According to the rhetoric of the Reagan administration much of the machinery of government should be dismantled and its functions, ideally, be limited to those intended by the Founding Fathers to serve the needs of the simple, agricultural community. Public opinion is becoming increasingly skeptical of this view and most professional economists, except for a small but very vocal group of “libertarians,” disagree with it. Even some of Reagan’s own economists claim that our immensely complex and highly vulnerable modern economy can be managed only through skillful manipulation of several powerful instruments controlled by the federal government—tax rates, budget deficits and surpluses, and the interest rates and monetary supply controlled by the Federal Reserve. I doubt it.

The different varieties of economic policies we hear about and try to practice are usually justified by theoretical constructions such as Phillips Curves, Laffer Curves, full-employment budgets, “rational expectation” theorems, and similar abstract notions. The builders of more and more intricate econometric models try in vain to compensate for their lack of hard, systematically organized factual information by relying on increasingly ingenious but utterly unreliable methods of indirect statistical inference.

While debate over these theories continues, the economy steadily deteriorates. A year ago when President Reagan issued his first budget I observed that the proposed combination of drastic tax cuts with unprecedented tightening of credit could very likely bring about a slump threatening to lead into a deep depression.1 The explosive rise in productive investment so confidently predicted a year ago failed to materialize notwithstanding all the tax concessions passed last summer.

Would refining the old theoretical schemes, or, as some propose, redefining the money supply to be controlled by the Federal Reserve, remedy the failure of recent economic policies? Can one seriously believe that a significant contribution to the solution of our economic problems will be made by efforts to increase the reliability of econometric forecasts through tinkering with already highly sophisticated statistical procedures, or by making marginal improvements in the accuracy of aggregate price and other indices that would identify the day and hour when recession ends and recovery begins? I doubt this can be done and so, I believe, does much of the informed public.

The inadequacy of the trial and error approach that still dominates our national economic policies will, I believe, become clear when we have to face the full consequences of the rising wave of new technology. The interdependence of the different sectors of a national economy is bound to grow with the increase in their internal complexity and the scale of operations. Many projects started today—whether for producing energy or for more advanced computers—will come to fruition only in ten, fifteen, or even twenty years. Their ultimate success will depend on whether they are coordinated with other developments taking place not only in our economy but in other parts of the world. Corporate decision making in this country, while effective so far as it goes, often suffers badly because it cannot take a long-run point of view. American corporate executives tend to lose their jobs if profits sag during three successive quarters.

The contrast in this respect with Japan and some Western countries is striking. During the past ten years American corporations earned an average of 18 percent on their investments while their Japanese counterparts earned only 11 percent. This means that in the US corporate managers are so cautious that they refuse to move until they can count on recovering newly invested capital in four and a half years. Managers in Japan are prepared to wait for eleven years. No wonder they continue to improve old plants and construct new ones while large US corporations often prefer to maintain liquidity and to diversify their investments by buying up each other’s stocks.

Much is said by conservatives about the need to compensate investors for the high risks they take, but not enough is said about reducing risks. One, and possibly the only, effective way of reducing risk—besides taking out insurance that simply redistributes risk without diminishing its level—is systematic coordination of investment based on long-range foresight. That is, planning. All too many corporate leaders bridle when they hear this word, although rational planning requires neither government ownership of industry, nor direct control of corporations, nor abandoning much of the play of the market.

The planning approach is less simplistic than the hit-or-miss approach; it is, however, more ambitious and because of that also more difficult. Under a planning approach national economic policies are not formulated as independent measures designed to solve a particular problem, whether it concerns environmental depredation, energy shortage, inflation, unemployment, or urban blight. They are conceived as coordinated actions intended to make the entire economy function more effectively and grow along a carefully projected path. Such policies cannot be derived from general theories, but must be based on the detailed observation and mapping of the actual state of economic life. Both the interdependence of the different parts of the economy and alternative paths along which they might advance in the coming years must be systematically explored. The picture is bound, of course, to become less and less detailed as different projections run from five to ten, from ten to twenty, and from twenty to, say, fifty years. Whether one path or another is followed is not a professional question but a political decision that must be reached by democratic processes.


