Getting Welfare Economics Right
Mises Review 8, No. 3 (Fall 2002)
THE ETHICS AND ECONOMICS OF MINIMALIST GOVERNMENT
Timothy P. Roth
Edward Elgar, 2002, viii + 134 pgs.
Professor Roth differs from most of his fellow economists. He finds the philosophical foundations of the standard model of welfare economics grossly deficient, and his book mounts a devastating criticism of the conventional view. James Buchanan claims that the book “exposes the contradictions [in welfare economics] that are present when the basic philosophical foundations are ignored” (quoted from the dust jacket), and he seems to me exactly right.
Defenders of the standard view might react with astonishment. Their view, they will say, makes no controversial philosophical assumptions at all. Quite the contrary, it rests on two fundamental theorems, the proofs of which admit of no doubt. According to the first of these, a fully competitive economy will in equilibrium result in a Pareto optimal state of affairs: in such a condition, no one can be made better off without making someone else worse off. The Pareto optimal condition thus achieved will usually be but one of many optimal states of affairs: it occupies only a single point on the “efficiency frontier.” The second fundamental theorem then tells us that the state, by suitably chosen taxes and subsidies, can achieve an “ethical equilibrium” without the surrender of Pareto optimality. Needless to say, the “ethical equilibrium” usually incorporates strongly egalitarian assumptions. In brief, a fully competitive economy can be both efficient and morally satisfactory.
Conventional welfare economists will no doubt admit that the choice of an ethical equilibrium demands resort to philosophy. But where are the philosophical assumptions that underlie the fundamental theorems themselves? Does not the welfare economist who advances these theorems act as a strict scientist? He simply tells people how to achieve their goals.
Roth maintains that the whole framework of the theorems rests on a mistaken philosophical doctrine. The welfare economist takes for granted the pursuit of efficiency. In doing so, he tacitly adopts a basic principle of utilitarianism: he aims to achieve the best consequences overall.
What is mistaken about this? Is it not self-evident that we want to achieve the best consequences? Roth by no means wishes to substitute pursuit of the mediocre, not to say the worst, for the utilitarian search for the best. But the consequentialist view, he thinks, ignores the basic principle of ethics: the dignity and moral worth of all human beings.
People are not mere tools whom we can manipulate at will to maximize efficiency. As moral beings, they have rights that cannot be violated. “Given that persons are morally equivalent, each is entitled to equal or impartial consideration. Each person has a duty to respect the intrinsic worth of others (and of himself)” (p. 5). On this view, which Roth derives from Kant, you cannot seek efficiency as a be-all and end-all. In particular, you cannot trample on people’s rights to achieve it.
A defender of the standard position might at this point acknowledge that there is more philosophy in welfare economics than he at first thought. But he will dissent from Roth’s main point: why need welfare economists reject rights and human dignity? On the contrary, a competitive economy rests upon a structure of property rights. Welfare economists do not oppose rights: they advocate them.
Roth finds this response unsatisfactory. Conventional welfare economics does not recognize the supreme importance of rights: “the problem is that, in the utilitarian framework, utility considerations trump rights” (p. 15). The structure of rights that the welfare economist favors counts as no more than a tool to achieve efficiency. If, in special cases, rights block an efficient outcome, they lapse.
Some defenders of the standard view agree, at least in part, that they need to give rights more attention. Accordingly, some welfare economists have developed Hobbesian social contract theories, in which rights result from bargaining among purely self-interested contractors. Roth finds these approaches inadequate. “The difficulty is that, given that goals and the means of achieving them are neither right nor wrong [in Hobbesian contract theories], and given that agents are narrowly self-interested, nothing in the theory implies a duty to obey the rules of justice” (p. 22).
Roth’s argument strikes me as promising but incomplete. He needs to show what specific rights follow from the recognition of human dignity. It does not suffice to say that each person has an equal right to participate in an original contract in the style of John Rawls. So long as Roth confines himself to invoking such a constitutional scheme, he leaves himself open to a counterthrust that may be fatal.
The welfare economist can argue that the rights that will be chosen in an adequate original position will be tailor-made to avoid conflicts with efficiency. John Harsanyi, a Nobel Laureate whose work Roth does not discuss, defended exactly this sort of view: he claimed that people in Rawls’s original position would choose to maximize average utility. To leave the scope of government up to an ideal constitutional convention, as Roth does, gets nowhere: how does Roth know that the convention will not choose Harsanyi’s plan, or one like it? Here I venture to suggest that attention to the works of Murray Rothbard would have greatly benefited our author.1
Roth’s assault on welfare economics has just begun. Following his mentor Buchanan, he raises another basic challenge. The sum and substance of welfare economics is the quest to maximize “social welfare”; but what exactly does this term designate? If one says that a state of affairs is “better” than an alternative, for whom is it better? Suppose that a Pareto improvement enables us to make one person better off, without ill effects on anyone else. Of course, the person “improved” has gained, but in what sense is society better off?
