Active Market Share: measuring competitiveness in retail energy markets

Active Market Share: measuring competitiveness in retail energy markets

Utilities Policy 8 (1999) 213–221 www.elsevier.com/locate/utilpol Active Market Share: measuring competitiveness in retail energy markets David Loomi...

892KB Sizes 0 Downloads 39 Views

Utilities Policy 8 (1999) 213–221 www.elsevier.com/locate/utilpol

Active Market Share: measuring competitiveness in retail energy markets David Loomis b

a,*

, Eric Malm

b

a Illinois State University, Campus Box 4200, Normal, IL 61790, USA Insights Unlimited Inc., 222 W. Lancaster Ave. #253, Devon, PA 19333, USA

Accepted 9 May 2000

Abstract As retail electric and gas markets deregulate, market share measurement becomes critical for marketers, regulators, and incumbent utilities. Yet traditional market share measures miss important features of these network industries. In this paper we model provider choice in network industries and develop two alternate market share measures—The Active Market Share (AMS) and the New Mover Market Share (NMMS), that are based on ‘active demand’. These measures are shown to provide more accurate real-time measures of market activity. The NMMS is a special case of the AMS which is easy to measure empirically. Numerical simulations are used to provide comparisons between each measure over time. Both the AMS and NMMS will be important tools for anyone interested in measuring the competitiveness of deregulating markets.  2000 Elsevier Science Ltd. All rights reserved. Keywords: Deregulation; Market share; Retail markets

1. Introduction As retail electric and gas markets deregulate, regulators, marketers and incumbent utilities are faced with a new challenge: accurately measuring the degree of competition in these newly formed markets. Existing tools, like aggregate market share measures, are easy to adopt, but may not convey an adequate picture of the market. We argue that the traditional measure of aggregate market share misses important elements of the retail gas and electric markets. Instead, we propose two new tools based on the concept of ‘active demand’ which provide more accurate real-time measures of market activity than do traditional aggregate measures. There are several features of network industries (like retail electric and gas) that confound traditional market share measures. Most significantly network industries are characterized by passive demand, the idea that a customer will continue to buy a service from the same provider each month unless they actively shop and/or * Corresponding author. Tel.: +1-309-438-7979; fax: +1-309-4385228. E-mail addresses: [email protected] (D. Loomis), [email protected] (E. Malm).

switch. This feature biases market share numbers towards past purchase decisions. The ‘time skew’ is particularly important since in most network industries there was an incumbent provider that served all customers. In some cases the incumbent also provides ‘default’ service for customers who do not select a provider. In this paper, we present two alternative measures of market share that more adequately reflect the underlying demand by consumers. We develop a measure called the Active Market Share (AMS) which adjusts for differences caused by active and passive demand. We then describe an empirically estimable proxy for AMS called the New Mover Market Share (NMMS). By focusing on the provider choice decisions of customers who are actively choosing providers during the new connect process, the NMMS provides a more accurate measure of market competitiveness than do measures based on aggregate numbers. The AMS and NMMS will be of particular importance to regulators who are attempting to measure the competitiveness of newly deregulated retail energy and local telephone markets. The paper is structured as follows. Section 2 presents an overview of provider choice in retail electric and gas markets, using competitive experiences in PA and GA

0957-1787/00/$ - see front matter  2000 Elsevier Science Ltd. All rights reserved. PII: S 0 9 5 7 - 1 7 8 7 ( 0 0 ) 0 0 0 0 2 - 3

214

D. Loomis, E. Malm / Utilities Policy 8 (1999) 213–221

as examples. Section 3 reviews standard market share measures and highlights issues of special importance for energy markets. Section 4 simulates a variety of market outcomes using a model of shopping and switching. The simulation results illustrate the effect of active and passive demand on traditional market share measures, and compares these standard measures against one based solely on customers in the active choice set. Section 5 presents an empirically measurable market share indicator which we call the New Move Market Share. Section 6 concludes and discusses the implications of the AMS and NMMS for regulators.

