《国际政治经济学》文献资料（Political Economy of Migration）Globalization, Convergence, and History.pdf

THE JOURNAL OF ECONOMIC HISTORY VOLUME 56 JUNE 1996 NUMBER 2 Globalization,Convergence,and History JEFFREY G.WILLIAMSON There were three epochs of growth experience after the mid-nineteenth century for what is now called the OECD"club":the late nineteenth century,the middle years between 1914 and 1950,and the late twentieth century.The first and last epochs were ones of overall fast growth,globalization,and convergence.The middle years were ones of overall slow growth,deglobalization,and divergence.Thus history offers an unambiguous positive correlation between globalization and convergence.When the pre-World War I years are examined in detail,the correlation turns out to be causal: globalization played the critical role in contributing to convergence. THEORY NEEDS HISTORY Wwo important features of the late twentieth-century international omy characterized the late nineteenth centuras elF earlier period was one of rapid globalization:capital and labor flowed across national frontiers in unprecedented quantities,and commodity trade boomed as transport costs declined sharply.Second,the late nineteenth century underwent an impressive convergence in living stan- dards,at least within most of what we would now call the OECD club.Poor countries at the periphery of the European club tended to grow faster than the rich industrial leaders at the center of the Old World and often even faster than the richer countries overseas in the New World.This club excluded,of course,most of the Third World and eastern Europe,and even around this limited periphery there were some who failed to catch up. But whereas Spain and Portugal lagged behind the leaders,others-like Ireland,Italy,and the Scandinavian countries-underwent a spectacular catch-up from the Great Famine to the Great War.To what extent were globalization and convergence connected? The Journal of Economic History,Vol.56,No.2(June 1996).The Economic History Association.All rights reserved.ISSN 0022-0507. Jeffrey G.Williamson is the Laird Bell Professor of Economics at Harvard University,Cambridge, MA02138. This article is an expanded version of the Presidential Address to the annual meeting of the Economic History Association,Chicago,8-10 September 1995. The research has been supported since 1990 by National Science Foundation grants SES-90-21951 and SBR-92-23002,for which I am grateful.I am also grateful for the excellent research assistance(and insightful comments)of Bill Collins.In addition,I want to thank my collaborators Tim Hatton,Kevin O'Rourke,and Alan Taylor with whom I have shared so many of the discoveries surveyed in this article.They,Moe Abramovitz,Don Davis,Mike Edelstein,Anne Krueger,Cormac O Grada,and Jeff Sachs have all offered advice that has greatly improved the final product. 277 17.15 PM

THE JOURNAL OF ECONOMIC HISTORY VOLUME 56 JUNE 1996 NUMBER 2 Globalization, Convergence, and History JEFFREY G. WILLIAMSON There were three epochs of growth experience after the mid-nineteenth century for what is now called the OECD "club": the late nineteenth century, the middle years between 1914 and 1950, and the late twentieth century. The first and last epochs were ones of overall fast growth, globalization, and convergence. The middle years were ones of overall slow growth, deglobalization, and divergence. Thus history offers an unambiguous positive correlation between globalization and convergence. When the pre-World War I years are examined in detail, the correlation turns out to be causal: globalization played the critical role in contributing to convergence. THEORY NEEDS HISTORY T wo important features of the late twentieth-century international economy characterized the late nineteenth century as well. First, the earlier period was one of rapid globalization: capital and labor flowed across national frontiers in unprecedented quantities, and commodity trade boomed as transport costs declined sharply. Second, the late nineteenth century underwent an impressive convergence in living standards, at least within most of what we would now call the OECD club. Poor countries at the periphery of the European club tended to grow faster than the rich industrial leaders at the center of the Old World and often even faster than the richer countries