Chapter 1. Regional productivity gaps and their consequences1

While there will always be some form of interregional gaps, those regions lagging behind should have opportunities to “catch up” in terms of social and economic development. This chapter considers the implications of the OECD trends of low levels of national labour productivity growth for different types of regions, including the differences between regions that are catching up to the “frontier” and those that are falling behind. It explores the dynamics of regions in the OECD and the extent to which certain regions are, or are not, catching up. It then addresses the implications of these trends for the well-being of people living in different cities and regions, as the regional and local level are at the nexus of productivity and inclusion. Finally, it outlines the three broad types of public action that can be used to boost productivity in lagging regions and address inclusion. They are: structural policies, public investment (including through regional development policies), and multi-level governance reforms.

  
Key Messages
  • While gaps in GDP per capita between countries have narrowed over the last two decades, within their own borders OECD countries are witnessing increasing gaps in GDP per capita between higher performing and lower performing regions. Leading cities and regions are increasingly competing with their global peers, rather than with others within national borders.

  • The gap within countries between the top 10% regions with the highest labour productivity and the bottom 75% has grown on average by almost 60% over the last two decades, from USD 15 200 to USD 24 000.

  • Three-quarters of “frontier” (highest productivity) regions in OECD countries are predominantly urban. Three-quarters of the regions that were catching up to their country’s frontier regions between 2000 and 2013 are intermediate and rural regions.

  • Tradable sectors emerge as a critical driver in regional catching-up dynamics, particularly tradable services, manufacturing and resource extraction and utilities. This is the case in both urban and rural regions, despite differences in their growth patterns.

  • Productivity growth is important for well-being as it has a significant impact on income, jobs and consequently several other non-material well-being dimensions such as health. One in four people in the OECD lives in a region that is falling behind in productivity growth, and that figure can climb as high as eight in ten people depending on the country. In terms of opportunity, catching-up regions register greater drops than in regions falling behind in terms of unemployment levels and the share of 18-24 year-olds who are not employed, in education or in training (NEETs).

  • Levels of well-being have improved across OECD regions on several indicators, however gaps have widened in many countries on some indicators. Interregional gaps in a multi-dimensional approach to well-being are even wider than for income alone. Complementary policies are important to ensure that productivity growth benefits different social groups and places, including within cities themselves.

  • Actions to boost productivity and social inclusion include: i) structural reforms combined with place-based approaches; ii) public investment drawing on subnational governments as well as regional, urban and rural development policies; and iii) multi-level governance reforms. Good governance is associated with higher levels of productivity and catching-up dynamics. Less fragmentation of local governments is associated with better performance in terms of productivity and inclusion.

Introduction

To address both productivity growth and inclusion, countries need to mobilise the catching-up potential of regions. The goal of regional development policy is to ensure that different types of regions are able to thrive and offer a high quality of life for their residents. There are enormous differences in productivity levels across regions in OECD countries. Often, those differences are much larger than those across countries. These differences may be the result of geographic conditions and cities (agglomeration forces). Therefore, one cannot expect that the gaps will entirely close over time, as it may happen with the process of convergence across countries. However, a productivity gap across regions always signals a potential for catching up. This “advantage of backwardness”, as often coined in economics textbooks, simply means that a lagging region can copy, imitate or import many of the innovations and discoveries produced in the frontier regions and, in this way, boost its productivity and increase growth, without necessarily requiring more labour or capital.

Over the last several decades, many countries have tried different approaches to promote the catching up of those regions lagging behind. The term convergence is often used, but it may imply that the values of different regions or countries are growing closer, but not necessarily for the right reasons. The term “catching up” implies a more dynamic view of regional performance whereby lagging regions are growing faster. However, in some cases regions may be converging, but the “frontier” itself is not growing. Policies should promote the growth of lagging regions, while not cutting off the ability of leading regions to continue to be successful. This chapter therefore explores the implications of firm productivity trends on the productivity performance of regions and the characteristics of those regions that are catching up, or not. It then considers the implications for interregional and inter-personal differences in well-being and inclusion, before outlining three broad areas of public action to consider for addressing both productivity and inclusion.

The role for regions and place-based policies in boosting aggregate productivity

The productivity gap between frontier firms and the rest has widened

Labour productivity growth has been on a downward trend over the last fifteen years across the OECD. By 2000, there was a notable labour productivity growth gap between the United States, Japan and the Euro area (Figure 1.1). It peaked at nearly a 2 percentage point difference between the United States and the Euro area in the early 2000s. Europe’s Lisbon Agenda was an attempt to address these trends, seeking to make Europe the most competitive knowledge-based society by 2010. However, starting in 2004, the United States joined Europe and Japan in their declining rates of labour productivity growth. Before the financial crisis, productivity in all major OECD economies was growing at approximately the same rate of around 1% per year.

Productivity experienced a temporary spike following the crisis, but its growth engine appears to be running out of steam in all major OECD economies. Crises are often processes through which unsustainable trends are stopped, such as over-investment or market price bubbles. It is therefore normal that, by disinvesting in declining productivity sectors and using these resources in more productive ones, average productivity tends to rebound after a crisis or during a recovery. The United States experienced a productivity rebound that peaked in 2010, given its flexible labour market permitting more rapid and deep labour reallocation across firms, sectors and places. Europe, with more rigid product and labour markets, did not peak again until a couple of years later. The same happened in Japan. However, the rebound in the United States was short-lived, and all three (United States, Europe and Japan) were back down to productivity growth levels of below 1% by 2014.

Recent OECD research on the “Future of Productivity” finds that the problem is not that all firms are experiencing slow productivity growth, but rather the diffusion of productivity from the top firms is not reaching the others (OECD, 2015a). A decomposition of productivity growth by type of firm shows that the top firms, those at the “frontier”, show continued increases in productivity (Figure 1.2).2 These findings are true for both manufacturing and service sectors. The service sector accounts for the bulk of the knowledge economy, and it displays the most dramatic productivity growth differentials at 5% per year for the top firms and 0.3% for all firms, with non-frontier firms actually showing negative productivity growth (-0.1% per year). Most of the contribution to aggregate labour productivity comes from the catching up of firms, sectors and regions. These findings may therefore explain why there has not only been an overall slowdown of labour productivity growth, but also why there are increasing inequalities (i.e. growth has been less inclusive).

Figure 1.1. Labour productivity growth trending downward even before the crisis
picture

Note: Values represent three-year moving averages (t, t-1, t-2) of labour productivity (GDP per hour worked) 1997-2014. GDP refers to the gross domestic product, in USD, at constant prices, constant PPPs, OECD base year 2010. Total hours worked are derived for all persons as average hours worked from the OECD Employment Outlook, OECD Annual National Accounts, OECD Labour Force Statistics and national sources, multiplied by the corresponding and consistent measure of employment for each country.

Source: Calculations based on OECD (2016a), Productivity Statistics (database), www.oecd.org/std/productivity-stats/ (accessed 17 March 2016).

 http://dx.doi.org/10.1787/888933411597

Figure 1.2. Productivity gaps between frontier firms and other firms are widening
Labour productivity; index 2001 = 0
picture

Note: “Frontier firms” corresponds to the average labour productivity of the 100 globally most productive firms in each 2-digit sector in the ORBIS database. “Non-frontier firms” is the average of all other firms. “All firms” is the sector total from the OECD STAN database. The average annual growth rate in labour productivity over the period 2001-09 for each grouping of firms is shown in parentheses.

Source: Andrews, D., C. Criscuolo and P.N. Gal (2015), “Frontier Firms, Technology Diffusion and Public Policy: Micro Evidence from OECD Countries”, OECD Productivity Working Papers, No. 2, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jrql2q2jj7b-en.

 http://dx.doi.org/10.1787/888933411605

The full explanations for this diffusion challenge are still to be found (Box 1.1). They may include the “winner-takes-all” markets surrounding new technologies or the fact that replication of certain innovations has become more difficult. Firms need to have many different capabilities to succeed, such as technological capacity; capabilities in branding, marketing, and managing; being part of global value chains (i.e. importing intermediate products, exporting parts or final products); etc. For regions in countries, the emergence of global value chains may shift productivity spillovers from leading regions to foreign countries rather than other regions of the same country. Indeed, one of the characteristics of the current wave of globalisation is the possibility to disconnect the creation of knowledge from its use. Lagging regions in high-cost countries compete increasingly with regions that have similar capabilities in middle-income countries.

Box 1.1. The global innovation “diffusion machine” for productivity

According to recent OECD research, the productivity problem is not the lack of innovation on a global scale, but rather the performance of the rest of the economy to adopt new technologies and best practices. Indeed, as stated in Criscuolo (2015), “… the main source of the productivity slowdown is not a slowing in the rate of innovation by the most globally advanced firms, but rather a slowing of the pace at which innovations spread throughout the economy: a breakdown of the diffusion machine.”

Why this process of diffusion may be more difficult during the current technological revolution (recently called digitalisation), than in the previous periods of major technical progress, is still a topic for much debate and research in economics. Both the global and national frontiers play a role in diffusion of innovation to other firms throughout the economy, as depicted in the figure below.

Stylised depiction of aggregate productivity growth
picture

The shift in the global frontier can be transmitted to national frontiers through the mobility of production factors (capital and labour) and trade flows. Within countries, the investment in knowledge-based capital (KBC) and all actions favouring spillovers and adoption may facilitate the diffusion of the frontier innovations to lagging firms, sectors or regions. This process is facilitated by a macro-structural environment that supports, rather than hinders, the shift of resources across sectors and the upscaling of best productivity practices.

Source: Criscuolo, C. (2015), “Productivity Is Soaring at Top Firms and Sluggish Everywhere Else”, Harvard Business Review, 24 August 2015, https://hbr.org/2015/08/productivity-is-soaring-at-top-firms-and-sluggish-everywhere-else; OECD (2015a), The Future of Productivity, http://dx.doi.org/10.1787/9789264248533-en.

The existence of persistent interregional disparities is not a new fact, but recent trends reveal greater differences within than across countries. As economic activities tend to concentrate in space, agglomeration economies (see later discussion) may create advantages leading to higher per capita GDP in urban regions over intermediate and rural regions. These disparities have been largely documented in the previous Regional Outlooks (OECD, 2011a, 2014a) and in OECD Regions at a Glance (OECD, 2016b). Economic disparities, measured in terms of per capita GDP, have slightly increased or remained stable, while there has been a steady reduction over the past decades of average disparities across countries (Figure 1.3). The same trend is found among metropolitan areas, as cities across the OECD are converging, while within countries, cities are diverging (Figure 1.4). The speed of convergence among the OECD countries’ metropolitan areas is slightly faster than that of countries as a whole, pointing further towards the importance of large cities for their national economies. The international links of these cities in the knowledge economy, combined with the international mobility of financial capital and of highly-skilled workers, also mean that large metropolitan areas have to adapt to competition at the global level.

Figure 1.3. Country convergence has been accompanied by divergence of regions within countries
picture

Note: Data refers to GDP PPP constant 2010 USD from the national accounts and the regional accounts; the disparity between countries is measured as the coefficient of variation of national GDP per capita across all countries in the sample; the disparity within countries is measured as the coefficient of variation of regional GDP per capita across regions within each country, and then is averaged across all countries. Data for 1995-2013. Countries included: Australia, Austria, Belgium, Canada, Chile, Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Japan, Korea, Netherlands, New Zealand, Poland, Portugal, Slovak Republic, Slovenia, Spain, Sweden, United Kingdom, and the United States (District of Columbia is excluded).

Source: Bartolini, D., H. Blöchliger and S. Stossberg (2016) “Fiscal Decentralisation and Regional Disparities”, Economics Department Working Paper (forthcoming).

 http://dx.doi.org/10.1787/888933411619

Figure 1.4. As metro areas across countries converged, metro areas within countries diverged
picture

Note: Data refer to per capita GDP PPP constant 2010 USD; the disparity between countries is measured as the coefficient of variation of national per capita GDP across all countries in the sample, between countries (metro area only) disparity is measured by the coefficient of variation for the country average of metro area per capita GDP; the disparity within countries is measured as the coefficient of variation of metro area per capita GDP across metropolitan areas within each country, which is then averaged across all countries. Data for 2001-12. Countries included: Australia, Austria, Belgium, Canada, Chile, Czech Republic, France, Germany, Italy, Japan, Korea, Netherlands, Poland, Spain, Sweden, United Kingdom and the United States.

Source: Calculations based on OECD (2016c), “Metropolitan areas”, OECD Regional Statistics (database). http://dx.doi.org/10.1787/data-00531-en (accessed 20 June 2016).

 http://dx.doi.org/10.1787/888933411622

Growing inequality across regions is mirrored by growing inter-personal income inequality in most countries. Since 1985, income inequality across households, measured as the Gini coefficient of disposable household income, has decreased in only 1 out of 22 OECD countries with available long-term data (Figure 1.5).3 In four more countries, inter-personal inequality changed only marginally. The majority of countries, as well as the OECD average, experienced a significant increase in income inequality. Aggregate inequality hides strong growth in the disparity between the top and the bottom of the income distribution. The recent crisis amplified the growing gap in some countries. In Spain, for instance, incomes of the poorest 10% dropped by almost 13% per year, compared to a drop of 1.5% for the richest 10%. In about half of the countries where incomes continued to grow, the gap nevertheless widened as the top 10% did better than the bottom 10%. In some countries, including Austria, Denmark, and the United States, top incomes grew, while bottom incomes declined in real terms (OECD, 2015b).

Figure 1.5. Income inequality increased in most OECD countries, but the crisis halted the trend in some countries
Gini coefficients of income inequality, mid-1980s and 2013 (or latest date available)
picture

Note: “Little change” in inequality refers to changes of less than 0.015 points in the Gini coefficient.

Source: OECD (2015b), In It Together: Why Less Inequality Benefits All, http://dx.doi.org/10.1787/9789264235120-en based on OECD (2016d), Income Distribution (IDD) (database), www.oecd.org/social/income-distribution-database.htm.

 http://dx.doi.org/10.1787/888933411632

The regional “catching-up machine” needs to be fixed

Regions with lower GDP per capita in their countries are not sufficiently benefiting from their catching-up potential. While the absolute convergence of regional per capita GDP and productivity is not an aim in itself, the fact that many “lagging” regions are not catching up indicates an untapped growth potential. Despite the overall increase of regional disparities across the OECD, in terms of per capita GDP, some convergence forces were at work in intermediate and rural regions during the period 1995-2007.4 In other words, those categories of regions with lower initial levels of per capita GDP experienced higher growth rates. In urban regions two different types of dynamics were observed: i) there was a catching up of lagging urban regions in some cases (convergence forces), while ii) certain leading regions experienced higher growth rates (agglomeration forces) (OECD, 2011a).

The crisis appears to have stalled interregional catching up. With the recent crisis, these regional convergence trends appear to have broken down (OECD, 2014a). The slowest growing regions (in terms of per capita GDP) experienced low growth rates in productivity and in labour utilisation that have hindered the catching-up process (OECD, 2016b). In contrast, labour productivity has continued to be the main driver of per capita GDP growth for the 50 fastest growing OECD regions. In 41 out of the top 50 regions, labour productivity growth accounted for 75% or more of the rise in per capita GDP during the period 2000-13.

The productivity gap between the “frontier” regions and the majority of other regions has widened over the last two decades. From 1995 to 2013, labour productivity (as measured by per worker GDP)5 across the OECD grew by 1.6% per year among those “frontier” regions with the highest labour productivity (Box 1.2). The lagging regions at the bottom of the labour productivity distribution fell further behind the frontier as productivity grew by less than 1.3% per year.6 While the difference (0.3% per year) may not seem high, over time, these productivity growth differentials have translated into substantial gaps. Over the last two decades (1995-2013), the gap widened by almost 50%, from USD 21 000 to USD 31 000 PPP per worker. However, it is not only the lagging regions that experienced lower growth rates, as productivity in the bottom 75% of regions (i.e. in the vast majority) also grew by only 1.3%, which widened the gap between the top 10% and the bottom 75% regions by nearly 60% (from USD 15 200 to USD 24 000) (Figure 1.6). In other words, the problem seems to be the lack of catching up, rather than a lack of growth in the frontier. The leaders are breaking away from the pack. This trend is consistent with the findings of the aforementioned OECD study on the “Future of Productivity” (OECD, 2015a).

Figure 1.6. Productivity growth of frontier regions in a country outpaces that of most other regions
Averages of top 10% (frontier), bottom 75%, and bottom 10% (lagging) regional GDP per employee, TL2 regions
picture

Note: Average of top 10% and bottom 10% TL2 regions, selected for each year. Top and bottom regions are the aggregation of regions with the highest and lowest GDP per employee and representing 10% of national employment. Due to lack of regional data over the period, only 19 countries are included in the averages. GDP per employee in constant PPP and constant prices 2010 USD.

Source: OECD (2016b), OECD Regions at a Glance 2016, http://dx.doi.org/10.1787/reg_glance-2016-en.

 http://dx.doi.org/10.1787/888933411648

Box 1.2. Defining the productivity frontier

Productivity measures how much can be produced with a given amount of inputs, i.e. capital and labour. Improvements in productivity therefore mean that more can be produced with the same inputs or the same amount of output can be produced with fewer inputs. For example, a delivery driver does not become more productive by working an additional hour, but rather by optimising the delivery route to finish in a shorter amount of time, which allows for additional deliveries without working more hours. In other words, it means working smarter, not working more.

Raising productivity is essential for long-term growth and increases in living standards. Capital investment and improvements in human capital (e.g. through raising educational attainment) can create growth, but returns to investment are typically decreasing, i.e. each additional increase yields a smaller benefit than the previous investment. This is why the Nobel Prize winning economist Paul Krugman famously remarked: “Productivity isn’t everything, but in the long run it is almost everything. A country’s ability to improve its standard of living over time depends almost entirely on its ability to raise its output per worker.” (Krugman, 1997).

The labour productivity “frontier” captures the potentially attainable productivity level of a region. “Potentially attainable”, as of course many factors, such as the sectoral composition, geographic characteristics and the distance to markets all influence the ability for a region to reach that frontier potential, at least in the short term. The frontier is based on observed levels of productivity in the most productive region(s) in each country. The focus on the within-country frontier (rather than a global frontier) accounts for institutional and country-level differences that might affect the productivity potential. The frontier is defined by the productivity level in the most productive region(s) in a country accounting for 10% of the country’s total employment. This option was chosen to ensure that the frontier in any country did not represent only one region with a small population size. In some countries, therefore, the calculation of the frontier is the weighted average (based on employment) of more than one region. The region(s) that form the frontier can change over time. To ensure that the group of “frontier regions” is not affected by time-dependent outliers, only regions that contribute a non-negligible percentage of their employment for several years during the 2000-13 period are labelled as “frontier regions”.

The discussion in this chapter focuses on labour productivity measured as real gross domestic product (GDP) per employee. There remains room to improve the measurement of labour productivity at the subnational level. Typically labour productivity is measured in terms of hours, rather than in terms of the number of employees. This measure takes into account productivity improvements that allow for the reduction in the time each employee spends at work. Differences between the two measures arise when there is a high incidence of part-time employment, such as in Germany or the Netherlands, or low statutory hours, such as in France (OECD, 2016e). But estimates for the total number of hours worked are often unavailable at the subnational level. As regions in this chapter are compared to their national frontier, cross-country differences (such as statutory hours) should not affect the analysis. A similar issue arises for subnational price indexes. Typically only national level price deflators are available and used to calculate real GDP at the regional level. This result can confound price changes with productivity changes if a region’s sectoral specialisation differs strongly from the national average. Price fluctuations that disproportionately affect some regions can erroneously be measured as changes in labour productivity. For the most part, the measurement error should be minor, but could be relevant for small and resource-intensive regions. In some cases, industrial level price deflators can be used to alleviate the potential error (e.g. when considering real gross value added by sector), but these are not consistently available for all OECD countries.

Source: Krugman, P. (1997), The Age of Diminished Expectations: U.S. economic policy in the 1990s, 3rd edition, MIT Press; OECD (2016e), OECD Compendium of Productivity Indicators 2016, http://dx.doi.org/10.1787/pdtvy-2016-en.

