Appreciation by Robert M. Solow

Leif Johansen (1930-1982): A Memorial

Robert M. Solow

Massachusetts Institute of Technology, Cambridge, MA, USA

Biographical note: Leif Johansen was born in Eidsvoll, Norway on May 11, 1930. He received his doctorate from the University of Oslo in 1961. He was Professor of Economics at the University of Oslo, 1965-1982, served as an adviser to the Norwegian Ministry of finance on economic planning, on the royal Commission on Money and Credit, 1960-1964, and on the Royal commission on Basic Research, 1980-81. Leif Johansen was co-editor of the Journal of Public Economics and for various periods Associate Editor or member of the Editorial Board of other scientific journals. He was awarded the Fridtjof Nansen Award for Outstanding Contributions to Economics in 1979 by the Norwegian Academy of Sciences and Letters.

Leif Johansen's first published article appeared in 1952, when he was just 22 years old. It must have been an early student paper, because his first degree in economics dates from 1952. I have no way of knowing what sort of student he was, or when his teachers and contemporaries first recognized the extraordinary and rather special talent that he had. Others will no doubt report at first hand on those biographical details. He began to publish regularly in 1955, as a research assistant at The University of Oslo. An article in English appeared in 1956, but it was published in Ekonomisk Tidsskrift. I certainly did not know of its existence, and I doubt that it was much read outside the Scandinavian countries. I cannot remember whether I read his "Not on the theory of Inter-industrial Wage Differentials" in 1958. But of course he burst on the world of English-speaking economics in 1959 with the famous "Substitution versus Fixed Production coefficients in the theory of Economic Growth".

I can easily remember what an exhilarating pleasure it was to read that article. (He sent me a copy of the manuscript and we corresponded before it appeared in Econometrica.) The idea and its execution were a particular delight to someone like me who was then working on the same set of problems. There was an additional source of excitement, however, the unmistakable experience of meeting a powerful, lively, and original mind. I do not suppose that it was obvious then that Norwegian economics had found a natural successor to Frisch and Haavelmo, someone who could carry on the great tradition at Oslo. But it must soon have become clear to everyone who read his work. So the sudden and unexpected death of Leif Johansen at the sadly early age of 52 is a tragedy not only for Norwegian economics but for economics everywhere. It is, nevertheless, a special and incalculable loss for Norway in particular. An intellectual tradition, even a firmly rooted one like that of Norwegian economics, is always fragile in a small country. I count among my own friends some younger Norwegian economists, people in their 20s and 30s, and I know that they feel they have lost a leader, a guide, someone who could not only offer them help and encouragement but also serve as an example of what a rigorous and serious-minded economist should be.

I speak also for myself in this article. I have the feeling of having lost a friend, an ally, almost a co-worker. This is strange, because I met Leif Johansen in person only a couple of times. He had a standing invitation to visit M.I.T. whenever and however he wished, but he was never ready to do that, presumably on political grounds that I understood perfectly. For my part, I do not often go to European meetings, and I have never set foot in Oslo. But we kept in touch; I usually knew what sort of thing he was working on, even when our scientific interests diverged. At the very end, as I shall describe later, we were again thinking about related problems. I had the same sense of excitement when I read his last paper as I had in 1959. It is sad to think that it will not happen again.

It is not possible for me to give a complete survey of Leif Johansen's work in a brief memorial article, if only because only the fraction that is in English is accessible to me. Instead I shall be frankly selective and discuss only a few of his works, chosen either because they are especially meaningful to me, or of general interest in the history of modern economics, or particularly revealing of his personal style as an economist.

Early Articles (1958, 1959)

"A Note on the Theory of Interindustrial Wage Differentials" was Johansen's third published article in English, but the first to appear in one of the standard English-language economics journals. (The first of all, as I have mentioned, appeared in Ekonomisk Tidsskrift, and the second in a journal read by demographers.) I want to say a word about it primarily because it was his introduction to the English-speaking part of the profession. There is besides, a second source of interest in it. The theory of wage structure was always a central issue in labor economics. Just about in the 1950s, however, it was eclipsed by discussion of the determination of the general level of wages, nominal and real. It seems possible that questions about relative wages, nominal and real, are now coming back to the center of the stage. In fact the 1958 Note has some connection with Leif Johansen's last published paper in English, the 1982 article on unemployment with heterogeneous labor which I shall discuss later.

