Power to Choose?An Analysis of Choice Frictions in the
Residential Electricity Market
Ali Hortasu, Seyed Ali Madanizadeh and Steven L. Puller
Many jurisdictions around the world that have deregulated formerly regulated utilities, suchas electricity and natural gas, have opened retail markets to competition and allowed customersto choose their retail provider. However, inertial decisionmaking can diminish consumer benefitsof this policy. Using household-level data from the Texas residential electricity market, wedocument evidence of inertial decisionmaking a majority of households continue to purchasepower from the incumbent despite the fact that switching to a new entrant retailer would reduceelectric bills by around 8%. We estimate an econometric model of decisionmaking to measurethe size of two sources of inertia: (1) consumer inattention, and (2) a brand advantage thatconsumers afford the incumbent. We find that both sources of inertia are prevalent. Householdsthat buy from the incumbent consider alternative retailers in less than 2% of months. Evenwhen they search for alternative providers, consumers attach a substantial brand advantage tothe incumbent that discourages switching to a lower-priced retailer, despite the fact that thepower is technically identical. However, this brand advantage diminishes substantially over timeand becomes relatively small after several years of retail choice. Using the parameters of ourmodel to conduct counterfactual experiments, we find that a low-cost information interventionhas the potential to increase consumer surplus. These findings suggest that careful marketdesign to address inertia during the transition to retail competition can substantially benefitconsumers.
Hortasu: University of Chicago and NBER, email@example.com; Madanizadeh: University ofChicago, firstname.lastname@example.org; Puller: Texas A&M University and NBER, email@example.com.We thank the University of Chicago Energy Initiative and EI@Haas for generous financial support. Weare grateful for assistance with data and institutional questions from Kelly Brink, Robert Manning, CalvinOpheim, and Jess Totten. We thank Tim Brennan, Severin Borenstein, Nancy Rose, Catherine Wolframand numerous seminar participants for very useful comments.
1 IntroductionThe deregulation of formerly regulated utilities has brought about more choice for energyconsumers. Across the world, electricity and natural gas consumers increasingly have theability to choose their retail provider. For example, households in over a dozen U.S. statesand over one-half of households in Europe have retail choice in electricity and/or natural gas.Consumers who previously purchased services from a utility are allowed to buy from otherretailers at prices that are not regulated. This creates new markets where entrant retailersprocure wholesale energy and market that energy to customers. This expansion of retailchoice has been touted to have several benefits. Creating competition for the provision ofutility services can lead to more competitive pricing in the short-run. In addition, introducingcompetition can create incentives to provide customers with new value-added services.
However, choice frictions can diminish the benefits of retail choice. Households who neverhave had the power to choose may not exercise the option to select an alternative lower-priced energy retailer. For example, households may not actively acquire information aboutother energy retailers, even if that information would indicate that better options exist.Alternatively, households may value the brand name of the incumbent the old utility andthis may reduce the amount of switching to new entrant retailers. Both of these sources ofchoice frictions can reduce consumer gains of retail choice.
In this paper, we study a particular retail choice program to measure the size of choicefrictions and to understand the underlying mechanisms. The Texas residential electricitymarket provides an excellent setting to investigate retail choice. Beginning in 2002, residen-tial electricity customers in Texas were allowed to choose their retail provider. Initially, allhouseholds were by default assigned to the incumbent. Then in every subsequent month,households had the option to switch to one of several new entrant electricity retailers. Inorder to inform consumers and provide transparency to the search process, the Public UtilityCommission of Texas created a website www.powertochoose.com. This website was intendedto create one stop shopping. Consumers can search among all retail options available, seesalient information on the plans offered, and easily switch retailers on-line.
Aggregate data from this market suggest strong evidence of choice frictions. Figure 1plots the prices being charged by both the incumbent and new entrant retailers.1 Althoughprices varied over time, the incumbents price was consistently higher than several of thenew entrants. This suggests that households could reduce their electricity bills by switching
1As we describe below, we study a particular service territory for which we could obtain household-leveldata; these figures represent aggregate data for that service territory.
from the incumbent to a new entrant retailer. However, Figure 2 shows that the incumbentheld on to its market share leadership throughout the first four years of the market. Theincumbent share only gradually eroded over time, and even after more than four years ofretail choice, the incumbents market share was over 60%. As we show in detail in section 3.1,a large number of households did not switch away from the incumbent, even though theywould have significantly reduced their electricity bills. Switching to a new entrant retailer a one-time action that would take approximately 15 minutes to complete would reduce theaverage electricity bill by around $100 in the first year, which represents about 8 percent ofelectricity expenditures.
Given this evidence of choice frictions, the goal of the paper is to understand the causesof this seemingly inertial behavior of residential electricity consumers. We develop a modelthat allows us to distinguish between two different mechanisms that could account for thisinertia:
1. Inattention / Status Quo Bias: Households may not choose to search for alternativeretailers. Even though www.powertochoose.com is a click away, the household may notbe aware of it, or the household may exhibit a status quo bias so that it does notconsider the offerings of alternative retailers.
2. Incumbent Brand Advantage / Product differentiation: Even if a household is awareof other retailers, consumers may view the retailers as vertically differentiated. Forexample, households may believe that service during power outage events or otherdimensions of power quality could differ across retailers. Such beliefs, even if technicallyincorrect, may be a source of vertical product differentiation. Alternatively, consumersmay believe that characteristics of customer support such as ease of paying bills varyacross retailers.
Understanding the mechanism driving the inertial behavior will inform the design ofpolicy to enhance the consumer benefits of retail choice. For example, if consumer inattentionto the availability of choice is the driver of inertia, then regulatory nudges for consumers tobecome informed about retail options could encourage switching and benefit consumers. Onthe other hand, suppose that the primary reason households continue to purchase from theincumbent is that they believe the incumbent provides more reliable power. (Technically,the delivery of power via local powerlines and all metering operations are not a function ofthe retailer, but consumers may not be aware of this.) In this case, information campaignsto inform consumers that its all the same power may induce households to choose lower
price retailers. Finally, if this incumbent brand effect diminishes with time, then any initialbrand advantage, and the resulting inertia, may be viewed by policymakers as part of thetransition costs to retail competition.
In order to estimate the magnitude of each source of inertia, we develop an econometricmodel of household choice in section 4 that nests both types of inertia within a two-stagediscrete choice framework. In each month, the household enters a two-stage model. Inthe first stage, the household decides whether to consider an alternative electric retailer.If the household does not consider alternatives, then the household stays with its currentretailer for the following month. However, if it does consider alternatives, the householdenters the second stage. In this stage, the household observes the retailers in the choiceset (which are easily available on www.powertochoose.com) and chooses the retailer thatmaximizes utility. In this second stage, we allow for vertical product differentiation to enterthe households decision, thus capturing potential brand advantage by the incumbent. Insection 4.3, we provide transparent conditions for the identification of model parametersfrom sample moments. We show that the first stage decision probabilities that captureinattention are separately identified from the second stage choice probabilities that captureincumbent brand advantage.
We estimate the model using household-level choice data from the first four years ofretail choice in Texas. Our results indicate that both inattention and the perception ofbrand differentiation explain the market dynamics shown in Figures 1 and 2. Householdinattention plays a role in the inertial behavior. We estimate that the average customer ofthe incumbent only searches for other retail options in about 2% of months, or approximatelyonce every 4-5 years. While the percentage of households who actively search in a