Tuesday, April 11, 2006

This weblog has moved to My New Blog

Thursday, April 10, 2003

Hello There Again

Thursday, February 27, 2003

Why THREE industrial revolutions?

Any reader of my China blog might well ask themself, why third industrial revolution? What were the other two? And why industrial, aren't we now in the information age? The purpose of this post will be to try and clarify a little what it is I am trying to say with this provocative expression. To start with the obvious, the first two industrial revolutions took place in the United Kingdom (at the end of the eighteenth century) and in the US (at the end of the nineteenth). Now in China we are seeing a process of transformation (funnily enough begining at the end of the twentieth century, but I'd say that's just a coincidence) which is likely to be every bit as significant and even dramatic in its consequences as its two predecessors. The reason I think the impact will be so significant is the sheer scale of the operation. Even if the value added per head might seem small in terms of international comparisons, the number of heads involved means that here is is a case of quantity means quality. And why industrial, is this not the information age. This is true, but China is only now really making the transition from a predominantly agricultural society to a highly industrialised one. If the US is the world's 'consumer of last resort', China is more than likely becoming the producer of first resort. But I am getting ahead of myself. Let us first go back to the end of the eighteenth century and ask the question: just what exactly was the first industrial revolution - or better put, in what did it consist?


In fact, far from being self-evident the very concept of industrial revolution itself has been subject to enormous scrutiny over the years. For some the very term 'Industrial Revolution" is, in fact, "a concept too many" on the grounds that that either it is too vague to be of any use, or that it produces false connotations of abrupt changes comparable in their suddenness to political ones like the French or Russian Revolution. What, then, was it that changed and in such a 'revolutionary' fashion in the process that we refer to as the first Industrial Revolution?

Joel Mokyr differentiates between four categories of response. The Social Change idea: that the industrial revolution constituted first and foremost a change in the way economic transactions between people took place. The emergence of formal, competitive, and impersonal markets in goods and factors of production is the basis for this view. Karl Polanyi for example judges the emergence of the market economy as the truly fundamental event, with everything else having been incidental. In this vein most modern social historians probably would view the central social changes as having to do with labor, and the relation of workers with their work environment, other laborers, and their employers. Secondly there is the Industrial Organization idea: the emphasis here is on the structure and scale of the firm, in other words on the rise of wage labour employment and eventually the factory system. The focal point is the emergence of large firms, such as industrial mills, mines, railroads, and even large retail stores, in which production was managed and supervised and where workers were usually concentrated under one roof, subject to discipline and quality control.
Thirdly comes what Mokyr terms the Macroeconomic School: here the emphasis is on aggregate variables, such as the growth of national income, the rate of capital formation or the aggregate investment ratio, or the growth and composition of the labor force. Some writers classified as belonging to this school, such as Gerschenkron prefer aggregation at the sectoral level, focusing on the rate of growth of the manufacturing sector rather than that of the economy as a whole. This sectoral emphasis will, as we shall see, once more acquire importance in the context of the impact of today's information technology revolution. Finally there is what might be termed the technological idea: the attribution of primacy to technological changes over all others, and the identification of invention and the diffusion of new technical knowledge as paramount. Technology as used here, of course, means more than just "gadgets", and encompasses techniques used for the organization of labor, marketing and distribution etc.

The Britain of the industrial revolution may usefully be regarded as a dual economy, in which two economies, one agricultural and domestic, the other industrial, coexisted. The former may also be termed the traditional economy, which developed gradually and along conventional lines, with slow productivity and slowly rising capital-labour ratios. This sector contained agriculture, construction, domestic industry, and many traditional "trades" that we would now classify as industrial but which in the eighteenth century and before were partially commercial: bakers, millers, tailors, shoemakers, hatters, blacksmiths, tanners, and other craftsmen. The modern sector consisted of cotton, iron smelting and refining, engineering, heavy chemicals, mining, some parts of transportation, and some consumer goods such as pottery and paper. At first, however, only segments of these industries underwent modernization, so that dualism existed within as well as between various products making estimations of relative sizes even more difficult.

The average size of agriculture and "all others" between 1780 and 1860 was 79 percent of the British economy, meaning that in 1760 it was likely to have composed close to 90 percent of the British economy. Productivity growth in this sector is estimated by McCloskey at about 0.6 per cent per annum. During the same period productivity in the modern economy grew at a rate of 1.8 percent per annum.Two-sector growth models imply that abrupt changes in the economy as a whole are a mathematical impossibility because the aggregate rate of growth of any composite is a weighted average of the growth rates of its components, the weights being the respective shares in output. Even if changes in the modern sector itself were discontinuous and its growth rate very high, its small initial size would limit its impact on the economy-wide growth rate, and its share in the economy would increase gradually. In the long run the force of compound growth rates was such that the modern sector swallowed the entire economy. How long was the long run? A numerical example is illuminating here. Assume two sectors in a hypothetical economy, one of which (the modern sector) is growing at the rate of 4 percent per annum while the other (the traditional sector) is growing at the rate of 1 percent per annum.15 Suppose that initially the modern sector produces 10 percent of total output. Then the aggregate growth rate is at first 1.3 (=.9x1 + .1x4) percent. After ten years the aggregate rate of growth will have increased to 1.39 percent per year. After thirty years of "dual growth" the share of the modern sector will have increased to 21 percent of the economy and after fifty years to one-third. Only afterseventy-four years will the two sectors be of equal size (at which point aggregate growth equals 2.5 percent per year), and a full century after the starting point the traditional sector will have shrunk to about 31 percent of the economy. The British economy as a whole was changing much more slowly than its most dynamic parts, because growth was diluted by slow-growing sectors (Pollard, 1981, p. 39). These hypothetical numbers fit the actual record rather well, and they indicate that it is hardly surprising that it took until 1830 or 1840 for the economy wide effects of the Industrial Revolution to be felt.


Utilizing the distinction between a modern and a traditional sector allows us to summarize what happened to the British economy during the Industrial Revolution as a three-pronged economic change. First, a small sector of the economy underwent quite rapid and dramatic technological change. Second, as a consequence, this sector grew at a rate much faster than the traditional sector so that its share in the overall economy continued to increase. Third, the technological changes in the modern sector gradually penetrated the membrane of the traditional sector so that parts of the traditional sector eventually became modernized. The economy grew, but because its sectoral composition changed, it did more than just increase in size, it was "growing-up".


In addition to the stormy developments in production technology, the British economy in the eighteenth century was subject to other, more gradual forces that affected the long-term growth of income. The most prominent of these forces were the growth of trade and the division of labor it brought with it. For Adam Smith, not surprisingly, the gains from trade and specialization were the main sources of economic growth. In the very long run, however, without continuous technological change, growth would slowly grind to a halt. The gains from trade and specialization, which in Smith's vision were the key to wealth, would have run into diminishing returns, as further declines in transportation or transactions costs would have yielded smaller and smaller marginal gains. Similarly, gains from improvements in the allocation of resources due to more effective economic institutions and the development of markets in factors and resources, eventually start yielding less and less as most of the easy gains are made early on. Changes in technology, that is, changes in human knowledge and ability to understand and utilize the laws of nature, is the the only dynamic element that seems thus far to be exempt from diminishing returns. (Talk more about the poulation technology dynamic here).

The consensus is that within the relatively narrow confines of production technology in a number of industries, more numerous and more radical inventions occurred during the Industrial Revolution than ever before in so short a period. It is equally uncontroversial that these changes had a farreaching effect on the lives of only a minority of Britons throughout our period. The Industrial Revolution was, above all, a regional affair, affecting Lancashire and parts of the adjoining counties and the Scottish Lowlands but leaving most of the rest of the country without visible marks. As late as 1851, only about 27 percent of the British labor force worked in the industries that were directly affected by the Industrial Revolution, although almost everyone had been touched by it indirectly as consumer, user, or spectator. Furthermore the Industrial Revolution coincided with the resumption of population growth in Britain, which until the middle of the eighteenth century had slackened off. There were ever more people who needed to be fed and clothed, threatening to materialize the dire predictions of the Reverend Malthus. The economic impact of population change was further complicated by the fact that it was in large part due to an increase in the birth rate. Like many underdeveloped countries today, this left Britain with an ever-younger population in which the proportion of small children who did not yet work was increasing.

A technological definition of the Industrial Revolution is a clustering of macroinventions leading to an acceleration in microinventions. The macroinventions not only increased productivity at the time but openedenough new technological vistas to assure that further change was forthcoming.The economic logic behind total factor productivity estimates is that output grows due to either increases in inputs or shifts of the production function (such as technological change). If the weighted contributions of the inputs are subtracted from the growth rates, the "residual" measures the rate of productivity growth, which is associated with innovation. Between 1760 and 1800, Crafts and Harley estimate, total factor productivity "explained" about 10 percent of total output growth; in the period 1801-1831 this went up to about 18 percent. This seems rather unimpressive, but it should be kept in mind that growth is concerned with output per worker (or per capita). If we look at output per worker, we observe that for the period 1760-1830 practically the entire growth of per capita income is explained by technological change. Economic growth was slow, as Harley and Crafts have shown, but what little there was is explained by the residual.

Even the most convinced detractors of the concept of the Industrial Revolution will concede two things. One is that although income per capita did not rise much between 1760 and 1830, it is hard to see how Britain could have sustained a more than doubling of its population while fighting a number of major wars had not its economic potential increased. Moreover, the undeniable sustained growth that occurred in the British economy after 1830 would not and could not have occurred without the changes in technology in the previous seventy years. Yet models of positive feedback have actually those characteristics, and it is easy to think of some historical processes in which increasing returns, induced technological change, frequency dependent processes, and the co-evolution of institutions and technology led to a spiral in which similar societies landed on quite different locations. Positive feedback can occur, for instance, when there are learning effects or under increasing returns (Arthur, 1994). In those cases technological change leads to lower prices, which could lead to the realization of scale economies, complementarities with other industries, demonstration effects, self-fulfilling expectations, bringing about even lower prices. Once the process had started, it fed on itself. Just as we have vicious circles in which backwardness breeds poverty and poverty breeds more backwardness, we have virtuous circles in which the reverse is true. If so, the role of contingency and accident in economic history may be far larger than people have supposed. In this approach, economic theory has to be complemented by insights from chaos theory: Comparatively minor differences in initial conditions can lead to major differences in historical outcomes. The Industrial Revolution in this interpretation was a "bifurcation point."


