Does Home Weatherization Pay for Itself?

Energy efficiency isn’t the sexiest environmental issue out there, and its advocates often complain that the subject doesn’t get enough media attention. Last week, they got a lesson in “be careful what you wish for.” A team of academics released a working paper that called into question a broadly accepted tenet of the field—that energy efficiency programs pay for themselves. The paper investigated a Michigan program to weatherize the homes of low-income families, and concluded that the state government was spending more than the families were saving. The Wall Street Journal, the Washington Post, Vox, and others covered the story, mostly buying into the paper’s thesis and raising the hackles of the energy efficiency community.

The real story, however, is more complex than the initial coverage let on. So, who’s right, who’s wrong, and what’s still up in the air? Here’s what you need to know.

Cost-effectiveness is in the eye of the beholder

The state of Michigan offers to pay the full cost of home weatherization, furnace replacement, and other energy efficiency measures for low-income families. Only a small fraction of the poor actually applies for the benefits, however, and an even smaller proportion completes the process.

The researchers saw this as an opportunity to conduct a randomized control trial, a study design traditionally used in drug trials and other medical research but currently in vogue among economists. The team compared the energy consumption of the families who took advantage of the program (the test group, in medical parlance) with that of those who did not (the control). The difference between the two groups represents the energy savings of weatherization. It is a very clever study design.

The results, however, did not look good for the government program. According to the study, the weatherization and upgraded appliances would eventually save the average participating family about $2,400. Since the program costs $4,143 per family, the headline-friendly finding was that the government is spending more than the beneficiaries were saving.

These findings appear damning at first. But, in the words of Esther Duflo, a pioneer of using randomized control trials in economics, “think again, again.” The Michigan program isn’t just aimed at energy savings. It’s also a program to help keep low-income families warm during the winter. That fact must be accounted for in the cost-effectiveness calculus, and it wasn’t.

Here’s an analogy. Say you’re in the market for a new car. You like the Honda Accord, which costs $20,517, but you’re willing to spend $24,701 on a Toyota Prius if the superior fuel efficiency will pay for itself. How do you decide whether to spring for the Prius? It’s a fairly simple calculation. If the amount of money you’ll save on gasoline over the car’s lifetime (let’s say $5,000) is more than the difference in price between the two cars ($4,184), then you buy the Prius.

But that’s not how the researchers calculated the cost-effectiveness of the Michigan program. They compared the value of energy savings to the total cost of the program. It’s a crucial distinction, because new furnaces appear to be the largest single expenditure in the Michigan program. This is the equivalent of comparing the gas savings of a Prius to the entire $24,701 price tag of the car.

Put slightly differently, most of car’s value comes from its ability to move you around. Hybrid technology is an add-on—a feature for which you pay a premium.

Similarly, most of a furnace’s value is warmth. Energy-efficiency is a bonus that costs a little more.

Does this mean the study’s numbers are wrong, and the program is an economic winner? Not necessarily. It’s not clear how the results would change if the researchers only considered the costs of the added efficiency, rather than the entire cost of the program. Just as importantly, it’s difficult to put a monetary value on keeping low-income families warm during winter.

How much energy do efficiency programs actually save?

While the cost-effectiveness numbers are a little misleading, in my view, the study still makes a very important point. The government will only pay for new furnaces and home weatherization if the forecasted savings outweigh the costs. However, by comparing the energy use in weatherized and non-weatherized homes, the researchers found that the forecasts were wrong—by a lot. The program’s forecasters allegedly overestimated the benefit by at least double, and possibly by a factor of four.

The researchers themselves can’t fully explain why the estimates appear to be so far off, since they didn’t have the resources to perform insulation tests at thousands of households. It’s possible that the weatherizing technicians did a bad job or that the ducts within the homes’ walls had undetected holes.

Or maybe—and this is the troubling scenario—some of our methods for estimating potential energy savings are simply wrong. The Michigan auditors, for example, rely on a piece of software called the National Energy Audit Tool to forecast the energy savings of weatherization. It’s one of many programs in the field, and all of them are likely reliable when it comes to estimating the benefits of individual appliances in isolation—we can compare the energy consumption of various furnaces, for example, under laboratory conditions. But houses are complicated: They have strange crevices and corners. The wind hits them at various angles. The structural elements wear down at different rates. All of these things, and more, affect how much heat will escape and how much electricity will be saved after weatherization.

Remember the climate

It’s important to keep a couple of things in mind. First, this is just one paper, and, so far, it looks like an outlier. Several other studies, including those from well-regarded sources like the Lawrence Berkeley National Laboratories, have found that efficiency programs perform admirably. Savings from ordinary residential programs—as opposed to the more generous subsidies available to low-income families—deliver nearly doubled the government’s outlay in the Berkeley research.

Still, the randomized control design used in last week’s study has many virtues—it’s transparent, it’s simple, and it can be repeated in other states. Dismissing its findings because they don’t jibe with our expectations would be a mistake. Michael Greenstone, one of the study’s coauthors, concedes that pursuing his research approach further “may require society to reexamine some sacred cows, but this seems a small price to pay to reduce the odds of disruptive climate change.”

And that leads to the second point. The environmental and health costs of carbon emissions aren’t reflected in our utility bills, if they’re calculable at all. Energy efficiency programs are a way to reduce those costs.

Last year, an opinion poll showed that two out of three Americans would pay more for electricity in order to lower carbon emissions. This study raises exactly the same question. Even if energy efficiency programs don’t immediately pay for themselves—and the jury is still out on that question—we have to consider the bigger picture. The goal is to cut carbon pollution, and energy efficiency measures do that.