Another New Yorker article: Jevons Paradox— Does Improving Efficiency Do Any Good?

The New Yorker has done much to introduce non-scientists to scientific thinking (eg, Kolbert’s articles on climate change), but now aims to confuse us, or so it appears, by presenting real concerns in a too simplistic manner. David Owen’s recent article discusses Jevons Paradox in The Efficiency Dilemma has been attacked by critics who object to his omissions. Truth sometimes lies in the middle, but in this case, Truth appears to be more towards the extremes, with a caveat, it depends on where and for what.

Jevons pointed out a century and a half ago that increased efficiency can lead to lower prices and thus to consumption greater than if there had been no improvement. Rebound effect is the term used when increased efficiency leads to lower consumption, but the decrease is made smaller by behavior change.

From the article:
• Consumption increases as costs go down. Because refrigerators are so much cheaper to operate, Owen says, they have spread to hotel rooms and gas stations. Additional energy loss (and increases in greenhouse gas emissions) occur as we increase the amount of food we buy and waste (and consume) as refrigerator size increases. Altogether, per capita energy consumption due to all these changes has presumably grown even as energy to power residential refrigerators has gone down. Other examples are the rapid increase in air conditioning and the size of (and number of) houses in the South, and increases in lighting use in the US so “that darkness itself is spoken of an endangered natural resource”—increases in efficiency mean that the typical person uses more energy for both lighting and air conditioning.
• Increased efficiency in automobiles has been devoted to increasing horsepower and weight rather than fuel economy.
• Decreases in cost increases both the number of car owners and the number of vehicle miles traveled per car per year.

Changing energy use in refrigerators
Changing energy use in refrigerators

The Jevons Paradox is still considered a factor in many parts of the world. For example, the introduction of cheap, efficient cars to India (the Nano) was expected to lead to increased consumption of oil. (Between March 2009 and January 2011, some 1 million cars were sold—see here for some reasons why the Nano hasn’t taken off, although this may still happen). Cheap solar panels (expensive compared to prices paid where there is a reliable grid, but cheap relative to the cost of a long ride in a motorcycle taxi to recharge the phone) and efficient light bulbs in Africa also lead to increased energy use, but a low-greenhouse gas form. There is great enthusiasm about the latter, but I have yet to hear policy experts wax enthusiastic about the Nano. The policy community appreciates the need to make more energy available to the poor. However, the need for more cars is either less clear than needs for phones and lightbulbs, or the downsides of adding more photovoltaics are smaller than the problems of using more oil, of which climate change is just one.

increased energy use in Kiptusuri
Increased energy use in Kiptusuri, Kenya

Of course, oil use is increasing in India anyway, as Indians become wealthier. Owen fails to discuss the effects of increased wealth on people’s choices, a fairly large omission, so would attribute the increase solely to the more efficient automobiles. (Nor does Owens consider the time for stock turnover.)

The rebound effects in 21st century US are of a different scale than the examples above. We already leave our lights on. A lot. We own considerably more than one car per licensed driver, 842 cars/1000 people (compared to 12/1000 in India). So it’s unlikely that the introduction of more efficient cars will lead to as dramatic an increase in fuel use as in India. Or that more efficient bulbs will produce the increase in lighting now being seen in Kiptusuri.

According to Effectiveness and Impact of Corporate Average Fuel Economy (CAFE) Standards, CAFE standards appear to have a 10 – 20% rebound effect, while changes in Europe produce a rebound effect of 20 – 30% (the difference is due to those shifting from public transit).

The rebound effect for cars today in the US may be greater than for refrigerators, now that the market for refrigerators is apparently saturated. (I’ve heard people in policy wonder when the US will reach saturation for automobiles—there has to be a point at which nothing can push Americans to drive more.)

There are three causes for the rebound effect, according to Energy Efficiency and the Rebound Effect: Does Increasing Efficiency Decrease Demand? (pdf)

Direct Effects – The consumer chooses to use more of the resource instead of realizing the energy cost savings. For example, a person with a more efficient home heater may chose to raise the setting on the thermostat or a person driving a more efficient car may drive more. This effect is limited since a person will only set the thermostat so high or have so many hours to spend driving.

Indirect Effects – The consumer chooses to spend the money saved by buying other goods which use the same resource. For example, a person whose electric bill decreases due to a more efficient air conditioner may use the savings to buy more electronic goods.

Market or Dynamic Effects
– Decreased demand for a resource leads to a lower resource price, making new uses economically viable. For example, residential electricity was initially used mainly for lighting, but as the price dropped many new electric devices became common. This is the most difficult aspect of the rebound effect to predict and to measure.

See the paper for the scale of the rebound effect, which is close to 0% for home appliances, 10 – 30% for cars, and 0 – 50% for space cooling.

Even advocates of energy efficiency see a need to do more. In Leaping the Energy Gap (subscription required), Dan Charles says,

Experience has shown that there is more to saving energy than designing better light bulbs and refrigerators. Researchers say it will need a mixture of persuasion, regulation, and taxation.

