Peter Klein points out that the term “transaction costs” doesn’t have a definitive definition, and there’s some tension between how economists use it. He links to a very good chapter by Douglas Allen, which discusses transaction costs and the two main approaches to them. My exposure to the literature is incomparable to that of Klein and Allen. But, I have the audacity to offer a definition, based on how I conceptualize the term. While I was thinking about this definition, I was aiming to define it to the non-expert. Still, I think that if there’s something to it, it can be built on.
Before defining transaction costs, let’s think about exchange in a “frictionless world.” There are two people, Person A (PA) and Person B (PB), and they can only exchange with each other. PA is a farmer who produces corn, and PB is a tailor who produces suits. The latter has a stock of one suit, and former a stock of 10 ears of corn. PA’s opportunity cost to selling his 10 ears of corn is the value he ascribes to consuming it, which is less than or equal to one of PB’s suits. On his end, PB’s opportunity cost to selling the suit is the foregone value of wearing it, which he thinks is less than or equal to eight ears of corn. There are gains from trade to be had, so PA approaches PB, and they exchange. Let’s focus on what costs mean in this situation. PA’s cost of doing trade is the need to reimburse PB’s opportunity cost, and the same is true the other way around.
Now, let’s add a friction: PB suffers from an intense hatred of PA. In order to exchange with PA, PB would need no less than 12 ears of corn — eight for the suit and four to compensate for the cost of having to trade with someone he despises. The requirement for four additional ears of corn represents a transaction cost. The additional cost is, in a qualified sense, external to the exchange being made. It’s a cost that has nothing to do with the opportunity cost of the item that PB really wants.
I see transaction costs as a very large group of costs: search, distance, technological constraints, et cetera. These are all additional obstacles that at least one party in an exchange has to bear the cost of. But, they are also costs that are separate from the ultimate goods these parties are interested in. To clarify my point, we can look at Ronald Coase’s use of transaction costs in “The Nature of the Firm” (1937) and “The Problem of Social Cost” (1960).
In his 1937 article, Coase defines transaction costs as the costs firms face when using the price mechanism. In 1960, by contrast, transaction costs are used in the context of externalities, where property owners can exploit mutual gains from trade, but don’t because of various transaction costs. I actually don’t see much of a difference in how the term is used. Whether it’s the firm or property owners, both agents are looking to exploit gains from trade. Both are experiencing costs to trade that are additional to the opportunity costs they are reimbursing for the products being traded themselves. These costs may be different in particulars, but they are similar once we generalize. I interpret these two articles as Coase using the same general concept to explain two different phenomena.
This definition also makes sense if we relate it to the literature on the “transaction cost sector” — economic activity that focuses on reducing transaction costs. Let’s say that the flow and stock of tradeable goods is fixed, so that we don’t have to worry about shifting margins and consequent changes in marginal rates of substitution. For PB, the opportunity cost of his suit will always be equal to or less than the value he ascribes to eight earns of PA’s corn. Because those eight ears of corn will always be part of the exchange (that is the ultimate good that PB wants), that cost, under our assumptions, cannot change. But, transaction costs can change, because it behooves people to specialize in reducing them, as long as the costs of reducing them (for those people) are less than the total cost of transacting.
Coming back to our example with PA and PB, assume that there’s now a third agent who enters the market, Person C (PC). PC only produces “labor,” or hours, and he specializes in exchange intermediation. The opportunity cost of the hours required to intermediate a trade between PA and PB is equal to or less than two earns of corn. PB only hates PA, not PC, so he is willing to sell PC his suit for eight ears of corn (not the 12 he wants PA to pay). It’s within PA’s interest to pay PC in between one and two earns of corn to intermediate the trade, because even after incurring that cost, it’s still cheaper than trading with PB directly, and now the cost is within the range PA is willing to pay for the suit (up to 10 earns) — specifically, eight ears to reimburse PB’s opportunity cost, plus one or two additional ears paid to PC, to reimburse PC’s opportunity cost. Thanks to PC, a mutually beneficial trade has occurred, whereas without PC, no trade would have happened — society’s welfare has increased.
How do institutions relate to this definition of transaction costs? Money is one example: one of the big benefits of having a general medium of exchange are the reduced additional costs that producers have to incur when seeking mutually beneficial exchanges. But, I want to bring attention to a more general relationship. Institutions are rules, and rules constrain the range of choice. Institutions can reduce transaction costs by removing choices from an agent’s set that would increase the transaction costs of exchange. For example, a rule that protects against fraud eliminates the added cost of the risk of being defrauded — a risk that otherwise could be high enough to make the exchange unprofitable, but that is, in principle, separable from the items being exchanged. Actually, interestingly, institutions can both decrease and increase transaction costs: they can increase costs to “undesirable” actions and lower costs to “desirable” actions (of course, poor institutions can do the exact opposite!).
To me, this seems like a simple definition of transaction cost. The problem is that if you think the profession is missing something simple — that is, if anybody actually disagrees with what I write here —, you’re probably wrong. There’s also the possibility that what you’re saying is obvious and you’re not going deep enough into the subject. But, I’ll give it a shot anyways.