Douglass C. North (1981, pp. 147, 158-170) has argued that the British Industrial Revolution was facilitated by better-specified property rights, which led to more efficient economic organization in Britain. The link between property rights and economic growth consists of the greater efficiency in the allocation of resources resulting from the equalization of private and social rates of return and costs. Property rights in innovation (patents and trademarks), better courts and police protection, and the absence of confiscatory taxation are examples of how the same phenomenon could raise the rate of innovative activity and capital accumulation. North points out that well-specified property rights are not the same as laissez-faire. (((If output increased and technology possibly changed because of a rise in demand for industrial goods, it has to be made clear why demand increased in the first place.Changes in demand are not exogenous to an economic system -- they occur for well-understood reasons. Population, of course, began to increase rapidly after 1750, but this was a worldwide phenomenon and it seems far-fetched to link it directly to the Industrial Revolution. In a technologically static world, population growth (as the Classical School firmly believed) would lead to declining living standards. Hence, population growth in and of itself would increase the demand for food products more than the demand for manufactured goods, and the combination of growing population, bad harvests, and disruption of foreign supplies led to sharply higher agricultural prices, hardly a stimulus for industrial demand.Export demand, too, although of some importance in some industries, does not seem to have been the crucial element in the Industrial Revolution that some scholars have claimed.!!!)))


Still, this does not mean that demand played no role in generating technological change. Adam Smith himself noted that the division of labor was limited by the extent of the market and strongly believed that the division of labor itself was the main agent of technological progress. He thought that highly specialized workmen were more likely to come up with inventions.66 Innovation usually involved substantial fixed costs, and thus a minimum level of sales was expected by the innovator. In 1769 Matthew Boulton wrote to his partner James Watt, "It is not worth my while to manufacture your engine for three counties only, but I find it very well worth my while to make it for all the world" (cited by Scherer, 1984, p. 13). Some minimum level of demand was thus necessary to cover the fixed costs of research and development. An expansion of demand, through the integration of markets or through a growth in population and income or through an increase in export demand, could thus have stimulated invention.



(To be continued as time permits.............)

A different approach to the "demand hypothesis" has been proposed recently in a duo of papers by Jan de Vries (1993, 1994) in which he argues that changes in preferences could be of importance in explaining some of the economic changes in eighteenth-century Britain. De Vries argues, essentially, that the period was characterized by two distinct but related events: a supply-driven Industrial Revolution and a demand-driven set of changes in household behavior that he calls an "industrious revolution." The idea focuses on the household asa decision-making process: The household can allocate its resources to production for the market or to household production. In premodern Europe, as is still true today, the existence of household work makes the concept of leisure hard to define. De Vries points out that market purchases and household production areimperfect substitutes for each other: Child rearing, food preparation, apparel making, and personal services can be purchased or homemade, but the products are not identical. An increased preference for the consumption of purchased goods requires cash, however, and thus implies greater labor force participation and market orientation. The resources thus reallocated were not idle before, nor were they absorbed by leisure, strictly speaking; they were simply deployed differently. The allocation between household and market depends simultaneously on preferences and on the relative efficiency of the household in producing for its own consumption or for the market.

At least some of the sharp differences of opinion that arise between O'Brien-Engerman and their opponents result from different formulations of the question. Foreign trade was necessary if Britain was to import goods she could not produce for herself or could produce only at enormous cost. Tropical groceries (sugar, tobacco, spices, tea), European foodstuffs (wine, dried fish, corn in years of high prices), and raw materials (timber, hemp, high-quality ores, tar, and of course raw cotton) had to be brought in from overseas. O'Brien and Engerman (1991, pp. 201-202) point out that for this reason, in a closed economy Britain's real income would have been substantially lower, though it is hard to know precisely by how much without specifying what the next best substitutes were. The first difference between an open and a hypothetical closed economy was the "gain from trade," and it was of course large because trade occurred in large part with economies whose factor endowments were radically different from Britain's. Harley’s calculation is important in underscoring the dangers of indispensability theorems in economic history, but it is difficult to see how Britons could have produced the raw materials for their textile industry and the tea and sugar for their breakfast from domestic resources.

Was the growth of exports an "engine of growth" in the period of the Industrial Revolution? The question seems somewhat moot, given that there is a growing consensus that growth itself was comparatively modest before 1831. The intellectual resources that have been dedicated to explain British economic growth before 1830 by growing exports may have been misallocated now that it turns out that this growth was far less impressive than was hitherto supposed. One way of testing the relationship is to see whether domestic supplies grew faster than foreign demand, so that foreign demand was in this sense more a passive than an active factor. After 1800, when more data become available, we can be more certain that British supply increased faster than foreign demand, because Britain's net barter terms of trade worsened continuously. At the same time, it seems plausible that Britain's single factoral terms of trade (in which the prices are weighted by the productivity of domestic factors of production) improved, so that the purchasing power of the average Briton to buy imports continued to rise due to growing productivity. Manufacturing products, of course, were exported in large quantities, and taken together foreign markets would have been difficult to replace. The ratio of industrial exports to gross industrial output increased sharply, from 24.4 percent to 35.2 percent between 1700 and 1760, a period in which output was growing only slowly. What happened subsequently? Whereas Crafts’s figures suggest a sharp decline in the subsequent decades (with a sudden and unexplained peak in 1801), Cuenca’s new computations draw a rather different picture.Turning to the dynamic question, as already noted it is much more difficult to connect the openness of the British economy with technological changes. It is transparent that technological advances will stimulate exports. But is there a feedback from rising exports to further technological progress? As Krugman (1995, p. 55) has pointed out, once we consider such a feedback effect the answers become much more controversial. Export demand may have been a consideration for some innovators, but almost every individual entrepreneur could cover his expenses by the domestic market. The growing dependence of the cotton industry on foreign markets was an ex post phenomenon, not something that caused technological change. Ralph Davis argues that cotton expanded overseas after it had earned its spurs in the domestic market and that the export-driven expansion of the industry in the 1790s simply called for a larger number of similar mills (Davis, 1979, p. 67). All the same, the microinventions that kept improving the quality and reducing the prices of the goods produced may have been a function of output and thus of the size of the market. Learning by doing and experience were the sources of productivity increase after the big breakthroughs had been made. Insofar as export markets permitted expandedII Asian economies have often been "high-tech" and so a large export market may produce a stimulus to the adoption of frontier technologies. The unresolved questions remain, however: Is this connection between exports and technological progress also true for a nation that is generating the new technologies, and not only adopting them? To what extent would the domestic market have been able to replace the foreign markets? What was the elasticity of cost with respect to sales (that is, how strong, really, were the marginal learning effects of overseas
sales)?

Even if the nexus between foreign trade and technological progress thus remains something of a mystery, the open-ness of the British economy was a central feature that determined her economic fate. Open-ness is not a yes-or-no variable: few economies have ever been hermetically closed and few have been "entirely open" (if that concept could be defined). While open-ness was thus a matter of degree, this degree was of great importance. One example is the role of agriculture in the industrialization process. In a recent paper, Matsuyama (1992) demonstrates rigorously an intuition long prevalent among economic historians, namely that the relation between agricultural productivity and the rate of industrialization depends on the open-ness of the economy. In a closed economy, manufacturing depends on productivity growth in agriculture to produce a surplus that will permit the reallocation of resources from farming to industry and to provide a market for manufactured products. It has often been thought that an "agricultural revolution" was a necessary precondition for industrialization. Yet in an open economy this is clearly false: food can be imported and paid for by industrial goods. In fact, in an open economy a highly productive agricultural sector signals to the economy that its comparative advantage lies in farming, thus losing the (unforeseen) advantages of industrialization. This is in fact what happened in the Netherlands between 1815 and 1870: an open, free-trade economy with a highly productive agricultural sector, the opportunity costs of labor was just too high to render manufacturing profitable (Mokyr, 1976a). In Britain, despite growing agricultural productivity (the dimensions of which are still heavily disputed) this did not happen.

Imports from the Celtic Fringe and the Continent made up the British food deficit (Thomas, 1985). Indeed, Matsuyama's model implies that in an open economy the Industrial Revolution occurred not because but despite the growth in agricultural productivity. The open-ness of the British economy also meant that technology was continuously stimulated by ideas from the outside. We have already seen the wide influence of French science and inventions on British technology. Throughout the period, close cooperation with French, German, and Swiss manufacturers led to the continuous exchange of technological knowledge. Arnold Pacey (1990, pp. 117-120) has argued that Asian stimuli were of primary importance to the Industrial Revolution. Indian calicoes and muslins could not be made in Britain using the laborious hand-spinning techniques of India, but they showed the British what could be done, and eventually Crompton's mule was able to produce yarns of Asian fineness. English entrepreneurs sent representatives to Smyrna to study the manufacturing of Turkey-red dye, and plants to produce it were set up in Manchester and Glasgow (Wadsworth and Mann, 1931, pp. 180-181). Technology was enriched by the infusion of foreign elements, and in the long run this exposure effect turned out to be one of the most lasting benefits of the open economy.There are two competing and apparently incompatible views of the role of labor in the Industrial Revolution. One of them sees labor as a scarce resource, in fact as the scarce resource, and therefore the Industrial Revolution had a better chance of succeeding in areas in which it was abundant and cheap. The other regards technology as responding to labor scarcity and thus implies that scarce labor was an advantage in the industrialization race.

The first model is based on a number of assumptions that should be spelled out.82 Because the model is not strictly speaking a growth model (it has few implications for the overall growth rate of the economy) and deals more with the composition and technological practices of some sectors, I termed it a "growing-up" model (Mokyr, 1976b). The assumptions are as follows: 1. Capital goods "embodied" the new technology. Then, as now, that assumption seems almost tooobvious to justify. Steam engines, mule jennies, blast furnaces, paper mills, chaff cutters, and threshers are all examples of a new technology requiring a large capital expenditure. One cannot have the new technology without sales, then, they led to productivity increases and lower costs. Export-oriented industries in the post World-War making an investment in the equipment that embodied it. Above all, there were factories that had to be built, maintained, heated, lighted, and guarded. The modern sector was physically located, by and large, in large buildings. And in contrast with France and Belgium, in Britain there were no more monasteries to confiscate and convert. This is not to deny the importance of disembodied technological change. It implies, however, that a lack of fixed capital could have retarded the transformation, as I shall argue later. The reverse does not hold: an abundant supply of capital did not guarantee the adoption of technological changes and the emergence of factories, because the owners of the capital could not be relied upon to lend it to aspiring factory owners. What mattered was venture capital, not aggregate savings. The growing-up model is different from the standard growth models in that it is a disequilibrium model.

