Proposal preview

Accounting for growth in global economic history

Growth accounting is as old as modern growth theory. Building on the seminal work of Copeland (1937), Tinbergen (1942) was the first to use the aggregate production function to decompose output growth into input contributions. Solow (1957) adapted the concept of Total Factor Productivity (TFP) from Stigler (1947) and used it as a parameter of Hicks-neutral neoclassical production function. Solow offered an attractive approach to measuring the dynamics of growth that has remained the template for growth accounting until today. The past half a century of growth accounting demonstrates the enormous legacy of Solow’s pioneering work (Hulten 2001). Different approaches have been proposed since, presenting economists and historians with both challenges and opportunities. Historical growth accounting has made one of the greatest contributions to comparative and global economic history (Crafts 2004; van Ark and Crafts 2007; Crafts 2010). Dual TFP accounting developed by Jorgenson and Griliches (1967) helped historians to extend the analysis of growth dynamics to periods preceding the establishments of national accounts using data on factor prices rather than quantities (Antràs and Voth 2003; Broadberry and Gupta 2009).

The following generation of growth accounting (Jorgenson, Gollop, and Fraumeni 1987; Jorgenson, Ho, and Stiroh 2005) improved the measurement of labour and capital inputs, incorporated labour and capital services, inputs, as well as labour composition. Disaggregated growth accounts suggested by Jorgenson, Gollop and Fraumeni (1987) have allowed researchers to measure returns to scale and the structural components of growth, rendering TFP a measure of efficiency and moving beyond a simple ‘measure of our ignorance’, as Abramovitz famously labelled it. The KLEMS methodology was introduced (Jorgenson, Ho, and Stiroh 2005; Timmer et al. 2010; Jorgenson, Fukao, and Timmer 2016) to exploit the richness of disaggregated statistics on output and inputs, including ICT capital, intangible assets and natural resources, available for more recent times, in order to better understand the aggregate growth process.

Economists have debated which approach is superior for measuring output, capital input and labour composition, which is the appropriate level of analysis, or how to best account for the effects of structural shifts in employment and output (Oulton 2016). Another issue is to what extent the neoclassical framework is applicable to economies, where free markets are absent or limited (Weitzman 1970; Easterly and Fischer 1995). Recent innovations in growth theory also question to what extent factor accumulation can be isolated from productivity growth (see models of endogenous and biased technical change) and whether the same production functions can be applied to different sectors of the economy (new structural economics). These insights have important implications for economic historians on how to measure the sources of growth and how to interpret the findings of historical growth accounts.

This session proposes three objectives. Firstly, we wish to provide an overview of current and recent contributions to historical growth accounting around the globe, where the literature stands, how far we have progressed, and what challenges we face. Secondly, we seek to build bridges to recently developed cross-country databases on comparative growth accounts, in particular World KLEMS, Total Economy Database and Penn World Tables, which provide the standard sources for economists and represent the methodological state of the art. This agenda calls for a second generation of historical national accounts and growth accounts for core western economies, too, since the seminal contributions from the 1950s and 1960s have become, in many aspects, out of date. In this context, we must investigate to what extent historical statistics allows us to analyse the role of intangibles (inventions), communication technologies (e.g. telegraph) or intermediate inputs (coal-steam-electricity energy transformations), using recent advances of the growth accounting framework. Finally, we wish to discuss how the most recent insights from growth theory can be integrated into growth accounting, which is vital for our work to continue to make an impact similar to what it has done over the last six decades. To blend these three perspectives, our session will include introductory and discussion notes from some of the internationally most prominent names on the topic.

References

Antràs, Pol, and Hans-Joachim Voth. 2003. “Factor Prices and Productivity Growth during the British Industrial Revolution.” Explorations in Economic History 40 (1): 52–77.

Ark, Bart van, and Nicholas Crafts. 2007. Quantitative Aspects of Post-War European Economic Growth. Cambridge: Cambridge University Press.

Broadberry, Stephen N., and Bishnupriya Gupta. 2009. “Lancashire, India, and shifting competitive advantage in cotton textiles, 1700–1850: the neglected role of factor prices.” Economic History Review 62 (2): 279-305.

Copeland, Morris A. 1937. “Concepts of National Income.” In The Conference on Research in Income and Wealth, 1:2–63. NBER Book Series Studies in Income and Wealth. New York: National Bureau of Economic Research. http://papers.nber.org/books/unkn37-1.