The 1958 article is a straightforward note of just five pages. It starts from the fact that the standard (American) labor-economics literature tends to make interindustrial differences in the level of wages dependant on corresponding differences in the level of productivity and employment. Johansen exhibits a plausible dynamic mechanism which would cause differences in the level of wages to be a function of differences in the rat of change of productivity. The mechanism is this. Suppose that the interindustrial mobility of labor is less than perfect, so that a persistent wage difference will attract a persistent flow (not a spike) of labor into the industry with a favorable differential. If one industry wishes to increase its employment more rapidly than another, which is a likely consequence -- under full employment, as Johansen points out -- of a faster growth of productivity, then a persistent wage differential will tend to emerge. So it is wage levels, not rates of change of wages, that will tend to be correlated across industries with rates of change of productivity (and employment).

It is a nice, neat piece of work. Anyone with a taste for simple and clean models will enjoy it, even today. And, as I mentioned, it may be on a subject whose time has come again. But I doubt that a reader would have seen, then and there, that the young author would soon evolve into one of the world's leading economic theorists.

There was not long to wait. The 1959 article came at a moment when the theory of economic growth was at the center of active discussion from the theoretical, empirical and practical points of view. It attracted instant attention. Everyone now knows what Johansen did, so I shall be very brief. Growth theorists suffered from a "guilty conscience", as he pointed out, quoting Domar. To assume a technologically-fixed ratio of labor to capital in production at the macroeconomic level is clearly an exaggeration, and not a harmless one because it eliminates an important adjustment mechanism from the economy. The assumption of smooth substitutability is also unrealistic, and may make adjustment to changing investment or saving habits look too easy. (I felt at the time, and still do, that the second alternative is better, it only because it can be thought of as allowing something for the possibility of automatic shifts in the composition of macroeconomic output as between labor-intensive and capital-intensive final goods.) Johansen proposed the clearly superior alternative that relatively free substitutability is possible only at the moment of gross investment, but that the operating labor requirement chosen at that moment is fixed afterwards for the life of that particular item of capital. (This sort of technology goes under the name of "putty-clay". The term was coined by E.S. Phelps who wrote on essentially the same subject at essentially the same time. It is a testimony to both Phelp's originality and his lack of acquaintance with the properties of putty!)

Johansen goes on to incorporate this putty-clay technology in the standard sort of one-sector full-employment growth model under a few choices of ex ante production function, depreciation pattern, and saving behavior. He is completely aware that the full-employment assumption forces the ex ante choice of technique always to be such as to provide employment on new capital for the increment of labour made available by natural growth and the retirement of old capital. His point is primarily to show that it is possible to work with the putty-clay technology in manageable models. In this he succeeded.

There has been continuing discussion of the theoretical value and the empirical significance of the putty-clay assumption. It has proved empirically elusive. Probably this merely illustrates the proposition that the smoother the development of the actual economy, the closer it is to steadystate behavior, the more nearly identical are the implications of putty-clay to those of universal substitutability. But I am convinced, as I think most interested economists are, that the putty-clay representation is the best one for analyzing the problems of investment and capital.

The mere idea of combining substitutability ex ante with fixed factor proportions ex post is not subtle or difficult. The thought itself must have occurred to many economists who were concerned with the theory of growth and capital in the 1950s, or even before. The mark of the find theorist is not so much to have such an idea, but to see that something can be done with it -- and then to do it. Johansen's 1959 paper has no very difficult or obscure mathematics, no long and tortuous proof. Sometimes those things are necessary for progress in economic theory, perhaps even more so nowadays. It seems to me, however, that the really important characteristics of the best sort of theorist are simply clear-mindedness, ingenuity, and patience. Leif Johansen had them from the very beginning until the end.

Production functions (1972)

The book on production functions -- subtitled "An Integration of Micro and Macro, Short Run and Long Run Aspects" -- was published in 1972, but clearly has its roots in the putty-clay article of 1959. The connection is worth describing. In the spirit of the putty-clay model, when productive capacity is first created and capital committed, inputs are substitutable. Depending on current and expected cost and demand conditions, and subject of course to technological constraints, a firm can build a plant of given capacity with a larger or smaller commitment of fixed factors, compensated by smaller or larger requirements for variable inputs during its operating lifetime. The particular mix of variable inputs can be chosen fairly freely, but is partially or completely predetermined at the time capacity is created. The simplest sharp embodiment of this idea allows fairly free substitutability ex ante among the fixed and variable inputs; but once a plant is built it requires variable inputs with fixed proportions and constant returns to scale up to capacity, and can not produce more. That was the message of the 1959 paper.

