Participation Trophies, IEX again, & Machine Augmented Intelligence

In Case You’ve Wondered WHY the Buy-Side Doesn’t Really Measure Trading Skill…

It is well known that buy side trading desks use both algorithms and benchmarks based on “participation” such as VWAP, TWAP, PWP, “Participate don’t Initiate”, etc.  To any rational outside observer, such methods seem strange, as they are designed to be the same, no matter what broker one uses.  As I described in ViableMkts most recent piece, there is a reason, but, the use of such benchmarks comes with a cost that goes a long way towards explaining why so many active managers underperform.  To read the article, download it here:  http://www.viablemkts.com/download-if-every-trader-gets-a-trophy/

 

IEX is at it again…

OK, this one is not their fault, since they are tweeting about BATS data on effective spreads.  In this case, they highlight the fact that IEX has the lowest effective spread, measured by trades, among exchanges.  As I have noted previously, this analysis is very misleading as it essentially measures the dark pool performance of IEX and compares them to exchanges that do most of their trading based on displayed quotes.  Since the link IEX tweeted starts with AAPL, let’s look at data for Apple, to put this in perspective.  For AAPL for the entire month of May, according to data provided by May Street, IEX only set either the National Best Bid or National Best Offer roughly 0.17%.  That is not a typo.   IEX created the best bid or best offer less than all other exchanges (excluding Chicago, which never set the best price), and by over 2 full orders of magnitude less than either Nasdaq or the Bats group of exchanges.   In addition, IEX was only AT the NBBO 5.6% of the time in AAPL, which compares to 92% of the time at Nasdaq, 87% of the time at BATS and 86% of the time at ARCA.  So, while it is true that “effective spreads” are lower at IEX, that is the direct result of people sending orders to them “blind”, and, in particular, from orders sent to them probing for midpoint fills (as compared to orders that are willing to cross the spread).   The data makes this extremely clear.

According to data from BestXStats, in May, for ALL Marketable orders, IEX had an effective/quoted spread of 127.5 which compares poorly to all the other exchanges, notably ARCA at 89.7, Bats at 94.9, and Nasdaq at 106.6.  Of course, the reason for this is that they executed more shares away (via their router) than they did on their own exchange, which is the opposite of the other exchanges.  To reconcile these numbers with the Bats website, however, the difference is that IEX executed a much higher % of AAPL shares as a result of “inside the spread” orders according to 605 data.  In fact, IEX executed 48.6% of their AAPL shares from these orders, compared to just 6.29% from Bats and 4.38% from ARCA.   This points to an clear conclusion:  If one is instructing a SOR to probe at the midpoint, IEX is a worthwhile destination, compared to other exchanges.  If, however, one wants to trade aggressively or take liquidity from displayed quotes (which is what exchanges are built to do), then IEX, based on data, is an inferior destination…

Active Management is misunderstood 

Rob Shapiro published a well-written, entertaining piece on active management, that does not account for a “middle ground” where analysts can provide inputs to a machine augmented process.  He describes the efforts of a buy-side analyst to recommend a stock based on in-person meetings as well as in-depth data analysis and concludes that quantitative models have likely beaten him to it.    The analyst, based on his perceptions of nuanced behavior and deep understanding of value, may well have a better idea of the relative value of this company to its peers.  He is not, however, qualified to predict the aggregate valuation of the market, or to construct an optimal portfolio based on his analysis.  Simply put, if the analyst was willing and able to express his conclusions as a relative performance number, and the asset manager used that as an input to a quantitatively constructed portfolio, if Rob’s fictional analyst was good at his job, his input could provide significant incremental value.  (of course, time and data analysis of his predictions will tell…)   This idea was the motivation of my quantamental investing article, which boils down to a case for “machine augmented intelligence” as the way forward for the active management industry.

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