Measuring the Power of Regulatory Regimes

Schumpeter (1942, pp. 104–106) contends that innovation is incompatible with the model of perfect competition and therefore should not inform the practice of economic regulation. He further observes that “perfect competition is and always has been temporarily suspended whenever anything new is being introduced—automatically or by measures devised for the purpose—even in otherwise perfectly competitive conditions” (p. 105). See also McNulty (1968).

Laffont and Tirole (1993, p. 11) provide three discrete characterizations of the power of regulatory incentive schemes: Cost-of-Service regulation is categorized as having “very low” power; incentive regulation (inclusive of cost and profit sharing) is characterized as having “intermediate” power; and price-cap regulation in which the regulated firm is the residual claimant is characterized as having “very high” power. Weisman (2019b) defines the power of the regulatory regime as the ratio of the regulated firm’s investment in cost-reducing effort and the first-best level of cost-reducing effort. See also Laffont and Tirole (1993, p. 40, Table 2).

PCR is commonly characterized as a fixed-priced contract because the price cap for the regulated firm is invariant (fixed) with respect to its performance for the duration of the regulatory regime (Laffont and Tirole, 1993, p. 40). Specifically, the regulated firm is typically limited in its ability to pass through cost increases in the form of rate increases, which mitigates the moral hazard problem that arises under RRR.

Similar issues arise with respect to capital investments if the regulator limits the returns on successful investments and forces the firm to bear the full financial consequences of unsuccessful investments. See Kolbe and Tye (1991) for further discussion of this important issue.

RRR is commonly characterized as a cost-plus contract because the rates that consumers pay tend to vary directly with the reported costs of the regulated firm. Specifically, the regulated firm is typically able to pass through a high proportion of cost increases in the form of rate increases (Laffont & Tirole, 1993, p. 40). Even under RRR, however, the regulated firm’s expenditures and capital investments are subject to prudence tests.

This is probably a more accurate characterization of the practice of PCR in Europe than in the United States. See, for example, Sappington and Weisman (2010) and Crew and Kleindorfer (1996).

These shared earnings may be used to effect various forms of redistribution (Posner, 1971). In the U.S. telecommunications industry, these shared earnings have been used to fund bill credits, rate refunds, and infrastructure improvements (Sappington and Weisman, 1996, Chapter 3). For example, in the U.S. telecommunications industry, shared earnings have been used to expand broadband infrastructure in rural areas to address “digital divide” concerns. The use of shared earnings for infrastructure improvements in other regulated industries—including electric power, natural gas, and water—is seemingly less common. From this perspective, shared earnings are a form of “consumer dividends” that regulators distribute to their constituencies in myriad ways. In addition, these “dividends” allow consumers to partake in the higher earnings of the regulated firm and therefore can serve to lessen the pressure on regulators to recontract when regulated firms report higher earnings (Sappington and Weisman, 1996, p. 334). The risk of regulatory recontracting and its effect on the power of the regulatory regime is explored in Sect. 5.

It can be shown that this metric is equivalent to the PRR metric that is employed in Weisman (2019b) when the latter is adjusted for output differences across regulatory regimes and effort costs are quadratic.

The regulator is typically not able to lower the price cap directly without breaching its commitment; but it can do so indirectly through entry-accommodation—policies that are advantageous to new entrants—that renders pricing below the cap a more profitable strategy for the regulated firm (Weisman, 2019a, 2019b). For example, under the provisions of the 1996 Telecommunication Act, incumbent providers are required to lease unbundled network elements to rivals at prices that are set by regulatory fiat and that allow competitors to resell their services at prescribed discounts (Kahn et al., 1999). The regulator can use these network element (input) prices and resale discounts to force the regulated firm to reduce retail prices below the price cap when its costs decrease. The ratchet effect can therefore arise even under pure PCR.

The Arrow effect (1962) recognizes that the firm’s incentive to undertake unit-cost reductions increases with output, ceteris paribus. See also Cabral and Riordan (1989).

In an unregulated setting, \(\rho\) is a measure of the share of cost savings that the market allows the innovating firm to retain. \(\rho = ( < ){\kern 1pt} {\kern 1pt} 1{\kern 1pt}\) for an unregulated monopolist (Cournot oligopolist) since s = 0, \({\kern 1pt} \frac{dp}{{dc}} > 0{\kern 1pt}\) and \(\frac{p - c}{p} = ( < )\frac{1}{\varepsilon }\) in equilibrium. For an unregulated monopolist, the ratchet and Arrow effects are perfectly offsetting, while the ratchet effect dominates the Arrow effect for a Cournot oligopolist. Finally, in the case of atomistic Cournot competition it can be shown that \(\rho \to {\kern 1pt} {\kern 1pt} 1{\kern 1pt} .\)

Weisman (2019b) considers the possibility that s endogenously influences the regulator’s choice of p. In that framework, the PRR can be higher under PCR with earnings sharing than under pure PCR.

