Virtu’s HFT Way To Play Crazy Oil Market

February 27, 2015

Would you ever sell something to yourself and pay someone else to be the middleman? Nobody's that dumb, right?

 

Virtu, the high-frequency trading firm (HFT) of the type profiled in "Flash Boys," did just that, to the tune of $32 million.

 

High-frequency traders are perhaps the most sophisticated players on Wall Street. Some might be scoundrels, but they're not fools. That's why their recent trades in oil futures-based ETFs are so fascinating.

 

Like hedge funds and mutual funds, HFT firms keep their portfolios under wraps, except when the Securities and Exchange Commission requires disclosure.

 

Virtu's most recent form 13F, the Securities and Exchange Commission's quarterly holdings report, revealed a $46 million position in United States Oil (USO | B-100) at the close of business on Dec. 31, 2014. USO buys front-month oil futures. By the end of 2014, with oil prices at a 10-year low, USO shares had taken a beating, as you can see in the chart below.

 

Offsetting Long And Short Positions      

 

USO Volume Weighted Average Price October-December 2014

Data: ETF.com

 

USO is a great vehicle for a turnaround bet—in this case, a bet that oil prices had bottomed, and were poised to rise.

 

But that's not the bet Virtu made. Virtu actually made a pair of bets—the $46 million in USO, and an additional $16 million in the PowerShares DB Crude Oil Double Short ETN (DTO).

 

Double short—that's key.

 

The word "short" in the fund's name means that DTO has sold oil futures. It's essentially taking the other side of USO's bet, by selling the futures that USO buys. The "double" part means that DTO is leveraged, designed to produce twice the daily returns of those front-month, short-sold futures.

 

Here's how it works out:

 

 

Virtu's Offsetting Oil Futures Positions

Ticker Fund Name Leverage Factor Value 2015 YTD
Flows ($M)
USO United States Oil 1 46,213,000 46,213,000
DTO PowerShares DB Crude Oil Double Short ETN -2 16,099,000 -32,198,000

As of Dec. 31, 2014. Data: Securities and Exchange Commission.

 

 

Virtu's combined oil exposure netted to $14 million on Dec. 31, 2014. Put another way, Virtu hedged its $46 million oil futures bet by … selling oil futures.

 

You read that right. Virtu has been selling oil futures to itself and paying US Commodity Funds and PowerShares for the privilege, not to mention any exchange fees or margin costs.

 

The people at Virtu aren't dumb. In fact, they're very, very clever.

 

Virtu's not actually making a bet on oil prices. Instead, it bet on oil volatility, and on the consistency of its price changes.

 

Let me explain.

 

Daily Precision, Long-Term Randomness

Geared funds—ETFs or ETNs applying leverage—are designed to give an exact multiple of an index's returns over a specified time period, usually a single day. But over longer stretches, say, two days or more, generally, they become less precise.

 

It's like this: The funds have to deliver their ratio, be it plus-200 percent or minus 300 percent—regardless of the day you buy the fund. It's 2 X or -1X or -3X every day. So at the end of the trading day, the portfolio managers have to rebalance their portfolios. And that's where things get wonky. Counterintuitive, even.

 

Inverse funds like DTO are pro-cyclical. This means that they need to buy more securities on days when their underlying indexes rise, and sell them on the days their indexes fall. Check out my colleague Dave Nadig's illustrations of this phenomenon.

 

If DTO's managers rebalance by buying oil futures, and the oil market rises again the next day, then they're ahead of the game, since they bought into a rising market. Ditto if they sell and the market keeps falling. But in a variable market, this rebalancing game can really cost, because DTO will repeatedly buy high and sell low.

 

Another way to explain this is that geared funds perform better than their leverage ratios would suggest in trending markets, but they get killed in a whipsaw.

 

You can see this effect clearly in the chart below, which plots DTO's returns against those of USO, doubled and inverse. In other words, DTO versus -2X USO. Returns are based on daily volume-weighted-average price (VWAP).

 

 

Data: ETF.com  

 

The two returns are not identical. Between Dec. 31, 2014 and Jan. 13, 2015, USO fell 13.76 percent, while DTO rose 28.95 percent, or 1.44 percent more than its leverage ratio would suggest. That's 56.83 percent annualized.

 

That's huge, especially for a position with no directional risk. If Virtu timed this trade just right, holding it for just one day, to Jan. 2, 2015, it would have captured a 1.32 percent spread between twice USO and DTO. And in just one day. On an annualized basis, 1.32 percent is a jaw-dropping 2,591 percent.

 

Lightning In A Jar

We call this trade a "volatility extraction." Time it right, and you can make a nice living by going long "the regular way," and also holding inverse leveraged exposure in equal notional amounts. Also, this works in many markets, not just oil.

 

But asset prices rarely march steadily in the same direction, day after day. As often as not, prices bounce around. When they do, volatility-capture trades turn ugly. Sometimes very ugly.

 

If Virtu had held the oil volatility extraction trade through Feb. 3, 2015, it would have lost 4.45 percent on the deal, or 40.63 percent annualized. That's true even though USO's price (VWAP) is within 3 percent of its starting VWAP, because DTO's VWAP picked up just under 1 percent. DTO lost approximately 5 percent to rebalances in a whipsaw market.

 

Virtu bet on two things, not three. It bet on high volatility and a continuing trend, but not specifically on price direction. It's a tricky bet, not suited to amateurs or anyone who's not looking at oil prices all day, every day.

 

Roll Out The Barrel

Yet I wonder if Virtu knew all the subtleties of this bet. You see, rebalancing aside, its short position didn't exactly offset its long position, because DTO rolls its contracts differently than USO.

 

More explanation is in order.

 

If USO's current WTI contract is within two weeks of expiration, the fund will roll—that is, switch—to the next month's contract. With February WTI contracts expiring on Jan. 20, 2015, USO had to have rolled to March contracts over the four business days between Dec. 31, 2014 and Jan. 6, 2015.

 

DTO's roll window ran between Jan. 5 and Jan. 9, 2015.

 

Therefore, on the critical day of Jan. 2, 2015, USO ended the day with about half its portfolio in the February WTI contracts, and half in the March ones. DTO held on to its February contracts for a few more days. That means Virtu didn't exactly hedge its bets on Dec. 31, 2014 and Jan. 2, 2015. The front-month versus second-month mismatch can be serious at times.

 

If Virtu knew about the differential roll periods, then it's fair to say it was making three bets: high volatility; trending markets; and outperformance of the March WTI crude contract versus the February. But if it didn't dive into the funds' rebalancing methodologies, then it unwittingly left quite a bit up to chance.

 

There's no knowing when Virtu put this trade on, or when it got out of it. Nor can we know the execution prices. That's proprietary information, and it's also ancient history, at least in the world of HFT.

 

The takeaway here is the same as always: Do your homework, and understand the bets you're taking.


At the time this article was written, the author held no positions in the securities mentioned. Contact Elisabeth Kashner, CFA, at [email protected].

 

Find your next ETF

Reset All