PTDoes the regulator’s score change energy utilities’ debt?
In the U.S., a commission in each state decides how much an electric utility may charge, and an agency scores the climate of these commissions, from 1 (lower risk) to 9 (higher risk). The thesis followed 137 utilities from 2012 to 2023 to see whether the score matters when they borrow. In 12 screens.
Illustration with the thesis coefficient (Table 16): the lit score moves from 1 to 9 and predicted debt, starting from the 46.9% average share of capital, changes by less than 0.1 percentage point per score.
A state commission decides how much the utility may charge
U.S. investor-owned electric utilities are regulated monopolies. To change rates, a case is opened at the state public utility commission (the rate case). Follow the three steps.
Tables 1, 2 and 4 of the thesis (EIA data, 2022). The stack uses the revenue of the major investor-owned utilities (US$381.1 billion); the split per US$100 is our calculation and adds up to 99.9 because of rounding.
Earlier studies point both ways
Classic theory says that firms facing more risk should borrow less. Under regulation, though, debt can also become a tool. Switch the arguments on and off and see where the scale tips.
illustration The scale only counts arguments; it does not weigh them. The arguments summarize sections 2.3 and 3 of the thesis. The thesis bets on the left side: more risk, less debt (hypothesis 1).
An agency gives each state a regulatory risk score from 1 to 9
Regulatory Research Associates (RRA, part of S&P Global) weighs more than 30 metrics, such as the returns the commission usually authorizes and the outcomes of rate cases. The score is written for investors: 1 is the most favorable climate; 9, the riskiest.
(lower risk)averagebelow average
(higher risk)
Tap a state. Appendix B: year-end score of the state where the utility mainly operates. Mean 4.72, standard deviation 1.67 (Table 14). Nebraska and the District of Columbia have no utility in the sample (hollow tiles).
In 17 states and the capital, power sales gained competition
Since the 1990s, some states have separated those who generate power from those who deliver it. The thesis splits the states into three groups.
Height = number of sample utilities in the state (Appendix A); tap a state to see which ones. Groups from Tables 7 and 8. Average score weighted by the number of utilities, 2012 to 2023: our calculation from Appendix B.
The thesis followed 137 utilities for 12 years
The search started from every regulated energy company in the S&P Capital IQ Pro database. What remained were the investor-owned electric utilities that actually operate, not their parent holdings, and that had data. Tap the steps.
Table 9, section 4.2 and Appendix A. The dots show counts: the thesis does not say which six utilities are listed. The categories in Table 9 add up to 3,478, two more than the printed total.
On average, debt is 47% of these utilities’ capital
The thesis measures leverage as debt’s share of capital: debt ÷ (debt + equity), at book values. Build a utility and compare it with the sample.
Table 14, Appendix C and section 5: mean 46.9%, standard deviation 7.8 points, minimum 3.0% and maximum 81.2%. In the 2012–2023 rate cases, authorized equity was 50.95% of capital, or about 49% debt (figure cited by the thesis).
The thesis tested three hypotheses on regulatory risk
The models compare each utility with itself over the years (fixed effects) and account for size, profit, fixed assets, investment and the industry’s typical debt. Before you look, place your bet: did the data confirm each hypothesis?
How much debt changes per extra point of the score
Tables 16, 19, 20 and 22 (coefficients × 100, our calculation). A band that crosses zero means the data cannot tell the effect apart from no effect. For comparison: 1 extra percentage point of profitability comes with 0.45 point less debt (screen 9).
What moves debt is the utility’s own profit and investment
In the same model, two of the utility’s own factors passed the test. With more profit, debt falls: the firm uses its own money first, as the pecking order theory predicts. With more investment, debt rises. Move the controls.
our illustration With the coefficients of the final model (Table 16 and Appendix D), starting from the sample mean (46.9%): it is the effect within the same utility over time. Size, fixed assets and the industry’s typical debt stay still (no clear effect at 5%; size was only a hint, p = 0.081). Striped bar: pulls debt down; gray: no clear effect, with the 95% band around it.
Debt returns to its target slowly, about 16% a year
When a utility drifts away from the debt level it considers ideal, it does not fix it all at once. In the thesis’ dynamic model, 84% of one year’s debt carries over to the next, and the gap to the target shrinks by 16.1% each year. The year-to-year change in the score did not explain this adjustment (p = 0.72).
our illustration A utility that starts 10 points above its target, using the coefficient on last year’s debt (Table 22 and Appendix F). The three options are the Appendix F models, which differ only in the instruments (the lags used to handle the back-and-forth between past and present debt). The author picked the model with the fewest instruments (42, below the 132 utilities); in it, the Hansen test (p = 0.007) flags a possible problem with the instruments, which the author acknowledges.
Looking at all of them together, more risk goes with slightly more debt
Why does the link vanish? First, a guess: in how many year-to-year turns did the score of the utility’s state change?
One column per state with a sample utility, one row per year (Appendix B). Counts of changes, utility weights and standard deviations: our calculations from Appendices A and B.
What changes for those who regulate, manage or invest
Regulators
The perception of regulatory risk did not move utilities’ leverage in the short run. For the author, this may come from institutional rigidity or from the shelter that group structures give to subsidiaries.
Utility managers
Capital structure followed internal fundamentals: profit (more profit, less debt) and investment (more investment, more debt), more than the state’s climate.
Investors
To assess a utility’s debt, the state’s score alone says little; the firm’s performance and investment plan say more.
Researchers
Suggested next steps: read rate cases with text analysis, measure political factors (commissioner turnover, public sentiment, lobbying) and follow money flows between subsidiaries of the same group.
What this research does not say
- One score does not tell everything. The RRA score may not capture every nuance of commission behavior, and the study covers one sector and one period. · according to the author
- The score barely changes. It changed in 18% of year-to-year turns, almost always by 1 point. When each utility is compared with itself, little movement is left to measure. · our observation
- No clear effect is not zero effect. Going from score 1 to 9, the 95% interval runs from 4.8 points less to 3.8 more debt (our calculation). · our observation
- The pace of adjustment depends on the model. With other instruments it falls from 16% to about 6% a year, and the chosen model fails the Hansen test. · our observation
Credits
SANTOS, Guilherme Spiazzi dos. Capital structure under regulatory risk in the U.S. electric industry. Doctoral thesis (Business Administration), Universidade Presbiteriana Mackenzie, São Paulo, 2025. Advisor: Prof. Dr. Wilson Toshiro Nakamura. Co-advisor: Prof. Dr. Manuel José da Rocha Armada.
Research supported by CAPES (Brazil’s federal agency for graduate education), with a research stay at the University of Minho (Portugal). Data from S&P Capital IQ Pro and the EIA; analysis in Stata. This version simplifies the work for a general audience; the formulas and full tables are in the original text.