Its dynamics depend on the coexistence and interaction of the "old" and the "new" technologies. It applies to the European as well as to non-European contexts (Pomeranz, 1998). The traditional sector, which produces the same good (or a close substitute) as the factories, can continue its existence for a long time after the process has started, because the modern sector is still too small to supplant it altogether. As long as the two sectors coexist, the modern sector earns a "quasi-rent," a disequilibrium payment that will eventually disappear when the manual industries have disappeared. Through continuous reinvestment, this rent in its turn provides the fuel for further growth of the modern sector. This model suggests that high-wage economies would have lower profits, lower rates of accumulation, and thus a slower and later Industrial Revolution. The model also predicts that wages in the modern sector would grow slowly if at all as long as the traditional sector remained a large employer. In this sense, the model is comparable to the labor surplus models of Lewis and Fei-Ranis popular in the 1970s. In contrast to those models, however, the "growing-up" model does not have to make any deus ex machina assumptions about the wage rate. The modern sector is small enough relative to the rest of the economy to take the wage parametrically (that is, the sector can hire workers at a wage rate that is unaffected by the number of workers it employs) and hence the lower the wage set in the traditional economy, the faster the modern sectorcould grow.

The second approach to the role of labor in the Industrial Revolution, most closely associated with thework of H. J. Habakkuk (1962), maintains that inventive activity in the nineteenth century was mostly labor saving and that scarce labor thus stimulated waves of technical change. This approach is based on a somewhat peculiar view of technological change, namely, that innovation was a process of choice between more or less equivalent alternatives, similar to the choice made by a firm facing an isoquant. Although Habakkuk was primarily concerned with the period after 1830, his approach extends naturally to the British Industrial Revolution. High wages and labor-supply constraints in Britain, in this view, stimulated the demand for labor-saving technological change (Landes, 1969, pp. 57-60). Yet the application of the model, at second glance, is fraught with difficulties.

To start with, it is far from obvious that technological change during the Industrial Revolution was, on balance, more labor saving than capital saving: Von Tunzelmann (1981, p. 165) believes that, on balance, it was about neutral. MacLeod, examining the declared motives of eighteenth-century English patentees, found that only percent of them stated that "labor saving" was the main purpose of the invention. Further, it always makes good sense to "search" for labor-saving innovations, even in low-wage economies, because labor always costs something, and thus innovations that reduce labor inputs increase profits. This is especially the case if, as was likely true in low-wage areas, production was highly labor intensive. In addition, as David (1975) has pointed out, the Habakkuk view implies that technological change is "localized" (that is, occurs in close proximity to the techniques actually used rather than over the entire range of feasible techniques). For the Habakkuk view to prevail, such localized technological change has to be stronger in the capital-intensive range of techniques than in the labor-intensive range. In that case a high-wage economy will naturally have chosen a less labor-intensive technique and will experience faster technological progress as the unintended by-product of this choice. Finally, although British wages were higher than on the Continent, some scholars (e.g., Flinn, 1966, p. 31) have insisted that the growth of population met the increased demand for labor and that there is no evidence for any labor scarcity.Economists have examined the assumptions on which the two alternative theses are based and have made them explicit. The seemingly obvious test of the low-wage hypothesis is that the areas of Britain that industrialized earliest should, at the outset, have had lower wages. The relevant variable here is nominal wages, because we are interested in the cost of labor, not in the standard of living. In this regard, at least, the hypothesis seems confirmed. The areas of Britain that industrialized first, the northwest counties of Lancashire and the northern midlands, had lower wages than the South in the middle of the eighteenth century (Hunt, 1986). During the Industrial Revolution this relation was reversed, so that by 1867 the industrial areas had higher wages. Yet although this pattern is repeated in a few other instances, such as the Low Countries (Mokyr, 1976a), it is far from universal. Ireland, by all accounts, had low wages but did not industrialize. Britain itself had higher wages than most of the European continent. Clark shows that the high labor cost in the Atlantic economies (always excluding Ireland) was essentially offset by the higher productivity of workers in high-wage countries. Clark concludes that "real labor costs turn out to be as high as those in Britain in most of the other countries except for the very low wage competitors in Asia. The per worker wage rate tells us very little about the true cost of labor"

Labor could vary in its productivity for a variety of reasons. Differences in education seem to have made relatively little difference in productivity, as Mitch's chapter below points out. Another interpretation emphasizes diet: Low-wage workers could not buy enough food, and their malnourishment caused their work to be of low quality. Poorly paid workers could be poorly fed workers. The connection between caloric intake and energy output of workers is well known. Workers on an insufficient diet do not necessarily get sick or die, their entire metabolism simply slows down, to the detriment of their productivity (Scrimshaw, 1983).86 The dietary model is attractive, because the so-called efficiency-wage model seems quite promising in explaining the failure of premodern, poor societies to develop. Unfortunately, the evidence produced thus far to support this promising idea is ambiguous.87 Although recent scholarship has concluded that French workers were, in all likelihood, worse fed than British workers (Fogel, 1989, 1991), the same is not true for the Irish, whose potato diets assured them of a plentiful if somewhat monotonous fare (Mokyr, 1983). Productivity, however, depended on more than nutrition. Adam Smith thought that "the wages of labour are the encouragement of industry, which like every other quality, improves in proportion to the encouragement it receives. A plentiful subsistence increases the bodily strength of the laborer . . . where wages are high, accordingly, we shall always find the workmen more active, diligent, and expeditious, than where they are low." (Smith, [1776] 1976, p. 91). What Smith seems to be describing, however, is an upward sloping supply curve of labor, which makes people work more if the wage is higher. The question is, however, what makes people work better or harder per unit of time?

Recent thinking about the efficiency-wage hypothesis has shown that labor productivity can depend on the real wage paid to workers in a variety of ways. A simple model of this type is the shirking model, in which it is expensive to monitor the effort the worker puts in. High wages are a mechanism by which the employer extracts more effort from the worker, because a worker caught shirking risks being fired and losing his or her high-paying job. High wages could also increase productivity through reduced turnover. Another model derives a correlation between productivity and wages through an "adverse selection" mechanism: the worst-quality workers agree to work for less (see Akerlof and Yellen, 1986; Weiss, 1990).88 Differences in productivity in the early stages of the Industrial Revolution were also likely to arise from differences in workers' attitudes. Concentrating large numbers of workers (of both sexes) in one room and subjecting them to discipline, regularity, and the increasing monotony of the more advanced technique were some of the most difficult problems encountered by early factory masters (Thompson, 1967). Cheap labor was no advantage unless it could be effectively transplanted from the traditional to the modern sector. Sidney Pollard (1965, chap. 5) has pointed to the central paradox of the labor-supply question during the Industrial Revolution: "The lack of employment opportunities . . . existing simultaneously with a labor shortage is in part explained by the fact that the worker was averse to taking up the type of employment being offered, and the employer was unwilling to tolerate the habits of work which the men seeking work desired" (p. 196). How a rural, mostly self-employed labor force was enticed to work in mostly urban mills is one of thein the literature produced by economists. One answer given, ironically, by the social historian

Perkin is purely economic: "By and large, it was the prospect of higher wages which was the most effective means of overcoming the natural dislike for the monotony and quasi-imprisonment of the factory" (Perkin, 1969, p. 130). Pollard (1965) and Thompson (1967) suggest a variety of alternative ways in which the factory owners educated their workers in their own image, trying to imbue them with an ethic that made them more docile and diligent. Punctuality, respect for hierarchy, frugality, and temperance were the qualities that the value system tried to convey onto the younger generation. The factory owners used a combination of approaches; they relied first and foremost on semi compulsory apprenticed child labor from workhouses ("pauper apprentices") and on women driven out of their cottage industries by the rapid mechanization of spinning. Gradually, they created a more balanced labor force by a combination of higher pay and social control. An example is provided by the research of Huberman (1986; 1991; 1992; 1996). Huberman points out that although in the pre-1800 period the labor market in Lancashire worked in the classical fashion, with flexible wages equating supply and demand, employers soon found that they needed more than a labor force that was available. They needed a labor force that was loyal, reliable, and motivated. To insure this they paid wages that soon became institutionalized as "fair wages" and lost their flexibility. The emergence of such wage rigidity in some industries meant that when demand fluctuated, theWhat about immigration? In Ireland, where the collapse of domestic industry in the 1830s was swift and brutal, migration of workers to England and Scotland was widespread (Collins, 1981), and these immigrants were an important supplement to the British labor force during the Industrial Revolution (Redford, [1926] 1964, pp. 132-164). As Pollard (1978, p. 113) puts it, "[Irish emigrants] were, in many aspects, the mobile shock troops of the Industrial Revolution, whose role consisted in allowing the key areas to grow without distorting the labor market unduly." Recently, Williamson (1986, 1990a) has questioned the importance of the Irish workers to British industrialization. His calculations assume that the Irish formed an unskilled labor force and that agriculture was more unskilled-labor-intensive than manufacturing. Consequently, he finds that the main impact of Irish immigration was on agricultural output. Although most Irish ended up in rural areas, Williamson points out that their arrival slowed down the migration of British rural workers from the countryside to the cities. It is possible to argue that further disaggregation could overturn this conclusion in some industries. The Irish tended to concentrate in certain sectors and industries, such as mining, construction, and transportation, and in these industries their labor may well have contributed more than Williamson's aggregate computations suggest.95 On the whole, however, there is little reason to doubt Williamson's conclusion, simply because the number of Irish in Britain, though considerable, was simply not large enough to make a decisive impact on Britain's economy. In 1841 it is estimated that there were 830,000 "effective Irish" in Britain, of whom 415,000 were Irish born and the rest descendants of Irish emigrants. If we assume that all the emigrants and half of the others were in the labor force, the Irish would have added 620,000 workers, which out of a total occupied labor force of about 6.8 million would have amounted to about 9 percent; not a trivial addition, but not large enough to change the parameters dramatically.

