As of the mid-1980s, there were two main choices for government regulation of large public utilities like electricity transmission companies and water mains, and neither seemed adequate. One option was "cost-plus" regulation, where the government regulator looked at the costs of the regulated firm and then lets the firm charge enough to make a modest profit. The problem, of course, is that this system gives a regulated firm no incentive to cut costs or even to provide quality service. Instead, the regulated firm has an incentive to build new plants and even to run up costs where possible, because the regulators will let the firms cover those costs--and make a bit more besides.
The other option was "price cap" regulation, where the government regulator set a price that the regulated firm could charge for the next few years. Sometimes the price was set on an downward trajectory: that is, the firm would be required to charge slightly lower prices each year. However, if the regulated firm could find a way to cut costs or innovate more rapidly, then the regulated firm could earn higher profits, at least for several years until the regulators reset the price. The problem, of course, is that the regulated firm now has incentives to deceive the regulator about its costs, to get the price cap set high, and then to find ways of slashing costs and making high profits for a few years, before then trying again to persuade the regulator that costs are high when the price cap comes up for renewal.
What advice does economic analysis offer for regulators in this situation? Jean Tirole has been awarded the Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel for 2014 “for his analysis of market power and regulation," which tackles this and a number of related questions. The Nobel committee always posts some background and supporting material at its website. I'll draw on their "Popular Information" and "Scientific Information" essays.
In work in the 1980s, Tirole and his co-author Jean-Jacques Laffont tackled the question of how to regulate firms by spelling out models which clarified what regulators could know about firms. Specifically, one basic model argued that a regulator can observe costs of production at a firm, but the regulator cannot observe either the potential technologies that a firm has available for reducing costs, nor can it observe how much effort a firm has put into reducing costs. Thus, the challenge for regulators is to give firms an incentive to reveal this kind of information. In turn, this led to a number of insights.
For example, one potential approach is for the regulator to combine a cost-plus plan and some incentives. Thus, the firm announces its costs, and regulator says: "Fine, we'll let you set prices in a way that lets you cover those costs. However, if you save money, we will let you add some portion of what you save to profits, and if you lose money, we will let you set prices in a way that recoups some of your losses." Under certain conditions, this kind of formula (the "optimal static mechanism") gives firms an incentive to reveal their true costs (and thus not to pump up costs in the way that pure cost-plus regulation would encourage) and also to seek out cost savings, but to share some of the gains of that cost saving with customers (because the firm doesn't get to keep 100% of its cost savings like it would under pure price cap regulation).
However, drawing up a specific contract for sharing of potential cost savings or cost overruns is a tricky business in practice. Thus, an alternative is for regulators to offer firms a choice: either the firm can choose to be regulated by a cost-plus approach, or it can choose to be regulated by a price cap approach. The idea here is that if a firm chooses the price cap approach, it is revealing to regulators that it sees a number of ways to cut costs; if it chooses the cost-plus approach, it is saying to regulators that it doesn't see a way to reduce costs.
Tirole's style of analysis is to work through the potential issues that can arise one at a time, modelling and analyzing each one separately and then in various combinations. Thus, another issue that arises is what happens if there is a dimension of quality of service that the regulator cannot observe. A price cap approach would encourage the firm to save money by reducing quality, and so when the regulator has a hard time observing quality of service, it should offer the firm only modest opportunities to add to profits by cutting costs.
Or what happens when we think of regulation not as a one-time choice, but as an agreement that both sides know will be renegotiated over time? The Nobel committee explains the potential problem that can arise: "Suppose the firm can make a sunk-cost investment in a technology which will generate
future cost savings. If the firm invests and its costs fall, the regulator may be tempted to expropriate the investment by reducing the transfer to the firm (or tighten its price cap). If the firm anticipates this kind of hold-up problem then it may prefer not to invest. This problem is the largest when long-run investments are essential, as in the electricity and telecommunications industries."
A counterintuitive finding results in this setting: "In practice, a regulator may employ various strategies to remain ignorant about the firm’s cost. For example, the regulator may try to commit to infrequent reviews of a price cap. If this commitment is credible, the firm will have a strong incentive to minimize its production cost. However, if the commitment is not credible, the firm expects that
any cost reductions will quickly trigger a tighter price cap, and the incentives for cost minimization
vanish."
