They say that the first year of law school is like an immersive course in a new language. At first, you don’t even know what the cases are about—let alone how to determine whether they’re still “good law” and in which jurisdictions. Once you learn some basic terms and rules, you begin to make some sense of the materials and even to imitate the style of reasoning they demonstrate. At some point you find you’re forming your own jargon-laden arguments on the fly.
That’s a hard enough task, but in many US law schools you’re required to learn an additional language: “economic analysis,” or at least the pidgin version that is taught in doctrinal courses. It’s not enough to be able to separate a holding from dicta; you also have to explain how that holding relates to an imaginary world in which all-seeing optimization machines bargain with each other, with and without various “transaction costs.”
Teaching this third language is not just a form of cruel amusement for law teachers. (For many of us, it is not even amusing.) Economic analysis (so called) has become part of the method of legal decision-making in many realms, so learning how to play some of its little language games has practical value. For LPE-ers who seek to overcome the hegemony of law and economics, it is essential to learn not just how to play these games but also how to critically analyze them and their premises.
Since such an education is not readily available at most law schools, a number of students have requested the Blog’s help. This post is one effort in that direction.
Our topic is “efficiency.” It’s a slippery term. Law-and-economics scholars generally use it to refer to outcomes that maximize total “welfare,” by which they mean outcomes for which people are willing to spend the most money. This is usually the sense referred to in discussions of “efficient breach” in contract or the “efficient standard of care” in tort. But “efficiency” also gets thrown around to refer to outcomes that minimize overall “cost” (as measured in money) or that maximize overall “output” (as measured in amount of goods or in money) or that all affected parties would have agreed to if there were no “transaction costs” preventing them from getting together and negotiating. None of these meanings is the same as the other, and each has serious problems. If that’s not confusing enough, economists and legal scholars also sometimes use “efficiency” in the familiar sense of achieving whatever outcome one desires in the most effective feasible manner. (And even these do not exhaust the usages: there’s also “static” vs. “dynamic” efficiency and “productive” vs. “operational” efficiency, for example.)
Economic analysis of law is constantly slipping between these meanings, cleaving them together and apart as is convenient. This lack of clarity, which at times borders on willful obfuscation, can inhibit students from challenging certain claims. To ask “but what do you mean by efficiency?” can make one appear unsophisticated or pedantic.
But that’s precisely the question we should be asking. Because there are good reasons to reject the notions of “efficiency” usually taught in 1L classes (even if—in fact, precisely because—we have good reasons to value other forms of efficiency). Here are some of them.
Efficiency, Intuitively
In its everyday use, “efficiency” is the opposite of “wastefulness.” To run a meeting efficiently is to avoid wasting time: to start it punctually, to keep everybody on topic, to keep breaks to a minimum. To run a washing machine efficiently is to avoid wasting water or electricity. To write efficiently is to avoid wasting words by adding needless examples to a list. Whatever one’s goal is, to achieve it efficiently is to do it without unnecessary use of resources, time, and/or effort.
Not coincidentally, one can also talk about efficiency in this everyday sense as “economizing.” An efficient washing machine economizes on water and so on.
Efficiency (or “economy”) of this intuitive sort is clearly a virtue, and a virtue that is especially important in systems of production and distribution. Finding ways to raise crop yields on a given plot of land with a given amount of effort can make the difference between feast or famine. Reducing the human labor needed to perform chores lessens drudgery (reduces demand for servants, creates free time for consciousness raising). Getting the most out of fisheries and fossil fuels is essential to maintaining functional provisioning systems in the future.
But the everyday usage of “efficiency” (and “economizing”) is also ambiguous. The same process can be efficient in one sense while inefficient in another. Is walking to work more efficient than driving? It depends on whether you care about wasting fuel or time. Is adopting an expensive new fertilizer efficient? It depends on whether you’re trying to save on money or effort. To get more precise about what is “efficient” requires specifying what resource you want to economize.
Or does it? Presumably, everybody who is deciding how to get to work wants to save both fuel and time and would be willing to trade off some fuel savings for some time savings or vice versa. Different people will surely prioritize the two differently, but everybody cares about each form of savings to at least some degree. What if we posited that both time and fuel usage are costs that one incurs to get to work, with the relative costliness of each depending on how much the person traveling cares about their time versus their budget (or environmental impact)? Then, couldn’t we posit that the most efficient way to get to work is one that minimizes the total cost of getting to work for that person? On this logic, couldn’t we say that to do something efficiently is just to achieve whatever benefits are desired from an activity at the least all-in cost?
Allocative Efficiency: Pareto
This is the basic intuition behind the law-and-economics notion of efficiency (and behind much of the neoclassical tradition from which law and economics draws). The intuition is sensible enough in a simple case, where one person with a given goal is considering different ways to achieve that goal. But economists take it to an extreme, developing it into a notion of “allocative efficiency” that replaces the intuitive concept with a set of highly problematic moral claims grounded on highly problematic models of how markets work. On close scrutiny, the logic falls apart.
