Is Competition Good

Review of Industrial Organization 19: 37–48, 2001.  © 2001 Kluwer Academic Publishers. Printed in the Netherlands. 37 Is Competition Such a Good Thing? Static Ef? ciency versus Dynamic Ef? ciency MARK BLAUG University of Amsterdam, Amsterdam, The Netherlands Abstract. This paper addresses the rationale for antitrust legislation. It is a striking fact that the legitimacy of antitrust law has been taken for granted in the United States ever since the Sherman Act of 1890 and, until the advent of the so-called Chicago School, it was even taken for granted by conservative American economists. Europeans, on the other hand, have always been lukewarm about legal action against trusts and cartels and this attitude is found right across the political spectrum in most European countries. Nevertheless, in both the U. S. A. and Europe, the ultimate justi? cation for antitrust law derives from economic doctrine regarding the bene? cial effects of competition. But what exactly are these bene? cial effects and how secure is the contention of economists that competition is always superior to monopoly? Surprisingly enough, competition, that central concept of economics, is widely misunderstood by many economists, both as a market phenomenon and as an organizing principle of economic reasoning. I. A Little History of Thought I begin by drawing what I believe is a fundamental distinction in the history of economics, as far back as Adam Smith or even William Petty, between two different notions of what is meant by competition, namely, competition as an end-state of rest in the rivalry between buyers and sellers and competition as a process of rivalry that may or may not terminate in an end-state. In the end-state conception of equilibrium, the focus of attention is on the nature of the equilibrium state in which the contest between transacting agents is ? nally resolved; if there is recognition of change at all, it is change in the sense of a new stationary equilibrium of endogenous variables in response to an altered set of exogenous variables; but comparative statics is still an end-state conception of economics. However, in the process conception of competition, what is in the foreground of analysis is not the existence of equilibrium, but rather the stability of that equilibrium state. How do markets adjust when one equilibrium is displaced by another and at what speed will these markets converge to a new equilibrium? But, surely, all theories of competition do both; existence and stability are tied up together and to study one is to study the other? By no means, however; it is easy to show that, for centuries, competition to economists meant an active process of jockeying for advantage, tending towards, but never actually culminating in, an 38 MARK BLAUG equilibrium end-state. Only in 1838, in Cournot’s Mathematical Principles of the Theory of Wealth was the process conception of competition totally displaced by the end-state conception of market-clearing equilibria. At ? rst this did not succeed in wiping the slate entirely clean of an interest in competitive processes but in the decade of the 1930s – those years of high theory as George Shackle called them – the Monopolistic Competition Revolution and the Hicks-Samuelson rehabilitation of Walrasian general equilibrium theory, forti? d by the New Welfare Economies, succeeded in enthroning the end-state conception of competition and enthroning it so decisively that the process view of competition was virtually buried out of sight. Let me elaborate. It is a striking feature of the language of The Wealth of Nations that the term â€Å"competition† invariably appears with a de? nite or inde? nite article preceding it: â€Å"a competition between capitals†; â€Å"the competi tion with private traders†, and so forth. For Smith, competition is not a state or situation, as it is for Cournot and for us, but a behavioural activity; it is a race – the original sense of the verb â€Å"to compete† – between two or more individuals to dispose of excess supply or to obtain goods available in limited quantities. What we nowadays call competition or the market mechanism was for him â€Å"the obvious and simple system of natural liberty†, meaning no more than an absence of restraints or ree entry into industries and occupations. Neither competition nor monopoly was a matter of the number of sellers in a market; monopoly did not mean a single seller but a situation of less than perfect factor mobility and hence inelastic supply; and the opposite of competition, was not monopoly, but co-operation. Producers in The Wealth of Nations treat price as a variable in accordance with the buoyancy of their sales, much like enterprises in modern theories of imperfect competition. This