Besides the question of the reallocation of labor from the traditional to the modern sector, there are many other loose ends to consider in the area of labor supply during the Industrial Revolution. One question is what happened to participation rates. We know little about these rates for the eighteenth century, and scholars have used population growth rates as a proxy for labor force growth rates. After 1801 the census provides figures for total occupied population that allow us to compute some very approximate participation rates. For what it is worth, the participation rate shows an initial decline from 1801 to 1831 and then rises until 1851 (Deane and Cole, 1969, pp. 8, 143). These changes are small and reflect primarily the changing age structure and measurement error. The concept of a participation rate is in any case something of an anachronism, because it requires alarge if declining percentage of the labor force was economically active in households (farms or workshops), this is not an unequivocal measure even if we had better data. It is thought that the Industrial Revolution mobilized a large part of its labor force by turning part-time workers into full-time workers and transferring workers from "disguised unemployment" to regular work (Pollard, 1978). What is clear is that in many of the more dynamic industries of the Industrial Revolution, including cotton, female and child labor predominated quantitatively in the early stages of the Industrial Revolution. What we do not know with enough precision is how many of these workers were drawn into the labor force altogether and how many were already active either in domestic manufacturing or agriculture. At some point after 1815 this reliance on non-adult-male labor began to decline, and by 1860 it was significantly lower than at the start of the century (Horrell and Humphries, 1995a, 1995b).

As the cottage industries had gone into decline, participation rates, however defined, must have been significantly lower already by the middle of the nineteenth century than at their peak during the heyday of the Industrial Revolution. The role of capital is not less controversial than that of labor. Recent work has concentrated on three issues. The first is the question of how capital markets worked during the Industrial Revolution and what effect they had on the process of technological change and accumulation. The other two issues have been raised primarily by economists, namely, the speed at which capital accumulated and the changes in its composition (circulating vs. fixed). On the issue of how capital markets worked, Larry Neal (1990) has recently pointed out that in the eighteenth century there was in fact an international capital market that funneled funds between different countries and that was clearly integrated, except when disrupted by war. There are also signs that British internal markets improved their operations during the Industrial Revolution: Buchinsky and Polak (1993) find that after 1770 there was a growing correlation between London interest rates and Yorkshire property transactions, though they find no sign of integration before that. Hoppit (1986) has reached a similar conclusion based on bankruptcy statistics. Although it would be premature to speak of a well-integrated capital market by 1800, clearly the capital market was becoming larger, more efficient, and more "modern" during the years of the Industrial Revolution.

Whereas the role of capital markets in the British economy as a whole is indisputable, their importance to the Industrial Revolution, properly speaking, is more difficult to assess. The biggest borrowers in Europe in this period were governments that needed to finance deficits. The demand for credit also came from merchants with bills to be discounted, entrepreneurs active in canal and road construction, landowners in need of funds for the purpose of enclosure and other improvements, and construction interests. There was some inevitable overlap between these borrowers and what we would consider the "modern sector," but it was relatively small. Moreover, the smallness of the modern sector relative to the entire British economy meant that its demand for loanable funds did not loom large relative to the needs of the economy. Dealing with the supply of savings on an aggregate level, however, is even more misleading than an aggregate analysis of labor markets. Such an analysis assumes the existence of a capital market that allocated funds to all competing users, presumably on the basis of an expected rate of return and riskiness. Certain developments, especially the growth of transport networks, would have been slowed down considerably, and possibly aborted, had it not been for capital markets. As far as the manufacturing sector is concerned, however, matters are quite complex. How did the Financial Revolution, which preceded the Industrial Revolution, affect it? The standard view of the interaction of "the two revolutions" has been that they had very little to do with each other. Postan argued in 1935 that "within industry almost every enterprise was restricted to its own supplies. The Industrial Revolution got under way while capital was not yet capable of moving between 'alternative employments'" (1935, p. 74). This view is now recognized as too simple: Financial markets were far more complex and subtle and their impact more pervasive than the earlier writers assumed. Yet there is little evidence that these financial markets were instrumental in helping modern industry more than vice versa.

Regarding the supply of capital, the most thorough work has been carried out by François Crouzet (1965, 1972, 1985b), complemented for the later period by Cottrell (1980). This work demonstrates that the capital needs of the modern sector during the Industrial Revolution were met from three sources. First were the internal sources in which the investor borrowed, so to speak, from himself using his private wealth (or that of his family) for start-up and plowing his profits back into the firm. Second, there were informal, or "personal," capital markets in which borrowers turned to friends, relatives, or partners for funds. Third, there was the formal capital market in which the borrower and the lender did not meet and in which attorneys, brokers, and eventually financial institutions (banks, insurance companies, stock markets) fulfilled their classic functions of intermediating between lenders and borrowers, concentrating information, and diversifying portfolios. The questions we must ask are, how important were these three forms of finance in the Industrial Revolution? and how can we explain this complex and seemingly inefficient mechanism? Students of the Industrial Revolution agree that most industrial fixed capital originated from internal finance. Crouzet (1965) concludes that "the capital which made possible the creation of large scale 'factory' industries came . . . mainly from industry itself . . . the simple answer to this question how industrial expansion was financed is the overwhelming predominance of self-finance"

In the early stages of the Industrial Revolution, the fixed-cost requirements to set up a minimum-sized firm were modest and could be financed from profits accumulated at the artisan level (Crouzet, 1965, p. 165; Pollard, 1964). Plow-back then provided a regular, almost automatic mechanism by which profits augmented the capital stock. As technology became more sophisticated after 1830, the initial capital outlays increased, and it became increasingly difficult to rely on internal finance to start a business. For railroads this was of course out of the question. For existing industrial firms, retained profits usually remained central to the accumulation of capital. Even in a world in which firms relied exclusively on retained earnings, an intersectoral capital market could function. Individuals who made their fortunes in commerce, real estate, or the slave trade could use these funds to diversify into manufacturing. There were examples of merchant princes entering modern manufacturing, such as the case of Kirkman Finlay, an overseas merchant who entered cotton spinning between 1798 and 1806, and the Wilson brothers who established the Wilsontown ironworks. On the whole, however, these cases were exceptional (Crouzet, 1985a, pp. 99-100).

The second source of funds, the informal capital market, can easily be illustrated with examples, but it is not known how important this form of finance was relative to other sources. Postan (1935) argues that capital was still a very personal thing, which most people wanted to keep under control. If one lent it out, it was only to an intimate acquaintance or to the government. Even partnerships, which were frequently resorted to in order to raise capital while avoiding the costly process of forming a joint-stock company, were usually closely tied to family firms. The taking in of strangers as sleeping partners merely for the sake of getting access to their wealth was relatively rare at first (Heaton, 1937, p. 89). This caution slowly dissipated during the Industrial Revolution, but active partners often bought out the others, and the advantages of partnership were as much in the division of labor as in the opportunity to raise credit. Many of the most famous characters in the Industrial Revolution had to resort to personal connections to mobilize funds. Richard Arkwright got his first loan from a politician friend, and James Watt borrowed funds from, among others, his friend and mentor, Dr. Joseph Black. Although the phenomenon was thus widespread (Crouzet, l965, p. 184; Mathias, 1969, pp. 150, 162-163), personal loans are as much of interest as a symptom of how the system operated as for the fact that they were a major channel through which funds were mobilized. Crouzet points out how exclusive and selective these personalized credit markets were: To have access to these informal networks one needed to be a member of them and be "known and well thought of in the local community" (1985a, p. 96). The market for capital can thus be seen once again to have depended on the market for information. As the modern sector grew, intrasectoral flows of funds between firms became more important, especially flows occurring within the same industry. Insofar as these mechanisms only reallocated funds among different industries in the modern sector, the upper bound that the rate of profit imposed on the rate of growth of that sector did not disappear. Instead of constraining the individual firm, the supply of funds now constrained the modern sector as a whole. Although there were important exceptions, by and large the modern sector pulled itself up by its own bootstraps.

The third mechanism for obtaining capital, the formal credit market, operated primarily through merchants, wholesalers, and country banks.100 The consensus on the role of the banks is that, with some exceptions, they rarely figured in the financing of long-term investment. Their importance was mainly in satisfying the need for working capital, primarily by discounting short-term bills and providing overdrafts (Flinn, 1966, p. 53; Pressnell, 1956, p. 326). Pollard has made a case for the reexamination of the importance of the banks on these grounds. Given that banks provided much short-term credit, firms short of capital could use all their internal funds on fixed investment (Crouzet, 1965, p. 193; Pollard, 1964, p. 155). Pollard, however, assumes that fixed capital grew at a rate much lower than implied by Feinstein's figures. His own earlier estimates imply a rate of growth of fixed capital of 2.4 percent per annum between 1770 and 1815, whereas Feinstein's fixed capital estimates grew at 4.2 percent per annum in the same period (Feinstein, 1978, p. 74). In manufacturing and trade the discrepancy is larger; according to Feinstein, gross fixed capital formation grew between 1770 and 1815 at 6 percent per annum, as opposed to Pollard's 3.4 percent (Feinstein, 1978, p. 74; see also Feinstein and Pollard, 1988). Thus financial constraints on capital accumulation may have been more stringent than Pollard originally presumed because he underestimated the needs. Moreover, substituting fixed capital for circulating capital may have been less simple than he thought, because as industrial output increased, the demand for circulating capital grew as well. Feinstein shows that between 176O and 1830 fixed capital in industry and commerce increased from 5 percent of domestic reproducible capital to 18 percent, whereas circulating capital in industry and commerce increased from 6 percent to 7 percent in the same period. Was the activity of banks enough to finance an increase of 164 percent in working capital over seventy years? Cottrell (1980, p. 33) concludes cautiously that there are indications that industrial growth before 1870 may have been blunted by shortages of circulating capital. Honeyman (1983, pp. 167-168) maintains that small businessmen found banks unreliable, and that even for circulating capital, kinship and friendship groups were preferred. The difficulty in obtaining funds led to the selective weeding out of the industry of entrepreneurs of humble origins who did not have access to these informal sources of funding and thus failed to survive crises during which working capital was hard to obtain. From a different point of view, Cottrell speculates that short-lived firms had better access to formal capital markets than firms that survived. The sharp fluctuations in the financial sector dragged into bankruptcy many industrial firms, and this effect may result in an underestimate of the importance of the plow-back of profit as a source of investment, because the firms that left records would tend to be less dependent on external finance (Cottrell, 1980, pp. 35, 253-255). Yet it remains to be seen whether enough evidence can be produced to jeopardize the widely held belief in the predominance of internal financing in this period. worker to be able to declare himself or herself as either being in the labor force or not.