Crafts, Nicholas. 2004. “Productivity Growth in the Industrial Revolution: A New Growth Accounting Perspective.” Journal of Economic History 64 (2): 521–35.

———. 2010. “The Contribution of New Technology to Economic Growth: Lessons from Economic History.” Revista de Historia Economica/Journal of Iberian and Latin American Economic History 28 (3): 409–40.

Easterly, William, and Stanley Fischer. 1995. “The Soviet Economic Decline.” The World Bank Economic Review 9 (3): 341–71.

Hulten, Charles R. 2001. “Total Factor Productivity: A Short Bibliography.” In , edited by Charles R. Hulten, Edwin R. Dean, and Michael J. Harper, 1–53. New Developments in Productivity Analysis. Chicago; London: The University of Chicago Press.

Jorgenson, Dale W., Kyoji Fukao, and Marcel P. Timmer, eds. 2016. Growth and Stagnation in the World Economy. Cambridge University Press.

Jorgenson, Dale W., Frank M. Gollop, and Barbara Fraumeni. 1987. Productivity and U.S. Economic Growth. Amsterdam: North-Holland.

Jorgenson, Dale W., and Zvi Griliches. 1967. “The Explanation of Productivity Change.” Review of Economic Studies 34 (3): 249–83.

Jorgenson, Dale W., Mun S. Ho, and Kevin J. Stiroh. 2005. Information Technology and the American Growth Resurgence. Vol. 3. Cambridge, MA: The MIT Press.

Oulton, Nicholas. 2016. “The Mystery of TFP.” International Productivity Monitor 31: 68–87.

Solow, Robert M. 1957. “Technical Change and the Aggregate Production Function.” Review of Economics and Statistics 39 (3): 312–20.

Stigler, George J. 1947. Trends in Output and Employment. National Bureau of Economic Research.

Timmer, Marcel P., Robert Inklaar, Mary O’Mahony, and Bart van Ark. 2010. Economic Growth in Europe. Cambridge: Cambridge University Press.

Tinbergen, Jan. 1942. “Zur Theorie der Langfristigen Wirtschaftsentwicklung.” Weltwirtschaftliches Archiv 55 (1): 511–49.

Weitzman, Martin L. 1970. “Soviet Postwar Economic Growth and Capital-Labor Substitution.” American Economic Review 60 (4): 676–92.

Organizer(s)

  • Leandro Prados de la Escosura Charles III University of Madrid leandro.prados.delaescosura@uc3m.es Spain
  • Tamás Vonyó Bocconi University tamas.vonyo@unibocconi.it Italy
  • Ilya B. Voskoboynikov National Research University Higher School of Economics ivoskoboynikov@hse.ru Russia

Session members

  • Dale W. Jorgenson, Harvard University
  • Joan Rosés, London School of Economics
  • José Díaz-Bahamonde, Pontifica Universidad Católica de Chile
  • Kyoji Fukao, Hitotsubashi University
  • Tatsuji Makino, Hitotsubashi University
  • Tokihiko Settsu, Hitotsubashi University
  • Claire Giordano, Banca d'Italia
  • Zollino Francesco, Banca d'Italia
  • Gert Wagner, Pontifica Universidad Católica de Chile
  • André Hofman, CEPAL/ECLAC
  • Jop Woltjer, University of Groningen
  • Gerben Bakker, London School of Economics
  • Harry X Wu, Hitotsubashi University

Discussant(s)

  • Nicholas Crafts Warwick University N.Crafts@warwick.ac.uk
  • Bart van Ark the Conference Board Bart.VanArk@conference-board.org

Papers

Panel abstract

To what extent historical statistics allows us to analyse the role of intangibles (inventions), communication technologies (telegraph) or intermediate inputs (coal-steam-electricity energy transformations), using recent advances of the growth accounting framework? The past half a century of growth accounting demonstrates the enormous impact of Solow’s (1957) pioneering study on the profession. The following generation of the growth accounting framework (Jorgenson et al. 1987; 2005) improved the measurement of inputs (labour and capital services, intermediate inputs, labour composition). This session proposes three objectives. Firstly, we wish to provide an overview of recent contributions to historical growth accounting around the globe. Secondly, we build bridges to recently developed cross-country databases on comparative growth accounts (e.g. World KLEMS, Total Economy Database and Penn World Tables). Finally, we discuss how the most recent insights from growth theory can be integrated into growth accounting.