A region’s productivity growth does not automatically benefit from strong frontier performance

Despite the widening average gap between the frontier and the bottom distribution of regions across the OECD, many regions are still catching up vis-à-vis their country-specific frontier. A specific indicator was computed to measure this within-country convergence effect. It is based on the Malmquist Index (see Box 1.3) and generalises the idea that a region needs to grow faster than the national frontier to reduce its productivity gap. Using this indicator, regions can be classified as those that are catching up (converging) and those that are falling further behind (diverging) (Figure 1.7, Table 1.A1.1). Furthermore, this productivity growth performance can be decomposed into a “frontier” and a “catching-up” effect. If productivity growth rates do not change, catching-up regions will not be able to close the gap to their frontier, on average by 2050. But without a change, this also means that during the same period diverging regions will have fallen to about 50% of the productivity in the frontier. To close the gap in the next 34 years, diverging regions would need to outgrow their frontier by about 1.2 percentage points. Put differently, the average labour productivity growth in diverging regions would need to increase to 2.8% per year, quadruple the current rate.

Box 1.3. How to measure regional catching up

A simple and often used way to measure the performance of a given region is to assess whether it grows faster than the country average. However, this measure can be quite misleading. Ultimately, the “frontier”, i.e. the most productive region, sets a precedent for the potential levels of productivity that regions can achieve. Suppose that in many regions productivity grows slowly, but the most productive region is outperforming, the average might indicate that there is general convergence, but regions are de facto diverging from the frontier. The right measure of performance in this case is whether there is any region growing faster than that frontier, potentially benefiting from innovations produced there.

A concise measure for the “catching up” of a region is the ratio between its own productivity growth and the productivity growth in the country’s frontier. The ratio measures by how much the “gap” between the frontier and the region has narrowed (or widened). Assuming that both regions produce with constant returns to scale, i.e. a doubling of inputs leads to a doubling of output, the “gap” has two interpretations. It captures how much more the frontier region would produce with the same input as the other region, and the inverse captures how much less factor inputs the frontier would have required to produce the same amount of output as the other region. The narrowing of the “gap” and therefore the concept of convergence or “catching up” towards the frontier can be computed as illustrated in the figure below.

Schematic representation of regional catching-up dynamics
picture

Assume that regional GDP is produced with employment only. In period t, a given lagging region is below the frontier, as displayed in the left panel. In the next period (t+1), both the frontier and the lagging region move upwards on the chart, increasing their productivity. If the lagging region is catching up, the distance between the frontier and the lagging region is smaller in period t+1. The productivity gap can be measured in two ways: i) productivity increased by increasing output, maintaining the same level of employment (output-oriented), or ii) the same level of output was obtained with less employment (input-oriented). If one assumes a linear frontier (in other words, if the technology exhibits constant returns to scale) both measures are equivalent. With constant returns to scale, this “catching up” of a region is also equivalent to the Malmquist Index (cf. Malmquist, 1953 and Caves, Christensen, Diewert, 1982) and can be computed as follows (right panel of the figure):

picture

When the index CU is bigger (smaller) than 1, the region is catching up to (diverging from) the frontier. Given that the shift in the frontier corresponds to the change in the slope from 0C to 0E, the increase in productivity of the region can be decomposed as:

1 + regional productivity growth = (1 + productivity growth of the national frontier) x CU

Taking the natural logarithm on both sides of the equation then results in a sum with each term measured in percentages.

These calculations are then used to derive the categories of regions based on their productivity performance. To avoid threshold effects around the value of 1 for the catching-up indicator, “catching-up regions” are defined as outgrowing the frontier by 5 percentage points and “diverging regions” as growing more slowly than the frontier by at least 5 percentage points between 2000 and 2013, i.e. the catching-up regions correspond to a Malmquist index of 1.05 or higher and the diverging regions to an index of 0.95 or less. Keeping-pace regions are those that had an indicator value of between 0.95 and 1.05.

Source: Elaboration based on Malmquist, S. (1953), “Index Numbers and Indifference Surfaces”, Trabajos de Estatistica, Vol. 4, pp. 209-242 and Caves, D., W.L. Christensen and E. Diewert (1982), “Multilateral Comparisons of Output, Input, and Productivity Using Superlative Index Numbers,” Economic Journal, Royal Economic Society, Vol. 92(365), pp. 73-86.

Figure 1.7. Patterns of catching up and divergence differ across countries
Classification of TL2 regions according to their labour productivity growth relative to their country’s frontier, 2000-13
picture

Note: The classification of regions is outlined in Boxes 1.2 and 1.3. The period covered is 2000 to 2013 (or closest available year) and countries included are Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Korea, Netherlands, New Zealand, Poland, Portugal, Slovak Republic, Slovenia, Spain, Sweden, United Kingdom and United States. For New Zealand TL3 regions and for Belgium, 10 provinces and the capital city region instead of TL2 regions are used. Exclusions of OECD countries are due to missing data or due to data only being available for a single region.

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 30 May 2016), using national boundaries provided by National Statistical Offices and FAO Global Administrative Units Layer (GAOL).

 http://dx.doi.org/10.1787/888933411658

Regions with high productivity growth are located in countries with fast-growing frontiers. The productivity growth of the top 50 regions across the OECD is decomposed between the effect of the national frontier shift and the catching-up dynamics (Figure 1.8). Most of the regions with high productivity growth rates have benefited from the potential pulling effect of the frontier region(s) to which they have converged. Many Polish regions, for example, experienced strong productivity growth alongside the strong growth of the country’s frontier. Only in Portugal and the United States is the frontier effect relatively small. In contrast, for the bottom 50 regions, most of their poor productivity performance is the combined result of a low performance of the national frontier region(s) and the lack of catching up. The notable exceptions are regions in the bottom 50 from Canada, Australia and the Netherlands that performed poorly due to the lack of catching-up effects. What unifies these regions is that the part of the frontier growth in productivity is driven by regions that are relatively specialised in resource extraction. Without an endowment in similar resources, imitation and adoption of frontier technologies is likely to yield lower returns in other regions since they require a transfer across sectoral boundaries. For example, optimisation of supply chain management in the mining sector, might be transferable to manufacturing, but likely not to its full extent and other innovations, e.g. a new drilling technology, might benefit an even smaller group of sectors in other regions.

However, a region’s productivity growth does not automatically benefit from strong frontier performance. Considering the top 50 and bottom 50 regions in terms of catching up to their country’s frontier, there are examples of strong catching-up dynamics that are supported by a strong frontier, such as in Poland and the United States (Figure 1.9). But there are also cases of regions catching up in countries where the frontier regions are underperforming and the region grows at a moderate pace, such as in Germany or Austria. Importantly, several Spanish and Portuguese regions are among the top 50 catching-up regions, defying the weak aggregate growth in their countries. This is contrasted by Greece, where most regions are slowly falling behind their frontier and two regions (Central Greece and South Aegean) are among the 10 fastest diverging regions. The bottom 50 regions also include those in Canada and Australia, where transfers from the frontier are less direct than in other countries. Examples of some of the catching-up regions are described in Box 1.4.

Figure 1.8. The top 50 OECD regions for productivity growth tend to be in countries with a strong frontier
Top 50 and bottom 50 regions in the OECD in terms of productivity growth by source, 2000-13
picture

Note: Lighter coloured bars indicate regions that are part of their country’s labour productivity frontier (see Box 1.2 for a detailed description). The productivity growth is decomposed into a frontier-shift and a catching-up effect (see Box 1.3 for details). In some countries, the frontier consists of more than one region. In those cases, frontier regions can catch up or diverge from the (composite) frontier if they grow faster or slower than the other frontier regions.

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411666

Figure 1.9. The frontier does not necessarily stimulate catching-up dynamics in all regions
Top 50 and bottom 50 regions in the OECD in terms of catching up and divergence, 2000-13
picture

Note: Lighter coloured bars indicate regions that are part of their country’s labour productivity frontier (see Box 1.2 for a detailed description). The productivity growth is decomposed into a frontier-shift and a catching-up effect (see Box 1.3 for details). In some countries, the frontier consists of more than one region. In those cases, frontier regions can catch up or diverge from the (composite) frontier if they grow faster or slower than the other frontier regions.

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411674

Box 1.4. Catching up regions: Examples from Poland and Spain

Productivity in Poland has increased rapidly between 2000 and 2013. Among the fastest growing TL2 regions is Małopolska Voivodeship (“Lesser Poland”) and its main city Krakow, the third largest metropolitan area in Poland (OECD, 2016c). Labour productivity in the region grew by nearly 4% per year, which was supported by a fast growing frontier (2.4%). Many factors contribute to the successful growth in Lesser Poland, but two factors stand out.

First, the sectoral structure is gradually shifting towards higher value-added activities in the tradable sector. By 2012, the share of employment in agriculture fell to less than half of the share in 2000 and the contribution of manufacturing and tradable services increased. Manufacturing plays a less important role in this shift than in other Polish regions. Instead IT and business services have been rapidly expanding in the region. Employment in the sector increased by 19% per year from 16 000 to 38 000 employees in 102 major IT and business service centres in Krakow alone (ASPIRE, 2015).

The second factor is the role of educational institutions in the region. The vocational school system in the region is leading across Poland in terms of achieving the highest results in professional examinations in Poland (OECD, 2013a). But, it is the university system that stands out for research and innovation. Lesser Poland’s R&D expenditure rate was 1.3% in 2013, higher than the Polish average and the second highest rate in the country. More than one-third of the total expenditure comes from the higher education sector, which used to be the main major contributor. Since 2012, business expenditure on R&D caught up to the same level as R&D in higher education and the two account for about 80% of the total (OECD Regional Statistics database). The higher education sector also accounted for more than 70% of funding allocated to projects in the region from the European Union’s 7th framework programme for European Research and Technological Development (EC JRC IPTS, 2015).

Labour productivity in Castile-La Mancha, a rural Spanish region in the centre of Spain that surrounds Madrid to the east and south, grew by 1.6% per year between 2000 and 2013. While overall productivity growth was significantly slower than in Poland, the “catching up” to the Spanish frontier was fast, at an annual rate of about 1% (compared to 1.5% in Lesser Poland). As in most rural areas, agriculture plays an important role, but the contribution and growth of manufacturing and services stands out compared to other parts of Spain and Europe. The concentration of traditional manufacturing (in particular textiles, food and beverages and products for the house) was a strong driver of growth in the 1990s. Out of 52 industrial districts in rural areas in Spain, 32 were located in the region and the employment growth in these clusters added 57 000 jobs between 1991 and 2001 to the regional economy (OECD, 2009).

Since the 2000s there has been a shift away from traditional manufacturing and the service sector has gained momentum. Between 2001 and 2011, the share of employment in knowledge intensive services increased by more than 60%, one of the fastest expansions of the sector across Europe, albeit starting from one of the lowest levels of concentration (23%) in 2001 (see the figure below). The shared border with Madrid (part of Spain’s productivity frontier, together with the Basque country) creates the potential for significant spillovers. The fastest productivity growth within Castile-La Mancha occurred in Guadalajara, where the western part of the region is part of the OECD metropolitan area of Madrid. But direct connections are not the only factor that supports growth in labour productivity, Albacete was the second fastest growing TL3 region within Castile-La Mancha and it is also the part of Castile-La Mancha that is the furthest away from Spain’s capital city (OECD Regional Statistics database).

Castile-La Mancha has one of the fastest growing knowledge-intensive service sectors in Europe
picture

Source: Calculations based on Boix, R. (2014), Background report: OECD Territorial Review of Bergamo, mimeo and EUROSTAT (2014), Structural Business Statistics – Regional Data – All Activities (database), EUROSTAT.

 http://dx.doi.org/10.1787/888933411680

Source: Elaboration based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 30 May 2016); OECD (2016c), “Metropolitan areas", OECD Regional Statistics (database). http://dx.doi.org/10.1787/data-00531-en (accessed 20 June 2016); ASPIRE (2015), “ASPIRE Headcount Tracker, 2015”, www.aspire.org.pl/ht2015/; EC JRC IPTS (2015), Stairway to Excellence – Facts and Figures: Małopolskie, Institute for Prospective Technologies, Joint Research Centre, European Commission; Perek-Białas J., C. Martinez-Fernandez and T. Weyman (2013), “Malopolska Region Demographic Transition: Working for the Future”, OECD Local Economic and Employment Development (LEED) Working Papers, No. 2013/06, http://dx.doi.org/10.1787/5k4818gwg2jk-en; Boix, R. (2014), Background report: OECD Territorial Review of Bergamo, mimeo; EUROSTAT (2014), Structural Business Statistics – Regional Data – All Activities (database), EUROSTAT; OECD (2009), OECD Rural Policy Reviews: Spain 2009, http://dx.doi.org/10.1787/9789264060074-en.

Rather than curtailing the performance of the high productivity (frontier) regions, policies should aim to encourage diffusion. The large share of French regions among the bottom 50 stands out: 12 of the 22 French regions are among the fastest diverging regions in the OECD. In contrast, only 2 of the 12 UK regions are part of the bottom 50, despite similar productivity growth in Greater London (1.3%) and Île de France (Paris, 1.15%). Nevertheless, the productivity gap between Greater London and Wales is 1.6 times the size of the gap between Île de France and Limousin. Both cases show how important it is to consider the system of regions when analysing and designing policies for regional convergence. Ensuring that the frontier regions play fully their role and continue to perform should be part of any strategy to promote catching up among the lagging regions. However, it is unlikely that catching up happens automatically. Unlocking catching-up potential requires policies that facilitate the diffusion of innovation and support regional development in general (see Chapter 2).

Different forms of proximity may facilitate diffusion of innovation from the frontier to increase productivity, such as geographic and technological proximity. For example, recent work finds that GDP per capita growth in regions7 is higher in regions within short drives from large metropolitan areas (Ahrend and Schumann, 2014a). The benefit declines by about 0.3 percentage points of annual GDP per capita growth with each doubling of the time it takes to access the metro area. But, distance is not only physical distance, sectoral distance can be a hindrance as well. There are many forms of proximity relevant to the innovation process in firms (Boschma, 2005). There is also a large amount of literature on the role of spatial proximity in innovation diffusion, typically referring to knowledge spillovers with respect to a concentration of firms (often in the same sector), human capital characteristics, R&D activities, or patents and patenting citations.8 General purpose technologies, including ICT, are able to boost productivity across sectors, facilitating innovation diffusion (Box 1.5) and inventions in one area can be adopted for others, such as the use of drones in agriculture. However, product innovations in high-tech and medium-high-tech firms are often complex and require a variety of specialised skills. Even the adaption of these innovations might not be straightforward and replication more complex than in the past (as discussed above). Taking again the case of Canada and Australia, the more limited impact of the frontier may be related to such factors. In both cases, the frontier regions in those countries are relatively specialised in mining, gas or oil production which may limit the transmission of productivity benefits. In addition, the economic centres of frontier regions are located far from other regional centres.

Box 1.5. ICT: Spillovers across sectors to boost productivity

The gains from ICT innovation are not limited to the ICT sector. More “traditional” sectors such as manufacturing are also absorbing innovations and are contributing to overall productivity gains while also innovating themselves. What do these innovations look like? In the manufacturing sector, a leading utility vehicle maker is incorporating new technologies and processes into its products. It is integrating sensors in existing equipment, which will help farmers reduce the downtime of their equipment, leading to more efficient use. Also, by synchronising tractors with GPS technology, the firm enables productivity increases and fuel savings due to better pathfinding. Other innovations allow better monitoring and tailoring of agricultural practices, leading to better use of resources and higher crop yields.

This example shows how some innovations are not limited to some sectors, but rather spillover into other sectors to boost productivity. ICT technologies improved products essential to agricultural production, improving productivity in the primary sector, linking services to manufacturing and finally agriculture. This only works, however, with sufficient absorption capacity. Firms and employees from various sectors need to have the skills and the institutional setup enabling them to make the most of new technologies.

Source: OECD (2015c), Data-Driven Innovation: Big Data for Growth and Well-Being, http://dx.doi.org/10.1787/9789264229358-en.

Drivers of growth and catching up are different in urban and rural areas

The productivity frontier is mostly urban. Three-quarters of the most productive regions in the 24 OECD countries with available data are mostly urban regions. Among mostly urban frontier regions 70% contain a large metropolitan area, which is often also the capital city (Figure 1.10). Another 20% also contain their capital city, but the capital itself has less than 1.5 million inhabitants. The remaining quarter of frontier regions are intermediate or mostly rural regions, which are typically resource rich regions (mining, oil and gas extraction), e.g. in Australia and Canada.

Figure 1.10. Frontier regions tend to be urban, but catching-up regions tend to be rural or intermediate
picture

Note: Numbers in parentheses indicate the number of regions in the group. For mostly urban regions the patterned part of the bar indicates the share of regions that contain (part of) a large metropolitan area with 1.5 million or more inhabitants. The classification of regions is outlined in Boxes 1.2 and 1.3. The period covered is 2000 to 2013 (or closest available year). The 24 countries included are: Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Ireland, Italy, Korea, Netherlands, New Zealand, Poland, Portugal, Slovak Republic, Slovenia, Spain, Sweden, United Kingdom and United States. For Belgium, 10 provinces and the capital city region are used. Exclusions of OECD countries are due to missing data or due to data only being available for a single region.

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411698

However, catching up and divergence dynamics arise in all types of regions. While the frontier is mostly urban, many regions with large rural populations tend to do well and have been catching up to the national frontier (see also Chapter 3). The potential for catching up is present in all types of regions, but the levers to unlock and sustain growth are quite distinct as economic models and local fundamentals differ significantly between regions. Among the regions that are catching up to their country’s frontier, 39% are mostly rural (less than 50% of residents live in a functional urban area with more than 50 000 inhabitants and no part of the region belongs to a large metropolitan area with more than 1.5 million inhabitants) and 35% are “intermediate” with between 50-70% of residents living in urban areas (Figure 1.10). Only 26% of catching-up regions are “mostly urban”, and conversely the mostly urban regions make up 43% of diverging regions. In other words, those falling behind the national frontiers include many urban regions.

Productivity growth in urban areas benefits from agglomeration economies. Workers in larger cities tend to be more productive. Partly this is due to the greater share of highly-skilled and educated workers in larger cities, but in part this is due to “agglomeration economies” that arise from living and working in large cities (see e.g. Ahrend et al., 2014). Three forces create agglomeration economies (Duranton and Puga, 2004). First, by locating in close proximity, firms can share suppliers, thereby allowing them to specialise and through that specialisation become more productive. Second, large cities are home to a variety of workers and firms, which creates more opportunities for workers to find the ideal job and for firms to find the “best-matching” – most productive – employee for a job. Third, informal interaction and learning from others is facilitated by proximity. This creates knowledge spillovers and therefore better diffusion of ideas and technologies. Especially in economies that move further into knowledge-intensive production, the availability of skilled workers and the knowledge that can be shared locally is becoming increasingly important. This poses a challenge for rural – by definition “low-density” – economies (Chapter 3).

Rural areas need to leverage other competitive advantages. Some agglomeration economies can be achieved by focusing on sectoral specialisation. But without an initial entrepreneurial impetus, the creation of a strong niche sector is difficult. To assess what strategies could be successful in creating convergence, it is useful to take a stylised view of the economy. In this view, output is produced with labour, natural resources and physical capital, which are combined using different production technologies. This stylised view offers insight into three avenues for convergence for both urban and rural areas: maximising the return to capital investment, technological diffusion and sectoral change.

First, the differential returns to capital investment can create convergence, but can fall short without adequate human capital development. In regions that attracted less investment in the past, the returns from new investments (capital deepening) are often greater. This means that poorer regions (in terms of GDP per capita output) should receive more investment, which raises their output potential and results in convergence. Importantly, investment often requires complementary human capital, i.e. the capacity and knowledge of workers and managers to ensure that the potential is fulfilled. Often this is where this first convergence force loses its momentum. For example, the reunification of Germany created tremendous investment opportunities as the capital stock of East German firms was often outdated and no longer competitive. But, convergence requires more than physical capital investment, it also requires the training of workers, the development of entrepreneurial spirit and managerial skills, for example.

