Has retail competition reduced residential electricity prices in Texas?

doi:10.21203/rs.3.rs-2864978/v1

A review of 19 extant studies reveals mixed effects of retail competition effect on retail electricity prices in various jurisdictions in the United States. This conflicting evidence is largely due to the differences in these studies’ market locations, pre-competition conditions, data types, estimation methods and features of the restructuring plan. Contributing to the debate on the impacts of electricity market reform on retail prices, we estimate the impact from the introduction of retail competition on retail electricity prices paid by residential consumers in Texas in two cities opened to retail competition. We find that retail competition raised average prices by $0.0112/kWh ($11.2/MWh) in the transition period from 2001 to 2006 and by $0.0134/kWh ($13.4/MWh) during the period of unfettered competition from 2007 to 2020. However, when the US wholesale natural gas prices are relatively low, actual retail electricity prices in areas opened to retail competition are close to the prices that would have prevailed had retail competition not been introduced, as measured by the counterfactual prices estimated using the synthetic control method.

JEL Codes: L51, L94, Z18, Q48

Retail competition

electricity market restructuring

residential electricity prices

synthetic control

Texas

In many regions of the world, the electricity industry has been restructured to foster competition. Every market reform has been different, pre-restructuring conditions have varied, and the outcomes have ranged from success to disaster, and everything in between (Sioshansi, 2006). An analysis of the retail price changes in Texas from industry restructuring is of obvious interest to the Texans paying electric bills. Further, lessons learned from Texas also inform market design efforts elsewhere.

In the immediate aftermath of Winter Storm Uri (UT Austin, 2021), The Wall Street Journal (WSJ) reported on 02/24/2021 that “Texas electric bills were $28 billion higher under deregulation” for the 16-year period of 2004-2019. The WSJ’s bill impact estimate of ~$1.75 billion per year was based on a comparison of the average prices of retailers in the competitive market to the average rates of “full-service” utilities that were primarily non-profit municipal utilities and rural electric cooperatives in Texas during that period. As simple comparisons that fail to control for other factors affecting prices may be misleading, this paper uses the method of synthetic controls (SC) (Abadie, 2021) to estimate the impact of retail competition on residential electricity prices in Texas. This approach yields a counterfactual of what prices would have been, absent the introduction of retail competition in Dallas and Houston, the two largest cities in Texas. Based on the optimal weights established through an SC algorithm, this counterfactual is a weighted average of the prices of utilities that were not exposed to retail competition but had similar pre-restructuring price patterns and attributes, using optimal weights established through an SC algorithm.

Motivated by the WSJ report and our literature review in Section 4, we use synthetic controls (SC) (Abadie, 2021) to find that the introduction of retail competition has generally increased residential electricity prices in Texas, though by a lesser amount than suggested by the WSJ’s simplistic comparison. Based on the annual data for 1990-2020, our statistically significant (p-value ≤ 0.05) estimates of average price impacts are a $0.0112/kWh ($11.2/MWh) increase for the transition period of 2001-2006 and a $0.0134/kWh ($13.4/MWh) increase for the unfettered competition period of 2007-2020. As explained in Section 6 below, these estimated price changes are largely due to relatively high wholesale natural gas prices in some of the years in the 2001-2020 period.

Responding to the WSJ article, we use the ~$0.0134/kWh ($13.4/MWh) estimate noted above to calculate the aggregate bill impact per year of retail competition on the residential customers in Texas directly affected by retail competition in the 2007-2020 period of unfettered competition. Our calculated impact is ~$1.09 billion per year,[1] which is ~$658 million or 38% less than the impact estimated by the WSJ of ~$1.75 billion per year.[2]

Our paper makes four contributions to the literature on the impact of retail competition on prices. First, we contribute further insights into how the introduction of retail competition has impacted residential electricity prices in Texas. This topic, or aspects related to it, have previously been examined by Zarnikau and Whitworth (2006), McKearin (2015), and Hartley et al., (2019), and Brown et al., (2020a), in addition to the WSJ article. Second, it shows how the SC method may be used to estimate the impacts of restructuring on electricity prices, complementing recent papers by Rose et al., (2021) and Hill (2023). Third, it remedies the bias inherent in using state-level or utility-level data from the US Energy Information Agency (EIA) to estimate the impacts of retail competition when only portions of a state have been opened to retail competition or the market shares of retailers within an area opened to competition are not known. Finally, we separately treat the transition period that frequently precedes full retail competition, following Rose et al., (2021) and Hill (2023).

Our paper proceeds as follows. Section 2 reviews electricity industry restructuring in Texas. Section 3 discusses how retail market reform may change how retail rates are established. Section 4 comments on the wide range of results found in previous studies of the impact of electricity market reform in the US. Section 5 introduces our approach and data sources. Section 6 provides our findings, the basis for Section 7 that contains our conclusions and their policy implications.

[1] The EIA reports that 142,100.5 million kWh is the average of total residential electricity sales in Texas in the 16-year period of 2014-2019. 6.198 million residential energy consumers had retail customer choice in 2017, per the PUCT “Report Cards on Retail Competition”, available at: https://www.puc.texas.gov/industry/electric/reports/rptcard/pastrc.aspx; while the EIA reports 10.809 million total residential electricity accounts in the state in that year, suggesting that retail restructuring affected 57.3411% of the state’s households. Thus, the aggregate bill impact per year is $0.0134/kWh increase × 57.3411% × 142,100.5 million kWh = $1.09 billion.

[2] Another benchmark is a 2015 study by the Texas Coalition for Affordable Power which made comparisons similar to the WSJ’s analysis but covered an earlier period of time. That analysis concluded that between 2002 and 2014, consumers exposed to retail competition paid $24 billion more for power, or $2 billion per year, on average. See Blum (2016).

To place these numbers in context, note that expenditures on electricity in the residential sector in Texas (including areas not opened to retail competition) were $15.9 billion in 2010, per the EIA: https://www.eia.gov/state/seds/data.php?incfile=/state/seds/sep_prices/res/pr_res_TX.html&sid=TX.

Texas accounts for over 9% of the nation’s total electricity generation, and the Electric Reliability Council of Texas (ERCOT) wholesale market serves roughly 85% of the electrical needs of the largest electricity-consuming state in the US.

Mirroring the global trend of electricity market reform and deregulation (Sioshansi, 2013), the electricity industry in Texas has undergone gradual restructuring since the mid-1990s (Zarnikau, 2005). Legislation enacted in 1995 required the Public Utility Commission of Texas (PUCT) to establish rules to foster wholesale competition and required the state’s independent system operator, ERCOT, to provide open and comparable transmission access to all market participants in the ERCOT market.

Enacted in 1999, Senate Bill 7 established a framework and timeline to allow retail electric providers to compete and offer diverse pricing plans to residential customers of the investor-owned electric utilities within the ERCOT power region. The same bill enhanced ERCOT’s centralized control by replacing ten utility-operated control centers with a single ERCOT-wide control center to coordinate the operations of generation and transmission providers and maintain reliability. A day-ahead market for ancillary services was created, as well as a real-time market for balancing energy.

On 06/01/2001, pilot programs were launched to offer retail choices to up to 5% of electricity consumers in areas of Texas to be opened to retail choice. Retail competition began on 01/01/2002, allowing consumers to choose among multiple retailers that offer many pricing plans with differing price and non-price terms (Brown et al., 2020a).

