Local Long-Run Growth Evolutions Across Britain:Some Exploratory Empirics

Local Long-Run Growth Evolutions Across Britain:Some Exploratory Empirics

Ben Gardiner*, Ron Martin**, Peter Sunley***and Peter Tyler****

*Cambridge Econometrics and Department of Geography, University of Cambridge, CB2 3EN, UK

** Department of Geography, University of Cambridge, CB2 3EN, UK

*** School of Geography and Environment, University of Southampton, Southampton, SO17 1BJ, UK

**** Department of Land Economy, University of Cambridge, CB3 9EP, UK

Corresponding author: Pete Tyler, email

April

2014

Acknowledgements:

The research for this paper is part of an ESRC-funded project on How Regions React to Recession: Resilience, Hysteresis and Long Run Impacts (Grant ES/1035811/1)

Abstract

Local Enterprise Partnerships (LEPs) have now been adopted in England as part of the UK Government’s new localism agenda. There is keen interest in understanding how they may be able to make an effective contribution to the narrowing of spatial disparities and, in line with current government thinking, help to spatially rebalance the English economy. If the LEPs are to develop and deliver effective policies it is important to know more about the factors that affect local growth rates. This article seeks to advance this agenda. It uses the shift share technique to assess the contribution of industry-mix relative to local-growth components in explaining variations in growth across LEP areas.

Key Words

Local economic growth policy Economic structure Localcompetiveness

Urbanisation

1. Introduction

Local Enterprise Partnerships (LEPs) have now been adopted as the basis for local economic policy in England as part of the UK Government’s new localism agenda (BIS, 2010). There is keen interest in understanding how they may be able to make an effective contribution to the narrowing of spatial disparities and, in line with current government aspirations help to spatially rebalance the English economy (See Gardiner et al (2013)).However, the dynamics of long-run local growth and their causes remain poorly understood, and many local growth models are untouched by empirical investigation. Yet if the LEPs are to develop and deliver effective policies then it is important to know more about the factors that determine local economic growth. This article begins by examining the economic performance of LEPs, together with thesub-areas of Scotland and Wales[1], over the last thirty years to provide something of a long-run perspective. The article uses the shift-share technique toidentify the contribution of local economic structureand local competitiveness factors to these local differential growth paths. Recent theoretical literature has repeatedly highlighted the importance of industrial structure and agglomeration economies as determinants of local growth, so this paper seeks to examine their relative significance to the growth trajectories of the UK’s LEPs.

2. Long-run evolutions in output, employment and productivity growth across LEPs and areas of Scotland and Wales

There has been considerable variation in the pattern of local growth across local areas in Great Britain since 1981, whether one considers output, employment, or labour productivity. Figure 1 presents this variation by ranking LEP area productivity annual growth rates over 1981-2011, and showing how these are broken down into their output and employment constituents. Whilst some LEPs recorded output (GVA) growth rates of around 3% in others it was below 1%. These differences have meant that a LEP like Northampton expanded its real GVA by over 150% over the entire period.[2] By way of contrast the slowest-growing local economy in terms of output was the Liverpool LEP which expanded by a mere 28% when the annual growth rate is compounded. These are significant differences. The physical distance between two areas is only two hundred miles but in economic terms they are poles apart. The best performing areas have enjoyed an annual increase in their employment of around 1-1.5% per annum. The weakest areas have hardly grown at all, and in Liverpool the employment base actually declined.

Figure 1 Average compound growth rates (1981-2012)

The combination of output and employment provides a measure of labour productivity. London has been the fastest-growing area on this metric with growth of around 2.25% pa, while at the other end of the spectrum regions such as Mid-Wales and Cumbria have grown by only 0.75%pa. Given the importance of productivity in determining long-run differences in GDP per capita (OECD, 2009) these are important performance gaps that need further exploration and explanation.

Figure 2 shows the variation in the growth of output across the local areas of Great Britain over the thirty years examined. In general, the fastest growth of output has been in the South of England although there are also some English LEP areas bordering with Wales that have done well and the effect of North Sea oil on Aberdeen is particularly noticeable.

Figure 2: Overall Output Growth across Great Britain 1981-2011

Figure 3 shows the spatial growth of employment over the period. The pattern is somewhat different to that of output. Although there has been relatively strong employment growth in the South of England including the area around Bristol, there has also been rapid growth in the East, parts of Yorkshire and Wales.

Figure 3:Employment growth in Great Britain over the period 1981-2011

The ability of areas to grow is the result of their relative competiveness in producing goods and services (Kitson, et al, 2004). The growth of their productivity is of central importance (OECD, 2009). Figure 4 shows how the growth of productivity has varied over the period 1981-2010. The strong performance of LEPs in the South of England and parts of the English Midlands is evident.

