With persistent energy, Leontief reminds us that this old problem will not go away; he continues to search actively for socially effective solutions. The following is a lightly edited transcript of a talk that he gave at the CV Starr Center for Applied Economics at New York University in the spring of 1996 along with the ensuing discussion. The session was introduced and moderated by William Baumol, and informal discussants included Ralph Baultjens, Edna Erlich, Robert Heilbroner and Robert Schwartz. We welcome the opportunity to share Leontief's insights on this important topic and to appreciate once again his enduring talents for clear analysis and expression.
Leontief: Among the many factors that have promoted economic change, I believe that technology or, rather, change in technology is the most prominent. I realize that it is dangerous to look for 'ultimate causes' in a world where everything seems to depend on everything else. But I believe that for the most part the economy, and ultimately the society, must adapt to the conditions that technology creates. If it cannot adjust to the challenges of changing technology, it fails.
To speak of the 'advance' of technology would involve a questionable value judgement. Rather, I try to speak of technology in as objective a sense as I can. A good way for me to describe technology is that it is like a recipe in a cookbook; it says what you need to produce something. It tells us not only the amounts of various ingredients but also, for example, the temperatures and cooking time.
A long time ago the development of technology was very slow. It took thousands of years for early technological innovations to be incorporated into agriculture. But the tempo changed abruptly after the steam engine was introduced. Over the last 200 years, technological development accelerated at least in Western countries. The invention of the steam engine in the mid-nineteenth century led to a real explosion! Looking into the future, the concrete effects of technological change on the economy and society will depend on other factors, such as population growth, availability of natural resources and so on. Nevertheless, for all plausible environmental and demographic scenarios, it seems likely that future technologies will tend to reduce the role of human labor in the process of production.
Many different approaches are used in analyzing change in technology. To some economists, mine may seem unorthodox. I find it somewhat naive to describe the productive process in terms of continuous production functions with the major focus on marginal productivity. This approach poses empirical difficulties because technological change is in reality anything but continuous. Instead, I am inclined to use the same conceptual framework, i.e. the same factual information, as used by the people who actually operate the technology. Particularly in our society, the profit motive leads to one set of technology replacing another, not necessarily in a continuous fashion. Change consists of introducing successive new technologies, which produces a mix of technologies at any given time and a succession of technology mixes over time.
Beginning with the second half of the nineteenth century, science began to play a very great role in technological change and now, of course, science and technology really work in tandem. Technological development now very quickly incorporates and really uses the devices, the information and the recommendations that scientists advance.
Economists and policy-makers characterize technological change in terms of increasing productivity of labor, but that is a somewhat ambiguous notion. If labor inputs are completely eliminated, labor productivity becomes infinite. Furthermore, output per unit of input has been increasing for many other types of input, such as electric power and machinery. To think of technical change in terms of labor productivity is a kind of hold-over from earlier times when labor was clearly the principal factor of production.
I think that labor is fast losing its position as a principal factor of production and this, in turn, will require or result in significant change in the social system. (I know that this is controversial and that is the reason that I look forward to open discussion.) Of course, for some time now we have faced problems of income distribution. The production process automatically determines the incomes of participants under the present system. In my view, a fundamental distinction should be made between income received as payment for labor and income dependent on ownership.
There are many types of income dependent on ownership. We often lump them into the general category of capital income. As technology calls for less and less labor, capital goods (in the most general sense) will have a growing role. The transition is not likely to be a smooth or linear process. Natural resources will have an important influence. With the discovery of oil, one type of natural resource gained enomous importance. Today, even the exploration for natural resource deposits is heavily dependent on developing technology. For that matter, what we usually call consumption is also affected by technology. In our time these matters are obvious.
Computerization is only the most recent stage of a process whereby the influence of scientific development and technological change has been growing. At every step along the way, the role of human labor was reduced and that of capital increased, which, of course, affected the distribution of income. Now what will be the consequences for society?
And now I come to the most controversial part of my observation: I see no limit to scientific development and I think this means that there is no limit to technological change. Technology is not likely to reach some particular stage and then stop. This means that the role of labor in society will continue to diminish. It has already diminished significantly.
Very often we accept the idea that uneducated labor will diminish in importance but assume that educated labor will not. Obviously, automation is one of the processes that will lead to the use of less educated labor but automation and now computerization are reducing the need for educated labor too. Years ago, Faye Duchin and I studied how automation affected management and we wrote a book called The Future Impact of Automation on Workers. We found that, in the last 50 years, the importance of middle management shrank because middle management was interpreting figures and conveying information to top managers who made the decisions. Now this function is performed with much less labor because of sophisticated computers and information systems. Thus the popular idea that if you have an education you will have a job has a shaky basis. Education is required to help labor adjust to changing technological requirements, but less and less educated labor is needed as processes are automated and computerized.