1st half

Accounting for productivity growth in the long run: Spain, 1850-2015

Leandro Prados de la Escosura, Joan Rosés

Labour productivity was the main determinant of per capita GDP growth between mid-nineteenth century in the early 1970s. Since then, GDP per head and per hour worked have evolved in opposite directions. This suggests that Spain’s economy has expanded in labour intensive sectors that attract less investment and, especially, innovation. The sources of labour productivity need to be explored to test this interpretation. It emerges that efficiency gains are the main driver of labour productivity growth and that during the last three decades TFP stagnated or declined while physical capital deepening accounts for the suluggish and uneven growth of output per worked hour.

Labour productivity was the main determinant of per capita GDP growth between mid-nineteenth century in the early 1970s. Since then, GDP per head and per hour worked have evolved in opposite directions. This suggests that Spain’s economy has expanded in labour intensive sectors that attract less investment and, especially, innovation. The sources of labour productivity need to be explored to test this interpretation. It emerges that efficiency gains are the main driver of labour productivity growth and that during the last three decades TFP stagnated or declined while physical capital deepening accounts for the suluggish and uneven growth of output per worked hour.

Long-run trends in labour and capital inputs and the ensuing productivity in Italy

Claire Giordano, Francesco Zollino

We first document data and methodologies we have employed to estimate : headcount and full-time equivalent labour input and net capital stock by main asset type in Italy since 1861. We then provide a reassessment of the origins of Italy’s economic development since the country unification. Italy failed to catch up on the European technological leader of the time, the U.K., until WWII, after which it embarked on an exceptional catching-up process during the Golden Age. As in many countries, this rapid growth could not be sustained after 1973, but the further slowing down since the Nineties has been more pronounced in Italy than elsewhere. Productivity growth rates even turned negative during the recent recessionary phase. The disappointing performance of the Italian economy since 1993 is explained largely by slow labour productivity growth in the now dominant services sector and by sluggish aggregate total factor productivity growth.

We first document data and methodologies we have employed to estimate : headcount and full-time equivalent labour input and net capital stock by main asset type in Italy since 1861. We then provide a reassessment of the origins of Italy’s economic development since the country unification. Italy failed to catch up on the European technological leader of the time, the U.K., until WWII, after which it embarked on an exceptional catching-up process during the Golden Age. As in many countries, this rapid growth could not be sustained after 1973, but the further slowing down since the Nineties has been more pronounced in Italy than elsewhere. Productivity growth rates even turned negative during the recent recessionary phase. The disappointing performance of the Italian economy since 1993 is explained largely by slow labour productivity growth in the now dominant services sector and by sluggish aggregate total factor productivity growth.

Accounting for growth in Latin America in an historical perspective

André Hofman

This article provides an assessment of the long-run economic performance of a group of Latin American from a comparative and historical perspective. The analysis of economic performance concentrates on the quantification of economic growth and the measurement of factor inputs and total factor productivity. A growth accounting analytical framework is presented using a total factor productivity analysis in which step-by-step explanatory factors are listed, and given their weight in ‘explaining’ economic growth in the sample of countries. Growth accounting shows the contribution of factor inputs (capital, labour and land) and total factor productivity to output growth. This kind of growth accounting exercise may serve different purposes such as explaining differences in growth rates between countries, illuminating the process of convergence and divergence, assessing the role of technical progress and calculating potential output losses. Growth accounting cannot provide a full causal explanation. It deals with ‘proximate’ rather than ‘ultimate’ causality and...

