“Did deregulation save Texans money?” has been argued in Texas for twenty-four years, and the argument has a structural defect. It is not one empirical question. It can mean an advertised offer, a bill actually paid, a residential revenue-per-kWh average, wholesale production cost, retail operating cost, a welfare counterfactual, or an avoided-investment claim. Four serious pieces of work have answered four of those, and each answer gets cited as though it settled all of them.
Here is what each one measured, with its own limits stated. Anyone who walks away from this able to name the outcome, population, period, and comparator behind a figure they’re quoting will be ahead of most of the people they’re arguing with.
TCAP: an area-average price comparison
The Texas Coalition for Affordable Power published a snapshot report in June 2016 comparing average residential prices inside and outside areas with retail deregulation, using EIA data for 2002 through 2014. The headline: “From 2002 through 2014 the imputed ‘lost savings’ resulting from higher average prices in deregulated areas exceeds $24 billion.” The report described the equivalent as more than $5,100 for a typical household and used an illustrative monthly usage assumption of 1,300 kWh.
Label it correctly: TCAP is a nonprofit public-entity electricity aggregation organization, and this is an advocacy-adjacent pricing comparison rather than a neutral study. That label is not a dismissal. The report is transparent about its arithmetic and its EIA source, and it acknowledges two qualifications frequently dropped in retelling — it found low individual competitive offers available in deregulated areas, and it noted that the average price gap narrowed after 2011.
The limitation follows from the comparison design. The exempt group includes municipal utilities, cooperatives, and investor-owned utilities outside ERCOT. Those providers differ from choice-area providers in ownership, cost structure, debt, governance, service territory, customer mix, fuel procurement, rate design, taxes, and local resource decisions. An area-average comparison can establish a descriptive price gap. It cannot alone establish that retail choice caused the whole gap, because the groups differ in ways other than choice.
And “lost savings” is an imputed counterfactual, not an observed loss. It is what the deregulated-area average would have produced had it matched the exempt-area average. That is a legitimate calculation with an explicit assumption inside it.
The PUCT: an available-offer comparison
The Commission’s 2019 Scope of Competition report presented a different measure and reached a different conclusion. It stated that rates in the competitive market had decreased by 31% since transition when compared against the last inflation-adjusted regulated retail rate. It reported an average lowest available residential price of 9.36 cents per kWh in September 2018, an average across all plans of 10.3 cents per kWh that month, and a 2018 national average of 13.02 cents per kWh.
The case for this comparison is that the last regulated rate is the actual regulatory starting point for the customers who became eligible to shop, and that an offer dataset shows what was selectable at a defined date.
The limitation is availability versus adoption. An offer snapshot cannot tell you what every customer paid, whether the lowest offer suited a given household’s usage, or whether households searched and switched. A market where the best available price fell 31% and most customers never moved would produce this statistic. So would a market where everyone captured it. The measure cannot distinguish them.
Hartley, Medlock, and Jankovska: price-cost transmission
The 2019 Energy Economics article by Peter Hartley, Kenneth Medlock III, and Olivera Jankovska asked a different question entirely — not whether prices fell, but whether post-reform retail prices better reflected wholesale prices and whether reform reduced retailer costs. Their conclusion: “We find clear evidence of both outcomes in competitive market areas but not in non-competitive areas.”
The authors were affiliated with Rice University’s Baker Institute, which should be disclosed when the work is cited in an argument, and which is not a defect in the finding.
This matters to the defense of restructuring because cost reflection and lower retailer cost are channels through which a competitive market can create efficiency even where a simple average-bill comparison is ambiguous.
The limitation is that cost reflection is not the same as lower bills. Better pass-through can expose customers to higher fuel, scarcity, congestion, delivery, or policy costs in an expensive period. Lower retailer costs can be offset elsewhere. And a study of the price-wholesale relationship does not value search burden, plan complexity, credit segmentation, or involuntary inertia. The strongest fair statement is not “the paper proves Texans saved.” It is that the paper finds competitive-area evidence of improved price-cost transmission and lower retailer cost under its model.
Zarnikau, Cao, Qi, and Woo: a synthetic control
The 2023 Utilities Policy article applied a synthetic control design to residential consumers in Dallas and Houston: “Retail competition raised average prices by $0.0112/kWh ($11.2/MWh) in the transition period from 2001 to 2006.” The authors reported a further increase of $0.0134/kWh for 2007 through 2020, and their abstract notes the result was small when U.S. wholesale natural gas prices were relatively low.
This design improves on a simple in-state utility-type comparison by constructing a stated external counterfactual rather than treating every Texas exempt provider as a natural control. It is the methodologically strongest entry in the set for the question it asks.