was not a conception invented by Smith because by 1776, competition had long been analyzed by a whole series of eighteenth century authors as a process which brings temporary â€Å"market† prices into line with cost-covering natural prices, those â€Å"natural† prices were indeed â€Å"the central price, to which the prices of all commodities are continually gravitating†, and in saying that Smith invoked Newtonian language to dignify a conception of price-determination that had a long tradition going back to the seventeenth century. To obtain that end-state in which market prices equal natural prices and the rate of pro? is equalized between industries, there had to be a considerable number of rivals, possessing common knowledge of market opportunities; they had to be free to enter and exit different lines of investment; but that was all and even that much was never spelled out explicitly as necessary prerequisites for competition – only once did Smit h ever mention the number of rival ? rms involved in competition. It was Cournot who ? rst had the notion of sellers facing a horizontal demand curve when their numbers become so large that none can in? uence the price of their own product. Competition, which once meant the way in which ? rms take account of how their rivals respond to their actions, now meant little more than the slope of the average revenue curve depriving ? rms in the limit of any power to make the price. Thus was born, decades before the Marginal Revolution of the 1870s what IS COMPETITION SUCH A GOOD THING? 39 one writer has wittily called â€Å"the quantity theory of competition† (quoted in Blaug, 1997, p. 68). Edgeworth’s Mathematical Psychics (1981) followed Cournot in providing all the trappings of the modern de? nition of perfect ompetition in terms of a large number of sellers, a homogeneous product, perfect mobility of resources and perfect knowledge on the part of buyers and sellers of all alternative opportunities. However, Marshall’s treatment of the competition always carefully labelled as â€Å"free competition† was much closer to Smith’s â€Å"simple system of natural liberty† than to that of C ournot and Edgeworth’s perfect competition. Even Walras hesitated to follow Cournot to the letter. Indeed, it was not until the 1920’s that the modern textbook concept of perfect competition was ? ally received into the corpus of mainstream economics, largely due to the impact of Knight’s classic, Risk, Uncertainty and Pro? t (1921). But it is doubtful whether the idea was in fact fully accepted in 1921 and a good case can be made for the thesis that it was Robinson and Chamberlain a decade later who hammered down the theory of perfect competition in the very process of inventing imperfect and monopolistic competition theory (Machovec, 1995). The replacement of the process conception of competition by an end-state conception, which was ? alized in 1933 or thereabouts, drained the idea of competition of all behavioural content, so that even price competition, the very kernel, of the competitive process for Adam Smith, David Ricardo and John Stuart Mill now had to be analysed as â€Å"imperfect† competition, a sort of deviation from the norm. Indeed, every act of competition on the part of a businessman was now taken as evidence of some degree of monopoly power, and hence a departure from the ideal of perfect competition, and yet pure monopoly ruled out competitive behaviour as much as did perfect competition. II. Perfect Competition, the Unattainable Ideal All I have said so far merely reiterates what Schumpeter said in 1942 and Hayek repeated in 1949: â€Å"perfect competition is not only impossible but inferior, and has no title to being set up as a model of ideal ef? ciency†; â€Å"what the theory of perfect competition discusses has little claim to be called ’competition’ at all and its conclusions are of little use as guides to policy† (quoted in Blaug, 1997, p. 69). But this message, delivered over a half-century ago, fell on deaf ears and the endstate theory of perfect competition is more ? mly in the saddle today than it ever was in the 1940s when Hayek and Schumpeter, not to mention John Maurice Clark (1949, 1961), were writing. And why? The answer is simple: it is that most of us were taught that although perfect competition is rarely if ever attained, nearly-perfect competition is said to be observable in some markets (agricultural markets being a favour ite example) and these approximations to the state of perfect competition somehow replicate many 40 MARK BLAUG f the desirable characteristics of perfect competition; in a word, second-best is so nearly ? rst-best that we may indeed employ ? rst-best as a standard. Open any textbook and what do we ? nd? The concept of perfect