This article provides an assessment of the long-run economic performance of a group of Latin American from a comparative and historical perspective. The analysis of economic performance concentrates on the quantification of economic growth and the measurement of factor inputs and total factor productivity. A growth accounting analytical framework is presented using a total factor productivity analysis in which step-by-step explanatory factors are listed, and given their weight in ‘explaining’ economic growth in the sample of countries. Growth accounting shows the contribution of factor inputs (capital, labour and land) and total factor productivity to output growth. This kind of growth accounting exercise may serve different purposes such as explaining differences in growth rates between countries, illuminating the process of convergence and divergence, assessing the role of technical progress and calculating potential output losses. Growth accounting cannot provide a full causal explanation. It deals with ‘proximate’ rather than ‘ultimate’ causality and records the facts about growth components: it does not explain the underlying elements of policy or circumstance, national or international, but it does identify which facts need further explanation. We subdivide long-run growth in sub-periods using as benchmarks, 1870, 1913, 1950 and 1973. The year in which the analysis starts varies from country to country but the starting point is at least 1900 and the initial year of the growth accounts is at least 1950. So the minimum period in common is the 1900-2017 period. However, for several countries we have been able to push the analysis further backwards. Quantitative analysis of this kind can provide guidance on the broad contours of development which is a very useful complement to qualitative analysis. There have been relatively few growth accounting studies for Latin America and most are for individual countries. One of the problems in comparing these countries is the great variety of methodologies used. This article makes use of a standardised approach applied both to Latin American and other countries with which the Latin America are compared. Increases in different factor inputs are measured in terms of average annual compound growth rates. Labour services are derived by employment, annual hours per person and level of education. The effects are weighted by the relevant factor shares to show the contribution of labour input. Capital services are derived by capital formation of machinery and equipment, residential and nonresidential structures. The average annual compound growth rates of the capital services were also weighted by its respective factor share to give the contribution of capital input. The sum of weighted capital and labour services contributions is an indicator of the impact of factor inputs on economic growth. The results of the growth accounting exercises, using capital and labour services, show a low or negative total factor productivity, for most countries and most periods in the Latin American countries.

Chile: Productivity and Capital Deepening in the Long Period

José Díaz-Bahamonde,Gert Wagner

Long period growth of the Chilean economy is decomposed for the 1833-2010 period highlighting the particular roles played by capital, labor, human capital and overall productivity growth. We distinguish among the whole time span and breakdowns following Maddison's world growth phases. Main findings for the long run are (1) Chilean long-run growth is almost entirely accounted for by factor accumulation when adopting a strict or net TFP concept and (2)physical capital accumulation plays and by far the main role in explaining income growth. However, average outcomes over long periods hide much heterogeneity when dividing according Maddison time breakdown. Finally, per capita income and per employee product occasionally diverge implying a role for the evolution of employment-population ratio.

Long period growth of the Chilean economy is decomposed for the 1833-2010 period highlighting the particular roles played by capital, labor, human capital and overall productivity growth. We distinguish among the whole time span and breakdowns following Maddison's world growth phases. Main findings for the long run are (1) Chilean long-run growth is almost entirely accounted for by factor accumulation when adopting a strict or net TFP concept and (2)physical capital accumulation plays and by far the main role in explaining income growth. However, average outcomes over long periods hide much heterogeneity when dividing according Maddison time breakdown. Finally, per capita income and per employee product occasionally diverge implying a role for the evolution of employment-population ratio.

Structural Change, Capital Deepening, and TFP Growth in Japan: 1885-1970

Kyoji Fukao, Tatsuji Makino, Tokihiko Settsu

After the Meiji Restoration of 1868, Japan modernized its institutions and economic growth gradually picked up. Growth accelerated especially during the so-called high-speed growth era from 1955 to 1970, when Japan rapidly caught up with Western economies. The long-term sustained high-speed growth recorded during this period was unprecedented not only in Japan but worldwide. While other East Asian countries such as Singapore, Taiwan, South Korea, and China subsequently also experienced remarkable growth over a prolonged period, Japan’s place in history as the first country to record such sustained high-speed growth means that its experience continues to garner worldwide interest. Using newly constructed Hitotsubashi estimates of Japan’s historical GDP statistics and a growth accounting flamework, we analyze the sources of Japan’s economic growth from 1885 to 1970 and try to answer why Japan was not able accomplish such high-speed growth before 1955. Since until the mid-1960s the primary sector accounted for...

After the Meiji Restoration of 1868, Japan modernized its institutions and economic growth gradually picked up. Growth accelerated especially during the so-called high-speed growth era from 1955 to 1970, when Japan rapidly caught up with Western economies. The long-term sustained high-speed growth recorded during this period was unprecedented not only in Japan but worldwide. While other East Asian countries such as Singapore, Taiwan, South Korea, and China subsequently also experienced remarkable growth over a prolonged period, Japan’s place in history as the first country to record such sustained high-speed growth means that its experience continues to garner worldwide interest. Using newly constructed Hitotsubashi estimates of Japan’s historical GDP statistics and a growth accounting flamework, we analyze the sources of Japan’s economic growth from 1885 to 1970 and try to answer why Japan was not able accomplish such high-speed growth before 1955. Since until the mid-1960s the primary sector accounted for a large share of economic activity and was a major determinant of overall economic growth, we use a Hayashi and Prescott (2008) type two-sector model in which the economy overall is divided into the primary sector and the non-primary sector.

