Implications: 2019 Payments For Order Flow Flat vs. 2018 View Larger Image “Historians study the past not in order to repeat it, but in order to be liberated from it.” – Yuval Noah Harari, Homo Deus: A Brief History of Tomorrow With three quarters worth of financial reports for calendar 2019 long in the bag, it is not much of a courageous leap for us to deliver an estimate for order routing revenue – otherwise more notoriously known as payment for order flow (PFOF) – for the full year. And, with the quarterly earnings season coming in the month ahead, it won’t be long before we are able to test the accuracy of this estimate. In the chart below, Alphacution extends our prior analysis not only to include 2011 and 2012 but also, more relevantly, to include the year just completed; thereby extending to nine years from six our focus on five of the primary players in retail order flow for US equity markets who also disclose order routing data: TD Ameritrade (soon to be acquired by Charles Schwab); E*Trade; the aforementioned Charles Schwab; the disruptive upstart, Robinhood; and Interactive Brokers. Here, we note that the year-over-year comparison – 2018 to 2019 – is essentially estimated to be flat. Now, while this new information may be mildly illuminating for some, it turns out not to be the central thrust of this Feed post. The more profound point – and the greater potential to expand your thinking on the matter – is in using this relatively benign figure to begin to demonstrate the interconnectedness of market actors and the chain reaction of factors that influence the strategic movements of those market actors (who are typically in a persistent hunt to defend and scale risk-adjusted returns harvested from a portfolio of adjacent opportunities). Consider that the capacity of these opportunities can be estimated (for any point in time) by the product of at least four general factors: inventory, liquidity, volatility, and information asymmetry. Note: Human, technical and financial capital are not direct capacity-defining factors; they are direct alpha extraction factors. (Let’s leave it there as fruit for a more thoughtful debate down the road. I can anticipate some collective winjing over this.) Anyway, for the chart below: Beginning January 2014, Alphacution presents 72 monthly averages of average daily share volumes (a proxy for aggregate US cash equity market liquidity) in parallel with the monthly average of daily closing levels for CBOE’s volatility index, VIX; which, btw, yield a correlation of 0.77. Here, the initial points are 1) to show that volatility drives liquidity, and 2) to imply that the evolutionary growth and decay of inventory (i.e. – the number of cash equity and equity-linked products in the market) is not having an impact on total liquidity. In other words, aggregate liquidity is flat – on average – during the 72-month period with gyrations around that average corresponding largely to volatility levels (as measured by VIX). Now, I fully realize that there is ample opportunity here to be caught with my oversimplifications down around the ankles of these points. For instance, when we shift perspective – such as with the annual average of average daily dollar volumes – the picture does imply growth in liquidity if for no other reason than a rising stock market. But, then again, there is some potential for illusion there, too: Dollar volume can increase without any changes in unit inventory. We only need an increase in unit inventory turnover for that, at any price level. The next two charts – annual averages of average daily dollar volume and average daily share volume – illustrate the disparity in these two perspectives, and help emphasize a stagnation of aggregate liquidity in the post-GFC period: This chain of logic helps to support the idea that with liquidity seemingly unaided by growth in inventory, it has become more dependent on volatility. And, the problem with that dynamic is that volatility has become more concentrated, for example, over the sample 72-month period since Jan 2014 relative to the full 30-year history of the VIX dataset. The chart, below, illustrates this relationship. So, though there are still several critical puzzle pieces missing – which gives us good reason to circle back with a more comprehensive sample of research – we can bring the PFOF figures, presented at the outset, back into the discussion under a much brighter spotlight: With aggregate liquidity flat, not being helped by growth in inventory and more dependent than ever on less-common volatility spikes, those players wielding industrial-scale technology and increasingly intelligent software (which is another way of saying that they have the power to harvest the value of information asymmetries at scale) have caused an increasingly competitive brawl over order flow (particularly considering how some of those players seek to expand beyond current scale and capacity constraints). As a result of this chain reaction, the growth in PFOF is not due to any organic growth in liquidity but instead due to those able to pay higher prices for increasingly concentrated sources of those order flows; which, in a short list of cases, is squeezing other players out of core areas of the market. Meaning: The fact that 2019 PFOF revenue is flat relative to 2018 may have profound implications for the current level of intensity in this chain of factors… What does it mean for the future if the intensity of competition for order flow has subsided? Now, mind you, as it seems with all things market related, there is much more to this story. Our PFOF figures only represent what we have observed among a consistent roster of players, not necessarily the absolute total. We also have not introduced the idea of net and gross PFOF because we have not yet factored in the impact of liquidity rebates offered by exchanges. And, what does it mean if, for example, Citadel Securities or Virtu runs a large order flow payable balance while Two Sigma Securities runs a more matched order flow payable / receivable book? While our modeling that underlies all of this becomes more robust, there remains much more to explore and assemble. Until next time…