Chapter II

The Role of Government in a Free Society

To the liberal, the appropriate means are free discussion and voluntary co-operation, which implies that any form of coercion is inappropriate. The ideal is unanimity among responsible individuals achieved on the basis of free and full discussion. This is another way of expressing the goal of freedom emphasized in the preceding chapter.

From this standpoint, the role of the market, as already noted, is that it permits unanimity without conformity; that it is a system of effectively proportional representation. On the other hand, the characteristic feature of action through explicitly political channels is that it tends to require or to enforce substantial conformity. The typical issue must be decided "yes" or "no"; at most, provision can be made for a fairly limited number of alternatives. Even the use of proportional representation in its explicitly political form does not alter this conclusion. The number of separate groups that can in fact be represented is narrowly limited, enormously so by comparison with the proportional representation of the market. More important, the fact that the final outcome generally must be a law applicable to all groups, rather than separate legislative enactments for each "party" represented, means that proportional representation in its political version, far from permitting unanimity without conformity, tends toward ineffectiveness and fragmentation. It thereby operates to destroy any consensus on which unanimity with conformity can rest.

There are clearly some matters with respect to which effective proportional representation is impossible. I cannot get the amount of national defense I want and you, a different amount. With respect to such indivisible matters we can discuss, and argue, and vote. But having decided, we must conform. It is precisely the existence of such indivisible matters -- protection of the individual and the nation from coercion are clearly the most basic -- that prevents exclusive reliance on individual action through the market. If we are to use some of our resources for such indivisible items, we must employ political channels to reconcile differences.

The use of political channels, while inevitable, tends to strain the social cohesion essential for a stable society. The strain is least if agreement for joint action need be reached only on a limited range of issues on which people in any event have common views. Every extension of the range of issues for which explicit agreement is sought strains further the delicate threads that hold society together. If it goes so far as to touch an issue on which men feel deeply yet differently, it may well disrupt the society. Fundamental differences in basic values can seldom if ever be resolved at the ballot box; ultimately they can only be decided, though not resolved, by conflict. The religious and civil wars of history are a bloody testament to this judgment.

The widespread use of the market reduces the strain on the social fabric by rendering conformity unnecessary with respect to any activities it encompasses. The wider the range of activities covered by the market, the fewer are the issues on which explicitly political decisions are required and hence on which it is necessary to achieve agreement. In turn, the fewer the issues on which agreement is necessary, the greater is the likelihood of getting agreement while maintaining a free society.

Unanimity is, of course, an ideal. In practice, we can afford neither the time nor the effort that would be required to achieve complete unanimity on every issue. We must perforce accept something less. We are thus led to accept majority rule in one form or another as an expedient. That majority rule is an expedient rather than itself a basic principle is clearly shown by the fact that our willingness to resort to majority rule, and the size of the majority we require, themselves depend on the seriousness of the issue involved. If the matter is of little moment and the minority has no strong feelings about being overruled, a bare plurality will suffice. On the other hand, if the minority feels strongly about the issue involved, even a bare majority will not do. Few of us would be willing to have issues of free speech, for example, decided by a bare majority. Our legal structure is full of such distinctions among kinds of issues that require different kinds of majorities. At the extreme are those issues embodied in the Constitution. These are the principles that are so important that we are willing to make minimal concessions to expediency. Something like essential consensus was achieved initially in accepting them, and we require something like essential consensus for a change in them.

The self-denying ordinance to refrain from majority rule on certain kinds of issues that is embodied in our Constitution and in similar written or unwritten constitutions elsewhere, and the specific provisions in these constitutions or their equivalents prohibiting coercion of individuals, are themselves to be regarded as reached by free discussion and as reflecting essential unanimity about means.

I turn now to consider more specifically, though still in very broad terms, what the areas are that cannot be handled through the market at all, or can be handled only at so great a cost that the use of political channels may be preferable.

It is important to distinguish the day-to-day activities of people from the general customary and legal framework within which these take place. The day-to-day activities are like the actions of the participants in a game when they are playing it; the framework, like the rules of the game they play. And just as a good game requires acceptance by the players both of the rules and of the umpire to interpret and enforce them, so a good society requires that its members agree on the general conditions that will govern relations among them, on some means of arbitrating different interpretations of these conditions, and on some device for enforcing compliance with the generally accepted rules. As in games, so also in society, most of the general conditions are the unintended outcome of custom, accepted unthinkingly. At most, we consider explicitly only minor modifications in them, though the cumulative effect of a series of minor modifications may be a drastic alteration in the character of the game or of the society. In both games and society also, no set of rules can prevail unless most participants most of the time conform to them without external sanctions; unless that is, there is a broad underlying social consensus. But we cannot rely on custom or on this consensus alone to interpret and to enforce the rules; we need an umpire. These then are the basic roles of government in a free society: to provide a means whereby we can modify the rules, to mediate differences among us on the meaning of the rules, and to enforce compliance with the rules on the part of those few who would otherwise not play the game.