2. Provider choices in network industries In this paper we are concerned with customer provider choices in retail electric and gas markets, industries which are typically considered network industries. We define a network industry as one in which: 앫 customers must be attached to a fixed network to buy service (local gas distribution system, the ‘grid’) 앫 once attached, customers decide on a provider of network services (gas supply, generation) 앫 a single regulated entity maintains the network 앫 after an initial choice is made, purchases recur automatically unless a customer chooses to switch providers While our analysis focuses on retail electric generation and gas supply markets, the analysis can be extended to other industries with similar characteristics.1 In network industries, customers can exercise choice in two distinct service phases. During the ‘initial sign-up’ phase, customers are getting attached to the network and must be connected to a provider of network services.2 Institutional rules and procedures govern the extent to which customers can exercise choice during this phase. During an ‘ongoing service’ phase, customers continuously purchase the service from the provider chosen in phase 1 unless they decide to switch providers. But the consumer will only engage in shopping if their expected savings are greater than switching costs.3 Even then, the consumer will switch if the savings exceed switching costs. The complete switching model is described in Section 4.

1 Others include cable, Internet and local telephone. For a more theoretical analysis, see Loomis and Malm (1999). 2 The manner in which customers are linked with providers during this phase is largely dependent on regulation. While the operational details surrounding new connect have important implications for market share, these issues are not dealt with at length in this paper. 3 Following Klemperer (1995) we define switching costs broadly to include shopping costs, psychological costs, and explicit costs.

In the remainder of the section we will detail the provider choice processes in deregulating electric and gas markets. We use Pennsylvania as a case study for electricity deregulation, and Georgia (AGL) to discuss gas deregulation. These case studies are important both because they provide early real world evidence about deregulation, and because they illustrate significantly different strategies for implementing retail choice. 2.1. Electricity generation Wide regional price variation has spurred legislative initiatives intended to increase competition and lower prices, particularly in high cost states. Changes at the wholesale level, made under FERC’s guidance and jurisdiction, will likely lead to lower generation costs. Many believe that wholesale competition is sufficient to extract the benefits of improved technology for generation and little will be gained through retail competition. In fact, they believe that high transaction costs at the retail level may cause a net increase in costs to end users. Others state that wholesale competition is insufficient and that customer choice is necessary at the retail level. They point to green energy and value added services as added benefits to retail choice. Regardless of the arguments for and against retail competition, it is apparent that states are moving beyond wholesale competition. Most states are pursuing legislative agendas that extend full customer choice to the retail level. While there is still debate about how much value (beyond lower prices) a competitive retail market would provide consumers,4 legislatures in 30 states are aligning themselves on the side of full retail competition.5 Prior to deregulation every customer was served by an incumbent vertically integrated electric utility that bundled generation, transmission and distribution charges into a single per kWh rate. The trend in electricity deregulation is to unbundle generation services from transmission and distribution (‘the wires’), with the generation component of the bill becoming competitive. While generators can differentiate their products using price, generation mix (‘green’ power), and billing options, the energy delivered to a retail customer will be identical regardless of the supplier chosen. While wholesale energy is regulated at the federal level, retail electric competition is controlled at the state level. Thus far, state utility commissions have taken an active role in developing retail energy markets. Typically, the state commission sets a timeline for competition, establishes rules, and oversees broad education programs to ensure that customers understand how they

4

See Joskow (2000) and Jacobs (1999) for an overview of the ‘wholesale/retail’ debate. 5 Energy Central News Release (www.energycentral.com) 2/5/99.