Second, agglomeration economies support the diffusion of innovation through knowledge spillovers, but in less-densely populated areas additional efforts might be required. Technological adoption raises productivity as less productive firms learn from frontier firms and imitate their processes. But imitation might not be straightforward. For example, new drilling technologies that create opportunities to tap into oil or gas reserves that were inaccessible before can create substantial productivity growth, but might not translate easily into innovative ideas in manufacturing or services. Another barrier can be found in local framework conditions. For example, a firm that considers developing an international product portfolio might need specialised marketing and sales employees. If employment protection legislation is stringent, the willingness to hire new staff might be reduced, as firing costs in the case of failure of the experiment are very high. Supporting the diffusion of best practices that can lead to innovation and improving the capacity to adopt them, are essential for catching up. National and regional level institutional frameworks and policies can cultivate this diffusion potential.

Third, a change in the sectoral composition of a region’s economy towards higher productivity sectors can also raise overall productivity levels. One of the secular trends that accompany economic growth is the gradual shift of employment into more productive sectors. Labour-intensive agricultural production was replaced by manufacturing, which is increasingly surpassed by knowledge-intensive services. This change in sectoral composition can be a strong driver for growth and convergence, but also poses challenges in lagging regions. The skills required in manufacturing do not lend themselves readily for work in knowledge-intensive services and workers moving from manufacturing into other sectors tend to end up in low-end consumer services.

Building on local competitive advantages is essential in all types of economies, as regions fall at different points in between the stylised extremes of metropolitan areas and the most remote rural regions (Table 1.1). At one end of the spectrum are the large urban centres, like London, New York or Tokyo, that are home to some of the most productive and innovative firms. They are mainly focused on services, often business services, but also health care, higher education and information and communications technologies (OECD, 2014b). Manufacturing firms located in large cities are typically focused on innovation- and skill-intensive production and often only parts of the company (e.g. the headquarters) remain in the city. The size of the city allows for a high degree of diversification and redundancy in the labour and local goods and service markets. At the other extreme are rural economies that are concentrated in agricultural production or natural resource exploitation. Manufacturing in these areas tends to be in “mature” parts of the product-cycle and the relatively small number of available workers requires specialisation in few activities. While it is certainly possible to find examples for the stylised extremes, the reality of rural areas is more diverse and most regions mix rural and urban elements (Chapter 3). But, a lack of diversification in less densely populated areas requires a careful assessment and support of local strengths and weaknesses that can be leveraged to support growth.

Proximity to large cities can support growth and catching up, but divergence in productivity is not necessarily driven by distance from those cities. Smaller cities and rural towns can “borrow” agglomeration effects by being more closely connected to other cities (Ahrend and Schumann, 2014a, OECD, 2015d). Importantly, functionally defined metro areas typically extend beyond their administrative boundaries and include significant parts of the surrounding – mostly rural – areas, which are connected to the local urban centre via daily commuting flows. These rural areas both benefit and support the growth of their core cities. But governance problems, such as a lack of local co-ordination, low levels of institutional capacity, the absence of a well-designed and implemented regional strategy or a piecemeal policy approach, can limit the benefits and hinder the catching-up process (OECD, 2012).

Table 1.1. Stylised models of urban and rural economies

Urban: The high-density economy

Rural: The low-density economy

  • Led by the service sector, especially producer services

  • Manufacturing is high end, innovation intensive

  • High diversification of economic activity and redundancy in markets

  • Network economy – internet, computers, telecommunication (ICT)

  • Highly skilled core workforce, with growing gap between skilled and unskilled

  • Economic growth driven by innovation and productivity

  • Job creation driven by entrepreneurs and SMEs

  • Growth driven by large conurbations

  • Competition is intense in most product/services markets, thanks to globalisation

  • Growth driven by internal factors (endogenous)

  • Most employment in services, but mainly low-end consumer services, with larger employment shares in the primary sector

  • Manufacturing tends to be “mature” in product-cycle terms

  • Limited diversification of economic activity, long supply chains

  • Weaker transport and communications connectivity, often lagging in internet connections and computer use

  • Weak skills, youth outmigration and an ageing workforce

  • Low productivity, except in the primary sectors, and limited entrepreneurial activity

  • Low levels of patenting and formal R&D

  • Local markets tend to be thin, with weak competition

  • Firm population dominated by SMEs but often low-growth firms

  • Growth driven by external factors (exogenous)

Source: Adapted from OECD (2014b), Innovation and Modernising the Rural Economy, http://dx.doi.org/10.1787/9789264205390-en.

A prominent tradable sector is a common characteristic of both urban and rural catching-up regions

Several characteristics could be associated with a stronger regional catching-up process. A larger share of the tradable sectors could favour productivity convergence. Tradable sectors are those that must compete in global markets and are therefore better able to catch up to the productivity frontier. The tradable sector allows greater opportunities to catch up through “unconditional convergence”, meaning convergence to the global frontier is less dependent on a country’s particularities or institutional weaknesses (Box 1.6). Population density could also determine the capacity of a given region to benefit from the diffusion of technology, in particular in the service sectors. Another element is the level of education of the regional workforce. R&D expenses should be a factor promoting the adoption of innovations. Finally, the quality of regional and local governments should contribute to the adoption of good policies and investment choices.

Box 1.6. Convergence and the tradable sector

Economic theory posits that countries that have access to the same production technologies have the potential to grow towards a common level of wealth. This “absolute” or “unconditional” convergence means that less-developed economies with initially lower levels of per capita income should experience faster growth than economies that have already reached higher income levels. But growth experienced by many countries is often less in line with absolute convergence, and rather supports “conditional” convergence towards different levels of GDP per capita.

While conditional convergence is predominant among economies as a whole, the tradable (industry) sector deviates from the pattern and shows absolute convergence across economies. Based on a sample of more than 100 countries, Rodrik (2013) finds absolute convergence in labour productivity in the manufacturing sector across the world. But the strong push of the tradable sector that supports the catching up of less-developed economies does not translate into absolute convergence for the economy as a whole, as the contribution of manufacturing to the economy tends to be small in less-developed economies. Recent evidence (Rodrik, 2016) suggests that the current shift away from manufacturing is not only a challenge for most OECD countries. Some developing countries seem to experience deindustrialisation at relatively low levels of wealth, which might prematurely reduce growth opportunities in the tradable sector. Notably, the shift away from manufacturing is strongest for Latin American countries and largely absent in Asian countries.

Several characteristics of the tradable (manufacturing) sector give rise to its special role for economies. First, it tends to be an innovative and dynamic sector, which adopts to and pushes the technological frontier. Second, manufacturing has traditionally employed not only the highly skilled, but also a large number of medium- and low-skilled workers at relatively high wages, which sets it apart from other high-productivity sectors such as mining or finance (Rodrik, 2016). Third, the growth and success of the tradable sector is not limited by the size of the local market, which decouples its growth, to a certain degree, from the rest of the economy. Fourth, the tradable sector creates significant spillovers to other, localised, sectors. Moretti (2010) finds substantial job creation multipliers associated with the tradable (manufacturing) sector in the United States. For each job created in manufacturing, the number of local jobs in non-tradable goods and services increases by 1.6. In Sweden, Moretti and Thulin (2013) find a smaller multiplier, with estimates ranging from 0.4 to 0.8 jobs.

Source: Moretti, E. (2010), “Local Multipliers”, American Economic Review, Vol. 100(2), pp. 373-77; Moretti, E. and P. Thulin (2013), “Local multipliers and human capital in the United States and Sweden”, Industrial and Corporate Change, Vol. 22(1), pp. 339-362; Rodrik, D. (2016), “Premature deindustrialization”, Journal of Economic Growth, Vol. 21(1), pp. 1-33; Rodrik, D. (2013), “Unconditional Convergence in Manufacturing”, Quarterly Journal of Economics, Vol. 128(1), pp. 165-204.

The tradable sectors seem to play a key role in all region types. The tradable share in gross value added (GVA) is (statistically significantly) higher in catching-up regions (Figure 1.11).9 This is not only the case in large regions (TL2), but also for smaller TL3 regions (Figure 1.13). Employment levels are similar and declined by the same margin in both types of regions, but catching-up regions experienced an increase in the contribution of the tradable sector to GVA. That contribution remained constant for diverging regions. This pattern highlights the increase in productivity in catching-up regions and its positive effect on the tradable sector. A closer look at the contribution from different tradable sectors shows two very different drivers at the TL2 level and a third when “zooming in” to the smaller TL3 level.

Figure 1.11. The tradable sector plays a critical role in regional productivity trends
picture

Note: Catching-up/diverging regions grew by at least 5 percentage points more/less than their national frontier over the 2000-13 period. The frontier is defined as the aggregation of regions with the highest GDP per worker and representing 10% of national employment. Due to lack of regional data over the period, only 24 countries are included in the averages. Tradable sectors are defined by a selection of the 10 industries defined in the SNA 2008. They include: agriculture (A), industry (BCDE), information and communication (J), financial and insurance activities (K), and other services (RSTU). Non-tradable sectors are composed of construction, distributive trade, repairs, transport, accommodation, food services activities (GHI), real estate activities (L), business services (MN), and public administration (OPQ).

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411700

Tradable services and resource extraction are the elements of the tradable sector that account for most of the difference in the catching-up and diverging regions. The contributions to GVA from tradable services and from resource extraction (i.e. mining and drilling) in catching-up regions exceed the contributions in diverging regions, tradable services by about 5 percentage points and resource extraction by even more in 2013 (Figure 1.12). In comparison, the gap in manufacturing between catching-up and diverging regions is comparatively small (1 percentage point). There is no difference with respect to the contribution of the agricultural sector, which accounts for only a small, and declining, percentage of total GVA in all types of regions. The importance of the tradable sector results in greater exposure to changing macroeconomic conditions. This can be an advantage when there is an increase in demand, but also a risk when prices are volatile. The vulnerability to global shocks might be particularly relevant for rural economies, being more reliant on local assets and global commodity prices, as they may lack a diversified economy to absorb those shocks by creating opportunities in other sectors.

Figure 1.12. Tradable services and resource extraction contribute to catching up
GVA and employment shares by type in TL2 regions, 2013
picture

Note: Catching-up/diverging regions grew by at least 5 percentage points more/less than their national frontier over the 2000-13 period. The frontier is defined as the aggregation of regions with the highest GDP per worker and representing 10% of national employment. Due to lack of regional data over the period, only 2013 values are used. Resource extraction and utilities are the aggregate of mining and quarrying (B); electricity, gas, steam and air conditioning supply (D); and water supply, sewerage, waste management and remediation activities (E). Tradable services are the aggregation of information and communication (J); financial and insurance activities (K); and other services (R to U).

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411716

The important role of the manufacturing sector becomes apparent at the smaller regional scale. When smaller regions (TL3 versus TL2) are considered, the contribution of manufacturing to GVA in catching-up regions is more than 5 percentage points higher than in diverging regions (Figure 1.13). Manufacturing activities are often locally concentrated and the combination of manufacturing-oriented TL3 regions with those specialised in other sectors results in a seemingly less important role for the manufacturing sector at TL2 level.

Figure 1.13. Manufacturing also observed to promote catching up, but at a smaller regional scale
GVA and employment shares by type in TL3 regions
picture

Note: Catching-up/diverging regions grew by at least 5 percentage points more/less than their national frontier over the 2000-13 period. The frontier is defined as the aggregation of regions with the highest GDP per worker and representing 10% of national employment. Due to lack of regional data over the period, only 24 countries are included in the averages. Tradable sectors are defined by a selection of the 10 industries defined in the SNA 2008. They include: agriculture (A), industry (BCDE), information and communication (J), financial and insurance activities (K), and other services (RSTU). Non-tradable sectors are composed of construction, distributive trade, repairs, transport, accommodation, food services activities (GHI), real estate activities (L), business services (MN), and public administration (OPQ).

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411729

The types of jobs that drive productivity growth differ depending on the type of tradable sector. The main divide is between resource extraction, utilities and manufacturing on the one side and tradable services (and agriculture) on the other. In manufacturing and resource extraction, workers are relatively more productive than the average in the region’s economy, as the percentage of employees working in the sectors is lower than the percentage contribution to GVA (Figure 1.12). The opposite is the case for tradable services in diverging regions and in agriculture. For tradable services in catching-up regions, the employment share is more than 8 percentage points lower than in diverging regions. In manufacturing and resource extraction, workers with low levels of formal education can achieve high levels of productivity. This is in contrast to the service sector and agriculture, where aggregate labour productivity is comparatively low. But tradable services combine a wide range of jobs and some tradable services have very high levels of productivity, and unlike in traditional manufacturing, these jobs are mainly focused on knowledge-intensive services that require highly-skilled workers. Therefore, transitioning workers from manufacturing or resource extraction sectors to knowledge-intensive services is not straightforward and requires substantial adjustments.

The differences concerning other characteristics are much less marked. The population density in catching-up regions is only slightly higher than in diverging regions (Figure 1.14). Levels of tertiary education are slightly higher in diverging regions, but both types of regions experienced reductions in the share of workers with only primary education, and the shares are virtually identical in 2013. This is consistent with prior evidence that often it is the large share of workers without secondary education that constitutes a greater bottleneck to growth than a relatively low share of tertiary educated workers (OECD, 2012).10 Total R&D expenditure was very similar, with business R&D slightly higher in diverging regions, while public R&D is higher in those that were catching up, and values for R&D expenditure in the frontier regions falling in between the two (Figure 1.14). These averages combine values for regions that are catching up using different types of growth models, some that are more intensive in skilled labour and other sectors that are more or less R&D intensive. Only for patents is a slightly more pronounced gap between diverging and catching-up regions evident, but both groups reduced their gap with the frontier. To sum up, apart from the share of tradable sectors and higher public R&D, the catching-up regions have very close fundamentals to the ones that have diverged.

Figure 1.14. Other growth-related factors do not differ between catching-up and diverging regions
picture

Note: The graph for population density depicts five characteristics of the distribution of population density across regions. The lower and the upper edge of the box show the first and last quartile, i.e. 25% of regions have density below (above) the value at the first (third) quartile. The horizontal line within the box shows the median (50% of regions have a population density below/above the value). The “whiskers” at the end of the vertical lines indicate the minimum and maximum values in the group.

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411736

The quality of government is good for productivity, but does not differentiate the catching-up regions from those that are diverging. Using European data, the governmental quality measured by the Gothenburg European Quality of Government Index (EQI) is higher in more productive regions (statistically significant correlation with a correlation coefficient between 0.3 and 0.4, Figure 1.15). However, there is no differential pattern in this positive relationship between regions that are catching up and those that are diverging. It is possible that the EQI is too narrowly defined to capture all the relevant institutional characteristics in a region. Recent empirical work on regional development suggests that informal governance factors, such as the participation in organisations that are inclusive, has a more significant impact than formal governance measures on sectoral changes in the regional economy (Cortinovis et al., 2016). However, an improvement in government quality, measured by the change in EQI during the period 2010-13, shows a positive correlation between the improvement in EQI and labour productivity growth for regions that are catching up, but not for those that are diverging (Figure 1.16). But, this result is not very robust in statistical terms. These results underscore the need for more empirical work on both formal and informal governance and the relationship with productivity.

Figure 1.15. Regions with high levels of productivity are also regions that are better governed
European Quality of Government Index (EQI) and labour productivity levels, 2013
picture

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016) and the University of Gothenburg (2013), European Quality of Government Index (EQI), http://qog.pol.gu.se/data/datadownloads/qog-eqi-data; Charron, N., L. Dijkstra and V. Lapuente (2014), “Mapping the Regional Divide in Europe: A Measure for Assessing Quality of Government in 206 European Regions”, Social Indicators Research, http://dx.doi.org/10.1007/s11205-014-0702-y.

 http://dx.doi.org/10.1787/888933411740

Figure 1.16. Catching-up regions for productivity also experienced modest improvements in governance quality
Change in the European Quality of Government Index (EQI) and productivity growth, 2010-13
picture

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016) and the University of Gothenburg (2013), European Quality of Government Index (EQI), http://qog.pol.gu.se/data/datadownloads/qog-eqi-data; Charron, N., L. Dijkstra and V. Lapuente (2014), “Mapping the Regional Divide in Europe: A Measure for Assessing Quality of Government in 206 European Regions”, Social Indicators Research, http://dx.doi.org/10.1007/s11205-014-0702-y.

 http://dx.doi.org/10.1787/888933411758

Countries display different patterns of catching-up dynamics

There is no unique pattern that links regional catching up or divergence with labour productivity growth in a country or its frontier. But a growing divide between productivity in the frontier and other regions raises the concern of a “two-tier” economy, with strong performance at the top and the rest of the country falling behind. In Figure 1.17, regional productivity growth patterns are compared with national average productivity growth.

  • Low productivity growth countries: towards the left of Figure 1.17 are countries with an overall low productivity growth performance. Nevertheless, in some countries the regional catching up was a key factor in driving national productivity growth. For example, Germany and Austria combined relatively low country-level productivity growth with a frontier that grew slowly, but the catching up of other regions sustained overall productivity growth.

  • High productivity growth countries: on the right side of the graph, there are countries like Poland and the Czech Republic that combined high aggregate productivity growth with strong catching-up dynamics in some regions. In contrast, in countries like the Slovak Republic, Hungary and Korea, it is the frontier region that dominates the productivity growth performance of the country.

  • Moderate productivity growth countries: in the middle of the graph there are countries such as France, the United Kingdom and Australia, that grew at moderate levels with relatively large gaps between the frontier and the other regions in the country. The value of the difference in growth rates may seem small, but over time they are quite substantial. For example, productivity growth in Île-de-France (Paris) was, on average, only 0.4 percentage points higher than in Nord-Pas-de-Calais, a large, traditionally industry-focused region in the north of France and the French entrance to the Eurotunnel which connects France with United Kingdom. But, over a period of 20 years, this difference adds up to almost 10 percentage points.

In addition, in many countries it is not only the gap between the frontier and other regions that tends to be wide, the range of growth rates across countries is often substantial. Greece’s top-performing region grew above the country average, but most other regions in the country have diverged. Large gaps are more the norm than the exception across the OECD, with some countries exceeding a productivity growth gap of 4 percentage points annually between their fastest and slowest growing regions (e.g. Canada and the United States).

Figure 1.17. Regions in both fast and slow growing countries can catch up (or fall behind) their frontier
Labour productivity growth in TL2 regions, 2000-13
picture

Note: Annual average growth in real per worker GDP between 2000 and 2013 (or closest year available). Catching-up/diverging regions are defined as those with 5 percentage points more/less cumulative labour productivity growth over the 14-year period compared to the frontier regions (defined as those with the highest GDP per employee until the equivalent of 10% of national employment is reached), with regions that are “keeping pace” falling within the +/- 5 percentage points band. The solid line indicates the country-level annual average productivity growth over the same period. Countries excluded due to lack of data or an insufficient number of regions include: Chile, Estonia, Iceland, Israel, Japan, Luxembourg, Mexico, Norway and Turkey.

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411768

Labour productivity growth is essential for GDP growth, but needs to be accompanied by employment growth to create substantial benefits and contribute to social inclusion. In the Netherlands, the three frontier regions, Groningen, North Holland (includes Amsterdam) and South Holland (includes Rotterdam and The Hague) exhibit very different trends (see Box 1.7 for details on the trend decomposition). Groningen and North Holland contributed more than 50% to the Dutch growth in GDP, while South Holland’s decline exerted a significant drag on growth (Figure 1.18). The right panel of Figure 1.18 shows the critical role that labour productivity plays in contributing to the trends. While productivity in Groningen and North Holland improved relative to the rest of the Netherlands, South Holland has started to fall behind. In Germany, more than 80% of GDP growth is generated by the regions that are catching up to the frontier’s productivity levels (Figure 1.19). The single largest contributor is Bavaria, which generated more than 25% of Germany’s GDP growth in the 2000-13 period, even though less than 17% of employees work in the Land.

