Rebound effect

Suppose a country subsidizes a device that would lead to more efficient energy use, then imagine this device is capable of reducing consumption of the fuel needed to run an energy conversion device (car, heater, etc) by 20%. After a year can we expect that the energy consumption related to the use of the energy conversion device be reduced by 20%?
Most probably not. It has something to do with microeconomics and consumer demand. Energy efficiency improvements generally mean a lowering of the cost of running the energy conversion device. Rational consumers would consume more of the good/service that is cheaper. If say, the new compact flourescent lamp (CFL) just needs 80% less electricity to provide the same lumens as an incandescent light bulb, would I after a year or a month save 20% electricity out of using CFLs? Probably not, because I will conveniently forget to turn it off if I am stepping out of the room. Or maybe I'll turn on the lights earlier in the afternoon.

What is a rough measure of this "rebound effect?" A good bit of thinking have been invested in estimating this effect. Check some of the academic energy journals...but it boils down (roughly) to the price elasticity of our demand for that fuel. Suppose you cut consumption of electricity by 50% everytime the price of electricity goes up by 100% (that is to say the elasticity of your electricity demand with respect to the price of electricity is -0.5), then the rebound effect in the case of the introduction of CFL is 50% which means, dont expect to reduce electricity consumption by 20%, most probably it would be less than that...perhaps only by 10%. Energy efficiency drives are not enough, policymakers still need to work with taxing and pricing policies to align the incentives of consumers to their policy goal of reducing energy consumption.