پویایی های کوچک در رشد بهره وری منطقه ای: منبع واگرایی در فنلاند

Despite a rapidly expanding theoretical and empirical literature emphasising the role of incessant intra-industry restructuring in productivity growth, few studies have gone beyond the framework of the representative firm in examining convergence or divergence in regional productivity. We use unique longitudinal plant-level data over a long period of time and apply a useful variant of productivity decomposition methods to study differences in productivity-enhancing restructuring within manufacturing industries among Finnish regions. Long-lasting differences in industry productivity growth between Southern and Eastern Finland can be attributed to the “creative destruction” components of productivity growth, mainly to the between and entry components.

Regional convergence and divergence have gained more notice in Europe in recent years, because deepening economic integration has emphasised the role of regions. These issues have almost exclusively been analysed by means of focusing on overall productivity growth through the use of aggregate data on regions and industries (e.g., Ezcurra et al., 2005 and Martin, 2005), and, as a consequence, the studies have been silent about what happens between firms or plants within industries. Only recently has the literature started to point to the role of firm heterogeneity, firm selection and resource reallocation between firms for economic development (e.g., Bartelsmans and Doms, 2000, Foster et al., 2001 and Klette and Kortum, 2004). In this paper, we investigate the role that these factors may play for regional productivity convergence/divergence. The novelty of this study is the use of unique longitudinal plant-level data over a long period of time and an application of a useful variant of productivity decomposition methods to analyse differences in intra-industry restructuring between Finnish regions.2 The analysis of micro-level dynamics of productivity growth requires longitudinal data on firms, or more preferably plants. Such data, together with a suitable decomposition method, allow one to examine mechanisms of productivity growth beyond the so-called “representative firm model” which has dominated research of regional economic growth. We use a decomposition formula which decomposes industry productivity growth into several distinct sources. The within component indicates the productivity growth rate of the average incumbent plant. The between component gauges the productivity-enhancing effect of intra-industry reallocation of inputs between heterogeneous plants. Other components, closely related to the between component, include the entry and exit components that capture the effect of the turnover of plants on productivity growth. The between, entry and exit components together indicate the role of “creative destruction” in industry productivity growth. In this paper we provide robust empirical evidence that the differences in the intensity of creative destruction within industries explain long-lasting differences in industry productivity growth among regions. Finland is an interesting case, because there has been large and increasing variation in regional performance. As the European Union average is standardised as 100, the level of GDP per inhabitant is 141 in the province of Uusimaa, which is located in the southern, more urbanised part of the country (Appendix: Table A1). Hence, the region of Uusimaa is among the richest regions in the whole of the European Union. In contrast, the same measure reveals that the level of GDP per inhabitant is 75 in Eastern Finland. It belongs to the club of the poorest regions in the EU 15 (Behrens, 2003).3 By using plant-level data we discover that there have been general and sustained differences in productivity growth among regions in 13 Finnish manufacturing industries over the period of 1975–1999. The richest region, Uusimaa, has had the fastest productivity growth. The growth rate of labour productivity (and, as happens to be the case, total factor productivity) for all plants has been 0.9 percentage points higher per year in Uusimaa than in Eastern Finland over the period. This gap does not derive from differences in the industry structures. We show that it has emerged from differences in micro-level dynamics within industries among regions, instead. Perhaps surprisingly, productivity growth of the average staying plant shows no advantage for Uusimaa, since the within component for annual labour productivity (TFP) growth has been even slightly larger in Eastern Finland than in Uusimaa, 2.8%-points (1.1%-points) vs. 2.5%-points (1.0%-points), respectively. One important aspect is that differences in regional productivity growth emerge essentially from the entry and between components, both being roughly equally significant. Both the entry and between components of annual labour productivity growth have been 0.5%-points higher in Uusimaa than in Eastern Finland. The respective numbers for TFP growth are essentially the same. These components are economically significant. For instance, the between component has cumulatively contributed to total factor productivity by 31% in Uusimaa over the period, whereas the corresponding number has been 15% for Eastern Finland. Analyses of trends reveal that the mid-'80s constituted a turning point in regional productivity dynamics. The micro-level restructuring started to fuel aggregate productivity growth, especially in Uusimaa. To sum up, the between and entry components for both labour and total factor productivity uniformly point out that the creative destruction process has been strongest in Uusimaa and weakest in Eastern Finland, especially since the mid-'80s. Because Eastern Finland has always been poorer and has had a lower productivity level than Uusimaa in most industries, creative destruction has contributed to the widening productivity level gap between these two regions since the mid-'80s. In addition to documenting differences in micro-level dynamics of regional productivity growth based on a decomposition method, we discuss at some length the relationship of our results to the theoretical and empirical literature on productivity dynamics that has emerged recently. It turns out that disparities in the level of agglomeration of economic activity and differences in exposure to international trade have most likely contributed to long-lasting regional differences in the between and entry components of industry productivity growth. The paper is organised as follows. Section 2 summarises the literature on regional productivity gaps. Section 3 introduces the productivity growth decomposition method. Section 4 describes the data. Section 5 documents the basic facts about the regional differences in productivity levels and productivity dispersion. Section 6 reports the results from the decomposition of regional productivity growth. Section 7 explores additional explanations for differences in regional industry productivity growth. The last section concludes.

By using plant-level data, large regional differences in productivity levels in manufacturing industries were found. For instance, the level of total factor productivity is roughly 10% higher in the province of Uusimaa, which is located in Southern Finland, compared with Eastern and Northern Finland. In addition, productivity dispersion between plants was found to be larger in Uusimaa than in Eastern Finland. When evidence was found of differences in the heterogeneity of plants between regions, micro-level dynamics of industry productivity growth were analysed by means of a decomposition method. Importantly, there are no statistically significant regional disparities in the average rate of productivity growth for continuing plants. However, the productivity-enhancing reallocation of resources within industries has been substantially stronger in Uusimaa compared with Eastern Finland. The same finding is obtained, irrespective of alternative gauges, for creative destruction (the between or entry components) and alternative measures of productivity (labour productivity or TFP). Hence, Schumpeterian creative destruction characterizes the micro-level dynamics of productivity growth in Uusimaa, which is the richest region. This process has led to an overall and long-lasting regional productivity growth difference in Finland during the past few decades. Broadly speaking, the dynamic perspective on competition and efficiency appears to provide a suitable theoretical framework for understanding the prevailing regional disparities in productivity growth. The finding that productivity dispersion across plants within industries is higher in Southern Finland is in keeping with the perspective that dynamic competition is more intensive in Southern Finland. This might explain why plants use more productive equipment and methods in this high productivity region. In contrast, sluggishness in dynamic competition explains why plants are equipped with low productivity technologies in Eastern Finland. Arguably, the differences in regional restructuring and dynamic competition have been induced by more fundamental forces. By excluding some candidates for factors that could account for the substantial differences in regional performance, such as the differences in the quality of the labour force in the population of plants, we did come to the tentative conclusion that the high level of agglomeration and increasing exposure to international trade have most likely supported more intensive restructuring and dynamic competition in Southern Finland. The prevalence of these effects is in line with the theoretical literature on reallocation. An in-depth analysis of the factors underlying the mechanisms of incessant micro-structural change and hence the sources of regional productivity differences is an important avenue for future research. Comprehensive plant-level data will be an invaluable tool in such work.