The key difference between the German and Dutch frontier performance lies in their employment dynamics. The frontier can be a driver for productivity growth, as is the case for North Holland and Groningen, two parts of the frontier in the Netherlands. Conversely, a lack of productivity growth in the frontier can hold back the whole country, as is the case in South Holland, the third part of the Dutch frontier. The frontier is even less dynamic in Germany: Hesse and Hamburg recorded the lowest productivity growth rates over the 2000-13 period. But the economic impacts of the slow growing frontiers in the Netherlands and in Germany are very distinct. South Holland contributed negatively to GDP growth and total employment declined between 2001 and 2013. The German frontier regions contributed positively to GDP growth, which was accompanied by an increase in employment. Hamburg and Hesse added more than 300 000 jobs between 2000 and 2013 and Hamburg’s employment was the fastest growing in Germany over that period. (See the country pages that accompany this publication for details in the different OECD countries.)

Box 1.7. The contribution of regions to GDP and labour productivity growth

The contribution of a region to its country’s economic (GDP) growth can be decomposed into the weighted sum of the GDP growth in each of its regions. The weights are equivalent to each region’s initial percentage contribution to the country’s total GDP. The relationship can be expressed as a formula, with Yr,t the GDP in region r in period t and picture the national GDP, which is the sum of the GDP produced in all R regions within the country.

picture

From this, the contribution of any region r to its country’s GDP growth can be expressed as:

picture

A similarly straightforward decomposition of a region’s contribution to labour productivity growth is not available. To calculate an alternative decomposition, it is useful to consider the hypothetical scenario “how much would labour productivity in the country have grown if the region had not contributed”. The difference between labour productivity growth in the hypothetical scenario and the actual labour productivity growth can then be used as an indicator for a region’s contribution. With E indicating total employment, and picture the sum of GDP from all regions except region r (analogously for employment), the contribution of each region to labour productivity growth can be expressed as:

picture

Figure 1.18. Frontier regions in the Netherlands experience both high and low rates of productivity growth
Contribution of Dutch regions to labour productivity and GDP growth, 2001-13
picture

Note: Regions are ordered by their labour productivity growth rates. The left panel depicts the percentage contribution of each region to national GDP growth. The right panel depicts the percentage point difference between actual labour productivity growth (GDP per employee growth) and a hypothetical scenario that considers the labour productivity growth that would have occurred if the region were not part of the country. See Box 1.5 for detailed definitions.

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411777

Figure 1.19. Hamburg and Hesse attract employment, but struggle to utilise it productively
Contribution of German regions to labour productivity and GDP growth, 2000-13
picture

Note: Regions are ordered by their labour productivity growth rates. The left panel depicts the percentage contribution of each region to national GDP growth. The right panel depicts the percentage point difference between actual labour productivity growth (GDP per employee growth) and a hypothetical scenario that considers the labour productivity growth that would have occurred if the region were not part of the country. See Box 1.5 for detailed definitions.

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411789

These country patterns result in very different distributions of the population according to catching-up and divergence patterns. Figure 1.20 provides the shares of the population living in frontier, catching-up, diverging or keeping-pace regions (those that neither gained nor lost significant ground relative to the frontier). The OECD countries split nearly completely into two groups: those countries where many people were living in regions catching up and those where people mainly saw their region diverging from the frontier. Few countries show intermediate positions, the notable exceptions being the Czech Republic, New Zealand and the United States. The percentage of the population living in regions falling behind ranges from 0% to over 80% depending on the country. The net result is that over one-quarter of the population across the OECD (26.4%) is living in regions that are falling behind (diverging) relative to their national frontiers, which translates into hundreds of millions of people.11 The rest of the OECD population is split between those living in frontier regions (15.6%), catching-up regions (19.7%) or regions that are keeping pace with growth in the frontier (38.3%).

The share of the population residing in regions falling further behind needs to be considered in light of interregional mobility trends. It is expected that some regions will gain and lose population based on the opportunities they provide to residents. Interregional mobility rates differ by country. For example, almost 5% of the population in Korea and Hungary, and less than 0.5% in the Slovak Republic, changed regions from 2011-13. The vast majority of migration occurs within a country. In total, around 2% of the OECD area’s population changed regions, which is more than four times the value of international migration to OECD countries (OECD, 2016b). Given the growing economic divide within countries, internal migration rates might seem low, but often there are a variety of impediments that curtail mobility. The less skilled are, at least on average, less likely to move within a country (e.g. Machin, Salvanes and Pelkonen, 2012). Local social ties and housing markets tend to also limit mobility (e.g. Antolin and Bover, 1997). Differential migration trends create an additional challenge for different types of regions. In most countries it is urban regions that benefit from net inflows, however in some countries rural regions are gaining (OECD, 2016b; BBSR, 2015).12 Those rural regions that lose too many of the higher-skilled workers may suffer from a cascading effect that limits growth potential for the remaining residents (Chapter 3).

Figure 1.20. One in four OECD residents lives in a region that is falling behind the frontier
Regional shares of population by type of convergence performance, 2014
picture

Note: Countries excluded due to lack of data or an insufficient number of regions include: Chile, Estonia, Iceland, Israel, Japan, Luxembourg, Mexico, Norway, Switzerland and Turkey. Numbers in parenthesis indicate the number of TL2 regions in the country.

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411795

Interregional differences in innovation-related factors have nevertheless narrowed in many countries and for several indicators

As noted in the analysis above, innovation-related factors (R&D and patents) do not seem to show marked differences between catching-up and diverging regions on an OECD-wide basis. However, an analysis at the country level indicates that innovation-related factors may play a role. For example, in the case of the United States, there is a clear pattern of stronger patenting activity in the frontier regions, given the spatial concentration of patenting activities in some metropolitan areas. There are also differences in the intensity of patenting in the catching-up and the diverging regions. Indeed, the intensity has increased over time in the catching-up regions, while it has remained stable in the diverging group (Figure 1.21). Concerning R&D intensity, it is only with respect to public R&D that catching-up regions display better performance than those that are diverging, as noted above on an OECD-wide basis. While private R&D was, on average, lower in catching-up regions than in diverging regions, the change between 2000 and 2013 within the two groups shows an increase in catching-up regions, but a decline both in the frontier and in diverging regions.

Figure 1.21. Innovation-related activities and productivity trends: United States
picture

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411804

The mechanisms by which innovation-related activities and policy influence productivity and regional growth are not fully understood (OECD, 2011b). As in the United States, regions at the frontier often outperform others on many innovation-related indicators, such as R&D investment and patents. In an analysis of regional drivers of per capita GDP growth, this was found to be particularly important for OECD regions with relatively high income levels in their country context. However, the relationships of R&D and patents with catching up were not as strong for lower income regions (OECD, 2012). Additional S&T-related investments may require more skilled human capital (a “social filter”) to make the most of these investments (Rodríguez-Pose and Crescenzi, 2008). However, even a combination of skilled workers (share of workforce with tertiary education) and R&D investment may not be enough for growth among less-developed regions (Sterlacchini, 2008). There may also be other bottlenecks to growth in those most lagging regions that need to be addressed to trigger catching-up dynamics and to create benefits from additional skilled workers and R&D investment.

Many other forms of innovation or firm practices which are important to productivity are simply not captured by measures of R&D and patents. The propensity to patent and conduct R&D does vary by economic sector. Organisational and marketing innovations, for example, are not typically captured by these statistics, but can have notable impacts on productivity. One study using a sample of 330 firms in the United States found that those engaging in data-driven decision-making could expect to have 5-6% increases in output and productivity compared to the others that did not (Brynjollfson, Hitt and Kim, 2011).The share of business investments in intangible assets, as opposed to other investments such as machinery and equipment, have been steadily rising over the last two decades. These different forms of knowledge-based capital can include software, organisational capital, and training, as well as R&D (OECD, 2015d).

There has been some convergence of the intensity of innovation-related activities across regions within countries. For example, gaps in the performance of the top 20% over the bottom 20% of regions (i.e. each accounting for 20% of the population) have narrowed between 2000 and 2013 in more than half of the countries for most innovation-related variables considered (Figure 1.22).13 While interregional gaps in tertiary education have declined in the vast majority of countries (23 out of 27 with data), the same is not true for R&D personnel (per 1 000 employees), an indicator for the most innovation-intensive workers. The gap in the performance of top and bottom regions actually increased in more countries than it declined (12 versus 7). This is in contrast to most forms of R&D intensity (R&D as a share of GDP), for which the gap narrowed in more countries than it increased. Patenting intensity (patents per million inhabitants) shows more mixed results, as only just over half of the countries showed a reduction in the gap (17 out of 31 countries), likely given the more technology-intensive requirements in patent-related innovation that are specific to certain industries and their spatial concentration. Where the gap narrowed, the reduction was generally due to greater improvements among the bottom 20% of regions relative to the top 20%.14 Public policy has likely contributed to this trend of reducing gaps across regions on these innovation-related factors. However, the results may not yet have fully materialised as there is always a time delay between investment in innovation inputs and the results in terms of productivity gains.

Figure 1.22. Interregional gaps in innovation-related performance show mixed results, often narrowing
Changes in performance between the top and bottom 20% of TL2 regions in a country, 2000-13
picture

Note: The top 20% of regions are defined as those with the highest value of the indicator until the equivalent of 20% of the national population is reached. The same calculation is made for the bottom 20%. For all graphs, Estonia and Luxembourg are excluded as both have only one TL2 region. Other countries are excluded due to lack of data or comparable years. For total, business, government and higher education R&D expenses: Chile, Denmark, Iceland, Israel, Japan (included for government R&D), Mexico, New Zealand, Switzerland (included for business and total R&D) and Turkey. R&D personnel also excludes: Australia, France, the United Kingdom and the United States. Tertiary educated labour force also excludes Australia. Patents per million also excludes New Zealand. The last year of available data for Greece, Japan, the Netherlands, Norway and Switzerland is 2011.

Source: OECD (2016b), OECD Regions at a Glance 2016, http://dx.doi.org/10.1787/reg_glance-2016-en.

 http://dx.doi.org/10.1787/888933411816

As a result of these trends, the spatial concentration of resources in the top 20% of regions has declined (Figure 1.23). This is particularly striking for business R&D expenditures, where that concentration in the top regions declined in 20 out of 24 countries with data from 2000-13.15 Using the HHI index16, the same general trends of declining levels of concentration are observed, and for this index it is true for an even greater number of countries relative to the indicator considering the top 20% of regions only.

Figure 1.23. Regional concentration of innovation-related resources within countries generally declining
Changes in the share found in the top 20% TL2 regions in a country, 2000-13
picture

Note: The top 20% regions are defined as those with the highest value of the indicator until the equivalent of 20% of the national population is reached. For all graphs, Estonia and Luxembourg are excluded as both have only one TL2 region. Other countries are excluded due to lack of data or comparable years. For total, business, government and higher education R&D expenses: Chile, Denmark, Iceland, Israel, Japan (included for government R&D), Mexico, New Zealand, Switzerland (included for business and total R&D) and Turkey. R&D personnel also excludes: Australia, France, the United Kingdom and the United States. Tertiary educated labour force also excludes Australia. Patents per million also excludes New Zealand. The last year of available data for Greece, Japan, the Netherlands, Norway and Switzerland is 2011.

Source: Maguire, K. and J. Weber (forthcoming), “Should we care about gaps in regional innovation capacities?”, OECD Regional Development Working Papers, OECD Publishing, Paris.

 http://dx.doi.org/10.1787/888933411821

However, venture capital (VC) follows a different path, implying that some aspects important for innovation activity are not easy to reproduce outside of the top regions. The case of the United States illustrates that the leading regions have continued to attract more VC than other regions. The share of VC in the United States in the top 20% regions (out of the PWC-defined 20 US regional groupings: Silicon Valley, New York Metro, New England and Los Angeles/Orange County) increased from 51% in 1995 to 76% in 2014 (PWC, 2015). Many of these same metro areas are found in states with a high rate of business R&D. A wide disparity in the share of total VC can also be observed across regions within other countries (OECD, 2016b).

From productivity to inclusion and well-being in regions and cities

Productivity growth matters because it helps determine wage levels and the types of jobs available that determine to a significant extent well-being. Material and non-material elements of well-being depend on one’s personal characteristics, as well as the characteristics of the places where one lives and works (OECD, 2014c). Interregional differences within countries in some areas of well-being, such as unemployment rates, are higher than across countries. Residents in some regions live, on average, several years longer than their counterparts in other regions. While there is mobility across different regions, often the least skilled are the least mobile and may remain in those places with fewer opportunities. The results are gaps for people living in different regions and cities in different dimensions of well-being, with some gaps being more persistent than others.

Well-being differences across regions have both increased and decreased depending on the dimension

The degree of interregional disparities in well-being and the progress made in closing interregional gaps depends on which dimension of well-being is considered. Across all regions in 33 OECD countries, the largest regional disparities are those associated with unemployment rates, household income levels and air quality (PM2.5 levels). In other factors, such as life expectancy, the interregional variations within the OECD are less pronounced in terms of the coefficient of variation (Figure 1.24). Such differences might seem small in statistical comparison, but may be substantial for individuals. Across OECD countries, the difference in life expectancy is eight years (between Japan and Mexico). Within countries, that interregional difference can be up to six years of greater life expectancy, e.g. in Australia between the Australian Capital Territory and the Northern Territory or in the United States between Hawaii and Mississippi. Another factor that also shows stark interregional variations across the OECD is that of safety (as measured by homicide rates) (OECD, 2016b).

The progress in reducing disparity in well-being also depends on the dimension, and some dimensions are easier to influence through policy. Over the last decade, regional disparities across the OECD have decreased in the two indicators that are most directly affected by policies: access to services (measured by broadband access) and education. Conversely, disparities in household income, air quality, unemployment and life expectancy have increased to different degrees (Figure 1.24). Safety is another well-being dimension that has experienced an increase in inequality across the OECD over the last decade (OECD, 2016b). Well-being differences are observed between urban and rural areas, the latter generally being better off in terms of housing and the environment (see Chapter 3).

Figure 1.24. The degree of interregional variation depends on the well-being dimension
Regional disparities in well-being dimensions OECD (TL2) regions
picture

Note: The higher the coefficient of variation, the higher the degree of regional disparities.

Source: Calculations based on OECD (2016f), OECD 2016, Regional Well-Being, OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411833

The catching-up performance of regions in terms of productivity supports economic dimensions of well-being and can create benefits in other dimensions of well-being. As expected, higher productivity performance in terms of catching up is associated with higher levels of household income, both in terms of levels and growth rates (Figure 1.25). More striking is the impact of catching up on employment. Average unemployment in catching-up regions declined between 2000 and 2014, while diverging regions and the frontier experienced increases in unemployment. The large increase in the frontier reflects a rise in unemployment across a range of countries in Southern, Western and Northern Europe, but also in the United States. Catching up benefited the young as well, as the average rate of 18-24 year olds who are not employed, in education or in training (NEETs) declined on average in such regions. For other dimensions of well-being, such as safety and environmental quality, all types of regions improved and in approximately the same proportions in different region types. The levels of life expectancy tend to be lower in the group of catching-up regions (on average 1 year less), however such an indicator changes much slower than other dimensions of well-being. With respect to air pollution, often a negative externality of economic growth, there was nevertheless a greater reduction in particulate matter in catching-up regions as compared to other regions.

Figure 1.25. Well-being indicators and productivity performance
picture

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411840

Frontier and catching-up regions tend to create income benefits for their residents. Nearly 60% of the most productive (frontier) regions are also regions that belong to the top 10% of regions in terms of disposable household income (Figure 1.A1.1 in Annex). The importance of productivity for household disposable income is also evident in the widening gap between diverging and catching-up regions. Diverging regions account for 45% of the regions where the income gap to the country’s wealthiest regions widened, but only for 20% of regions that narrowed the disposable income gap. This is in stark contrast to regions that are catching up in productivity. These regions make up only 15% of regions where income gaps widened, but more than 30% of regions that improved relative to the highest income regions. The patterns indicate general trends, but are not necessarily deterministic, e.g. redistribution can alleviate the impact of diverging economic growth on household incomes. But labour productivity is essential to create redistributable wealth.

Productivity growth can be accompanied by job creation. As firms become more productive, their competitiveness increases, leading to more demand and incentives to hire more workers. At the regional level, productive firms create more demand for local products and therefore more employment opportunities. These effects can be very large. For US metropolitan areas, estimates suggest that a new job in the tradable sector eventually results in the creation of up 1.6 jobs in the non-tradable sector (Moretti, 2010). For Swedish labour market regions the impact is slightly lower, 0.5 jobs per new job in the tradable sector (Moretti and Thulin, 2013). In both the United States and Sweden, the strongest impact, or “local multiplier”, is found for the creation of high-skilled jobs in the tradable sector. A new employee with tertiary education is, on average, associated with about 3 new jobs in the non-tradable sector in the city or labour market region (ibid.). The reasons are two-fold. First, firms that create tradable-sector jobs require services from other local industries, services such as catering or maintenance at the lower end of the skill spectrum, but also legal advice and marketing at the high end. In addition, the workers who fill the new jobs create demand for local services, such as nannies, cafés or personal trainers. In some cases however, productivity growth may be accompanied by job destruction.

While information and communications technologies have created a large number of jobs, many highly-valued companies create their services with relatively small numbers of in-house staff. Young and fast growing firms account for a disproportionate share of total job creation (Criscuolo, Gal and Menon, 2014). This is also the case in ICT sectors, but many innovative internet-focused firms create rapid economic growth with a small workforce. Skype S.a.r.l., the voice-over-ip service company, for example, had around 140 employees with stock options when it sold to eBay in 2005-06. The price eBay paid was USD 2.6 billion.17 In 2011, just 5 years later, Skype was sold to Microsoft for USD 8.5 billion. As of June 2010, Skype had less than 840 staff and contractors.18 The example of Skype is by no means unique, but its incorporation in the larger tech-companies also shows that diversification of product portfolios and continued expansion of successful and fast-growing companies leads to significant job creation. This is also true for companies that focus on facilitating contact between clients and service providers, such as Airbnb for holiday rentals or Uber for transport services. Founded in 2008, Airbnb had already around 600 employees in 2013 and about 1 600 by 2015.19 But these jobs are not directly related to the services that are being traded on the company’s platform and pale in comparison to the number of users and listings.20

Most productivity growth is accompanied by employment growth, but in some regions the crisis led to significant job losses and recovery is not complete. The total number of jobs increased in most OECD regions and countries. But in many regions in Southern and Eastern Europe (in Greece, Hungary, Italy and Portugal), the number of employees declined significantly between 2000 and 2013. In some cases this decline was accompanied by increased labour productivity. In the majority of cases, labour productivity growth is accompanied by employment growth (see Chapter 3) and often the increase in productivity through reduced employment is a transitory effect that fades in the aftermath of a crisis. The high level of unemployment and the lack of job recovery suggest that this transitory effect has not been completed in some parts of Europe.

Catching-up dynamics are also possible for different dimensions of well-being, and this trend is observed in several dimensions. For the most part, well-being increased in both the top and bottom 10% of regions across the majority of OECD countries, but gaps between the best- and worst-performing regions widened in some cases. Considering the gap, defined as the ratio of the top and bottom 10% regions (i.e. the regions that account for 10% of the country’s population at the top and bottom), for several indicators and in many countries there is a narrowing of the gaps in well-being (Figure 1.26). In at least one-quarter of the countries where gaps widened, the regions that were initially part of the top and bottom 10% of regions changed between 2000 and 2013. The regional disparities in several labour-force related variables (unemployment, the education level of the workforce, and the gap between the employment rate for women compared to that for men) decreased in most countries. But in nine countries a narrowing gap in unemployment was due to increased unemployment, with larger unemployment increases in the top 10% than in the bottom 10% of regions. In other words, the gap narrowed, but it was not due to a catching-up dynamic. In contrast, the disparity between the regions with the highest and lowest levels of PM2.5 air pollution increased for 45% of the countries (13 out of 29 countries). Nevertheless, in all but two countries, pollution decreased in both the regions with the highest and lowest levels of PM2.5 air pollution, thus making everyone better off. For per capita disposable household income and life expectancy, the changes are similarly split with slightly more than half of the countries showing a narrowing gap.