To assume that it is, Roth contends, is to treat society as if it were a person with values of its own. “The problem is that the utilitarian conception of the good ‘embodies a definition of “good” in application to the whole community of persons rather than to individual members.’ It is, in short, a supraindividual ‘good’ to which is attributed an independent existence and which, moreover, must be promoted” (p. 30, quoting Geoffrey Brennan and Buchanan). Our author informs us that Amartya Sen, one of the key figures of mainstream welfare economics, finds this objection “quite persuasive” in certain circumstances.
Here Roth has entered a vast philosophical debate. His argument rests on the premise that all value or good must be a value held by someone. When this premise is combined with the further premise that society is not an entity capable of holding values, Roth’s conclusion at once follows: the notion of a social good lacks sense. I think Roth ought to have recognized that his premise about values is controversial; to Plato, e.g., to speak of the good as such makes the highest sense, although it is not a value held by a person. Into these puzzles I resolutely refuse to enter here.
Defenders of the standard view will at this point be tempted to launch an attack. “Roth has raised some interesting philosophical issues,” they may say; “but what is that to us? Our models work: let us leave Roth to the metaphysicians.”
Even those with the Philistine views just portrayed cannot so easily turn Roth away. Our author meets the conventional position on its own terms, and raises a large number of technical criticisms against its use in practice. One of these I found especially ingenious: it uses one part of welfare economics to show the limitations of the rest.
The standard model assumes that all markets are perfectly competitive. In practice, this condition can only be approximated. What then follows for the attempt to generate Pareto improvements? According to the theorem of second-best, all bets are then off. “[I]f one accepts social welfare theory on its own terms, the general theorem of the second-best show[s] the futility of ‘piecemeal welfare economics’ “... policy prescriptions are, inevitably, ‘piecemeal’” (p. 44, quoting Lipsey and Lancaster, the developers of the theorem). In brief, the theorem shows that, in practice, welfare economics is useless. Incidentally, those like Roth who support a rights-based view can consistently support piecemeal changes. A violation of rights demands immediate correction: it is unnecessary to meet the difficult requirement that the correction be Pareto superior.
Roth has the singular virtue of being willing to kick his opponent when he is down. Even without considering the theorem of second-best, it is nonsensical to endeavor to use the standard model. How are we supposed to move along the “efficiency frontier” to reach “ethical equilibrium”? A benevolent government, fully apprised of all necessary information, has the task of making the necessary adjustments. Is it not, as Shakespeare says, a “very midsummer madness” to suppose the state capable of such well-intentioned behavior? Everyone acts to promote his own interest — except, of course, for the benevolent despots who rule over us.
Our author is not yet done. Not only is the standard model useless in practice: it is unsatisfactory as it stands. The model assumes that there is a determinate “efficiency frontier,” but the conditions under which this could be true are extremely unlikely to obtain. “Fundamental features of observable reality militate against the specification of the efficiency frontier ... the logic of the production possibility frontier requires, inter alia, that all producers, both within and across industries, employ the same productive service flows ... this analytically convenient assumption is irreconcilable with a technical environment characterized by cognitive limitations, information asymmetries and positive decision costs. ... If the production-theoretic foundations of the efficiency frontier may be called into question, the same is true of its utility-theoretic underpinnings” (p. 65).
Readers who want the technical details of these points will find them in abundance in Roth’s well-written book. (Like Keynes, though, he has an inordinate fondness for the Latin phrase “pari passu”.) At only one place do I think that he has fallen into outright error. In support of his contention that utilitarians cannot give rights their proper standing, he notes that some economists argue that corruption can be efficient. If so, do they not in some cases endorse rights violations?
This argument moves too fast. The economists in question may endorse attempts to evade the law. But in so doing, they need not abandon moral rights. To assume otherwise confuses the moral and legal orders. Only if the economists are willing to endorse corruption when moral rights are involved does Roth’s argument succeed. On the whole, though, Roth’s arguments are convincing — or at least provocative. This book is one of the best discussions of welfare economics since Murray Rothbard’s classic paper of 1956, “Toward a Reconstruction of Utility and Welfare Economics.”
- 1Roth would, I think, respond that the generality constraints that he thinks would govern laws under a proper constitutional order suffice to eliminate most governmental invasions of rights. But what about rights violations that come about through proper procedures? I do not think that Roth’s “Kantian-Rawlsian” procedures rule these out.