Unlike other states where price caps were imposed on the incumbent utility providers of retail electricity during a transition period, the PUCT set “price-to-beat” rates for the retailers affiliated with the former vertically integrated utilities that were “high enough” to attract competitors into the market. With sufficient “headroom,” the price-to-beat rates allowed retailers wishing to enter the market an opportunity to compete on the basis of price against retailers who were successors of the former vertically integrated utilities during the transition period of 01/01/2002 -12/31/2006. Additionally, various stranded costs were recovered from ratepayers during this period.

Since 01/01/2007, retail electricity prices have been largely unregulated, notwithstanding that the PUCT continues to regulate other components of residential electricity bills, including charges for transmission, distribution, and energy efficiency programs.

An important structural change to the ERCOT market was completed on 12/01/2010. Locational marginal pricing (Stoft, 2002) was introduced to better manage intrazonal transmission, thus reducing the real-time energy prices used by ERCOT for wholesale energy bill settlement with market participants (Zarnikau et al., 2014; Triolo and Wolak, 2022).

As a result of the blackouts prompted by Winter Storm Uri in February 2021 (UT Austin, 2021), market design is being reviewed. Most of the proposed reforms will affect the wholesale market, rather than retail competition.

Municipal utility systems and rural electric cooperatives may “opt in” to retail competition. Nueces Electric Cooperative has done so, and the City of Lubbock is in the process of joining the ERCOT power region and plans to offer retail competition beginning in the fall of 2023. Nevertheless, most municipal and cooperative systems (including the large municipal systems serving San Antonio and Austin) have not elected to offer customer choice. Accordingly, roughly 57% of the state’s residential energy consumers can exercise a choice among various retailers for their electricity service.

Prior to restructuring, the investor-owned utilities in the ERCOT power region were vertically integrated and operated their power plants to meet their retail sale obligations. These utilities had limited dependence upon the wholesale market for generation. Under traditional cost-of-service regulation, fuel factors and/or the generation component of retail electric rates reflected the weighted average cost of all fuels used by the integrated utility to generate electricity, as well as purchased power costs. Average cost pricing prevailed (Rose et al., 2021), implying that the revenues received from the utility’s ratepayers were close to the reasonable and necessary costs incurred by the utility. Moreover, there was typically some regulatory lag which slowed the rate changes necessary to match fuel cost changes.

In Texas and some other states, vertically integrated utilities were unbundled into separate generation, transmission and distribution, and retail organizations at the start of retail restructuring. While organizations in these different industry segments could be held under common ownership, limits were placed upon communications and transactions between affiliates operating in different segments of the industry. The generation companies associated with the incumbent utilities were initially required to auction-off the rights to a share of their generation output, ensuring that retailers entering the market could obtain an adequate supply of generation sufficient to complete retail sales. Over time, the volumes traded through ERCOT’s formal wholesale markets grew, and the independent power producer segment of the market expanded.

The variable operating cost of generators “on-the-margin” set market prices in a competitive wholesale market (Stoft, 2002; Biggar and Hesamzaden, 2014). The ERCOT market is highly reliant upon natural gas as a generation fuel, and natural gas plants are typically on-the-margin (Zarnikau et al., 2019). As a result of unbundling and growing dependence upon wholesale markets, retail prices in areas of Texas opened to retail competition have become extremely sensitive to wholesale natural gas prices (Woo and Zarnikau, 2009; Brown et al., 2020a; Rose et al., 2021; Hartley et al., 2019; Hill, 2023).

Prior to retail market restructuring, a change in wholesale natural gas prices might have affected half of the generation component of a consumer’s bill, if half of the utility’s generation mix was from natural gas-fired power plants. But under retail market restructuring, changes in the generation component of retail prices tend to track changes in wholesale natural gas prices. Retail prices in areas of Texas opened to competition have been comparable to prices in areas not opened to customer choice when natural gas prices have been low. When wholesale natural gas prices have been high, a noticeable gap emerges between prices charged by retailers in competitive areas versus prices charged by utilities under traditional regulation or non-profit ownership. Moreover, the high volatility in the US wholesale natural gas prices has led to similar volatility in electricity prices in areas with retail competition and natural gas as the marginal generation fuel.

When retail competition replaces cost-of-service regulation, certain benefits may be anticipated. Restructuring is often undertaken with the hope that competition among retailers will lower prices for consumers.[3] Supplier choice and product differentiation under retail competition are likely to benefit residential customers (Lancaster, 1990; Goett et al., 2000). Faster adjustment of retail rates to changes in wholesale costs may foster efficiencies (Hartley et al., 2019). However, some aspects of a competitive retail market may place upward pressure on prices. Increased prices may result if some retailers possess market power (MacKay and Mercadal, 2020), a particular concern following mergers of large retailers (Brown et al., 2020b).

The details associated with each reform effort will affect its success, and each retail restructuring initiative has had different features with respect to the role of the incumbent utility, the recovery of stranded costs, qualification requirements placed upon retailers wishing to enter the market, the presence and design of a competitive wholesale market, acceptance of gentailer relationships, energy efficiency requirements placed on retailers, transition periods, price caps or price floors placed upon the incumbent utility provider, and social obligations placed upon retailers. Hence, it is unclear a priori whether the introduction of retail competition in a particular market will raise or lower prices, given that competitive market forces may lead to lower prices while other pressures may increase prices. As discussed in the following section, the results of prior studies have been mixed, thus presenting an empirical question that will be addressed here for Texas.

[3] The WSJ article mentioned earlier cites a quote from then-Gov. George W. Bush in 1999: "Competition in the electric industry will benefit Texans by reducing monthly rates."

To provide a contextual background for our paper, we review 19 studies of the impact of retail competition on retail electricity prices in the US. Some of these studies have sought to directly answer the question: has the introduction of retail competition led to higher or lower prices? These papers include Tabor et al., (2006), Swadley and Yücel (2015), Su (2015), McKearin (2015), Ros (2017), and Rose et al., (2021).

Other studies explore the related question of whether there was an increase or decrease in the markup of retail prices over wholesale prices when retail competition was introduced. A reduction in the markup of retail prices over wholesale costs might suggest a savings to consumers. Studies exploring the question of pass-throughs include Simeone and Hanger, 2016; Simeone et al., 2023; Hill, 2023). An examination of pass-through costs of competitive retailers may be of interest when a vertically integrated utility remains after competitors are permitted to enter a retail market. It might also prove helpful in isolating the impacts of retail competition from the impacts of wholesale market restructuring. Yet, revenue shifts resulting from charges authorized by regulatory authorities sometimes make it difficult to use the results from a pass-through analysis to address questions regarding the price impacts of retail competition (Dormady et al., 2019; Bowen et al., 2023; Simeone et al., 2023).

The last column in Table 1 paints a picture of mixed findings. Specifically, a couple of studies unequivocally report finding price reductions due to retail competition (Ros, 2017; Bowen et al., 2023). Fagan (2006) finds savings, while noting that his analysis is only preliminary. In contrast, seven or eight studies cannot find evidence of price reductions. Hill (2023) identifies small price reductions for some customer sectors and increases for others. The remaining six studies are also inconclusive.

In many cases, the authors have qualified their conclusions and reported that price reductions were found during some periods of time and that price increases were found in others. Some sectors appeared to experience price reductions, while other customer sectors saw price increases. In response to the mixed findings, we offer some additional observations.

First, mixed results are not surprising. As noted in Section 3, retail market restructuring has taken different paths in different states. Initial conditions tend to vary, with discernible differences in general energy price levels, the systems and rules through which utilities were formerly regulated, and wholesale market conditions.