Figure 1.26. Gaps between the top and bottom performing regions in many well-being dimensions generally narrowed
Top and bottom 10% performing regions (i.e. 10% of the country’s population), 2000-13
picture

Note: The relative gap is the percentage difference between the top and bottom TL2 regions, defined as the regions with the highest/lowest value in the indicator that account for no less than 10% of the country’s population in the reference year. Changes are for the period from 2000 to 2013/14, or closest year available. Only countries with at least 3 regions with data from 2005 or earlier are included. Numbers in parenthesis indicate the number of countries for which data is available. Numbers in the bar indicate the number of countries in the category.

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411856

While overall convergence is observed on individual well-being dimensions at the country level, when several dimensions are combined, interregional disparities can be exacerbated. One way to combine dimensions is through a calculation of multi-dimensional living standards. The OECD has recently produced a composite indicator that combines income, unemployment and health.21 This calculation at the regional level in a sample of 26 OECD countries finds that existing disparities in health and unemployment rates accentuate the gaps relative to household income only. When considering the changes over time, from 2003 to 2012, regional disparities between the top and bottom regions were mainly driven by trends in income and jobs (Veneri, and Murtin, 2015). Regional characteristics are also highly relevant in explaining variation in self-reported life satisfaction. In a recent study, 40% of the explained variation of OECD residents’ self-reported life satisfaction can be accounted for by regional characteristics, with individual characteristics accounting for the other 60% (Brezzi and Diaz Ramirez, 2016).

Figure 1.27. Regional disparities in multidimensional living standards are higher than for income alone
Coefficient of variation (higher values mean larger disparities), 2012
picture

Note: Income=disposable household income.

Source: Veneri,P. and F. Murtin (2016), “Where is inclusive growth happening? Mapping multi-dimensional living standards in OECD regions”, OECD Statistics Working Papers, No. 2016/01, http://dx.doi.org/10.1787/5jm3nptzwsxq-en. Calculations based on OECD (2016g), Regional Well-Being (database), http://www.oecdregionalwellbeing.org/ (accessed 12 June 2016) and national income surveys.

 http://dx.doi.org/10.1787/888933411860

Cities face particular challenges for inclusion across income levels and populations, such as immigrants

Cities generally have higher rates of productivity than other regions for a variety of reasons, but may face some trade-offs that require more integrated policy approaches. One of the drivers of these higher rates is that co-location creates “agglomeration economies”. Agglomeration economies confer a productivity “bonus” to workers that depends on the size of the city. Metropolitan areas also benefit from other advantages, such as a diversity of firms in close proximity, above and beyond the density of firms in the same sectors, which can also spur more innovation. An additional factor that contributes to the success of cities is the concentration of highly educated workers. These workers are not only more productive themselves, but create “human capital spillovers”, i.e. a higher percentage of highly educated workers increases productivity (measured by individual earnings) for all workers (e.g. Moretti, 2004). In a sample of five OECD countries (Germany, Mexico, Spain, the United Kingdom and the United States) a 10 percentage point increase in a city’s share of university graduates, is associated with productivity increases of about 3% (Ahrend et al., 2014). In addition, knowing that there are greater returns to education provides an incentive for further investment in one’s education, creating a virtuous cycle.

However, the opportunities afforded to high-skilled workers by larger cities may also exacerbate the degree of income inequality among workers in metropolitan areas. The disparities in terms of both wage and total income can indeed be very high, and vary considerably across metropolitan areas, with some being more unequal than others (Figure 1.28). Larger cities have, on average, higher levels of income inequality (Boulant, Brezzi and Veneri, 2016). In part, this is due to metropolitan areas providing opportunities at both ends of the skills spectrum; they attract some of the highest wage earners (“bankers”), as well as workers for many lower-skilled jobs, often in the non-tradable sector (“baristas”).22 Metropolitan areas also tend to attract immigrants, whose skills might not be directly transferable to their new environment or might be undervalued in the labour market for various reasons, including lack of qualification recognition.

Figure 1.28. The degree of metropolitan area income inequality can vary a lot in some countries
Metropolitan areas with minimum and maximum Gini coefficients, by country, 2014 or latest year available
picture

Note: Elaboration based on national data from tax records and household income surveys. Countries are ordered by the width of range between minimum and maximum Gini coefficients. The national Gini coefficients are estimated using the same source of data employed for the metropolitan areas. Due to sampling errors, they might slightly deviate from values provided by national surveys. Data do not allow the calculation of national Gini coefficients for Mexico. Numbers in parenthesis indicate the number of metropolitan areas (500 000+ inhabitants) in each country.

Source: Boulant, J., M. Brezzi, and P. Veneri (2016), “Income Levels and Inequality in Metropolitan Areas: A Comparative Approach in OECD Countries”, OECD Regional Development Working Papers, No. 2016/06, http://dx.doi.org/10.1787/5jlwj02zz4mr-en.

 http://dx.doi.org/10.1787/888933411877

In some metropolitan areas, average household income in a jurisdiction in the same metropolitan area can be double that of another, such as in the United States (Figure 1.29). The county level covers many localities; therefore those differences would be much more striking by municipality or neighbourhood. Such segregation by income or social background can limit access to jobs, particularly in cases where there is traffic congestion and poor public transport serving those locations where the lower income groups live. There may also be lower quality public services in the locations with a larger share of lower-income residents. Recent evidence for the United States shows that the impact of a lack of opportunities can have profound and lasting effects, not only for the current, but also for future generations. The intergenerational “upward mobility” of children, i.e. children improve their position in the income distribution, relative to their parents, is determined by the characteristics of the neighbourhoods in which they grew up (Chetty and Hendren, 2015). Among the characteristics, lower levels of segregation and income inequality, as well as better primary schools, improve children’s chances to move up in the income distribution (Chetty, et al., 2014). As a result, policies to address inclusion need to consider not only the distribution in income across individuals, but also the disparities generated by segregation according to income level or other socio-economic factors.

Figure 1.29. Average household income varies significantly across jurisdictions in a metropolitan area
County-level variation of household disposable income in US metropolitan areas in 2014: constant 2010 prices USD
picture

Note: Metropolitan areas are ordered by increasing value of the difference between the maximum and the minimum county values. The figure includes the 26 largest American metropolitan areas according to the OECD definition of functional urban areas. Data come from American Community Survey; 2014. Numbers in parenthesis after the metropolitan area’s name indicate the number of counties included in a metro area.

Source: Based on Boulant, J., M. Brezzi, and P. Veneri (2016), “Income Levels and Inequality in Metropolitan Areas: A Comparative Approach in OECD Countries”, OECD Regional Development Working Papers, No. 2016/06, http://dx.doi.org/10.1787/5jlwj02zz4mr-en.

 http://dx.doi.org/10.1787/888933411883

Another source of inequalities in cities is related to the integration of migrants. National government budgets may benefit from having additional workers, such as through additional contributions to pension systems, but subnational budgets bear a lot of the burden. Many of the public services that play a critical role in making integration happen are financed and delivered by local governments.23 The right framework conditions at the local level to facilitate integration include housing, the local labour market, the supply of education, health care, and the presence of communities and strong civil societies, among others. Co-ordination with national-level policies is needed, but also across stakeholders at the local level. Fragmentation among municipal governments in the same metropolitan area can accentuate this challenge, where both “immigration-friendly” and “immigration-hostile” local governments can be co-located (Walker and Leitner, 2011).24

The challenge of integrating migrants has moved beyond the “gateway” or global cities to include secondary cities. While many global cities have a high share of foreign-born residents, such as Paris, London, or Brussels, dozens of smaller urban areas with between 50 000 to 500 000 inhabitants are increasingly home to a population that is 10% to 20% foreign born (Figure 1.30). Some cities like Detroit or Cleveland are pursuing aggressive pro-immigration agendas in order to fight urban blight (Tobocman, 2014). In other cases, immigrants are seen as an opportunity to fulfil needs in rural areas. In Italy, for example, migrants are a fundamental part of the emerging silver economy in suburbs and in rural areas (Çağlar, 2014). Policy approaches that allocate immigrants to locations with less expensive housing may face longer-term problems if there are insufficient jobs available in those same locations (OECD, 2016h). Among the main challenges of migration policies is the need to combine thorough integration efforts with a quick delivery, as the speed of integration allows greater benefits (Aiyar et al., 2015). Understanding the role of a locality in the global migration flows, as a destination or transit point, also allows for better policy making to support integration (IOM, 2015).

Figure 1.30. Many small- and medium-sized cities have a significant share of foreign-born residents
Share of population that is foreign born by size of functional urban area, Europe 2011
picture

Source: EU Urban Audit data for 2011 or closest year available for 15 out of 28 EU countries. EU Urban Audit (n.d.), “Functional urban areas (urb_luz)”, (database), http://ec.europa.eu/eurostat/web/cities/data/database (accessed June 2016).

 http://dx.doi.org/10.1787/888933411891

Public action to promote catching up and inclusion: structural reforms, public investment (including through place-based policies) and governance reforms

The integration of policies across sectors, levels of government and places which underpin regional development can promote catching-up dynamics and boost aggregate productivity. It is neither desirable, nor is there fiscal space, for interregional productivity gaps to be entirely compensated by redistribution. The regional development paradigm outlined in the first Regional Outlook (OECD, 2011a) highlights the importance of combining policies across different sectors to unlock regions’ growth potential and ensure that growth is both inclusive and sustainable. It also stresses that policies need to take the local context, the local “eco-systems”, into account. This can support the effectiveness of policy implementation, but also ensure that policy makers recognise that policies that are not spatially targeted can nevertheless have a differential impact across different types of regions and design them accordingly. Strategies that focus solely on the weakest regions are likely to miss out on potential growth compared to strategies that take an integrated regional view. The “pull” that frontier regions can exert is one of the forces that could support greater catching up, but this force does not necessarily arise automatically. In other words, what is needed are policies that boost productivity in all regions while guarding against potential adverse effects on equity, both in terms of income and non-income outcomes that matter for well-being.

Economy-wide structural reforms help regional catching up, more so if complemented by regional development policies

The current approach to structural reforms should be complemented by a regional perspective to boost productivity and the extent to which it is inclusive. The traditional view of structural policies is that the degree of structural reform will determine, in large part, the level and growth of productivity. This policy package generally concerns product markets, financial markets and labour markets, as well as selected other policies, such as for health systems and pensions. The expected impacts of these structural reforms assume that all factors are mobile. However, evidence shows that some factors are particularly “sticky” to places, most notably workers. For the United States, 89% of applications sent through an online job portal were sent to firms in the same state as the applicant, which indicates a strong “distaste for distance” (Marinescu and Rathelot, 2016). In empirical work for other countries, the estimated “distaste” tends to be even stronger, e.g. in the United Kingdom (Manning and Petrongolo, 2015). This is why active labour market policies, including those that facilitate moving to jobs, are needed to complement a labour market reform.

Certain structural reforms can benefit lagging regions even more than leading regions. Several structural policies are a greater bottleneck for the growth of lagging regions relative to frontier regions.25 For example, product market regulations in wholesale and retail trade have greater negative impacts on the productivity growth of lagging regions. Rigid employment regulations also penalise lagging regions more than leading regions, which amplifies the challenge of lagging regions that tend to be less urban and thus have thinner labour markets with fewer high-skilled workers. Conversely, trade openness appears to help lagging regions disproportionately more than other regions. In other cases, regulatory barriers may have a greater impact on leading regions, for example product market regulations reduce labour productivity growth in financial intermediation and business services more strongly in regions that are closer to the productivity frontier (D’Costa, Garcilazo and Oliveira Martins, 2013).

To be more effective, structural reforms may also require complementary policies that take a place-based dimension into account. For example, labour market reforms will be of lesser benefit if there are no complementary measures to support better matching of workers to jobs or to facilitate physical access to jobs. Many of the labour market matching considerations, particularly for low-skilled workers, may involve efforts to tailor worker training to the needs of firms located in the area. Transport infrastructure is another tool, in both rural and urban areas, which can increase the effective size of a local labour market and therefore boost the productivity of firms and individual workers.

Well-designed and well-implemented public investments also support regional catching up

Public investment can make important contributions to growth, albeit research highlights some caveats. Considering 68 studies for the 1983-2008 period, the findings of one meta-analysis suggest an under-supply of public capital in OECD economies, with potential (gross) returns of USD 16-40 in GDP per USD 100 investment (Bom and Ligthart, 2014). Other findings show that the returns to investment depend on the initial level of public capital, as when already high, these effects may not be as strong (Arslanalp, 2010). The timeframe matters for observing an impact on growth, as well as whether investments are not only made within a region but also in neighbouring regions (Bom and Ligthart, 2014; Creel and Poilon, 2008). Studies on investments in transport infrastructure have shown mixed results, albeit maintenance of existing infrastructure may prove to be particularly important (Congressional Budget Office, 1991; Cullison, 1993). Different forms of network infrastructure may have growth impacts above and beyond their contributions to capital stock (Sutherland et al., 2009) and supporting investment in human capital might be important to make the most of physical capital investment.26

There are notable projected financing gaps to meet investment needs across the OECD and the globe. The OECD estimates annual global investment requirements by 2030 for telecommunications, road, rail, electricity (transmission and distribution) and water are likely to total around 2.5% of world GDP. If those estimates were to include electricity generation and other energy-related infrastructure investments in oil, gas and coal, the figure would increase to 3.5% of GDP (OECD, 2007). In the developing economies, where the population is expected to grow by 2 billion people between now and 2050, much new infrastructure is required. In advanced countries, the bulk of those infrastructure financing needs are for the maintenance of existing infrastructure. For example, in EU countries, maintenance and renewal of existing infrastructure accounts for around 70% of public investment.27 There is a challenge of financing gaps from both the private and public sectors.

There has been a decline in public investment relative to pre-crisis levels with possible underinvestment. Since the crisis, year-to-year changes in public investment have been negative. Private investment year-to-year changes started to rebound in 2014 (Figure 1.31). Both public and private investment appear to follow mostly opposing trends (as one goes up, generally the other goes down). However, since private investment is more than five times greater than public, those fluctuations do not necessarily compensate for one another. For the OECD area as a whole, government spending on gross fixed capital formation as a share of total general government outlays declined from 9.5% in 1995 to 7.7% in 2014 (Figure 1.32). In EU countries, total investment in Q2 2014 was 15% less than in 2007 despite a GDP that has rebounded to pre-crisis levels, which translates into a decline of investment of EUR 430 billion. In 2013, investment accounted for 19.3% of GDP, around 2 percentage points below the longer-term average of typical years, translating into levels of EUR 230 to 370 billion below historical rates. The problem is attributed to lower investor confidence in European financial markets and not a lack of capital, per se (EC, 2015).

Figure 1.31. Trends of weakened public and private investment may undermine productivity goals
picture

Note: Investment is gross fixed capital formation as reported in the national account statistics. OECD total excludes the following countries due to lack of data over the period 1996-2014: Chile, Mexico and Turkey.

Source: Calculations based on OECD (2016i), National Accounts Statistics (database), www.oecd-ilibrary.org/economics/data/oecd-national-accounts-statistics_na-data-en (accessed 2 June 2016).

 http://dx.doi.org/10.1787/888933411902

Figure 1.32. Public investment as a share of government expenditure on a downward trend over the last 20 years
picture

Note: OECD total excludes the following countries due to lack of data over the period, 1995-2014: Chile, Iceland, Mexico and Turkey.

Source: Calculations based on OECD (2016i), National Accounts Statistics (database), www.oecd-ilibrary.org/economics/data/oecd-national-accounts-statistics_na-data-en (accessed 2 June 2016).

 http://dx.doi.org/10.1787/888933411912

The impact of public investment on growth depends also on how it is managed. One calculation estimates that globally USD 1 trillion per year could be saved from better governance of expected public infrastructure investment needs (McKinsey, 2013). In addition, the capacity for public investment to leverage private investment, rather than crowd out such investment directly or through the way public investment is financed, is also critical. Several studies have also highlighted the role of different aspects of institutional quality in the effectiveness of public investment and its impact on growth at national and regional level (OECD, 2013b).

The subnational level is a critical financing partner but has additional capacity constraints. Subnational governments accounted for 40% of public expenditure, 50% of public procurement, 59% of public investment and 63% of public staff expenditure in 2014. In the period of fiscal stimulus after the crisis began, subnational governments partnered with national governments to increase public investment. However, from 2010, public investment by subnational governments was squeezed and served as the “adjustment variable”. The steep drop stopped in 2013, albeit it declined another 1.2% in real terms between 2013 and 2014. Overall investment has not reached pre-crisis levels in terms of volume and as a share of GDP. In a recent survey of subnational governments in Europe, the financing gaps reported by subnational governments were driven mainly by drops in allocations from central governments and a lack of recourse to private financing. Many of the infrastructure financing gaps are in areas related to productivity drivers, such as roads, educational institutions and economic development investments (Box 1.8).

Box 1.8. Results of the OECD-COR 2015 survey on public investment of subnational governments

The 2015 consultation identifies the main financing and governance challenges for infrastructure investment of subnational governments (SNGs) in the European Union.1 It also seeks to test the degree to which SNGs are facing challenges with the implementation of the principles contained in the OECD Recommendation on Effective Public Investment across Levels of Government.2 This Recommendation sets out 12 principles to help governments assess the strengths and weaknesses of their public investment capacity across levels of government and set priorities for improvement.

Virtually all SNGs report investment spending gaps (96%), whether for new infrastructure or operations and maintenance. Nearly 45% of SNGs reported a drop in investment since 2010, of which more than 70% experienced a drop by more than 10%. These cuts appear to be present more in regions and counties than in municipalities. While tax revenue has not changed significantly for most SNGs, over half (53%) experienced cuts in the grants they receive from central governments and 39% indicate stable or declining use of loans. Only a minority of cities and regions (7%), essentially metropolitan areas and regions, report increasing private sources of financing since 2010. A problematic legal and regulatory environment for public-private partnerships is another major challenge for accessing private financing, as reported by 35% of SNGs.

Several of the areas that are most affected by subnational funding cuts are areas that directly influence drivers of productivity. Indeed, around three-quarters of SNGs surveyed are not able to finance their road infrastructure needs, half are not able to build relevant educational infrastructure and 40% of SNGs surveyed report funding gaps for infrastructure destined for economic development purposes (see figure below).

Sectors subject to investment cuts in the last five years
picture

 http://dx.doi.org/10.1787/888933411928

Notes:

1. The consultation took place between 31 March and 15 July 2015 and was available in all the official EU languages. In total, there were 296 respondents, 255 of which are SNGs in 27 EU Member States (Luxembourg did not participate in the survey). They represent all categories of SNGs: regions and provinces (25%); intermediary entities (e.g. county or department) (10%); small municipalities i.e. under 50 000 inhabitants (33%); medium municipalities i.e. between 50 000 and 500 000 inhabitants (22%); large municipalities with more than 500 000 inhabitants (2%); and inter-municipal co-operation bodies (8%).

2. For more information on these Principles, see www.oecd.org/effective-public-investment-toolkit.

Source: OECD-COR (2016), Results of the OECD-CoR Consultation of Sub-national Governments: Infrastructure planning and investment across levels of government: current challenges and possible solutions, https://portal.cor.europa.eu/europe2020/pub/Documents/oecd-cor-jointreport.pdf.

Multi-level governance and territorial reforms can unlock productivity potential and support inclusion

While subnational governance is not typically part of the productivity discussion, it should be. Given that urban areas often encompass many localities, not only the core city, the metropolitan scale is critical for policy. The more complex the metropolitan area is in terms of the number of jurisdictions, the harder it may be to reap the agglomeration benefits associated with size that translate into higher levels of productivity. Out of the 281 metropolitan areas of 500 000+ inhabitants in the OECD, one-quarter contain at least 100 municipalities, and that rises to one-third when considering those that have at least 60 localities.28

There is indeed a productivity penalty associated with administrative fragmentation, as measured by the number of jurisdictions. A doubling of the degree of fragmentation results in a penalty of 6% for productivity. That penalty is halved when there is a governance body for the metropolitan area (Ahrend et al., 2014). This is one of the reasons why many countries have already, or soon will, implement metropolitan governance arrangement reforms, which are typically designed for the largest cities (OECD, 2015d). A given level of municipal fragmentation has a greater negative impact on growth in urban regions due to the higher density of interactions than in rural areas (Bartolini, 2015).

