Total Factor Productivity Growth in Local Enterprise Partnership Regions in Britain, 1997–2008

Abstract

Harris R. and Moffat J. Total factor productivity growth in Local Enterprise Partnership regions in Britain, 1997–2008, Regional Studies. This paper decomposes aggregate total factor productivity (TFP) growth in Britain for 1997–2008 to show the contribution of different Local Enterprise Partnerships (LEPs) as well as manufacturing and services and UK- and foreign-owned plants within LEPs. These contributions are further decomposed to show the role of productivity growth in continuing plants vis-à-vis reallocations in output shares. The results show that the largest LEPs, in population terms, with higher levels of job density, a greater reliance on manufacturing, more skilled and better qualified workers, and a lower turnover of businesses achieved the highest TFP growth. This is mostly the result of reallocations of output shares towards high productivity plants.

Harris R. and Moffat J. 英国在地企业伙伴区域的全要素生产率成长，1997 年至 2008 年，区域研究。本文分解英国1997 年至2008 年间的集合全要素生产率成长（TFP），展现不同的在地企业伙伴关係（LEPs）、製造与服务业，以及在地企业伙伴关係中英国厂商与外资厂商的贡献。本研究将进一步分解这些贡献，以展现相对于产量比例再配置而言，生产率成长在续存工厂中的角色。研究结果显示，包含最多人口、有着较高程度的就业密集度、较为倚赖製造业、较多的高技术与较佳素质的员工，以及较少企业转移的在地企业伙伴关係，能够获致最高的全要素生产率成长。此则多半是迈向高生产率的工厂时对产量比例进行再分配的结果。

Harris R. et Moffat J. La croissance de la productivité totale des facteurs dans les régions en Grande-Bretagne où il existe un Local Enterprise Partnership, de 1997 à 2008, Regional Studies. Ce présent article cherche à ventiler la croissance globale de la productivité totale des facteurs en Grande-Bretagne de 1997 à 2008 pour démontrer la contribution de divers Local Enterprise Partnerships (LEP – partenariats en faveur de l'entreprise locale) ainsi que de l'industrie et des services et des établissements à capital étranger situés au sein des LEP. Ces contributions sont davantage ventilées afin de démontrer le rôle de la croissance de la productivité dans les établissements existants vis-à-vis la réaffectation des parts de la production. Les résultats montrent que les LEP les plus peuplés, dotés des densités de l'emploi plus élevées, manifestant une dépendance accrue de l'industrie, témoignant d'une dotation plus élevée de travailleurs compétents et mieux qualifiés, et avec une rotation moins élevée des entreprises, ont réalisé la croissance la plus élevée de la productivité totale des facteurs. Cela s'explique dans une large mesure par la réaffectation des parts de la production vers des établissements à haute productivité.

Harris R. und Moffat J. Wachstum der Gesamtfaktorproduktivität von lokalen Unternehmenspartnerschaftsregionen in Großbritannien, 1997–2008, Regional Studies. In diesem Aufsatz wird das Wachstum der Gesamtfaktorproduktivität in Großbritannien im Zeitraum von 1997 bis 2008 aufgeschlüsselt, um den Beitrag von verschiedenen lokalen Unternehmenspartnerschaftsregionen sowie von Produktions- und Dienstleistungsbetrieben und von den Werken in britischem oder ausländischem Eigentum in den Unternehmenspartnerschaftsregionen zu verdeutlichen. Anschließend werden diese Beiträge weiter aufgeschlüsselt, um die Rolle des Produktivitätswachstums in weiterbestehenden Betrieben im Vergleich zu Neuzuweisungen von Produktionsanteilen zu verdeutlichen. Aus den Ergebnissen geht hervor, dass die in puncto Bevölkerung größten lokalen Unternehmenspartnerschaftsregionen mit einer höheren Beschäftigungsdichte, einer wichtigeren Rolle der Produktion, besser ausgebildeten und qualifizierten Arbeitnehmern sowie einem selteneren Unternehmenswechsel das höchste Wachstum der Gesamtfaktorproduktivität erzielten. Dies ist zum größten Teil auf die Neuzuweisung von Produktionsanteilen an Werke mit hoher Produktivität zurückzuführen.