2nd half

The Sources of Growth in a Technologically Progressive Economy: the United States, 1899-1941

Gerben Bakker,Nicholas Crafts,Jop Woltjer

We develop new aggregate TFP growth estimates for the United States between 1899 and 1941, and sectoral estimates at the most disaggregated level so far, 38 industries. We include hard-to-measure services, and a refined measure of sectoral labour quality growth. The resulting dataset supersedes Kendrick (1961), showing TFP growth lower than previously thought, broadly based across industries, and strongly variant intertemporally. The four ‘great inventions’ that Gordon (2016) highlighted were important but less dominant in TFP growth than their predecessors in the British Industrial revolution. The findings also make it unlikely the 1930s had the twentieth century’s highest TFP growth.

We develop new aggregate TFP growth estimates for the United States between 1899 and 1941, and sectoral estimates at the most disaggregated level so far, 38 industries. We include hard-to-measure services, and a refined measure of sectoral labour quality growth. The resulting dataset supersedes Kendrick (1961), showing TFP growth lower than previously thought, broadly based across industries, and strongly variant intertemporally. The four ‘great inventions’ that Gordon (2016) highlighted were important but less dominant in TFP growth than their predecessors in the British Industrial revolution. The findings also make it unlikely the 1930s had the twentieth century’s highest TFP growth.

Sources of Growth in China's Pre-Communist Industrialization, 1912-1949

Harry X. Wu

Based on a newly constructed output and employment accounts by major sector, and a new estimation of capital stock since 1912, this study conducts an industry-origin growth accounting analysis of China's early industrialization following the Republican Revolution (1912) and up to the Communist Revolution. It highlights the impact of the Revolution, WWII and the Civil War on industrial development in terms of capital accumulation, labor transformation, structural changes in output and productivity performance.

Based on a newly constructed output and employment accounts by major sector, and a new estimation of capital stock since 1912, this study conducts an industry-origin growth accounting analysis of China's early industrialization following the Republican Revolution (1912) and up to the Communist Revolution. It highlights the impact of the Revolution, WWII and the Civil War on industrial development in terms of capital accumulation, labor transformation, structural changes in output and productivity performance.

Why did socialist economies fail? The role of factor inputs reconsidered

Tamás Vonyó,Alexander Klein

This article presents new estimates for investment and new growth accounts for three socialist economies between 1950 and 1989. Government statistics reported distorted measures for both the rate and the trajectory of productivity growth in Czechoslovakia, Hungary, and Poland. Researchers have benefited from revised output data, but have continued to use official statistics on capital input, or estimated capital stock from official investment data. Investment levels and rates of capital accumulation were much lower than officially claimed and over-reporting worsened over time. A setback in factor accumulation—both investment in equipment and labour input—contributed very significantly to the socialist growth failure of the 1980s.

This article presents new estimates for investment and new growth accounts for three socialist economies between 1950 and 1989. Government statistics reported distorted measures for both the rate and the trajectory of productivity growth in Czechoslovakia, Hungary, and Poland. Researchers have benefited from revised output data, but have continued to use official statistics on capital input, or estimated capital stock from official investment data. Investment levels and rates of capital accumulation were much lower than officially claimed and over-reporting worsened over time. A setback in factor accumulation—both investment in equipment and labour input—contributed very significantly to the socialist growth failure of the 1980s.

Regional development under socialism: Evidence from Yugoslavia

Leo Kukic

This paper analyses regional patterns of growth in Yugoslavia, under the most decentralized socialist system that ever existed. My analysis reveals that, despite government efforts to the contrary, socialist economic development in Yugoslavia resulted in divergence rather than in convergence between the constituent regions. I find that regional income divergence was caused by the failure of the less developed regions to converge towards the employment rates and total factor productivities of the more developed regions. I interpret these failures as symptoms of a single underlying problem: a capital intensity bias inherent to the governing objective of labour-managed firms.

This paper analyses regional patterns of growth in Yugoslavia, under the most decentralized socialist system that ever existed. My analysis reveals that, despite government efforts to the contrary, socialist economic development in Yugoslavia resulted in divergence rather than in convergence between the constituent regions. I find that regional income divergence was caused by the failure of the less developed regions to converge towards the employment rates and total factor productivities of the more developed regions. I interpret these failures as symptoms of a single underlying problem: a capital intensity bias inherent to the governing objective of labour-managed firms.