At least some of the sharp differences of opinion that arise between O'Brien-Engerman and their opponents result from different formulations of the question. Foreign trade was necessary if Britain was to import goods she could not produce for herself or could produce only at enormous cost. Tropical groceries (sugar, tobacco, spices, tea), European foodstuffs (wine, dried fish, corn in years of high prices), and raw materials (timber, hemp, high-quality ores, tar, and of course raw cotton) had to be brought in from overseas. O'Brien and Engerman (1991, pp. 201-202) point out that for this reason, in a closed economy Britain's real income would have been substantially lower, though it is hard to know precisely by how much without specifying what the next best substitutes were. The first difference between an open and a hypothetical closed economy was the "gain from trade," and it was of course large because trade occurred in large part with economies whose factor endowments were radically different from Britain's. Harley’s calculation is important in underscoring the dangers of indispensability theorems in economic history, but it is difficult to see how Britons could have produced the raw materials for their textile industry and the tea and sugar for their breakfast from domestic resources.

Was the growth of exports an "engine of growth" in the period of the Industrial Revolution? The question seems somewhat moot, given that there is a growing consensus that growth itself was comparatively modest before 1831. The intellectual resources that have been dedicated to explain British economic growth before 1830 by growing exports may have been misallocated now that it turns out that this growth was far less impressive than was hitherto supposed. One way of testing the relationship is to see whether domestic supplies grew faster than foreign demand, so that foreign demand was in this sense more a passive than an active factor. After 1800, when more data become available, we can be more certain that British supply increased faster than foreign demand, because Britain's net barter terms of trade worsened continuously. At the same time, it seems plausible that Britain's single factoral terms of trade (in which the prices are weighted by the productivity of domestic factors of production) improved, so that the purchasing power of the average Briton to buy imports continued to rise due to growing productivity. Manufacturing products, of course, were exported in large quantities, and taken together foreign markets would have been difficult to replace. The ratio of industrial exports to gross industrial output increased sharply, from 24.4 percent to 35.2 percent between 1700 and 1760, a period in which output was growing only slowly. What happened subsequently? Whereas Crafts’s figures suggest a sharp decline in the subsequent decades (with a sudden and unexplained peak in 1801), Cuenca’s new computations draw a rather different picture.Turning to the dynamic question, as already noted it is much more difficult to connect the openness of the British economy with technological changes. It is transparent that technological advances will stimulate exports. But is there a feedback from rising exports to further technological progress? As Krugman (1995, p. 55) has pointed out, once we consider such a feedback effect the answers become much more controversial. Export demand may have been a consideration for some innovators, but almost every individual entrepreneur could cover his expenses by the domestic market. The growing dependence of the cotton industry on foreign markets was an ex post phenomenon, not something that caused technological change. Ralph Davis argues that cotton expanded overseas after it had earned its spurs in the domestic market and that the export-driven expansion of
the industry in the 1790s simply called for a larger number of similar mills (Davis, 1979, p. 67). All the same, the microinventions that kept improving the quality and reducing the prices of the goods produced may have been a function of output and thus of the size of the market. Learning by doing and experience were the sources of
productivity increase after the big breakthroughs had been made. Insofar as export markets permitted expandedII Asian economies have often been "high-tech" and so a large export market may produce a stimulus to the adoption of frontier technologies. The unresolved questions remain, however: Is this connection between exports
and technological progress also true for a nation that is generating the new technologies, and not only adopting them? To what extent would the domestic market have been able to replace the foreign markets? What was the elasticity of cost with respect to sales (that is, how strong, really, were the marginal learning effects of overseas
sales)?

Even if the nexus between foreign trade and technological progress thus remains something of a mystery, the open-ness of the British economy was a central feature that determined her economic fate. Open-ness is not a yes-or-no variable: few economies have ever been hermetically closed and few have been "entirely open" (if that
concept could be defined). While open-ness was thus a matter of degree, this degree was of great importance. One example is the role of agriculture in the industrialization process. In a recent paper, Matsuyama (1992) demonstrates rigorously an intuition long prevalent among economic historians, namely that the relation between agricultural productivity and the rate of industrialization depends on the open-ness of the economy. In a closed economy, manufacturing depends on productivity growth in agriculture to produce a surplus that will permit the reallocation of resources from farming to industry and to provide a market for manufactured products. It has often been thought that an "agricultural revolution" was a necessary precondition for industrialization. Yet in an open economy this is clearly false: food can be imported and paid for by industrial goods. In fact, in an open economy a highly productive agricultural sector signals to the economy that its comparative advantage lies in farming, thus losing the (unforeseen) advantages of industrialization. This is in fact what happened in the Netherlands between 1815 and 1870: an open, free-trade economy with a highly productive agricultural sector, the opportunity costs of labor was just too high to render manufacturing profitable (Mokyr, 1976a). In Britain, despite growing agricultural productivity (the dimensions of which are still heavily disputed) this did not happen.

Imports from the Celtic Fringe and the Continent made up the British food deficit (Thomas, 1985). Indeed, Matsuyama's model implies that in an open economy the Industrial Revolution occurred not because but despite the growth in agricultural productivity. The open-ness of the British economy also meant that technology was continuously stimulated by ideas from the outside. We have already seen the wide influence of French science and inventions on British technology. Throughout the period, close cooperation with French, German, and Swiss manufacturers led to the continuous exchange of technological knowledge. Arnold Pacey (1990, pp. 117-120) has argued that Asian stimuli were of primary importance to the Industrial Revolution. Indian calicoes and muslins could not be made in Britain using the laborious hand-spinning techniques of India, but they showed the British what could be done, and eventually Crompton's mule was able to produce yarns of Asian fineness. English entrepreneurs sent representatives to Smyrna to study the manufacturing of Turkey-red dye, and plants to produce it were set up in Manchester and Glasgow (Wadsworth and Mann, 1931, pp. 180-181). Technology was enriched by the infusion of foreign elements, and in the long run this exposure effect turned out to be one of the most lasting
benefits of the open economy.There are two competing and apparently incompatible views of the role of labor in the Industrial Revolution. One of them sees labor as a scarce resource, in fact as the scarce resource, and therefore the Industrial Revolution had a better chance of succeeding in areas in which it was abundant and cheap. The other
regards technology as responding to labor scarcity and thus implies that scarce labor was an advantage in the industrialization race.

The first model is based on a number of assumptions that should be spelled out.82 Because the model is not strictly speaking a growth model (it has few implications for the overall growth rate of the economy) and deals more with the composition and technological practices of some sectors, I termed it a "growing-up" model (Mokyr, 1976b). The assumptions are as follows: 1. Capital goods "embodied" the new technology. Then, as now, that assumption seems almost tooobvious to justify. Steam engines, mule jennies, blast furnaces, paper mills, chaff cutters, and threshers are all examples of a new technology requiring a large capital expenditure. One cannot have the new technology without sales, then, they led to productivity increases and lower costs. Export-oriented industries in the post World-War making an investment in the equipment that embodied it. Above all, there were factories that had to be built, maintained, heated, lighted, and guarded. The modern sector was physically located, by and large, in large buildings. And in contrast with France and Belgium, in Britain there were no more monasteries to confiscate and convert. This is not to deny the importance of disembodied technological change. It implies, however, that a lack of fixed capital could have retarded the transformation, as I shall argue later. The reverse does not hold:
An abundant supply of capital did not guarantee the adoption of technological changes and the emergence of factories, because the owners of the capital could not be relied upon to lend it to aspiring factory owners. What mattered was venture capital, not aggregate savings. The growing-up model is different from the standard growth models in that it is a disequilibrium model.

Its dynamics depend on the coexistence and interaction of the "old" and the "new" technologies. It applies to the European as well as to non-European contexts (Pomeranz, 1998). The traditional sector, which produces the same good (or a close substitute) as the factories, can continue its existence for a long time after the process has
started, because the modern sector is still too small to supplant it altogether. As long as the two sectors coexist, the modern sector earns a "quasi-rent," a disequilibrium payment that will eventually disappear when the manual industries have disappeared. Through continuous reinvestment, this rent in its turn provides the fuel for further
growth of the modern sector. This model suggests that high-wage economies would have lower profits, lower rates of accumulation, and thus a slower and later Industrial Revolution. The model also predicts that wages in the modern sector would grow slowly if at all as long as the traditional sector remained a large employer. In this sense, the model is comparable to the labor surplus models of Lewis and Fei-Ranis popular in the 1970s. In contrast to those models, however, the "growing-up" model does not have to make any deus ex machina assumptions about the wage rate. The modern sector is small enough relative to the rest of the economy to take the wage parametrically (that is, the sector can hire workers at a wage rate that is unaffected by the number of workers it employs) and hence the lower the wage set in the traditional economy, the faster the modern sectorcould grow.

The second approach to the role of labor in the Industrial Revolution, most closely associated with thework of H. J. Habakkuk (1962), maintains that inventive activity in the nineteenth century was mostly labor saving and that scarce labor thus stimulated waves of technical change. This approach is based on a somewhat peculiar view of technological change, namely, that innovation was a process of choice between more or less equivalent alternatives, similar to the choice made by a firm facing an isoquant. Although Habakkuk was primarily concerned with the period after 1830, his approach extends naturally to the British Industrial Revolution. High wages and labor-supply constraints in Britain, in this view, stimulated the demand for labor-saving technological change (Landes, 1969, pp. 57-60). Yet the application of the model, at second glance, is fraught with difficulties.