One of the critical elements of inclusion, as well as productivity, is the ability of workers to reach jobs from their homes. Commuting costs in the form of time and money influence the distance by which workers can readily reach jobs. This is true for both private transport as well as public transport options. A larger effective labour market allows workers to find better matching jobs or firms. Reduction in the time and cost of commuting can also improve quality of life. Metropolitan governance authorities typically focus on regional development, transport and spatial planning (Ahrend, Gamper and Schumann, 2014). Co-ordination across municipalities or regions can be used to improve the cost-effectiveness of public services, the quality of those services, and coherence of overall planning, among other rationales. Those metropolitan areas with a transportation authority had higher levels of satisfaction with public transport relative to those living in cities that did not (Ahrend, Gamper and Schumann, 2014).29

Administrative fragmentation is also associated with greater levels of segregation by income, that in turn influences access to opportunities. Across several OECD countries, a stable and positive association is found between administrative fragmentation and spatial segregation across local jurisdictions within metropolitan areas. This finding is consistent across estimates when considering a range of measures of both segregation and fragmentation.30 Administrative fragmentation thus can contribute to the virtuous or vicious cycles associated with segregation by income (see previous section).

Conclusion

If aggregate labour productivity derives in large part from the catching up of regions, given recent trends it is not surprising that there has been slowdown of labour productivity growth and an increase in inequalities. Creating the conditions for regions to improve productivity and generate more and higher quality jobs is, however, not straightforward. As countries continue to seek structural reforms, they should consider the complementary approaches both in terms of national policies and the respective roles that national and subnational governments can play. There is also a role for public investment and reforms of subnational governance.

These three areas of public action to promote regional catch up can offer a double dividend to countries in terms of productivity and individual well-being. There is no magic bullet, and in some cases these goals will not be jointly attainable. However, it is clear that the place-based dimension of policies is insufficiently accounted for in many policy areas. The next chapter will therefore consider the objectives and instruments for regional, urban and rural development policies to guide place-based public investment, as well as some of the governance tools and reforms to accompany them, as strategies for boosting productivity and promoting inclusion.

Bibliography

Ahrend, R., E. Farchy, I. Kaplanis and A.C. Lembcke (2014), “What makes cities more productive? Evidence on the role of urban governance from five OECD countries”, OECD Regional Development Working Papers, No. 2014/05, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jz432cf2d8p-en.

Ahrend, R., C. Gamper and A. Schumann (2014), “The OECD Metropolitan Governance Survey: A Quantitative Description of Governance Structures in large Urban Agglomerations”, OECD Regional Development Working Papers, No. 2014/04, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jz43zldh08p-en.

Ahrend, R. and A. Schumann (2014a), “Does Regional Economic Growth Depend on Proximity to Urban Centres?”, OECD Regional Development Working Papers, No. 2014/07, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jz0t7fxh7wc-en.

Ahrend, R. and A. Schumann (2014b), “Approaches to Metropolitan Area Governance: A Country Overview”, OECD Regional Development Working Papers, No. 2014/03, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jz5j1q7s128-en.

Aiyar, S., B. Barkbu, N. Batini, H. Berger, E. Detragiache, A. Dizioli, C. Ebeke, H. Lin, L. Kaltani, S. Sosa, A. Spilimbergo and P. Topalova (2016), “The Refugee Surge in Europe: Economic Challenges”, IMF staff discussion note, February 2016, SDN/16/02, Washington DC, USA.

Andrews, D., C. Criscuolo and P.N. Gal (2015), “Frontier Firms, Technology Diffusion and Public Policy: Micro Evidence from OECD Countries”, OECD Productivity Working Papers, No. 2, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jrql2q2jj7b-en.

Antolin, P. and O. Bover (1997), “Regional Migration in Spain: The Effect of Personal Characteristics and of Unemployment, Wage and House Price Differentials using Pooled Cross-sections”, Oxford Bulletin of Economics and Statistics, Vol. 59, pp. 215-235.

Ars Technica (2013), “How can they be so good?: The strange story of Skype”, http://arstechnica.com/business/2013/09/skypes-secrets/4/ (accessed 20 June 2016).

Arslanalp, S. et al. (2010), “Public Capital and Growth”, IMF Working Paper, WP/10/175, International Monetary Fund, Washington, DC, July.

ASPIRE (2015), “ASPIRE Headcount Tracker, 2015”, www.aspire.org.pl/ht2015/ (accessed 20 June 2016).

Bartolini, D. (2015), “Municipal Fragmentation and Economic Performance of OECD TL2 Regions”, OECD Regional Development Working Papers, No. 2015/02, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jrxqs60st5h-en.

Bartolini, D., H. Blöchliger and S. Stossberg (2016), “Fiscal Decentralisation and Regional Disparities”, Economics Department Working Paper (forthcoming).

BBSR (2015), “Where the Population in Europe is Growing or Shrinking”, Federal Institute for Research on Building, Urban Affairs and Spatial Development, Germany, www.bbsr.bund.de/BBSR/EN/SpatialDevelopment/SpatialDevelopmentEurope/AnalysesSpatialDevelopment/Projects/PopulationDevelopment/PopulationDevelopment.html?nn=392744 (accessed 16 June 2016).

Boix, R. (2014), Background report: OECD Territorial Review of Bergamo, mimeo.

Bom, P. and J.E. Ligthart (2014), “What have we learned from the three decades of research on the productivity of public capital?”, Journal of Economic Surveys, Vol. 28(5), pp. 889-916.

Boschma, R. (2005), “Proximity and Innovation: A Critical Assessment”, Regional Studies, Vol. 39(1), pp. 61-74, http://dx.doi.org/10.1080/0034340052000320887.

Boulant, J., M. Brezzi, and P. Veneri (2016), “Income Levels and Inequality in Metropolitan Areas: A Comparative Approach in OECD Countries”, OECD Regional Development Working Papers, No. 2016/06, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jlwj02zz4mr-en.

Brezzi, M. and M. Diaz Ramirez (2016), “Building subjective well-being indicators at the subnational level: A preliminary assessment in OECD regions”, OECD Regional Development Working Papers, Vol. 2016/03, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jm2hhcjftvh-en.

Brynjolfsson, E., L.M. Hitt and H.H. Kim (2011), Strength in Numbers: How Does Data-Driven Decisionmaking Affect Firm Performance?, http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1819486.

Business Insider (2013), “Check Out The Extreme Lengths Airbnb Will Go To In Order To Please Customers”, published 14 August 2013, www.businessinsider.com/insane-lengths-airbnb-will-go-to-in-order-to-please-customers-2013-8?IR=T (accessed 20 June 2016).

Çağlar, A. (2014) “Urban Migration Trends, Challenges and Opportunities”, background paper for The World Migration Report 2015: Migrants and Cities: New Partnerships to Manage Mobility, International Organization for Migration (IOM), Geneva, Switzerland.

Caponio, T. (ed.) and M. Borkert (ed.) (2010), The local dimension of Migration policy-making, IMISCOE Reports, Amsterdam University Press, Amsterdam, www.imiscoe.org/publications/library/4-imiscoe-reports-series/66-the-local-dimension-of-migration-policymaking.

Carlino, G. and W. Kerr (2014), “Agglomeration and Innovation”, Harvard Business School Working Paper, 15-007, 5 August 2014.

Caves, D., W. L. Christensen and E. Diewert (1982), “Multilateral Comparisons of Output, Input, and Productivity Using Superlative Index Numbers”, Economic Journal, Royal Economic Society, Vol. 92(365), pp. 73-86.

Charron, N., L. Dijkstra and V. Lapuente (2014), “Mapping the Regional Divide in Europe: A Measure for Assessing Quality of Government in 206 European Regions”, Social Indicators Research, http://dx.doi.org/10.1007/s11205-014-0702-y.

Chatry I. and C. Hulbert (forthcoming), “Multi-level governance reforms: Overview of OECD country experiences”, OECD Publications, Paris.

Chetty, R. and N. Hendren (2015), “The impacts of neighborhoods on intergenerational mobility: Childhood exposure effects and county-level estimates”, mimeo.

Chetty, R., N. Hendren, P. Kline and E. Saez (2014), “Where is the land of Opportunity? The Geography of Intergenerational Mobility in the United States”, The Quarterly Journal of Economics, Vol. 129(4), pp. 1 553-1 623.

Congressional Budget Office (1991), How Federal Spending for Infrastructure and Other Public Investments Affects the Economy, Congressional Budget Office, Washington, DC.

Cortinovis, N., J. Xiao, R. Boschma and F. van Oort (2016), “Quality of government and social capital as drivers of regional diversification in Europe”, Papers in Evolutionary Economic Geography, No. 16.10, Urban and Regional Research Centre Utrecht (URU), Utrecht University.

Creel, J. and G. Poilon (2008), “Is Public Capital Productive in Europe?”, International Review of Applied Economics, Vol. 22(6).

Criscuolo, C. (2015), “Productivity Is Soaring at Top Firms and Sluggish Everywhere Else”, Harvard Business Review, 24 August 2015, https://hbr.org/2015/08/productivity-is-soaring-at-top-firms-and-sluggish-everywhere-else.

Criscuolo, C., P.N. Gal and C. Menon (2014), “The Dynamics of Employment Growth: New Evidence from 18 Countries”, OECD Science, Technology and Industry Policy Papers, No. 14, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jz417hj6hg6-en.

Cullison, W. (1993), “Public Investment and Economic Growth”, Federal Reserve Bank of Richmond Economic Quarterly, Vol. 79(4).

D’Costa, S., E. Garcilazo and J. Oliveira Martins (2013), “The Impact of Structural and Macroeconomic Factors on Regional Growth”, OECD Regional Development Working Papers, No. 2013/11, OECD Publishing, Paris, http://dx.doi.org/10.1787/5k451mplq9lw-en.

Dupont, J., D. Guellec and J. Oliveira Martins (2010), “OECD Productivity Growth in the 2000s: A Descriptive Analysis of the Impact of Sectoral Effects and Innovation”, OECD Journal: Economic Studies, Vol. 2011/1, http://dx.doi.org/10.1787/eco_studies-2011-5kgf3281fmtc.

Duranton, G. and D. Puga (2004), “Micro-foundations of urban agglomeration economies”, Handbook of Regional and Urban Economics in: Henderson, J.V. and J.F. Thisse (eds.), Handbook of Regional and Urban Economics, Vol. 4, Ch. 48, pp 2 063-2 117.

EC (2015), “Why the EU needs an investment plan?”, Factsheets on the Investment Plan, European Commission and European Investment Bank, 8 December 2015, https://ec.europa.eu/priorities/publications/why-eu-needs-investment-plan_en.

EC JRC IPTS (2015), Stairway to Excellence Facts and Figures: Malopolskie, Institute for Prospective Technologies, Joint Research Centre, European Commission.

EUROSTAT (2014), Structural Business Statistics – Regional Data – All Activities (database), EUROSTAT.

EU Urban Audit (n.d.), “Functional urban areas (urb_luz)” (database), http://ec.europa.eu/eurostat/web/cities/data/database (accessed June 2016).

IOM (2015), “World Migration Report 2015: Migrants and Cities, New Partnerships to Manage Mobility”, International Organization for Migration, Geneva, Switzerland.

Krugman, P. (1997), The Age of Diminished Expectations: U.S. economic policy in the 1990s, 3rd edition, MIT Press.

Machin, S., K.G. Salvanes and P. Pelkonen (2012), “Education and Mobility”, Journal of the European Economic Association, Vol. 10(2), pp. 417-450.

Maguire, K. and J. Weber (forthcoming), “Should we care about gaps in regional innovation capacities?”, OECD Regional Development Working Papers, OECD Publishing, Paris.

Malmquist, S. (1953), “Index Numbers and Indifference Surfaces”, Trabajos de Estatistica, Vol. 4, pp. 209-242.

Manning, A. (2004), “We Can Work it Out: The Impact of Technological Change on the Demand for Low-Skilled Workers”, Scottish Journal of Political Economy, Vol. 51(5), pp. 581-608.

Manning, A. and B. Petrongolo (2015), “How Local Are Labor Markets? Evidence from a Spatial Job Search Model”, mimeo.

Marinescu, I. and R. Rathelot (2016), “Mismatch unemployment and the geography of job search”, mimeo.

Market Watch (2015), “Opinion: What really keeps Airbnb’s CEO up at night”, published 13 February 2015, www.marketwatch.com/story/what-really-keeps-airbnbs-ceo-up-at-night-2015-02-13 (accessed 20 June 2016).

McKinsey (2013), Infrastructure productivity: How to save $1 trillion a year, McKinsey & Company, January 2013, www.mckinsey.com/industries/infrastructure/our-insights/infrastructure-productivity (accessed 19 June 2016).

Moretti, E. (2010), “Local Multipliers”, American Economic Review, Vol. 100(2), pp. 373-77.

Moretti, E. (2004), “Estimating the Social Return to Higher Education: Evidence from Longitudinal and Repeated Cross-Sectional Data”, Journal of Econometrics, Vol. 21(1-2), pp. 175-212.

Moretti, E. and P. Thulin (2013), “Local multipliers and human capital in the United States and Sweden”, Industrial and Corporate Change, Vol. 22(1), pp. 339-362.

OECD (2016a), Productivity Statistics (database), www.oecd.org/std/productivity-stats/ (accessed 17 March 2016).

OECD (2016b), OECD Regions at a Glance 2016, OECD Publishing, Paris, http://dx.doi.org/10.1787/reg_glance-2016-en.

OECD (2016c), “Metropolitan areas”, OECD Regional Statistics (database), http://dx.doi.org/10.1787/data-00531-en (accessed 20 June 2016).

OECD (2016d), Income Distribution (IDD) (database), www.oecd.org/social/income-distribution-database.htm, (accessed 18 June 2016).

OECD (2016e), OECD Compendium of Productivity Indicators 2016, OECD Publishing, Paris, http://dx.doi.org/10.1787/pdtvy-2016-en.

OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

OECD (2016g), Regional Well-Being (database), www.oecdregionalwellbeing.org/ (accessed 12 June 2016).

OECD (2016h), Making Integration Work: Refugees and others in need of protection, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264251236-en.

OECD (2016i), National Accounts Statistics (database), www.oecd-ilibrary.org/economics/data/oecd-national-accounts-statistics_na-data-en (accessed 2 June 2016).

OECD (2015a), The Future of Productivity, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264248533-en.

OECD (2015b), In It Together: Why Less Inequality Benefits All, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264235120-en.

OECD (2015c), Data-Driven Innovation: Big Data for Growth and Well-Being, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264229358-en.

OECD (2015d), The Metropolitan Century: Understanding Urbanisation and its Consequences, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264228733-en.

OECD (2015e), OECD Science, Technology and Industry Scoreboard 2015: Innovation for growth and society, OECD Publishing, Paris, http://dx.doi.org/10.1787/sti_scoreboard-2015-en.

OECD (2014a), OECD Regional Outlook 2014: Regions and Cities: Where Policies and People Meet, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264201415-en.

OECD (2014b), Innovation and Modernising the Rural Economy, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264205390-en.

OECD (2014c), How’s Life in Your Region?: Measuring Regional and Local Well-being for Policy Making, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264217416-en.

OECD (2013a), OECD Regions at a Glance 2013, OECD Publishing, Paris, http://dx.doi.org/10.1787/reg_glance-2013-en.

OECD (2013b), Investing Together: Working Effectively across Levels of Government, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264197022-en.

OECD (2013c), Regions and Innovation: Collaborating across Borders, OECD Reviews of Regional Innovation, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264205307-en.

OECD (2012), Promoting Growth in All Regions, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264174634-en.

OECD (2011a), OECD Regional Outlook 2011: Building Resilient Regions for Stronger Economies, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264120983-en.

OECD (2011b), “Regions and Innovation Policy”, OECD Reviews of Regional Innovation, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264097803-en.

OECD (2009), OECD Rural Policy Reviews: Spain 2009, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264060074-en.

OECD (2007), Infrastructure to 2030 (Vol.2): Mapping Policy for Electricity, Water and Transport, OECD Publishing, Paris, http://dx.doi.org/10.1787/9789264031326-en.

OECD-COR (2016), Results of the OECD-CoR Consultation of Sub-national Governments: Infrastructure planning and investment across levels of government: Current challenges and possible solutions, https://portal.cor.europa.eu/europe2020/pub/Documents/oecd-cor-jointreport.pdf.

Perek-Białas,J., C. Martinez-Fernandez and T. Weyman (2013), “Malopolska Region Demographic Transition: Working for the Future”, OECD Local Economic and Employment Development (LEED) Working Papers, No. 2013/06, OECD Publishing, Paris, http://dx.doi.org/10.1787/5k4818gwg2jk-en.

PWC (2015), Pricewaterhouse Coopers/National Venture Capital Association MoneyTree™ Report (2015).

Rhys A., G. Boyne, L. O’ Toole, K. Meier and R. Walker (2013), “Managing Migration? EU enlargement, local government capacity and performance in England”, Public Administration, Vol. 91(1), pp. 174-194.

Rodríguez-Pose, A. and R. Crescenzi (2008), “Research and development, spillovers, innovation systems, and the genesis of regional growth in Europe”, Regional studies, Vol. 42 (1). pp. 51-67.

Rodrik, D. (2016), “Premature deindustrialization”, Journal of Economic Growth, Vol. 21(1), pp. 1-33.

Rodrik, D. (2013), “Unconditional Convergence in Manufacturing”, Quarterly Journal of Economics, Vol. 128(1), pp. 165-204.

Skype (2010), Form S-1, Securities and Exchange Commission, United States, https://www.sec.gov/Archives/edgar/data/1498209/000119312510182561/ds1.htm (accessed 20 June 2016).

Sterlacchini, A. (2008), “R&D, higher education and regional growth: Uneven linkages among European regions”, Research Policy, Vol. 37, pp. 1 096-1 107.

Sutherland, D. et al. (2009), “Infrastructure Investment: Links to Growth and the Role of Public Policies”, OECD Economics Department Working Papers, No. 686, OECD Publishing, http://dx.doi.org/10.1787/225678178357.

Tobocman, S. (2014) “Revitalizing Detroit: is there a role for migration?”, Migration Policy Institute, August, Washington DC, USA.

University of Gothenburg (2013), European Quality of Government Index (EQI), http://qog.pol.gu.se/data/datadownloads/qog-eqi-data.

Veneri,P. and F. Murtin (2016), “Where is inclusive growth happening? Mapping multi-dimensional living standards in OECD regions”, OECD Statistics Working Papers, No. 2016/01, OECD Publishing, Paris, http://dx.doi.org/10.1787/5jm3nptzwsxq-en.

Walker, K. and H. Leitner (2011), “The Variegated Landscape of Local Immigration Policies in the United States”, Urban Geography, Vol. 32(2), pp. 156-178.

Wall Street Journal (2013), “Airbnb Slows Hiring Overseas, Shifts Strategy”, published 15 July 2013, www.wsj.com/news/articles/SB10001424127887323394504578608192000978414 (accessed 20 June 2016).

