You know what’s a surprisingly not-great business? High frequency trading, for Getco anyway. A couple months back Getco signed a deal to acquire Knight Capital after one of Knight’s computers had an unfortunate couple of minutes. Yesterday Knight and Getco filed a joint S-4 containing a passel of merger details and Getco financials, and they’re sort of sad:
The sadness is really brought home by the fact that, if current trends hold, this year the average Getco employee will take home less than the average Goldman Sachs employee for the first time in modern memory. The Getco employees will still be a bit more productive, revenue-wise, though it’s getting closer:
The trend is not great, but Getco seems to be taking it in stride. In the course of their fairness opinion, Getco’s bankers at BAML disclosed some earnings projections for standalone Getco:1
There is a whole genre of V-shaped charts where everything gets consistently worse right up until today, and then gets consistently better starting tomorrow, because that is how trends work.
Anyway the S-4 is pretty interesting, though also pretty long. One question you could ask is, is Knight getting a good deal? (In the sense of “is the deal worth the $3.75 that it claims to be worth?” Its rescue-y elements arguably make any deal a good deal; unhorsed knights can’t be choosers and all that.) We talked about that a while back, but we did it in the absence of any information, which is how we roll around here; now there’s information so I suppose we can update our discussion. Meet me downstairs if that sounds like something you might like.2
Another question you could ask is, whither high-frequency trading. Zero Hedge’s take involves the usual global central-bank conspiracy to force retail investors back into the casino, plus declining margins from increased competition. Here’s Getco’s take:
Internalized order flow, which occurs when a broker fills customer orders from its own inventory, particularly in U.S. cash equities, currently constitutes a significant portion of total market volume, which limits the market making opportunity for companies like GETCO. In periods with low volatility, such as 2012, there has historically been a strong inverse correlation between off-exchange activity and implied volatility. GETCO expects this relationship to hold, prompting higher levels of off-exchange activity during periods of lower volatility, which could negatively impact GETCO’s market share. Conversely, periods of increasing volatility should bring a greater portion of consolidated market volume to public venues and GETCO would expect to benefit. …
The decline in consolidated revenues [in 2012] was primarily attributable to lower trading revenues from the Market Making segment, which was negatively impacted by industry-specific trends such as lower market volumes and volatilities across all asset classes, and the increasing internalization of order flow in the U.S. cash equity markets, as well as market share contraction primarily due to changes in the competitive landscape in the U.S., Europe and Asia-Pacific. For the nine months ended September 30, 2012, the average realized 20-day volatility of the S&P 500 and the 10-year Treasury Note Future, and the average implied volatility of the VIX declined approximately 28%, 28%, and 19%, respectively, as compared to the nine months ended September 30, 2011. Compounding the impact of these volatility declines was lower market volumes. For the nine months ended September 30, 2012, the average daily market volumes for U.S. cash equities, U.S. equity options, and CME interest rate products declined approximately 18%, 13%, and 21%, respectively, as compared to the nine months ended September 30, 2011 …
This correlation of volatility with Getco earnings seems to be true; until this year, Getco was reliably making in the ballpark of $40mm a year per point of VIX:3
One simple thing you could take away from this is that it’s pretty good evidence that, despite occasional complaints, high-frequency market making really is a force for reducing volatility. (Except when the computers break, but what are the odds of that happening? Again, I mean.) Getco is in the business of owning options – not so much buying options, but creating them out of computers. An order to buy stock is a tiny free option – the market-maker has the option to fill you if the price moves in his favor over the microseconds your order is open – and Getco is an accumulator of those options.
Like any hedged option owner, Getco makes more money when volatility is higher. But also like any hedged option owner, its trading itself reduces volatility: the way you make money trading volatility is by buying low, selling high, and thus creating demand when prices are low and supply when they’re high. You harvest volatility; your trading naturally reduces the supply of what makes you money. Which is nice for the market, I suppose, if low volatility is nice for the market. It’s okay for you, too, as long as outside events supply enough volatility to keep you comfortable. And as long as there aren’t lots of other option-accumulators competing with you and reducing volatility. When those conditions aren’t true, though, it’s trouble.
Knight Holdco Form S-4 [Edgar]
How GETCO Went From HFT Trading Giant To Dwarf, And Raked Up Over $50 Million In T&E Expenses Along The Way [ZH]
1. There is also an adjusted case, which is … bad. Net income never gets above $77mm in that one. “The adjusted case scenario reflects no improvement in revenues attributed to technology improvements and business investments and no execution on contemplated strategic initiatives,” which does not sound like a crazy downside case.
2. The simple answer is that KCG’s stock is basically unch’d since the filing, so the new information hasn’t changed anyone’s view that the deal is, basically, worth its $3.75/share sticker price. Sandler O’Neill’s fairness opinion has a pro forma implied valuation analysis based on pro forma (2012? not entirely clear) GAAP and non-GAAP EPS of $0.25 and $0.33 and a range of multiples from 8 to 13x – Knight ended 2011 at a little under a 10 P/E – suggesting that newco stock is worth more like $2.50-$3.30, and that the one-third-stock-two-thirds-cash deal is worth more like $2.90-$3.45 on that metric.
Maybe to put it another way, the deal is constructed to be “worth” $3.75 on a tangible book value basis, which means it’s worth $3.75 if, among other things, Getco’s equity is worth its $785mm tangible book value. $785mm is 24x Getco’s annualized 2012 earnings, which seems high, though only 7x its estimated 2013 earnings, which seems low, though those earnings are just an estimate. I dunno.
3. Defining VIX for a year as just the simple average of closing VIX levels each day that year (through 9/30 for 2012), with data from the CBOE.