Does the long position in contango always incur a loss?


New Member
Hi David,

the roll yield for a long futures position in contango is negative (I.e. futures prices are always decreasing until they converge to the spot at maturity so a long futures position that closed out in the interim or at maturity will always incur loss).

Does this mean that if I were to go long a futures position that, from roll yield, I would always incur a loss? This seems somewhat counterintuitive as I'm sure people enter into long positions all the time hoping to gain (i.e. they enter a futures position expecting the spot price to be higher than the agreed-upon contract price)

Thanks in advance!

David Harper CFA FRM

David Harper CFA FRM
Staff member
Hi @jcklam This is considered a truism to my knowledge (my basis is actually CAIA syllabus). In fact, I think the statement can be inverted in this way: negative roll return implies contango. See below, negative roll implies that F(t,T) < F(t-1,T); i.e., contango at least in the recent historical timeframe! However, to your point, there are at least two caveats, it seems to me (based on the years of forum conversations on this topic), that we could/should attach to the truism that "contango (backwardation) implies negative (positive) roll return:"
  1. Roll return is not the only component in total return. As I understand, total return = spot return + roll return + collateral return + rebalancing return (source: handbook of commodity investing)
  2. Related to the spot return component (if not the same thing) and to the thread for which i generated this image (, this truism (it seems to me) in general presumes a somewhat static forward curve. A curve can remain in contango while both the spot and forward price increase (to your point) such that the roll is profitable simply because the forward price is converging to an increasing spot price. So the truism is not unconditional, as pointed out by @akcfa447 (although as I think about it, it really depends on specifically how we define "contango" and what time frame we are talking about right? To me, if we define contango as below, as a feature of the roll return, then by definition the truism must be true; but if we define contango as a instaneous shape of the forward curve, which we seem to do, then the truism is not always true). We have search tags for contango and backwardation if you want to see more; e.g., ,
  3. Less central, I suppose. Maybe let's not forget that hedgers often or generally have a negative expected payoff. If the motivation is to hedge, there is nothing wrong per se with negative expected payoff on the hedge position per se (similar to how insurance has negative expected PV). I hope that's helpful!