Second, except for McKearin (2015), Ros (2017), Dormady et al., (2019), and Simone et al., (2023) that use data at the utility level, most studies have used data at the state level that mask the reality that not all electricity consumers in a particular state have retail choices. For example, retail choice in Michigan is capped at 10% of an electric utility’s retail sales.[4] As noted in Section 2, only ~57% of Texans reside in areas where retail competition has been introduced. As a state’s average price under retail competition is based on the prices paid by all retail customers, the impact of retail competition at the state level declines with the state’s share of customers with retail choice, possibly contributing to the mixed findings shown in the last column of Table 1.

Third, estimation sophistication varies across studies. For example, six studies in our sample mainly rely upon simple price comparisons and fail to control for how factors other than retail competition may impact prices, implying that their findings may be inaccurate. In contrast, eleven studies use various regression setups, ranging from ordinary least-squares to difference-in-difference, to account for changes in the factors affecting rates across time and space.

Fourth, the findings of the regression studies are conflicting. At the national level, three studies report no price reduction (Tabor et al., 2006; Borenstein and Bushnell, 2015; Rose et al., 2021), one study is inconclusive (Swadley and Yücel, 2011), and two studies report price reduction, though Fagan (2006) characterizes his results as preliminary and Su (2015) reports that price reductions are quite small. Different estimation periods and differences in the treatment of transition periods may account for some of the discrepancies. Focusing on Texas, McKearin (2015) finds price increases in the period of 1994-2012, thus suggesting a potential gain from using a longer sample period to understand the price impact of retail competition in Texas.

Table 1. Summary of 15 studies of retail competition’s impact on electricity prices in the US

Study	Geographic coverage	Data description	Customer class	Methodology	Used synthetic controls?	Has retail competition reduced prices?
Apt (2005)	The US	Annual state data for 1990-2003	Industrial	Comparison: before-competition rates vs. after-competition prices	No	No
Fagan (2006)	The US	Annual state data for 1993 to 2003	Industrial	Regressions using ordinary least-squares	No	Yes, based on “preliminary” results
Joskow (2006)	The US	Annual state data for 1996-2004	Residential and industrial	Comparison: restructured vs. unrestructured states	No	Inconclusive
Tabor et al., (2006)	The US	Annual state data for 1990-2003	Industrial	Regressions with heteroscedasticity and autocorrelation	No	No
Pfeifenberger et al., (2007)	The US	Annual state data from 1985 to 2006	Average of all customer classes	Graphical analysis	No	Inconclusive
Kwoka (2008)	The US	Not applicable	Not applicable	Review of ten prior studies	No	Inconclusive
Lien (2008)	The US	Annual state data for 1995-2006	Average of all customer classes	Comparison: restructured vs. unrestructured states	No	Inconclusive
Swadley and Yücel (2011)	16 states and Washington DC	Monthly state data for January 1990 to May 2010	Residential	Dynamic panel data analysis	No	Inconclusive
McKearin (2015)	Texas	Annual utility data for 1994 to 2012	Residential	Regressions based on difference-in-differences	No	No
Su (2015)	The US	Annual state data for 1990-2011	Residential, commercial, industrial	Regressions based on difference-in-differences	No	Yes, but only in the short run
Borenstein and Bushnell (2015)	The US	Annual state data for 1998-2017	Average of all customer classes	Retail price change regression	No	No
Simeone and Hanger (2016)	Pennsylvania	Annual state data for 2001-2015	Residential and small commercial	Comparison: before-competition rates vs. after-competition prices	No	Inconclusive
Ros (2017)	The US, aside from ERCOT	Annual data for 72 electricity distribution companies for 1972-2009	Residential, commercial, industrial	Panel data analysis	No	Yes
Zummo (2018)	The US	Annual state data for 1997-2017	Average of all customer classes	Comparison: restructured vs. unrestructured states	No	No
McKay and Mercadal (2020)	The US	Annual data for 1994-2016.	All	Difference-in-differences	No	No
Rose et al., (2021)	The US	Annual state data for 1990 to 2019	Residential, commercial, industrial	Synthetic control; Regressions based on difference-in-differences	Yes	No
Simeone et al., (2023)	Pennsylvania	Utility-specific data from 2011-2019	Residential	Examined the pass-through of wholesale costs by competitors to the utilities	No	Unlikely
Hill (2023)	Nine states, including Texas	Annual state data for 1990-2016.	Residential, commercial, industrial	Synthetic control	Yes	Yes, for residential and commercial sectors. No, for industrial sector. However, impacts are small and change over time.
Bowen et al., (2023)	Midwest US	Monthly state-level data for 1990-2019; and data from an energy aggregator	All sectors for some tests. Residential sector only for other analyses.	Three methods: difference-in-differences, repeated t-tests, and ratio analysis	No	Yes

Fifth, the studies differ with respect to how the beginning date of retail competition is defined and how transition periods are treated. During transition periods, price caps (or price floors, in the case of Texas) may be in place, and various stranded costs may be recovered through rates. Some studies use the enactment date of restructuring legislation that led to prompt retail price changes for recovery of stranded costs (Woo et al., 2003). Further, unfettered competition is often preceded by retail choice pilot programs, implying that the price data collected via these programs are of limited usefulness in the estimation of the full impact of retail competition upon prices. Even after retail choice is offered on a widespread basis, there may still be pricing restrictions to foster retail market entry and prevent market power abuse by affiliated retailers. While prices may be impacted during any of these transition periods, unfettered retail competition’s price impact is not completely realized.

Similarly, the period of estimation matters. Wholesale natural gas prices were particularly high in the US in late 2006, 2008, and early 2021. In Texas and other states with a high reliance upon natural gas as a marginal generation fuel, electricity prices in areas opened to retail competition were unusually high during those periods. Thus, the inclusion or exclusion of periods with high wholesale natural gas prices will impact any estimation of the impacts of retail competition.

Finally, only two studies employ SC (Rose et al., 2021; Hill, 2023), notwithstanding that “[p]robably because of their interpretability and transparent nature, synthetic controls have become widely applied in empirical research in economics and the social sciences” (Abadie, 2021, p.391). Along with those stated above, this remark motivates our methodology, which uses the city-level price data for Dallas and Houston to estimate the impact of retail competition upon prices, thereby removing the potential inaccuracy caused by using state-level data.

[4] https://www.michigan.gov/mpsc/-/media/Project/Websites/mpsc/regulatory/reports/elec-comp/2021_Electric_Competition_Status_Feb2022.pdf rev=81a305f56a744ecd88b08f73e2a59f82&hash=A9CF6A07B0520812643A6A99F9981005#:~:
text=In%20Michigan%2C%20Public%20Act%20286%20caps%20electric%20choice,average%20weather-adjusted%20retail%20sales%20for%20the%20preceding%20year.

In this section, we first provide a conceptual overview of our approach, followed by a discussion of the empirical steps that we followed for estimating the impact of introducing retail competition in Texas.

5.1 Taxonomy of price impact estimation

Table 2 shows the taxonomy of estimating the price impact, defined herein as the price difference between a control group that has not experienced retail competition and a treatment group that has. Specifically, row (1) in Table 2 applies to the pre-retail-competition period in which the regulated price or rate of the control group is R₁ and the regulated rate of the treatment group is P₁.[5] The rate difference is D₁ = P₁ – R_1. Variations in the rate difference across states or utilities are mainly caused by the geographic differences in rate drivers (e.g., features of the restructuring plan, customer mix, generation mix, fuel prices, transmission and distribution charges, weather, etc.).