competition is said to be like the assumption of a perfect vacuum in physics; descriptively inaccurate, to be sure, but nevertheless productive of valid insights about actual economies. Thus, Samuelson and Nordhaus (1992, p. 295) in the 14th edition of their Economics concede that a perfect and absolutely ef? ient competitive mechanism has never existed and never will â€Å"but the oil crisis of the 1970s† is only one of their many examples of how an empirically empty competitive model can nevertheless produce the right answers to a concrete imperfectly competitive situation (for other textbook treatments, see Blaug, 1997, pp. 69–70). This is prec isely what Reder (1982, p. 12), called the notion of â€Å"tight prior equilibrium†, which he thought was characteristic of the Chicago School of Economics: â€Å"one may treat observed prices and quantities as good approximations to their long-run equilibrium values†. Call this the good-approximation assumption. Unfortunately, the idea of a near or far approximation to perfect competition has absolutely no logical meaning. We seem conveniently to have forgotten the famous Lipsey–Lancaster (1996) second-best theorem published in 1956, according to which we are either at a ? rst-best optimum or it matters not whether we are at second-best or tenth-best because we cannot rigorously demonstrate that doing away with a tax or a tariff that put us at tenth-best will bring us closer to ? st-best in a welfare sense of these terms. This theorem has not been conveniently forgotten; it has been deliberately forgotten because it wreaks havoc with the end-state, ? rst-best conception of competition. Must we therefore cease to give advice on competition policy? I think not; but what it does mean is that instead of gnostic pronouncements about the desirability of any move in the direction of ? st-best perfect competition, we must engage instead in qualitat ive judgements about piecemeal improvements, embracing a dynamic process-conception of competition, which is precisely the old classical conception that Schumpeter, Hayek, Clark and modern neo-Austrians have urged us to adopt. To grasp why it was necessary to revive this tradition, we must spend a moment explaining why modern price theory is so strong on the nature of the competitive equilibrium end-state and so weak on the process by which competition drives a market towards a ? al equilibrium. III. The Awful Legacy of General Equilibrium Theory When Walras literally invented general equilibrium (GE) in 1871, he was just as much concerned with the process-conception of competition known as â€Å"the stability problem† as in what we have called the end-state interpretation of equilibrium known as â€Å"the existence problem† – is simultaneous multimarket-equilibrium possible in a capitalist economy? But gradually, in successive editions of his Elements of Pure Economics, the existence problem came ever more to the fore, while the sta- IS COMPETITION SUCH A GOOD THING? 41 bility problem receded in the background (Walker, 1996). Even so, Walras’s view of how markets adjust in disequilibrium was always somewhat naive. It is a story which we all learn in our ? rst course of economics: in response to the appearance of excess demand and supply, prices adjust automatically as independently acting buyers and sellers â€Å"grope† their way to a ? al equilibrium. When this tatonnement story is well told, it sounds utterly convincing and at such times we are apt to forget that many markets, particularly labour markets and â€Å"customer markets†, react faster in terms of quantities than in terms of prices (as Marshall always insisted in opposition to Walras) and sometimes only in terms of quantities (see Blaug, 1997, pp. 71–75). But prices and quantities aside, what about product ifferentiation and competition by maintenance and service agreements, what about Schumpeterian competition in terms of new products and processes, new methods of marketing, new organizational forms and new reward structures for employees? In short, all the forms of rivalry between producers which Chamberlain and Robinson have taught us to call monopolistic or imperfect competition (the irony of calling what cannot exist, perfect competition, and what always exists, imperfect competition, never ceases to amuse me! . Walras struggled manfully to provide a rigorous solution to the existence problem but never got much beyond counting equations and unknowns to ensure that there were enough demand and supply equations to solve for the unknown equilibrium prices and quantities in the economy. As for the stability problem, he solved that after much hesitation by simply eliminating disequilibrium transactions as â€Å"false trading† (another wonderfully ironic piece of rhetoric). Although