This makes "do the cheapest thing possible" a surprisingly good proxy for "do the most energy efficient thing".

Of course, if you know something about the manufacturing chain, you can use that knowledge to do something better, but that is usually taking second order effects into account.

Taking cars as an example: buying a $100,000 Model S Tesla instead of a $35K gas guzzler to save $10K on gas is probably bad for the environment, if you only look at the marginal effect of that one car (and ignore subsidizing Tesla R&D, etc).

If the Tesla is only $35K, then you saved $10K of gas and the Tesla manufacturing process had a much lower energy footprint, so it is probably a huge win.

That would be if markets were efficient :).

As it is, the costs of goods are still dominated by middle men - people who have nothing to do with neither manufacturing nor selling the thing to the end customer. Prices today aren't good signals, because they're pretty arbitrary - shaped by supply/demand in long term, yes, but in short term directed by various business tricks (like market segmentation, brand value, contracts, etc.).

Your Tesla example actually shows it pretty well - in order to quantify its environment impact, you need to know at least that a) Tesla is electric, b) where your energy comes from.

Price would be a good proxy for a lot of things, if it wasn't so messed up by various parties trying to profit at each level of the supply chain.