Transaction Costs and Markets in Networks

Coase's work on radio spectrum policy is relevant to current electric grid issues

Last week my friend and co-author Mike Munger published our conversation on his podcast, The Answer is Transaction Costs: Smart Grids, DERs, and the Economics of Energy. He and I are both deeply interested in the work of Ronald Coase and applying Coase’s extensive insights to a range of topics. Our interests focus on applying Coase’s transaction cost logic and framework to emerging technology issues, his in the sharing economy and mine in electric grid digitization. In both instances Coase’s insights get us to thinking about digital platform business models.

But I’m getting ahead of myself. I shared the podcast link on various social media platforms, and in the comments section on my post on LinkedIn, my friend and co-author (and long-time collaborator on the old Knowledge Problem blog) Mike Giberson made the following insightful comment:

Coase’s study of the FCC was among the analyses that resulted in the FCC auctioning access to the airwaves instead of the old regime of the federal government awarding licenses to “deserving” applicants. I’m not at all sure he would have opposed moving the electric power industry – an institution rife with special government privileges in the last century – toward use of markets in areas where competition can operate.

Mike is exactly right in his interpretation of Coase.

Coase was an institutional economist in the same classical liberal intellectual tradition as Adam Smith and F.A. Hayek, and focused on reintroducing institutional analysis to the body of ideas that neoclassical economics had made institutionally sterile. Thus to understand and apply Coase we have to understand and apply both transaction costs and comparative institutional analysis, comparing one realistic set of arrangements with another set of realistic arrangements.

I’d like to use his observation as a jumping-off point to do three things: define transaction costs more accurately, discuss the history of the U.S. Federal Communications Commission's approach to radio spectrum and how it changed over time thanks to Coase’s work, and draw some relevant parallels between radio spectrum and electric grids/wires networks. To help me in this I am going to excerpt extensively from my 2021 book in the Fraser Institute's Essential Scholars series, The Essential Ronald Coase. Fraser did a fantastic job of producing materials to accompany the book, which is available online at zero price so you have no excuse for not reading it, and I'll link to one of those as well.

What are transaction costs?

Transaction costs are not the easiest concept to understand.

Coase defined transaction costs as consisting of all costs of using markets, contracts, and the price system. [Simon Fraser University economist Douglas] Allen defines transaction costs as the costs of “establishing and maintaining property rights” (1999: 898). Transaction costs affect the distribution of property rights across all types of governance structures and organizations. Coase never defined transaction costs explicitly, relying instead on examples to illustrate how they affect contracts, incentives, and outcomes. (Kiesling 2021, p. 6)

Think of transaction costs as things that make it harder or more costly for two parties to engage in mutually beneficial exchange. Some transaction costs are essential to being able to trust the underlying property rights, such as requiring car owners to have a legal title and to transfer that title when selling the car. Search costs, the cost of finding trading partners, come immediately to mind as a transaction cost (although some institutional scholars distinguish search costs from transaction costs, but I don’t). The institutional structure of a market will affect how high search costs are, and how much of the available surplus they can create. We can apply Coase’s logic using experimental economics to test the effects of different market institutions, as experimental economists have been doing since the 1960s.

For example, compare a bilateral market arrangement and a double auction where the underlying market supply and demand values are the same. If you are unfamiliar with experimental economics, we model the supply and demand in advance, developing an analytical prediction of the market-clearing price and quantity. We then take each unit of output and the willingness to pay (buyers)/willingness to accept (sellers) and divide them up among everyone in the room, so each person is either a buyer or a seller and each one has their own set of units on the demand or supply curve. We can then use this analytical prediction as a benchmark against which to compare the actual performance in the room, since we have used the same underlying market model and data with both institutions (thus the comparative institutional analysis).

In a bilateral market buyers and sellers are spread out. They have to find each other and engage in 1:1 negotiation, which may or may not yield an agreeable exchange. In a classroom setting, you can even make the sellers stand in place (like having a fixed store location) while the buyers can move around. The time and effort spent moving and negotiating are transaction costs, and in bilateral markets like these they can be high enough that the participants cannot create close to 100 percent of the possible surplus/gains from trade. The distributed nature of the negotiation and having to find each other limits the rate of information creation and transmission in that particular knowledge ecosystem.

