Examination of Utility Betas - An Update

Utility betas increased sharply as part of the Covid-19 related market volatility experienced in the early 2020s and have only now, five years later, dropped back to much lower levels, albeit not back down to the level of earlier beta estimates. It has been an incredible turn of events, one not seen since the financial crisis of 2008¹. This matters because the beta estimate² is a core part of the Capital Asset Pricing Model (“CAPM”), an analytical model relied upon by many regulatory commissions to inform the allowed return on equity (“ROE”), which utilities earn on their investment in utility plant, i.e., rate base. This change will likely put significant downward pressure on the ROEs authorized by regulatory commissions.

For the purposes of this article, we use a computational approach accepted by FERC, where the beta for an individual stock is measured using the S&P 500 as the market return index over a five-year period based on weekly realized return data and is Blume-adjusted. We also use the S&P 500 Utility sector as a proxy for individual utility stocks.

EVOLUTION OF UTILITY BETAS

As described in a previous article, utility betas generally trended downwards towards 0.55 for several years leading up to the Covid-19 market volatility and then shot up to around 0.85-0.90 in a matter of a few short weeks in early 2020³. The utility betas have only now declined significantly to 0.70. Similar to the sharp increase, this drop happened quickly, in about a two-month period. The catalyst for this change is that the rolling five-year beta study period no longer captures the extreme Covid-19 related financial data. This is consistent with the trend seen with shorter-term betas and how those betas similarly declined following the removal of the early 2020 financial market data from the computation, e.g., betas computed using two years of financial data⁴. Figure 2 below depicts rolling five-year betas for the S&P 500 Utility sector using market data since 2000.

Figure 2: S&P 500 Utility Sector Five-Yr. Betas

BETA ESTIMATES - BEHIND THE SCENES

It’s beneficial to a take a look at the different calculation components behind the beta estimate to better understand how the beta has evolved. For reference, the beta can be determined using the following mathematical expression⁵:

Where β = beta, Ri = the return on the individual stock, Rm = the return on the market, σi = standard deviation of the return on the individual stock, and σm = the standard deviation of the return on the market. This produces a “raw” beta. Regulators commonly accept the use of Blume adjusted betas, which adjusts a stock’s raw beta towards 1.0 based on the premise that the beta for all stocks trend towards 1.0. Therefore, the three key components of a beta are: (1) correlation between the individual stock and the market index; (2) the standard deviation in returns of the individual stock as a factor of the standard deviation of the market index; and (3) the Blume adjustment⁶.

Figure 3: S&P 500 Utility Sector Beta Calculation Components

In reviewing the changes in these key components, we identify several aspects that bear mentioning regarding the change in five-year utility betas:

1. The standard deviation factor experienced swings of approximately 20%, with the factor changing from 1.03 at the end of 2019 to 1.21 by the end of April 2020. It dipped to 1.15 during part of the interval period, and is now back down to 1.03 at the end of May 2025. This indicates that the utility sector’s standard deviation of weekly returns increased (decreased) at a quicker rate than the S&P 500 Index’s standard deviation.
2. The correlation of returns between the utility sector and S&P 500 Index saw an even more dramatic rate of change. The correlation between the two was 0.30 at the end of 2019, increased noticeably to 0.65 by the end of April 2020, went to around 0.70 in the interval period, and now declined to 0.54. It appears this is the main reason why betas have not drawn back to the pre-early 2020 estimates.
3. A unique feature of the impact on betas from the early 2020 period is that both the standard deviation and correlation moved in lock-step, i.e., both increased at the same time. In previous years the two components generally moved in different directions.
4. The Blume adjustment’s impact of increasing utility betas was considerably less since early 2020 to present, given that the underlying raw beta was much closer to 1.0 than before.

It is expected that the decline seen in betas will put downward pressure on the return on equity results measured using the CAPM model, all else being equal. The CAPM measures the systematic risk of a company and its expected return. The beta measures the systematic risk of the company⁷. To illustrate the impact on CAPM ROE results, we developed an illustrative example using a risk-free rate of 4.60% and market risk premium of 7.50%, together with two beta scenarios of 0.89 to represent the five-year utility beta calculation that includes the early 2020 period and 0.70 to represent the now-lower utility betas. As shown in Figure 4 below, the CAPM ROE result falls from 11.28% to 9.85%, a decline of 143 basis points.

Figure 4: Beta Impact on Utility CAPM ROEs⁸

It will be of significant interest to observe how a change of this magnitude in CAPM results will impact authorized ROEs for regulated utilities. In many instances, a regulator decision is informed by additional analytical models, over and above the CAPM, together with its judgement to determine the appropriate ROE. Therefore, the full decline in CAPM results may be negated somewhat but will nevertheless be expected to be impactful.

For more information or to comment on
this article, please contact:
BREANDAN MAC MATHUNA, PRINCIPAL
GDS Associates, Inc. - Marietta, GA
770.799.2391 or
breandan.macmathuna@gdsassociates.com

References
1. Moreover, the previous spike seen in utility sector betas was short-lived in nature.
2. The “beta” term measures the volatility of a company’s stock return relative to the market return. The price of a stock that has a beta value greater than 1.0 is assumed to be more responsive to a change in the market returns than a stock
that has a beta value of less than 1.0.
3. See Examination of Utility Betas, GDS Associates, August 30, 2023, available at https://blog.gdsassociates.com/examination-of-utility-betas.
4. See id., at Figure 2: S&P 500 Utility Sector Betas At Different Study Periods.
5. Another common approach of expressing the beta formula is Covariance (Ri, Rm) / Variance (Rm).
6. As a general guide, please note that holding all else equal, if the correlation increases, the beta estimate increases; if the standard deviation factor increases, the beta estimate increases; and the farther away the raw beta is from 1.0, the
larger the impact will be from the Blume adjustment to bring the beta closer to 1.0.
7. The CAPM methodology is mathematically expressed as: ERi = Rf + βi (ERm – Rf), where: ERi = expected return on investment, Rf = risk-free rate, βi = beta, ERm = expected return of market, (ERm – Rf) = market risk premium.
8. The illustrative assumptions used in this example are generally representative of estimates computed at the time of writing using FERC’s preferred methodological approaches.