Yet another issue that arises is the problem of "regulatory capture," which refers to a situation where over time, the regulators end up looking out for the regulated industry, rather than for consumers. This dynamic is a common one. After all, the regulated industry pays a huge amount of attention to the regulatory agency, doing its best to get sympathetic folks chosen. The regulated industry necessarily provides and shapes the information on which regulators rely. The regulated industry focuses with laser intensity on the fine print of every word and comma in the regulations. And after regulators have served for a few years, they often can end up working for consulting firms or for the regulated industry, helping deal with the same regulations they wrote in the first place. In contrast, most consumers don't have time or energy to focus on regulatory agencies in a way that would counterbalance these forces.
Tirole's model argues that unsophisticated price-cap regulation--where the firm gets 100% of every dollar it saves under the price cap--offers the biggest financial incentives for regulatory capture. (Imagine regulators who set a price cap at a high level, enabling the firm to make high profits, and later are rewarded in their careers for having done so.) Thus, when concerns about regulatory capture are especially high, giving regulated firms a chance to earn very high profits--even by saving money and cutting costs!--may be unwise.
Tirole is meticulous in going through possible factors, situations, and contingencies. I've focused here on the issue of regulating large firms like electricity companies, which is in some ways at the heart of his work. But Tirole also looks at how these lessons apply for regulation across a range of other industries, including "too big to fail" financial firms, the pricing of telecommunications networks, and others. Tirole has lots to say about how large firms might compete or subtly cooperate with each other, and applies these lessons across "horizontal" markets where similar firms compete with each other and "vertical" markets that involve supply chains between firms.
One interesting branch of Tirole's work looks at the "patent race" problem, which is the issue that if many firms feel that they have a good chance to get an important patent, they may spend so much on duplicative research and development that their efforts are a poor deal for society as a whole. On the other side, if firms feel that only one company is really well-positioned to get an important patent, they may choose not to try, and without the pressure of competition, insufficient research and development may be done in developing that technology. One of the policy controversies in this areas is whether competitive firms should be able to set up "patent pools," in which firms pay a fee to use all the patents int he pool. Tirole's models suggest that patent pools are a good idea, but only if the patents in the pool can also be licensed individually--which prevents the patent pool from becoming a way to shut out small competitors who only need access to one or a few patents.
Since the 1950s, economic work in the field of industrial organization has gone through three waves, with Tirole's work serving as a canonical representation of the third wave. As the Nobel essay explains, in the 1950s the standard approach was called "Structure-Conduct-Performance (SCP) paradigm. The basic idea was that industry conditions (the number of sellers, the production technology, and so forth) determine industry structure, which determines firm conduct (pricing, investment and so forth), which in turn determines industry performance." Thus, a standard study in this approach might look at a measure of market concentration--like the share of total market output for the top four firms--and see how was correlated with some measure of profitability for the industry. "Prescriptions for government policies, particularly with regard to horizontal mergers, reflected the SCP paradigm and were largely based on these concentration measures."
Then starting in the 1960s, a "Chicago School: approach pointed out that correlation from these earlier studies didn't mean causation. Say that there is an industry with a small number of large firms, who are earning high profits. Does this mean that the large firms are earning profits by unfairly squeezing the competition? Or earning profits by using their size to be very efficient? The typical structure-conduct-performance study couldn't really separate those possibilities.
Tirole and others brought analytical tools of game theory and mechanism design to industrial organization. The Nobel committee writes: "In the 1980s, the game-theory revolution in IO [Industrial Organization] closed the circle by supplying the tools necessary to take these industry-specific conditions into account. Since then, game theoryhas become the dominant paradigm for the study of imperfect competition, providing a rigorous and flexible framework for building models of specific industries, which has facilitated empirical studies and welfare analysis." In this way, all of Tirole's specific models, illustrations, and investigations add up to fundamental change in how economics think about regulation, competition policy, and antitrust--which is what makes this body of work Nobel-worthy.