Endogenous Growth in Historical Perspective: From Adam Smith to Paul Romer
书评回顾了从亚当·斯密到现代的内生增长理论，探讨分工、报酬递增等概念，并批评数学化使理论脱离现实，适合对增长理论历史感兴趣的学者。
The book under review examines endogenous growth from a historical perspective, starting with Adam Smith and coming forward into the twenty-first century. What it tries to illuminate is not one thing but many. Growth we know of, but why endogenous? How does it relate to economic development? Is there any interaction between growth and development, and do they have a common history? These are general questions also asked by others. Why has the author entered this field? Chandra's originality lies in presenting ideas that he traces to Adam Smith and Allyn Young. The division of labor is fundamental to both Smith and Young, and the role of increasing returns, albeit in a new sense, was made central to Young's vision of capitalism. Chandra believes that the insights of Smith and Young have been neglected and need to be resurrected. The editor was kind enough to permit me to review the book despite my skepticism about growth “theory.” Let me discuss the general issues first and then turn to the merits of the Smith-Young framework. I will try to be fair to the concepts, but my tone may not be respectful.What makes growth theory endogenous? Endogenous must be endogenous to something. What is this something that endogeneity arises from? To understand this in plain English, we need to go back to the famous formulation of Robert Solow, which, if I may conflate two of Solow's papers, was meant to explain the growth of output in the United States from the mid-nineteenth century on. We posit that aggregate output, GDP, is produced by a production function whose inputs are land and labor: Y = F (L, K). The Y, F, L, and K mark the system, and what their interactions produce is properly called “endogenous.”Since we do not wish to be adding apples and oranges, the inputs have to be homogeneous. The growth of output between 1875 and 1950 is explained by the growth of some sort of (homogeneous) labor and some sort of (generic) capital. Since we are traversing a century, what will be the benchmark of such labor or capital? The most plausible answer is unskilled labor and some rudimentary tool such as a spade. So the original exercise began by assuming that the growth of the US economy could be explained by simply multiplying farm labor and the spades farmers used. Needless to say, such an exercise explains very little of the actual growth. So the observed facts came to be explained by assuming that technological progress fell like manna from heaven and enriched the powers of both labor and capital on a continuous basis. This technological change was properly called exogenous because nothing in the economic system explained how such technological change came about. The hunt was now afoot for finding ways to modify F, L, or K so that technological change became endogenous. The field for explaining growth using endogenous factors is vast: education, knowledge, invention, machinery, science, and organization are a few of the plethora of factors that will enrich and enliven such a barren plain. Since this (rather obvious) exercise is what endogenous growth does, why is so much made of the introduction of a word whose presence was assumed by everyone, including no doubt Robert Solow himself?There are really two meanings of the term economic growth. The first is the growth of advanced economies such as the economies of the United Kingdom and the United States. The second is the problem of development for poor countries. Somehow, it is pretended that the two problems are the same. This is a recent development, as most of the development economists between 1940 and 1960 were convinced that development involved changes in attitudes, consumption patterns, governance, and institutions. Even to raise the question of development was to place endogeneity front and center. Growth on the other hand was summarized in three variables, Y, L, and K, and the function F. Since the content of Y is generally agreed upon, all the jiggling of growth theory involves manipulating L, K, and F. However, since theory is meant to be all-encompassing, those who engaged in it felt free to talk about all countries, whether rich or poor, thus suggesting that Y, L, K, and F were globally applicable. Relevance, a local virtue, was trapped by the pretensions of science, a global concept.Why has “theory” been so prolific in a field in which endogeneity matters, whose aim is to penetrate the “veil” of society? The question probably has a simple answer. Once mathematics is introduced and its theorems are said to explain reality, the mathematics creates its own momentum. Twice differentiable functions can be replaced with differentiable functions, which can be replaced by continuous functions, which can again be replaced by measurable functions, and so on and so on. Such abstractions give empirically empty theorems. So we impose conditions such as separability or additivity in order to obtain predictions from the theory, a process known as tractability. The limits of our ingenuity are tested not by the extraction of insights from the real world but from the stock of available mathematical knowledge. Questions of growth and especially of development are very serious. It is a pity that they become subsumed within a framework that is endogenous to mathematics and not endogenous to the society being