As a first step we should ask whether local economies have changed and shifted their relative performance through time, or whether their rankings have been immutable. Focusing on productivity, it is striking how persistent productivity differences have been across the LEP areas over the period of analysis. Figure 5. shows that, with a few exceptions, there has been little change in the relative rankings of LEP areas over the thirty years examined, with a correlation coefficient of 0.75 between the 1981 and 2012 rankings. There is also a hint of divergence in

Figure 4: LEP Average Productivity Growth Across Great Britain 1981-2011

performance, as the regions at the top end of the scale mostly improved their position by the end of the period, while those at the bottom fell further behind.

There has been much interest in examining the relationship between the growth of output and productivity. Verdoorn’s law posits a positive relationship because of the belief that there are increasing returns. These can take a number of forms that include scale economies and possible effects on the absorption of technology (Scott, 1991). The Verdoorn relationship has been investigated using a variety of different approaches with a particular focus on models that regress productivity growth on output growth, and also employment growth on output growth, the argument for the latter approach being that it tends to avoid the problem of output being on both sides of the regression models (productivity being output per worker employed). Although not directly of the form examined in the main literature we regressed the growth of productivity across our local areas against the corresponding growth of output. Figure (6) shows the results. There is a statistically significant relationship with the coefficient on the output growth term being around 0.2 which is somewhat lower than the 0.5 found in much other estimation.

Figure 5: Productivity levels in LEP areas in 2012 compared to 1981.

Figure 7 examines how the local areas secured relative productivity growth, by examining whether this productivity was obtained at the expense of employment growth. It shows the growth of productivity in each LEP relative to the national average over the entire period and compares it with the respective relative growth in employment. Clearly, a few areas (e.g. Northamptonshire) have the virtuous position of combining above-average productivity growth while also managing a relatively strong employment performance (top-right quadrant). In stark contrast, Liverpool stands out as one of the LEPs that have relatively under-performed on both productivity and employment, placing it in the bottom left quadrant. In-between these two extremes, some areas like Greater Cambridgeshire and Peterborough (bottom-right quadrant) have experienced strong employment growth but have under-performed on productivity. Finally, some areas like the Black Country have displayed an above-average productivity performance, but have under-performed on employment growth (top left quadrant).

Figure 6. Relationship between Productivity Growth and Output Growth across LEP areas. (Average Annual Growth Rates for 1981-2012).

Figure 7:Relative productivity and employment growth across the LEPs

3. The factors underpinning geographical variations in economic growth: the role of industrial structure.

The evidence in the previous section identified significant differences across local areas in their rate of economic growth since 1981. In the modern world, local economic growth is of course strongly shaped by factors that are national or international in their origin and scale of impact. Such factors include the nation’s exchange rate, the growth of the world economy, industrial lifecycles and changes in technology. However, their impact is nonetheless mediated and conditioned by local economic structures and processes.

Figure 8 shows how growth has varied across economic sectors over the period 1971-2010. It is clear that manufacturing output has declined relative to the general growth of the UK economy. Infact, in 2009 the level of manufacturing outputhad only returned to where it was in 1973-three and a half decades before. Deindustrialization has affected all local areas in the United Kingdom to some degree, although not to the same extent. In contrast, there has been very significant growth in the financial services sector such that real output in that sector is now some 350% above its 1971 base. Local economies with a strong presence in that sector have benefitted from this expansion, with the most notable example, of course, being London (Gardiner, et al, 2013; Martin, 2014).

Figure 8: Sectorally Unbalanced Growth in the UK Economy,1971-2010.

The consequences of Britain’s’ relative decline in manufacturing have been felt the most intensively in country’s long-established industrial cities which led the Industrial revolution.Many of the first casualties were the ‘smoke-stack’ industries which produced large volumes of goods with small unit value added and which competed mainly on price. Over the past fifty years the UK has lost virtually all of these industries. To the extent that these activities were concentrated in specific places there have been very different spatial consequences. To put it simply, the more an area had a structure that was biased to these sectors the more it tended to suffer relative decline.

Box 1. A Dynamic Shift Share Procedure for Decomposing LEP Growth Evolutions

The contribution of industrial structure more generally to economic growth in local areas can be examined using a dynamicshift share decomposition procedure (Gardiner et al (2013).Figures 9-11 show the results of applying thisapproach. The analysis has only been possible at a relatively low level of industrial disaggregation (23 sectors[3]); in part due to the relatively small spatialscale of the LEPs. However, even allowing for this, it is clear that industrial structure effects, while significant in many local areas, are generally far outweighed in importance by locally-specific factors. Traditionally in shift-share analyses, such locally-specific ‘shift’ effects are attributed to ‘competitiveness’, that is to a range of local factors that tend to raise or lower the performance of particular sectors in a local area. Thus while the industrial structure effect captures the extent to which a locality has above average shares of nationally faster and slower growing sectors, the ‘competitiveness’ effect reflects the extent to which locally those same sectors are growing even faster (or more slowly) than their national counterparts.