The next step is the one analyzed in detail in the 1972 book. Within any industry there will exist an array of plants (or smaller productive units) built at various times in the past. The various choices of input proportions will have been made under different technological constraints, more recent capacity being more advanced, and also with different experience and expectations of the prices of variable input. The current short-run state of the industry is thus described by a k-variate capacity distribution if there are k variable inputs. This distribution shows how much capacity there is corresponding to each technologically feasible collection of input coefficients. Its total volume is the current capacity of the industry. If we assume that input requirements do not change as a plant ages, then the capacity distribution changes through time only as old capacity is retired and new capacity is built. Otherwise the situation is more complex.

Now consider a given instant. One can do two important exercises with this model. First, take as given the k+1 prices of output and the variable inputs. Only a certain easily calculable subset of the possible vectors of input coefficients is viable, in the sense that such a plant could break even or earn a positive profit at the given prices. The capacity associated with the viable plants represents the competitive supply of output at the given configuration of prices. By varying the prices, one calculates the supply function for the industry as function of the prices of output and the variable inputs. It is easy to see that this supply function is generated by an efficient allocation of variable inputs to units of capacity. If it were possible to produce a given output with less of some variable inputs and no more of any of them than in the configuration given by the supply function, there would have to be idle capacity than can earn a profit, and that is excluded by construction.

This suggests the second exercise. Choose a level of output. Fix the input prices and find the output-price that generates a supply equal to the stipulated output. This will be possible if the stipulated output is not too large. Calculate the input requirements of the plants viable at the price configuration thus reached. Since the allocation of resources is efficient, this input-bundle represents a point on the isoquant for the level of output with which we started. Hold to the same level of output, try a different input-price configuration and find a second point on the same isoquant. By going on in this way one can in principle generate the complete isoquant, and then every isoquant, and thus the complete short-run production function. (For very high outputs, the range of possible variable-input-bundles will narrow; at the extreme, the largest output producible is the capacity of the industry, and since it requires that every plant be fully used, there is only a single bundle of variable inputs capable of producing it -- the isoquant shrinks to a point.)

This idea of generating a production function by aggregating over elementary cells with a distribution of input-coefficients was originally due to Houthakker. He proved the neat result that a modified Pareto-distribution of input-coefficients gives rise to a Cobb-Douglas production function when aggregated. There is related material in W.E.G. Salter's Productivity and Technical Change (1960). Johansen carries the analysis much further, with great patience, care and ingenuity. I will give only one example. Any economist would be curious about the derivatives of the aggregate production function and their relation to the characteristics of the capacity distribution over the underlying micro-cells. On reflection, those derivatives must depend on the characteristics of the marginal cells (i.e. those just breading even at the initial aggregate inputs and output) because small changes in output will occur through the activation of deactivation of marginal capacity. Johansen works out all the details an d shows how the first- and second-order properties of the aggregated production function, including the elasticities of substitution between variable factors, can be written nearly in terms of the "statistical" properties -- the means, variances and covariances -- of the existing capacity distribution over the marginal cells. It is very satisfying.

It seems to me that this "distribution approach" to the vintage putty-clay model of production must be the right foundation for any serious theory of investment intended for application at something less than the highest level of aggregation. Think, for example, of the need to deal currently with the existing mix of capacity with varying degrees of energy-efficiency and the complicated incentives set up by the prospect of rising (or falling) energy prices. The model is hard to work with, and that is probably why it has not been used much. (There was an early application of the putty-clay apparatus to U.S. investment data by Charles Bischoff, and I have the impression that this approach is favored by the Dutch Planning Bureau.)

All those who wish to deal with this approach can learn from Johansen's treatment of it. Apart from important details, they will learn something about the way a certain sort of fine theorist works. I am thinking here of two characteristics. One is supremely important: it is the constant check of common sense. It is always in his mind that the strict assumptions of theory will not hold in practice: ex post input-coefficients are not quite fixed; they do change a bit as capacity ages; the allocation of variable factors to units of capacity may not be quite efficient. There are marvelous theorists who are not troubled by such intrusions; Leif Johansen was not one of them. The second characteristic is a kind of playfulness, a willingness to try out special cases and specific assumptions in the hope that something interesting will turn up. In his case, it often does. That can not be mere luck.