The need for government in these respects arises because absolute freedom is impossible. However attractive anarchy may be as a philosophy, it is not feasible in a world of imperfect men. Men's freedoms can conflict, and when they do, one man's freedom must be limited to preserve another's -- as a Supreme Court Justice once put it, "My freedom to move my fist must be limited by the proximity of your chin."

The major problem in deciding the appropriate activities of government is how to resolve such conflicts among the freedoms of different individuals. In some cases, the answer is easy. There is little difficulty in attaining near unanimity to the proposition that one man's freedom to murder his neighbor must be sacrificed to preserve the freedom of the other man to live. In other cases, the answer is difficult. In the economic area, a major problem arises in respect of the conflict between freedom to combine and freedom to compete. What meaning is to be attributed to "free" as modifying "enterprise" ? In the United States, "free" has been understood to mean that anyone is free to set up an enterprise, which means that existing enterprises are not free to keep out competitors except by selling a better product at the same price or the same product at a lower price. In the continental tradition, on the other hand, the meaning has generally been that enterprises are free to do what they want, including the fixing of prices, division of markets, and the adoption of other techniques to keep out potential competitors. Perhaps the most difficult specific problem in this area arises with respect to combinations among laborers, where the problem of freedom to combine and freedom to compete is particularly acute.

A still more basic economic area in which the answer is both difficult and important is the definition of property rights. The notion of property, as it has developed over centuries and as it is embodied in our legal codes, has become so much a part of us that we tend to take it for granted, and fail to recognize the extent to which just what constitutes property and what rights the ownership of property confers are complex social creations rather than self-evident propositions. Does my having title to land, for example, and my freedom to use my property as I wish, permit me to deny to someone else the right to fly over my land in his airplane ? Or does his right to use his airplane take precedence? Or does this depend on how high he flies ? Or how much noise he makes? Does voluntary exchange require that he pay me for the privilege of flying over my land? Or that I must pay him to refrain from flying over it? The mere mention of royalties, copyrights, patents; shares of stock in corporations; riparian rights, and the like, may perhaps emphasize the role of generally accepted social rules in the very definition of property. It may suggest also that, in many cases, the existence of a well specified and generally accepted definition of property is far more important than just what the definition is.

Another economic area that raises particularly difficult problems is the monetary system. Government responsibility for the monetary system has long been recognized. It is explicitly provided for in the constitutional provision which gives Congress the power "to coin money, regulate the value thereof, and of foreign coin." There is probably no other area of economic activity with respect to which government action has been so uniformly accepted. This habitual and by now almost unthinking acceptance of governmental responsibility makes thorough understanding of the grounds for such responsibility all the more necessary, since it enhances the danger that the scope of government will spread from activities that are, to those that are not, appropriate in a free society, from providing a monetary framework to determining the allocation of resources among individuals. We shall discuss this problem in detail in chapter iii.

In summary, the organization of economic activity through voluntary exchange presumes that we have provided, through government, for the maintenance of law and order to prevent coercion of one individual by another, the enforcement of contracts voluntarily entered into, the definition of the meaning of property rights, the interpretation and enforcement of such rights, and the provision of a monetary framework.

The role of government just considered is to do something that the market cannot do for itself, namely, to determine, arbitrate, and enforce the rules of the game. We may also want to do through government some things that might conceivably be done through the market but that technical or similar conditions render it difficult to do in that way. These all reduce to cases in which strictly voluntary exchange is either exceedingly costly or practically impossible. There are two general classes of such cases: monopoly and similar market imperfections, and neighborhood effects.

Exchange is truly voluntary only when nearly equivalent alternatives exist. Monopoly implies the absence of alternatives and thereby inhibits effective freedom of exchange. In practice, monopoly frequently, if not generally, arises from government support or from collusive agreements among individuals. With respect to these, the problem is either to avoid governmental fostering of monopoly or to stimulate the effective enforcement of rules such as those embodied in our anti-trust laws. However, monopoly may also arise because it is technically efficient to have a single producer or enterprise. I venture to suggest that such cases are more limited than is supposed but they unquestionably do arise. A simple example is perhaps the provision of telephone services within a community. I shall refer to such cases as "technical" monopoly.

