Firm-level evidence on productivity differentials and turnover in Taiwanese manufacturing

https://doi.org/10.1016/S0304-3878(01)00155-9Get rights and content

Abstract

High rates of firm entry and exit have accompanied the rapid and sustained growth of output in Taiwan's manufacturing sector. A high rate of firm turnover can contribute to industry productivity growth if it reflects a transfer of resources from less efficient to more efficient producers. Using comprehensive firm-level panel data from the Taiwanese Census of Manufactures for 1981, 1986, and 1991, we measure total factor productivity for entering, exiting, and continuing cohorts of firms and quantify the contribution of firm turnover to industry productivity improvements. Across manufacturing firms, we find significant differences in productivity that are reflected in turnover patterns. Cohorts of new firms have lower average productivity than incumbents but are themselves a heterogeneous group. The more productive members of the group, on average, survive and, in many cases, their productivity converges to the productivity level of older incumbents. Exiting firms are less productive than survivors. The productivity differential between entering and exiting firms is an important source of industry-level productivity growth in Taiwanese manufacturing, accounting for as much as one-half of industry improvement in some industries and time periods.

Introduction

The substantial growth of Taiwan's manufacturing sector since 1960, and the growth in world demand for its exports which spurred this expansion, has been well documented. While the growth of manufactured output after 1980 slowed relative to the 20% annual rate that was common in the previous two decades, it still averaged 6.5% per year from 1981–1991, with annual growth rates in double digits for several major industries. Output expansions of this magnitude are likely to require substantial changes in market structure but there is little evidence on the micro-level changes which have generated this supply growth.

One of the interesting features of Taiwan's supply expansion is that a substantial portion is attributable to new firm entry. The total number of manufacturing firms grew by 7.7% per year between 1981 and 1991. However, this large net change in the number of firms is still only a lower bound on the magnitude of structural adjustment that the sector has undergone. Even in slowly growing economies, the micro-level reallocation of resources that occurs as firms enter, expand, contract, and exit production is many times larger than the net change in sectoral output, employment, or the number of firms.1 While there is a considerable literature debating the magnitude and sources of productivity growth in Taiwan, the evidence is based on aggregate or sectoral data and no studies have examined the role of micro-level reallocations as a determinant of productivity change.2 The linkage is direct: if firms differ in their productivity, and their own efficiency is a determinant of their choice to enter, exit, or grow, then firm turnover contributes to changes in sectoral or aggregate productivity. The size of this contribution is an empirical question, but is potentially important in rapidly growing economies like Taiwan.

In this paper, we use comprehensive firm-level panel data from the Taiwanese Census of Manufactures for the years 1981, 1986, and 1991, to document the role of firm entry, exit, and growth in Taiwan's productivity growth. We quantify the extent of firm turnover, measure differences in total factor productivity between entering, exiting, and continuing firms, and determine the contribution of firm turnover to industry productivity growth. Our analysis parallels a number of recent studies of firm or plant turnover and productivity growth for manufacturing industries in the United States Baily et al., 1992, Olley and Pakes, 1996, Haltiwanger, 1997, Foster et al., 2001, Israel (Griliches and Regev 1995), and Chile and Colombia Liu, 1993, Tybout, 1996a, Liu and Tybout, 1996. With the exception of Olley and Pakes (1996), these studies find that entry and exit of firms or plants within an industry contribute little to productivity growth, generally because there are only small productivity differences between entering and exiting producers and/or these groups account for a very small share of industry output.

Research on the institutional environment in Taiwan has emphasized that the low capital intensity of much of the production combined with the manufacturing sector's dense network of subcontracting relationships and trading firms result in low sunk entry and exit costs. Recent theoretical models of firm dynamics have emphasized that low sunk entry costs result in higher entry flows and increased pressure on inefficient firms to exit, and thus, hasten the reallocation of resources from inefficient to more efficient producers. Finally, as we will document, entering and exiting firms account for a substantial share of industry production in Taiwan. Together these factors suggest that if turnover and market share reallocations reflect productivity differentials across firms, they could be a much more important source of productivity growth in Taiwan than has been found in other countries.

A consistent finding of our empirical study is that there are significant differences in productivity across manufacturing firms and these differences are reflected in turnover patterns. Cohorts of new firms have lower average productivity than incumbents but are themselves a heterogeneous group. The more productive members of the group survive and, in many cases, their productivity converges to the productivity level of older incumbents. Firms that will exit in the future are less productive than their counterparts that survive. These patterns are consistent with the view that markets sort out high from low productivity firms and that the most efficient firms tend to survive.

For many industries, but not all, we find evidence of productivity growth at modest annual rates, 2% per year, over the entire decade. In most cases, the productivity improvement is widespread across the whole distribution of firms, suggesting that it may be less related to individual firm investments than it is to common improvements in worker quality and infrastructure. One implication of this general shift in the firm-level productivity distribution is that entering firms also shared in the higher productivity so that the turnover of firms through entry and exit made a significant positive contribution to industry productivity growth. Productivity growth of incumbent firms was also a significant source of industry productivity growth.

In Section 2, we summarize a dynamic model of firm turnover and market selection due to Hopenhayn (1992) with an emphasis on the model's implications for productivity differences between entering, exiting, and continuing cohorts of firms. Section 3 summarizes the importance of firm entry and exit in the market for nine major manufacturing industries in Taiwan. 4 Firm total factor productivity, 5 Productivity differentials and firm turnover describe the index number methods used to measure firm-level total factor productivity and quantify the differences in productivity among entering, exiting, and continuing firms. Section 6 disaggregates industry productivity growth into the contributions of incumbents and the reallocation of production from less efficient to more efficient producers.