Under the PCR regimes that are commonly employed in the U.S., the firm’s prices are set to reflect prevailing costs at the start of the PCR regime, and then are permitted to increase, on average, at the rate of economy-wide retail price inflation (I), less an offset (X)— the “X factor”— for the remainder of the regime. Under this form of PCR, the X factor measures the extent to which productivity in the regulated industry is expected to increase more rapidly and industry input prices are expected to increase less rapidly than in the general economy (Bernstein & Sappington, 1999). Because the X factor is developed on the basis of a peer group of regulated firms, there is no guarantee that it delineates a “competitive” price trajectory. Rebasing therefore provides an additional competitive check (correction) on earnings. As the industry becomes more competitive, the length of the PCR regime increases along with the time between rebasing until it may no longer be necessary (Sappington & Weisman, 2010).

Earnings sharing enables consumers to share in realized efficiency gains throughout the PCR regime (Lyon, 1996; Schmalensee, 1989), albeit at the cost of lower overall efficiency gains (Cabral & Riordan, 1989; Weisman, 1993).

To address this issue, regulators in select jurisdictions have adopted efficiency-carryover mechanisms that allow the regulated firm to carry over a stipulated proportion of its excess (deficient) returns into the subsequent PCR regime. See, for example, Alberta Utilities Commission (2012, Section 9). These mechanisms would be expected to exhibit the same qualitative effect as increasing the duration of the regulatory regime.

Sappington (2002, pp. 251–2) discusses the factors that inform the length of time between regulatory reviews, while Armstrong et al. (1995) explore the optimal regulatory lag. There are similarities between the optimal regulatory lag and the optimal patent life (Scherer, 1972; Nordhaus, 1969, ch. 5): In both cases, the objective is to encourage firms to undertake innovation without subjecting consumers to supra-normal prices indefinitely. Nonetheless, there are also differences: Patents shield firms from competition pro tempore to encourage innovation by preventing instantaneous replication by rivals. Regulatory lag does not shield the regulated firm from competition as much as it emulates the competitive process. See note 3 supra. It does this by permitting the regulated firm that innovates successfully to retain the benefits of that innovation for a transitory period of time before they are appropriated by the regulator and passed on to consumers in the form of lower rates. I am grateful to Lawrence White for this observation.

The regulated firm and the competitive firm are assumed to employ identical discount rates. In addition, it is implicitly assumed that \(\rho = 1{\kern 1pt}\) for the benchmark competitive firm. See note 12 supra.

It is straightforward to show that \({\kern 1pt} \forall n \le l,{\kern 1pt}\) \(DAF(n,r,l) = 1\) only when n = l = 1 or n = l = ∞. Hence, except in degenerate cases, \(\rho^{D} < 1.\)

In order to properly emulate the competitive process, the regulator would opt not to reflect in current-period rates any firm-specific cost savings realized less than l periods prior. This may be referred to as sequential regulatory lag.

In an early survey of the performance of incentive regulation in the telecommunications industry, Kridel et al. (1996) report only modest efficiency gains in the transition from RRR to PCR. These early forms of incentive regulation were of relatively short duration and frequently incorporated earnings-sharing mechanisms. Uri (2001) reports similar lackluster efficiency gains from the early adoption of PCR. In subsequent years, these efficiency gains improved concomitantly with an increase in the length of the regulatory regime and the elimination of earnings sharing (Abel, 2000; Ai and Sappington, 2002; Eckenrod, 2006; Resende, 2000; Sappington and Weisman, 2010, pp. 236–239; Seo and Shin, 2011). Nonetheless, the overall empirical evidence on the efficiency gains associated with transitioning from RRR to PCR is best characterized as “mixed” (Sappington, 2002, pp. 278–282). The empirical evidence on performance-based regulation (PBR) in electric power is less comprehensive than that for telecommunications. However, Hellwig et al. (2020) report efficiency gains from high-powered incentives in German electricity networks. See also Crowley and Meitzen (2021) and Domah and Pollitt (2001). In general, the performance gains from adopting PCR are less dramatic than the theory would have predicted. This may be the case because the differences between RRR and PCR in practice are less pronounced than they are in theory (Laffont and Tirole, 1993, p. 19).