Two recent articles have argued that as the energy efficiency of products improve, it becomes less expensive to operate these products and as a result, people increase their use of these products, increasing energy use and potentially wiping out the energy savings caused by the efficiency gains.

In “Solid-State Lighting: An Energy-Economics Perspective,” Tsao and his co-authors look at lighting energy use and efficiency over the past 300 years and conclude that over this period, the world spent about 0.72% of its GDP on light. As lighting has improved in efficiency and as incomes have risen, use of light has increased. The authors then extrapolate these trends into the future to conclude “that there is a massive potential for growth in the consumption of light if new lighting technologies are developed with higher luminous efficacies and lower cost of light.”

Likewise, David Owen, in an article published in The New Yorker entitled “The Efficiency Dilemma,” notes that “[a] growing group of economists and others have argued that [correlations between energy efficiency and energy use] aren’t coincidental… [and that] efforts to improve energy efficiency can more than negate environmental gains…” As support for this idea, he discusses the rising amount of refrigeration and air conditioning since the 1950’s.

Both articles make some useful points. For example, Tsao and his co-authors note that the increases in lighting and lighting efficiency contribute to increased human productivity and quality of life. But both articles tend to conflate correlation with causality and underestimate the impacts of a growing economy on these trends.  

To help illustrate these points, it is useful to differentiate between direct rebound and indirect rebound.

Direct rebound is the impact of a purchase of an efficient product by the purchaser’s use of that product. For example, a car buyer may drive an efficient car more often than an inefficient car. But, as Owen quotes in his article, an engineer from Stanford University argues “that processes, products, and activities where energy is a very high part of the cost — in this country a few metals, a few chemicals, air travel — are the only ones whose variable cost is very sensitive to energy.” Owen goes on to note: “Most economists and experts have come to similar conclusions. For example, some of them say that when you increase the fuel efficiency of cars you lose no more than about ten percent of the fuel savings to increased use.”

Indirect rebound, on the other hand, is the impact of re-spending the money that consumers and businesses save from improved energy efficiency. For example, a household that cuts its heating bill may use a little of the savings on higher thermostat settings, but much of the extra income might be spent on eating out or buying a new flat screen television. These latter items contribute to extra economic growth, but also contribute to increased energy use. ACEEE economist Skip Laitner recently did an illustrative analysis on this and found that energy efficiency policies will increase energy savings, helping to lower energy prices modestly and increase household income. These price and income effects will in turn take back about 25% of the energy saved from the efficiency policies, before accounting for these indirect effects.

Owen’s example of the rising saturation of air conditioning in homes is an example of indirect rebound, but indirect rebound is only part of the story. For example, in the U.S., Owen notes that in 1960 only 12% of American households had air conditioning. Today, that figure is about 84% according to the Energy Information Administration.  Over these 50 years, the efficiency of air conditioners has more than doubled, decreasing energy use per unit of cooling by just over 50%. But the average American home with central air conditioning spends about $300 annually on air conditioning, so even if this is half of what it would have been if we all had air conditioners with 1960 efficiency levels, the $300 per year in energy cost savings is not nearly enough to drive the large increase in use of air conditioners.

Much more likely, the causes of rising use of air conditioners are due to rising household incomes and the declining price of air conditioners. To illustrate, according to government statistics, the average American household made $52,029 in 2008, up substantially from the $6,691 in 1960 ($48,956 in 2008 dollars using the Consumer Price Index, but still more than $3,000 less than 2008 household income). Likewise, the average central air conditioner had an average value (selling price) to the manufacturer of $738 in 2009. Costs to consumers are roughly four times this amount, after including markups from the manufacturer to the customer plus installation, totaling a cost to a homeowner of about $2,952 for a new air conditioner in 2009. Adjusting this 2009 value to manufacturers for changes in the production cost index for air conditioners since 1960, the typical 1960 central air conditioner would have a cost to consumers of about $884 in 1960 dollars and $6,396 in 2009 dollars, In other words, air conditioners now cost less than half what they cost in 1960 in constant dollars.

Similarly, rising incomes and declining costs are driving growing saturations of microwave ovens, personal computers, and flat screen televisions. But for natural gas on the other hand, growing efficiencies are driving absolute declines in consumption since there are not significant new uses of natural gas. In 2005, the average household used 67 thousand cubic feet of natural gas, down from 87.5 in 1993. Since homes are getting larger, the decline in gas use per square foot of floor area is even more dramatic. Owen conspicuously did not mention this major trend. And for residential refrigerators, Owen is overstating the case, since the 75% lower energy use for new refrigerators that he cites is only partially offset with the growing saturation of second refrigerators (e.g., an increase from 14% of households with two or more refrigerators in 1993 to 25% in 2005). But Owen does correctly note that higher energy costs, such as through higher energy taxes, would contribute to reducing energy use.

Returning to Tsao and his co-authors, they do note that much of the increase in lighting efficiency and lighting use has helped to increase economic growth and individual well-being. The same can be said for growing use of air conditioning and household refrigeration, at least to some point. Tsao and his co-authors also note that while lighting energy use in developing countries is likely to continue growing, developed countries may be approaching saturation in the demand for light. We agree with this latter point. For example, with the increasing use of computer screens and increasing concerns about glare, the Illuminating Engineering Society now recommends lighting to 30 footcandles in offices with computer work, down from the 75 footcandles typically recommended for offices in their 1981 Handbook. Similarly, a recent analysis by Schipper and Millard-Ball found that in eight developed countries (including the U.S.), while transportation use per capita grew until about 2000, transportation use per capita since then has stopped growing.

Summing up, the use of lighting, air-conditioning, and personal vehicles has been growing as incomes grow and products come down in cost. Improved energy efficiency has contributed only marginally to the growing use of these services. Energy efficiency has helped to moderate (but not eliminate) the associated increases in energy use as these services grow. But for some energy uses, such as natural gas use for heating homes, improved energy efficiency has resulted in real declines in energy use. Overall, energy efficiency has helped to moderate growth in energy use, with energy use in the U.S. climbing much more slowly than GDP since the 1973 oil crisis. However, eliminating growth in energy use while the economy grows has proven much more challenging.