Now what will be the long-run effect: the effect in 20, 50 or 100 years? I think in the long run we will greatly reduce the role of labor and the relative importance of ownership of capital, and also of natural resources, will increase. While capital income will continue to grow, the price of labor will be pressed down. I cannot resist the temptation to point out that these trends are already apparent. Real wages in this country [the US] have not only ceased to increase but have begun to decrease. The wages of unskilled labor are already too low to provide minimum living standards.
Society now responds to this problem through income transfers: free education, free healthcare and income to old people, and so on, but this is just the beginning. Ultimately, this distribution problem will become very serious and we must be very careful. I think the best solution is to make sure that, starting now, and to a growing extent in the future, working people have some capital as well as wage income. Since social security is mandatory, it could be a vehicle for inducing wage-earners to save a certain part of their income which would be invested in capital with the income distributed as social security payments. This, I think, will be the future.
If everyone were now deriving income from capital and labor in the same proportions, shifting in the direction of greater reliance on capital would be very easy. In practice, capital holdings are far from proportional to labor income and the process of transition will not be easy at all. The fact that incomes of a large fraction of the population are being squeezed will become a serious problem. It is likely to generate considerable social tension. Ultimately the structure of our free competitive market economy must face this very difficult challenge: How can we effect a substantial increase in the proportion of workers' ownership income in a system where the dominant objective is cost reduction through the substitution of new for older technologies? How will the economic system accommodate these imperatives?
Of course, there are also questions of how evolving technology will affect developing countries. Here I can only make one observation. It is quite clear that less developed countries do not follow the same path of technological change that developed countries do. Some technology can just 'float over' from the sophisticated industrial to the developing world. Even in the tropical parts of Brazil, trees are cut down not by traditional methods but by extremely powerful machinery which replaces the very low wage labor alternative. Thus it would appear that our problem might arise in developing countries as well.
In summary, I remind you that science very quickly leads to change in technology and that change in technology is the driving force of development; it will continue to be so for a long time. At this point I hope that I have provoked you sufficiently to ensure a response. Please raise your objections: I'm not sure I'm right but I would like very much to see.
Discussant: I am very much in accord with Wassily's arguments and I'd like to put a question to him. It strikes me that new technology generally affects discrete groups: in the process of creative destruction, some industries are rising while others are going down. But I am struck by one particular feature of the bunching of the characteristic form of technological change in our time, which is automation/ computerization. This new form of technology appears in and affects a wide range of industries at once, which bears on Wassily's expectation.
In the past, looking back 100 years or so, every time massive technological change took place in railroads, automobiles or jet planes, there was an escape mechanism for individuals displaced from their jobs because of new technology. The escape mechanism was to enter a highly labor-intensive field called the 'service sector'. When the percentage of employment in agriculture plummeted and later on when the percentage of employment in manufacturing came to a ceiling and then began to decline, there was this constantly growing 'economic sponge' called the service sector which was characterized by low productivity of labor. What strikes me as peculiar and ominous about the present wave of technology is that it invades the service sector. It affects service jobs of all kinds, which previous technologies did not.
This suggests to me that, in the medium run of 25 years or so, there is only one adjustment mechanism at hand, and that would be to make a change in the conventional number of hours of work per week. In 1890, 60 hours per week was the norm; it wasn't until the 1930s that our present 40-hour week became established. I suspect that if and when the service industry adopts displacing technology we will have to shift to a 30-hour week. Along with that we'll need other kinds of changes such as an increase in the age of entry into the labor force and legislation that provides some sorts of sabbaticals for skills enhancement and mandates a lower age of retirement. In other words, it seems to me that government will have to be the intervening mechanism to adjust the supply of labor to the demand by changing the entry mechanism.
Leontief: Good observation: you do not disagree with me! You suggest possible adjustments to the fact that human labor loses its importance in the production of goods and even of services, if you take computerization into account. I agree completely. However, your solution would require a very different social organization. Right now, in the United States, there is a move toward the elimination, or at least the drastic reduction, of government, which would be a disaster. In your proposal, the government must be very strong because I doubt that people will reduce their working time, particularly if they are well paid. Of course, a reduction in the pay of labor will bring social tensions. In the ideal situation--and this illustrates the problem rather than giving the answer--everybody would derive 50% of his income from supplying labor services and 50% from income on capital, directly or indirectly. In the case of social security there could be a lovely smooth transition. I think that, possibly, this would be an easier institutional change than restriction of labor hours.
At the end of the nineteenth century the working week in the United States was 70 hours. Reducing it was a spontaneous and natural change because families and individuals decided that they wanted a larger part of their 'income' as leisure. For a while that mechanism works but finally I think it will stop working. After a while, if per capita income increases, people would choose more commodities rather than more leisure. This produces distributional problems. Because technological change has no limits, the problem will be how to change the structure of a society so as to provide a reasonable standard of living to all parts of the population. How can we do this? I would really like to raise this general question without giving a definite answer.