The effects of any major policy cannot be assessed without also taking account of the mostly unintended indirect effects of other policies applying to other problems. Consider, for instance, a policy by which both government agencies and private corporations want to promote the export of West Virginia coal. Special credits or tax breaks may be provided to do this (as they are for producing oil). Such measures will likely affect the demand for mining equipment produced, say, in Minnesota and also encourage the construction of new, specialized port facilities in Delaware. This in turn will require stepped-up output of cranes and other equipment as well as of cement, steel, and other construction materials. Some of the electric power stations now using West Virginia coal may well shift to natural gas that has to be piped from Texas, leading to an increase in the price of electric power used, say, by textile mills in nearby Tennessee. But that is not all: pollution is usually a byproduct of productive activities and increased coal exports are bound to affect efforts to maintain the quality of water near the steel and cement plants, and to protect the fishing grounds lying close to the shore of Delaware.

Thus it is not surprising that actions intended to solve one particular problem create new problems elsewhere. If policy makers act in ignorance of such indirect interrelationships, measures taken by one government or corporate office will tend to cancel out the effects of actions taken by another. The costs of these measures will increase without their contributing to the solution of either one or another problem. As things stand, one group of policy makers often becomes aware that it works at cross purposes with another group; but neither one possesses sufficient knowledge about the combined effect of each others’ policies. A jurisdictional dispute then takes place; experts from both sides find themselves engaged in the costly process called “adversary fact finding.” The apparently interminable fight over Westway in New York is a case in point.

The method called “input-output” analysis was developed some fifty years ago to deal with such problems involving relations between different sectors of the economy. The “input-output” tables are designed to show in great detail the flow of goods and services between various producing and consuming sectors of the economy. In our example above, the tables would show how the steel, cement, coal, and other industries would each be affected by the West Virginia project. Originally, such analysis was intended as a tool for tracing the effects of technological change on national income and in particular on the levels of production and employment in each of the many sectors of the economy, including that in which the original technical change took place. The same approach proved useful in determining the direct and indirect effects of many other changes such as, for example, an increase or a reduction in military spending, shortages of some raw material, or increases in the price of a particular import, say crude oil.

The power and the applications of input-output analysis as a means of establishing economic facts have much increased. Large multinationals such as the International Telephone and Telegraph Company use it to plan corporate policy. At the time of the Marshall Plan, France was the first advanced Western country to use input-output methods to plan the postwar reconstruction of its economy. The Japanese government uses them to plot national economic growth in close cooperation with organized business leadership, Hungary has employed them to guide the difficult transition from a rigid, centrally planned economy to the present system which tries to combine encouragement of private initiative with planning national development. Norway uses them more extensively than any other country.

The aversion to such planning in Washington has been curious and marked. The Labor and Commerce Departments and the Pentagon made use of input-output computations during World War II and in the early postwar years. Suddenly in the early 1950s, the use of input-output methods was prohibited; even compiling official input-output tables was temporarily discontinued. This was the period when the head of General Motors testified to Congress, “What is good for General Motors is good for the country.” The availability of detailed input-output information would, it was thought, be conducive to national economic planning and, in effect, bad for General Motors (whose executive vice president, Roger Keyes, as soon as he became deputy secretary of defense cut off the funds for compiling US input-output tables). I remember a young staff member of the Council of Economic Advisers confiding to me that he and his colleagues secretly made input-output computations in order to be able to answer questions posed to them by Arthur Burns, the council’s cautious chairman.


Now the Departments of Labor, Commerce, and Energy, and the Pentagon, are again using routine input-output computations; so does the Economics Division of the Congressional Research Service in the Library of Congress. In Texas important decisions on allocation of scarce water resources among different users are made with the help of a regional input-output table.

But what remains particularly self-defeating in the United States is the inability—or occasionally the reluctance—of official statistical organizations to collect the readily available specialized information that is needed not only for input-output analysis but for other projects as well. The US is the only advanced, industrialized country that still does not possess an effective central statistical office responsible for systematically obtaining information about population, natural resources, technology, and other aspects of the national economy and society. The Census Bureau hardly qualifies as a central statistical office. As things now stand, each department and each agency of the federal government, and of most local governments, compiles data it happens to need for its own immediate purposes. Confronted with what amounts to a huge statistical jigsaw puzzle, economists and statisticians working in the government or private business, as well as those in the universities, spend a large part of their time trying to put its pieces together. They must try to reconcile incompatible figures coming from different sources and to fill in as best they can the gaping holes in the overall picture.

