Comparing Apples to Oranges: Productivity Convergence and Measurement Across Industries and Countries
Year: 1997
Type: preprint
Abstract: This paper examines the role of sectors in the convergence of aggregate productivity levels in 14 OECD countries from 1970-1987.The major finding is that manufacturing shows little evidence of either labor productivity or multi-factor productivity convergence while other sectors, especially services, are driving the aggregate convergence result.The paper introduces a new measure of multi-factor productivity which avoids problems inherent to traditional TFP measures when comparing productivity levels.A model of trade, learningby-doing, and spillovers is developed which can explain convergence in some sectors and divergence in others.JEL Classification: 041, 047 was converging over the period, the sectors show disparate behavior.For all measures of productivity, the manufacturing sector shows no or little convergence, while other sectors, especially services, show strong evidence in favor of convergence.This finding for services together with the declining share of manufacturing in all 14 countries contributes to the convergence found at the aggregate level.The lack of convergence within manufacturing over this seventeen year period indicates that convergence is not an automatic phenomenon.Most theories of economic growth predict that openness and spillovers from R&D investment would contribute to convergence across countries and thus are not easily reconciled with these findings.However, we interpret this result in the context of a simple model of trade and learning-by-doing and argue that the lack of convergence within manufacturing may not be all that surprising.These results are especially pertinent to the study of convergence in countries at more heterogeneous levels of development.In a recent paper Young (1992) showed that while Hong Kong and Singapore apparently followed similar growth paths, their productivity performances were quite dramatically different.Our results suggest further that convergence of aggregate productivity may mask substantial differences at the sectoral level.Previous work on convergence across countries has concentrated almost exclusively on labor productivity using GDP per capita as the measure.This is due largely to a lack of data on labor and capital inputs necessary to construct broader measures of productivity.Using cross-section regressions, Baumol (1986), Barro and Sala-i-Martin (1991, 1992) and Mankiw.Romer and Weil (1992) argue that countries and regions are converging, or catching-up, since initially poor areas grow faster than their richer counterparts.However, the cross- section evidence is not uniform.Barro (1991) and DeLong (1988) show that the particular sample of countries determines whether catch-up holds.Time series results on longer series for OECD countries also show evidence of common trends but no tendency for convergence in levels (for example, see Bernard and Durlauf 1991).The use of labor productivity necessarily entails restrictions on the depth of analysis.By its very definition, a change in labor productivity confounds potential changes in tech- ' For example, see Dertouzos et al. (1989) and Baumol et al. (1989).'^Y/ L is constructed as output per worker.TFP is constructed as a weighted average of capital and labor productivities vnth the weights being the average factor shares over time and across countries.All analysis with productivity is done using the logs of the levels.Throughout the paper we maintain the assumptions of constant returns to scale and perfect competition.This means we calculate our labor shares using output rather than cost shares, not an innocent assumption.Here we hold the labor share constant across countries and time.In Section 2 we discuss this assumption in greater detail.The countries in the sample are the U.
Source: SSRN Electronic Journal
Authors: Andrew B. Bernard, Charles I. Jones
Language: English
Subfield: Economics and Econometrics
Domain: Social Sciences
Open Access status: green
OpenAlex ID: W3124964689
