If the success of scholarship is based on merit, Richard G. Lipsey, Kenneth I. Carlaw, and Clifford T. Bekar's book, Economic Transformations: General Purpose Technologies and Long-Term Economic Growth will be recognized as an important turning point in the debates on economic growth. The book tries to answer the “big” economic history question: why the West grew rich? The logic of the book is as follows: growth cannot occur without technological change. This claim is supported with extensive examples of profoundly important innovations and thought experiments of economic development in their absence (e.g., imagine freezing technological advance in 1900, to what extent could growth be sustained based solely on capital accumulation?) However, technological trajectories are inherently Knightian uncertain and given the limitations of human decision making and foresight, progress proceeds only through trial and error. The West produced a social system in which new ideas and methods were (more or less) accepted or rejected based on their economic merits as opposed to other criteria, such as religious prejudice. This social system is driven by the prospect of super-normal entrepreneurial profits. Critically, given the accumulative nature of technological knowledge, informed experimentation required a uniquely Western development: modern scientific institutions. The authors focus on the development of scientific institutions, and the historical development of Christian, Islamic, and Chinese philosophies, which varied in their tolerance of scientific inquiry. They argue that a path-dependent sequence of events and decisions led to the theological accommodation and rise of science in the West. The scientific foundation to the First Industrial Revolution, Newtonian Mechanics, was most advanced and, critically, diffused among practitioners in Britain, which therefore became the most technologically advanced economy in the nineteenth century. Each claim is supported thoroughly and exhaustively with historical fact, psychological theory, or original argument, as appropriate. In this respect, the book is a delight.

The book also contains a damning indictment of neoclassical orthodoxy and the unconditional application of its policy implications, in particular as related to neoclassical growth models.Footnote ¹ These two components are tied together with appreciative theory, which both provides an alternative to neoclassicism and facilitates the interpretation of history. To complement Joel Mokyr's excellent Eh.Net review's emphases on the contribution to the “big” question, this review will focus on the book's somewhat unique juxtaposition of history and theory.Footnote ²

It is rare to see such a rich understanding of economic history coupled with strong skills in economic theory. The juxtaposition is leveraged to expose the weaknesses of the ahistorical analysis inherent in neoclassical theorists' efforts trying to understand growth in industrialized and industrializing capitalist economies, and in its place the authors advance an alternative “Structuralist-Evolutionary” theory (S-E theory). S-E theory labels different parts of the economy and institutions in helpful ways. It is useful, for example, to distinguish between organizational and technological changes—something from which modern growth theorists abstract away, and that distinction in turn provides insight into the nature of production changes that improve efficiency and enhance output possibilities, and it has implications for how such changes might be measured. Other assumptions underlying neoclassical conceptions of technology and behavior are thoughtfully refuted as well (e.g., distributions can be unknowable, or uncertain in the Knightian sense, and therefore large profits are necessary to encourage experimentation, not rents that serve no social purpose. Alternatively, the Solow residual does not accurately measure technological change because such change is embodied in capital whose contribution to growth may be realized only in the long term. Or assumptions of constant returns to scale ignore diminishing returns and economies of scale inherent in different technologies). The authors are careful to state that neoclassical analysis of growth can be useful and basic insights should not be ignored. However, underlying mechanisms of technology, institutional settings, and human behavior matter a great deal, and it is critical to take these into account. This often leads to policy recommendations that are at odds with neoclassical inspired recommendations. For example, since learning takes time, trade barriers can be useful if they allow vulnerable economies to incubate technologies. To wit, absent trade barriers Americans would have overwhelmed the Japanese automobile market and this would have left us without (or delayed the introduction of) the Toyoda system of automobile production (aka lean production). That is, the pursuit of short-run efficiency may be detrimental to the advancement of the human condition because it ignores the long-run dynamics of technological change. More profoundly, dynamic technological systems are path dependent, nonergodic processes that are influenced by chance historical events; the absence or delay of lean production was a very possible outcome avoided by a particular policy choice, not rational foresight. Incorporating a more informed understanding of the randomness and non-smoothness of technological processes allows for modeling the impact of rapid growth facilitated by particularly good technological draws. Generally, at the price of generating messier models, the authors incorporate assumptions that are significantly more realistic and historically informed. Thus they provide a preemptive though tentative response to Timothy Bresnahan's critique that GPT growth models are difficult validate empirically and therefore of limited use.Footnote ³

The strengths of the book are unhappily hindered by its structure, which attempts to include all the authors' related work. It is long, meandering and filled with tangents that are not germane to the authors' main points. That imposes greater burdens on the reader and will reduce the book's impact, to the detriment of all. For example, the authors devote considerable space to determining which technologies qualify as GPTs. The concept is certainly useful and important in focusing our attention on how successful use of most technologies requires some adjustment of adjacent sectors and related institutions, and that this adjustment process is dynamic and reciprocal.Footnote ⁴ GPTs are technologies for which there are a plethora of application sectors and critically, there are development complementarities. That is, a GPT is a GPT because it appears in a primitive state and improves over time through feedback from the application sectors, which leads to further development in the application sectors. However stratifying the continuum of technologies into GPTs and their lessers is both arbitrary and unnecessary, except in so far as it focuses our attention on the boundaries of the definition.Footnote ⁵ Indeed, identification of those technologies that qualify as GPTs by definition relies on ex-post success criteria informed by the historical record. Under the widespread assumption that technology fuels growth, this is dangerously close to tautology: GPTs are a subset of ex-post identified growth drivers that are critical to growth! Nonetheless, the concept is still useful because it recognizes and unpacks the process of complementarities and the implication that feedback loops between application and GPT sectors can lead to general increasing returns to scale and sustained growth even when decreasing returns prevail in all industries. Indeed, when the authors use the concept in their models, complementarities are a key feature. But unless we are able to show that the relationship between the importance of a technology and rate of progress is not continuous or at minimum non-monotonic, stratification is a distraction.

Nevertheless, for those willing to wade through these weeds, the book provides great rewards: readers will leave with a dimmer view of neoclassical growth theory, but they will benefit from an alternative formal approach that is enlightened by actual facts and that manages to put structure on an excellent historical analysis designed to answer the “big” question. I have not read another book that marries these varied approaches so skillfully and convincingly.