The variations in price difference D₂ = P₂ – R₂ in row (2) of Table 1 are chiefly due to the geographic differences in rate drivers and implementation of retail competition. Hence, using D₂ as the basis for a side-by-side price comparison without accounting for geographic differences in the variables affecting rates may not accurately reveal the price impact solely attributable to retail competition.

Table 2’s last row shows that DP = P₂ – P₁ is the price change between the pre-retail-competition and competition periods. Without controlling for the changes in rate drivers over time, a before-and-after comparison based on DP may not accurately reveal the price impact solely attributable to retail competition.

Table 2. Taxonomy of estimating price impact of retail competition made possible by electricity market restructuring

Period definition	Control group	Treatment group	Price comparison	Remark
(1) Pre-retail-competition	R₁ = rate based on the cost-of-service regulation	P₁ = rate based on the cost-of-service regulation	Pre-retail-competition difference = D₁ = P₁ – R₁	D₁’s variations across states/utilities are caused by the geographic differences in rate drivers (e.g., customer mix, generation mix, transmission and distribution charges, weather, etc.).
(2) Retail competition	R₂ = rate based on the cost-of-service regulation	P₂ = competitive price based on plans offered by multiple retailers	Post-retail-competition difference = D₂ = P₂ – R₂	D₂’svariations across states/utilities are caused by the geographic differences in rate drivers *and* implementation of retail competition.
(3) Period change from (1) to (2)	Rate change = DR = R₂ – R₁	Price change = DP = P₂ – P₁	Net change = N = D₂ – D₁	DP’svariations across time are caused by changes in rate drivers over time *and* implementation of retail competition. Further, N is the difference-in-differences estimate for retail competition’s price impact.

Complicating the simple comparisons in Table 2’s rows (1) and (2) is the treatment of a transition period. Generally, retail competition is introduced in phases. While price deviations from the rate level in the pre-retail-competition period can occur in the transition period, they are not the result of an unfettered competitive retail market.

Suppose the price impact estimation is a panel data analysis of retail prices by state/utility and year (Wooldridge, 2010). Such an analysis aims to explain price variations using regressors as rate drivers and state/utility shares of customers eligible for retail choices in the competition period. If the coefficient estimate of the share variable is statistically significant and negative, it lends empirical support to the contention that retail competition leads to lower prices. Alternatively, the net price change N = D₂ – D₁ can be estimated using a difference-in-differences regression (Donald and Lang, 2007).

5.2 Approach

Our approach begins by defining the three periods in the evolution of retail competition in the ERCOT power region:

The 11-year pre-retail-competition period of 1990 - 2000. During this period, electricity prices in Dallas and Houston were set by the PUCT based on the cost-of-service regulation of Texas Utilities Electric Company that served Dallas and Houston Lighting and Power Company that served Houston.
The 6-year transition period of 2001 - 2006. Pilot customer choice programs began on 06/01/2001. Further, prices were adjusted by the PUCT in 2001 to enable the investor-owned utilities to recover their uncollected fuel costs. As a result, residential electricity prices jumped in 2001, although retail competition had not yet been fully implemented. The prices charged by the retailers affiliated with the traditional utilities were subject to the PUCT’s price-to-beat regulation until a sufficient level of customer switching was reached at the end of 2006.
The 14-year unfettered competition period of 2007-2020. Since 01/01/2007, there has been minimal regulatory oversight of the residential pricing plans offered by retailers in the areas opened to retail competition.

Each of the three SC listed below is a weighted combination of the candidate control group members listed in Table 3, against which actual prices in Houston and Dallas are compared. We use Stata’s Synth package to determine the SC’s optimal weights.

Our SC-based methodology comprises three steps. The first step uses annual data published by the EIA for a large set of integrated utilities that did not face retail competition within the 1990-2020 period to construct SC for a treatment city (i.e., Houston or Dallas).

Importantly, this step exploits the reality that all residential customers in Houston and Dallas had retail choices in 2007-2020, thus circumventing the empirical challenge posed by the EIA’s utility-level data or state-level data. This challenge arises for the following reasons. First, once a utility service area has started to have retail choices, average prices for that area are no longer available. The EIA might continue to track the prices charged by the incumbent utility serving the area if that utility continues to make retail sales. But the average price paid by consumers in a utility service area will be the average of the prices charged by all retailers serving that area, not just the incumbent utility. The market shares of retailers in an area might not be known, as sales data for competitive retailers are reported on a state-level basis by the EIA. Further, multiple pricing plans with various terms and conditions offered by each retailer complicate average retail price calculations.

Table 3. Electricity utilities included in our residential data sample for the period of 1990 – 2020

Panel A: Integrated utilities not exposed to retail competition for constructing the synthetic controls of a city in Texas