Section snippets

A theoretical framework relating productivity and turnover

Recent theoretical models of industry dynamics by Jovanovic (1982), Lambson (1991), Hopenhayn (1992), and Ericson and Pakes (1995) have been developed to explain the divergent paths of growth and failure that characterize microdata on individual producers. These models all begin with the assumption that producers within the same industry differ in their productive efficiency and are subject to idiosyncratic shocks or uncertainty. Differences in the evolution of their productivity over time, in

Firm turnover in the Taiwan manufacturing sector

To assess the importance of entering and exiting firms to Taiwan's supply expansion, we classify firms in each two-digit manufacturing industry into entry cohorts and summarize their cumulative output contributions over the 1981–1991 period. Specifically, we classify each producer into one of three cohorts based on when they first appeared in the industry. The first cohort consists of firms observed in operation in 1981, the second cohort is firms first observed in the industry in the 1986

Measurement issues

For each firm in the Taiwanese manufacturing data, we construct an index of total factor productivity (TFP) in each of the three census years 1981, 1986, and 1991.7 We use this index as a single measure of the firms' relative efficiency, a proxy for ϕt in the theoretical model.8

Productivity differentials and firm turnover

In this section, we document whether firms that undergo transitions in or out of production tend to be located in different parts of the productivity distribution. We do this by quantifying (unweighted) mean differences in productivity across groups of firms with different transition patterns. The goal is to determine if turnover patterns reflect the underlying differences in productivity as even the simplest models of firm heterogeneity and market selection predict. To clarify our productivity

Implications for industry productivity growth

We begin by defining industry productivity as the market-share weighted sum of the firm productivity levelslnTFPt=fθftlnTFPftwhere firm productivity is defined in Eq. (1) and θft is the value of firm f sales relative to total industry sales in year t. With this formulation, shifts of output from low productivity to high productivity firms will contribute positively to industry productivity growth, even if no individual firm experiences a productivity increase. This is appropriate because our

Summary and conclusions

In this paper, we use micro panel data for Taiwan's manufacturing sector to measure firm total factor productivity, to study the movement of the productivity distribution over time, to examine if patterns of firm turnover are related to productivity differences, and to measure the contribution of firm turnover to industry productivity growth.

A broad conclusion from our examination of the data is that there are significant differences in TFP across firms and that these differences are reflected

Acknowledgements

We are grateful to two anonymous referees, Rick Bond, Tim Bresnahan, Dick Caves, Erwin Diewert, Wayne Gray, Zvi Griliches, John Haltiwanger, J. David Richardson, Jim Tybout and participants at the NBER Productivity Summer Institute, the Conference on “Empirical Investigations in International Trade” at Purdue University, the Conference on “Trade and Technology Diffusion” at the Fondazione Eni Enrico Mattei, Milan, Italy, and the Conference on “The Impact of Technological Change on Firm and

References (39)

  • D.W. Caves et al.

    US Trunk Air Carriers, 1972–1977: a multilateral comparison of total factor productivity

  • D.W. Caves et al.

    Output, input, and productivity using superlative index numbers

    Economic Journal

    (1982)
  • S.J. Davis et al.

    Job Creation and Destruction

    (1996)
  • T. Dunne et al.

    Patterns of entry and exit in US manufacturing industries

    The Rand Journal of Economics

    (1988)
  • R. Ericson et al.

    Markov-perfect industry dynamics: a framework for empirical work

    Review of Economic Studies

    (1995)
  • L. Foster et al.

    Aggregate productivity growth: lessons for microeconomic evidence

  • S. Gee et al.

    Export success and technological capability

  • D.H. Good et al.

    Index number and factor demand approaches to the estimation of productivity

  • J.C. Haltiwanger

    Measuring and analyzing aggregate fluctuations: the importance of building from microeconomic evidence

  • Cited by (222)

    • Wage dispersion with heterogeneous wage contracts

      2023, Review of Economic Dynamics
    • Too much energy The perverse effect of low fuel prices on firms

      2022, Journal of Environmental Economics and Management
    • Bundling and exporting: Evidence from German SMEs

      2021, Journal of Business Research
      Citation Excerpt :

      Entering export markets requires internalizing sunk costs (Melitz, 2003), and overcoming these costs is challenging, as the competition with incumbent foreign firms quite often results in lower mark-ups than those obtained in domestic markets (Bughin, 1996). The attributes and quality of the product are the main determinants of a firm’s capacity to increase profit margins abroad and strengthen its presence in foreign markets (Aw et al., 2001). Most differences in the ability to penetrate foreign markets or in organizational innovation can be explained by differences in product/service quality (Golovko & Valentini, 2011).

    • Assessing the drivers of changes in aggregate fuel economy in Massachusetts: The role of vehicle reallocation

      2021, Technological Forecasting and Social Change
      Citation Excerpt :

      Hence, this paper employs alternative decomposition approaches to incorporate the structural changes in the vehicle population. The techniques used here draw on literature about the firm productivity decomposition using microdata, as well-documented in prior studies (Aw et al., 2001; Baily et al., 2001; Christensen et al., 1981; Griliches and Regev, 1995; Olley and Pakes, 1996). The heterogeneity and reallocation of vehicles will determine the dynamics of aggregate fuel economy in a complex way.

    View all citing articles on Scopus
    View full text