Here’s how it goes. If “cost” is just anything that somebody does not like and “benefit” is anything that somebody does like, then to produce benefits at the least overall cost is just to produce outcomes that are most desirable to that person. To use resources, or legal rules, “efficiently” is to arrange them so that they make the people they impact as satisfied as possible: maximizing “welfare.”
How do you know if you’ve maximized welfare? In the pure theory of neoclassical welfare economics, you’re supposed to imagine a world in which everybody has a Buzzfeed-style list of The Top Trillion Possible States of the World (their “preferences”) and knows which actions to take to get as close to the top of their list as possible. On these assumptions, every time people trade with each other, they are made better off—creating “surplus” welfare relative to the previous state of the world. This is a “Pareto improvement,” meaning that at least one person prefers the new situation and nobody objects to it. (If I agree to pay you $50 for your old Contracts casebook, I must value the casebook more than I value continuing to have that $50. If you agree to accept $50 for your casebook, you must value the money more than the book. When we make the trade, we each find ourselves with things we value more than the things we had before. Overall “welfare” increases.)
With such lists in hand, these people are then imagined to enter a “perfectly competitive” market in which literally everything can be bought and sold and everybody can effortlessly find the best deal available. In such a world, people will continue to trade until they reach an “equilibrium” in which nobody can find anybody else to trade with—everybody will have discovered the best possible state of the world for them (best job, best production processes, best mix of goods and services) that is compatible with the best possible state of the world for everybody else. Markets “clear.” “Surplus” cannot increase. A “Pareto optimum” is reached, because no more Pareto improvements are possible. This process of trading is also a process of “price discovery.” Everybody will be forced to value every possible trade-off: between saving time and saving fuel, between work and leisure, between paper and plastic. The resulting equilibrium is one in which the price of everything perfectly reflects its (marginal) social value/cost (“cost” in its broad sense: every impact is accounted for.)
What does this utopic vision have to do with efficiency? It is supposed to represent “allocative efficiency” in that each resource is being allocated to the part of the provisioning system where it best contributes to the outcomes that maximize welfare (its price is said to equal its marginal cost to society as a whole). The sense in which we “reduce waste” is by putting resources (and efforts) to their “highest and best use.” To fail to use resources in this way is thus to “waste” them relative to how they could be used in a more “optimal” scenario.
Allocative efficiency according to the Pareto criterion has no necessary relationship to the intuitive concept of efficiency with which we began. Increases in various forms of intuitive efficiency—productive efficiency, energy efficiency—may or may not be allocatively efficient, because they might produce outcomes that are more objectionable to at least one person than previous arrangements. (In a world of perfect competition, any such change would arguably be competed away such that relative prices and quantities readjust to their optimal levels, but that’s a “fact” about that imagined world, not the concepts of efficiency used to evaluate them.) Conversely, an increase in allocative efficiency may or may not have anything to do with the intuitive sense of efficiency. If I sell you my casebook for $50 and we’re both happier, we have not changed anything about production processes (though we may have incidentally contributed to efficiency in resource use by reusing a book). If I convince you that you should be happier with what you have or less desirous of what you don’t, I have caused you to rank your current situation higher up in your Buzzfeed list of Top Trillion Possible States of the World, increasing the “allocative efficiency” of the provisioning process without changing anything about how things are produced or even who gets them.
And the Pareto criterion is useless in analyzing the actual world. It would take at least a whole other post to explain the problems with the “price theory” built on the concepts of perfect competition.1 For present purposes what you need to know is that we don’t live in a world of perfect competition, we never could, and most of us would prefer not to. (Can you imagine constantly running auctions on everything in your life while monitoring others’ auctions for everything in their life?) Our world is so different from the imagined paradise of perfect competition that every existing social arrangement is already Pareto optimal, since any meaningful change would invariably be objectionable to at least one person. In a sense, each arrangement (once you’re in it) is superior to every other arrangement, so there is no practical way to compare them. Meanwhile, we have no reason to assume that prices in our world even come close to approximating all-in marginal cost. So while allocative efficiency, Pareto optimality, and price equating to marginal cost come neatly together if you assume multiple impossibilities, they come apart in dealing with actual policy questions.
Allocative Efficiency in Law and Economics
Realizing at least the impracticality of Pareto, law and economics generally relies on a different notion of allocative efficiency–a notion that treats actual payment decisions in the real world as good-enough reflections of how price and cost would line up in the perfectly competitive ideal.