To start with, it is far from obvious that technological change during the Industrial Revolution was, on balance, more labor saving than capital saving: Von Tunzelmann (1981, p. 165) believes that, on balance, it was about neutral. MacLeod, examining the declared motives of eighteenth-century English patentees, found that only percent of them stated that "labor saving" was the main purpose of the invention. Further, it always makes good sense to "search" for labor-saving innovations, even in low-wage economies, because labor always costs something, and thus innovations that reduce labor inputs increase profits. This is especially the case if, as was likely true in low-wage areas, production was highly labor intensive. In addition, as David (1975) has pointed out, the Habakkuk view implies that technological change is "localized" (that is, occurs in close proximity to the techniques actually used rather than over the entire range of feasible techniques). For the Habakkuk view to prevail, such localized technological change has to be stronger in the capital-intensive range of techniques than in the labor-intensive range. In that case a high-wage economy will naturally have chosen a less labor-intensive technique and will experience faster technological progress as the unintended by-product of this choice. Finally, although British wages were higher than on the Continent, some scholars (e.g., Flinn, 1966, p. 31) have insisted that the growth of population met the increased demand for labor and that there is no evidence for any labor scarcity.Economists have examined the assumptions on which the two alternative theses are based and have made them explicit. The seemingly obvious test of the low-wage hypothesis is that the areas of Britain that industrialized earliest should, at the outset, have had lower wages. The relevant variable here is nominal wages, because we are interested in the cost of labor, not in the standard of living. In this regard, at least, the hypothesis seems confirmed. The areas of Britain that industrialized first, the northwest counties of Lancashire and the northern midlands, had lower wages than the South in the middle of the eighteenth century (Hunt, 1986). During the Industrial Revolution this relation was reversed, so that by 1867 the industrial areas had higher wages. Yet although this pattern is repeated in a few other instances, such as the Low Countries (Mokyr, 1976a), it is far from universal. Ireland, by all accounts, had low wages but did not industrialize. Britain itself had higher wages than most of the European continent. Clark shows that the high labor cost in the Atlantic economies (always excluding Ireland) was essentially offset by the higher productivity of workers in high-wage countries. Clark concludes that "real labor costs turn out to be as high as those in Britain in most of the other countries except for the very low wage competitors in Asia. The per worker wage rate tells us very little about the true cost of labor"

Labor could vary in its productivity for a variety of reasons. Differences in education seem to have made relatively little difference in productivity, as Mitch's chapter below points out. Another interpretation emphasizes diet: Low-wage workers could not buy enough food, and their malnourishment caused their work to be of low quality. Poorly paid workers could be poorly fed workers. The connection between caloric intake and energy output of workers is well known. Workers on an insufficient diet do not necessarily get sick or die, their entire metabolism simply slows down, to the detriment of their productivity (Scrimshaw, 1983).86 The dietary model is attractive, because the so-called efficiency-wage model seems quite promising in explaining the failure of premodern, poor societies to develop. Unfortunately, the evidence produced thus far to support this promising idea is ambiguous.87 Although recent scholarship has concluded that French workers were, in all likelihood, worse fed than British workers (Fogel, 1989, 1991), the same is not true for the Irish, whose potato diets assured them of a plentiful if somewhat monotonous fare (Mokyr, 1983). Productivity, however, depended on more than nutrition. Adam Smith thought that "the wages of labour are the encouragement of industry, which like every other quality, improves in proportion to the encouragement it receives. A plentiful subsistence increases the bodily strength of the laborer . . . where wages are high, accordingly, we shall always find the workmen more active, diligent, and expeditious, than where they are low." (Smith, [1776] 1976, p. 91). What Smith seems to be describing, however, is an upward sloping supply curve of labor, which makes people work more if the wage is higher. The question is, however, what makes people work better or harder per unit of time?

Recent thinking about the efficiency-wage hypothesis has shown that labor productivity can depend on the real wage paid to workers in a variety of ways. A simple model of this type is the shirking model, in which it is expensive to monitor the effort the worker puts in. High wages are a mechanism by which the employer extracts more effort from the worker, because a worker caught shirking risks being fired and losing his or her high-paying job. High wages could also increase productivity through reduced turnover. Another model derives a correlation between productivity and wages through an "adverse selection" mechanism: the worst-quality workers agree to work for less (see Akerlof and Yellen, 1986; Weiss, 1990).88 Differences in productivity in the early stages of the Industrial Revolution were also likely to arise from differences in workers' attitudes. Concentrating large numbers of workers (of both sexes) in one room and subjecting them to discipline, regularity, and the increasing monotony of the more advanced technique were some of the most difficult problems encountered by early factory masters (Thompson, 1967). Cheap labor was no advantage unless it could be effectively transplanted from the traditional to the modern sector. Sidney Pollard (1965, chap. 5) has pointed to the central paradox of the labor-supply question during the Industrial Revolution: "The lack of employment opportunities . . . existing simultaneously with a labor shortage is in part explained by the fact that the worker was averse to taking up the type of employment being offered, and the employer was unwilling to tolerate the habits of work which the men seeking work desired" (p. 196). How a rural, mostly self-employed labor force was enticed to work in mostly urban mills is one of thein the literature produced by economists. One answer given, ironically, by the social historian

Perkin is purely economic: "By and large, it was the prospect of higher wages which was the most effective means of overcoming the natural dislike for the monotony and quasi-imprisonment of the factory" (Perkin, 1969, p. 130). Pollard (1965) and Thompson (1967) suggest a variety of alternative ways in which the factory owners educated their workers in their own image, trying to imbue them with an ethic that made them more docile and diligent. Punctuality, respect for hierarchy, frugality, and temperance were the qualities that the value system tried to convey onto the younger generation. The factory owners used a combination of approaches; they relied first and foremost on
semi-compulsory apprenticed child labor from workhouses ("pauper apprentices") and on women driven out of their cottage industries by the rapid mechanization of spinning. Gradually, they created a more balanced labor force by a combination of higher pay and social control. An example is provided by the research of Huberman (1986; 1991; 1992; 1996). Huberman points out that although in the pre-1800 period the labor market in Lancashire worked in the classical fashion, with flexible wages equating supply and demand, employers soon found that they needed more than a labor force that was available. They needed a labor force that was loyal, reliable, and motivated. To insure this they paid wages that soon became institutionalized as "fair wages" and lost their flexibility. The emergence of such wage rigidity in some industries meant that when demand fluctuated, theWhat about immigration? In Ireland, where the collapse of domestic industry in the 1830s was swift and brutal, migration of workers to England and Scotland was widespread (Collins, 1981), and these immigrants were an important supplement to the British labor force during the Industrial Revolution (Redford, [1926] 1964, pp. 132-164). As Pollard (1978, p. 113) puts it, "[Irish emigrants] were, in many aspects, the mobile shock troops of the Industrial Revolution, whose role consisted in allowing the key areas to grow without distorting the labor market unduly." Recently, Williamson (1986, 1990a) has questioned the importance of the Irish workers to British industrialization. His calculations assume that the Irish formed an unskilled labor force and that agriculture was more unskilled-labor-intensive than manufacturing. Consequently, he finds that the main impact of Irish immigration was on agricultural output. Although most Irish ended up in rural areas, Williamson points out that their arrival slowed down the migration of British rural workers from the countryside to the cities. It is possible to argue that further disaggregation could overturn this conclusion in some industries. The Irish tended to concentrate in certain sectors and industries, such as mining, construction, and transportation, and in these industries their labor may well have contributed more than Williamson's aggregate computations suggest.95 On the whole, however, there is little reason to doubt Williamson's conclusion, simply because the number of Irish in Britain, though considerable, was simply not large enough to make a decisive impact on Britain's economy. In 1841 it is estimated that there were 830,000 "effective Irish" in Britain, of whom 415,000 were Irish born and the rest descendants of Irish emigrants. If we assume that all the emigrants and half of the others were in the labor force, the Irish would have added 620,000 workers, which out of a total occupied labor force of about 6.8
million would have amounted to about 9 percent; not a trivial addition, but not large enough to change the
parameters dramatically.

Besides the question of the reallocation of labor from the traditional to the modern sector, there are
many other loose ends to consider in the area of labor supply during the Industrial Revolution. One question is
what happened to participation rates. We know little about these rates for the eighteenth century, and scholars
have used population growth rates as a proxy for labor force growth rates. After 1801 the census provides figures
for total occupied population that allow us to compute some very approximate participation rates. For what it
is worth, the participation rate shows an initial decline from 1801 to 1831 and then rises until 1851 (Deane and
Cole, 1969, pp. 8, 143). These changes are small and reflect primarily the changing age structure and measurement
error.96 The concept of a participation rate is in any case something of an anachronism, because it requires alarge if declining percentage of the labor force was economically active in households (farms or workshops), this
is not an unequivocal measure even if we had better data. It is thought that the Industrial Revolution mobilized
a large part of its labor force by turning part-time workers into full-time workers and transferring workers from
"disguised unemployment" to regular work (Pollard, 1978). What is clear is that in many of the more dynamic
industries of the Industrial Revolution, including cotton, female and child labor predominated quantitatively in
the early stages of the Industrial Revolution. What we do not know with enough precision is how many of these
workers were drawn into the labor force altogether and how many were already active either in domestic
manufacturing or agriculture. At some point after 1815 this reliance on non-adult-male labor began to decline,
and by 1860 it was significantly lower than at the start of the century (Horrell and Humphries, 1995a, 1995b).
As the cottage industries had gone into decline, participation rates, however defined, must have been significantly
lower already by the middle of the nineteenth century than at their peak during the heyday of the Industrial
Revolution.
The role of capital is not less controversial than that of labor. Recent work has concentrated on three
issues. The first is the question of how capital markets worked during the Industrial Revolution and what effect
they had on the process of technological change and accumulation. The other two issues have been raised
primarily by economists, namely, the speed at which capital accumulated and the changes in its composition
(circulating vs. fixed). On the issue of how capital markets worked, Larry Neal (1990) has recently pointed out
that in the eighteenth century there was in fact an international capital market that funneled funds between
different countries and that was clearly integrated, except when disrupted by war. There are also signs that British
internal markets improved their operations during the Industrial Revolution: Buchinsky and Polak (1993) find
that after 1770 there was a growing correlation between London interest rates and Yorkshire property
transactions, though they find no sign of integration before that. Hoppit (1986) has reached a similar conclusion
based on bankruptcy statistics. Although it would be premature to speak of a well-integrated capital market by
1800, clearly the capital market was becoming larger, more efficient, and more "modern" during the years of the
Industrial Revolution.