explained.The novelty of the new, endogenous growth theory has been challenged by many, some of whom, such as Robert Fogel (2009) and Nicolai Foss (1998), have escaped Chandra's wide net, as have some criticisms directed at Young, for example by Alice Amsden (1977). Fogel, in particular, should be a mandatory item in any course on economic growth. It is technology that is the prize in this venture. Can economists find a way to make technology arise from their formulations, or does it have to be left as a tabula rasa? Fogel pointed out that no one has studied technological change more assiduously than the economic historians. Among some of the findings are that the lag between invention and innovation can be over fifty years, that only two out of ten products that emerge from R&D become commercial successes, and that the cost of bringing an innovation to market is usually ten to twenty times the original research cost (Fogel 2009: 20, 27). Nothing in the mathematical formulations gives us an inkling about these institutional realities or the process by which they change. It may well be that the numbers we obtain after the battle is over, “when the hurly burly's done,” fit some convenient mathematical form. But the theory based on such forms is merely summarizing a complex reality post hoc—we learn little about how and why change does take place. It may be better to relabel the entire exercise as “concise summaries of economic data over centuries.”It is valuable to look at explanations of growth and development prior to the mathematization of the subject because such explanations will not be constrained by the needs of tractability. It is therefore welcome news that the subject is being treated from a historical perspective. The work of Professor Chandra is careful, competent, and comprehensive, covering almost three centuries of economic thought. He deals with a subject that everyone believes to be central to economics. It is an ambitious attempt, and he is to be applauded for having undertaken it. His book is well researched; every specific point dealt with has some support in the literature. Whether this scholarly care suffices to rescue our concern for endogenous growth is an open question. Chandra provides very readable summaries of all the authors he deals with. This includes references to the grand “stylised facts” that guide growth theory, first those of Kaldor, then those of Romer. Some chapters are excellent by themselves, such as those on Lauchlin Currie and on Rosenstein-Rodan. The concluding chapter provides all the ideas again in a nutshell. This makes it a convenient book to read and refer to.If we are to look at it from the viewpoint that Chandra has chosen, the historical perspective, then one obvious criticism lies in its acceptance of the Adamite heresy, which considers Adam Smith as the origin of economic knowledge, without even a peep about Smith's predecessors. Among the authors who have been neglected are George Berkeley and Josiah Tucker. Even before Tucker there was the tradition of what has been mislabeled as “mercantilism” (a convenient part of the heresy) but what is better called “pragmatic political economy.” A comprehensive summation of all that was best in this tradition is provided in the writings of George Berkeley. Subsequent to Ricardo and Malthus there was the wandering Scotsman John Rae. These are the scholars who cast a wide net and whose writings better embody the past of endogenous growth theory.One can also subject the internal logic of the book to scrutiny. Chandra pushes to interpret endogenous growth in terms of the division of labor and increasing returns. He is required to do this because he wishes to rescue Allyn Young's vision (11, 295). As Young was a devotee of the Adamite heresy, with the Wealth of Nations seemingly Young's only eighteenth-century reading, Adam Smith is involved as well. Charles Blitch correctly argues that Young's entire presidential address was motivated by a desire to save the capitalism he was familiar with and to argue that the presence of increasing returns, with its concomitant large-scale production, did not invalidate any of the essential virtues of capitalism (96). Young may well be right about the conclusion he wished to reach—capitalism flourishes and will continue to evolve and flourish—but his arguments are so strained that they need close examination. I will focus my critique on the division of labor and then discuss Young's own adaptation.That Adam Smith made the idea of the division of labor a household phrase is undoubted; whether he also added clarity to the subject is to be doubted. Consider for example the division of labor in a pin factory; the increasing returns envisioned are increasing returns to scale in the physical sense, clearly demonstrated by Smith's numerical example. This is effective only because the entrepreneur has already chosen those aspects of production whose subdivision and separation into tasks will be conducive to greater productivity. In other words, the planned recombination of labor is as important, if not more so, than the initial division of labor. Who does this planning? It is the entrepreneur, the capitalist. This insight, hinted at by Dugald Stewart, was later explicitly stated by E. G Wakefield. Smith's affection for the worker is undoubted, but it misled him seriously in terms of the role of the capitalist, particularly in evaluating inventions. To