(August 14, 2009 Science)

A frequently touted statistic is that while per capita US electricity use increased 40% over the last 3 decades, it remained flat in California. Some credit the efficiency mandates in California. That appears to be true only in part:

Anant Sudarshan and James Sweeney of Stanford University’s Precourt Energy Efficiency Center (PEEC) recently calculated that the state’s energy policies can take credit for only a quarter of California’s lower per capita electricity use. The rest is due to “structural factors” such as mild weather, increasing urbanization, larger numbers of people in each household, and high prices for energy and land that drove heavy industry out of the state.

Art Rosenfeld
Art Rosenfeld

An old economic assumption is that if scientists add efficiency, the consumer will come.

[Art] Rosenfeld [who was the most important person pushing California’s push toward higher efficiency] and Edward Vine had a friendly, long-running argument during their 2 decades as colleagues at [Lawrence Berkeley National Laboratory]. Rosenfeld believed in technology. When he testified before the U.S. Congress, as he did frequently in the early 1980s, he always came with props in hand: compact fluorescent light bulbs, heat-shielding windows, or computer programs for predicting the energy use of new buildings. But Vine, whose Ph.D. is in human ecology, wasn’t convinced of technology’s power. “We can’t assume, if we have a great technology, that people will rush to stores and buy it,” Vine says. “We need to find out how people behave, how they make decisions, how they use energy, and we need to work with them.”

For the most part, energy-efficiency programs around the country have followed Rosenfeld’s line. They offer financial incentives for adopting energy-saving, cost-effective technology, and trust that consumers will follow their economic self-interest.

Yet many researchers are now coming around to Vine’s point of view. Consumers don’t seem to act like fully informed, rational decision-makers when they make energy choices. Many avoid making choices at all. Give them a programmable thermostat, and they won’t program it. Offer them an efficient light bulb that pays for itself in 2 years, and they won’t buy it.

Some points made by the article:
• The goal is to decrease energy use per person—stable energy use is not enough.
• Even for-profit companies don’t realize how much money can be saved on energy [and companies do much better than individuals].
• In a crisis, people respond to a need for “good citizens”. Some percentage of that change in behavior remains after the crisis ends.
• We see waste in others reflecting their “inner characters” and “own wasteful practices as the product of circumstances”, so information about the need rarely helps.
• Role models do help.
• We care what others are doing. Sacramento Municipal Utility District included information with the bills about how one’s energy use compares to one’s neighbors, and energy use declined 2%. [Information about saving energy left on your door knob is ineffective if accompanied by the importance of saving money or saving the earth, but is effective if we are told that our neighbors are doing it.]
• The current market option, more efficient and more expensive appliances targeting high-end customers, is less effective than selling these appliances at Costco or Walmart.
• Social marketing works, at least in some places, such as Hood River, OR, where 85% of the homes got energy audits and free efficiency upgrades.

[Hugh] Peach compared the process to a political campaign. The utility sat down with local leaders, followed their advice, and relied heavily on local volunteers. The process was time-consuming and labor-intensive but, Peach says, a pleasure. There was “a lot of community spirit. People just saw it as the right thing to do.”

• Feedback helps, eg, the Prius dashboard showing car drivers their rate of energy use. There is hope that Smart Meters will lead consumers to reduce energy use in their home, first by cutting use, eventually shifting to more efficient appliances.
• Green buildings don’t do nearly as well as advertised, and architects get too little feedback on how energy use changes as a result of their work. In a response to this article, several examples are given for projects where actual energy use came in at least double predicted energy use.
• There are a number of perverse incentives: people away from home have little incentive to reduce energy and water use. Landlords have little incentive to purchase more expensive more efficient appliances. Cable services provide boxes which use 40 W 24 hours/day and have no incentive to spend a tad more on reducing energy use. These perverse incentives might be responsible for 1/4 of US residential energy use. In Japan, on the other hand, vending machine suppliers pay for the electricity, and vending machines are more efficient.
• Really, adding a cost to energy is necessary, because we need to see the cost of our behavior, which goes beyond the price we pay today for energy.

In Behavior and Energy Policy, (subscription required, March 5, 2010 Science), there is more discussion of how to combine greater energy efficiency with changed behavior.

Summary: Jevons Paradox appears more important in less saturated markets and other factors, such as increased wealth should be considered. Increased efficiency does reduce energy use in the US, but if our goal is to mitigate greenhouse gas emissions quickly enough, we may want to move to “mixture of persuasion, regulation, and taxation.”

Comments from others: See The National Geographic blog for the comments of James Barrett (Clean Economy Development Center) and Matthew Kahn (UCLA)

One Response to “Another New Yorker article: Jevons Paradox— Does Improving Efficiency Do Any Good?”

  1. […] When we follow a path of a variable, we also see causal loops present in the system. Causal loops are an important learning tool in systems thinking and shows how some variables can continuously reinforce, or balance each other in a loop. For instance, if technology efficiency increases, resource use decreases. Unhindered by resource availability, business activity grows and more investments are made in technology efficiency – thus putting these variables in a reinforcing loop. This seems like a good thing, since, after all, efficiency is often tooted as the “low-hanging fruit” of sustainability strategy. However, the increase in overall resource use due to higher business activity can overcome the initial decrease gained from efficiency (this is also known as the Jevon’s Paradox. See more concrete examples here). […]