Forging Ahead, Falling Behind and Fighting Back Read online

  To this extent, McCloskey’s position has been vindicated. Yet, it relies fundamentally on the proposition that stronger TFP growth was not possible. Moreover, McCloskey was writing before the advent of endogenous growth economics, which complicates the evaluation of British growth performance. On the one hand, an analysis of this kind might explain the unmatched acceleration of American TFP growth as unavoidable given an economic environment that was more conducive to innovative effort and tended to produce innovations that were unsuited to British conditions. On the other hand, it might point to policy interventions that could have raised British TFP growth but were not pursued.

  The estimates in Table 3.7 suggest that the United States had attained a higher TFP level than the United Kingdom by 1911 except in the service sector. The large TFP gap in manufacturing is especially striking, and at first sight this may seem to connote British failure. However, this is probably misleading since these estimates are impacted by the direct impacts of scale economies and natural resources and also by technologies that were developed in the American environment to exploit scale and cheap energy (Abramovitz and David, 2001) but were not appropriate for British conditions. Cain and Paterson (1986) found that from 1850 to 1920 technological change in American manufacturing generated economies of scale and entailed pervasive materials and capital using and labour saving biases. Moreover, the network of cumulative technological learning was essentially a national one at this time (Nelson and Wright, 1992).7 In the Second Industrial Revolution this underwrote clear American advantages in much cheaper electricity, which promoted the diffusion of electric motors and the associated transformation of American factories (David, 1991), and in mass production of cars, which was not viable in the much smaller (and more working-class) British market (Bowden, 1991).

  Table 3.7 USA/UK productivity levels in 1911 (UK = 100)

  Labour productivity TFP

  Agriculture 181 208

  Industry 206 161

  Manufacturing 214 185

  Services 119 79

  GDP 138 106

  Source: Broadberry (1998) revised in accordance with Woltjer (2013).

  In the early twentieth century, the United States had several obvious advantages that endogenous innovation theories might stress because they influenced the expected profitability of costly innovative effort. In the terminology of Figure 1.1, the United States was a relatively high-λ economy. These include a much larger domestic market and capital stock and a greater availability of engineers and science/technology graduates. Large markets encouraged independent inventors (Nicholas, 2010) and partly explain the higher R & D spending of firms in the United States, perhaps 0.25 per cent of GDP (in a much bigger economy) compared with 0.02 per cent in Britain (cf. Table 3.6). American factor endowments encouraged ‘directed technical change’, which was labour saving and materials using and was often not appropriate for use in British conditions but increased the transatlantic TFP and income gaps. Labour scarcity in the United States may even have increased the rate of technological change (Acemoglu, 2010). These arguments give further reasons to doubt the notion of an avoidable failure.

  Unlike later periods, allegations of government failure at this time are about errors of omission rather than commission. The British state continued to have very limited ambitions in terms of economic policy including support for innovation. Although small beginnings were made in promoting scientific research, for example, through the National Physical Laboratory (1899) and the Medical Research Council (1913), public expenditure on science and technology was only 0.06 per cent of GDP in 1914 (Pollard, 1989). This undoubtedly implied that there was too little government support for R & D, a pro-growth activity where social returns exceed private returns. An obvious contrast with the United States was the investment there by state governments in new universities with a greater emphasis on research, professional schools and industrial connections including, notably, MIT (founded in 1862) with its strengths in chemical and electrical engineering (Goldin and Katz, 1999). However, large-scale federal support for R & D in the United States had to await the Second World War (Mowery and Rosenberg, 2000), and the much more important differences between the two countries were in industrial R & D.

  3.4 The ‘Over-Commitment’ Hypothesis

  Critics of the pre-First World War British economy have frequently argued that while its industrial structure may have reflected comparative advantage in the short run it was not well equipped for long-run productivity growth and, in this regard, compared unfavourably with the United States. The ‘old’ industries of the First Industrial Revolution were too big and the new industries of the Second Industrial Revolution were too small – for example, too much cotton textiles and too few cars being produced. A very well-known version of this argument was made by Richardson (1965), who claimed that Britain suffered from ‘over-commitment’ to the old industries and that government should have intervened to speed up the transfer of resources to new industries.

  In essence, this is an ‘early-start’ hypothesis about British relative economic decline in which the market economy was ‘locked-in’ by its pattern of specialization in international trade to a structure that was suboptimal.8 It can also be seen as an argument for an ‘industrial policy’ of the kind which would be adopted in the 1960s and 1970s.9 In particular, Richardson (1965) suggested that the government should have abandoned free trade and implemented policies of tariff protection for ‘infant industries’ to facilitate a re-balancing of the economy. The United Kingdom was unusual at this time in its devotion to free trade, despite the presence of a vocal protectionist lobby, partly because the agricultural producer interest was so weak (O’Rourke, 1997), itself a reflection of the early start.