imagine that a worker would invent a machine that would make his task redundant is the height of absurdity from the point of view of self-interest. Such a worker has just invented unemployment. Furthermore, to attribute most inventions to workers and only a minority role to inventors goes against the entire grain of capitalist production. It is the capitalist who is devoted to profitability, hence in dividing and recombining labor, as well as in machines that will increase profitability, and hence in all improvements that would make his machines better. That Smith could think of workers inventing machinery to relieve themselves of the division of labor in book 1 and write equally stirring passages in book 5 on the idiocy induced by repetitive labor is a marvel. With all his love for the free market, Smith never quite understood or appreciated the primacy of the entrepreneur in capitalism (Rashid 1986).The difficulties in interpreting Young's famous presidential address of 1928 on increasing returns are many. It is very imaginative and stimulating but hard to obtain clarity from. To be fair, Young himself did not wish to be clear in the sense theorists consider economic concepts to be clear. It is perfectly all right for Young to argue against the trend of the profession and suggest that they are asking the wrong questions. He makes many interesting claims en passant: British managers are just as efficient as American ones, technology arises effortlessly, population growth has helped the economy, supply provides its own demand, and so on. However, Young seems to want the language of the economist at his disposal in arguing for his new vision. He uses phrases like “increasing returns” and “division of labor” in a wide and nebulous sense but expects these amorphous arguments to convey the same force as standard usage. Young was honest. He told us clearly not to expect “economic logic” from his speech. It is the fault of his readers for having attributed meanings he did not intend. Let me give examples of each.A little later, after telling us that he means to address the global changes occurring under capitalism, those beyond the firms and industries economists focus upon, he writes, “Change in this external field is qualitative as well as quantitative. No analysis of the forces making for economic equilibrium, forces which you might say are tangential at any moment of time, will serve to illumine this field. . . . Instead we have to go back to a simpler and more inclusive view” (Young [1928] 1969: 229).Young thinks that recent economists have “added nothing” to Smith's famous theorem that the division of labor depends upon the extent of the market. If Young had only recognized that the increasing returns Smith's pin factory deals with is a feature of the technology and stayed with that use, all would be well. But Young wishes to use the division of labor and increasing returns in other senses. Thus increasing returns means a monetary phenomenon for Young ([1928] 1969: 233). As long as greater supply means more revenue, Young is convinced that progress will continue—even without any technological progress! (234). To go further into this classic paper—I remember being inspired by it as a student—is perhaps needless. Young has conflated several distinct concepts in his eagerness to rescue capitalism. Marshall's view of industries composed of firms enjoying external economies dependent upon the size of the industry may rescue the use of perfect competition, but it has little to do with the division of labor and the technological increasing returns of Smith. That the economies of large-scale production in one firm can be duplicated by having many small firms is only a logical possibility, for which Young has provided very little support. Young felt that Marshall had rescued the neglect of increasing returns at the level of the firm, but how does one identify the industry? This last problem was one that Young's student, E. H. Chamberlain, had to face when he followed Young's insights into the theory of monopolistic competition.It is the aggregate market that matters for Young, and, owing to the benefits of capitalism, this aggregate somehow grows by feeding on itself and rising in a virtuous spiral. “Modified then in the light of this broader conception of the market Adam Smith's dictum amounts to the theorem that the division of labor depends in large part upon the division of labor.” (Young [1928] 1969: 233). If Young had only read Josiah Tucker, he would have found someone who gave good reasons why a rich country not only has more knowledge and infrastructure than a poor country but also has the capacity to create more. Furthermore, Tucker sees economic progress as indefinite. Young would even find an almost exact statement of his own thesis: “In the richer Country, where the Demands are great and constant, every Manufacture that requires various processes . . . is accordingly divided and subdivided into separate and distinct branches” (Tucker, 1774: 25). Had Young not been enthralled by the Adamite heresy, he would have known that and that Tucker fully anticipated Young.Chandra has chosen to climb uphill in rescuing Young's vision and seeking its compatibility with endogenous growth. It is a tribute to his scholarship that the book is as readable and valuable as it is. I would definitely recommend it when teaching a course in growth or development, and I hope others will too.