Integrated utility	State	Investor-owned?	BA with RTO in bold	Regulated rates ($/kWh) from the EIA
Integrated utility	State	Investor-owned?	BA with RTO in bold	1990-2000	2001-2006	2007-2020
Alabama Power Co	AL	Yes	SOCO	0.070	0.078	0.119
Arkansas Power & Light Co; aka Entergy Arkansas LLC	AR	Yes	MISO	0.089	0.081	0.098
Carolina Power & Light Co; aka Duke Energy Progress	NC	Yes	CPLE	0.081	0.084	0.106
Central Louisiana Elec Co Inc; aka Cleco Power LLC	LA	Yes	MISO	0.068	0.089	0.114
Chattanooga City of	TN	No	TVA	0.060	0.067	0.099
Colorado Springs City of	CO	No	WACM	0.061	0.076	0.111
Duke Power Co; aka Duke Energy Carolinas	NC	Yes	DUK	0.073	0.077	0.097
El Paso Electric Co	TX	Yes	EPE	0.103	0.114	0.119
Fayetteville Public Works Comm	NC	No	CPLE	0.076	0.080	0.104
Florida Power & Light Co	FL	Yes	FPL	0.079	0.093	0.109
Florida Power Corp; aka Progress Energy Florida Inc	FL	Yes	FPC	0.081	0.096	0.129
Gainesville Regional Utilities	FL	No	GVL	0.076	0.089	0.131
Georgia Power Co	GA	Yes	SOCO	0.076	0.078	0.115
Gulf Power Co	FL	Yes	SOCO	0.066	0.078	0.124
Huntsville City of	AL	No	TVA	0.055	0.063	0.093
Idaho Power Co	ID	Yes	IPCO	0.050	0.063	0.089
Indiana Michigan Power Co	IN	Yes	PJM	0.068	0.067	0.101
Indianapolis Power & Light Co	IN	Yes	MISO	0.059	0.064	0.093
Interstate Power Co	IA	Yes	MISO	0.074	0.096	0.138
Jacksonville Electric Auth; aka JEA	FL	No	JEA	0.070	0.074	0.115
Kansas City Power & Light Co	KS	Yes	SWPP	0.074	0.070	0.113
Kansas City Power & Light Co; aka Evergy Metro	MO	Yes	SWPP	0.082	0.074	0.116
Kansas Gas & Electric Co; aka KG&E a Western Resources Co	KS	Yes	SWPP	0.090	0.078	0.110
Kentucky Utilities Co	KY	Yes	LGEE	0.045	0.049	0.088
Knoxville, City of; aka Knoxville Utilities Board	TN	No	TVA	0.059	0.067	0.097
Lakeland City of	FL	No	FMPP	0.077	0.100	0.112
Louisville Gas & Electric Co	KY	Yes	LGEE	0.060	0.061	0.094
Madison Gas & Electric Co	WI	Yes	MISO	0.079	0.111	0.158
Memphis City of	TN	Yes	TVA	0.061	0.068	0.094
Minnesota Power & Light Co; aka ALLETE, Inc.	MN	Yes	MISO	0.061	0.069	0.095
Mississippi Power & Light Co; aka Entergy Mississippi LLC	MS	Yes	MISO	0.079	0.088	0.093
Monongahela Power Co	WV	Yes	PJM	0.071	0.073	0.099
Montana Power Co; aka NorthWestern Energy LLC - (MT)	MT	Yes	NWMT	0.061	0.079	0.110
Nashville, City of; aka Nashville Electric Service	TN	No	TVA	0.059	0.070	0.106
Nevada Power Co	NV	Yes	NEVP	0.062	0.096	0.123
Northern Indiana Pub Service Co	IN	Yes	MISO	0.100	0.099	0.126
Northern States Power Co; aka Northern States Power Co - Minnesota	MN	Yes	MISO	0.075	0.082	0.122
Oklahoma Gas & Electric Co	OK	Yes	SWPP	0.070	0.076	0.094
Orlando Utilities Comm	FL	No	FMPP	0.078	0.088	0.113
PacifiCorp	UT	Yes	PACE	0.069	0.070	0.101
PacifiCorp	WA	Yes	PACW	0.049	0.049	0.082
Public Service Co of Colorado	CO	Yes	PSCO	0.074	0.082	0.110
Public Service Co of IN Inc; aka PSI Energy Inc, Duke Energy Indiana Inc	IN	Yes	MISO	0.061	0.070	0.105
Public Service Co of NM	NM	Yes	PNM	0.094	0.083	0.119
Public Service Co of Oklahoma	OK	Yes	SWPP	0.060	0.070	0.089
Puget Sound Power & Light Co; aka Puget Sound Energy Inc	WA	Yes	PSEI	0.058	0.065	0.103
Seattle City of	WA	No	SCL	0.039	0.066	0.087
Sierra Pacific Power Co	NV	Yes	NEVP	0.086	0.112	0.117
South Carolina Electric & Gas Co; aka Dominion Energy South Carolina, Inc	SC	Yes	SCEG	0.075	0.089	0.132
Southwestern Electric Power Co	TX	Yes	SWPP	0.065	0.067	0.088
Southwestern Public Service Co	TX	Yes	SWPP	0.061	0.071	0.096
Tallahassee City of	FL	No	TAL	0.085	0.103	0.123
Tampa Electric Co	FL	Yes	TEC	0.081	0.097	0.114
Tucson Electric Power Co	AZ	Yes	TEPC	0.092	0.091	0.108
Union Electric Co	MO	Yes	MISO	0.075	0.070	0.096
UtiliCorp United Inc; aka Aquila Inc, KCP&L Greater Missouri Operations, Evergy Missouri West	MO	Yes	SWPP	0.078	0.070	0.105
Wisconsin Electric Power Co	WI	Yes	MISO	0.073	0.092	0.143
Wisconsin Power & Light Co	WI	Yes	MISO	0.068	0.093	0.126
Wisconsin Public Service Corp	WI	Yes	MISO	0.067	0.092	0.129

Notes: (1) Appendix 1 explains our selection of these integrated utilities.

(2) BA = Balancing authority of an electric grid; RTO = Regional transmission organization; MISO = Midcontinent Independent System Operator; PJM = PJM Interconnection; SWPP = Southwest Power Pool.

Panel B: Cities and pre-competition utilities used to obtain regulated rates and retail prices at the city level

Pre-retail-competition investor-owned integrated utility	City	Regulated rates ($/kWh) from the EIA: 1990-2000	Retail prices ($/kWh) from the BLS: 2001-2020
Texas Utilities Electric Company	Dallas	0.074	0.118
Houston Lighting and Power Company	Houston	0.083	0.126

We avoid the empirical challenge related to the EIA’s data by using the average price data collected by the US Bureau of Labor Statistics (BLS) for large cities for the purpose of constructing inflation indices. Limited by the city-level price data available from the BLS, we use Dallas and Houston to represent the areas of Texas that opened to retail competition. The population of Dallas in 2020 was 1.38 million (ranking ninth in city population in the US), and Houston’s was 2.31 million (ranking fourth in city population in the US). Because these cities have large populations, their retail electricity prices are deemed reasonable proxies of those in other competitive areas in the ERCOT retail market.

The second step graphically compares the counterfactual rates for the SCs found in the first step with actual residential prices in Houston and Dallas from the BLS in the transition period of 2001-2006 and the unfettered competition period of 2007-2020. This step informs price impact of retail competition based on the gap between actual prices in Houston and Dallas and counterfactual rates from the SC algorithm.

The third step uses the price and rate data from the last two steps to perform a panel data analysis (Wooldridge, 2010) based on a dummy variable regression specification: per kWh charge for residential consumption = a + b × binary indicator for the transition period of 2001-2006 + c × binary indicator for the unfettered competition period of 2007-2020 + error.

For empirical implementation, per kWh charges for Houston and Dallas are the actual rates in 1990-2000 and the actual prices in 2001-2020. The SC’s per kWh charges in the 1990-2022 period are the synthetic rates. Finally, coefficients b and c, respectively, are the estimated average price impacts of retail competition in the transition period and the period of unfettered competition.

[5] For the remainder of this paper, “competition” refers solely to competition in the retail market, as opposed to the wholesale market. Similarly, restructuring refers to the changes enabling retail customer choice.

6.1 Construction of synthetic controls

The key problem in estimating the impacts of this policy change is constructing a counterfactual to represent what prices would have been had retail competition not been introduced. Actual prices should be compared to this counterfactual. Using simple “before and after” comparisons of prices to measure the introduction of retail competition is problematic since energy prices often fluctuate over time for many unrelated reasons.

A simple comparison of the prices of non-profit utilities (and, perhaps, intra-state utilities) in Texas to prices in areas opened to competition is also problematic. Before competition was introduced in parts of Texas, municipal utilities and rural electric cooperatives tended to have low relatively rates. This is perhaps because they paid few government taxes and did not need to earn a profit for shareholders, although the prices charged by municipal systems were often set above cost-of-service to provide transfers to fund city services. Thus, including non-profit systems in a comparison group (as was done by the WSJ and the Texas Coalition for Affordable Power) is questionable. The control group’s lower prices could merely reflect historical differences in prices between groups.

Using prices from a single utility with price levels and patterns resembling the region where retail competition was introduced might be advantageous, if such a utility can be found. Averaging of a group of states with similar pre-restructuring price levels and patterns might be an improvement, but the results will be sensitive to which utilities are included in the control group.

Aiming to transparently produce a better counterfactual, the SC method creates a weighted average of the prices of similar utilities to form the counterfactual.[6] Weights are assigned to each of the candidate control group utilities, based on the similarity of their prices to the pre-restructuring price levels and patterns in the city (in this case) of interest. Additionally, covariates may be included in the SC algorithm to better assess the qualifications of the candidate utilities for membership in the control group and the weight that a control group utility’s prices should be assigned.

Corresponding to our methodology’s first step, Table 4 reports the optimal non-zero weights found via the algorithm of Abadie (2021) for constructing the SC for Houston and Dallas. These weights are based on the regulated rates of the integrated utilities listed in Table 3 and selected characteristics of the utilities in the pre-retail-competition period.