In a double auction, in contrast, buyers and sellers simultaneously submit bids and offers, other participants can see their bids and offers, and trades are consummated quickly. This is an extremely information-rich knowledge ecosystem. Double auction prices converge to the analytical prediction quickly, and participants create close to or exactly 100 percent of the surplus almost all of the time. There are costs associated with establishing the market platform, and in an actual market things like certifying their identities and credit-worthiness are transaction costs, but if we want to assess the relative transaction costs in these two institutions we have to compare the low setup costs and higher search costs in the bilateral market with the higher setup costs and very low transaction costs in the double auction market. If the setup costs of the double auction are low enough, then that’s worth incurring to get the reduction in transaction costs. This type of analysis is Coase in action.

Regulatory polices are another source of transaction costs:

Importantly, Coase stressed that various forms of government regulation, including state-owned property, are alternative ways of performing the coordinating tasks of property rights and entail their own transactions costs. Chief among these transactions costs are the barriers that regulation often puts in the way of firms that wish to reorganize to operate more efficiently or to introduce new innovations. These costs are often unseen and, more dangerously, ignored in traditional economic analyses. (Kiesling 2021, pp. 6-7)

The Fraser Institute video What Are Transaction Costs? provides more background on the definition of transaction costs. I like the way we defined them in the intro to the video: “Transaction costs include the time it takes to find someone with which to trade, the time and costs to negotiate the conditions of the exchange, and the costs to enforce the contract afterwards, if necessary.”

The Federal Communications Commission and Radio Spectrum

The most knowledgeable economist writing about spectrum history, property rights, and policy is Thomas Hazlett at Clemson, and most of what I know about this issue and discuss below I have learned from him. With the refinement in the early 20th century of technologies for communicating using the radio spectrum, both point-to-point communications and broadcast (radio) communications grew and the radio spectrum was scarce by the 1920s. Scarcity means interference, using a band of the spectrum that is too close to that of another user, so the sound waves of the two streams interact and reduce the fidelity of the signal in each. What we know now is that radio spectrum is a classic case of a common-pool resource, a situation in which defining and enforcing property rights is difficult and costly. But in the 1920s these conflicts were seen as political conflicts between government spectrum users like the Navy (which argued for a government spectrum monopoly that they controlled) and the burgeoning private radio industry, among others. In 1927 Congress passed legislation establishing the Federal Radio Commission, which in 1934 would be folded into the Federal Communications Commission (FCC), authorizing them to regulate radio spectrum in accordance with “public interest, necessity, and convenience”.

The FCC chose to do so first by hearings and then by spectrum license lotteries:

Between 1927 and 1981, the FRC/FCC awarded licenses using comparative public interest hearings, a process that according to the US Congressional Budget Office “weighs the relative merits of the contending applicants”—and a process that telecommunications economist Thomas Hazlett called “socially wasteful and politically charged” (1998: 530). In 1981 the FCC switched from the hearings to using lotteries to allocate spectrum licenses, which de-politicized the process but did not ensure efficient license allocation and continued the process of wasteful rent seeking (as lottery applicants had to fill out voluminous documents to establish their “public interest” credentials). (Kiesling 2021, p. 28)

The permissions granted in the FCC hearings meant that the agency stipulated the license-holder’s specific service offered, technology used, and business model, which reduced competition among license-holders and denied licenses to many potential competitors. Those granted licenses profited handsomely while as a whole the radio spectrum resource was underutilized. No innovation came to market, including innovations that could reduce interference and thus accommodate more, and more varied, services within the radio spectrum. It’s easy to see why Hazlett called this process socially wasteful and politically charged!

Coase critiqued the FCC’s hearings-based license allocation procedures in what was ultimately an extremely influential paper, The Federal Communications Commission (Journal of Law & Economics Vol. 2, pp. 1-40). As an institutional economist Coase excelled in highlighting the history of a particular business practice, law, or regulation, and in this case his historical narrative of the evolution of the FCC’s spectrum license policy led him to argue for a policy change.