he never mentioned the concept of a ? tional auctioneer announcing different prices until an equilibrium price is discovered, whereupon trade is allowed to take place – this is one of those historical myths that subsequent generations have invented – it is dif? cult to avoid the conclusion that he simply gave up the effort to provide a convincing account of how real-world competitive markets achieve GE. Such an account has in fact never been provided even to this date. In 1954, Arrow and Debreu ? nally solved the existence problem by modern mathematical techniques – topological properties of convexity, ? ed point theorems, Nash equilibria, etcetera – of which Walras could never have dreamt but, in so doing, they travelled even further than Walras had from anything smacking of descriptive accuracy: there are forward markets in their GE model for all goods and services in the economy, including all locations and conceivable contingen t states in which these goods and services might be consumed, and yet no one holds cash to deal with the likelihood that income and expenditure may fail to synchronize. They were perfectly candid about this failure to describe actual economies. Indeed, they made a virtue of the purely formal properties of their model. 1 1 As Debreu (1959, p. x) expressed it in his Theory of Value: â€Å"The theory of value is treated here with the standards of rigor of the contemporary formalist school of mathematics . . . . Allegiance to rigor dictates the axiomatic form of the analysis where the theory, in the strict sense, is logically entirely disconnected from its interpretation†. And yet this book claimed to be a work in economics! 42 MARK BLAUG They cracked the existence problem, not to mention the uniqueness problem – is there one unique vector of prices at which GE exists? but they never tackled the stability problem. In other words, after a century or more of endless re? nements of the central core of GE theory, an exercise which has engaged some of the best brains in twentieth-century economics, the theory is unable to shed any light on how market equilibrium is actually attained, not just in a real-world decentrali zed market economy but even in the toy economies beloved of GE theorists. We may conclude that GE theory as such is a cul de sac: it has no empirical content and never will have empirical content. Moreover, even regarded as a research program in social mathematics, it must be condemned as an almost total failure. That is not to say that highly aggregated computable GE models, such as IS-LM, are pointless or that a GE formulation of an economic problem, emphasizing the interdependence of all sectors of the economy, may not prove illuminating but simply that Walrasian GE theory – the notion that the existence of multi-market equilibrium may be studied in a way that is analogous to solving a set of simultaneous equations – has proved in the fullness of time to be an utterly sterile innovation. The real paradox is that the existence, uniqueness and stability of GE should ever have been considered an interesting question for economists to answer: a complete satisfactory proof of all three aspects of the problem would no doubt have been a considerable intellectual feat in logic but would not in any way have enhanced our understanding of how actual economic systems work. IV. The Welfare Implications of GE Of course, Walras hoped to show, not just that GE is possible, but that it is good. But here too he never got much beyond the idea that voluntary exchange between two parties improves both of their welfares – otherwise, why would they have traded? What is true of bilaterial exchange will also be true of competitive exchange between a large number of traders if individual producers cannot themselves set prices, so that all consumers face identical prices for identical homogeneous commodities. This is precisely where the notion of perfect competition as an end-state of rest comes into welfare economics grounded in GE theory. Pareto, who was a much better technician than Walras, carried on where Walras left off. He too was convinced that GE is good for everyone but as a follower of Ernest Mach in philosophy, he hated such metaphysical ideas as maximising happiness, utility, welfare, or call it what you will, and he strenuously objected to interpersonal comparisons of utility (ICU) on the grounds that such comparison could not be operationalised. Pondering these issues, he realised that the one circumstance that avoids ICU is a social state which meets with unanimous approval or at least with the absence of con? ict in which one person is only made better off at the expense of another person. In other words, we want a state which is so ef? cient that there is no surplus, no waste, no slack, â€Å"no such thing as a free lunch†. But is not perfect competition just such a state? Of course, it may leave some people rich IS COMPETITION SUCH A GOOD THING? 