Consider the following scenario:
You have two cars, one is a very inefficient van (giving you on average 5 MPG) and one is a relatively efficient sedan (giving you on average 20 MPG). Due to your work and obligations you have to drive each of them the same distance every month.

You need both types of cars and for now you can replace only one of them. What should you replace?
Option 1: Replace the 5 MPG van with a 10 MPG van
Option 2 Replace the 20 MPG sedan with a 50 MPG sedan

What would you select?

As a new paper by Rick Larrick and Jack Soll shows many people select option 2, where in fact option 1 would be better for them (also see this story about the research).

Does this sound odd? Lets look at it more carefully: Lets assume that people drive 100 miles a month. This means that the 5 MPG van uses 20 gallons a month while the 20 MPG sedan uses 5 gallons a month. Now what if we change them? If we change the van we would change from using 20 gallons a month to using 10 gallons a month (saving 10 gallons a month). If we change the sedan we would change from using 5 gallons a month to using 2 gallons a month (saving 3 gallons a month). Now it is clear that changing the van is a much better move.

Why is this not obvious to people from the MPG information? It is because comparing MPG don’t directly reflect the cost differences. In Europe they present efficiency measures I liters per 100 kilometers, and this seems to be a much more intuitive measure.

One advice is clear – if you think about changing a car, change first the least efficient car.