Are Transaction Costs just Costs?
Are Transaction Costs just Costs?
A Tribute to Yoram Barzel (1931-2022)
Another giant of applied price theory, Yoram Barzel, has passed.
In a tribute a few years back, Doug Allen described Barzel as one of the three giants of property rights economics, alongside Armen Alchian and Harold Demsetz. We often talk about “UCLA price theory,” and although Barzel spent his career at the University of Washington, he was very much in the UCLA tradition that informs this newsletter. Here are our previous posts discussing the insights of Alchian, Demsetz, and Barzel.
Of the three, Barzel has been the most underrated, by us and others. This week’s newsletter is my small tribute to Barzel’s intellectual legacy.
Between Property Rights, Transaction Costs, and Back Again
Put yourself in 1961 as an economist. Paul Samuelson is king. Arrow-Debreu is spreading; game theory is largely relegated to weird math and operations departments. Economic theory is what Pete Boettke calls “institutionally antiseptic.” Institutions are driven out of the discussion in favor of focusing on production sets and marginal conditions.
Against this backdrop, property rights economics starts developing. With property rights taken into account, now the law matters for economics, so the field of law and economics starts up at the same time. (A field that we are still promoting today at the International Center for Law and Economics, where I work, btw.) The Journal of Law and Economics started in 1958, and it just published Coase’s 1960 article on “The Problem of Social Cost,” but it will be a few years before Demsetz’s “Toward a Theory of Property Rights” comes out.
The goal of property rights economics, not that it was an explicit goal at the time, was to explain two things: the cause of property rights and the effects of property rights. That is, these economists want to explain property rights as an economic phenomenon and to understand the effects of property rights on other economic outcomes like prices, quantities, and distributions.
This is the world in which Barzel gets his PhD at Chicago under Arnold Harberger. From there, we went to Washington, where Doug North was, and later others like Steven Cheung and Walter Oi joined. He’d spend the next 40 years leading this property rights revolution. Well, maybe I should say attempted revolution since we are still working on it.
Like Alchian and Demsetz, Barzel’s work was wide-ranging and too much to cover in a single newsletter. His most famous work is a book called “Economic Analysis of Property Rights,” which synthesized much of his earlier work on the topic.
Between Property Rights and Transaction Costs, and Back Again
As soon as you start thinking about property rights, it becomes clear you need to worry about the transaction costs of the property. The reverse is true too. If there are no transaction costs, then Stigler’s Coase Theorem holds, and the property rights don’t matter. For example, if there are people who know how to drill for oil, any land with oil is extra valuable to them. Without transaction costs, regardless of who starts with the property rights to the land with oil, the land will end up with the people who value it most.
That’s not true with positive transaction costs. If transaction costs are positive, then Coase’s Coase Theorem holds, and the property rights do matter for who ends up with what. In the extreme, if they are sufficiently high, no trade will happen, and the land with oil will not end up owned by people who can use it. It will remain relatively low-valued, such as for cattle grazing.
Property rights and transaction costs are intimately tied together. In my head, property right economics is a subset of transaction cost economics. I’m not sure if that’s the agreed-upon taxonomy in the history of thought.
Transaction Costs are Costs
The most important contribution of transaction cost economics was to point out to Samuelsons, Arrows, and Debreus of the world that transaction costs are very real, they matter, and if we ignore them, we will say and do dumb stuff.
Start with the benchmark, Arrow-Debreu model; producers grow apples, and consumers eat apples. Producers have a cost to growing an apple, and consumers have some willingness to pay. Prices adjust so that the amount of apples grown equals the amount of apples consumed, and the marginal cost to grow an apple equals the marginal value of eating an apple. All the usual stuff.
But after the apple is grown, it doesn’t magically appear in the consumer’s belly.
The most obvious additional cost (after the physical apple exists) is that the apple needs to be transported from the orchard to the consumer. That is a cost of transacting that is different from the cost of production. However, that is maybe just a semantic issue. We need to redefine the “good” as not a physical apple but a good in the consumer’s belly. After all, we are subjectivists, and “apple in belly” or maybe “apple within reach” is what the consumer really wants. Now, the appropriate notion of marginal cost is the cost of production plus the cost of shipping.
But what if the consumer doesn’t even know that the apple exists? Then the apple has no value. In response, the seller may run ads to spread the word about their apples. Or they may pay grocery stores for a good location in the store with “trade promotions” so the consumer sees the apple and learns about it. Again, we need to redefine the good being produced. It is now an “apple that the consumer knows about and is within reach.” All of these costs (production, shipping, advertising, and trade promotions) make up the cost of the good.
