I think the scale would be too hard to measure at a single-calorie level, but I don’t think it’s an invalid question. In fact, it’s interesting, I think. If we measure it at a kilo-calorie level then I think that we could measure it in dollars and cents.
And finding an “average” or mean cost could be done, I suppose, but since there’s going to be such an incredibly wide range of values, this won’t be a normal bell curve of values. It’s a curve with a constraint at zero, since you’re unlikely to find any food costing less than nothing (where you’d be paid for consumption, that is), but you can find foods with priced values in the thousand-dollar-per-kilocalorie range, I expect. But not so many of those. So: a curve with a long tail to the right representing a lot of dollars per kilocalorie, and a big hump to the left, closer to zero, representing the cheaper calories that most of the world consumes in the form of rice, sorghum and other grains.
When you find the average price, be sure to pay attention to the other measures of central tendency: the median and mode. Those will tell you more, I think, about “what the world actually spends” on its food calories. I’ll be interested to see the values.
You’ll want to differentiate, too, I think, between human and domesticated animal feeds. (As well as bacteria, insects, worms, arachnids, wild animals, birds and fishes, who get their calories ‘for free’, but… a calorie is a calorie.)