D. Loomis, E. Malm / Utilities Policy 8 (1999) 213–221

are impacted by competition. Since the choice of a generator is an entirely new purchase decision for all but the largest commercial customers, states have a formidable task in both educating customers about choice and ensuring that energy generators and marketers have equal access to the market. Pennsylvania was one of the first states to deregulate its retail energy markets, and has been widely cited as an example of early deregulation success. Following a limited pilot program in late 1997, Pennsylvania extended the option of buying electricity on the competitive market to two-thirds of customers by January 1999. During the summer of 1998, the state commission developed and implemented a statewide consumer education program. The media campaign informed customers about choice, taught them how to shop for electricity, and provided each customer with a list of licensed suppliers. Approximately 2 million customers enrolled in the program, although only about 8% of customers have switched to a competitive supplier.6 All customers who are new to the network, however, must initially buy power from the incumbent. Requiring all new customers to buy from the incumbent utility, PA ‘new connect’ procedures do not promote participation in competitive markets. Since retail customers need their incumbent utility account number in order to choose a competitive provider, they must wait two to six weeks until they receive their first bill before they can choose and switch. The waiting period, and requirements to provide their incumbent account number and signed contract effectively increase switching costs and thus impede competition.7 2.2. Natural gas Retail natural gas markets have also begun to open up.8 Like electric, retail gas customers receive the gas commodity through a distribution system (here a set of pipes) that are maintained by a local regulated monopoly. In deregulated service areas customers may choose the company that supplies the gas commodity, although the local distribution company (LDC) continues to maintain the local distribution network. Deregulation of the Atlanta Gas Light (AGL) service area (in Atlanta, GA) attracted national attention in 1999 because of the speed and aggressiveness of the effort Atlanta Gas Light Company, 1999. Unlike most other deregulation efforts in electric or gas, AGL actually exited the merchant function (they ceased to sell the gas commodity to retail customers). During a brief transition

period marketers competed for customers.9 At the end of the transition period, customers who did not choose were randomly assigned to competitive suppliers. Suppliers received customers based on the percentage of ‘switchers’ they received in the transition period.10 The AGL case illustrates strongly pro-competitive rules. While some have criticized the ‘random assignment’ of passive customers to marketers, the absence of a single default provider makes ‘active’ choice inevitable. New customers, for example, can only establish service by contacting a competitive supplier. For AGL customers, deregulation was hard to miss. The PA and GA experiences illustrate how different regulatory frameworks result in dramatically different (short-term) outcomes. After several years of competition, about 8% of PA customers are buying electricity from a competitive supplier. In contrast, within a year all AGL customers switched to a competitive provider. Similarly, rules surrounding ‘new connect’ procedures have big impacts on choice—while no new customer in PA can exercise choice, all new customers in AGL must participate in the competitive market. It is hard to tell whether AGL’s aggressive policies will result in more vibrant markets—or will benefit customers more—than PA’s more conservative policies; yet it is clear that the regulatory environment will have important impacts on deregulation where ever it occurs.

3. Standard measures of market competitiveness Market share information is an important gauge of the competition in a market. At its most basic level, market share for firm i is typically measured by dividing the sales of firm i, by total sales.11 Yet complications in the electric industry challenge tried-and-true measures of market share such as the Herfindahl–Hirschman Index and the Gini coefficient. At the wholesale level, market power and strategic behavior in the real-time power trading market challenge traditional concentration measures.12 At the retail level, these measures are similarly deficient but for different reasons. Passive demand, imperfect information, and institutional factors similarly confound traditional retail market share measures, and give misleading and erroneous indications of the true state of retail competition.

9

See www.aglc.com/deregulation for a summary and timeline. For example, if supplier A won 20% of customers who switched during the transition phase, they would also receive 20% of customers who were randomly assigned a default provider. 11 The term ‘sales’ can be expressed using total customers, load, or revenue. 12 See Borenstein et al. (1999) or Sweetser (1998) for a discussion of alternative measures of wholesale competitiveness. 10

6

PA PUC, July 1999. Oddly enough, these consumer protection provisions seem to be ‘protecting’ customers from competition. 8 See the US Energy Information Administration for current details. www.eia.doe.gov/oilFgas/naturalFgas/restructure. 7

215

216

D. Loomis, E. Malm / Utilities Policy 8 (1999) 213–221

3.1. Passive vs active demand In network industries, a consumer makes a decision once and will continue to ‘demand’ service from this provider until he/she chooses to switch. In every month that the consumer does not switch, he/she passively chooses the provider from the previous month. In this arrangement the consumer is not required to actively decide which provider to have each month—if he does nothing, he continues to buy from the same provider.13 This is further complicated by the fact that an incumbent provider typically provides ‘default’ service. Since, in a monopoly environment, consumers had no choice of provider, there was little consequence to this type of arrangement. However, this type of market institution has serious complications for market share measurement. An example illustrates the point. When measuring the competitiveness of car manufacturers, one usually considers the market share of new car purchases in a given month or quarter. This market share would reflect the demand of ‘active demanders’ in the marketplace (showroom) during that time. However, one could also measure market share based on car registrations or cars driven on the road. This ‘registrations’ measure of market share would include ‘passive demanders’ who did not purchase a new car during the period. A ‘registrations’ measure of market share would include information about consumers’ past decisions about car purchases as well as current purchases. Given the institutional arrangements in network industries, we believe that an aggregate market share in network industries is closer to taking the market share of car registrations than the market share of new car purchases. Market share of network industries includes passive and active demanders reflecting current and past choices of service providers.14 A more accurate measure would be to adjust market share to include only active demanders. This measure would be closer to the market share of new car purchases. In most public policy decisions, a current or even prospective measure of market competitiveness is desirable. Fig. 1 can be used to illustrate and compare the traditional aggregate measure with two new measures based on active demand. Section (A) represents the entire network customer base. During a given period of time, a subset of customers will be actively reviewing