Multisectoral Models and macroeconomic Planning (1960-1978)

I have not found it possible to summarize in any neat way Leif Johansen's lifelong interest in the use of fairly aggregative -- but still multisectoral -- long-term models as a guide to macroeconomic planning. His main contributions to this subject are A Multisectoral Study of Economic Growth (1960) and the two volumes of Lectures on Macroeconomic Planning (1977 and 1978). A good way to get the feel of his approach is a fifty-page article "Explorations in long-term projections for the Norwegian economy" written in collaboration with Harvard Alstadheim and Asmund Langsether and published in Economics of Planning, 1968. There are also incidental items in English and no doubt much more material, published and unpublished, in Norwegian.

The underlying model is in most respects an equilibrium construction. For example it provides always enough consumer demand to maintain full employment. This is a reasonable way to proceed in a model whose purpose is to provide "a background and a framework for the continuous flow of decisions which have to be taken with respect to those elements of the economy which are under more or less direct influence by the government, in addition to serving its purpose of providing a background for the general four-year programs". Equilibrium is presumed to be maintained by active policy, not by automatic forces. The structure of the model is a sort of hybrid. Interindustry flows of materials and services are governed by a fixed-coefficient input-output matrix. (Johansen describes it as "disappointing" that he has not yet been able to introduce more technological flexibility here.) For capital and labor, however, each sector is given its own Cobb-Douglas production function, including an industry-specific exponential technological change factor. Thus, technological progress does not economize on flows or raw materials and intermediate goods.

Aggregate investment is allocated among sectors through rates of profit, but it is not assumed that the flow of investment equalizes the rate of profit across sectors. Instead there is a stationary structure of relative rates of profit by sector. Johansen regards this assumption as a "weak link" in the model. He tests it by fitting an autoregressive model to the observed rates of return in sixteen separate manufacturing industries, 1950-1963. It is found that the convergence of relative rates of return to their estimated stationary levels is quite slow if it occurs at all. Johansen makes the interesting observation that in general the rate of return in "sheltered" industries was rising relative to that for "exposed" industries during the sample period. I do not know when that particular dichotomy entered discussion in Norway; I first heard about it in connection with the "Scandinavian model" of inflation, some years later.

There is another interesting aspect of the empirical results that struck me on rereading the 1968 paper, but on which Johansen offers no comment. The direct price elasticities of demand are estimated for each sector using Ragnar Frisch's "complete scheme". (Income elasticities are taken over from the Central Bureau of Statistics.) The price elasticities are all very small: the largest (-0.4) is for electricity consumption and the largest among manufactured products is for the mechanical and electrotechnical industry (-0.3).

These few details are not meant even to give the flavour of the model. My goal is just to say something about Johansen the economist. What is striking in this part of his work is its pragmatic approach. There is always the discipline of a formal model in the background It is too much to say that anything goes; but he is always willing to experiment, to listen to the data and their limitations, but to be as sophisticated in conceptual approach as they will allow, to relax an a priori constraint to see if doing so will improve the fit of model to reality. This pragmatic attitude is visible also in the Lectures on Macroeconomic Planning, which have that reassuring ability to suggest the controlling presence of analytical bones in the background while allowing for the cushion of common sense in applying them to the everyday problems of the real economy.

Public Economics (1962-1965)

Although the English translation of Public Economics appeared in 1965, the Norwegian original was published in two parts, in 1962 and 1964. It is based on a course of lectures on public finance at the University of Oslo, evidently for undergraduate students. What did they learn?

The course starts with a discussion of targets and instruments in the Tinbergen manner. The chapter is about evenly divided between an elementary exposition of the analytical framework and a brief commentary on the main targets of economic policy. The next topic is fiscal policy in its economic-stabilization aspects. The treatment is clear and straight-forward. It is so clear, in fact, that it quite naturally contains a brief treatment of the "government budget constraint" although the well-known papers by Ott and Ott and Christ did not appear until 1965 and 1966. It is too bad that the fairly elementary level of the course required Leif Johansen to limit himself almost entirely to static theory. Otherwise he might have developed further a study of the longer-run issues connected with the financing of the public-sector deficit. Probably for the same reason, the very short chapter on the public debt does not go very far, though it covers its ground with great lucidity. At this date I am left wishing that he had had the time and space to discuss the demand for bonds and money and their relevance for stabilization policy. It occurs to me that his students may have had better preparation in microeconomic theory than in macroeconomics. That would explain why he writes down an IS-LM model as a vehicle for discussing the relationship between fiscal and monetary policy, but does not actually solve and use it.