Whereas the role of capital markets in the British economy as a whole is indisputable, their importance
to the Industrial Revolution, properly speaking, is more difficult to assess. The biggest borrowers in Europe in
this period were governments that needed to finance deficits. The demand for credit also came from merchants
with bills to be discounted, entrepreneurs active in canal and road construction, landowners in need of funds for
the purpose of enclosure and other improvements, and construction interests. There was some inevitable overlap
between these borrowers and what we would consider the "modern sector," but it was relatively small. Moreover,
the smallness of the modern sector relative to the entire British economy meant that its demand for loanable
funds did not loom large relative to the needs of the economy. Dealing with the supply of savings on an aggregate
level, however, is even more misleading than an aggregate analysis of labor markets. Such an analysis assumes
the existence of a capital market that allocated funds to all competing users, presumably on the basis of an
expected rate of return and riskiness. Certain developments, especially the growth of transport networks, would
have been slowed down considerably, and possibly aborted, had it not been for capital markets. As far as the
manufacturing sector is concerned, however, matters are quite complex.
How did the Financial Revolution, which preceded the Industrial Revolution, affect it? The standard
view of the interaction of "the two revolutions" has been that they had very little to do with each other. Postan
argued in 1935 that "within industry almost every enterprise was restricted to its own supplies. The Industrial
Revolution got under way while capital was not yet capable of moving between 'alternative employments'" (1935,
p. 74). This view is now recognized as too simple: Financial markets were far more complex and subtle and their
impact more pervasive than the earlier writers assumed. Yet there is little evidence that these financial markets
were instrumental in helping modern industry more than vice versa.

Regarding the supply of capital, the most thorough work has been carried out by François Crouzet (1965,
1972, 1985b), complemented for the later period by Cottrell (1980). This work demonstrates that the capital
needs of the modern sector during the Industrial Revolution were met from three sources. First were the internal
sources in which the investor borrowed, so to speak, from himself using his private wealth (or that of his family)
for start-up and plowing his profits back into the firm. Second, there were informal, or "personal," capital
markets in which borrowers turned to friends, relatives, or partners for funds. Third, there was the formal capital
market in which the borrower and the lender did not meet and in which attorneys, brokers, and eventually
financial institutions (banks, insurance companies, stock markets) fulfilled their classic functions of intermediating
between lenders and borrowers, concentrating information, and diversifying portfolios. The questions we must
ask are, how important were these three forms of finance in the Industrial Revolution? and how can we explain
this complex and seemingly inefficient mechanism? Students of the Industrial Revolution agree that most
industrial fixed capital originated from internal finance. Crouzet (1965) concludes that "the capital which made
possible the creation of large scale 'factory' industries came . . . mainly from industry itself . . . the simple answer
to this question how industrial expansion was financed is the overwhelming predominance of self-finance"
In the early stages of the Industrial Revolution, the fixed-cost requirements to set up a minimum-sized
firm were modest and could be financed from profits accumulated at the artisan level (Crouzet, 1965, p. 165;
Pollard, 1964). Plow-back then provided a regular, almost automatic mechanism by which profits augmented the
capital stock. As technology became more sophisticated after 1830, the initial capital outlays increased, and it
became increasingly difficult to rely on internal finance to start a business. For railroads this was of course out
of the question. For existing industrial firms, retained profits usually remained central to the accumulation of
capital. Even in a world in which firms relied exclusively on retained earnings, an intersectoral capital market
could function. Individuals who made their fortunes in commerce, real estate, or the slave trade could use these
funds to diversify into manufacturing. There were examples of merchant princes entering modern manufacturing,
such as the case of Kirkman Finlay, an overseas merchant who entered cotton spinning between 1798 and 1806,
and the Wilson brothers who established the Wilsontown ironworks. On the whole, however, these cases were
exceptional (Crouzet, 1985a, pp. 99-100).
The second source of funds, the informal capital market, can easily be illustrated with examples, but it
is not known how important this form of finance was relative to other sources. Postan (1935) argues that capital
was still a very personal thing, which most people wanted to keep under control. If one lent it out, it was only
to an intimate acquaintance or to the government. Even partnerships, which were frequently resorted to in order
to raise capital while avoiding the costly process of forming a joint-stock company, were usually closely tied to
family firms. The taking in of strangers as sleeping partners merely for the sake of getting access to their wealth
was relatively rare at first (Heaton, 1937, p. 89). This caution slowly dissipated during the Industrial Revolution,
but active partners often bought out the others, and the advantages of partnership were as much in the division
of labor as in the opportunity to raise credit. Many of the most famous characters in the Industrial Revolution
had to resort to personal connections to mobilize funds. Richard Arkwright got his first loan from a politician
friend, and James Watt borrowed funds from, among others, his friend and mentor, Dr. Joseph Black. Although
the phenomenon was thus widespread (Crouzet, l965, p. 184; Mathias, 1969, pp. 150, 162-163), personal loans
are as much of interest as a symptom of how the system operated as for the fact that they were a major channel
through which funds were mobilized. Crouzet points out how exclusive and selective these personalized credit
markets were: To have access to these informal networks one needed to be a member of them and be "known
and well thought of in the local community" (1985a, p. 96). The market for capital can thus be seen once again
to have depended on the market for information.
As the modern sector grew, intrasectoral flows of funds between firms became more important,
especially flows occurring within the same industry. Insofar as these mechanisms only reallocated funds among
different industries in the modern sector, the upper bound that the rate of profit imposed on the rate of growth
of that sector did not disappear. Instead of constraining the individual firm, the supply of funds now constrained
the modern sector as a whole. Although there were important exceptions, by and large the modern sector pulled
itself up by its own bootstraps.

The third mechanism for obtaining capital, the formal credit market, operated primarily through
merchants, wholesalers, and country banks.100 The consensus on the role of the banks is that, with some
exceptions, they rarely figured in the financing of long-term investment. Their importance was mainly in satisfying the need for working capital, primarily by discounting short-term bills and providing overdrafts (Flinn,
1966, p. 53; Pressnell, 1956, p. 326). Pollard has made a case for the reexamination of the importance of the
banks on these grounds. Given that banks provided much short-term credit, firms short of capital could use all
their internal funds on fixed investment (Crouzet, 1965, p. 193; Pollard, 1964, p. 155). Pollard, however, assumes
that fixed capital grew at a rate much lower than implied by Feinstein's figures. His own earlier estimates imply
a rate of growth of fixed capital of 2.4 percent per annum between 1770 and 1815, whereas Feinstein's fixed
capital estimates grew at 4.2 percent per annum in the same period (Feinstein, 1978, p. 74). In manufacturing
and trade the discrepancy is larger; according to Feinstein, gross fixed capital formation grew between 1770 and
1815 at 6 percent per annum, as opposed to Pollard's 3.4 percent (Feinstein, 1978, p. 74; see also Feinstein and
Pollard, 1988). Thus financial constraints on capital accumulation may have been more stringent than Pollard
originally presumed because he underestimated the needs. Moreover, substituting fixed capital for circulating
capital may have been less simple than he thought, because as industrial output increased, the demand for
circulating capital grew as well. Feinstein shows that between 176O and 1830 fixed capital in industry and
commerce increased from 5 percent of domestic reproducible capital to 18 percent, whereas circulating capital
in industry and commerce increased from 6 percent to 7 percent in the same period. Was the activity of banks
enough to finance an increase of 164 percent in working capital over seventy years? Cottrell (1980, p. 33)
concludes cautiously that there are indications that industrial growth before 1870 may have been blunted by
shortages of circulating capital. Honeyman (1983, pp. 167-168) maintains that small businessmen found banks
unreliable, and that even for circulating capital, kinship and friendship groups were preferred. The difficulty in
obtaining funds led to the selective weeding out of the industry of entrepreneurs of humble origins who did not
have access to these informal sources of funding and thus failed to survive crises during which working capital
was hard to obtain. From a different point of view, Cottrell speculates that short-lived firms had better access
to formal capital markets than firms that survived. The sharp fluctuations in the financial sector dragged into
bankruptcy many industrial firms, and this effect may result in an underestimate of the importance of the
plow-back of profit as a source of investment, because the firms that left records would tend to be less dependent
on external finance (Cottrell, 1980, pp. 35, 253-255). Yet it remains to be seen whether enough evidence can be
produced to jeopardize the widely held belief in the predominance of internal financing in this period.



worker to be able to declare himself or herself as either being in the labor force or not. In a society in which a
adjustments would take place through quantity adjustments: layoffs and short-time became commonplace.
most interesting questions in the debate on the Industrial Revolution, and yet it has not received much attention

Why THREE industrial revolutions?

Any reader of my China blog might well ask themself, why third industrial revolution? What were the other two? And why industrial, aren't we now in the information age? The purpose of this post will be to try and clarify a little what it is I am trying to say with this provocative expression. To start with the obvious, the first two industrial revolutions took place in the United Kingdom (at the end of the eighteenth century) and in the US (at the end of the nineteenth). Now in China we are seeing a process of transformation (funnily enough begining at the end of the twentieth century, but I'd say that's just a coincidence) which is likely to be every bit as significant and even dramatic in its consequences as its two predecessors. The reason I think the impact will be so significant is the sheer scale of the operation. Even if the value added per head might seem small in terms of international comparisons, the number of heads involved means that here is is a case of quantity means quality. And why industrial, is this not the information age. This is true, but China is only now really making the transition from a predominantly agricultural society to a highly industrialised one. If the US is the world's 'consumer of last resort', China is more than likely becoming the producer of first resort. But I am getting ahead of myself. Let us first go back to the end of the eighteenth century and ask the question: just what exactly was the first industrial revolution - or better put, in what did it consist?


In fact, far from being self-evident the very concept of industrial revolution itself has been subject to enormous scrutiny over the years. For some 'Industrial Revolution" is in fact "a concept too many." (eg Coleman 1983) on the grounds that that the term is either too vague to be of any use at all, or that it produces false connotations of abrupt change comparable in suddenness to political ones like the French Revolution. What, then, was it that changed in the years that we refer to as the first Industrial Revolution?

