Equity in Residential Electricity Pricing.

• Author: Horowitz, Shira

1. INTRODUCTION

   Wholesale and retail electricity prices are decoupled for most residential electricity customers. Wholesale prices change in real time to reflect marginal cost and can range from negative values to $1000/MWh (1) (PJM Interconnection, 2009). Residential retail rates are typically flat rates, which are load weighted averages of expected price over a certain period of time (2) (typically a year or more). Flat rate (FR) pricing is inefficient because price does not reflect marginal cost, so customers may be under- or over-consuming at any point in time (Borenstein and Holland, 2005; Spees and Lave, 2008). Customers with a high coincident peak relative to their average demand are, on average, paying below marginal cost (3) while customers with flatter usage or those whose peak demand occurs at off-peak prices are paying above the marginal cost they impose, on average (Spees and Lave 2008). Customers that add to peak load impose high costs on the system, but under FR pricing, all customers pay the same amount. This is a policy where customers with high coincident peaks are receiving a cross-subsidy from the remaining customers.

   Real-time pricing (RTP) has the potential to address these problems by directly coupling retail and wholesale prices. The energy charge in a residential RTP tariff changes hourly to reflect either the day-ahead or real-time locational marginal energy price (LMP). This provides a signal for customers to use only the amount of power that they value at or above the current marginal price of power. If customers respond to high prices by lowering usage, RTP can potentially lead to lower peak demand and price. Even if only some customers respond, all customers can potentially benefit from lower marginal price and lower capacity costs due to lower peak demand. Charging customers the RTP is no guarantee that customers will reduce or shift load when price is high. The potential savings must be large enough for customers to invest in the time, education and technology necessary to effectively reduce peak demand. Several utilities including Commonwealth Edison (ComEd) and Ameren currently offer optional RTP tariffs to residential customers. Other utilities offer approximations of RTP such as time-of-use (TOU), where days are divided into peak and off-peak prices for electricity or critical-peak pricing (CPP) where higher prices are triggered by high wholesale prices or a correlated metric such as temperature.

   Borenstein and Holland (2005) show that increasing the share of customers on real-time pricing is likely to improve efficiency. Borenstein (2005) shows that even with small elasticities, gains in economic efficiency from RTP can be substantial. We will not elaborate any further on the inefficiencies of flat rate pricing and the potential efficiency gains from dynamic pricing, since there is already substantial literature on this subject (see: Borenstein et al., 2002; Borenstein and Holland, 2005; Holland and Mansur, 2006; Spees and Lave, 2007; Borenstein, 2005). Instead we will focus on the distributional impacts of dynamic pricing.

   Here, we address a question of practical importance to electric utilities and public utility commissions who are considering a move to dynamic pricing: which consumers "win" (will save money under RTP compared to FR) and which consumers "lose" (lose money under RTP compared to FR) when switching from FR to RTP? Because of the inherent cross-subsidies between customers under FR pricing, when a utility switches to dynamic pricing, the cross subsidy will be reduced (CPP) or disappear entirely (RTP), and the cost burden will shift from customers with flatter loads or non-coincident peaks to those with high coincident peaks. Some customers may experience significant changes in their bills--both increases and decreases if they don't shift their usage. It will be important for utilities and PUC's to know in advance which customers will have large bill increases, so they can supply those customers with information and tools to help mitigate the increased bill by increasing energy efficiency or shifting or curtailing load, or create policies to tax the "winners" and subsidize the "losers".

   The question can be reframed for those utilities that are not considering a switch to RTP: which customers are currently providing cross-subsidies to other customers under FR pricing? Is the wealth transfer caused by the cross-subsidies an acceptable policy from an equity perspective?

   We address these questions by taking a sample of customers and calculating their bill difference under RTP and FR under both inelastic and elastic demand. We treat the scenario with inelastic demand as a zero-sum game used to explore cross-subsidies: one customer's loss is another's gain. This is also a "worst case scenario" for RTP programs, where consumers don't respond, so there are no net savings to consumers. Under elastic demand, there are net savings to consumers due to avoided energy usage, lower marginal prices and lower capacity costs. We then analyze customer characteristics including income and demand. We obtained data from a sample of ComEd customers.