Table 1.A1.1. Categorisation of OECD regions by within-country catching-up dynamics
Productivity and employment indicators for frontier (F), catching-up (C), keeping-pace (K) and diverging (D) regions

Country

Region

Productivity typology

Productivity and Employment

2013 or closest year available

Employment growth

2000-13

Productivity growth

2000-13

Tradable share

2013

Productivity (USD PPP)

Employment (000)

Frontier Productivity (USD PPP)

Absolute change (000)

Employment growth (%)

Productivity annual growth (%)

Frontier shift effect

Catch-up effect

GVA (%)

Employment (%)

AUS

Australian Capital Territory

F

109 278

212

128 716

39

1.56

2.28

3.23

-0.95

13.0

12.8

New South Wales

D

88 477

3 586

128 716

616

1.46

0.39

3.23

-2.83

36.1

26.2

Northern Territory

F

112 262

126

128 716

35

2.52

1.74

3.23

-1.48

29.5

21.7

Queensland

D

82 966

2 301

128 716

665

2.66

1.35

3.23

-1.88

34.5

25.0

South Australia

D

77 663

805

128 716

131

1.38

0.74

3.23

-2.49

35.9

27.2

Tasmania

D

69 590

232

128 716

33

1.19

0.87

3.23

-2.36

36.1

25.9

Victoria

D

79 072

2 855

128 716

651

2.01

0.13

3.23

-3.10

34.0

27.1

 

Western Australia

F

128 716

1 324

128 716

410

2.89

3.55

3.23

0.32

47.0

30.1

AUT

Burgenland

C

68 439

124

92 398

13

0.84

0.91

0.11

0.80

33.1

33.2

Carinthia

K

72 632

275

92 398

22

0.64

0.40

0.11

0.29

35.8

33.3

Lower Austria

C

78 299

729

92 398

68

0.76

0.66

0.11

0.55

34.6

32.6

Salzburg

C

84 159

322

92 398

43

1.10

0.76

0.11

0.64

29.5

28.0

Styria

C

74 339

632

92 398

67

0.86

0.73

0.11

0.62

39.4

35.6

Tyrol

C

80 451

404

92 398

61

1.27

0.64

0.11

0.53

31.3

28.8

Upper Austria

C

80 250

771

92 398

96

1.03

0.64

0.11

0.53

43.9

37.1

Vienna

F

92 398

1 013

92 398

119

0.96

0.11

0.11

0.00

29.2

24.0

 

Vorarlberg

C

87 364

193

92 398

25

1.07

0.66

0.11

0.55

42.0

35.4

BEL

Antwerp

K

106 924

779

115 131

82

0.86

0.42

0.56

-0.13

35.7

24.8

Brussels-Capital Region

F

115 131

689

115 131

48

0.55

0.56

0.56

0.00

36.7

22.8

East Flanders

K

92 185

570

115 131

65

0.94

0.74

0.56

0.18

29.1

24.4

Flemish Brabant

K

104 585

440

115 131

45

0.83

0.76

0.56

0.20

24.5

20.5

Hainaut

K

79 965

429

115 131

37

0.70

0.27

0.56

-0.29

26.4

21.9

Liège

K

83 375

381

115 131

28

0.59

0.56

0.56

0.01

27.3

21.9

Limburg (BE)

K

84 304

331

115 131

32

0.78

0.43

0.56

-0.13

29.0

26.0

Luxembourg (BE)

K

75 598

93

115 131

9

0.80

0.20

0.56

-0.35

20.7

21.8

Namur

K

81 722

167

115 131

20

0.98

0.70

0.56

0.14

20.5

19.6

Walloon Brabant

D

107 970

154

115 131

33

1.91

-0.05

0.56

-0.61

40.0

24.2

 

West Flanders

K

86 907

508

115 131

38

0.59

0.59

0.56

0.03

28.9

27.5

CAN

Alberta

F

117 755

2 211

117 853

628

2.60

1.65

1.65

-0.01

42.6

29.6

British Columbia

D

76 615

2 308

117 853

377

1.38

0.57

1.65

-1.08

28.2

24.5

Manitoba

D

74 562

633

117 853

81

1.06

1.11

1.65

-0.54

32.9

28.6

New Brunswick

D

69 932

351

117 853

20

0.44

0.72

1.65

-0.93

28.4

25.5

Newfoundland and Labrador

F

118 502

233

117 853

34

1.22

3.69

1.65

2.04

47.8

24.9

Nova Scotia

D

66 412

454

117 853

42

0.75

0.30

1.65

-1.35

25.3

23.0

Ontario

D

77 859

6 879

117 853

1 065

1.30

-0.14

1.65

-1.80

32.3

28.2

Prince Edward Island

D

60 138

74

117 853

11

1.28

0.62

1.65

-1.03

27.4

26.3

Québec

D

69 281

4 032

117 853

631

1.32

-0.04

1.65

-1.69

32.8

28.1

 

Saskatchewan

C

115 384

555

117 853

82

1.24

3.35

1.65

1.70

49.5

30.0

CZE

Central Bohemian Region

D

56 378

559

78 211

71

1.04

1.54

2.03

-0.48

46.6

41.4

Central Moravia

K

49 641

544

78 211

2

0.02

2.23

2.03

0.20

46.7

43.9

Moravia-Silesia

C

52 511

534

78 211

9

0.13

2.51

2.03

0.49

49.6

41.9

Northeast

K

50 098

675

78 211

-21

-0.23

1.98

2.03

-0.05

49.5

46.4

Northwest

K

48 157

481

78 211

-13

-0.20

1.73

2.03

-0.30

47.2

40.5

Prague

F

78 211

914

78 211

131

1.19

2.03

2.03

0.00

39.0

24.8

Southeast

C

53 691

802

78 211

36

0.35

2.59

2.03

0.56

44.6

41.6

Southwest

K

50 885

576

78 211

10

0.13

1.96

2.03

-0.07

46.4

43.5

DNK

Capital

F

92 078

973

92 078

51

0.42

0.59

0.59

0.00

35.2

25.1

Central Jutland

K

77 286

620

92 078

4

0.05

0.45

0.59

-0.14

31.0

26.9

Northern Jutland

D

74 738

276

92 078

-4

-0.11

0.15

0.59

-0.44

31.1

27.2

Southern Denmark

K

79 653

558

92 078

-24

-0.32

0.62

0.59

0.03

31.0

26.5

 

Zealand

K

74 425

316

92 078

-8

-0.19

0.39

0.59

-0.19

26.9

22.2

EST

Estonia

.

53 797

601

53 797

16

0.20

3.42

3.42

0.00

37.3

35.5

FIN

Åland

C

76 472

18

93 088

2

0.87

1.40

0.70

0.70

24.1

25.1

Eastern and Northern Finland

K

74 265

550

93 088

44

0.69

0.72

0.70

0.01

34.8

32.1

Helsinki-Uusimaa

F

93 088

837

93 088

89

0.95

0.70

0.70

0.00

32.3

27.3

Southern Finland

K

77 891

509

93 088

32

0.55

0.40

0.70

-0.30

36.1

30.8

 

Western Finland

K

76 809

623

93 088

73

1.04

0.76

0.70

0.06

41.1

35.7

FRA

Alsace

D

82 436

764

117 670

3

0.03

0.40

1.15

-0.75

32.1

..

Aquitaine

D

79 523

1 339

117 670

122

0.74

0.60

1.15

-0.55

26.6

..

Auvergne

D

76 205

530

117 670

2

0.03

0.50

1.15

-0.65

28.2

..

Brittany

D

75 697

1 301

117 670

100

0.62

0.33

1.15

-0.82

29.0

..

Burgundy

D

77 610

642

117 670

-11

-0.13

0.28

1.15

-0.87

28.1

..

Centre-Val de Loire

D

78 591

995

117 670

-3

-0.03

0.35

1.15

-0.80

31.2

..

Champagne-Ardenne

D

82 745

520

117 670

-28

-0.41

0.35

1.15

-0.80

34.8

..

Corsica

D

78 653

124

117 670

30

2.17

0.58

1.15

-0.57

15.0

..

Franche-Comté

D

74 703

439

117 670

-18

-0.31

0.16

1.15

-0.99

31.1

..

Île-de-France

F

117 670

6 081

117 670

298

0.39

1.15

1.15

0.00

30.9

..

Languedoc- Roussillon

D

79 370

961

117 670

123

1.06

0.69

1.15

-0.46

20.4

..

Limousin

D

71 647

280

117 670

-7

-0.20

0.26

1.15

-0.89

24.1

..

Lorraine

D

77 977

825

117 670

-50

-0.45

0.33

1.15

-0.82

26.3

..

Lower Normandy

D

77 017

572

117 670

-5

-0.06

0.66

1.15

-0.49

27.2

..

Midi-Pyrénées

K

79 099

1 214

117 670

140

0.95

0.79

1.15

-0.36

26.8

..

Nord-Pas-de-Calais

D

79 567

1 509

117 670

25

0.13

0.74

1.15

-0.40

26.7

..

Pays de la Loire

D

79 454

1 508

117 670

120

0.64

0.72

1.15

-0.43

31.2

..

Picardy

D

81 363

661

117 670

-19

-0.22

0.42

1.15

-0.73

29.1

..

Poitou-Charentes

D

76 437

691

117 670

26

0.29

0.65

1.15

-0.50

30.5

..

Provence-Alpes-Côte d’Azur

D

86 596

1 997

117 670

227

0.93

0.55

1.15

-0.60

22.1

..

Rhône-Alpes

D

86 341

2 715

117 670

195

0.57

0.73

1.15

-0.42

28.8

..

 

Upper Normandy

D

83 442

700

117 670

-8

-0.09

0.48

1.15

-0.67

32.2

..

DEU

Baden-Württemberg

C

84 826

5 945

92 266

538

0.73

0.42

-0.01

0.42

..

..

Bavaria

C

85 336

7 082

92 266

690

0.79

0.69

-0.01

0.70

..

..

Berlin

K

75 895

1 774

92 266

182

0.84

0.08

-0.01

0.09

..

..

Brandenburg

C

66 841

1 082

92 266

4

0.03

1.01

-0.01

1.01

..

..

Bremen

K

84 640

417

92 266

24

0.46

0.31

-0.01

0.32

..

..

Hamburg

F

101 034

1 180

92 266

125

0.86

-0.19

-0.01

-0.19

..

..

Hesse

F

88 872

3 272

92 266

192

0.47

0.05

-0.01

0.06

..

..

Lower Saxony

C

76 048

3 887

92 266

332

0.69

0.41

-0.01

0.42

..

..

Mecklenburg- Vorpommern

C

61 269

730

92 266

-37

-0.38

0.90

-0.01

0.91

..

..

North Rhine-Westphalia

C

80 367

9 038

92 266

464

0.41

0.43

-0.01

0.43

..

..

Rhineland-Palatinate

K

76 281

1 952

92 266

167

0.69

0.32

-0.01

0.32

..

..

Saarland

C

75 208

519

92 266

6

0.09

0.70

-0.01

0.70

..

..

Saxony

C

62 445

2 010

92 266

17

0.07

1.12

-0.01

1.12

..

..

Saxony-Anhalt

C

64 568

1 015

92 266

-54

-0.40

1.08

-0.01

1.09

..

..

Schleswig-Holstein

K

73 040

1 334

92 266

73

0.43

0.23

-0.01

0.24

..

..

 

Thuringia

C

60 218

1 047

92 266

-39

-0.28

1.23

-0.01

1.23

..

..

GRC

Attica

F

85 585

1 471

85 585

-124

-0.62

0.80

0.80

0.00

24.1

28.0

Central Greece

D

61 242

188

85 585

-17

-0.66

-1.22

0.80

-2.02

40.9

44.3

Central Macedonia

K

58 093

612

85 585

-110

-1.26

0.78

0.80

-0.03

27.7

34.7

Crete

K

55 132

235

85 585

-36

-1.10

0.97

0.80

0.16

23.8

35.4

East Macedonia – Thrace

K

52 075

206

85 585

-34

-1.18

0.80

0.80

0.00

26.7

42.6

Epirus

K

54 132

115

85 585

-13

-0.82

0.55

0.80

-0.25

23.7

33.6

Ionian Islands

D

56 746

83

85 585

-3

-0.24

-0.43

0.80

-1.24

16.7

31.0

North Aegean

D

57 863

68

85 585

2

0.27

0.33

0.80

-0.47

17.3

26.6

Peloponnese

D

55 975

207

85 585

-22

-0.76

0.32

0.80

-0.48

33.4

46.9

South Aegean

D

67 219

129

85 585

6

0.39

-0.82

0.80

-1.62

14.1

22.7

Thessaly

K

53 951

254

85 585

-25

-0.73

0.41

0.80

-0.39

30.6

40.2

West Greece

K

56 364

223

85 585

-46

-1.43

1.03

0.80

0.23

27.1

39.7

 

West Macedonia

K

66 171

87

85 585

-14

-1.12

0.51

0.80

-0.29

49.4

42.2

HUN

Central Hungary

F

61 976

1 680

61 976

85

0.43

2.24

2.24

0.00

38.3

29.8

Central Transdanubia

D

48 789

424

61 976

-29

-0.56

1.49

2.24

-0.75

51.4

49.8

Northern Great Plain

K

45 963

452

61 976

-43

-0.75

2.08

2.24

-0.16

40.9

42.4

Northern Hungary

D

44 718

347

61 976

-38

-0.87

1.46

2.24

-0.78

43.8

41.9

Southern Great Plain

D

43 551

437

61 976

-59

-1.05

1.67

2.24

-0.58

41.1

44.4

Southern Transdanubia

D

45 793

298

61 976

-51

-1.30

1.81

2.24

-0.43

38.3

39.3

 

Western Transdanubia

D

51 439

421

61 976

-43

-0.81

1.76

2.24

-0.48

52.7

48.2

IRL

Border, Midland and Western

C

81 455

433

114 234

16

0.29

1.65

1.14

0.50

43.9

30.5

 

Southern and Eastern

F

114 234

1 449

114 234

170

0.97

1.14

1.14

0.00

51.9

32.8

ITA

Abruzzo

K

72 045

518

94 756

1

0.02

-0.26

-0.38

0.12

33.1

38.5

Aosta Valley

F

92 735

60

94 756

3

0.34

0.22

-0.38

0.60

28.0

27.7

Apulia

K

63 657

1 293

94 756

-73

-0.42

-0.38

-0.38

-0.01

27.7

35.8

Basilicata

K

67 862

194

94 756

-15

-0.56

-0.22

-0.38

0.15

36.1

38.6

Calabria

K

60 884

608

94 756

-14

-0.18

-0.60

-0.38

-0.23

22.6

36.2

Campania

K

64 931

1 863

94 756

-21

-0.09

-0.15

-0.38

0.23

27.2

33.3

Emilia-Romagna

K

85 437

2 061

94 756

69

0.26

-0.18

-0.38

0.19

38.5

39.8

Friuli-Venezia Giulia

K

80 493

538

94 756

13

0.19

-0.55

-0.38

-0.17

36.3

39.3

Lazio

F

87 705

2 569

94 756

363

1.18

-0.88

-0.38

-0.50

31.8

32.1

Liguria

F

90 325

650

94 756

1

0.02

-0.22

-0.38

0.15

25.2

29.5

Lombardy

F

94 756

4 647

94 756

416

0.72

-0.38

-0.38

0.00

40.1

39.7

Marche

K

74 757

635

94 756

14

0.17

-0.48

-0.38

-0.10

37.0

42.8

Molise

K

70 575

103

94 756

-6

-0.45

-0.66

-0.38

-0.29

28.0

34.5

Piedmont

K

84 266

1 834

94 756

63

0.27

-0.73

-0.38

-0.36

38.8

40.6

Province of Bolzano-Bozen

K

90 817

278

94 756

40

1.20

-0.25

-0.38

0.12

29.9

29.6

Province of Trento

K

87 088

256

94 756

28

0.88

-0.67

-0.38

-0.30

30.4

31.6

Sardinia

K

67 596

563

94 756

-7

-0.10

-0.33

-0.38

0.05

24.8

32.3

Sicily

K

69 413

1 466

94 756

-50

-0.26

-0.40

-0.38

-0.02

24.1

32.2

Tuscany

K

80 919

1 639

94 756

121

0.59

-0.45

-0.38

-0.08

34.1

37.1

Umbria

72 507

370

94 756

20

0.43

-0.74

-0.38

-0.36

32.7

39.3

 

Veneto

K

83 827

2 150

94 756

126

0.47

-0.63

-0.38

-0.26

38.1

41.3

KOR

Capital Region

D

63 932

12 528

77 259

1 631

1.56

2.09

3.33

-1.24

43.5

63.8

Chungcheong Region

F

77 259

2 648

77 259

402

1.85

3.33

3.33

0.00

59.2

67.9

Gangwon Region

D

58 214

698

77 259

28

0.46

2.59

3.33

-0.73

31.0

64.6

Gyeongbuk Region

D

59 611

2 581

77 259

43

0.19

2.52

3.33

-0.81

56.2

66.2

Gyeongnam Region

D

71 903

3 809

77 259

237

0.72

2.55

3.33

-0.78

57.7

64.5

Jeju

K

49 809

305

77 259

27

1.02

2.82

3.33

-0.51

34.0

64.4

 

Jeolla Region

D

61 895

2 499

77 259

141

0.65

2.56

3.33

-0.77

52.3

66.7

NLD

Drenthe

D

66 817

226

107 172

-7

-0.24

0.46

1.24

-0.79

33.6

22.1

Flevoland

K

77 192

172

107 172

18

0.90

0.91

1.24

-0.34

24.7

20.6

Friesland

K

68 285

304

107 172

-2

-0.05

0.88

1.24

-0.36

36.2

23.2

Gelderland

D

73 359

1 026

107 172

23

0.19

0.75

1.24

-0.50

30.3

21.6

Groningen

F

131 283

278

107 172

-9

-0.28

3.20

1.24

1.95

61.7

22.6

Limburg (NL)

D

72 008

556

107 172

-21

-0.31

0.71

1.24

-0.53

33.5

24.3

North Brabant

K

82 476

1 336

107 172

20

0.12

0.88

1.24

-0.36

36.5

23.7

North Holland

F

95 829

1 517

107 172

22

0.12

1.00

1.24

-0.24

32.4

22.0

Overijssel

D

69 682

590

107 172

17

0.25

0.81

1.24

-0.44

32.5

23.0

South Holland

F

87 043

1 791

107 172

-28

-0.13

-0.49

1.24

-1.73

28.8

19.5

Utrecht

D

90 910

715

107 172

-4

-0.05

0.79

1.24

-0.45

32.0

21.9

 

Zeeland

C

73 290

181

107 172

0

0.02

1.87

1.24

0.63

37.1

23.6

NZL

Auckland Region

F

71 878

712

78 341

576

14.76

0.50

0.51

-0.04

38.0

27.0

Bay of Plenty Region

C

66 064

113

78 341

60

6.46

1.97

0.51

1.30

44.0

32.2

Canterbury Region

C

58 845

324

78 341

207

8.90

1.60

0.51

0.96

43.0

29.7

Gisborne/Hawke’s Bay

D

48 987

103

78 341

-300

-10.76

-0.57

0.51

-1.04

47.0

35.9

Manawatu-Wanganui Region

D

51 352

112

78 341

-89

-4.76

-0.18

0.51

-0.68

40.8

31.2

Northland Region

K

54 281

67

78 341

37

7.00

0.88

0.51

0.30

47.9

33.7

Otago Region

K

53 242

116

78 341

-230

-8.69

0.58

0.51

0.03

42.8

28.8

Southland Region

K

62 054

56

78 341

-148

-10.21

0.14

0.51

-0.38

61.9

44.5

Taranaki Region

F

91 767

63

78 341

-250

-12.50

0.28

0.51

-0.25

72.1

41.3

Tasman-Nelson-Marlb./West Coast

C

55 236

93

78 341

46

5.86

1.18

0.51

0.57

46.6

35.1

Waikato Region

C

67 040

194

78 341

141

11.54

1.79

0.51

1.14

52.1

35.5

 

Wellington Region

K

70 729

270

78 341

223

15.71

0.34

0.51

-0.20

39.6

23.5

POL

Greater Poland

C

61 200

1 359

78 130

-45

-0.25

4.13

2.41

1.72

..

45.3

Kuyavian-Pomerania

C

50 589

759

78 130

-158

-1.44

4.42

2.41

2.01

..

44.5

Lesser Poland

C

50 497

1 306

78 130

-24

-0.14

3.87

2.41

1.46

..

42.3

Lodzkie

C

41 950

1 246

78 130

120

0.78

2.80

2.41

0.39

..

