Koichiro Ito again has used a discrete event to develop a “control” for an economic experiment. In this case, he has studied PG&E’s 20/20 rebate program in 2004. The “event” he uses is the eligibility date for the program–he uses new customers who connected to service just before and after that date. He finds that the program had almost no effect on coastal customers but that it was effective in reducing energy use for low-income inland consumers. 

Previously, he had looked whether tiered-block rates were better at inducing conservation across the entire pool of customers. The final version of his paper was published February in the American Economic Review. Discerning the true effects of tiered-rates has been very difficult due to the endogeneity problem–consumers essentially set their own marginal price by choosing their consumption level. Many studies have been conducted in both water and electricity trying to tease out this effect, but the results have always been questionable for this reason. 

Ito was able to use two key facts in his latter study: 1) the 2001 California electricity crisis caused rates to rise rapidly and 2) the SCE and SDG&E service areas are closely interlocked across similar communities in southern Orange County. He was able to run an after-the-fact experiment with two treatment groups that had similar socio-economics and were exposed to the same media market. It’s as if two groups of customers were presented with two different sets of rates from the same utility–a truly unique situation that probably can’t be duplicated. He found that the tiered rates induced no more change in energy use than simple average rates.

These well-done studies can cause policymakers to ask whether complicated proposals that seem to mitigate various concerns are truly effective. In these two cases, the answers are largely “no”.