Plant-level Productivity Research Articles

Factor Adjustments after Deregulation: Panel Evidence from Colombian Plants

In this paper, we analyze employment and capital adjustments using a panel of plants from Colombia. We allow for nonlinear adjustment of employment to reflect not only adjustment costs of labor but also adjustment costs of capital, and vice-versa. Using data from the Annual Manufacturing Survey, which include plant-level prices, we generate measures of plant-level productivity, demand shocks, and cost shocks, and use them to measure desired factor levels. We then estimate adjustment functions for capital and labor as a function of the gap between desired and actual factor levels. As in other countries, we find non-linear adjustments in employment and capital in response to market fundamentals. In addition, we find that employment and capital adjustments reinforce each other, in that capital shortages reduce hiring and labor shortages reduce investment. Moreover, we find that the market oriented reforms introduced in Colombia after 1990 increased employment adjustments, especially on the job destruction margin, while reducing capital adjustments. Finally, we find that while completely eliminating frictions from factor adjustments would yield a dramatic increase in aggregate productivity through improved allocative efficiency, the reforms introduced in Colombia generated only modest improvements.

International outsourcing and productivity: evidence from the Irish electronics industry

This paper examines the effect of international outsourcing of intermediate inputs on productivity at the level of the plant, with a focus on the electronics industry in the Republic of Ireland. The effect of international outsourcing on plant-level productivity depends critically on the nature of the outsourced inputs (services or tangibles) and on the plant's export intensity. Outsourcing of materials provides significant productivity gains, but this effect holds only for plants with low export intensities.

Workers' Education, Spillovers, and Productivity: Evidence from Plant-Level Production Functions

I assess the magnitude of human capital spillovers by estimating production functions using a unique firm-worker matched data set. Productivity of plants in cities that experience large increases in the share of college graduates rises more than the productivity of similar plants in cities that experience small increases in the share of college graduates. These productivity gains are offset by increased labor costs. Using three alternative measures of economic distance—input-output flows, technological specialization, and patent citations—I find that within a city, spillovers between industries that are economically close are larger than spillovers between industries that are economically distant.

Productivity Dynamics with Technology Choice: An Application to Automobile Assembly

During the 1980’s, all Japanese automobile producers opened assembly plants in North America. Industry analysts and previous research claim that these transplants are more productive than incumbent plants and that they produce with a substantially different production process. I compare the production processes by estimating a model that allows for heterogeneity in technology and productivity, both of which are intrinsically unobservable. The model is estimated on a panel of assembly plants, controlling for capacity utilization and price effects. The results indicate that the more recent technology uses capital more intensively and it has a higher elasticity of substitution between labour and capital. Hicks-neutral productivity growth is estimated to be lower, while capital-biased (labour-saving) productivity growth is higher for the new technology. Using the estimation results, I decompose industry-wide productivity growth and find plant-level changes in lean plants to be the most important contributor. Plant-level productivity growth is further decomposed to reveal the importance of capital-biased productivity growth, increases in the capital‐labour ratio, and returns to scale. Measuring technological change, which is defined as the shift of a production function over time, becomes harder if plants can choose which production technology to adopt. In the automobile assembly industry there is evidence that plants choose to operate with mass or lean technology. I estimate technical change for the industry using a rich plant-level data set for the U.S. Firms choose which technology to employ, but their choice is not observable to me. The shape of both technologies and the rate and factor-bias of their shift over time is estimated jointly with the technology decision. I use the model to decompose aggregate productivity growth and find that firm-level change in lean production plants is the most important contributor to aggregate productivity growth for the industry, but it only appears after controlling for the technology choice. In contrast with previous results in the literature, relocation of resources between plants is only the second most important effect. Labour productivity growth at lean plants is driven by increased capital use and capital-biased technical change. In mass production plants, on the other hand, almost all labour productivity advance is the result of Hicks-neutral technological growth. Traditional productivity measures suggest that the automobile industry enjoyed a considerable productivity improvement since the 1980’s, the period when the first Japanese producers entered the industry in the U.S. The number of vehicles produced per worker, an often quoted statistic, is on average higher in Japanese transplants than in plants owned by American producers. Researchers have concluded that entry of more productive plants caused industrywide productivity to increase. Compositional changes at the intensive margin—relocation of shares between active plants—and the extensive margin—plant turnover—are thought to be the principal sources of aggregate productivity change. 167

Marshall's scale economies

Using panel data this paper estimates plant level production functions for machinery and high-tech industries that allow for scale externalities from other plants in the same industry locally and from the scale or diversity of local economic activity outside the own industry. The paper finds that the count of other own industry plants, representing a count of local information spillover sources, has strong productivity effects in high tech but not machinery industries. Single plant firms both benefit more from and generate greater external benefits than corporate plants, given their greater reliance on external environments. On dynamic externalities, high-tech single plant firms benefit also from the scale of past own industry activity. I find little evidence of economies from the diversity or scale of local economic activity outside the own industry.