Mokyr differentiates between four categories of response. The Social Change idea: that the industrial revolution constituted first and foremost a change in the way economic transactions between people took place. The emergence of formal, competitive, and impersonal markets in goods and factors of production is the basis for this view. Karl Polanyi ([1944] 1985, p. 40) for example judges the emergence of the market economy as the truly fundamental event, with everything else having been incidental. Most modern social historians probably would view the central social changes as having to do with labor and the relation of workers with their work environment, other laborers, and employers. secondly there is the Industrial Organization idea: the emphasis here is on the structure and scale of the firm, in other words on the rise of wage labour employment and eventually the factory system. The focal point is the emergence of large firms, such as industrial mills, mines, railroads, and even large retail stores, in which production was managed and supervised and where workers were usually concentrated under one roof, subject to discipline and quality control. was unable for social reasons to adopt factories, the Industrial Revolution came late to it.Then there is the so-called Macroeconomic School: here the emphasis is on aggregate variables, such as the growth of national income, the rate of capital formation or the aggregate investment ratio, or the growth and composition of the labor force. Some writers in this school, such as Gerschenkron (1962), prefer to aggregate on a sectoral level, dealing with the rate of growth of the manufacturing sector rather than the growth of the entire economy. Most practitioners of what is known as the New Economic History tend to belong to this school, since by its very nature it tends to ask questions about large collections of individuals rather than about individuals (Fogel, 1983, p. 29) and because of its natural interest in quantitative analysis. Lastly there is the Technological idea: this considers changes in technology to be primary to all other changes and thus focuses on invention and the diffusion of new technical knowledge. Technology as used here, of course, means more than just "gadgets", It encompasses techniques used for the organization of labor, consumer manipulation, marketing and distribution techniques, and so forth. There is typically a long lag between the occurrence of changes in technology, even those of fundamental importance, and the time they start affecting aggregate statistics such as industrial production and national income per capita.


It is useful for present purposes to regard Britain during the period of the Industrial Revolution as a dual economy in which two economies coexisted although the argument would be no different if we considered a continuum of many sectors. One was the traditional economy, which, although not stagnant, developed gradually along conventional lines, with slow productivity and slowly rising capital-laborratios. This sector contained agriculture, construction, domestic industry, and many traditional "trades" that we would now classify as industrial but which in the eighteenth century and before were partially commercial: bakers, millers, tailors, shoemakers, hatters, blacksmiths, tanners, and other craftsmen. The modern sector consisted of cotton, iron smelting and refining, engineering, heavy chemicals, mining, some parts of transportation, and
some consumer goods such as pottery and paper. At first, however, only segments of these industries underwent modernization, so that dualism existed within as well as between various products, which makes calculations about the performance of the modern sector rather tricky. According to McCloskey's (1985) computations, the traditional economy was large, if relatively shrinking.

The average size of agriculture and "all others" between 1780 and 1860 was 79 percent of the British economy, meaning that in 1760 it was likely to have composed close to 90 percent of the British economy. Productivity growth in this sector is estimated by McCloskey at about 0.6 per cent per annum. During the same period productivity in the modern economy grew at a rate of 1.8 percent per annum.Two-sector growth models imply that abrupt changes in the economy as a whole are a mathematical impossibility because the aggregate rate of growth of any composite is a weighted average of the growth rates of its components, the weights being the respective shares in output. Even if changes in the modern sector itself were discontinuous and its growth rate very high, its small initial size would limit its impact on the economy-wide growth rate, and its share in the economy would increase gradually. In the long run the force of compound growth rates was such that the modern sector swallowed the entire economy. How long was the long run? A numerical example is illuminating here. Assume two sectors in a hypothetical economy, one of which (the modern sector) is growing at the rate of 4 percent per annum while the other (the traditional sector) is growing at the rate of 1 percent per annum.15 Suppose that initially the modern sector produces 10 percent of total output. Then the aggregate growth rate is at first 1.3 (=.9x1 + .1x4) percent. After ten years the aggregate rate of growth will have increased to 1.39 percent per year. After thirty years of "dual growth" the share of the modern sector will have increased to 21 percent of the economy and after fifty years to one-third. Only afterseventy-four years will the two sectors be of equal size (at which point aggregate growth equals 2.5 percent per year), and a full century after the starting point the traditional sector will have shrunk to about 31 percent of the economy. The British economy as a whole was changing much more slowly than its most dynamic parts, because growth was diluted by slow-growing sectors (Pollard, 1981, p. 39). These hypothetical numbers fit the actual record rather well, and they indicate that it is hardly surprising that it took until 1830 or 1840 for the economy wide effects of the Industrial Revolution to be felt.


Utilizing the distinction between a modern and a traditional sector allows us to summarize what happened to the British economy during the Industrial Revolution as a three-pronged economic change. First, a small sector of the economy underwent quite rapid and dramatic technological change. Second, as a consequence, this sector grew at a rate much faster than the traditional sector so that its share in the overall economy continued to increase. Third, the technological changes in the modern sector gradually penetrated the membrane of the traditional sector so that parts of the traditional sector eventually became modernized. The economy grew, but because its sectoral composition changed, it did more than just increase in size, it was "growing-up".


In addition to the stormy developments in production technology, the British economy in the eighteenth century was subject to other, more gradual forces that affected the long-term growth of income. The most prominent of these forces were the growth of trade and the division of labor it brought with it. For Adam Smith, not surprisingly, the gains from trade and specialization were the main sources of economic growth. In the very long run, however, without continuous technological change, growth would slowly grind to a halt. The gains from trade and specialization, which in Smith's vision were the key to wealth, would have run into diminishing returns, as further declines in transportation or transactions costs would have yielded smaller and smaller marginal gains. Similarly, gains from improvements in the allocation of resources due to more effective economic institutions and the development of markets in factors and resources, eventually start yielding less and less as most of the easy gains are made early on. Changes in technology, that is, changes in human knowledge and ability to understand and utilize the laws of nature, is the the only dynamic element that seems thus far to be exempt from diminishing returns. (Talk more about the poulation technology dynamic here).

The consensus is that within the relatively narrow confines of production technology in a number of industries, more numerous and more radical inventions occurred during the Industrial Revolution than ever before in so short a period. It is equally uncontroversial that these changes had a farreaching effect on the lives of only a minority of Britons throughout our period. The Industrial Revolution was, above all, a regional affair, affecting Lancashire and parts of the adjoining counties and the Scottish Lowlands but leaving most of the rest of the country without visible marks. As late as 1851, only about 27 percent of the British labor force worked in the industries that were directly affected by the Industrial Revolution, although almost everyone had been touched by it indirectly as consumer, user, or spectator. Furthermore the Industrial
Revolution coincided with the resumption of population growth in Britain, which until the middle of the eighteenth century had slackened off. There were ever more people who needed to be fed and clothed, threatening to materialize the dire predictions of the Reverend Malthus. The economic impact of population change was further complicated by the fact that it was in large part due to an increase in the birth rate. Like many underdeveloped countries today, this left Britain with an ever-younger population in which the proportion of small children who did not yet work was increasing.

A technological definition of the Industrial Revolution is a clustering of macroinventions leading to an acceleration in microinventions. The macroinventions not only increased productivity at the time but openedenough new technological vistas to assure that further change was forthcoming.The economic logic behind total factor productivity estimates is that output grows due to either increases in inputs or shifts of the production function (such as technological change). If the weighted contributions of the inputs are subtracted from the growth rates, the "residual" measures the rate of productivity growth, which is associated with innovation. Between 1760 and 1800, Crafts and Harley estimate, total factor productivity "explained" about 10 percent of total output growth; in the period 1801-1831 this went up to about 18 percent. This seems rather unimpressive, but it should be kept in mind that growth is concerned with output per worker (or per capita). If we look at output per worker, we observe that for the period 1760-1830 practically the entire growth of per capita income is explained by technological change. Economic growth was slow, as Harley and Crafts have shown, but what little there was is explained by the residual.

Even the most convinced detractors of the concept of the Industrial Revolution will concede two things. One is that although income per capita did not rise much between 1760 and 1830, it is hard to see how Britain could have sustained a more than doubling of its population while fighting a number of major wars had not its economic potential increased. Moreover, the undeniable sustained growth that occurred in the British economy after 1830 would not and could not have occurred without the changes in technology in the previous seventy years. Yet models of positive feedback have actually those characteristics, and it is easy to think of some historical processes in which increasing returns, induced technological change, frequency dependent processes, and the co-evolution of institutions and technology led to a spiral in which similar societies landed on quite different locations. Positive feedback can occur, for instance, when there are learning effects or under increasing returns (Arthur, 1994). In those cases technological change leads to lower prices, which could lead to the realization of scale economies, complementarities with other industries, demonstration effects, self-fulfilling expectations, bringing about even lower prices. Once the process had started, it fed on itself. Just as we have vicious circles in which backwardness breeds poverty and poverty breeds more backwardness, we have virtuous circles in which the reverse is true. If so, the role of contingency and accident in economic history may be far larger than people have supposed. In this approach, economic theory has to be complemented by insights from chaos theory: Comparatively minor differences in initial conditions can lead to major differences in historical outcomes. The Industrial Revolution in this interpretation was a "bifurcation point."


Douglass C. North (1981, pp. 147, 158-170) has argued that the British Industrial Revolution was facilitated by better-specified property rights, which led to more efficient economic organization in Britain. The link between property rights and economic growth consists of the greater efficiency in the allocation of resources resulting from the equalization of private and social rates of return and costs. Property rights in innovation (patents and trademarks), better courts and police protection, and the absence of confiscatory taxation are examples of how the same phenomenon could raise the rate of innovative activity and capital accumulation. North points out that well-specified property rights are not the same as laissez-faire. (((If output increased and technology possibly changed because of a rise in demand for industrial goods, it has to be made clear why demand increased in the first place.Changes in demand are not exogenous to an economic system -- they occur for well-understood reasons. Population, of course, began to increase rapidly after 1750, but this was a worldwide phenomenon and it seems far-fetched to link it directly to the Industrial Revolution. In a technologically static world, population growth (as the Classical School firmly believed) would lead to declining living standards. Hence, population growth in and of itself would increase the demand for food products more than the demand for manufactured goods, and the combination of growing population, bad harvests, and disruption of foreign supplies led to sharply higher agricultural prices, hardly a stimulus for industrial demand.Export demand, too, although of some importance in some industries, does not seem to have been the crucial element in the Industrial Revolution that some scholars have claimed.!!!)))


Still, this does not mean that demand played no role in generating technological change. Adam Smith himself noted that the division of labor was limited by the extent of the market and strongly believed that the division of labor itself was the main agent of technological progress. He thought that highly specialized workmen
were more likely to come up with inventions.66 Innovation usually involved substantial fixed costs, and thus a minimum level of sales was expected by the innovator. In 1769 Matthew Boulton wrote to his partner James Watt, "It is not worth my while to manufacture your engine for three counties only, but I find it very well worth my while to make it for all the world" (cited by Scherer, 1984, p. 13). Some minimum level of demand was thus necessary to cover the fixed costs of research and development. An expansion of demand, through the integration of markets or through a growth in population and income or through an increase in export demand, could thus have stimulated invention.



(To be continued as time permits.............)