  It is certainly the case that Britain’s revealed comparative advantage (RCA) looked very different from that of the United States and indeed from what one might expect of an advanced economy of the second half of the twentieth century, as can be seen in Table 3.8. This is reflected, for example, in the rankings of textiles and cars and aircraft. The United Kingdom had a weak position in R & D intensive sectors and the rank correlation of RCA and TFP growth by sector in the United States over 1899–1937 was significantly negative (Thomas, 1988). Cotton textiles remained a very strong export with a world market share of about 80 per cent on average during 1880–1911 despite having wage rates which were six times Asian levels because unit costs were held down by the productivity benefits of the Lancashire agglomeration, which had accrued from the early start (Crafts and Leunig, 2005).

  Table 3.8 Revealed comparative advantage rankings

  UK 1913 UK 1937 USA 1913 USA 1937

  Agricultural equipment 10 16 2 1

  Cars and aircraft 12 11 4 2

  Industrial equipment 5 7 3 3

  Electricals 8 5 5 4

  Iron and steel 3 9 9 5

  Non-ferrous metals 16 15 1 6

  Book and film 13 8 10 7

  Chemicals 11 12 12 8

  Metal manufactures 7 13 6 9

  Brick and glass 14 10 11 10

  Wood and leather 15 14 7 11

  Rail and ship 1 3 8 12

  Fancy goods 9 4 13 13

  Apparel 6 6 14 14

  Alcohol and tobacco 4 1 15 15

  Textiles 2 2 16 16

  Note: Rankings are made on the basis of world market share with 1 = highest.

  Source: Crafts (1989).

  Nevertheless, this ‘over-committed’ structure did not entail a productivity growth penalty before the First World War. Table 3.9 reports a calculation of the difference that United Kingdom (rather than American) employment shares would have made to overall American labour productivity growth in manufacturing assuming that sectoral productivity levels and growth rates remained unchanged. The difference is trivial but actually goes in the direction of raising American productivity growth. If productivity growth was disappointing in British manufacturing in the Edwardian period, this was a problem of within-sector performance not compositio
n of activity.

  Table 3.9 Impact of changing sectoral weights on labour productivity growth in US manufacturing, 1899–1909 (% per year)

  USA weights 1.65

  UK weights 1.73

  Post-tariff UK weights 1.76

  Note: Calculations based on fixed 1900 weights in each case where the weight for sector i uses the formula in Nordhaus (1972) and is [(Y/L)i/(Y/L)manf]si where Y/L is labour productivity and s is the share of manufacturing employment. In each case the term in square brackets is based on productivity data for the United States and the employment shares are as follows: row 1 is actual United States in 1900, row 2 is actual United Kingdom in 1907 and row 3 is counterfactual United Kingdom with Chamberlain tariff in 1907.

  Sources: Derived from Kendrick (1961), Niemi (1974) and Thomas (1984).

  It is well-known that, although Britain maintained policies of free trade, the United States was a high-tariff country with an average tariff rate on manufactures of around 40 per cent. This does not mean, however, that American overtaking was the result of protectionism. Irwin (2001) pointed out that faster productivity growth in the United States mainly accrued in non-traded sectors (cf. Table 3.2) which did not benefit from protection and that American policies did not add up to a successful infant-industry strategy. More generally, careful econometric analysis of the cross-country evidence does not support the hypothesis that higher tariffs raised growth rates in this period – if anything it points to the opposite conclusion (Schularick and Solomou, 2011). This may reflect the tendency for protection not to be tightly focused on a few selected sectors with excellent growth prospects or positive externalities for the rest of the economy (Tena-Junguito, 2010).

  Similar problems would surely have undermined any British attempt to use protectionism to promote faster growth. The political economy of tariff protection was such that the proposals that had the most political support such as those made by Chamberlain would have actually tended to divert employment towards traditional sectors such as agriculture and textiles, which were relatively labour intensive rather than new growth industries (Thomas, 1984).10 A general tariff policy would have weakened competition in product markets with potentially adverse effects on productivity performance. In any case, if the real problem was market failures that implied too little investment in human capital and R & D, then the right response was policy intervention by government to address these failures directly.

  3.5 Institutional Trajectories

  As Chapter 2 highlighted, there were important institutional legacies of the Industrial Revolution, especially with regard to corporate governance and industrial relations, which had the potential to affect subsequent economic performance. At this point, it is opportune to review the state of play in these areas in the early twentieth century. The picture that emerges is that Britain had moved further down the paths which were already mapped out in the mid-nineteenth century.