It is also a study of two cities, residential average prices, a particular donor pool and model construction, and a historical period. It does not establish identical effects for every Texas service territory, for commercial accounts, or for every retail product, and it does not measure what customers may value in product choice, contract type, renewable content, or bill stability.
What the transition years contaminate
There is a fifth data point that is regularly misused as a sixth study.
The PUCT’s January 2003 report described the market’s opening: customers who did not select a new REP were transferred to the affiliated REP in January 2002, and residential and small nonresidential customers with peak demand of 1 MW or less who remained were charged a regulated price to beat. The report said customers received more than a 6% reduction in total rates compared with December 31, 2001 rates after fuel-surcharge termination and lower fuel factors, with an affiliate-specific range from 3.6% to 18.1%.
That report also describes a fuel-factor dispute in which REPs argued the initial factors were too low to permit effective competition while consumer representatives argued they were too high, and notes that all but one affiliated REP requested fuel-factor adjustments after gas prices rose in early 2002.
Those are regulated transition comparisons under a price to beat. They are not evidence of a permanent competitive dividend, and folding them into a long-run treatment period contaminates the estimate.
The strongest objection
The objection to this whole approach is that it is a sophisticated way of refusing to answer. Policymakers and customers need a verdict. “Four studies measured four things” is an academic posture that lets an analyst avoid taking a position while sounding rigorous, and in practice it defends the status quo, because the burden of proof always falls on whoever wants to change something.
That criticism has force, and it is worth answering rather than deflecting.
It is right that methodological pluralism can become an evasion, and right that it tends to favor incumbency. It is also right that people deciding policy cannot wait for a perfect panel.
But the response is not to pick a winner the evidence does not support. It is to say precisely what the evidence does support, which is narrower than either camp’s slogan and is not nothing:
Competitive Texas customers have demonstrable opportunities to choose among offers, and there is evidence from competitive areas of improved transmission between wholesale cost and retail price and of lower retailer cost. At the same time, published descriptive and synthetic-control work provides serious evidence that average residential price outcomes may have been worse under the counterfactuals those studies chose.
Both halves of that sentence are supported. Neither cancels the other, because they are answers to different questions. That is a position, not a refusal to take one — and it is a position that will survive the next study, which the slogans will not.
What would actually settle more of it
A reproducible, parent-adjusted, territory-matched panel of customer bills, usage, plan and EFL versions, switching history, credit and eligibility conditions, delivery charges, taxes, and wholesale hedge and cost proxies, with a pre-specified counterfactual and separate results for active shoppers and inert customers. It would report residential, small commercial, and large commercial and industrial populations separately rather than transferring a residential result to businesses.
One figure that circulates in this debate should be dropped until it can be sourced. A widely repeated 2021 Wall Street Journal dollar estimate of deregulation’s cost to Texas consumers could not be authenticated with its original dated methodology and data extract in the research behind this piece. Do not repeat the number without the original analysis and a reproducible calculation.
Amerigy Energy is a registered electricity broker in Lufkin, PUCT registration #BR190721. We’re a participant in this market, not a disinterested party, which is exactly why we’d rather cite the studies than summarize them.
1. Texas Coalition for Affordable Power, “Electricity Prices in Texas: Snapshot Report June 2016,” June 2016. https://static.texastribune.org/media/documents/TCAP-ElectricityPricesinTX-Snapshot-A-1.1.pdf
2. Public Utility Commission of Texas, “2019 Scope of Competition in Electric Markets of Texas Report,” 2019. https://ftp.puc.texas.gov/public/puct-info/industry/electric/reports/scope/2019/2019scope_elec.pdf
3. Peter R. Hartley, Kenneth B. Medlock III, and Olivera Jankovska, “Electricity Reform and Retail Pricing in Texas,” Energy Economics, 2019. https://doi.org/10.1016/j.eneco.2018.12.024
4. Jay Zarnikau, Kang Hua Cao, H.S. Qi, and C.K. Woo, “Has Retail Competition Reduced Residential Electricity Prices in Texas?,” Utilities Policy, 2023; SSRN preprint page published July 6, 2023. https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4502238
5. Public Utility Commission of Texas, “2003 Scope of Competition in Electric Markets in Texas,” January 2003. https://ftp.puc.texas.gov/public/puct-info/industry/electric/reports/scope/2003/2003scope_elec.pdf
6. U.S. Energy Information Administration, “Form EIA-861 Annual Electric Power Industry Report,” accessed September 12, 2026. https://www.eia.gov/electricity/data/eia861/