Capacity constraint and changing seasonality over business cycles: evidence from plant-level production data

The seasonal variability of production shrinks during a boom, not only at the industry level, but also at the plant level. Since this relationship is significant especially among plants with high capacity utilization, our finding is consistent with the capacity constraint hypothesis.

Plant Level Productivity, Efficiency, and Environmental Performance of the Container Glass Industry

This paper presents a methodology and empirical results based on theMalmquist productivity index. We measure productivity while treatingpollution as an undesirable output. Our estimates show that technicalchange has contributed to productivity and environmental performancegrowth in the container glass industry, an energy and pollution intensivesector. Changes in inter-plant efficiency over time have made thisproductivity growth more rapid than otherwise would have occurred withthe underlying technical change. The efficiency estimates show that thereare both opportunities to improve productivity and reduce pollution in thisindustry, as well as productivity losses associated with the emissionscontrol. The shadow prices for NOx, the undesirable output we analyze,is quite high compared to other regulated sectors.

The set of just-in-time management strategies: An assessment of their impact on plant-level productivity and input-factor substitutability using variable cost function estimates

Many manufacturers in the automobile industry around the world have adopted the just-in-time (JIT) set of management strategies in an effort to improve productivity, efficiency and product quality. This paper provides empirical evidence that supports the idea that JIT manufacturing environments are, in fact, more productive than their non-JIT counterparts. Plant-level cross-sectional data from auto-parts manufacturing firms are used to estimate variable cost functions for a JIT group as well as for a non-JIT group of plants. Differences in cost function characteristics between the two groups are examined and discussed.

Trade Liberalization, Exit, and Productivity Improvements: Evidence from Chilean Plants

This paper empirically investigates the effects of liberalized trade on plant productivity in the case of Chile. Chile presents an interesting setting to study this relationship since it underwent a massive trade liberalization that significantly exposed its plants to competition from abroad during the late 1970s and early 1980s. Methodologically, I approach this question in two steps. In the first step, I estimate a production function to obtain a measure of plant productivity. I estimate the production function semiparametrically to correct for the presence of selection and simultaneity biases in the estimates of the input coefficients required to construct a productivity measure. I explicitly incorporate plant exit in the estimation to correct for the selection problem induced by liquidated plants. These methodological aspects are important in obtaining a reliable plant-level productivity measure based on consistent estimates of the input coefficients. In the second step, I identify the impact of trade on plants' productivity in a regression framework allowing variation in productivity over time and across traded- and nontraded-goods sectors. Using plant-level panel data on Chilean manufacturers, I find evidence of within plant productivity improvements that can be attributed to a liberalized trade for the plants in the import-competing sector. In many cases, aggregate productivity improvements stem from the reshuffling of resources and output from less to more efficient producers.

Labor Productivity: Structural Change and Cyclical Dynamics

A longstanding issue in empirical economics is the behavior of average labor productivity over the business cycle. This paper provides new insights into the cyclicality of aggregate labor productivity by examining the cyclical behavior of productivity at the plant level as well as the role of reallocation across plants over the cycle. We find that plant-level productivity is even more procyclical than aggregate productivity, because short-run reallocation yields a countercyclical contribution to labor productivity. At the plant level, we find that cyclicality of productivity varies systematically with long-run employment growth. Over the course of the cycle, plants that are long-run downsizers exhibit significantly greater procyclicality of productivity than do long-run upsizers. When we control for the direction of a cyclical shock, we find that the fall in productivity from an adverse cyclical shock for long-run downsizers is significantly larger in magnitude than is the fall in productivity from an equivalent adverse cyclical shock for long-run upsizers. We argue that these findings raise questions about one of the most popular explanations of procyclical productivity: changing factor utilization over the cycle.

Estimating the linkage between energy efficiency and productivity

Many analyses have referred to industrial productivity benefits associated with energy efficiency that are at least as great or larger than the energy benefits. If plants with higher energy intensity also tend to have lower productivity, then energy policy needs to consider this. This study examines this issue for two segments of the glass industry, using plant level data from the Census Bureau. Productivity is defined by the difference in `best practice’ production efficiency, as measured by data envelopment analysis (DEA). This study uses regression analysis to estimate how differences in plant level electricity and fossil fuel intensity, i.e. energy use per unit of production, are attributable to differences in plant level productivity and other economic variables, like energy prices and cumulative production. In every case, productivity differences between plants are statistically significant in explaining differences in plant energy intensity. The coefficient that links productivity to energy efficiency yields a less than proportional impact for only one industry and fuel type. For others the relationship is implies that a 1% increase in productivity increase energy efficiency by more than 1%. This effect is statistically significant for flat glass, but not container glass. The estimates of the price and learning-by-doing coefficients seem reasonable, but without imposing the assumption that all plants in the industry are equally productive. This lends further credence to a significant relationship between productivity and energy efficiency.

