Pilot projects are demonstrating the potential of smart metering and smart rates to make the most of supply and demand resources. But as empirical studies show, not all pricing designs are equally...
Low-Income Reality Check
Evaluating the impact of dynamic pricing.
As a result, the Track A pilot suffers from sample bias, and doesn’t provide a reasonable platform for making generalizations. Generalizing to a larger population from information generated from a biased sample won’t result in reasonable projections.
In addition, as was the case with the CL&P pilot, the Wood/Faruqui article arbitrarily treats a lower income classification associated with the CSPP as “low income.” In the Charles River report, the results of the CSPP were evaluated in part by grouping participants into two income classifications—those with “average annual income equal to $100,000” and those with “average annual income equal to $40,000.” Here too, the Wood/Faruqui article defines the lower income threshold as “low income” and reports its conclusions as if all customers in the lower income classification are verified low-income customers. However, the “lower” income group could have many households that don’t exhibit hardship.
The CSPP also studied customers that were enrolled in the state’s CARE program. The CARE program provides energy assistance to qualified low-income customers. Given the income qualification of the CARE customers, the results relating to the CARE program are important to consider. The Charles River report on the CSPP summarizes as follows regarding the CARE program participants:
Thus, according to the Charles River report, the true low-income customers studied in the CSPP are much less price responsive. The Wood/Faruqui article’s conclusion that the CSPP shows that high and low-income customers don’t exhibit a “substantial” difference isn’t consistent with the data from verified low-income consumers.
Similarly, evaluation of the PG&E SmartRate Tariff program has found the participating low-income customers in the CARE program to be much less demand-responsive than non-CARE customers. The 2009 report on the SmartRate tariff, prepared by Freeman, Sullivan & Co., concludes:
The 2010 Freeman, Sullivan & Co. report states:
The year-to-year decline in low-income CARE customer response is also important to note—CARE customer response declined from 11 percent in 2008 to 7.5 percent in 2009, while non-CARE customers’ response was unchanged. The persistence of customer response to dynamic pricing among low-income consumers is important to the long-term success of the programs and may need to be addressed in the design of dynamic pricing programs.
Further, in analyzing the Pepco PowerCentsDC pilot, the Wood/Faruqui article correctly notes that the observed response of low-income customers is lower than high-income customers. However, the underlying data appears to be less than ideal. The final report on the Pepco PowerCentsDC Program states:
Thus, the inclusion of low-income consumers in the Pepco pilot ran into problems, and low-income representation was less than planned. The Pepco pilot also illustrates the policy problem of incorporating a more transitory low-income population into a dynamic pricing program.
Winners and Losers
The Wood/Faruqui article states that it simulates the impact on the average monthly bill from the implementation of critical peak pricing (CPP), and concludes that more than half of low-income consumers will benefit immediately from CPP rates.
Some detail regarding the nature of the samples of residential and residential low-income customers used in the simulations would be informative. For example, it isn’t clear whether