overseas in the New World. This club excluded, of course, most of the Third World and eastern Europe, and even around this limited periphery there were some who failed to catch up. But whereas Spain and Portugal lagged behind the leaders, others-like Ireland, Italy, and the Scandinavian countries-underwent a spectacular catch-up from the Great Famine to the Great War. To what extent were globalization and convergence connected? The Joumnal of Economic History, Vol. 56, No. 2 (June 1996). ? The Economic History Association. All rights reserved. ISSN 0022-0507. Jeffrey G. Williamson is the Laird Bell Professor of Economics at Harvard University, Cambridge, MA 02138. This article is an expanded version of the Presidential Address to the annual meeting of the Economic History Association, Chicago, 8-10 September 1995. The research has been supported since 1990 by National Science Foundation grants SES-90-21951 and SBR-92-23002, for which I am grateful. I am also grateful for the excellent research assistance (and insightful comments) of Bill Collins. In addition, I want to thank my collaborators Tim Hatton, Kevin O'Rourke, and Alan Taylor with whom I have shared so many of the discoveries surveyed in this article. They, Moe Abramovitz, Don Davis, Mike Edelstein, Anne Krueger, Cormac 6 Grada, and Jeff Sachs have all offered advice that has greatly improved the final product. 277 This content downloaded from 211.80.95.69 on Sun, 20 Oct 2013 19:57:15 PM All use subject to JSTOR Terms and Conditions

Globalization,Convergence,and History 279 CONVERGENCE IN THE PAST What does history have to say about convergence?To answer that question,we have to agree on the meaning of convergence.The critical bottom line for me is whether the living standard gap between rich and poor countries falls over time.Convergence implies an erosion in this gap, at least in percentage terms.New growth theorists call this sigma- convergence.To get sigma-convergence,poor countries must grow faster than rich,an event new growth theorists call beta-convergence.Second, what history?My interest has always been in what Simon Kuznets called modern economic growth,and that translates here into the century and a half since about 1850.Third,convergence of what?There appear to be two data sets that can be used for such analysis.Angus Maddison's GDP per worker-hour estimates offer one (originally published in 1982,now super- seded by his 1991 book and by even more recent revisions).My real wages of the urban unskilled offer another.2 Fourth,convergence among whom? As the introduction suggests,my net will only capture members of the present OECD club with European origin (plus Argentina and Brazil).3 All of us know that much of the convergence since 1870 disappears when the net is widened to include Eastern Europe,and if it were widened still further to include the Third World,convergence would totally evaporate. Why the small net?Because I think the sources of convergence in the OECD club are themselves misunderstood,and it matters to get the facts right. Figure 1 and Table 1 document real wage convergence from midcentury to the Great War.Although the convergence was not as fast as that of the late twentieth century,it was pronounced,and about as fast as average convergence over the full 150 years since 1850.Figure 2 and Table 2 show that the late nineteenth-century real wage convergence was replicated by gross domestic product (GDP)per worker-hour.However,real wage convergence was a lot faster than GDP per worker-hour,and the global- ization arguments that follow offer some reasons why.Table 3 reports the A's that emerge when so-called unconditional convergence equations are applied to this historical epoch.They imply a peak rate of real wage convergence between 1870 and 1890 of 1.2 percent per annum,and about 1 percent per annum over the 1870 to 1913 period as a whole. Although impressive,the late nineteenth-century rate of convergence Maddison,Phases,Dynamic Forces,and"Explaining the Economic Performance." 2These wages are purchasing-power-parity adjusted;see Williamson,"Evolution." 3 The full real wage 17-country sample is Australia,Argentina,Belgium,Brazil,Canada,Denmark, France,Germany,Great Britain,Ireland,Italy,the Netherlands,Norway,Portugal,Spain,Sweden, and the United States.The full(1991)GDP per worker-hour 15-country sample subtracts Argentina, Brazil,Ireland,Portugal,and Spain,but adds Austria,Finland,and Switzerland.Thus,Maddison's sample has no Latin observations,either Old World or New,and does not treat separately one of the best examples of catching-up,Ireland.His 1994 sample repairs this damage.See his"Explaining the Economic Performance." 4 DeLong,.“Productivity Growth.” 17.15 PM