Harris R. y Moffat J. Crecimiento de la productividad total de factores en las regiones con asociaciones de empresas locales en el Reino Unido, 1997–2008, Regional Studies. En este artículo descomponemos el crecimiento agregado de la productividad total de los factores (PTF) en el Reino Unido para 1997–2008 con el objetivo de mostrar la contribución de las diferentes asociaciones de empresas locales así como la manufactura y los servicios y las instalaciones británicas y de propiedad extranjera en las asociaciones de empresas locales. Descomponemos aún más estas contribuciones para mostrar el papel del crecimiento de productividad en las empresas restantes con respecto a las reasignaciones de cuotas de producción. Los resultados muestran que las asociaciones de empresas locales más grandes, en términos de población, con niveles más altos de densidad laboral, mayor dependencia de la manufactura, trabajadores más especializados y cualificados y un volumen más bajo de cambios de empresas consiguieron un crecimiento más alto de PTF. Esto es principalmente el resultado de las reasignaciones de cuotas de producción en plantas con alta productividad.

• Productivity decomposition
• Regional productivity growth

• 生产率分解
• 区域生产率成长

• Ventilation de la productivité
• Croissance régionale de la productivité

• Aufschlüsselung von Produktivität
• Regionales Produktivitätswachstum

• Descomposición de la productividad
• Crecimiento de la productividad regional

JEL classifications:

• C23
• D24
• R12

Acknowledgment

This work contains statistical data from the Office for National Statistics (ONS) which is Crown copyright and reproduced with the permission of the controller of HMSO and the Queen's Printer for Scotland. The use of ONS statistical data in this work does not imply the endorsement of the ONS in relation to the interpretation or analysis of the statistical data. This work uses research datasets that may not exactly reproduce National Statistics aggregates. The authors also would like to thank UK Trade & Investment (UKTI) which supported this research.

Notes

1. According to Krugman (1997, p. 11), ‘Productivity isn't everything, but in the long run it is almost everything.’ Baumol (1984, p. 5) similarly states that ‘without exaggeration in the long run probably nothing is as important for economic welfare as the rate of productivity growth’. Using standard growth-accounting methods, large-scale country and industry studies tend to confirm the importance of TFP and its dominance in explaining differences in output growth across different economies (for example, Organisation for Economic Co-operation and Development (OECD), 2003, fig. 1.2; BERR, 2008a, fig. 6.3; Mourre, 2009, fig. 10; O'Mahony and Timmer, 2009, table 2).

2. Wales and Scotland also saw the devolution of most micro-based economic policies to newly elected legislative bodies in 1999; and in 2002 the UK government proposed further devolution in England to Regional Assemblies (UK Government, 2002) although this did not happen.

3. LEPs were set up in England after 2010 to replace the RDAs, which were created by the new Labour government post-1997. The UK government defines them as follows: ‘Local enterprise partnerships are led by local authorities and businesses across natural economic areas. They provide the vision, knowledge and strategic leadership needed to drive sustainable private sector growth and job creation in their area’ (BIS, 2012b).

4. For a recent overview of the literature on regional economic growth, see Harris (2011). While the 1997–2010 UK government did give priority to the role of productivity, this is not to imply that actual policy instruments (such as Regional Selective Assistance – RSA) were necessarily fully aligned with this position. However, there was clear movement in this direction. For example, in April 2004 the RSA scheme in England was replaced by the Selective Finance for Investment in England (SFIE) scheme, which had a greater focus on increasing productivity (alongside the traditional job creation and maintenance aims of the RSA scheme) in the Assisted Areas.

5. Prior to 1997, the Conservative government in the 1990s operated at both a national (centralized) level, through policy devised and operated through the Department of Trade and Industry (DTI), and at a local level (in England and Wales) with the operation of Training and Enterprise Councils (TECs). In April 2001, the seventy-two English and four Welsh TECs were disbanded. Scotland operated with twenty-one local enterprise companies (LECs), which had a similar remit to TECs, but these were abolished in September 2007 and replaced with Enterprise Regions (ERs). There are currently six ERs and these cover the whole of Scotland. No similar bodies exist in Northern Ireland.

