Beyond the Rebound Effect: Energy Efficiency Gaps

The Paris Climate Agreement has renewed a focus on American state and local policies which aim to reconcile economic growth with the declared goal of limiting global warming to two degrees Celsius above pre-industrial temperatures. Energy efficiency support programs seek to reduce the consumption of fossil fuels through technological improvements to insulation technology and electronic devices. These programs are seen as a rare “win-win” opportunities because upfront investments are believed to be recovered through the avoided costs of energy consumption alone. The global popularity of these perceptions is evidenced by the fact that the International Energy Association’s pathway to limiting global warming allocates 49% of greenhouse gas abatement to energy efficiency improvements.

In practice, though, a persistent difference has begun to emerge between projected returns on investments in energy efficiency and realized returns. This is known as the “efficiency gap” and is commonly attributed to imperfect information, capital market failures, and behavioral explanations such as consumer short-sightedness or inattentiveness. However, a 2018 paper by Meredith Fowlie (UC Berkeley), Michael Greenstone (University of Chicago), and Catherine Wolfram (UC Berkeley) examines what the authors call a “pedestrian possibility” that in the real world, returns on energy efficiency are just lower than engineering models indicate.

The authors analyzed large-scale field evidence on the returns of energy efficiency in a randomized control trial involving over 30,000 households in Michigan participating in the Weatherization Assistance Program (WAP). WAP provides low income households with federal funds for home improvements such as furnace replacement and attic insulation. A quarter of these households were randomly chosen, encouraged to apply and aided in applications. This was the treatment group. The remaining households were also free to apply but received no explicit advice to do so. WAP pays for energy efficiency measures that pass a cost benefit test based on engineering estimates using simulation tools such as the National Energy Audit Tool (NEAT). The energy efficiency measures implemented in treated households were projected (by engineering estimates) to unlock savings of $10,000 per household over an average design lifespan of 16 years. However, using real energy consumption data from these households from 2011 to 2014, the average cost savings per household over the same lifespan was calculated to be only between $1,000 and $3,500.

In fact, while the engineering estimates project savings of up to twice the upfront costs, the study finds savings of only between 31% and 76% of the upfront costs, depending on the how much we discount savings realized in later years. For a household that doesn’t receive weatherization assistance, projected energy savings from ex-ante engineering simulations would yield a rate of return of about 12% with a lifespan of 16 years compared to the -2.3% from the realized savings in this experiment. One may argue that programs such as WAP are encouraged by government agencies in cognizance of social benefits of reduced carbon emissions, which renders this private rate of return less relevant than a social rate of return incorporating the benefits accrued to consumers by avoided emissions. Using peer-reviewed literature as the source for monetary figures of these benefits, the study finds that this social rate of return on energy efficiency is about 0.1% after 16 years and about 2.3% after 20 years. Compared to standard rates of market returns, the costs of weatherization assistance still seem to yield significantly lower social rates of return.

This might be attributed to the “rebound effect” as awareness of increased efficiency may lead to higher energy consumption. However, the study does not find a significant increase in indoor temperature in weatherized homes, even during winter months with lower than mean cold temperatures. If the “rebound effect” had explanatory power, this should have been heightened particularly since the households did not need to pay to institute energy efficiency measures.

The gap might also be attributed to inflated costs of weatherization assistance when done through government contracting. However, recorded unit costs of WAP interventions to private market costs of energy efficiency measures do not yield any significant differences. The authors delineate one legitimate possibility: the projected energy savings estimated by ex-ante simulations overestimate the potential gains from energy efficiency because feedback from implemented efficiency programs is not incorporated into simulated performance. The authors’ opinion favors this notion; they note that tools used by engineering firms originate in a culture lacking penalties for failing to deliver promised results. Therefore, they do not incorporate ex-post performance into models. Given the debate that the Green New Deal has kicked up on allocating federal investments towards climate change, the rather unglamorous need of the hour is to look more closely into the assumptions and feedback mechanisms of technical simulations and studies like NEAT that inform cost-benefit analyses of programs geared towards climate change mitigation.

---

Fowlie, Meredith, Michael Greenstone, and Catherine Wolfram. 2018. “Do Energy Efficiency Investments Deliver? Evidence from the Weatherization Assistance Program.” The Quarterly Journal of Economics 133, Issue 3 (August): 1597–1644. https://doi.org/10.1093/qje/qjy005.