Automotive engines steadily improved in efficiency by roughly 60 percent from 1980 to 2006, according to a new study by MIT economist Christopher Knittel. That means we could already be driving cars that get an average of 37 miles per gallon (MPG), well above today's average of 27 MPG. The catch, points out Reason's Ronald Bailey: we'd have to be driving cars with the same average weight and power as the average car on the road in 1980.

Instead, consumers took the majority of the improvements in engine efficiency over the last three decades to enjoy larger and more powerful cars (e.g. increasing their use of energy services) rather than reduce energy use, according Knittel's paper, published in American Economic Review.

As Reason's Bailey notes, "This seems an example of the energy rebound effect in which increased energy efficiency encourages people to use even more energy; in this case to fuel bigger and peppier cars."

Indeed it does. (Click here for an introductory FAQ to rebound effects)

If vehicle weight and average power had held constant from 1980 to 2006, Knittel estimates that vehicles today would be roughly 60 percent more efficient than they were in the '80s. Instead, average fuel economy of new vehicles sold in the United States improved just 15 percent over this period.

The reason is clear: consumers chose to take these improvements in engine efficiency as a major increase in average vehicle weight, which rose 26 percent, and a doubling of average horsepower, which rose 106 percent from 1980 to 2006.

"Most of that technological progress has gone into [compensating for increased] weight and horsepower," notes Knittel.

There is definitely a direct rebound effect at work here: improvements in engine efficiency lower the cost of the energy services derived from vehicle engines -- acceleration, vehicle size, etc. -- which consumers respond to by increasing the use of those energy services, eroding the reduction in energy use expected if rebound effects were absent.

It's not straightforward to disentangle this rebound response to engine efficiency from other factors that may contribute to consumer preference for larger, peppier cars. Yet Knittel's study illustrates that a full 75 percent of the technical engine efficiency improvements are "lost" to making cars larger and more powerful rather than reducing fuel use: vehicle fuel economy improved just 15 percent, instead of 60 percent, with three quarters of the engine efficiency improvements taken as increases in energy services (e.g. power, vehicle size) rather than improvements in vehicle miles per gallon.

Given the magnitude of this "takeback" of efficiency improvements, the direct rebound at work here must be quite large. Furthermore, the effect of growing consumer incomes can only account for a portion of the increase in vehicle size and power; average U.S. incomes rose by about 20 percent in real terms from 1980 to 2007, while average horsepower more than doubled and vehicle weight increased by more than a quarter over this period.

What is clear is that 75 percent of expected technical efficiency improvements in vehicle engines were lost to accommodate consumer preference for larger, more powerful cars, rather than reduce energy use.

And that is before we even consider any increase in average vehicle miles travelled -- the main energy service derived from vehicle engines -- which also rose over this period.

As Knittel aptly notes, if you "end up reducing the cost of driving ... you actually get what's called 'rebound,' and they drive more than they would have."

As we reported in "Energy Emergence" our comprehensive 2011 review of the expert literature on rebound effects, direct rebound effects for personal vehicles have been generally estimated within the range of 10 to 30 percent of initial energy savings (see table below, from page 15 of our report).

Direct_Rebounds.pngSource: "Energy Emergence: Rebound and Backfire as Emergent Phenomona," Breakthrough Institute, February 2011. Click to enlarge.

Yet the estimates we surveyed typically only measured the increase in driving (vehicles miles travelled) in response to improvements in on-road vehicle fuel economy. As our table notes, "Unmeasured in these studies are changes in automotive attributes, particularly heavier vehicles and more powerful engines."

Now those attributes have been measured as well, and if Knittel's research is accurate, the total share of technical efficiency gains lost to these factors appears to be far more substantial if we start from vehicle engine efficiency gains, rather than on-road fuel economy gains.

When measuring the end result of a roughly 60 percent improvement in vehicle engine efficiency, we can estimate that 75 percent of these gains were "lost" to larger and more powerful vehicles (45 percentage points out of 60), while increases in vehicle miles travelled eroded another 2.5 to 7.5 percent of these gains (1.5 to 4.5 percentage points out of 60).

In total, 77.5 to 82.5 percent of the initial technical improvement in engine efficiency is taken to support larger, more powerful cars, driven further (see graph below). In other words, only about one-fifth of the improvement in engine efficiency translated to a reduction in energy use.

Engine_Rebound_graph.pngClick to enlarge

Note that in the future, we may see some degree of saturation in consumer demand for larger or more powerful vehicles. This measurement of eroded energy savings from 1980 to 2006 may not be a perfect predictor of future rebound effects in response to improvements in vehicle engine efficiency.

That said, the magnitude of this apparent rebound effects here would counsel significant caution in estimating the impacts of future technical improvements in vehicle engine efficiency. If one were to consider only technical efficiency improvements alone, without accounting for rebound effects and consumer preference, it would be easy to vastly over-estimate the likely reductions in fuel use.

"The fact that cars have muscled up rather than become more efficient in the last three decades is known, but Chris [Knittel] has done the most credible job of measuring that tradeoff," notes University of California at Berkeley resource economist Severin Borenstein. "This paper should get a lot of attention when policymakers are thinking about what is achievable in improved automobile fuel economy."

As Knittel himself cautions, given consumer preferences, larger changes in fleet-wide gas mileage will occur only when policies change, too. "It's the policymakers' responsibility to create a structure that leads to these technologies being put toward fuel economy," he says.

Rebound effects simply must be accounted for.