Gasoline Trading Shouldn't Be This Easy. [Code Included]
It's about time we gave commodities some love.
Gasoline — you know, that sweet smelling liquid that you’re somehow always paying too much for? It’s one of the largest influences in our day-to-day lives, but when it comes to actually trading it, it’s kind of a sleepy product:
Despite this low interest, the specs on gasoline futures are quite attractive. Its minimum tick value is $0.0001 (e.g., 2.4000 to 2.4001), with each tick representing $4.20 in PnL. This means that capturing just a $0.50 move in the price of gasoline will result in a $21,000 gross profit. So, let’s just say we have quite the incentive to give this market a deeper look.
To get started, we first need a bit of inspiration.
Who’s Trading This Stuff Anyway?
CTAs, or Commodity Trading Advisors, are firms that, well— advise on commodity trading. Generally, these firms help clients such as large farmers who want to get stable wheat prices, but they also engage in a good bit of proprietary trading — and they’re damn good at it:
So, now that we know about a major player in the space, we want to get inside data on their activities, model them, and see just how much of an edge we can eke out from this small corner of the market.
Luckily, many months ago, we began looking into the hedging pressure effect that will give us a head start in our pursuit. Essentially, by using published data on the positioning of farmers, swap dealers, CTAs, and speculators, we were able to get a robust estimate of where a particular futures market would be heading by the contract maturity date:
While farmers aren’t major players in gasoline markets, it’s likely that the major participants are far more informed than with say a typical S&P futures market. The few participants in this market are trading with factors in mind such as regional supply and demand, refining margins (e.g., cost of converting crude oil to gasoline), and even the weather (particularly in the Gulf Coast region).
So, in theory, data on their positioning should have some quantitative edge. Our earlier experiment was done before we developed our big boy tools, so let’s create a new hypothesis:
Based on the positioning of presumably informed traders in the gasoline markets, can the subsequent direction of gasoline prices be predicted and profitably traded?
Let’s find out.