   Borenstein (2012) and Faruqui et al. (2010) also perform empirical analyses of the distributional effects of dynamic pricing. Our analysis differs in several ways: we use an RTP tariff while the other analyses focus on variations of TOU and CPP; we focus on different geographic regions which have different load and price characteristics; our analysis assumes a mandatory tariff, while the other analyses assume opt-in tariffs for their distributional calculations.

   We find that under inelastic demand, only 36% of consumers would save money under RTP. With elastic demand of -0.2 (an upper bound), roughly 50% of customers would save money from reductions in energy usage and energy price. Many more customers save if we assume reductions in capacity costs due to demand response. The customers who save tend to be the largest consumers, while those who would lose money under RTP tend to be smaller consumers and represent a disproportionate amount of low-income customers.

   The remainder of this paper is organized as follows. Section 2 describes the data set. Section 3 explains the analysis for inelastic demand and 4 models elastic demand. Section 5 gives the policy implications of the analysis.

2. DATA SET

   2.1 Usage

   ComEd serves Northern Illinois and the greater Chicago area and is part of the Regional Transmission Organization PJM Interconnection. ComEd has an optional residential RTP tariff currently in use, so we were able to use actual tariffs (adjusted to be revenue neutral, see Appendix B) in our calculations. We have hourly electricity data from a stratified sample of 1260 residential customers from 2007 and 2008. Some of the strata were oversampled, however corrections were made for this in all statistics using the bootstrap method (Appendix C).

   These customers were all on a residential FR tariff, so there are no confounding behavioral factors due to exposure to RTP. We know which of four customer classes each customer belonged to: (1) single family (SF); (2) multi-family (MF); (3) single family with electric space-heat (SFH); and (4) multi-family with electric space heat (MFH) (see Table 1 for summary statistics). We also have data on whether customers received any need-based subsidies (Table 1). There are several income-based subsidies customers can qualify for. (4) We classified any customer that received any need-based subsidy at any point over 2007 or 2008 as "low income". Approximately 6% of customers in the population are low income by this definition.

   The raw data, consisting of hourly household electricity usage were cleaned and verified so that all remaining data were valid. The protocol used for cleaning the data is in Appendix D.

   2.2 Tariffs

   ComEd residential electricity bills are monthly bills and consist of three sections: (1) electricity supply services, (2) delivery services, and (3) taxes and other. There are several different charges in each section. Some charges are the same for both RTP and FR customers; some charges are different and in some cases a charge is exclusive to either RTP or FR. Charges can be either fixed monthly costs or based on the amount of electricity consumed that month (i.e. a cost per kWh). The one exception is the capacity charge, which is applied only to RTP bills. Customers are billed per kW-month of demand, where demand is the customer's average usage during the 10 hours of highest system usage. Appendix A gives details on the rates and how bills are calculated. Table 2 shows the average annual bill for each customer class.

   For customers that are on RTP, the only portion of the bill that changes hourly is the energy supply charge, which corresponds to the wholesale LMP. Over 2007 and 2008 the RTP energy supply charge ranged from -25Cents/kWh to 50Cents/kWh with a mean value of 5Cents/kWh, a median of 4Cents/kWh and a standard deviation of 3Cents/kWh. 90% of prices were between 1Cents/kWh and 10Cents/kWh, while 50% of prices were between 3Cents/kWh and 7Cents/kWh. Prices exceeded 15Cents/kWh 1% of the time. The RTP energy supply charge represents 45% of the total average annual electricity bill. (5)

   The flat rate energy supply charge ranged from 4.4Cents/kWh to 7.6Cents/kWh depending on the month and customer class. (6) All other rates (both marginal and fixed), for both FR and RTP are constant throughout each billing cycle, but may be adjusted, no more frequently than monthly, to reflect changes in cost.

3. ANALYSIS: NO BEHAVIOR CHANGE

   In this section we calculate the difference in annual electricity bill for the sample, had the customers been on RTP, compared to what they actually paid under FR. It should be noted that all of the customers in the sample were on the FR, and were at no point on RTP during this time period. We apply the RTP tariff that was optional for ComEd customers...