  With regard to corporate governance, equity finance of joint-stock limited liability companies had been embraced on quite a wide scale despite the very weak levels of shareholder protection provided by company law. Recent research has established that many companies had diffuse share ownership and it is possible already to see a significant separation of ownership and control. For example, Acheson et al. (2015) found that in companies registered between 1881 and 1902 the median holding of the largest shareholder and the directors was 6.4 and 9.0 per cent of the capital, respectively. For very large companies in 1911, Foreman-Peck and Hannah (2012) found that the median value of shares held by the directors was only 2.4 per cent. That said, it was possible for control in voting rights to stay quite tightly held in many companies whose shares were traded on the stock exchange; for Initial Public Offering (IPO)s between 1900 and 1911, insiders on average controlled over 55 per cent of voting rights and dispersion of initial holdings was fairly slow (Cheffins et al., 2013).

  It seems clear that the principal–agent problems of the railways, which stemmed from very weak shareholders and barriers to entry, were not typical of this period. Foreman-Peck and Hannah (2013) found no evidence that diffuse shareholding in large companies was associated with under-performance in terms of profitability while managerial discipline together with shareholder protection seems to have been underpinned by high levels of dividend payouts (Campbell and Turner, 2011). Shares were still nearly always held by individuals, many of whom lived locally, rather than institutions while coalitions of a relatively small number of shareholders may have sufficed to provide an important check on managers.11 At the same time, the way was open for a future evolution towards a much more problematic separation of ownership and control later on.

  By the early twentieth century the British system of industrial relations had developed considerably from the Industrial Revolution era but retained distinctive features inherited from the mid-nineteenth century. Many more workers were members of trade unions – almost 25 per cent of the labour force in 1913 compared with 4 per cent as recently as 1890 – and collective bargaining over wages had become quite widespread – 2.4 million workers covered by such agreements in 1910 (Aldcroft and Oliver, 2000). By then trade unionism was spreading to unskilled workers, but its defining characteristic remained craft unionism. So in many important industries multi-unionism prevailed and the typical union represented a small subset of the workforce. Despite the founding of the Trade Union Congress (TUC) in 1868, industrial relations remained decentralized.

  The continued strength of craft unionism entailed craft control of work on the shop floor and the use of piece rates to elicit effort. This had emerged as a solution to appropriation problems during the Industrial Revolution and was still the preferred management style of many British employers, at least partly because they perceived the ‘switching costs’ of imposing greater direct managerial control as too high. There were periodic disputes over the details of these arrangements, including, most notably, the engineering lockout of 1897/1898, but, in the aftermath, employers did not seek to end craft control on this and other occasions when they ‘won’ the dispute.12 Arguably, these institutional arrangements served Britain well through the nineteenth century (Lazonick, 1994). When, however, the Fordist technologies of the Second Industrial Revolution came along with requirements for large sunk-cost investments in fixed capital, firms were inhibited from adopting them by exposure to ‘hold-up’ problems. Unions lacked an ability to commit to cooperative behaviour and the effort levels required, for example, by mass production methods in the car industry (Lewchuk, 1987).

  Together with the expansion of the franchise in 1867 and 1884, the value that employers placed on Victorian industrial relations informed the legal privileges given to trade unions in the 1906 Trade Disputes Act, which granted complete legal immunity to trade unions from all actions at tort and went beyond the already extensive immunities thought to have been granted in 1875 (Phelps Brown, 1983).13 This gave trade unions considerably enhanced bargaining power and remained essentially unreformed for over sixty years.

  The strength of craft unionism in Britain may also have had political consequences by precluding a move to proportional representation. Cusack et al. (2007, 2010) argue that its presence meant that there was no incentive to move away from a majoritarian system of voting in an economy where there was no prospect of gains for the right from collective institutions to deliver investment in skills. And, if winning elections meant wooing the median voter who was a skilled worker, in turn this underpinned craft unionism, as 1906 underlined. Thus, an equilibrium was sustained that mitigated against moving towards a coordinated market economy.

  3.6 Conclusions

  Overall, it seems reasonable to conclude that there was no massive growth failure in the pre-1914 economy. Any decline in the trend growth rate was slight and American outperformance probably had its roots in unique American advantages based on a favourable configuration of factor endowments and market size, rather than serious errors by British business or governments. That said, the
neoclassical exoneration of British performance epitomized in the title of the well-known survey article by McCloskey and Sandberg (1971) surely went too far, as is illustrated by the example of the railways. Arguably, also British governments might have done more to establish a stronger national innovation system.

  There is also no strong argument that the early start was a serious handicap on British growth performance in the decades before the First World War. Certainly, British comparative advantage was still concentrated in the ‘old’ nineteenth-century industries, but this structure was not yet a big disadvantage; American outperformance was intra-sectoral and in services. Early industrialization and the strength of London as a financial centre meant that Britain was an enthusiastic participant in the globalizing economy, but policies of openness served the economy well both in capital and product markets. The early start mitigated against protectionist policies, but this did not entail a pre-1914 growth penalty, although it increased the economy’s exposure to globalization risks and future adjustment problems.