Confidence and Fear Why Quantitative Models Win

Confidence and Fear…  two factors that are strong motivators. When it comes to the quantitative world, traders and quants are looking for the confidence that their strategies will turn a profit and are fearful that what they deploy into production could become headline news as the latest rogue algo to wreak havoc.  Of course those quant firms employ both advanced technology and human capital to have the confidence in achieving alpha and to minimize the fear.

I recently attended the Battle of the Quants conference where I was both an attendee and a panelist. While this conference had that Battle title, overall I think there was more consensus than confrontation among the panelists and audience.  The afternoon’s main battle was a panel that pitted a quant team dedicated to automated algorithms against a team that (presumably) considers human discretionary decision making as a better tool for alpha. In other words, like the Jeopardy challenge of IBM’s Watson, it was human vs. machine. That particular event I was completely fascinated by.  An interesting pre-game commentary relates the Jeopardy match to the Singularity by author Ray Kurzwell, a convergence of human and machines.  In the battle pitting algo’s against humans the outcome was decided by an audience vote.  The voting was not simply two choices: “for the machines” or “for humans”, but a third choice was offered more aligned with Ray Kurzwell’s Singularity – “a combination of human and machine decision making”.  As you might have guessed, that third choice was the overwhelming favorite. I believe the majority have the confidence to let machines decide many things but are wanting of human intuition or that proverbial finger on the button as a measure of risk control so fear does not overwhelm.

At the Battle, I was a co-panelist on the topic of High Frequency Trading and Quant Strategies. Myself and the distinguished cast of co-panelists, CEO’s and Directors of numerous Quant firms were guided by a series of questions that aimed to educate the audience on HFT, what makes it unique among automated trading, the types of strategies employed by firms including those on the panel itself, the influence of current market structures (i.e. equities vs. futures vs. FX) and the impact of coming regulation.  There was clearly consensus on many of the strategy types especially arbitrage models that seek to benefit from market liquidity imbalances/price inefficiencies across the fragmented venues, both within an asset class (i.e. equities) and across them (equities to futures & options).

The CFTC-SEC Advisory Committee report wasn’t released before our HFT panel session; it would have been the fodder for more detailed discussion on regulation’s impact.  One motivation behind the document is clearly high frequency trading and its impact on volatility. This document outlined 14 points or considerations as focal points for the SEC and CFTC. From the research conducted, they mention 2 major themes: 1) liquidity problems arise as the result of automated algorithms 2) the exchange markets are highly inter-connected exacerbating volatility.   But rather than condemn HFT or other types of algorithmic trading the underlying message in the report is that automated, low latency is here to stay so let’s find ways to live within this next evolutionary step in our markets.

Of the 14 recommendations many provide refinements to existing controls. An example is the single-stock circuit break rule in existence since the May 6th event, the suggestion is to expand its reach to all but the most inactively traded securities but integrate a limit up/down rule and increase the pause period beyond 5 seconds contingent upon the market’s reaction to the up/down rule’s ability to stabilize contra-side liquidity.  A more complex definition of a single-stock trading halt, but likely an improvement that will mitigate mini-crashes.

Three of the recommendations are designed to reduce market volatility and point directly at HFT.

8.  Benefits for ‘peak load’ pricing models.   This advisement is designed to provide incentives for market participants on the “maker” side to stay in market during volatile times by offering them increased rebates. HFT firms already leverage the Exchange rebate model via Passive Rebate Arbitrage. The fee structures from the exchanges allow High Frequency Traders to remain profitable on trades where their net take is limited to the exchange’s passive rebate.   The ‘peak load’ pricing model is designed further encourage that incentive.

9.  Incentives or regulations for market making.  This topic is likely considered most controversial, specifically the role of HFT in liquidity provisioning and how to obligate or incent HF firms to stay in the market during times of extreme volatility. In the end, the advisory committee suggested the SEC/CFTC should do something but make no specific recommendations. I guess I shouldn’t be surprised; this is a very difficult topic to tackle.   They’re probably thinking all the other points designed to reduce market volatility through trading pauses and rebates may be enough incentive and be the best first steps to take. There is no magic bullet and defining specific obligations and understanding their side-effect is an enormously difficult task.

10.  Fees for high levels of Order Cancellations. This suggested ruling imposes a tax on market participants for order cancellations. However, as Themis Trading points out there is a real subtly to the language used. While participants should absorb the cost of their activity it wouldn’t be a simple fixed fee, but imposed based on deviations to past behavior. I recently read a research paper on Low Latency Trading by Joel Hasbrouck and Gideon Saar. The authors’ goal was to determine the impact of HFT on market quality (where market quality is defined as liquidity and short-term volatility). The authors clearly defined their objectives, outlined the criteria to measure, defined the measurements and showed the results.  Their research did discover numerous cases of high rates of order submissions immediately followed by order cancellations. I suspect these are triggering strategies attempting to detect market liquidity. Often called order-anticipation strategies designed to seek out the typical (large) institutional order. The conclusion of the research was that HFT improved market quality by providing increased liquidity and lower volatility.

Overall the Advisory report provides numerous suggestions for the SEC and CFTC to focus on. They do so without condemnation of any style of trading such as HFT. I think it’s a realization that automated low-latency strategies are here to stay. Regulation’s role is to assure all participants whether human or machine that Exchange markets are fair and equitable.

Once again thanks for reading.
Louis Lovas

For an occasional opinion or commentary on technology in Capital Markets you can follow me on  twitter, here.

About Louis Lovas

Director of Solutions, OneMarketData, with over 20 years of experience in developing cutting edge solutions and a leading voice in technology and trends in the Capital Markets industry.
This entry was posted in Algorithmic Trading, Analytics, Equities, Foreign Exchange, Futures and Options, HFT, HFT Regulation, High Frequency Trading, OneMarketData, OneTick. Bookmark the permalink.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s