Services constitute the greatest part of the wealthiest economies today. The efficiency with which they are supplied determines economic performance as a whole. Putting this proposition into historical perspective, Steve Broadberry investigates the nineteenth-century export success of British services and their subsequent supposed difficulty in adjusting to modern business enterprise. He is particularly concerned with the overtaking of Britain by the United States and Germany, which cannot be explained by the broadly unchanging comparative performance of industry. Industrialization of services is critical to understanding Britain's relative economic decline according to Broadberry. He maintains that British service firms were too slow in adapting to changes originating in the United States.
Market Services and the Productivity Race advances this argument in three sections. Chapters 2–4 establish the relative productivity facts over time, chapters 5–7 explain these facts thematically, and chapters 8–11 provide a more chronological narrative, with epochal breaks in 1914, 1950, and the 1990s. As with all Broadberry's work, the analysis utilizes economics in an intelligent and imaginative fashion to illuminate important historical questions, the data are carefully assembled and reconstituted on the basis of a thorough appreciation, and the context is provided by a very comprehensive review of the secondary and official sources. The approach is rigorously quantitative; chapter 8 with 50 tables only just beats chapter 10 with 48 tables, in second place, and chapter 9 comes in third, with 47.
Interpretation of the measurement section (chapters 2–4) could be facilitated by a little more explicit theory. As long as capital is freely available everywhere at a world price, and output is measured net, Broadberry's productivity index, output per worker, should reflect efficiency or genuine productivity. When output is measured gross, not taking into account possible differing costs of materials and energy, then international differences in product per worker may capture variations in input costs and appropriate substitutions, as well as efficiency. Perhaps this is one reason that Broadberry finds huge relative labor productivity gaps by sector between the United States and the United Kingdom in 1869/71. While U.S. manufacturing was 82 percent more productive than in the United Kingdom, U.S. utilities achieved little more than half of the output per person engaged of the U.K. sector (Table 4.1, p. 37).
Low physical productivity industries can obtain workers in competition with high physical productivity industries such as manufacturing by charging higher relative prices. Ideally, to take into account these relative prices, Broadberry's computational method is to consider one or more benchmark years for relative physical productivity, with value added at sector-specific purchasing power parities, adjusted in line with purchasing power parities, and then to extrapolate with time series. However the 1910 benchmark (Appendix 4.2), for instance, apparently uses only physical measures for utilities, transport, and communications outputs. A question is then whether the absence of prices biases some of the results.
Broadberry (p. 45) reports as generally accepted that at the end of the nineteenth century the price level in United States was higher than in the United Kingdom and that the U.S./U.K. real exchange rate was stationary over the century as a whole. If this is so, Balassa-Samuelson reasoning implies that aggregate productivity was also higher in the United States throughout the century—contrary to Broadberry's figures. One country's price level is higher than another's typically because it is more productive in traded goods. Greater physical productivity in traded goods is usually not matched in the nontraded sector—either because of more restricted competition, or for contingent technological reasons. But nontraded-sector workers must be paid sufficiently to attract them from traded-sector employment. Their wages then are higher in the country with the more productive traded sector, with the consequence that nontraded and aggregate prices are also higher in that country.
A real exchange rate can be defined in terms either of foreign and domestic prices, or of the traded and nontraded domestic prices. If the U.S./U.K. real exchange rate was stationary in the nineteenth century, at first sight there can have been no trend in their relative aggregate productivity growths (because their traded-nontraded price ratios were changing at the same rate). Hence higher U.S. productivity at the end of the nineteenth century must have been matched by higher productivity earlier. Greater U.S. aggregate productivity through most of the century is consistent with the pattern of international migration. Would 2.8 per 1,000 of the English (not the U.K.) population have migrated to a lower real wage economy in the Civil War decade?
That U.S. real wages and productivity were early much higher in one service industry is clearly brought out in a rare contemporary comparison. The Chief Engineer of the British Post Office investigated telegraphers in London and New York in 1877. He found productivity in New York, measured by telegrams per head, was twice that in London. Hours were 25 percent longer in the American city, but this was more than compensated by wages two and a half times greater than those paid in England.
The similar productivity growth rate implication of the stationary real exchange rate presents more of a problem. As Broadberry (p. 44) notes, accepted time series indicate a faster growth of productivity in the United States than in the United Kingdom during the later nineteenth century. Improvement of the terms of trade for an exporter need not be matched by deterioration in those for the corresponding importer if trade costs are falling, as they were in the century before 1914. This, then, may be why the U.S./U.K. real exchange rates remained untrended while U.S. productivity grew faster.
If the United States was already more productive than Britain when Broadberry's story begins, Britain's relative economic decline is less memorable than he would have the reader believe.
Broadberry contends Britain found it challenging to transform low volume, high margin services into high volume and low margin products. The required shift was from network organization to hierarchies. Networks are groups of persons who make independent decisions but share the output of the activity in proportions fixed according to their initial investment. Reward for their exertion includes future network participation, the return to which depends on agents' time discount factors. A hierarchy is where one agent becomes the principal, investing in monitoring technology, and dictating the actions of other agents, who in return accept a fixed wage. Venture types—standardized or customized—are distinguished by the information available, respectively venture-wide or agent-specific.
Standardized ventures are more suitable for hierarchies because the principal can obtain information for monitoring at a cost, whereas agents' information in customized ventures is not accessible. The availability of appropriate monitoring technology (such as the submarine telegraph for shipping) therefore influences whether a network or a hierarchy is a more efficient organization. To the extent that monitoring is a fixed cost, larger markets are conducive to hierarchy; costs per agent or transaction are lower than when the number of customers is smaller.
An oddity of the model for British market services is the democratic and egalitarian approach to economic activity in what was such a class-ridden, or class-blessed, society. Everybody is an “agent” and nobody a principal in the exposition. Broadberry deduces that low-education agents are more suitable for networks (result 4, p. 105). Yet the opposite could be the case; high relevant-education or knowledge workers were often necessary in networks. British apprenticeship was a network phenomenon as defined, with training delegated to the workforce. Information and intelligence were decentralized. Workers were self-educating in the venture. Management needs to know more in a hierarchy; in a network, information is privately managed.
A theme of Broadberry's earlier volume on manufacturing (The Productivity Race: British Manufacturing in International Perspective 1850–1990. Cambridge: Cambridge University Press, 1997) was that, in contrast to the United Kingdom, the U.S. system of production could not draw upon already trained or educated workers. The United States, therefore, developed an approach where the intelligence or the knowledge was in the system or hierarchy, rather than in the individuals. But managing this system required understanding it, a novel approach for British network “managers” and one to which they often experienced difficulty changing, perhaps because of their lack of relevant education. So Broadberry's result 4 may apply to management, but not to workers in the same way.
Broadberry believes that deep-seated institutional or other reasons ensured that the United Kingdom was more comfortable with customized, small-scale ventures, supplied in networks, rather than with hierarchies, even in the 1990s. Britain's improved relative performance he attributes then not to supply side reforms of the 1980s but to Information and Communication Technology (ICT) being better adapted to U.K. social capabilities (p. 357). Yet the productivity figures suggest ICT was even more appropriate for U.S. competencies, which he judges are rather different.
Readers need not agree with all Broadberry's conclusions to be thoroughly engaged by them. For they are all supported by a wealth of informative analysis and evidence that makes this an important book, both for students of the British economy and economic history, and for those concerned more generally with economic growth.