The second part of the book -- in fact exactly the contents of the 1964 Norwegian publication -- discusses "The extent of collective consumption and public activity" and "Forms of taxation and their effects". I will call attention here to only two points, because that reflect on the common sense that I consider to be one of Leif Johansen's great virtues.

There is, as one would expect, an exposition of the Lindahl pricing of public goods. It is nicely done; the idea is given full credit for the insight it provides, but its value as a practical device is not overstated. It is followed by a section describing the possibilities and difficulties of majority rule as a device for allocating public resources. Johansen evidently took the problem seriously, as a "real" problem, not technical exercise. Over a decade later, in 1977, he wrote a six-page note ("The theory of public goods: misplaced emphasis?") suggesting that the fashion for finding and describing complicated mechanisms to elicit accurate revelation of preferences might itself be a misallocation of intellectual resources. He gave two main arguments for believing that the free-rider problem might be less intractable in practice than in theory. The first reason is that individuals, especially when acting in a sort of public capacity, might feel that they ought to meet certain standards of honesty and responsibility, and might behave accordingly. The second reason is more structural, Johansen points out that public-good decisions are in actual fact done in a two-stage way: voters elect representatives who then, in turn, make the decisions in full view of their constituents. The dissembling that would have to be an inevitable part of the free-riding is not so easy. For example, a politician who wishes to appeal to those who favor a particular expenditure will have to speak out on that side of the issue in electioneering. Once elected, he or she can not very well vote Yes but pretend not to care. There is more in the same vein. Johansen makes a very good case.

The second example is almost trivial: the chapter on taxation contains the conventional sort of exposition of descriptive concepts about income taxation, especially progressivity. But Leif Johansen's students were not treated only to a boring technical discussion of fine points. They learned enough about the technical concepts to be able to think about the social-political issue of providing needed tax revenue in a reasonably fair way. The practical issues are well discussed. Students must have been forced to think of the theory of public finance as a tool for good government. It is another example of seriousness without dullness.

Last Articles (1982)

As he reached his fiftieth year, Leif Johansen did not lose his interest in, or his capacity to do, technical work in economics. But one can sense a deepening of his concern with broader questions of economic life and of the strategy of modelling economic life. One interesting example is his article "The Bargaining Society and the Inefficiency of Bargaining" (1979). I think this development is both desirable and inevitable. A theorist with a serious interest in the real world will eventually have seen a lot of the world and a lot of theory. He or she is bound to accumulate ideas about priorities in research and about the likely significance of alternative approaches to research. It is, in fact, a duty to pass these ideas along to students. I am not referring to mere pontification; one cannot imagine Leif Johansen being pontifical. I shall discuss two articles published in 1982. They may not have been his last papers; there may be others in the pipeline. But they are the last I read before hearing of his death.

"On the Status of the Nash Type of Noncooperative Equilibrium in Economic theory" appeared in the Scandinavian Journal of Economics in 1982. Johansen starts from the observation that the Nash equilibrium is far and away the preferred concept of noncooperative equilibrium in economics. He then goes on to consider the various justifications for this preference that are given in the literature. He finds that most of them leave the impressions that Nash equilibrium describes the outcome of rather naïve behavior or foolish assumptions about others on the part of the agents concerned. If that were really so, it would be hard to justify the widespread use of the Nash concept (except on the unsatisfactory ground that it is available and workable).

But Johansen thinks that the foundations of the Nash concept are much deeper, and the popular explanations miss the point. He hoes on to argue that the Nash equilibrium -- provided it is unique -- has a claim to being coextensive with individual rationality. To do so he proposes five postulates for rational behavior in the noncooperative-game-type situation. They are, roughly, that: players know only the feasible actions and the preference functions of the other players; thus if there is a rational decision for a player, other players will be able to predict it, and each player will realize that; given such predictions each player chooses selfishly. The fifth postulate -- that no player regrets his decision -- in the sense that he would behave differently if given another chance along with the others -- is according to Johansen essentially implied by the first four. He then goes on to claim that the Nash equilibrium is the only way to satisfy these requirements. Everything turns on the postulates. Johansen tries to persuade the reader that the postulates are necessary and sufficient for what we mean by individual rationality in this sort of setting. If he succeeds, then the Nash equilibrium concept does indeed have deep foundations.