When technical conditions make a monopoly the natural outcome of competitive market forces, there are only three alternatives that seem available: private monopoly, public monopoly, or public regulation. All three are bad so we must choose among evils. Henry Simons, observing public regulation of monopoly in the United States, found the results so distasteful that he concluded public monopoly would be a lesser evil. Walter Eucken, a noted German liberal, observing public monopoly in German railroads, found the results so distasteful that he concluded public regulation would be a lesser evil. Having learned from both, I reluctantly conclude that, if tolerable, private monopoly may be the least of the evils.

If society were static so that the conditions which give rise to a technical monopoly were sure to remain, I would have little confidence in this solution. In a rapidly changing society, however, the conditions making for technical monopoly frequently change and I suspect that both public regulation and public monopoly are likely to be less responsive to such changes in conditions, to be less readily capable of elimination, than private monopoly.

Railroads in the United States are an excellent example. A large degree of monopoly in railroads was perhaps inevitable on technical grounds in the nineteenth century. This was the justification for the Interstate Commerce Commission. But conditions have changed. The emergence of road and air transport has reduced the monopoly element in railroads to negligible proportions. Yet we have not eliminated the ICC. On the contrary, the ICC, which started out as an agency to protect the public from exploitation by the railroads, has become an agency to protect railroads from competition by trucks and other means of transport, and more recently even to protect existing truck companies from competition by new entrants. Similarly, in England, when the railroads were nationalized, trucking was at first brought into the state monopoly. If railroads had never been subjected to regulation in the United States, it is nearly certain that by now transportation, including railroads, would be a highly competitive industry with little or no remaining monopoly elements.

The choice between the evils of private monopoly, public monopoly, and public regulation cannot, however, be made once and for all, independently of the factual circumstances. If the technical monopoly is of a service or commodity that is regarded as essential and if its monopoly power is sizable, even the short-run effects of private unregulated monopoly may not be tolerable, and either public regulation or ownership may be a lesser evil.

Technical monopoly may on occasion justify a de facto public monopoly. It cannot by itself justify a public monopoly achieved by making it illegal for anyone else to compete. For example, there is no way to justify our present public monopoly of the post office. It may be argued that the carrying of mail is a technical monopoly and that a government monopoly is the least of evils. Along these lines, one could perhaps justify a government post office but not the present law, which makes it illegal for anybody else to carry mail. If the delivery of mail is a technical monopoly, no one will be able to suceed in competition with the government. If it is not, there is no reason why the government should be engaged in it. The only way to find out is to leave other people free to enter.

The historical reason why we have a post office monopoly is because the Pony Express did such a good job of carrying the mail across the continent that, when the government introduced transcontinental service, it couldn't compete effectively and lost money. The result was a law making it illegal for anybody else to carry the mail. That is why the Adams Express Company is an investment trust today instead of an operating company. I conjecture that if entry into the mail-carrying business were open to all, there would be a large number of firms entering it and this archaic industry would become revolutionized in short order.

A second general class of cases in which strictly voluntary exchange is impossible arises when actions of individuals have effects on other individuals for which it is not feasible to charge or recompense them. This is the problem of "neighborhood effects". An obvious example is the pollution of a stream. The man who pollutes a stream is in effect forcing others to exchange good water for bad. These others might be willing to make the exchange at a price. But it is not feasible for them, acting individually, to avoid the exchange or to enforce appropriate compensation.

A less obvious example is the provision of highways. In this case, it is technically possible to identify and hence charge individuals for their use of the roads and so to have private operation. However, for general access roads, involving many points of entry and exit, the costs of collection would be extremely high if a charge were to be made for the specific services received by each individual, because of the necessity of establishing toll booths or the equivalent at all entrances. The gasoline tax is a much cheaper method of charging individuals roughly in proportion to their use of the roads. This method, however, is one in which the particular payment cannot be identified closely with the particular use. Hence, it is hardly feasible to have private enterprise provide the service and collect the charge without establishing extensive private monopoly.

These considerations do not apply to long-distance turnpikes with high density of traffic and limited access. For these, the costs of collection are small and in many cases are now being paid, and there are often numerous alternatives, so that there is no serious monopoly problem. Hence, there is every reason why these should be privately owned and operated. If so owned and operated, the enterprise running the highway should receive the gasoline taxes paid on account of travel on it.