Globalization, Convergence, and History 279 CONVERGENCE IN THE PAST What does history have to say about convergence? To answer that question, we have to agree on the meaning of convergence. The critical bottom line for me is whether the living standard gap between rich and poor countries falls over time. Convergence implies an erosion in this gap, at least in percentage terms. New growth theorists call this sigmaconvergence. To get sigma-convergence, poor countries must grow faster than rich, an event new growth theorists call beta-convergence. Second, what history? My interest has always been in what Simon Kuznets called modern economic growth, and that translates here into the century and a half since about 1850. Third, convergence of what? There appear to be two data sets that can be used for such analysis. Angus Maddison's GDP per worker-hour estimates offer one (originally published in 1982, now superseded by his 1991 book and by even more recent revisions).1 My real wages of the urban unskilled offer another.2 Fourth, convergence among whom? As the introduction suggests, my net will only capture members of the present OECD club with European origin (plus Argentina and Brazil).3 All of us know that much of the convergence since 1870 disappears when the net is widened to include Eastern Europe, and if it were widened still further to include the Third World, convergence would totally evaporate.4 Why the small net? Because I think the sources of convergence in the OECD club are themselves misunderstood, and it matters to get the facts right. Figure 1 and Table 1 document real wage convergence from midcentury to the Great War. Although the convergence was not as fast as that of the late twentieth century, it was pronounced, and about as fast as average convergence over the full 150 years since 1850. Figure 2 and Table 2 show that the late nineteenth-century real wage convergence was replicated by gross domestic product (GDP) per worker-hour. However, real wage convergence was a lot faster than GDP per worker-hour, and the globalization arguments that follow offer some reasons why. Table 3 reports the A's that emerge when so-called unconditional convergence equations are applied to this historical epoch. They imply a peak rate of real wage convergence between 1870 and 1890 of 1.2 percent per annum, and about 1 percent per annum over the 1870 to 1913 period as a whole. Although impressive, the late nineteenth-century rate of convergence 1 Maddison, Phases, Dynamic Forces, and "Explaining the Economic Performance." 2 These wages are purchasing-power-parity adjusted; see Williamson, "Evolution." 3 The full real wage 17-country sample is Australia, Argentina, Belgium, Brazil, Canada, Denmark, France, Germany, Great Britain, Ireland, Italy, the Netherlands, Norway, Portugal, Spain, Sweden, and the United States. The full (1991) GDP per worker-hour 15-country sample subtracts Argentina, Brazil, Ireland, Portugal, and Spain, but adds Austria, Finland, and Switzerland. Thus, Maddison's sample has no Latin observations, either Old World or New, and does not treat separately one of the best examples of catching-up, Ireland. His 1994 sample repairs this damage. See his "Explaining the Economic Performance." 4 DeLong, "Productivity Growth." This content downloaded from 211.80.95.69 on Sun, 20 Oct 2013 19:57:15 PM All use subject to JSTOR Terms and Conditions

