With volatility spiking in Q1 of 2018 – and the successful porting of KCG’s intellectual property (IP) prior to that in late 2017 – VIRT earned a welcomed reprieve from the conditions of recent quarters, as we predicted here (and elsewhere prior to that). In the exhibit below, Alphacution’s as-if modeling of the combined entity – Virtu + KCG pre-Q3 2017 – yields a level of net trading income that would not have been seen since Q1-2016.
Meanwhile, Alphacution’s tracking of adjusted net trading income per employee – a proxy for our common look at revenue per employee (RPE) – starkly illustrates the path through the most recent maneuvers: Persistent declines in top line “productivity” since its most recent peak in early 2015 ultimately led to the acquisition of KCG, which closed in July 2017. Swift transfer of KCG’s IP onto Virtu’s infrastructure along with elimination of redundant technology and human capital allowed this productivity measure to bounce off its lows in Q3 2017 to finish the year as strongly as possible given the persistence of historically low volatility.
The aforementioned spike in vol starting in early February 2018 provided the newly integrated platform with enough raw market fuel to bring this picture back to the lower end of the range that Virtu enjoyed for the 3 years prior to 2017 (see Exhibit below). In short, this was a series of savvy navigational moves that produced meaningful upticks in this measure of productivity against a backdrop of declining theoretical capacity in their core US equity strategies…
Now, typically, when we update a particular model with the latest data, we also try to expand what we are able to expose within that model. An example of this type of model expansion is the analytic illustrated below – Asset under Management per Employee (AuM/e) – which we showcased in our latest asset management technology spending study, “The Context Machine.”
Alphacution counts market-makers, high-frequency trading firms and other proprietary trading groups (or, “prop shops”) as members of the broader asset management continuum (which also includes hedge funds). However, one of the challenges inherent to firms like Virtu as members in our operational benchmarking framework is that they don’t think of or report assets under management (AUM) in the same way as hedge funds or traditional asset managers.
These players aren’t aggregating third-party capital into traditional fund vehicles. Instead, their “AUM” is the sum of line items like cash on hand, short and long term borrowings, net securities borrowed/loaned, repo / reverse-rep securities, payables to / receivables from brokers, and trading assets / liabilities. In other words, the sum of trading assets less the sum of trading liabilities yields a measure of total trading capital.
So, in the latest update to our Virtu model, we use quarterly disclosures of “trading capital” as a proxy for AUM in order to calculate our proprietary AuM/e analytic. Of course, one could argue with the idea that trading capital is a reasonable proxy for AUM, however, because we have validated the hypothesis that AuM/e levels are a good indicator of underlying strategy selection(s) – and that the highest turnover strategies tend to yield the lowest AuM/e readings – a slight shift in the numerator of this analytic in the case of firms like Virt is not going to change the fact that they yield the lowest readings of AuM/e relative to any other “asset manager” in the world.
In fact, Virtu is the lowest we have found yet – which makes sense in the context of our framework: It turns out that AuM/e is a reliable measure of “human capital leverage,” which is strategy dependent, and tends to correspond inversely with total technology spending per employee (TCO/e), which is also strategy dependent. This is an easy translation: High-turnover strategies require significant technology spending, minimal capital and can be successfully executed with small teams thus yielding high TCO/e and low AuM/e.
Though appearing somewhat volatile due to the scaling, the chart below illustrates an incredibly consistent AuM/e level for Virtu of between $2 – 3 million per employee over the past 14 quarters. This level is one extreme of the asset management continuum. By contrast, Alphacution has discovered that the other extreme, BlackRock, generates recent AuM/e levels of nearly $400 million per employee.
With this, Alphacution has determined that every other asset manager in the world yields AuM/e levels that are somewhere in between these Virtu-BlackRock poles. Moreover, a related hypothesis that we are further validating now is that specific investment strategies tend to be correlated with specific AuM/e levels because of the underlying workflow automation inherent to each of those strategies.
The last of our updates on extremes takes us back to where Alphacution began: technology spending, and more specifically, total technology spending per employee (TCO/e). In the chart below, Virtu continued to define the highest level of TCO/e among all asset managers (that we have been able to observe) through Q2-2017 at nearly $513,100 per employee (per year), which is another of our strategy-dependent analytics; sometimes called a “technical signature.”
In Q3-2017, the combined technical signatures of old Virtu and KCG – including their variant allocations to human capital – resulted in a decline in TCO/e by nearly half (-48.6%), down to $263,200; a level we might expect given the combination of what amounts to two different business models each with its own inherent technical signature.
With significant reductions in redundant technology and human capital since the deal closing in July 2017, the new Virtu TCO/e reading has rebounded to $366,300 as of Q1-2018. While still incredibly high relative to all other asset managers (and even other business models within financial services), this latest level reflects an amalgamation of the higher market-making tech spend plus the relatively lower level tech spend of the remaining customer-facing business of the former Knight Capital.
Alphacution’s expectation going forward is that, while Virtu’s TCO/e level may have stabilized in a new range, the intensely competitive nature of this subset of trading strategies will force it to continue to inch up over time.
At the end of the day, this story – or, the profitability of this story – hinges on volatility, particularly US equity volatility. Will volatility return to single-digit VIX levels in Q2 and beyond? And, if so, what new and savvy navigations will Virtu need to make in order to stay ahead of the grinding and bleeding that those levels entail?
And, as always, if you value this work: Like it, share it, comment on it – or discuss amongst yourselves – and then send us firstname.lastname@example.org.
As our “feedback loop” becomes more vibrant – given input from clients and other members of our network, especially around new questions to be answered – the value of this work will accelerate.
Don’t be shy…