Parks are an interesting example because they illustrate the difference between cases that can and cases that cannot be justified by neighborhood effects, and because almost everyone at first sight regards the conduct of National Parks as obviously a valid function of government. In fact, however, neighborhood effects may justify a city park; they do not justify a national park, like Yellowstone National Park or the Grand Canyon. What is the fundamental difference between the two ? For the city park, it is extremely difficult to identify the people who benefit from it and to charge them for the benefits which they receive. If there is a park in the middle of the city, the houses on all sides get the benefit of the open space, and people who walk through it or by it also benefit. To maintain toll collectors at the gates or to impose annual charges per window overlooking the park would be very expensive and difficult. The entrances to a national park like Yellowstone, on the other hand, are few; most of the people who come stay for a considerable period of time and it is perfectly feasible to set up toll gates and collect admission charges. This is indeed now done, though the charges do not cover the whole costs. If the public wants this kind of an activity enough to pay for it, private enterprises will have every incentive to provide such parks. And, of course, there are many private enterprises of this nature now in existence. I cannot myself conjure up any neighborhood effects or important monopoly effects that would justify governmental activity in this area.

Considerations like those I have treated under the heading of neighborhood effects have been used to rationalize almost every conceivable intervention. In many instances, however, this rationalization is special pleading rather than a legitimate application of the concept of neighborhood effects. Neighborhood effects cut both ways. They can be a reason for limiting the activities of government as well as for expanding them. Neighborhood effects impede voluntary exchange because it is difficult to identify the effects on third parties and to measure their magnitude; but this difficulty is present in governmental activity as well. It is hard to know when neighborhood effects are sufficiently large to justify particular costs in overcoming them and even harder to distribute the costs in an appropriate fashion. Consequently, when government engages in activities to overcome neighborhood effects, it will in part introduce an additional set of neighborhood effects by failing to charge or to compensate individuals properly. Whether the original or the new neighborhood effects are the more serious can only be judged by the facts of the individual case, and even then, only very approximately. Furthermore, the use of government to overcome neighborhood effects itself has an extremely important neighborhood effect which is unrelated to the particular occasion for government action. Every act of government intervention limits the area of individual freedom directly and threatens the preservation of freedom indirectly for reasons elaborated in the first chapter.

Our principles offer no hard and fast line how far it is appropriate to use government to accomplish jointly what it is difficult or impossible for us to accomplish separately through strictly voluntary exchange. In any particular case of proposed intervention, we must make up a balance sheet, listing separately the advantages and disadvantages. Our principles tell us what items to put on the one side and what items on the other and they give us some basis for attaching importance to the different items. In particular, we shall always want to enter on the liability side of any proposed government intervention, its neighborhood effect in threatening freedom, and give this effect considerable weight. Just how much weight to give to it, as to other items, depends upon the circumstances. If, for example, existing government intervention is minor, we shall attach a smaller weight to the negative effects of additional government intervention. This is an important reason why many earlier liberals, like Henry Simons, writing at a time when government was small by today's standards, were willing to have government undertake activities that today's liberals would not accept now that government has become so overgrown.

Freedom is a tenable objective only for responsible individuals. We do not believe in freedom for madmen or children. The necessity of drawing a line between responsible individuals and others is inescapable, yet it means that there is an essential ambiguity in our ultimate objective of freedom. Paternalism is inescapable for those whom we designate as not responsible.

The clearest case, perhaps, is that of madmen. We are willing neither to permit them freedom nor to shoot them. It would be nice if we could rely on voluntary activities of individuals to house and care for the madmen. But I think we cannot rule out the possibility that such charitable activities will be inadequate, if only because of the neighborhood effect involved in the fact that I benefit if another man contributes to the care of the insane. For this reason, we may be willing to arrange for their care through government.

Children offer a more difficult case. The ultimate operative unit in our society is the family, not the individual. Yet the acceptance of the family as the unit rests in considerable part on expediency rather than principle. We believe that parents are generally best able to protect their children and to provide for their development into responsible individuals for whom freedom is appropriate. But we do not believe in the freedom of parents to do what they will with other people. The children are responsible individuals in embryo, and a believer in freedom believes in protecting their ultimate rights.