Coase asked if there was a feasible way to allocate the use of radio spectrum to create the most possible value out of it, which the then-current public interest hearings method did not accomplish. The policy objective should be not to minimize interference along the spectrum, but to maximize output from the spectrum, treating interference as a constraint to be managed (or something that innovation would reduce). Why not define a property right in a specific part of the spectrum for each user, and make those rights tradable? Coase here followed the suggestion of Leo Herzel (1951), who proposed defining spectrum ownership rights and allocating them through auctions.

Coase claimed that despite arguments to the contrary, the scarcity of spectrum does not necessitate its administrative allocation, ongoing regulation, or government ownership. (Kiesling 2021, p. 29)

Coase argued that moving from hearings-based allocation of licenses to a license market in which the licenses were tradable would generate more value and reduce barriers to the types of innovation that could make spectrum more plentiful by reducing interference. He called it the FCC defining a property right, but it’s more accurate to apply Ostrom’s nomenclature to say that the license defines a well-specified use right in the common-pool resource. The radio spectrum is still a CPR, but through using markets as governance institutions we can create more value and neither under-exploit nor over-exploit the CPR. From his 1959 article:

Certainly, it is not clear why we should have to rely on the Federal Communications Commission rather than the ordinary pricing mechanism to decide whether a particular frequency should be used by the police, or for a radiotelephone, or for a taxi service, or for an oil company for geophysical exploration, or by a motion-picture company to keep in touch with its film stars or for a broadcasting station. Indeed, the multiplicity of these varied uses would suggest that the advantages to be derived from relying on the pricing mechanism would be especially great in this case. (Coase 1959, p. 16)

It took 34 years, in part because as a network resource spectrum licenses are complicated and spectrum license market design is extremely challenging (involving complicated combinatorial auctions to create what became known as package bidding for groups of licenses) and in part because of the political economy of persuading people in agencies who did not want to relinquish control over the resource. But in 1994 the FCC held the first spectrum license auctions, a process that has played a fundamental role in unleashing the vibrant economic and technological dynamism of the digital era and has created the foundation of the digital and digitized economy.

How Does This History Help Us Understand Electric Systems?

I want to expand on this more later but get the ball rolling now. As Mike Giberson said, both the communications industry and the electricity industry have been “an institution rife with special government privileges in the last century”. That means that the comparative institutional analysis has to start with the actual, historical regulatory institutions, in much the same way that Coase started his FCC analysis by evaluating the performance of the regulatory institution compared to what's possible by making the incremental move to using markets “where competition can operate”, as Mike said. The technologies and the physics are much more complicated in electric networks than they are in communications networks, so the analogy is not perfect, but the challenge is to use markets “where competition can operate” and to do so in a value-creating way. The establishment of wholesale power markets in the 1990s was motivated by precisely this logic. While these markets have created substantial value and have reduced energy costs for customers (recall that wholesale markets affect only the energy portion of the regulated rate, not the distribution wires charges or other charges), they are still in many respects more political institutions than the FCC's spectrum license auctions. And it can’t just be because the RTO markets are monopoly platforms, since the FCC auctions are also in a monopoly setting, with no one able to compete with the FCC as the market platform provider

The other way the radio spectrum history is relevant is that the electric wires network, as an alternating current real-time network, is more of a common-pool resource than most people realize. The physics of alternating current makes it impossible for a contract path and a physical path to match (except for simple cases like radial lines and even then it’s iffy), and there are several steps in the physical supply chain of electricity where ill-defined property rights are unavoidable. I think that means that we should be thinking about electricity policy MUCH more as an Ostrom-style CPR governance problem, where we have to have an institution that defines and enforces use rights in the CPR. Like Coase, I think the best way to do that is to make as much use of markets and of tradable rights as is possible. There will be cases, such as sub-5-minute actions and emergency situations, where that won’t be feasible, although with innovation it may become so in the future and we should not foreclose that greater move to markets. But we should identify those boundary cases and work to reduce transaction costs and lower entry barriers for the rest.

Coase gets the last word here, from his Nobel address in 1991:

In fact, a large part of what we think of as economic activity is designed to accomplish what high transaction costs would otherwise prevent or to reduce transaction costs so that individuals can freely negotiate and we can take advantage of that diffused knowledge of which Hayek has told us.