280 Williamson TABLE 1 COEFFICIENTS OF VARIATION OF REAL WAGES,1854-1939 Full Sample Less Full Sample Less Full Sample North America North America and Iberia C(13) C(17) C(16) C(12) C(15) C(14) C(10) C(13) 1854 0.326 0.308 0.340 1870 0.254 0.255 0.224 0.223 0.229 0.232 1890 0.199 0.114 0.102 1913 0.191 0.068 0.039 1914 0.103 0.085 0.068 1926 0.148 0.146 0.138 1927 0.188 0.147 0.186 0.142 0.131 1939 0.285 0.200 0.138 Notes:The "full sample"includes the following 13 countries until 1870:Australia,the United States Belgium,France,Germany,Great Britain,Ireland,Netherlands,Norway,Spain,Sweden,Brazil,and Portugal.In 1870 the following four countries are added to the sample:Argentina,Canada,Denmark. and Italy.Portugal drops from the sample from 1914 to 1926 and then rejoins.The "full sample less North America"excludes Canada and the United States,implying that we start with 12 countries and then increase to 15 in 1870.Again,Portugal drops from the sample between 1914 and 1926.The"full sample less North America and Iberia"excludes the United States,Canada,Spain,and Portugal, implying that we start with 10 countries and expand to 13 in 1870. Source:Williamson,"Evolution,"with Great Britain revised. TABLE 2 COEFFICIENTS OF VARIATION OF GDP PER WORKER-HOUR,1870-1938 Full Sample Less Full Sample North America C(15) C(13) 1870 0.153 0.169 1890 0.118 0.122 1913 0.107 0.088 1929 0.110 0.080 1938 0.090 0.054 Notes:The "full sample"includes Australia,Austria,Belgium,Canada,Denmark,Finland,France Germany,Italy,the Netherlands,Norway,Sweden,Switzerland,the United Kingdom,and the United States.It does not include Japan.The"full sample less North America"drops Canada and the United States from the sample. Source:Maddison,Dynamic Forces. TABLE 3 THE ESTIMATED RATE OF CONVERGENCE(A)1854-1939 Real Wages GDP per Worker Hour Less North Less Iberia and Less North Epoch Full Sample America North America Full Sample America 18541870 +0.005 +0.006 +0.004 1870-1890 +0.012 +0.020 +0.021 +0.004 +0.005 1890-1913 +0.008 +0.017 +0.033 +0.007 +0.011 19141926 -0.011 -0.016 -0.030 1927-1939 -0.003 +0.002 -0.001 1913-1929 +0.002 +0.005 1929-1938 +0.019 +0.024 Notes:The rate of convergence is A=1/t In (B+1),where B is the coefficient in convergence equation on log of initial real wages or GDP per worker-hour,and t is the time span. Source:Data underlying Tables 1 and 2. 219575

280 Williamson TABLE 1 COEFFICIENTS OF VARIATION OF REAL WAGES, 1854-1939 Full Sample Less Full Sample Less Full Sample North America North America and Iberia C(13) C(17) C(16) C(12) C(15) C(14) C(10) C(13) 1854 0.326 0.308 0.340 1870 0.254 0.255 0.224 0.223 0.229 0.232 1890 0.199 0.114 0.102 1913 0.191 0.068 0.039 1914 0.103 0.085 0.068 1926 0.148 0.146 0.138 1927 0.188 0.147 0.186 0.142 0.131 1939 0.285 0.200 0.138 Notes: The "full sample" includes the following 13 countries until 1870: Australia, the United States, Belgium, France, Germany, Great Britain, Ireland, Netherlands, Norway, Spain, Sweden, Brazil, and Portugal. In 1870 the following four countries are added to the sample: Argentina, Canada, Denmark, and Italy. Portugal drops from the sample from 1914 to 1926 and then rejoins. The "full sample less North America" excludes Canada and the United States, implying that we start with 12 countries and then increase to 15 in 1870. Again, Portugal drops from the sample between 1914 and 1926. The "full sample less North America and Iberia" excludes the United States, Canada, Spain, and Portugal, implying that we start with 10 countries and expand to 13 in 1870. Source: Williamson, "Evolution," with Great Britain revised. TABLE 2 COEFFICIENTS OF VARIATION OF GDP PER WORKER-HOUR, 1870-1938 Full Sample Less Full Sample North America C(15) C(13) 1870 0.153 0.169 1890 0.118 0.122 1913 0.107 0.088 1929 0.110 0.080 1938 0.090 0.054 Notes: The "full sample" includes Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Italy, the Netherlands, Norway, Sweden, Switzerland, the United Kingdom, and the United States. It does not include Japan. The "full sample less North America" drops Canada and the United States from the sample. Source: Maddison, Dynamic Forces. TABLE 3 THE ESTIMATED RATE OF CONVERGENCE (A) 1854-1939 Real Wages GDP per Worker Hour Less North Less Iberia and Less North Epoch Full Sample America North America Full Sample America 1854-1870 +0.005 +0.006 +0.004 - 1870-1890 +0.012 +0.020 +0.021 +0.004 +0.005 1890-1913 +0.008 +0.017 +0.033 +0.007 +0.011 1914-1926 -0.011 -0.016 -0.030 - 1927-1939 -0.003 +0.002 -0.001 - 1913-1929 - - +0.002 +0.005 1929-1938 - - - +0.019 +0.024 Notes: The rate of convergence is A = 1 / t ln (X3 + 1), where 3 is the coefficient in convergence equation on log of initial real wages or GDP per worker-hour, and t is the time span. Source: Data underlying Tables 1 and 2. This content downloaded from 211.80.95.69 on Sun, 20 Oct 2013 19:57:15 PM All use subject to JSTOR Terms and Conditions