Figure 9: Contribution of industry mix and locally specific factors to long-run employment change 1981-2012

Figure 10: Contribution of industry mix and locally specific factors to long-run output change 1981-2012.

Figure 11: Contribution of industry mix and locally specific factors to long-run productivity change 1981-2012.

To further examine the impact of industrial composition we calculated a Krugman Specialisation Index for selected LEP areas that have been the subject of particularly extensive change over the period.[4] Figure12 presents the results for four such areas;two that had experienced the most significant growth (Greater Cambridge and Greater Peterborough and Oxfordshire) and two areas experiencing the most dramatic decline (the Black Country and Liverpool). It is clear that differences in industrial structure across these local areas have been steadily declining over time (i.e. their economic structure have become more similar). In fact this convergence is common across almost all LEPS. This reinforces the findings from the shift-share analysis: in general specific local‘competitiveness’factors appear to be the major influence behind the longer–run dynamics of local growth.

Figure 12: The Convergence of Local Economic Structures: Krugman’s Specialisation Index, Selected LEPs, 1981-2012.

4. Possible determinants of the local ‘competitiveness’ component

The evidence presented in the previous section indicates that in most local areas factors other than industrial mix appear to account for the major part of differences in economic growth,and the importance of industrial structure seems to be declining. The local shift or competitiveness component has been responsible for an increasing share of growth and our analysis now turns to the factors that underpin its movement. As argued above, it reflects the extent to which industries in an area grow faster or slower than their national counterpart. Hence it is capturing a local ‘competitiveness’ effect and may thus reflect a number of locally-specific factors like the existence of positive (or negative) externalities of various kinds, including the skill level of a region’s workforce, local market size effects, knowledge spillovers, comparative advantages in access to capital and perhaps the effects of specific government policies (Moore, et al, 1986), as well as other functional and competitive differences between firms within the same industry sector.

The complexity of the local growth process suggests that a search for any one underlying unified theory with which to explain the determinants of the local growth component is unlikely to be successful. Moreover, the causal factors are likely to alter in relevance and intensity through time against a backdrop of the increasing pace of economic and technological change. However, a considerable body of research has been undertaken that provides some helpful pointers to its explanation. Thus, as a recent comprehensive study into the determinants of local growth undertaken by the OECD (2009) argued:

The conclusions from the very significant amount of research and econometric modelling undertaken by the OECD and other researchers is that regional economic development is the result of ‘the interplay between physical capital, human capital and the business environment’ and the ‘benefits of strong interaction between different types of regional assets’(page 46).

Our results align with similar results by Ogus and Skinner (2010) who, using standard shift-share analysis, found that for most of the UK regions differential output growth (relative to the national average) over the boom period 1995-2007 was mainly attributable not to industrial structure but to regional shift effects. This effect implies that firms in different areas differ in technology, management, market access, and labour productivity. In the remainder of this article we extend our analysis in two further directions to help to throw light on the factors that are contributing to variations in local growth. The first is to consider the role of agglomeration. The second is to assess the contribution of enterprise formation.

The impact of agglomeration

In recent years there has been an extensive body of research into the contribution that agglomeration economies make to local economic growth (see for example Fujita, Krugman and Venables, 1999); Fujita and Thisse, 2002; World Bank, 2009; Storper, 2013). The ability of companies to gain increasing returns in urban areas and thus falling average costs has been extensively discussed (see Gardiner, et al, 2011, for a discussion). The increasing returns are argued to reflect the benefits to firms from economies of scale (large markets) and economies of scope (specialisation) and the gains that arise from labour, knowledge spillovers and spillovers with specialist suppliers. NEG models build on the trade-offs between transport costs and the ability of companies to be able to realise the increasing returns associated with agglomeration. An extensive amount of research into the possible benefits of agglomeration and density in the United Kingdom has been undertaken by the Department for Transport (DfT, 2005). The research has sought to identify the productivity gain arising from the change in the level of agglomeration. This is based on the elasticity of productivity with respect to “effective density”. Calculating such elasticities can help to assess the impact of agglomeration enhancing transport investments. Rosenthal and Strange (2004) find that in the USA a doubling of urban size increases productivity by between 3 and 8 percent.