What a contrast with the statistical organization of Japan or even that of a small country like Norway, which has decided recently to discontinue its periodic census because all the information needed for government and business planning is collected and brought up to date, month by month, and year by year. In the US the task of compiling a decennial US input-output table is assigned to a small team tucked away in one of the many bureaus of the Department of Commerce. The most recent Japanese input-output table is produced by the combined efforts of thirteen ministries under the general supervision of a committee of the Council of Ministers. The five hardcover folio volumes containing the Japanese table are several times as large as their US counterpart. The table was compiled much faster, and is indispensable to the Japanese government and businesses in deciding, among other things, which industries have good prospects for growth and which do not.

Creating and maintaining a comprehensive supply of data would permit a drastic reduction in the amount of guesswork and idle theorizing that goes into our policy making now. But providing the requisite data is not enough. The time has come to take a decisive step by setting up a strong, autonomous research organization that would be analogous to the Congressional Research Service, but much larger; its task would be to provide all agencies of government with the information needed to work out a systematic, coordinated approach to the main problems of national and local economic policy.

This organization would also be responsible for monitoring in great detail what is happening to the different parts of the US economy—the decline and growth of different industries, services, and regions, for example, and how they affect one another. It should be able to anticipate potential trouble spots, the parts of the economy where, to name only a few examples, energy shortages, technological unemployment, population movements, or sudden needs for longterm credits may arise. This agency should not make grand predictions but should elaborate different scenarios, each describing likely effects of any particular combination of national, regional, and local economic policies. This would, in fact, be the only means by which the government and the electorate would be enabled to make informed choices among differing policies.

While providing information to legislators and administrators responsible for national economic policies, and outlining appropriate methods to carry them out, such a technical organization should have no more control over final decisions than, say, the Bureau of Labor Statistics in the Department of Labor, or the Bureau of Economic Analysis in the Department of Commerce. But this independent agency should have a decisive voice in determining the direction and scope of federal data-gathering activities and should advise state and local governments on their own data collection.

If we had such an agency to gather facts and project different directions for the economy and their consequences, this would sharply reduce one of the most wasteful and futile aspects of the present policy-making process—what I have called “adversary fact finding.” Studying the supposedly factual reports from interested parties on such matters as energy policy or the future of the airlines, I cannot help being reminded of conflicting witnesses to an auto accident arguing before a judge. The policy-making process would be much more effective if it did not imitate a traffic court but rather were modeled along the lines of the arbitration of labor disputes. The arbitrator first establishes the relevant facts and only then explores alternatives for a workable agreement.

Of all the problems we will face in the coming years, the inefficiency of the present policy-making process seems to me the greatest obstacle of all. David Stockman’s instructive interview in the Atlantic Monthly testified to this. So long as that obstacle is not removed, most of the difficulties we are facing now will stay with us.

Inflation is a case in point. Following Mrs. Thatcher’s lead, the administration has, with considerable success, tried to suppress inflation by beating the entire economy into the ground. An old joke talks of a gypsy who eked out a meager living by renting out the services of a horse he owned. One day he decided to increase the profitability of this enterprise by training the old nag gradually, step by step, to get by on smaller and smaller rations of oats. For a couple of weeks this seemed to work until, to the poor fellow’s great surprise, the horse suddenly died.

To judge by past experience, what reason do we have to expect that after inflation has stopped rising, at the cost of prolonged and severe depression, the old Keynesian or monetarist policies will not again be put in effect and prices will not again begin to rise as the familiar cycle recommences? I do not believe the inflation that has long been plaguing the US economy can be dealt with effectively by the familiar combination of fiscal and monetary policies alone. Even the most skillful manipulation of taxes and interest rates will not work, in my view, unless it is accompanied by institutionalized day-by-day cooperation—not simply “mutual understanding”—between business, labor, and government.