and selecting a provider. This ‘active choice’ set is composed of two groups—‘movers’ who must select a provider at time of hook-up (B), and non-movers who are reviewing their service (C). While it is possible to identify all customers who switch (D+E), it is impossible to identify non-movers who are reviewing service but choose not to switch. Table 1 illustrates how each market share measure would be calculated using the areas from Fig. 1.

4. Simulation A simple model of customer shopping and switching can be used to simulate consumer shopping and switching in both static and dynamic states. First, a base case is established to illustrate how the two-stage process operates in a single period, like the initial period of deregulation. Shopping and switching (market) shares are calculated. Next, the effects of customer education programs affecting shopping costs are examined. Customer education programs that effectively reduce shopping costs are shown to increase shopping, but do not necessarily increase switching. Conversely operational changes that reduce the effort needed to switch providers are shown to increase both shopping and switching. The model is then extended by incorporating multiple periods to model market share dynamically and illustrate the effects of customer movement within and between service areas. As customers enter the network, they face a different set of choice costs than do existing customers. Here we see that changes in aggregate market share can be driven by the choices of customers who are entering the system. 4.1. The model During ongoing service, shopping and switching is a two stage process. Stage 1, shopping, is described using Eq. (1). Customers will shop if their expected savings is greater than switching costs. Here PI is the price of the incumbent (or current provider) and is known with certainty.15 The term E(PC) is the expected price of the competitor and is not known with certainty prior to shopping.16 The usage of customer i is represented by qi. The

13

While the overall market elasticity of electricity is inelastic, the individual firm elasticities could be elastic if consumers freely switch from one provider to another. Yet this institutional arrangement of default service each month makes the firm elasticities more inelastic than they would otherwise be. 14 One important difference is that passive demanders in the car market do not continue to pay the manufacturer for the car. In network industries, consumers continue to pay rates set by the service provider that they have passively chosen.

15 The term PI is often referred to as the ‘shopping credit’. See Jacobs (1999), Tschamler (1999) or Regulatory Assistance Project (1999). 16 In the simulation below we assume that the expected price of the competitor is related to the price of the incumbent through a linear expectations operator.

D. Loomis, E. Malm / Utilities Policy 8 (1999) 213–221

Fig. 1.

A graphic depiction of customer switching.

Table 1 Market share calculation using areas from Fig. 1 Measure

Area

Active Market Share New Mover Market Share Traditional Market Share

(D+E)/(C+B) D/B (D+E)/A

217

New customers choose the competitive provider if : [PI⫺E(PC)]qi⬎W

(3)

Once a customer shops, he or she may choose to switch providers. Eq. (2) shows the switching decision. The price of the competitor is now known with certainty. The H term drops from the right hand side since the customer has already exerted the effort to gather information.

Price expectations, ‘shopping credits’ and shopping and switching costs drive the model. Prices are not known with certainty prior to shopping, thus customers form price expectations for competitive suppliers based on general advertising, media, etc. The price reduction given if the customer chooses an alternative provider, sometimes called the ‘shopping credit’, is set in the regulatory environment and can be considered a policy variable. Even if this credit does not reflect underlying costs, the shopping credit becomes the implicit price charged by the incumbent. Expected and actual savings are then compared with customer shopping and switching costs to produce shopping and switching behavior. It is important to note that shopping and switching costs are also policy variables, since they are heavily impacted by customer sign-up procedures.19