Wages, Productivity, and Worker Characteristics: Evidence from Plant‐Level Production Functions and Wage Equations

We use a unique new data set that combines individual worker data with data on workers' employers to estimate plant-level production functions and wage equations, and thus to compare relative marginal products and relative wages for various groups of workers. The data and empirical framework lead to new evidence on numerous questions regarding the determination of wages, questions that hinge on the relationship between wages and marginal products of workers in different demographic groups. These include race and sex discrimination in wages, the causes of rising wages over the life cycle, and the returns to marriage. First, workers who have ever been married are more productive than never-married workers and are paid accordingly. Second, prime-aged workers (aged 35-54) are equally as productive as younger workers, and in some specifications are estimated to receive higher wages. However, older workers (aged 55+) are less productive than younger workers but are paid more. Third, the data indicate no difference between the relative wage and relative productivity of black workers. Finally, with the exception of managerial and professional occupations, women are paid about 25-35% less than men, but estimated productivity differentials for women are generally no larger than 15%, and significantly smaller than the pay differential.

Plant Size and Industry-mix Effects on UK Regional Productivity, Wage Costs and Operating Surplus

ROPER S. (1998) Plant size and industry-mix effects on UK regional productivity, wage costs and operating surplus, Reg. Studies 32 , 325-332. The ability of a region to create wealth depends both on the competitiveness of its individual firms and its industrial structure. This paper examines the effect of industry-mix and plant size structure on the wealth generating capacity of the UK regions from 1980-91. Reflecting the findings of similar US studies, the analysis suggests that, in general, UK regions' structural disadvantages act to reinforce rather than counter plant level productivity, wage cost and operating surplus (profitability) differentials. Thus lagging regions face a three-fold disadvantage: poorly performing plants; disadvantageous industry-mix; and plant size structure. This suggests a need for a strategic approach to policy which must aim to counter regions' structural disadvantages as well as to improve the competitiveness of existing plants. ROPER S. (1998) Les effets de la taille des établissements et de la structure industrielle sur la productivité, les coûts salariaux et le résultat d'exploitation au Royaume-Uni, Reg. Studies 32 , 325-332. La capacité d'une région à créer de la richesse dépend à la fois de la compétitivité de ses entreprises et de sa structure industrielle. Cet article examine les effets de la structure industrielle et de la taille des entreprises sur la capacité des régions au Royaume-Uni à créer de la richesse entre 1980 et 1991. En accord avec les résultats des études similaires faites aux Etats-Unis, l'analyse laisse supposer qu'en règle générale les inconvénients structurels des régions au Royaume-Uni servent à renforcer plutôt qu'à neutraliser les écarts de productivité, des coûts salariaux et du résultat d'exploitation (rentabilité) au niveau des établissements. Par la suite, les régions défavorisées doivent affronter un désavantage à triple tranchant; la présence des établissements qui ont de mauvais résultats, une structure industrielle et une ventilation des établissements par taille défavorables. Cela laisse supposer qu'il faut une approche stratégique qui doit chercher à neutraliser les inconvénients structurels des régions aussi bien qu'améliorer la compétitivité des établissements existants. ROPER S. (1998) Die Auswirkung des Umfangs von Werksanlagen und Industriemischung auf regionale Produktivität, Lohnkosten und Betriebsrentabilität im Vereinigten Königreich, Reg. Studies 32 , 325‐332. Die Fähigkeit einer Region, Vermögen zu schaffen, hängt sowohl von der Wettbewerbsfähigkeit ihrer einzelnen Firmen als auch von ihrer industriellen Struktur ab. Dieser Aufsatz untersucht die Auswirkung von Industriemischung und Größenstruktur von Werksanlagen auf die Fähigkeit der Regionen des Vereinigten Königreichs, im Zeitraum 1980‐91 Vermögen zu schaffen. Wie auch die Befunde ähnlicher Studien in den USA, legt die Analyse nahe, daß die strukturellen Vorteile der Regionen des Vereinigten Königreichs im allgemeinen eher dazu dienen, Leistungshöhe, Lohnkosten und Rentabilität von Betrieben zu unterbauen als ihnen zuwiderzulaufen. So sehen zurückgebliebene Regionen sich einem dreifachen Nachteil gegenüber: ihre Aufgaben schlecht erfüllenden Betrieben, nachteiliger Industriemischung und Betriebsgrößenstruktur. Dies weist auf die Notwendigkeit hin, strategisch auf Vorgehensweisen hinzuarbeiten, die darauf abzielen müssen, sowohl die strukturellen Nachteile der Regionen zu überwinden wie auch Leistungsfähigkeit der bestehenden Betriebsanlagen zu verbessern.