Austria, a highly industrialized modern democracy, has successfully resisted inflationary pressure by setting up an institution for such cooperation. During the annual negotiations between trade unions and employers’ organizations, the government serves as an impartial fact-finder by providing detailed input-output information on the actual state of the national economy, and a systematic analysis of the effects that a proposed settlement will have on its future growth, as well as on the prospects for particular industries.

A specific example: The Institute for Socio-Economic Studies, a branch of the Austrian Academy of Sciences, carried out at the government’s request a detailed analysis of how new data-processing technology would affect the newspapers and related industries. In the US and UK the unions have resisted such computerized substitutes for old-fashioned printing. In Austria, the report provided the basis for an agreement by the printers, the editors, and the “union of art, media, and self-employed professions” to follow a step-by-step plan to introduce the new technology. This meant phasing out jobs, setting up training programs, etc.; but the hardships imposed on the employees, on the employers, and, one might add, on consumers, have been minimized and the overall effect has not been not inflationary. The negotiations were stiff, but there were no strikes and no lockouts. What was crucial was that the negotiators had before them a careful factual study whose authority they could accept.

Encouraged by that experience, the Austrian government has directed the National Academy of Sciences to analyze the potential effects on the economy and society of the electronic revolution. The academy has worked out scenarios for different economic and social policies that would permit the country to reap the benefits of the new technology while mitigating its disruptive effects.2

In the US, linking wage restraints with price reductions, as in the recent negotiations between the UAW and GM, seems to point in the right direction. However, what is needed is a general agreement between organized labor and organized business that would apply to major sectors of the economy and would be based on a carefully designed, comprehensive, and voluntary plan. Separate wage contracts made by the UAW without such an agreement and such a plan can contribute little to the solution of national inflationary problems. In fact Japanese imports, rather than agreement with the unions, will do more than the union contract to set the prices for GM’s cars.

The situation in the automobile industry brings us to what is probably the greatest challenge facing not only our economy but also the economies of other advanced countries: the effects of automation.

Something quite new is happening. All previous technological revolutions—except perhaps the one that caused the so-called enclosure movement in sixteenth-century England—enhanced the commanding role of labor as the dominant, indispensable factor of production. Compared to the demand for capital and natural resources, the demand for labor was steadily maintained and thus secured reasonably full employment at steadily increasing real wages.

The growth of total output was accompanied by rising per capita consumption and, up until the middle 1940s, a progressive shortening of the normal working day, week, and year. At the end of World War II, the situation changed. Successive waves of technological innovation continued to overtake one another, and the real-wage rate continued to go up; but the length of the normal work week today is practically the same as it was thirty-five years ago.

Machinery, however, continues to replace human labor, and as it does so some sectors of the economy, and some types of labor, are much more affected than others. Less skilled workers usually go first; skilled workers, later. And even if a drastic general wage cut could temporarily arrest the adoption of laborsaving technology, the trend is bound to resume unless Luddite barriers are set up against introduction of the new technology. Even a principled libertarian might hesitate to have the wage level settled by cutthroat competition among workers using continued pressure from steadily improving labor-saving machinery.

In the long run a reasonable response to the incipient technological unemployment caused by the spread of automation should aim at an equitable distribution of gradually shrinking employment opportunities on the one hand and, on the other, of the gradually increasing national product. This product will continue to rise, however, only if technological advance is not obstructed directly or indirectly. This means that employment policies allowing workers to be displaced by technology will have to be combined with income policies protecting those who are being displaced.

The income policies I have in mind would have to supplement the declining real income that many blue- and white-collar workers receive from the sale of their services on the labor market. So also would they have to protect the modest incomes of some independent craftsmen, professionals, and self-employed small entrepreneurs. In effect, we have willy-nilly installed such income policies by gradually changing the tax system, and increasing Social Security, medical insurance, welfare payments, and unemployment benefits. Instead of being hastily curtailed, these systems will have to be redesigned and expanded so as to reduce the contrast between those who are fully employed, partially employed, retired, or simply out of work.

This prospect of an increasingly automated society shows all the more forcefully the need to provide the foundation of factual analysis and economic projection that would make democratic national planning possible. Our political economy will continue to flail blindly unless we can uncover its interacting empirical realities and consider in what general directions it should move.

This Issue

August 12, 1982