6. The 39th LEP (Buckinghamshire Thames Valley Berkshire) was established in February 2012 before the analysis was carried out. It was therefore omitted from this paper. Unpublished appendix table UA.1 provides details of the local authorities comprising each LEP. The authors have also created five ‘equivalent’ LEP-type areas covering Scotland and Wales; these are defined in unpublished appendix table UA.2 and comprise Aberdeen, Greater Glasgow, Greater Edinburgh, South East Wales and Swansea Bay. Advice on which local authorities should be included in these additional LEP-type areas was provided by BIS and the Welsh Assembly government.

7. It is recognized that some element of RGF spending ‘might include part-funding research and development, training or productivity-boosting technology’ (added emphasis).

8. The definition of the ‘public sector’ is provided by HM Treasury (2010). This paper considers below employment in public administration, defence, health and education across LEPs; the two measures are not the same but it could be argued that both depend significantly on public sector spending (and are largely non-market based).

9. This is confirmed by considering the type of projects that have been funded in rounds 1 and 2 of the operation of the Fund. As National Audit Office (2012, fig. 5) shows, only 23% of funding has gone to support investment in new capital assets, R&D or training; the rest has gone to help business gain access to finance, in support of infrastructure, and to automobile manufacturers. A small amount has gone to programmes and smaller projects.

10. Action 2.3 was to close the RDAs.

11. Once with reference to public sector efficiency; the other reference is included in the vision statement at the start of the document: ‘by removing bottlenecks and making it easier for new businesses to start, we can free opportunities for investment and help generate productivity growth’.

12. There are of course other factors that impact on economic growth beyond the microeconomic determinants of productivity already mentioned. The macro-economy (through short-run aggregate demand for goods and services) will also have an important impact, particularly on short-run growth prospects (short-run because macroeconomic factors are usually geared to the business cycle, although the current fiscal imbalance faced by many governments belonging to the OECD is leading to a longer period of tight fiscal policy until or unless economic growth increases enough to boost tax revenues sufficiently to cover the currently unsustainable level of government spending). Partly this depends on how recession impacts on demand in the private sector of the economy and partly how governments react to downturns in the business cycle. In the usual boom-and-bust scenario, governments tend to tighten their fiscal stance to stop the economy over-heating, while they spend relatively more in recession to underpin the fragile demand in the private sector of the economy.

13. The inclusion of fixed effects is necessary as empirical evidence using plant- and firm-level panel data consistently shows that plants are heterogeneous (productivity distributions are significantly ‘spread’ out with large ‘tails’ of plants with low TFP) and that the distribution is persistent – plants typically spend long periods in the same part of the distribution. Evidence using the ARD has been presented by, for example, Haskel (2000) and more recently by Martin (2008). Evidence from other countries is presented by Baily et al. (1992) and Bartelsman and Dhrymes (1998). Such persistence suggests that plants have ‘fixed’ characteristics (associated with access to different path dependent (in)tangible resources, managerial and other capabilities) that change little through time, and thus need to be modelled.

14. Harris and Moffat (2011, table 1) set out the variables comprising X_it and also review the literature that justifies their inclusion in equation (1)(1) .

15. Output, intermediate inputs, labour, capital, R&D and ‘brownfield’ FDI are treated as endogenous. Thus, lagged (predetermined) values of these variables in levels and first differences are used as instruments and their validity is tested.

16. Further details are given by Harris and Moffat (2011), although the present paper has extended the original analysis based on 1997–2006 to include the years 2007–2008. The range of variables included in X_it in equation (1)(1) cover the age of the plant, external economies of scale, various dimensions of foreign ownership, intra-area spatial spillovers (such as agglomeration and diversification), the undertaking of R&D, and other location and industry effects.

17. See http://epp.eurostat.ec.europa.eu/cache/ITY_SDDS/Annexes/htec_esms_an3.pdf/.

18. For the unpublished appendix, see https://docs.google.com/file/d/0B5AkLHykqQ9VRXJrTlFRb3lNMFk/edit?usp=sharing/.

19. There is also an issue of whether there is a statistical case for disaggregation by industry subgroup. The authors test their a priori expectation – that the eleven industry subgroups as defined are likely to operate under different technologies, and thus that imposing the same factor input coefficients is not appropriate – using simple Welch (1947) t-tests which show that the parameter estimates are statistically different. Across the eleven subgroups, the factor input parameters are statistically different at the 10% or lower significance level (using the high-technology sector as the benchmark) in 77% of cases.