As the construction of the SC is sensitive to the selection of pre-treatment years and the decision of whether to include covariates (Kaul et al., 2021), Table 4 contains Case 1: no covariates and multiple years of pre-restructuring price data. To avoid using “too many” time-series observations to form the SC (Kaul et al., 2021), we limit, after extensive experimentation, the number of pre-retail-competition years to three (1990, 1995, and 1999) for Dallas and five (1990, 1992, 1994, 1995, and 2000) for Houston.

Table 4 also contains Case 2A: covariates and multiple pre-retail-competition years and Case 2B: covariates and single pre-retail-competition year of 2000. These two cases aim to show how the use of covariates and the selection of pre-retail-competition year may affect the comparison between actual prices in Dallas and Houston and counterfactual rates. Informed by Rose et al., (2021), the list of covariates used in Cases 2A and 2B includes (a) state-level data for per capita income, per capita employment, cooling degree days, heating degree days, and shares of coal-fired, natural-gas-fired, hydro, solar, and wind generation; (b) utility-level data for MWh sales and MWh shares by customer class; (c) whether the utility is an investor-owned utility; and (d) whether the utility belongs to a regional transmission organization or independent system operator.

Table 4. Optimal non-zero weights (sum = 1.0) found via the algorithm of Abadie (2021) for constructing the synthetic control of a city in Texas based on the pre-retail-competition period of 1990-2000

Panel A: Dallas

Candidate utility	Case 1: No covariates and multiple years of pre-restructuring price data	Case 2: Covariates
Candidate utility		A: Multiple years of pre-restructuring price data	B: Single year of pre-restructuring price data
El Paso Electric Co	0.075	0.252	0.285
Florida Power Corp; aka Progress Energy Florida Inc		0.010	0.081
Gulf Power Co		0.310
Nevada Power Co	0.595	0.410
Southwestern Public Service Co		0.018	0.612
Arkansas Power & Light Co; aka Entergy Arkansas LLC	0.330
Central Louisiana Elec Co Inc; aka Cleco Power LLC			0.004
Public Service Co of Oklahoma			0.018

Panel B: Houston

Candidate utility	Case 1: No covariates and multiple years of pre-restructuring price data	Case 2: Covariates
Candidate utility		A: Multiple years of pre-restructuring price data	B: Single year of pre-restructuring price data
Central Louisiana Elec Co Inc; aka Cleco Power LLC		0.091
El Paso Electric Co	0.042	0.228	0.215
Mississippi Power & Light Co; aka Entergy Mississippi LLC	0.255	0.120
Northern Indiana Pub Service Co	0.104	0.084	0.130
Sierra Pacific Power Co	0.210	0.122
Southwestern Electric Power Co		0.322	0.289
Southwestern Public Service Co		0.002	0.040
Tucson Electric Power Co		0.032
Gainesville Regional Utilities	0.389
Florida Power & Light Co			0.311
Northern States Power Co; aka Northern States Power Co - Minnesota			0.014

Note: The covariates are (a) state-level data for per capita income, per capita employment, cooling degree days, heating degree days, and shares of coal-fired, natural-gas-fired, hydro, solar, wind generation; (b) utility-level data for MWh sales and shares of MWh sales by customer class; (c) whether the utility is an investor-owned utility (IOU); and (d) whether the utility belongs to a regional transmission organization (RTO).

The placebo test results presented in Appendix 2 further support the contention that restructuring had a discernible effect on prices in Dallas and Houston.

6.2 Residential price comparison

Corresponding to our methodology’s second step, Figure 1 based on Case 1 shows that regulated prices in Dallas and Houston closely match the associated regulated rates of the SCs in the pre-retail-competition period of 1990-2000. As expected, electricity prices in Dallas and Houston jumped in 2001, reflecting the PUCT’s order that allowed the investor-owned utilities affected by Senate Bill 7 to recover the previously uncollected fuel costs prior to the start of full-scale customer choice on 01/01/2002. Moreover, actual retail prices exceed SC rates in 2001-2020, thus indicating retail competition’s price-increasing effect. Figures 2 and 3 based on Cases 2A and 2B affirm the key message conveyed by Figure 1: retail competition has generally raised retail prices in Dallas and Houston. Finally, Figures 1 to 3 seem to suggest that the price increases in Dallas are below those in Houston. However, our regression results for Model 2 in Section 6.3 indicate that the negative difference between the Dallas’ and Houston’s price increases is highly insignificant (p-value > 0.5).

We end this section by identifying the main cause of retail competition’s price-increasing effect portrayed in Figures 1 to 3. As suggested in Figure 4, the gaps between actual prices in Dallas and Houston and counterfactual rates increase with wholesale natural gas prices. This positive correlation is expected, as ERCOT’s wholesale electricity prices increase with wholesale natural gas prices (Zarnikau et al., 2019) and are passed through to residential pricing plans (Brown et al., 2020a). In contrast, the weighted average of regulated rates of the SC’s integrated utilities reflects the average fuel costs of electricity generation, which are below ERCOT’s marginal energy costs when natural gas prices are high. In summary, relatively high wholesale natural gas prices are a main contributing factor to the discernible gap between actual prices and counterfactual rates in Houston and Dallas, particularly around the 2008 spike in the US wholesale natural gas prices. When wholesale natural gas prices were low in 2016, retail electricity prices in Dallas and Houston were equal to or lower than the counterfactual.

6.3 Regression results

Corresponding to our third step, we use a panel data analysis to estimate the impact of retail competition on prices for two reasons. First, the price comparisons portrayed in Figures 1 to 3 are sensitive to how SC is constructed. However, we cannot a priori ascertain which synthetic control is more empirically reasonable for creating the counterfactual rates for Dallas and Houston, as all three SC are constructed to match the two cities in the pre-retail-competition period. Second, the panel data analysis estimates the average price impact from retail competition for the unfettered competition period that is useful for computing retail competition’s aggregate bill impact. A good case in point is the comparison of the WSJ’s reported bill impact of ~$1.75 billion per year and our calculated bill impact of ~1.09 billion per year.

Our panel data analysis has two models. Model 1 assumes constant price impacts and Model 2 assumes city-dependent price impacts. Importantly, the regression results from each model presented in Table 5 are insensitive to the assumption of fixed vs. random effects of city and SC construction.

Thanks to our use of SC, Model 1’s intercept estimates in Table 5 are highly insignificant (p-value > 0.9) and virtually equal to zero. The highly significant (p-value < 0.01) coefficient estimate for the binary indicator for the transition period is $0.0112/kWh ($11.2/MWh), which is smaller than the coefficient estimates of $0.0134/kWh ($13.4/MWh) for the binary indicator for the period of the unfettered competition period.