Switch if : [PI⫺PC]qi⬎W⫹E

4.2. The base case

terms H, W, and E represent shopping costs, brand equity, and switching effort, respectively.17 Shop if : [PI⫺E(PC)]qi⬎H⫹W⫹E

(1)

(2)

New customers face a slightly different decision, which is captured in Eq. (3). Since each customer must select a provider of network services when they are signing up with the regulated network provider the E term drops from the right hand side. Customers choose the competitive supplier if the expected savings exceed the psychological costs of switching (W).18 17 Shopping costs can be thought of as the costs of collecting information about alternate offers. Brand equity captures psychological costs of switching away from the incumbent; it also captures customer uncertainty about buying from an ‘unknown’ competitive supplier. The effort term refers to the effort customers must go through to actually switch. This may include filling out forms, finding account numbers, and contacting the regulated network provider. 18 In the simulation we assume that E(PC)] is a linear function of the price of the incumbent. The parameter ⬀ describes the anticipated price difference.

By assigning numerical values to the elements in Eq. (1) and Eq. (2) we simulate market outcomes in a network industry. Table 2 shows the parameters for the base case.20 Prices for the incumbent and competitor are set at 5.6 and 4.9, respectively.The expected savings (⬀) are assumed to be 25%. Customers face shopping costs of 400 and must exert switching effort of 500. Customers vary both in their level of usage (qi) and in their perceived brand equity in the incumbent (wi). Customers decide whether or not to shop according to Eq. (1), and decide whether or not to switch providers according to

19

For example, the NJ BPU responded to sluggish customer signup statistics by announcing that sign up procedures would soon be modified to allow for sign-ups over the Internet. See Johnson (2000). 20 The parameters for the base case were chosen to mirror the prices and outcomes of the PA electric pilot.

218

D. Loomis, E. Malm / Utilities Policy 8 (1999) 213–221

Table 3 Cases 2 and 3: education and procedure

Table 2 The Base case Case 1: Base case PI PC ⬀ Wi is distributed Qi is distributed H E Shop Traditional market share Active market share

5.6 4.9 0.75 N(700,300) N(1000,300) 400 500 31% 8% 26%

Eq. (3). In the base case 31% of customers ‘shop’, and 8% of customers ‘switch’. According to traditional market share measures, the competitive provider would be assigned a market share of 8%.21 We argue that the market share should be framed with reference to the active choice group. The Active Market Share measure of 26% shows that over one quarter of customers who shopped chose the competitive provider. By focusing on the ‘shopping’ group, the measure also calls attention to how many people are in the active choice set. In practice a regulator would be concerned with both the numerator and denominator of AMS—(the number of people switching and the number of people shopping.) 4.3. The roles of education and procedure Anyone familiar with deregulation in network industries is aware of the importance of customer education efforts, and the challenges associated with implementing sign-up procedures that allow for choice in previously monopolistic markets. In Cases 2 and 3 we model the impact of two types of programs that are typically under regulatory jurisdiction: customer education programs and switching rules/procedures. The parameters for these cases are shown in Table 3. Case 2 models the effects of an education program which focuses on reducing customer shopping costs through the provision of provider information. The shopping costs parameter (H) is lowered from 400 to 100. This results in an increase in the amount of shopping, but does not necessarily lead to increased switching. The Active Market Share decreases from 26% to 12%. In Case 3 the switching effort parameter (E) is reduced from 500 to 100, simulating a change in operational procedures that makes it easier for customers to sign up

21 The parameters of the base case were chosen to mirror the initial outcomes of the PA electric market.

PI PC ⬀ Wi is distributed Qi is distributed H E Shop Traditional market share Active market share

