel Tower. It will provide trading firms with a completely unobstructed optical and radio line of sight, never previously offered in Europe, increasing signal transmission speed. In the meantime, traders in the U.S. have been switching from microwave transmissions to more reliable, yet costly, laser links. Our findings shed light on the possible consequences of these developments.
Another way of putting it: heavy rainfall literally adds to market liquidity, for one simple reason: it reduces the overall adverse influence of HFTs, predating market orders, and soaking up what little liquidity is left.
The implications are fascinating: if one were so inclined, the market could once again return to its more stable, “liquid” state if HFTs – which it has now been showned definitively do not provide or in any way add to market liquidity – were eliminated. Alternatively, if one were so inclined, tearing down the microwave towers linking Chicago to NJ and NYC, would achieve the same effect. Which, of course, should not be interpreted as call to arms against microwave – or laser – towers in order to restore normalcy to the market. After all, even if one eliminates HFTs, there are still central bankers whose impact on markets is just as dire and profound.
* * *