Table 5. Panel data analysis of per kWh charge for residential electricity consumption; sample period = 1990-2020; sample size = 31 years × 2 cities × 3 SC cases = 186 observations; statistically significant (p-value ≤ 0.05) estimates in bold based on robust errors clustered by city and SC case

Variable	Model 1: Constant price impacts						Model 2: City-dependent price impacts
	Fixed effects			Random effects			Fixed effects			Random effects
	Estimate	Standard error	p-value	Estimate	Standard error	p-value	Estimate	Standard error	p-value	Estimate	Standard error	p-value
R²: within	0.2752			0.0000			0.2765			0.2765
R²: between	0.0000			0.0000			0.0147			0.0147
R²: overall	0.2462			0.2462			0.2489			0.2489
Intercept	-0.0000	0.0020	0.982	-0.0000	0.0002	0.797	-0.0000	0.0020	0.982	-0.0000	0.0002	0.798
Binary indicator for the transition period = 1 if 2001-2006, 0 otherwise	0.0112	0.0021	0.003	0.0112	0.0021	0.000	0.0124	0.0030	0.009	0.0124	0.0030	0.000
Binary indicator for the period of unfettered competition = 1 if 2007-2020, 0 otherwise	0.0134	0.0035	0.013	0.0134	0.0035	0.000	0.0139	0.0050	0.039	0.0139	0.0050	0.005
Binary indicator for the transitional period × Binary indicator for Dallas							-0.0024	0.0040	0.575	-0.0024	0.0039	0.537
Binary indicator for the period of unfettered competition × Binary indicator for Dallas							-0.0011	0.0071	0.887	-0.0011	0.0070	0.878

Table 6. Panel data analysis of electricity price gap ($/kWh) vs. Henry Hub’s wholesale natural gas price ($/MMBtu); sample period = 2001-2020; sample size = 20 years × 2 cities × 3 SC cases = 120 observations; statistically significant (p-value ≤ 0.05) estimates in bold based robust errors clustered by city and SC case

Variable	Fixed effects			Random effects
Variable	Estimate	Standard error	p-value	Estimate	Standard error	p-value
R²: within	0.2998			0.0000
R²: between	0.0000			0.0000
R²: overall	0.2295			0.2295
Intercept	-0.0004	0.0021	0.857	-0.0004	0.0040	0.919
Binary indicator for the transition period = 1 if 2001-2006, 0 otherwise	-0.0083	0.0015	0.003	-0.0083	00015	0.000
Henry Hub’s wholesale natural gas price	0.0035	0.0005	0.001	0.0035	0.0005	0.000

Note: The coefficient estimates in the last row measure the estimated effects of a $1/MMBtu increase in wholesale natural gas price on the electricity price gap in 2001-2020.

Intercept estimates for Model 2 are highly insignificant (p-value ≥ 0.8) and virtually equal to zero. The highly significant (p-value < 0.01) coefficient estimate for the binary indicator for the transition period is $0.0124/kWh ($12.4/MWh), which is smaller than the coefficient estimate of $0.0139/kWh ($13.9/MWh) for the binary indicator for the period of unfettered competition. Finally, the interaction terms in the last two rows have highly insignificant (p-value > 0.5) coefficient estimates, indicating that the estimated impacts of retail competition upon prices are not city-dependent.

We end this section by performing a panel data analysis of the relationship between the electricity price gap and the wholesale natural gas price set at Henry Hub, a major trading hub for natural gas in the neighboring state of Louisiana. The coefficient estimates in the last row of Table 6 are 0.0035 (p-value < 0.01), thus reinforcing the key takeaway from Figure 4: the gaps between actual prices in Houston or Dallas and counterfactual rates increase with wholesale natural gas prices.

[6] We note one limitation of this approach. Hill (2023) reports difficulty in using this method to assess market reforms in states with high electricity prices (e.g., the Northeast US). We have had similar difficulty in our attempts to examine the impacts of retail competition on prices in Boston, New York, and other cities in that region. The problem is finding control group candidates with high prices that did not undergo retail market restructuring. There may be too few candidates to make this approach work.

Our price impact estimates suggest that retail competition has tended to increase residential electricity prices in Texas, at least during the years of high wholesale natural gas prices after the commencement of retail competition in June 2001.

This conclusion deserves elaboration. As noted earlier, restructuring in the wholesale segment of the industry has increased the sensitivity of retail electricity prices to natural gas prices. The share of generation from natural gas plants in Texas has been particularly high. Under the cost-of- service regulation, retail electric rates reflect the weighted average cost of all fuels used by an integrated utility to generate electricity. In contrast, retail prices in Texas increase with ERCOT’s wholesale electricity prices that track the marginal fuel cost of natural gas (Brown et al., 2020a). Thus, high wholesale natural gas prices in some years of the 2001-2020 period largely explain the gaps between actual prices in Houston and Dallas and counterfactual rates. Had wholesale natural gas prices been at the pre-retail-competition period’s level of ~$3/MMBtu, residential electricity prices under retail competition could have been far lower.

The impact of retail competition on residential electricity prices in Texas may not reflect those experienced elsewhere. Texas was an unusual case of a state that traditionally enjoyed low-to-moderate electric rates pursuing extensive restructuring. States in the northeast and west coast of the US with traditionally high electricity costs may have had a greater potential to achieve savings through market reforms. Moreover, market reforms vary greatly among different states. This may, in part, explain why some antecedent studies that exclude the ERCOT region (e.g., Ros, 2017) find no price increase after restructuring.

Our conclusions have the following policy implications. First, retail competition via market restructuring in a region that traditionally enjoys low electricity rates and is highly dependent upon natural gas for electricity generation may see higher electricity prices – particularly during periods of high wholesale natural gas prices. Second, Texas should exercise caution when considering the merger of large REPs that may reduce the retail electricity market’s extent of price competition. Finally, any actions that might reduce the cost of natural gas to power plants would reduce electricity costs to consumers, particularly in areas of a state opened to retail competition. Indeed, efforts are underway to improve the coordination between the natural gas and electric industries in Texas, particularly during winter storms, and this coordination might reduce future natural gas price spikes.

We would be remiss had we failed to acknowledge that our paper only considers one aspect of the electricity market restructuring: the impact of retail competition on prices. Hence, it ignores consumer benefits of retail competition, such as additional value-added services (Rai and Zarnikau, 2016) and choices of providers and services (Lancaster, 1990; Goett et al., 2000). Addressing these non-price benefits, however, is well outside the scope of our paper.

Declarations, Compliance with Ethical Standards, and Data Availability

The authors have no competing interests to declare that are relevant to the content of this article.

The authors did not receive support from any organization for the submitted work.

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request. Moreover, all data were acquired from U.S. government agencies are and publicly-available on websites.