Case 2

Case 3

5.6 4.9 0.75 N(700,300) N(1000,300) 100 500 67% 8% 12%

5.6 4.9 0.75 N(700,300) N(1000,300) 400 100 64% 40% 63%

with a new provider.22 The reduced level of switching effort is shown to increase both shopping and switching. This increases the Active Market Share to 63%, indicating that a higher proportion of ‘shoppers’ are electing to choose the competitive provider. While big advertising budgets can boost customer awareness of competition, cumbersome sign-up procedures can hinder participation.23 These simulation results suggest that simplifying customer sign up procedures may be at least as important as a widespread customer education program. This is especially true over time; sign up procedures persist long after the ad campaigns are over. 4.4. ‘Movers’: customer entry into the network Nationally about 15–20% of customers move each year, making entry into the network an important determinant of market share dynamics. In Cases 4, 5, 6 we examine the impact of customer movement and service transfer on choices and market shares. Since customers who are requesting new service do not incur extra ‘effort’ costs for choosing a provider during the new service process, we model customer choice using Eq. (3). In the simulation we assume that 20% of customers (randomly distributed among the customer base) move each period. We make the additional assumption that ⬀, the expected savings, converges to actual savings in the long run.24 Fig. 2 shows traditional market share numbers when new service procedures are structured in a way which allows movers to choose a provider during the new con22

These operational issues could include, for example, regulations that require a customer signature to switch providers, or regulations that require that customers find and give their regulated account number to a competitive provider in order to switch. 23 While NJ’s ‘NJ Energy Choice’ ad campaign achieved high levels of program awareness, initial residential sign-ups totaled under 3000 in the first three months of the program. 24 This eliminates the impacts of a persistent divergence between expected and actual prices.

D. Loomis, E. Malm / Utilities Policy 8 (1999) 213–221

Fig. 2.

219

Residential market share at old and new residence.

nect process. This places each new customer into the ‘active choice’ set described in Section 3. Since every customer who connects to the network must establish service with a provider, customers may make an initial choice without overcoming an ‘effort’ barrier associated with switching from an existing provider. These ‘aggregate’ market share measures are compared with the AMS and NMMS in Section 5, below. A more complicated simulation could combine, for example, the impacts of moving and diminishing incumbent brand equity in a single dynamic model. This would effectively create two types of customers who are in the ‘active choice’ set—(i) ‘movers’ and (ii) current customers who decide to shop because of changes in underlying parameters (like brand equity, the competitive price, etc.). The empirical challenge—and a practical solution—for monitoring the decisions of customers in the active choice group are developed below. From a dynamic perspective, it is important to recognize the importance of customers who are entering the network. Since all customers were initially ‘new customers,’ making choice available (and easily accessible) to customers during the new connect process is important to the long run health of competitive markets. Assigning people to a ‘standard offer’ or ‘default supplier’, with the option of switching to a competitive provider is likely to favor the default supplier.25

Table 4 Simulated Market Share comparison Measure

Year 1

Year 2

Year 3

Year 4

Traditional market share New mover market share

8% –

18% 60%

29% 60%

38% 50%

Table 4 utilizes results from the simulation in Section 4 to show the lagging nature of the aggregate market share. Here we see that healthy levels of customer choice during the move process slowly elevate the traditional market share of new competitors. Yet waiting years to determine whether the market is functioning properly is not acceptable to the practitioner. The New Move Market Share (NMMS) provides an immediate gauge of the actions of customers who are in the active choice set. Using the automotive market share analogy developed in Section 3, the NMMS provides a measure of the decisions of people in the showroom. Most importantly, the New Move Market Share is easy to estimate empirically. Every network provider knows when a customer enters the system. Because each customer must be associated with a supplier, the network provider also knows what supplier choices each new customer made. Alternately movers can be easily identified and surveyed outside of the network provider’s systems. Fig. 3 shows the slow shift in market share in the long

5. The New Mover Market Share: an empirical measure The New Mover Market Share provides an important gauge of the current health of a deregulated market.

25 See Johnson et al. (1993) for a discussion of default service options in the insurance industry.

Fig. 3.

Residential market share at old and new residence.

220

D. Loomis, E. Malm / Utilities Policy 8 (1999) 213–221

distance market.26 As a group, movers shifted away from AT&T toward alternate providers. Yet mover activity is even more dramatic than Fig. 3 suggests. About 24% of residential customers switched providers when they signed up for new service. This high switching rate— both to and from AT&T—signals that the long distance market is experiencing healthy competition. The initial connection into the network is the one time that all customers must choose (or be assigned) a provider. Understanding the choices of customers entering the network is important if one is to understand the market as a whole. The results in Table 4 illustrate that high levels of ‘new customer’ switching gradually result in changes in traditional aggregate market share measures. Given the passive nature of customer demand, the availability of choice during the new connect process should be an important consideration for those involved with deregulation.