3 and some people poor but that will be the consequence of the fact that we started with unequal endowments of the individuals in our economy – some people are born clever and some people have rich parents – but, given those endowments that are not themselves explained by GE theory – no theory ever explains everything – the GE model will grind out the rental prices of all the services of land, labor and capital as well as the prices of all goods , produced with those services. Once we have somehow arrived at the end-state of perfectly competitive equilibrium, it will be impossible to make one person better off without making another person worse off except by interfering with the initial endowments of agents. In this way, Pareto thought that he had ? nally found an admittedly narrow de? nition of the bene? cial effects of competition that was totally free of that positivist bugbear, ICU. The idea, only later called â€Å"Pareto optimality†, fell into oblivion as soon as it was announced but was rescued along with Walrasian GE theory in the 1930s by John Hicks and Nicholas Kaldor. They extended the scope of Pareto optimality by arguing that any economic change, whether from a position of competitive equilibrium or not, was welfare improving if the gains to bene? ciaries of that change were large enough to enable them at least in principle, to bribe the losers voluntarily to accept the change. The existences of such potential Pareto improvement (PPI), as they are nowadays called, still involves no ICU because it is grounded on the voluntariness of market exchange. In short, Hicks and Kaldor (with a prodding from Lionel Robbins) stayed true to the Paretian conception of how an economist should study welfare economics. At ? rst glance, the Hick–Kaldor compensation test does seem virtually to pull a rabbit out of a hat but further re? ection soon showed that the achievement was semantic, not substantive. Why is it a potential and not an actual PI? The moment we try to implement PPI by encouraging gainers and losers to negotiate a bribe, they will engage in strategic bargaining and even without fancy game theory, it is easy to see that they may never reach an agreement. If the change has political signi? cance, the state may then intervene to force the parties to agree – in which case we have said goodbye to our taboo on ICU. No matter how we slice it, in the end we cannot avoid (1) a qualitative judgement from on high of the size of the PPI – remember that there is no objective way short of voluntary trade to measure the magnitude of a gain or a loss to the parties concerned – and (2) an interpersonal comparison of that gain and loss to the respective parties. But all that brings us back to Marshall and Pigou whose Economics of Welfare (1921) had none of Pareto’s compunctions about ICU and was perfectly content to declare that a pound sterling taken from a rich man by a progressive income tax hurt him less than the pleasure it gave the poor man when it was handed over to him. We have not quite reached the end of the story. The Arrow–Debreu proof of the existence of GE in 1954 was almost contemporary with Arrow’s proof of what he labelled the First and Second Fundamental Theorems of welfare economics. The ? st theorem demonstrates that every competitive equilibrium in a decentralized economy is Pareto-optimal, which we have already discussed, and the second 44 MARK BLAUG theorem demonstrates that a Pareto-optimum can always be achieved via perfect competition if lump-sum taxes and transfers are feasible, so that whatever were the original endowments of agents, we can still make everyone better off with a perfectly compe titive economy. Immense pains are taken in every textbook of microeconomics to persuade readers of the validity of those two theorems. And they are valid – as mathematical exercises. Lump-sum taxes and transfers are changes which do not affect economic behaviour and even the most ingenious modern welfare economists have never been able to come up with a convincing example of such things. 2 I think that we may safely conclude that the First and Second Fundamental Theorems of welfare economics are just mental exercises without the slightest possibility of ever being practically relevant. They are what Ronald Coase (1988) called â€Å"blackboard economics†, an economics that is easy to write on a blackboard in a classroom but that bears no resemblance to the world outside the classroom. V. Why Is Competition Good? I contend that perfect competition is a grossly misleading concept whose only real value is to generate examination questions for students of economics. 