Abadie, A., Diamond, A. and Hainmueller, J. (2010). Synthetic control methods for comparative case studies: Estimating the effect of California’s tobacco control program. Journal of the American Statistical Association, 105(490), pp.493-505.
Abadie, A., Diamond, A. and Hainmueller, J. (2015). Comparative politics and the synthetic control method. American Journal of Political Science, 59(2), pp.495-510.
Abadie, A., and Vives-i-Bastida, J. (2021). Synthetic Controls in Action. Working paper.
Abadie, A. (2021). Using synthetic controls: feasibility, data requirements, and methodological aspects. Journal of Economic Literature, 59(2), 391-425.
Apt, J. (2005). Competition has not lowered US industrial electricity prices. The Electricity Journal, 18(2), 52-61.
Biggar, D., Hesamzaden, M.R. (2014). The economics of electricity markets. Wiley, IEEE Press.
Blum, J. (2016). Texas consumers pay more in deregulated electricity markets, study says. Houston Chronicle. June 8, 2016. Accessed at: https://www.houstonchronicle.com/business/energy/article/Texas-consumers-pay-more-in-deregulated-7972017.php
Borenstein, S., Bushnell, J. (2015). The U.S. electricity industry after 20 years of restructuring. Annual Review of Economics, 7, 437-463.
Bowen, W.M., Hill, E., Thomas, A., Henning, M. (2023). Consumer price effects of deregulated electric generation markets: The case of Ohio and the Midwestern United States. SSRN preprint available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4269903#:~:text=Bowen%2C%20William%20and%20Hill%2C%20Edward%20and%20Thomas%2C%20Andrew,at%20SSRN%3A%20https%3A%2F%2Fssrn.com%2Fabstract%3D4269903%20or%20http%3A%2F%2Fdx.doi.org%2F10.2139%2Fssrn.4269903%20Download%20This%20Paper
Brown, D.P., Tsai, C.H., Woo, C.K., Zarnikau, J., Zhu, S. (2020a). Residential electricity pricing in Texas’s competitive retail market. Energy Economics, 92, 104953.
Brown, D.P., Zarnikau, J., Adib, P., Tsai, C.H., Woo, C.K. (2020b). Rising market concentration in Texas’s retail electricity market. The Electricity Journal, 33, 106848.
Donald, S., Lang, K. (2007). Inferences with differences-in-differences and other panel data. Review of Economics and Statistics, 89, 221-233.
Dormady, N., Hoyt, M. Roa-Henriquez, A., Welch, W. (2019). Who Pays for Retail Electric
Deregulation?: Evidence of Cross-Subsidization from Complete Bill Data. Energy Journal, 40(2): 161-194.
Fagan, M.L. (2006). Measuring and explaining electricity price changes in restructured states. The Electricity Journal, 19(5), 35-42.
Goett, A.A., Hudson, K., Train, K.E. (2000). Customers’ choice among retail energy suppliers: The willingness-to-pay for service attributes. Energy Journal, 21(4), 1–28.
Hartley, P., Medlock, K., Jankovska, O. (2019). Electricity reform and retail pricing in Texas. Energy Economics, 80, 1-11.
Hill, A. (2023). Price freezes and gas pass-through: An estimation of the price impact of electricity market restructuring. Journal of Regulatory Economics, 63, 87–116.
Joskow, P.L. (2006). Markets for power in the United States: An interim assessment. The Energy Journal, 27(1), 1-36.
Kaul, A., Klößner, S., Pfeifer, G., Schieler, M. (2021). Standard synthetic control methods: The case of using all preintervention outcomes together with covariates. Journal of Business and Economic Statistics, 40(3), 1362-1376.
Kwoka, J. (2008). Restructuring the U.S. electric power sector: A review of recent studies. Review of Industrial Organization, 32, 165-196.
Lancaster, K. (1990). The economics of product variety: A survey. Marketing Science, 9(3), 189-278.
Lien, J. (2008). Electricity restructuring: What has worked and what has not, and what is next. US DOJ Working Paper. Available at: http://www.justice.gov/atr/electricity-restructuring-what-has-worked-what-has-not-and-what-next.
McKearin, T. (2015). Essays on the effects of government intervention in Texas’ electricity market and the health insurance markets in Missouri and Oklahoma. Ph.D. dissertation, University of Texas at Austin. Available at: https://repositories.lib.utexas.edu/bitstream/handle/2152/32827/MCKEARIN-DISSERTATION-2015.pdf;sequence=1
McKay, A., Mercadal, I. (2020). Shades of integration: The restructuring of the U.S. Electricity Markets. Available at: https://www.semanticscholar.org/paper/Shades-of-Integration%3A-The-Restructuring-of-the-Mackay-Mercadal/4e86a7904377f62d9bbf15f288c01258140f8b76
Pfeifenberger, J.P., Basheda, G.N., Schumacher A.C. (2007). What we can learn from retail-rate increases in restructured and non-restructured states. Public Utilities Fortnightly. Available at: http://www.fortnightly.com/fortnightly/2007/06/restructuring-revisited?page=0%2C1
Rai, V., Zarnikau, J. (2016). Retail competition, advanced metering investments, and produce differentiation: Evidence from Texas. In Future of Utilities, Utilities of the Future. Ed. Sioshansi, F. Academic Press. 153-173.
Ros, A. (2017). An econometric assessment of electricity demand in the United States using utility-specific panel data and the impact of retail competition on prices. The Energy Journal, 38(4), 73-99.
Rose, K., Tarufelli, B., Upton, G. (2021). Electricity market restructuring and retail rates, USAEE Working Paper No. 20-462. Available at: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3660974
Simeone, C., Hanger J. (2016). A case study of electric competition results in Pennsylvania: Real benefits and important choices ahead. Kleinman Center for Energy Policy, University of Pennsylvania.
Simeone, C., Lange, I., Gilbert, B. (2023). Pass-through in residential electricity competition: Evidence from Pennsylvania, Utilities Policy, 80(9): 101479.
Sioshansi, F.P. (2006). Electricity market reform: What have we learned? What have we gained? The Electricity Journal, 19(9): 70-83.
Sioshansi, F.P. (2013). Evolution of global electricity markets: New paradigms, New challenges, New approaches. Elsevier.
Stoft, S. (2002). Power system economics: Designing markets for electricity. Wiley-IEEE Press.
Su, X. (2015). Have customers benefited from electricity retail competition? Journal of Regulatory Economics, 47(2), 146-182.
Swadley, A., Yücel, M. (2011). Did residential electricity rates fall after retail competition? A dynamic panel analysis. Energy Policy, 39(12), 7702-7711.
Tabor, J., Chapman, D., Mount, T. (2006). Examining the effects of deregulation on retail electricity prices. Working Paper, Dept. of Applied Economics and Management, Cornell University. Available at: http://publications.dyson.cornell.edu/research/researchpdf/wp/2005/Cornell_Dyson_wp0514.pdf
Triolo, R.C., Wolak, F.A. (2022). Quantifying the benefits of a nodal market design in the Texas electricity market. Energy Economics, 112, 106154.
UT Austin (2021). The timeline and events of the February 2021 Texas electric grid blackouts. Available at: https://energy.utexas.edu/sites/default/files/UTAustin%20%282021%29%20EventsFebruary2021TexasBlackout%2020210714.pdf
Woo, C.K., Lloyd, D., Karimov, R., Tishler, A. (2003). Stranded cost recovery in electricity market reforms in the US. Energy, 28(1), 1-14.
Woo, C.K., Zarnikau, J. (2009). Will electricity market reform likely reduce retail rates? The Electricity Journal, 22(2), 40-45.
Wooldridge, J.M. (2010). Econometric analysis of cross section and panel data. MIT Press.
Zarnikau, J. (2005). A review of efforts to restructure Texas’ electricity market. Energy Policy, 33, 15-25.
Zarnikau, J., Whitworth, D. (2006). Has electric utility restructuring led to lower electricity prices for residential consumers in Texas? Energy Policy, 34(1), 2191-2200.
Zarnikau, J., Woo, C.K., Baldick, R. (2014). Did the introduction of a nodal market structure impact wholesale electricity prices in the Texas (ERCOT) market? Journal of Regulatory Economics, 45, 194-208.
Zarnikau, J., Woo, C.K., Zhu, S., Tsai, C.H. (2019). Market price behavior of wholesale electricity products: Texas. Energy Policy, 125, 418-428.
Zummo, P. (2018). 2017 retail electric rates in deregulated and regulated states. American Public Power Association.

No competing interests reported.

Appendix.docx

Has retail competition reduced residential electricity prices in Texas?

Status:

Version 1

Abstract

Figures

1. Introduction

2. Electricity industry restructuring in Texas

3. The change in industry structure affected the determinants of retail prices

4. Review of antecedent US studies

5. Methods

6. Empirics

7. Conclusions and policy implications

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1