It is absolutely characteristic of Leif Johansen that he ends by pointing out that Nash equilibria may be very hard to achieve in practice, and may be socially destructive if achieved; this is likely to lead the agents to try to change the situation. They may postpone action, seek further information, redefine the rules. Nevertheless Johansen states his belief that Nash equilibria will be important and typical in practical situations. (I think he might have pointed out that the specification of choice variables could be part of such a redefinition -- prices and quantities being an obvious example.)

Finally I would like to call attention to a fascinating paper called "Some Notes on Employment and Unemployment with Heterogeneous Labour" which appeared in a Festskrift for Jorgen Gelting published by the Danish Economic Association in 1982. One must go out of one's way to read it. But the level of unemployment, its dynamics and its intergroup structure are now among the central social problems of our time and — among the central intellectual problems of our science. Fresh and promising theoretical insights into the problem are rare. This is one.

Suppose that the members of the labor force fall into grades or categories by decreasing order of skill, talent, or "effective labor" per hour worked. The labor input generated by any group of workers is thus a weighted sum of the numbers in each grade, with higher weights for the better grades. Suppose that workers must all be paid the same wage, regardless of grade, for historical or institutional or political reasons. Employers do not generally know the grade of each worker—who may not know either—but it may be that employers learn something, though not necessarily everything, about a worker's grade by employing that worker. The marginal value product of employment is the current marginal value product of effective labor multiplied by the amount of effective labor generated by the increment of employment. If that exceeds the wage, profit will increase with employment. If it is less than the wage, it will pay to lay workers off.

If employers are unable to be selective when they hire, they will hire a random sample of the unemployed, and the effective labor per unit of employment will be essentially the average grade of the unemployed. If some selectivity can be exercised, the average grade of those hired will be better than that of the unemployed at large; and the average quality of the unemployed will decrease as their better members are creamed off. An employer may be able to be fairly selective in laying off workers, in which case those laid off will be of poorer average quality than the employed work force at large, and the average quality of those employed will improve as layoffs go on. If "efficiency-in", as Johansen calls it, exceeds "efficiency-out" by virtue of greater selectivity in layoffs, it is possible that an employer can increase profits by simultaneously hiring workers of approximately average quality and laying off interior workers. Johansen calls this process "filtering". A robust sort of equilibrium arises if the marginal value process "filtering". A robust sort of equilibrium arises if the marginal value product for hires is smaller than that for layoffs, and the wage falls strictly in between. Then it is unprofitable to add workers, because the average quality of the intake is poor, and unprofitable to lay off workers—despite selectivity—because the average quality of all those who are employed is high. Small changes in data will not disturb this situation.

Here are two stories Johansen can tell using this model. Imagine a demand-driven upswing with employment rising and little or no filtering. Eventually a peak is reached and falling demand leads to layoffs. Now employers exercise whatever selectivity they can. The result is that when employment reaches the initial level, the average grade of those still employed is higher than it was at the start, and the average grade of the unemployed is lower. If the original burst of demand now repeats itself, efficiency-in will be lower at each stage and the marginal value product for hires may fall to the level of the real wage rather quickly and choke off the boom. What used to be Keynesian unemployment has been converted to classical unemployment at the old real wage.

Secondly, this effect can be strengthened if a worker's quality improves with employment and deteriorates with unemployment. The dynamics may create or intensify a strong class barrier between high-grade and low-grade workers. By this route, a prolonged recession can create ordinary cyclical unemployment and convert it into what looks like "structural" unemployment.

If the reader has a feeling that this story may actually capture something of the recent history of industrial economies, then the reader has the same feeling I do. There is more in the paper, but it is clearly only a first and incomplete step. There could by no better tribute to Leif Johansen than for someone to carry it further.

A Last Word

Economics has many mansions, and it takes all sorts of economists to occupy them. Leif Johansen represents the best of one sort. He was always perfectly willing to get wrapped up in the hardest sort of technical calculation and he was very good at it. But he never once lost sight of the underlying goal of social betterment. The rest of us could not ask for a better example.

URL of this topic: www.copsmodels.com/webhelp/viewhar/hc_johansen1.htm

Link to full GEMPACK Manual

Link to GEMPACK homepage