3 It is misleading because it breeds the view that economics is a subject like Euclidean geometry, whose conclusion may be rigorously deduced from fundamental axioms of behaviour plus some hard facts about technology. But of course this does not imply that competition is bad. I, along with most economists, believe that competition is good. But if perfect competition is impossible, and Pareto-optimality almost impossible, what is the basis of this belief in the desirability of competition? It is based on a concept of dynamic ef? ciency, the outcome of competitive processes, and not the static ef? ciency of Walras, Pareto and the First and Second Fundamental Theorems of welfare economics. The schizophrenia of economists on this issue is simply extraordinary. The manin-the-street favours capitalism because it is ultimately responsive to consumers’ demands, technologically dynamic and produces the goods that are wanted at low cost; of course, it also suffers from periodic slumps, more or less chronic unemployment even in booms, and frequently generates a highly-unequal distribution 2 They would have to be randomly assigned to individuals or else to re? ect some personal noneconomic characteristic, such as more consonants than vowels in one’s last name. It used to be thought that a uniform poll tax was a perfect example of a limp-sum tax but as Mrs. Thatcher discovered it had a most profound effect on economic behaviour: almost a million people disappeared from the electoral roll in Britain because the poll tax could not be collected without a home address. 3 I concede reluctantly that it has its uses for purposes of answering comparative statics questions on taxes and subsidies but even these have much less practical signi? cance than is usually assumed (see Vickers, 1995). IS COMPETITION SUCH A GOOD THING? 5 of income. 4 Still, on balance the good outweighs the bad and without becoming Panglossian, he or she votes for capitalism – and so do virtually all economists. But is this what we teach in our textbooks? To ask the question is to already answer it. Can one actually teach the principles of dynamic ef? ciency? Of course, one can and that is what we do in every course in industrial organization (and in every course in man agement schools), where, alas, we have to undo the brainwashing that students have undergone in their courses on microeconomics. In so doing, we employ historical comparisons and case studies, and these can only cultivate the ability to make informed judgements about speci? c attempts at what Popper called â€Å"piecemeal social engineering†, making the world a little better here and there, because we do not know enough to make the whole world best once and for all. VI. Some Conclusions: Coase and Posner Beliefs in the ef? cacy of antitrust law ? ts neatly into the concept of dynamic ef? ciency, or what Clark called â€Å"workable competition†. A question like: should we break up Microsoft or just reprimand and perhaps ? e the company? does not lend itself to a precise answer by the edicts of economists and it is just as well that it does not. Empirical science frequently proceeds on the untidy basis of what is plausible rather than what can be formally demonstrated beyond any doubt. The structureconduct-performance paradigm of yesteryear, associated with names of Edward Mason and Joe Bain, did j ust that but that has since been superseded by game theory and transaction cost on the one hand and the Chicago School of Richard Posner and Robert Bork on the other hand. In between we ? d Ronald Coase and the widely misunderstood Coase Theorem as the very centre piece of the law and economics movement. Since this so-called inappropriately named theorem picks up a number of the themes in welfare economics that we have discussed above, let us close with a brief discussion of it. As stated by its inventor, George Stigler (1966, p. 113), the Coase Theorem is the proposition that â€Å"under perfect competition private and social costs will be equal† and hence â€Å"the composition of output will not be affected by the manner in which the law assigns liability for damage†. This combines two claims in one, the ? rst of which will be familiar to us: (1) an ef? ciency claim that perfect competition is always optimal if voluntary bargaining between the affected parties to their mutual advantage is possible at zero transaction costs, de? ned as the costs of making deals, negotiating contracts, and policing the enforcement of those contracts (Allen, 2000), and (2) an invariance claim that the ? nal allocation of resources is invariant to different initial assignments of property rights provided these are in fact clearly de? ed. A voluminous literature has shown that both propositions are either highly contentious or else a tautology if perfect competition, perfect information and zero 4 In an instructive essay, Richard Nelson (1981 reiterates my charge of schizophrenia and adds to my list of the bene? ts of a private enterprise system of capitalism that of â€Å"administrative parsimony†, an echo of Hayek’s discussion of the merits of competiti ve prices as information signals. 