If the real wages of labor in this country began to fall very fast I would anticipate very serious social problems. On the other hand, if total income increases because of technological change then there is no problem. But the rate of economic growth depends on many things--pollution and so on. To try to compensate for the reduced demand for labor due to technological change by manipulating the overall growth rate is not realistic.
Discussant: Of course, Wassily has been discussing these concerns for many, many years now. What makes this period and present concerns so very different from other periods of major technological change? The Bureau of Labor Statistics has been operating since the mid-1880s. Their first assignment for Congress was to do a story on the impact of technological change. There is a long report that talks not about automobiles or the steam engine, but about the manufacture of bread boxes. For the first time, in my reading, it actually used the word 'displacement' to characterize the kind of technological change that was sweeping the country right after the Civil War, and traced out its many impacts. Is there some demarcation, something that makes today's change different from that of the past?
Leontief: The problem of distribution of income between what we call labor and capital has intensified. At the end of the nineteenth century, in this country, there were institutional factors that helped to maintain labor income. Trade unions, I think, play a very great role in maintaining the income of labor, essentially by reducing the supply of labor, as any monopolist can do, while increasing wages. In Europe, trade unions are still very important. Now one reads in the newspaper that an attempt is being made in the United States to revive trade unions. I am afraid that profit-maximizing business (and you cannot blame them for that) will choose automation, requiring less labor and more capital and that will reduce the effectiveness of the trade unions.
So long as technology sustains strong economic growth the social conflict arising from the declining relative importance of labor is muted. But I have a feeling that the rate of economic growth cannot be maintained forever. I believe that there will be no easy way out. Someone should compute what part of the income of the average American is really capital income. I consider social security to be essentially capital income because you are induced to save part of it and it is invested in capital.
Discussant: I agree with what you have said but I think we should look at how the arms industry fits into this. This is an industry that has benefited, if that's the word, from all the automation and other technological improvements, producing faster and more destructive weapons. It is a unique product and we hope that it will never be used and never perfected so it constantly changes before it is produced. It is also an industry that we depend on for creating jobs. So it is similar to digging gold in the ground and burying it someplace else in the ground. This is an industry we do not want. What would it do if we related it to the theme of today and did away with it? How would that permit the reduction of hours of work and the production of facilities for human need. What a waste! But at the same time, it employs pople.
Leontief: Thirty years ago I made a rather detailed computation of how the reduction of the military budget would affect the American economy, including effects on labor, in some regional detail. I agree with your assessment of the function of military spending, but military spending, or its reduction, doesn't solve the basic problem. It is just one element that will have to be taken into account; hopefully it will be taken into account very soon--although I'm not sure.
Discussant: We must anticipate internal conflicts as a result of the transformations that are already going on. Many people today are both workers and capital owners. As workers they are concerned about their jobs. At the same time, to some extent, they are capital owners through their ownership of pensions and other non-pension forms of investment in stocks and corporate equities. As the CEO of Chrysler Corporation was reported to say in the New York Times, "They are we". For example, Chrysler and GM are being censured because of the reduction in their employment. At the same time they are pressured by institutions that hold workers' pensions as well as by other asset managers to be more competitive so as to produce better results for their investors. So, in effect, the same workers who are worrying about reducing employment are also insisting that their firms become more competitive so that they can earn more through their capital ownership. At least in the medium term, I see this as a growing problem as we undergo the changes you have enumerated. Do you agree that this problem will grow?
Leontief: Profit is, of course, the main motivation of business now. Some kind of competitiveness on the job is not bad training, but often competitiveness is used as a nice friendly word to encourage other people to cut. My feeling is that, under the present system of pricing capital and labor, because technology is changing fast, a smooth transition is not likely to be achieved without institutional adjustment. In this country the problem is particularly serious because there is such great hostility to any government activity. The government is possibly the only agent who could do it here.
Discussant: Thirty years ago President Kennedy said "A rising tide lifts all boats". Now the tides are rising but 80% of the boats are sinking, i.e. the incomes of non-supervisory workers. Dr Heilbroner suggested that we might have to legislate to prevent people from working. That raises another question of social values: Is work good or not good? Professor Leontief spoke about social unrest.
All this suggests that there are some deep political problems around the corner. I would like to ask Professor Leontief, who has always taken a broader view of economics than simply technical expertise, whether democracy can survive in an environment of increasing dissatisfaction, with large numbers of people out of work, even legislated out of work.
Leontief: I have a feeling that there are all kinds of democracies. Many of them are adapted and I hope that adapted democracy will fare better. Dictatorship generally does not solve the problem unless the situation is very limited. But what really interests me is your question. I ask myself, and I do not visualize any answer, how in the less developed countries the introduction of new technology can find popular approval. Technology is so powerful that it can replace all kinds of labor.