20. This negative sign is imposed in the following tables to make it easier to interpret the results.

21. The ‘between plant’ and cross-plant effects obtained from the Haltiwanger approach were combined into one ‘between plant’ effect. While the separate information is of some interest, of main concern is whether there were changes in TFP due to resource reallocations within or between plants, or through entry and exit.

22. This is more of an issue when dealing with subgroups (for example, based on UK and foreign ownership) where there have been large changes in output shares over the period considered. It is less prominent in Table 2, but becomes more so the greater the disaggregation involved in defining subgroups (for example, unpublished appendix table UA.1).

23. To help understand better the difference between the Haltiwanger and standard approaches, a simple hypothetical example is provided in unpublished appendix table UA.8.

24. The last row of Table 2 also indicates that plants that were opened throughout did not generally contribute through ‘within plant’ productivity improvements (column 3); while on average relatively more (and not fewer) productive plants were closed (column 6).

25. Disney et al. (2003) use a different decomposition approach to that used here. They also cover a different (non-overlapping) time period, but their data source is the ARD (although there are many differences in terms of how the data are constructed and analysed).

26. The sum of all the values in column (1)(1) equals the value in the final row of 1.6% per annum.

27. Harris and Moffat (2012) show that FDI plants significantly outperformed UK-owned plants in terms of TFP growth, and thus there is an expectation that FDI concentrations may be important in explaining the results.

28. Jobs' density is defined as the total number of filled jobs in an area divided by the resident population of working age in that area.

29. After experimentation, it was chosen to measure this as the ratio of the stock of businesses in 2007 divided by 1997 for the following sectors: manufacturing, transport, storage and communications, financial intermediation, and real estate, renting and business activities.

30. Other economic characteristics might also be considered but the paper is limited to the data available in BIS (2012b).

31. A plant could be UK owned in 1997 (or open post-1997 as UK owned), but by 2008 have been acquired by a foreign-owned firm (and vice versa). The authors have checked to ensure that the use of 2008 as the basis for deciding ownership does not present misleading results, since it is possible to classify plants by who owned them in 1997 or 2008. These results are not presented here because (1)(1) they differ little from those in Table 4; and (2)(2) presenting such disaggregated results would lead to a much larger and more complicated table.

32. The Black Country (and to a lesser extent Stoke and Staffordshire) overall did well (Table 2) but poorly when considering foreign-owned plants (Table 4). This shows that the share of LEP output accounted for by foreign-owned plants was relatively small in these areas.

33. The order in which these have been ranked starts with the worst on the basis of the sum of their relative performance in the UK- and foreign-owned subgroups.

34. A simple test was to regress the values in Table 4 (column 2) for UK-owned plants on those for foreign-owned plants; the result was an estimated parameter value of –0.07, which was not significantly different from zero at the 20% level or better (adjusted R² = 0.01).

35. Other work using these relative TFP growth estimates (Harris and Moffat, 2012) has found that foreign-owned plants perform better than UK-owned plant (as discussed in the above text). The results also show that relative TFP growth in manufacturing was 1% per annum on average over 1997–2008, while it was 1.7% per annum for services.

36. A larger set of variables, based on the data available for the LEPs, was in fact included (BIS, 2012b). On the demand side, jobs' density, employment rate, unemployment rate, the percentage of total employees working in manufacturing and in public administration, education and health, and the growth in the stock of value added tax (VAT)-registered business in 1997–2007 (covering SIC codes D, I, J and K) were included; on the supply side, the population aged 16–64 years, the percentage of employees in each SOC (1–9), the percentage of the working-age resident population with NVQ4+ qualifications, and the percentage of the working-age resident population with no qualifications were included. In all cases the earliest date for which data were available (with 1997 the earliest date used where available) was taken.

37. That is, whether β < 0 was tested in the following regression:

•  where X_jt is a (statistically significant) subset of the variables included in Table 3. The actual speed of convergence is measured as: This statistic is often used in conjunction with the so-called half-life time it takes to achieve an elimination of 50% of the gap in the TFP levels across regions. Half-life is given by:

38. These were suggested by an anonymous referee.

39. The data on public sector employment comprised the percentage of the workforce in the public administration, health and education sectors in 1997 and 2008; the other variables were available when estimating equation (1)(1) .