46 MARK BLAUG transaction costs are rigorously de? ned (Medema and Zorbe, 2000). Lo and behold, however, Coase has argued ever more vehemently that transaction costs can be reduced by appropriate judicial decisions but that they can never be reduced to zero even under Cournot-type perfect competition. Of course, if we de? ne perfect information as literally foreseeing every alternative opportunity under all possible contingencies, now and in the future, it follows immediately that we can write and enforce contracts at zero costs (zero in ? nancial outlays, in time and even in cognitive effort), in which case only increasing returns to scale will prevent us achieving perfect competition. Once transaction costs are zero and competition is perfect, it follows immediately that the distribution of property rights cannot matter. In short, the Coase Theorem is just a logical corollary of perfect competition and perfect information but that does little to persuade us that it is much more than a logical theorem. 5 As for the more controversial invariance claim, income and wealth effects in consumption patterns and the strategic behaviour of the injured and injuring parties as they enter into voluntary bargaining (the old objection to Hicks–Kaldor compensation payments) will certainly make the ? al allocation of resources sensitive to the way in which the law of the moment assigns liability for damage. Are we really to believe that my claim against the American Tobacco Company for giving me lung cancer will be decided in 2002 in exactly the same way it would have been decided in 1940? Coase (1964, p. 105) said it all 35 years ago: Contemplation of an optimal system may provide techniques of analysis that would otherwise have been missed and, in certain special cases, it may go far to providing a solution. But in general its in? uence has been pernicious. It has directed economists’ attention away from the main question, which is how alternative arrangements will actually work in practice. It has led economists to derive conclusions for economic policy from a study of an abstract of a market situation. Richard Posner, in his in? uential textbook, Economic Analysis of Law (1998), now in its ? fth edition, subsumes Pareto optimality and the Coase Theorem in an ef? ciency logic of â€Å"wealth maximization†. He claims not only that common law, statute law and judge-made law should serve to maximize wealth, so that for example entitlements in property law should be shifted to the more productive litigants as evidenced by their willingness to pay, but that legal entitlements and hence resources actually tend to gravitate towards their most valuable use if voluntary exchange is permitted. Without saying so, Posner clearly believes that we can 5 Moreover, as Allen (2000, pp. 904–905) argues quite rightly, the famous Modigliani-Miller Theorem of corporate nance – if capital markets are perfect, the value of a ? rm is invariant to its debt-equity ratio – and the Ricardo Equivalence Theorem of government ? nance – if capital markets are perfect, the level of household wealth is invariant to the ratio of taxes to the size of the public debt – are both special cases of the Coase Theorem because all taxes, debt obligations and equity shares are simply delineation s of property rights; in a world of zero transaction costs, both ? rms and governments could decide on debt levels by tossing a coin. IS COMPETITION SUCH A GOOD THING? 47 isolate PPI, divorcing ef? ciency from equity without committing ourselves to ICU, in short, he believes in classic or rather neoclassical Paretian welfare economics. Although he deals at length with distributional issues arising from liability rules and various forms of taxation, he never lays down any general principles about income redistribution, such as, for example, Pigou did: any transfer of income from the rich to the poor that does not diminish national income was deemed desirable by Pigou. What he argues, when criticized, is simply that users of distributive justice will have to be addressed outside the framework of standard economic analysis (Parisi, 2000). But this is exactly what Pareto, Kaldor and Hicks said years ago. Orthodox welfare economics, including the â€Å"ef? ciency of the common law hypothesis† upheld by Posner, has simply stood still ever since the 1930s. This notion of a neat divorce of ef? ciency from equity, of an objective value-free de? nition of ef? iency, has haunted economics from its outset but it is, of course, a will-o’-the-wisp: there is in fact a different ef? ciency outcome for every different distribution of income, and vice versa. Ef? ciency is necessarily a value-laden term and welfare economics is necessarily normative, that is, a matter of good or bad and not true or false. 