According to this “Kaldor-Hicks criterion,” it’s okay to change production processes or legal rules in a way that somebody objects to so long as the benefits caused by the changes outweigh the harms. How do you know when the benefits outweigh the harms? By asking how much the people who benefit are willing to pay for those benefits and how much those who are harmed are willing to pay to avoid those harms. If all benefits are understood in terms of their monetary value to the people experiencing them, then we can think of all benefits and harms as effectively equivalent to receiving payment (economists refer to people being “indifferent” as between a monetary payment and the benefit or harm under consideration). If we assume that everybody values money the same way, we can compare benefits and harms across people while allowing them to make their own judgments about which things are beneficial to them. And if the “payoffs” to the beneficiaries of a given change are larger than the losses of those harmed by the change, then total wealth/welfare is increased. More “surplus”! (But a different type of surplus than we were just talking about.)
This “wealth maximization” principle does squarely confront the question of how to deal with conflicts and trade-offs, but it does so in a very unappealing way. And it doesn’t get us any closer to the everyday concept of “efficiency.”
Perhaps you have already spotted one core problem: it is impossible to separate out how much people are willing to pay from how much they are able to pay. If a law professor outbids a law student in an auction for tickets to the NBA Finals, should we therefore assume the law professor will enjoy the tickets more than the student? What if the law student is a lifelong Lakers fan and the professor thinks that basketball involves tackling and sliding into home plate? The trouble is that any attempt to identify how much of a manifest willingness to pay reflects desire and how much reflects cash on hand will require some method for determining how valuable something is to each person independently of how much money they are willing to pay for it, and once one has such a method, it’s not clear why one should look to willingness to pay at all. Why not just determine who should get the tickets based on how well they answer basketball trivia?
Often, economists will acknowledge this “distributional” problem with the Kaldor-Hicks criterion but claim that it can be dealt with via directly “distributional” solutions such as taxes and transfer payments. The notion is that redistributing money will surgically excise the wealth-flaunting aspects of willingness to pay while preserving payment as a way for people to express their preferences with precision in a common currency. Then we do not need to resort to interventions that “distort” the economic processes by setting priorities through some other means—money can be made neutral.
This is a sleight of hand. The amount of redistribution necessary to make it at all plausible that differences in willingness to pay do not reflect differences in ability to pay would need to be unprecedently massive (and international!) and it would need to happen all the time—constantly rebalancing money incomes to correct for accumulated advantages. Obviously, we are nowhere near such a system. Any focus on willingness to pay in this system is “distorted” relative to the purportedly ideal one of heavy taxation—so why focus on willingness to pay in the name of reducing “distortions”?
(Even if we were on the cusp of worldwide revolution, it is not at all obvious how to design a system of redistribution that attempts to make money “neutral” in this sense. Would everybody have to have equal amounts of money at all times? Wouldn’t that eliminate money’s role as an incentive to compete and innovate? Wouldn’t it also fail to recognize that some forms of living may be more expensive than others—both people in need of ongoing assistance with everyday tasks and people who have expensive tastes? Attempting to arrive at a balance between these issues inevitably requires moral and political reasoning about when to defer to willingness to pay as an expression of value, which necessarily contradicts treating it as the foundation of all value. It also undermines the concept of monetary neutrality.)
A more subtle problem is that attempting to reduce the value of everything to a price distorts our valuation processes. What, for instance, is the value of having loving elders during childhood? How would we even attempt to answer this question in terms of willingness to pay? Should we look at differences in earnings between adults who did have such elders as children and those who didn’t? Should we ask the amount of money that all government agencies spend on intervening in cases it adjudicates to be childhood neglect? You might think “of course not, those are idiotic and racist and insulting.” If so, I regret to inform you that these are the sorts of inquiries that administrative agencies have been forced to undertake when they attempt to do a fully priced “cost-benefit analysis” in the name of allocative efficiency.
Back to Basics
Sometimes, neoclassical economists respond to these sorts of critiques by saying something like: “I concede that other values are important, but you can’t throw out efficiency. Whatever goals we have for our provisioning system (equality, inclusion, ecological stewardship), we should always try to accomplish them in the most efficient way possible. Otherwise, we’re just wasting resources–and that’s not valuable to anybody.”
Do they have a point? No! In fact, this type of response is perhaps the most frustrating aspect of trying to make sense of “efficiency” as economists use the term. It relies on a slippage between allocative efficiency and the more intuitive notions we started with.
As we discussed, the sense in which we “reduce inefficiencies” when pursuing allocative efficiency is by moving toward a more “optimal” allocation of resources according to some criterion of optimality. If you are not convinced that either Pareto or Kaldor-Hicks describe optimal scenarios (as I have just argued you shouldn’t be), then you have no reason to care about what counts as “waste” according to their notions of efficiency, even as a halfway point. Allocative efficiency—in either sense—is an end goal (that may or may not be compatible with other end goals). It is a category error to think of it as a way to accomplish other goals.
