When customers sell demand response into a regional capacity market (such as PJM’s Reliability Pricing Model, known as the RPM), how much credit should they earn for agreeing to curtail demand and...
The Trouble with Freeriders
The debate about freeridership in energy efficiency isn’t wrong, but it is wrongheaded.
What if the services offered under a program induced participants to take further conservation actions? What if they encouraged other consumers to adopt conservation measures without taking advantage of the program’s incentives? They might take action because the program changed their perceptions about the benefits of conservation, or because the increase in demand induced a shift in supply, making energy-efficient products more available.
These behavioral effects on participants (participant spillover) and consumers in general (non-participant spillover or market transformation), although they’re hard to quantify, can be sizable. Joskow and Marron recognized the validity of this proposition, but didn’t explicitly account for these effects in their analysis.
Motivation and Social Desirability
A variety of methods have been used to either measure or account for freeridership. These methods fall into one of two general categories. The first is the general difference-in-differences approach, which involves comparing actual energy consumption of participants before and after they participate in a program to change consumption among a comparable group of non-participants in the same period.
Implemented properly and with a well-chosen comparison group, this quasi-experimental research design produces reasonably reliable results for net savings, but doesn’t provide separate estimates for the components of NTG, freeridership, spillover, and market transformation effects, individually. The method is often implemented using regression-based techniques to control for residual difference between the two groups, evaluate the sensitivity of savings to various factors, and estimate savings for individual measures for programs that bundle measures.
The main limitation of this approach is that it isn’t well suited for measuring savings for programs involving large commercial and industrial consumers. These consumers tend to be unique in many ways, identifying a comparable group of non-participants is often impractical. Savings, relative to total consumption, may also tend to be too small to measure against the many unpredictable factors that affect energy consumption of these consumers. It’s also less effective in new construction programs, where the lack of pre-program data doesn’t allow a complete comparison.
The second, and by far the more commonly used, group of methods rely on “self-report.”At a basic level, self-report involves asking participants a series of questions about what they would have done in the absence of the program. Responses are then scaled, weighted, and combined to produce a composite freeridership score (or index) for each respondent. The scores for individual respondents are then weighted (by their savings) and averaged to produce a program-level freeridership fraction.
The obvious limitation of the self-report approach is that it doesn’t produce an NTG ratio. Other components of NTG—spillover and market transformation effects—have to be estimated separately and then factored into the calculations. But eliciting reliable information about intentions and motivations can be thorny.
Using surveys to assess freeridership also raises concerns about response bias, particularly those biases involving social desirability, which is the tendency of respondents to gauge their responses to conform to socially acceptable values. This issue is well recognized in social sciences, and it’s discussed in a vast body of academic and professional literature, including conservation program evaluation manuals. 16
One aspect of social desirability is the