6 However, there is real merit in treating ef? ciency and equity questions lexicographically, so that we can be as explicit as possible about our di stributional judgements, but that is not because we can ever decisively separate them. My complaint about Posner is that he evades all these fundamental questions in applied welfare economics. Not only does he fail to tell us how to add equity to ef? ciency but he does not even tell us whether ef? ciency means static ef? ciency or dynamic ef? ciency. There is an almost deliberate fuzziness of language in all his writings, which smacks of ideology rather than science. If we are going to employ the economist’s language of ef? ciency, we ought to be told just how to apply it and why ef? ciency should be our standard for judging the consequences of the law. One of Clark’s old rules of â€Å"workable competition†, such that entry into industries should be kept as free as is technically feasible taking due account of sunk costs, if necessary by antitrust legislation, is more relevant for public policy than Posner’s continual appeal to the principle of wealth maximization. The Chicago school does not deny that there is a case for antitrust law but they doubt that it is a strong case because most markets, even in the presence of high concentration ratios, are â€Å"contestable† (Bork, 1978). How do we know? We know because the good-approximation assumption: the economy is never far away from its perfectly competitive equilibrium growth path! Believe it or not, that is all there is to the â€Å"antitrust revolution† of the Chicago School. 6 Some economists believe, extraordinarily enough, that welfare economics is positive and not evaluative at all (see Hennipman, 1992; Blaug, 1992, chap. 8, 1993). 48 References MARK BLAUG Allen, Douglas W. (2000) ‘Transaction Costs’, in Bouckaert and De Geest, eds. , pp. 893–926. Blaug, Mark (1992) The Methodology of Economics, 2nd edn. Cambridge: Cambridge University Press. Blaug, Mark (1993) ‘Pieter Hennipman on Paretian Welfare Economics: A Comment’, De Economist, 141, 127–129. Blaug, Mark (1997) ‘Competition as an End-State and Competition as a Process’, in Not Only an Economist. Recent Essays. Cheltenham: Edward Elgar, pp. 66–86. Bork, Robert H. (1978) The Antitrust Paradox: A Policy at War with Itself. New York: Basic Books. Bouckaert, Boudewijn, and Gerrit De Geest (2000) Encyclopaedia of Law and Economics, 3 Vols. Cheltenham: Edward Elgar. Clark, John Maurice (1961) Competition as a Dynamic Process. Washington, DC: Brookings Institution. Coase, Ronald G. (1964) ‘The Regulated Industries: Discussion’, American Economic Review, 54, 194–197. Coase, Ronald G. (1988) The Firm, the Market and the Law. Chicago: University of Chicago Press. Debreu, Gerard (1959) Theory of Value. An Axiomatic Analysis of Economic Equilibrium. New Haven: Yale University Press. Hennipman, Pieter (1992), ‘Mark Blaug on the Nature of Paretian Welfare Economics’, De Economist, 140, 413–445. Lipsey, Richard C. , and Kelvin Lancaster (1996) ‘The General Theory of Second Best’, Review of Economic Studies, 24, 1956, pp. 11–32, reprinted in Richard C. Lipsey, Microeconomics, Growth and Political Economy. Selected Essays, Vol. 1. Cheltenham: Edward Elgar, pp. 153–180. Machovec, Frank (1995) Perfect Competition and the Transformation of Economics. London: Routledge. Medema, Steven G. , and Richard O. Zerbe (2000), ‘The Coase Theorem’, in Bouckaert and De Geest, eds. , pp. 36–92. Nelson, Richard R. (1981) ‘Assessing Private Enterprise: An Exegesis of a Tangled Doctrine’, Bell Journal of Economics, 12, 93–100, in Peter Boetke, eds. , The Legacy of Friedrich von Hayek, Vol. III. Cheltenham: Edward Elgar, pp. 80–98. Parisi, Francesco, ed. 2000) The Economic Structure of the Law: The Collected Essays of Richard A. Posner, Vol. I. Cheltenham: Edward Elgar. Reder, Melvin W. (1982) ‘Chicago Economics: Permanence and Change’, Journal of Economic Literature, 20, 1–38. Stigler, George J. (1966) Theory of Price, 3rd edn. New York: Macmillan. Van Cayseele, Patrick, and Rog er Van den Bergh (2000) ‘Antitrust Law’, in Bouckhaert and De Geest, eds. , Vol. III, pp. 467–498. Vickers, John (1995) ‘Concepts of Competition’, Oxford Economic Papers, 47, 1–23. Walker, Donald A. (1996) Walras’s Market Models. Cambridge: Cambridge University Press.