If efficiency is to be useful as a way to identify methods for pursuing other goals more effectively, it must be a more modest notion. Sometimes, economists refer to “productive efficiency” as a way to capture this notion, but even there we have to be careful. One temptation is to think of productive efficiency as minimizing “costs,” which is usually understood in terms of money: the cheapest way to produce something is the most efficient. But, as we saw when exploring Kaldor-Hicks, we cannot be so confident about money’s ability to capture value. Focusing only on monetary cost will make it difficult to differentiate between changes in production methods that reduce the amount of labor required to produce a product and forcing workers to work longer hours for the same pay (or outsourcing to workers who will). And it will not help us in drawing lines between low cost production methods that use more physical resources (because they rely on old technology or exist in low-regulation jurisdictions) from those that use more at a higher cost.
If we want to take efficiency seriously as a virtue, we have to resist the economist’s temptation to generalize by means of price. We must pierce the monetary veil. That means being more precise and more pluralistic about what we want to economize on—about what counts as wasteful, why, and in which contexts.
We could, for instance, focus on technical efficiency—on producing more outputs with the same (or fewer) inputs. Such a notion would not count lengthening the workday as an efficiency increase (even if it increases output), but it would count technological improvements, changes in production techniques, and efforts to create worker morale that increase the amount of a good produced per worker per hour. We could break this concept into components: resource versus labor efficiency, operational versus productive efficiency, and so on.
Exploring the ins and outs of these concepts goes well beyond our purpose here. And tracing the relationship to law will not be straightforward. But it is worth noting that any of these more practical notions of efficiency that are not defined in terms of “welfare” should not be treated as unalloyed virtues. Most obviously: if a process produces harmful outputs, making it more efficient only facilitates harm (think of finding economies of scale in scamming people out of their money). More subtly: making a production process more efficient may well throw people out of work or cause a shift in production to a different location or lead to an excess amount of goods that we need to find new uses for. Accounting for all of these impacts will require not just careful empirical work but also moral deliberation about which things matter and whose interests to prioritize. There are no moral shortcuts to be found in asking what people would pay for, either in the real world or in some imaginary scenario of universal preference maximization.
***
This is a long and dense post, but its central message can be communicated simply: whenever you see or hear a claim that something is “efficient,” ask yourself “efficient in what sense?” and then “why should I care about that type of efficiency in this context?” (You don’t have to care about allocative efficiency of any sort in any context.) The term gets thrown around loosely in part because even sophisticated scholars do not ask these questions often enough. It is not an exaggeration to say that failure to do so has facilitated an enormous amount of harm and upward redistribution of wealth.
Further Reading
The best introduction I am aware of to the basic moral concepts at play here is Hausmann, McPherson and Satz. A more technical (but still accessible and non-mathy) explanation of welfare criteria and why law and economics cannot justify those it uses is Glick and Lozada.
For some discussion of how efficiency slippage works in antitrust and surrounding doctrinal areas, see Paul and Paul.
A great introduction to the problems with willingness to pay is Ackerman and Heinzerling. I also synthesized some critiques of even reformed cost-benefit analysis here at the Blog.
Dworkin, Kronman, and Coleman offer classic takedowns of Richard Posner’s articulation of “wealth maximization” as a value. For what it’s worth, here is Posner’s response.
Calabresi is a good explanation of why Pareto criteria can’t be useful to legal problems, with some consideration of the issues with Kaldor-Hicks.
Some classic explanations of issues with the assumptions of neoclassical welfare economics are Sen, Sen, and Sen.
Sanchirico is a detailed and somewhat technical explanation of why the problems with efficiency can’t be dealt with by “tax and transfer.” See also McCluskey.
A masterful explanation of some of the core conceptual problems with using money as a measure of value is Desan.
For a taste of some of the issues with the core assumptions about markets that ground the notion of “perfect competition,” see Lee and Keen and Ackerman (both are fairly technical, but it’s something of the nature of the beast.)
- Explaining why not goes beyond the scope of this post, but the basic point is that once you actually start to describe the sorts of assumptions you have to make to make a world of perfect competition actual, you realize you’re describing multiple physical impossibilities, as well as several conceptual confusions. Think, for instance, of what would be required for people to have those Buzzfeed-style lists guiding their actions. Or, to take a less intuitive example, in order for us to be sure that capital goods (that is to say, the things used in production processes) are always going to the production process that makes the highest and best use of them, we have to imagine that they can be constantly and costlessly repurposed. Machines would have to constantly disassemble themselves and ship their parts to the various factories at which they could potentially produce greater incomes, uniting with other parts to become new machines that can be used just until demand shifts again and a factory somewhere else becomes a higher bidder. ↩︎