6. Conclusions In most industries aggregate market share is a satisfactory measure of market activity. Yet the role of ‘passive demand’ in deregulating retail electric and gas markets makes aggregate market share a poor ‘real-time’ measure of activity in these new markets. In this paper we propose two alternative market share measures—Active Market Share (AMS) and New Move Market Share (NMMS)—which provide a more accurate gauge of activity in these newly deregulated markets. Customers going through the ‘new connect’ process must choose (or be assigned) an electric or gas supplier, in what we call the initial service phase. Once service is established customers continue to passively demand service from their provider unless they actively shop (gather information on prices, etc.) and possibly switch (which also requires effort). Passive demand, and in some cases the presence of an incumbent default provider, makes the traditional aggregate market share a poor measure of the health of a newly deregulated electric or gas market. By using the Active Choice Set (those people who are shopping) as the denominator in a market share calculation we produce a real-time measure which we call the Active Market Share (AMS). AMS reflects the decisions of people who were actively choosing during the period in question. AMS also focuses attention on the size of both the ‘shopping’ and ‘switching’ groups. Simulation results show a lagging relationship between AMS and aggregate market share. Over time market share levels

26 Data is from Insights Unlimited Inc., 1998 survey of residential new-connects.

among the active choice set gradually impact the aggregate measure. Simulation results also have important implications for those involved in developing rules and customer education programs in deregulating states. Cumbersome sign-up procedures act as a barrier to choice. Our results show that while customer education campaigns may increase the number of people in the active choice set, they do not necessarily result in switching. Changes in procedure that make it easier to sign up with a competitive supplier are shown to increase customer switching. The New Move Market Share (NMMS) is a special case of the AMS that focuses on customers who are signing up for service in the initial service phase. Since new customers are easy to identify and each new customer must choose (or be assigned) a supplier, the NMMS is an easy-to-estimate proxy for the AMS. Because 15– 20% of all customers move each year, virtually all customers are included in the new customer group over time. Simulation results show how healthy levels of choice among new customers gradually results in changes in aggregate market share. Real world measures of NMMS in the long distance phone industry reflect high levels of switching and a gradual deterioration of the incumbent’s aggregate market share. Active Market Share and New Move Market Share are important tools for anyone who is interested in understanding newly deregulated retail electric or gas markets. In addition to providing real-time measures of market activity, AMS/NMMS reveal the importance of newly evolving rules and regulations surrounding choice in these markets. Making choice available during the initial service phase, and reducing burdensome switching procedures during on-going service, will help spur competition.

References Atlanta Gas Light Company, 1999. Deregulation Information. www.aglc.com/deregulation Borenstein, S., Bushnell, J., Knittel, C., 1999. Market power in electricity markets: beyond concentration measures. The Energy Journal 20 (4), 65–88. Insights Unlimited Inc., 1998. New Residents Survey—1998. Devon, PA. Jacobs, J.M., 1999. Setting a retail generation credit. The Electricity Journal 12(4), 80-87. Johnson, E.J., Hershey, J., Meszaros, J., Kunreuther, H., 1993. Framing, probability distortions, and insurance decisions. Journal of Risk and Uncertainty 7, 35–51. Johnson, T., 2000. BPU craves more rivalry in utility sector. Newark Star-Ledger, March 17, pp. 1,55. Joskow, P.L., 2000. Why Do We Need Electricity Retailers? Or Can You Get It Cheaper Wholesale? MIT University Discussion Paper, February. Klemperer, P., 1995. Competition when consumers have switching costs: an overview. Review of Economic Studies 62, 515–539.

D. Loomis, E. Malm / Utilities Policy 8 (1999) 213–221

Loomis, D., Malm, E., 1999. Measures of Market Competitiveness In Deregulating Industries. Working Paper, 18th Annual Advanced Workshop in Regulation and Competition, Newport, RI. Regulatory Assistance Project, January, 1999. Setting Rates For Default Service: The Basics. www.rapmaine.org/ilet9901.html. Sweetser, A., 1998. Measuring a dominant firm’s market power in a

221

restructured electricity market, a case study of Colorado. Utilities Policy 7, 243–257. Tschamler, T., 1999. Designing Competitive Electric Markets: The Importance of Default Service and Its Pricing, XENERGY White Paper.