To put this in a different and what may seem a more callous way, children are at one and the same time consumer goods and potentially responsible members of society. The freedom of individuals to use their economic resources as they want includes the freedom to use them to have children -- to buy, as it were, the services of children as a particular form of consumption. But once this choice is exercised, the children have a value in and of themselves and have a freedom of their own that is not simply an extension of the freedom of the parents.

The paternalistic ground for governmental activity is in many ways the most troublesome to a liberal; for it involves the acceptance of a principle -- that some shall decide for others -- which he finds objectionable in most applications and which he rightly regards as a hallmark of his chief intellectual opponents, the proponents of collectivism in one or another of its guises, whether it be communism, socialism, or a welfare state. Yet there is no use pretending that problems are simpler than in fact they are. There is no avoiding the need for some measure of paternalism. As Dicey wrote in 1914 about an act for the protection of mental defectives, "The Mental Deficiency Act is the first step along a path on which no sane man can decline to enter, but which, if too far pursued, will bring statesmen across difficulties hard to meet without considerable interference with individual liberty."¹ There is no formula that can tell us where to stop. We must rely on our fallible judgment and, having reached a judgment, on our ability to persuade our fellow men that it is a correct judgment, or their ability to persuade us to modify our views. We must put our faith, here as elsewhere, in a consensus reached by imperfect and biased men through free discussion and trial and error.

A government which maintained law and order, defined property rights, served as a means whereby we could modify property rights and other rules of the economic game, adjudicated disputes about the interpretation of the rules, enforced contracts, promoted competition, provided a monetary framework, engaged in activities to counter technical monopolies and to overcome neighborhood effects widely regarded as sufficiently important to justify government intervention, and which supplemented private charity and the private family in protecting the irresponsible, whether madman or child -- such a government would clearly have important functions to perform. The consistent liberal is not an anarchist.

Yet it is also true that such a government would have clearly limited functions and would refrain from a host of activities that are now undertaken by federal and state governments in the United States, and their counterparts in other Western countries. Succeeding chapters will deal in some detail with some of these activities, and a few have been discussed above, but it may help to give a sense of proportion about the role that a liberal would assign government simply to list, in closing this chapter, some activities currently undertaken by government in the U.S., that cannot, so far as I can see, validly be justified in terms of the principles outlined above:

1. Parity price support programs for agriculture.
2. Tariffs on imports or restrictions on exports, such as current oil import quotas, sugar quotas, etc.
3. Governmental control of output, such as through the farm program, or through prorationing of oil as is done by the Texas Railroad Commission.
4. Rent control, such as is still practiced in New York, or more general price and wage controls such as were imposed during and just after World War II!
5. Legal minimum wage rates, or legal maximum prices, such as the legal maximum of zero on the rate of interest that can be paid on demand deposits by commercial banks, or the legally fixed maximum rates that can be paid on savings and time deposits.
6. Detailed regulation of industries, such as the regulation of transportation by the Interstate Commerce Commission. This had some justification on technical monopoly grounds when initially introduced for railroads; it has none now for any means of transport. Another example is detailed regulation of banking.
7. A similar example, but one which deserves special mention because of its implicit censorship and violation of free speech, is the control of radio and television by the Federal Communications Commission.
8. Present social security programs, especially the old-age and retirement programs compelling people in effect (a) to spend a specified fraction of their income on the purchase of retirement annuity, (b) to buy the annuity from a publicly operated enterprise.
9. Licensure provisions in various cities and states which restrict particular enterprises or occupations or professions to people who have a license, where the license is more than a receipt for a tax which anyone who wishes to enter the activity may pay.
10. So-called "public-housing" and the host of other subsidy programs directed at fostering residential construction such as F.H.A. and V.A. guarantee of mortgage, and the like.
11. Conscription to man the military services~in" peacetime. The appropriate free market arrangement is volunteer military forces; which is to say, hiring men to serve. There is no justification for not paying whatever price is necessary to attract the required number of men. Present arrangements are inequitable and arbitrary, seriously interfere with the freedom of young men to shape their lives, and probably are even more costly than the market alternative. (Universal military training to provide a reserve for war time is a different problem and may be justified on liberal grounds.)
12. National parks, as noted above.
13. The legal prohibition on the carrying of mail for profit.
14. Publicly owned and operated toll roads, as noted above.

This list is far from comprehensive.

1 A. V. Dicey, Lectures on the Relation between Law and Public Opinion in England during the Nineteenth Century (2d. ed.; London: Macmillan & Co., 1914), p. li.