Technology Locks, Creative Destruction, and Nonconvergence in Productivity Levels

Plant-level data from U.S. textile industries indicate (1) significant cross-sectional dispersion in plant-level productivity within narrowly defined industries, (2) that highly productive plants grow faster and are less likely to exit, (3) dispersion in productivity is larger in industries with more rapid productivity growth, (4) older plants are bigger, and (5) plant births and closures are common to all four-digit textile industries. This paper presents an extended vintage model to explain these facts. The model assumes that a plant incurs a fixed cost of adopting the current best practice and convex costs of adjusting its capital stock. The model provides alternative explanations for the phenomena of investment spikes and S-shaped diffusion. Finally, the model interprets the fact that entry and exit are positively correlated across industries as evidence that variation in plant turnover across industries is driven by variation in technology rather than variation in demand growth.Journal of Economic LiteratureClassification Numbers: L6, O3, O4.

Productivity Adjustments and Learning-by-Doing as Human Capital

This paper measures plant-level productivity gains associated with learning curves across the entire manufacturing sector. We measure these gains at plant startups and also after major employment changes. We find: 1) The gains are strongly associated with a variety of human capital measures implying that learning-by-doing is largely a firm-specific human capital investment. 2) This implicit investment is large; many plants invest as much in learning-by-doing as they invest in physical capital and much more than they invest in formal job training. 3) This investment differs persistently over industries and is higher with greater R&D. 4) Consistent with a learning-by-doing interpretation, the human capital investment is much larger following employment decreases than increases. We conclude that learning-by-doing is a major factor in wage determination, technical progress and asymmetric employment adjustment costs.

Productivity dynamics: U.S. manufacturing plants 1972-1986

This paper presents an analysis of the dynamics of total factor productivity measures for large plants in SICs 35, 36 and 38. Several TFP measures, derived from production functions and Solow type residuals, are computed and their behavior over time is compared, using non-parametric tools. Aggregate TFP, which has grown substantially over the time period, is compared with average plant level TFP, which has declined or remained flat. Using transition matrices, the persistence of plant productivity is examined, and it is shown how the transition probabilities vary by industry, plant age, and other characteristics. This paper describes the results of investigations into the dynamics of total factor produc- tivity (TFP) in large manufacturing plants in high-tech industries from 1972 through 1986. In this paper we build on the work of Dhrymes (1991) and compare econometric with Solow-type measures of TFP. We try to answer fundamental questions regarding the move- ment of plants within the cross sectional distribution of productivity, as time progresses: If productivity in the aggregate improves, does that mean that the constituent plants (for this aggregate) share, nearly uniformly, in this improvement? Does plant level productivity improve steadily over time? Do plants move up or down the productivity rank in complete jumps? Is there a tendency for plants to converge, in their productivity characteristics, either towards the best or worst practice plants, or towards the mean? We present evidence on nearly all of these issues, viz., aggregate versus (plant) average productivity behavior, * The views expressed herein are solely the authors' and do not necessarily reflect those of the Board of Governors of the Federal Reserve System or its staff. This paper was originally written in the summer of 1991 and was presented in the NBER summer workshop (Cambridge) summer of 1991, the summer meeting of the ASA (Boston) in the summer of 1992, and the December meetings of the AEA. We wish to thank the staff and visiting researchers at the Center for Economic Studies, U.S. Bureau of the Census, for providing a stimulating research environment. We would also like to thank two anonymous referees, participants of the NBER Summer '91 workshop, the summer (1992) meetings of the ASA, and the December (1993) meetings of the AEA for valuable comments and suggestions.

A productivity measurement model using a programmable calculator

A comprehensive total factor productivity measurement model has been developed by the American Productivity Center for use in firm-level and plant-level productivity measurement. The computer program for this model is available from the APC in Houston, Texas for use on a main-frame computer system. However, this model has now been programmed for use on a Texas Instruments TI-59 programmable calculator with a PC-100 printer. This makes possible several inexpensive applications of the model, including the use of the model for small business with limited computer availability, use by consultants for developing complex measurement systems, and for training purposes in the application of the model and analysis of results. The article describes total factor productivity, discusses the APC model, and presents the program listing and operating instructions for the programmable calculator version of the model.

Measuring Plant Level Productivity in Batch Manufacture

This paper demonstrates the application of three productivity measurement techniques—added value, productivity costing and total factor productivity. It takes the form of a case study within a British engineering firm producing high precision products in medium/large batches. The advantages and disadvantages of each method are discussed and the empirical results for a five year period analysed. The main conclusion reached is that no single measurement technique is sufficiently comprehensive to meet all requirements.