Instead of advertising and trade promotions being wasteful, as many economists argued since they aren’t in the Arrow-Debreu model, they are costs of the good, just like any other. Banning advertising, for example, would be like banning any other factor of production, say fertilizer. It wouldn’t eliminate the production of apples, true, but it would shift production to less efficient inputs and raise the cost to society. We can’t assume away transaction costs.
Transaction Costs are not Just Costs
But transaction costs are slightly different than other costs, according to Barzel. They aren’t just “costs those other dummies are ignoring,” as if they ignored capital and forgot to measure it.
Barzel, building on Alchian’s work, stresses informational and measurement costs as the key transaction costs. It may seem like our newsletter keeps returning to information, week in and week out. Because it does! That’s the key behind everything.
Barzel’s focus on information is evident throughout his book, but an easier starting point for readers is his 1985 paper “Transaction Costs: Are They Just Costs?” Barzel’s answer is no. Just is too far; they are like other costs, as I argued above, but there is a key difference.
For Barzel, property rights over an asset mean that a person can gain from that asset, either by using it or by selling it. However, it is costly to measure every aspect of an asset; we can never say what the asset is from a God’s-eye or Arrow-Debreu perspective. Information is the key friction that causes property rights to be imperfectly defined. Barzel writes:
It is my contention that costly (intermediate and final) product information is the central problem of transacting, leading to all other transacting problems.
Again, contrast this with the Arrow-Debreu model without property rights disputes. Everyone knows what commodity X is.
When we don’t perfectly know (or maybe don’t agree) what the item is, then other people may try to appropriate the property right. The most obvious example is theft. Again, there is zero theft in Arrow-Debreu. But imperfect information also leads to adverse selection, free riding, or shirking, which are not in the standard Arrow-Debreu model. All of these are attempts to appropriate value from others. With transaction costs, people can gain at the expense of others. It is not all mutually beneficial trade.
Also, in a world where theft (for example) can occur, people will incur transaction costs to better protect and delineate the property rights of their assets. Maybe we could shoehorn these costs into just the costs of production. Now the apple is the “apple that won’t get stolen,” but now we are really stretching the Arrow-Debreu framework.
There is one prediction of positive transaction costs that, according to Barzel, cannot occur in a modified Arrow-Debreu. Since people can gain at each other’s expense, in order to minimize that loss, people will restrain themselves. I am willing to lock my house door, which constrains me somewhat, in order to protect my assets from appropriation.
More importantly, people will “erect social institutions to impose and enforce the restraints.” Now, since it isn’t costless to know the apple is mine since it could have been stolen, we will set up complex social institutions to settle disputes. Courts, police, and a myriad of other things that constrain us develop to deal with disputes over property rights, which are disputes of information. Agreeing to these institutions imposes a cost on me.
Barzel uses the example of expiration dates for dairy products in the store. Why do they exist? Why does the government force the store to get rid of expired milk? From an Arrow-Debreu perspective, this imposition is arbitrary and wasteful. The market can adjust prices down as time goes on. Maybe there reaches a point where no one buys the milk, and it must be thrown out, but why impose a restraint against trade?
But transaction costs do exist! When people may not know the exact quality and date of the milk, this imposition reduces the transaction costs for consumers. They no longer need to exert as much effort to inspect the milk, which is costly in this world. Now, the government mandate may be too loose or too stringent, but it isn’t “wasteful” in the same way that it would be in an Arrow-Debreu world where it literally has no benefit.
As with all of the property rights giants, there is much more depth than I can convey in 1,800 words. If I took 20,000, I still wouldn’t get there. I’m working toward it myself.
But I hope this introduction to property rights and Yoram Barzel encourages you to read his articles or books. Here is a thread with a few more of his classics.
RIP, Professor Barzel.
Clearly written summary, which is extremely valuable. You connect transaction cost economics to the issue of institutions well here.
For years I've been saying we need a popular book, with the reach of Malcolm Gladwell, that explains New Institutional Economics. Many "ideological" debates became much less ideological when one understands transaction costs and then observes different institutional frameworks as various imperfect approaches to the issue of transaction costs. When one adds the institutional diversity of the Ostroms, public choice theory, and the literature on political ignorance (see Mark Pennington's "Robust Political Economy" as a nice integration of public choice and political ignorance issues), then one has a powerful toolkit for discussing institutions. Surprisingly few economists and few political scientists have adequate mastery of these tools.
But starting with property rights and transaction costs as you do here is a graceful introduction to the set of tools that will allow for more insightful and productive discussions of institutional choices and challenges.
I'd love to see you write such a book someday.