45.2

Lower Silesia

C

69 707

1 041

78 130

102

0.79

3.38

2.41

0.97

..

40.5

Lublin Province

C

35 493

956

78 130

5

0.04

3.23

2.41

0.82

..

47.9

Lubusz

C

47 838

398

78 130

-3

-0.07

3.19

2.41

0.78

..

39.7

Mazovia

F

78 130

2 403

78 130

480

1.73

2.41

2.41

0.00

..

38.5

Opole region

C

54 458

332

78 130

-11

-0.25

3.04

2.41

0.64

..

45.2

Podkarpacia

C

42 127

799

78 130

47

0.46

2.99

2.41

0.58

..

47.3

Podlasie

C

42 702

454

78 130

16

0.28

3.09

2.41

0.68

..

47.1

Pomerania

K

55 814

883

78 130

178

1.75

2.04

2.41

-0.36

..

37.4

Silesia

K

56 491

1 889

78 130

173

0.74

2.35

2.41

-0.06

..

40.9

Swietokrzyskie

D

37 556

549

78 130

89

1.37

1.47

2.41

-0.94

..

48.5

Warmian-Masuria

K

43 981

527

78 130

59

0.91

2.16

2.41

-0.24

..

40.7

 

West Pomerania

C

57 207

563

78 130

-80

-1.01

3.39

2.41

0.98

..

35.1

PRT

Alentejo

K

63 160

275

76 493

-44

-1.14

0.71

0.49

0.22

38.2

38.9

Algarve

C

62 320

181

76 493

-10

-0.41

1.14

0.49

0.65

18.5

22.1

Azores

C

58 781

97

76 493

-6

-0.45

1.46

0.49

0.97

26.8

31.8

Central Portugal

C

50 856

974

76 493

-206

-1.46

1.19

0.49

0.70

35.1

44.6

Lisbon

F

76 493

1 288

76 493

-71

-0.41

0.49

0.49

0.00

30.6

24.4

Madeira

C

60 681

104

76 493

-21

-1.43

2.28

0.49

1.79

18.0

26.9

 

North

C

49 087

1 528

76 493

-228

-1.06

1.00

0.49

0.51

36.4

44.5

SVK

Bratislava Region

F

91 576

423

91 576

61

1.20

3.76

3.76

0.00

36.5

25.5

Central Slovak Republic

D

53 086

519

91 576

30

0.46

3.26

3.76

-0.50

40.1

35.2

East Slovak Republic

K

55 440

515

91 576

12

0.18

3.47

3.76

-0.29

39.6

33.0

West Slovak Republic

D

60 602

735

91 576

65

0.71

3.15

3.76

-0.61

46.7

38.8

SVN

Eastern Slovenia

C

54 459

439

63 577

-34

-0.57

2.00

1.34

0.66

42.8

46.4

 

Western Slovenia

F

63 577

485

63 577

44

0.73

1.34

1.34

0.00

35.7

34.4

ESP

Andalusia

K

73 660

2 647

89 402

230

0.70

0.74

0.65

0.10

28.7

28.0

Aragon

C

81 663

557

89 402

16

0.23

1.25

0.65

0.60

38.1

34.9

Asturias

K

76 516

384

89 402

14

0.28

0.46

0.65

-0.19

33.1

30.9

Balearic Islands

K

79 118

459

89 402

52

0.93

0.31

0.65

-0.34

18.7

20.0

Basque Country

F

89 791

978

89 402

46

0.37

0.72

0.65

0.07

38.1

35.2

Canary Islands

K

76 788

732

89 402

69

0.76

0.36

0.65

-0.29

20.6

19.4

Cantabria

K

78 645

213

89 402

7

0.26

0.67

0.65

0.02

34.4

32.4

Castile and León

K

78 389

944

89 402

-10

-0.08

0.88

0.65

0.23

35.7

32.6

Castile-La Mancha

C

78 263

672

89 402

29

0.34

1.64

0.65

0.99

39.3

34.9

Catalonia

K

84 886

3 209

89 402

172

0.43

0.93

0.65

0.28

33.8

30.8

Ceuta

K

79 063

27

89 402

2

0.59

0.34

0.65

-0.31

11.5

12.9

Extremadura

C

70 223

335

89 402

-6

-0.13

1.33

0.65

0.68

30.2

30.1

Galicia

C

74 014

1 021

89 402

54

0.42

1.04

0.65

0.39

35.0

33.9

La Rioja

C

84 466

126

89 402

0

0.00

1.18

0.65

0.53

43.3

40.4

Madrid

F

88 935

3 042

89 402

440

1.21

0.61

0.65

-0.04

32.0

28.3

Melilla

K

77 067

25

89 402

0

-0.06

0.97

0.65

0.32

11.4

12.7

Murcia

K

69 117

537

89 402

97

1.55

0.27

0.65

-0.38

31.8

34.2

Navarra

C

87 159

280

89 402

-2

-0.06

1.27

0.65

0.62

43.3

40.1

 

Valencia

K

77 343

1 747

89 402

43

0.19

0.91

0.65

0.26

31.3

30.6

SWE

Central Norrland

D

80 690

170

106 206

-2

-0.09

1.10

1.61

-0.51

40.4

27.6

East Middle Sweden

K

81 544

714

106 206

38

0.42

1.40

1.61

-0.21

36.6

27.1

North Middle Sweden

D

78 865

361

106 206

0

0.00

0.89

1.61

-0.72

39.0

29.6

Småland with Islands

D

74 459

395

106 206

6

0.12

1.09

1.61

-0.52

39.4

33.8

South Sweden

D

77 965

648

106 206

58

0.72

0.80

1.61

-0.81

29.6

25.5

Stockholm

F

106 206

1 208

106 206

149

1.02

1.61

1.61

0.00

38.1

25.8

Upper Norrland

K

83 907

245

106 206

20

0.66

1.45

1.61

-0.16

42.9

26.3

 

West Sweden

D

82 100

931

106 206

103

0.91

0.85

1.61

-0.76

34.4

27.4

GBR

East Midlands

K

64 548

2 116

133 506

141

0.53

1.04

1.30

-0.26

31.2

..

East of England

D

68 791

2 914

133 506

273

0.76

0.59

1.30

-0.71

28.8

..

Greater London

F

133 506

3 898

133 506

539

1.15

1.30

1.30

0.00

37.6

..

North East England

K

61 035

1 143

133 506

73

0.51

0.95

1.30

-0.34

30.8

..

North West England

K

69 834

3 118

133 506

159

0.40

0.97

1.30

-0.32

31.0

..

Northern Ireland

D

63 434

796

133 506

111

1.16

-0.02

1.30

-1.31

28.0

..

Scotland

K

72 990

2 467

133 506

141

0.45

1.19

1.30

-0.10

33.3

..

South East England

K

81 954

4 263

133 506

251

0.47

1.07

1.30

-0.23

30.7

..

South West England

K

68 753

2 545

133 506

201

0.63

1.06

1.30

-0.23

30.2

..

Wales

D

59 611

1 343

133 506

115

0.69

0.73

1.30

-0.56

30.9

..

West Midlands

D

69 270

2 447

133 506

62

0.20

0.82

1.30

-0.47

31.1

..

 

Yorkshire and The Humber

K

63 589

2 459

133 506

200

0.65

0.90

1.30

-0.39

31.2

..

USA

Alabama

K

72 617

2 542

108 796

147

0.46

1.20

0.83

0.37

51.2

25.8

Alaska

F

117 589

462

108 796

72

1.31

2.47

0.83

1.64

52.7

22.7

Arizona

D

76 819

3 392

108 796

583

1.46

0.38

0.83

-0.45

41.6

21.9

Arkansas

C

71 264

1 578

108 796

94

0.48

1.64

0.83

0.81

48.6

24.8

California

K

97 845

21 449

108 796

2 169

0.82

0.79

0.83

-0.04

45.4

24.4

Colorado

K

81 585

3 352

108 796

428

1.06

0.61

0.83

-0.22

45.1

24.0

Connecticut

F

104 838

2 233

108 796

123

0.44

0.53

0.83

-0.30

46.7

26.9

Delaware

F

106 094

544

108 796

43

0.64

0.24

0.83

-0.59

53.7

25.9

District of Columbia

F

125 468

844

108 796

110

1.08

1.61

0.83

0.78

53.1

16.7

Florida

D

71 935

10 556

108 796

1 638

1.31

0.43

0.83

-0.40

36.2

22.0

Georgia

D

78 653

5 504

108 796

631

0.94

0.13

0.83

-0.70

45.4

23.4

Hawaii

C

81 313

876

108 796

123

1.17

1.40

0.83

0.57

37.5

15.9

Idaho

D

64 156

903

108 796

127

1.17

0.37

0.83

-0.46

46.0

22.6

Illinois

K

91 561

7 507

108 796

148

0.15

0.77

0.83

-0.06

44.2

25.4

Indiana

K

80 130

3 683

108 796

37

0.08

1.06

0.83

0.23

54.0

28.1

Iowa

C

78 350

2 019

108 796

105

0.41

1.88

0.83

1.05

56.0

27.5

Kansas

C

72 465

1 864

108 796

105

0.45

1.27

0.83

0.44

50.1

26.6

Kentucky

K

72 124

2 414

108 796

107

0.35

1.20

0.83

0.37

50.4

24.6

Louisiana

C

88 866

2 632

108 796

245

0.75

1.89

0.83

1.06

53.2

24.0

Maine

K

64 443

804

108 796

20

0.19

0.83

0.83

0.00

40.0

22.7

Maryland

C

92 639

3 475

108 796

387

0.91

1.40

0.83

0.57

42.8

19.0

Massachusetts

K

96 866

4 322

108 796

244

0.45

0.76

0.83

-0.07

40.4

23.1

Michigan

D

77 646

5 309

108 796

-299

-0.42

0.01

0.83

-0.82

46.9

26.1

Minnesota

K

82 037

3 552

108 796

225

0.51

1.04

0.83

0.21

44.6

26.1

Mississippi

K

64 288

1 536

108 796

59

0.30

1.15

0.83

0.32

50.5

24.2

Missouri

K

73 306

3 580

108 796

106

0.23

0.72

0.83

-0.11

45.7

23.6

Montana

C

63 852

639

108 796

86

1.12

2.05

0.83

1.22

44.5

21.6

Nebraska

C

82 465

1 258

108 796

85

0.54

2.36

0.83

1.53

50.6

24.7

Nevada

D

77 827

1 560

108 796

306

1.70

0.31

0.83

-0.52

36.3

19.7

New Hampshire

K

78 010

835

108 796

53

0.50

0.86

0.83

0.03

41.2

24.3

New Jersey

F

99 871

5 103

108 796

366

0.57

0.46

0.83

-0.37

37.2

22.0

New Mexico

K

79 826

1 079

108 796

117

0.89

0.85

0.83

0.02

51.9

20.1

New York

F

110 106

11 555

108 796

1 163

0.82

0.89

0.83

0.06

48.9

24.2

North Carolina

K

81 240

5 452

108 796

559

0.84

1.15

0.83

0.32

52.6

23.6

North Dakota

C

83 505

580

108 796

139

2.14

3.87

0.83

3.04

49.6

23.4

Ohio

K

80 111

6 663

108 796

-117

-0.13

0.89

0.83

0.06

47.3

25.6

Oklahoma

C

74 202

2 255

108 796

261

0.95

2.02

0.83

1.19

52.8

26.5

Oregon

C

85 778

2 265

108 796

175

0.62

1.58

0.83

0.75

53.9

24.3

Pennsylvania

K

82 936

7 322

108 796

427

0.46

0.94

0.83

0.11

42.4

24.6

Rhode Island

K

84 696

597

108 796

18

0.23

1.10

0.83

0.27

42.3

23.9

South Carolina

K

69 230

2 499

108 796

223

0.72

0.67

0.83

-0.16

47.6

24.7

South Dakota

C

73 610

576

108 796

66

0.95

2.04

0.83

1.21

55.6

25.1

Tennessee

K

74 168

3 710

108 796

246

0.53

0.99

0.83

0.16

42.9

24.3

Texas

C

95 244

15 505

108 796

3 366

1.90

1.68

0.83

0.85

52.9

25.5

Utah

K

73 438

1 743

108 796

366

1.83

1.14

0.83

0.31

47.4

25.4

Vermont

K

64 194

426

108 796

25

0.47

1.00

0.83

0.17

42.1

22.9

Virginia

K

88 070

4 899

108 796

502

0.84

1.17

0.83

0.34

42.5

20.1

Washington

K

96 899

3 985

108 796

460

0.95

1.09

0.83

0.26

50.3

24.1

West Virginia

C

73 118

916

108 796

40

0.35

1.52

0.83

0.69

50.7

23.0

Wisconsin

K

76 486

3 530

108 796

124

0.27

1.14

0.83

0.31

49.9

28.8

 

Wyoming

F

100 236

395

108 796

73

1.58

3.16

0.83

2.33

59.6

25.0

Note: Productivity is measured as GDP per worker. GDP and GVA are measured at constant PPP 2010 USD. Catching-up/diverging regions grew by at least 5 percentage points more/less than their national frontier over the 2000-13 period. The frontier is defined as the aggregation of regions with the highest GDP per worker and representing 10% of national employment. Regions are labelled “frontier regions” if they contribute a non-negligible percentage of their employment for several years during the 2000-13 period. The productivity of the frontier in 2013 includes only the frontier regions of 2013. Due to lack of regional data over the period, only 24 countries are included in the averages. Tradable sectors are defined by a selection of the 10 industries defined in the SNA 2008. They include: agriculture (A), industry (BCDE), information and communication (J), financial and insurance activities (K), and other services (RSTU). Non-tradable sectors are composed of construction, distributive trade, repairs, transport, accommodation, food services activities (GHI), real estate activities (L), business services (MN), and public administration (OPQ).

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016).

 http://dx.doi.org/10.1787/888933411933

Figure 1.A1.1. Labour productivity is mostly positively associated with economic aspects of well-being
picture

Note: Bars indicate the number of regions in each of four categories with regards to the well-being indicator. The top 10% are the regions with the best values that account for 10% of the country’s population. Regions that are improving/falling behind are those where the gap to the frontier narrowed/widened by more than 5 percentage points over the indicated period (or the closest available years). Regions with a constant gap are those within the 5 percentage point band. Colours indicate the regions that belong to the labour productivity frontier, the group of catching-up/diverging regions and those that are keeping pace with the frontier (see Boxes 1.2 and 1.3 for detailed definitions).

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016) and OECD (2016g), Regional Well-Being (database), www.oecdregionalwellbeing.org/ (accessed 12 June 2016).

 http://dx.doi.org/10.1787/888933411948

Figure 1.A1.2. The relationship between labour productivity and well-being is often complex
picture

Note: Bars indicate the number of regions in each of four categories with regards to the well-being indicator. The top 10% are the regions with the best values that account for 10% of the country’s population. Regions that are improving/falling behind are those where the gap to the frontier narrowed/widened by more than 5 percentage points over the indicated period (or the closest available years). Regions with a constant gap are those within the 5 percentage point band. Colours indicate the regions that belong to the labour productivity frontier, the group of catching-up/diverging regions and those that are keeping pace with the frontier (see Boxes 1.2 and 1.3 for detailed definitions).

Source: Calculations based on OECD (2016f), OECD Regional Statistics (database), http://dx.doi.org/10.1787/region-data-en (accessed 18 June 2016) and OECD (2016g), Regional Well-Being (database), www.oecdregionalwellbeing.org/ (accessed 12 June 2016).

 http://dx.doi.org/10.1787/888933411951

Notes

← 1. The statistical data for Israel are supplied by and under the responsibility of the relevant Israeli authorities. The use of such data by the OECD is without prejudice to the status of the Golan Heights, East Jerusalem and Israeli settlements in the West Bank under the terms of international law.

← 2. Andrews, Criscuolo and Gal (2015) use harmonised cross-country firm level data to identify the globally most productive firms in each 2-digit industry based on a number of definitions (e.g. the top 100 firms in each industry etc.) from 2001-09.

← 3. The data for inter-personal inequality within TL2 regions is currently only available for the year 2010 in the OECD Regional Database (OECD, 2016f).

← 4. The OECD typology classifies regions at the TL3 level into urban, intermediate and rural regions. See OECD (2016b) for details.

← 5. The best measure of labour productivity available at the regional level.

← 6. Frontier and lagging regions are defined as the regions with the highest (lowest) GDP per employee until the equivalent of 10% of national employment is reached.

← 7. In this case, the analysis was on OECD TL3 regions, which are smaller in size than TL2 regions.

← 8. See Box 1.1 in OECD (2013c) for more details.

← 9. Tradable sectors are defined by a selection of the ten industries defined in the SNA 2008. They include: agriculture (A), industry (BCDE), information and communication (J), financial and insurance activities (K), and other services (R to U). Non-tradable sectors are composed of construction, distributive trade, repairs, transport, accommodation, food services activities (GHI), real estate activities (L), business services (MN), and public administration (OPQ).

← 10. This analysis considered GDP per capita growth.

← 11. This analysis excludes several countries due to lack of data or an insufficient number of regions: Chile, Estonia, Iceland, Israel, Japan, Luxembourg, Mexico, Norway, Switzerland and Turkey.

← 12. This is evident in the population changes at highly disaggregated levels. In Europe, for example, a recent mapping exercise of population changes between 2001 and 2011 shows clear concentration in and around cities, with some rural areas (and their towns and villages) defying the trend (BBSR, 2015). From 2000-14, net changes in the share of the population in OECD countries by region type (TL3) reveal that predominantly rural regions experienced an increase in population shares in Belgium, Chile, Ireland, Switzerland and the United States, (OECD, 2016b).

← 13. The top 20% of regions are defined as those with the highest value of the indicator until the equivalent of 20% of the national population is reached. The same calculation is made for the bottom 20%. For example, if the value for the top 20% was 50% of the labour force with tertiary education, and the rate for the bottom 20% was 25%, that ratio would be 2. The performance of this ratio over time is then assessed.

← 14. Faster catching up was the main force behind the values for total R&D expenditures, higher education R&D expenditures, R&D personnel and the share of the labour force with tertiary education. For patents, the reasons for convergence of regions were more mixed. For business R&D intensity, convergence was driven, in part, by worsening performance in the leading regions along with a simultaneous increase in values in the bottom regions.

← 15. Note that in two-thirds of OECD countries with data, the total volume of business R&D declined after 2008, however most of those countries showed a rebound within three years. The intensity of business R&D (i.e. as a share of GDP) shows more fluctuation, and 17 out of 31 countries showed a decline after 2008. In a few countries, the business R&D intensity in 2011 was at a lower level than in 2000.

← 16. The normalised HHI ranges from 0 to 1 (concentrated in one region).The normalised Herfindahl-Hirschman Index formula is picture, for N > 1, and picture, where Si denotes the share of the chosen indicator in region i.

← 17. Based on ars technica (2013), available at: http://arstechnica.com/business/2013/09/skypes-secrets/4/ (accessed 20 June 2016).

← 18. According to Skype S.a r.l. S-1 filing to the Securities and Exchange Commission (Skype, 2010).

← 19. Data from MarketWatch (2015), available at: www.marketwatch.com/story/what-really-keeps-airbnbs-ceo-up-at-night-2015-02-13 (accessed 20 June 2016) and Wall Street Journal (2013), available at www.wsj.com/news/articles/SB10001424127887323394504578608192000978414 (accessed 20 June 2016).

← 20. Many of the created jobs are related to customer services to ensure 24/7 availability and support in local languages (Business Insider, 2013), available at: www.businessinsider.com/insane-lengths-airbnb-will-go-to-in-order-to-please-customers-2013-8?IR=T (accessed 20 June 2016).

← 21. The calculation of multidimensional living standards is based on the equivalent income approach, where, for different income groups, the monetised value of health status and unemployment are added to disposable income and aggregated with a generalised mean function to allow inequality to be taken into account.

← 22. Manning (2004) shows that employment rates of low-skilled workers are higher (and unemployment rates lower) in US metro areas with higher percentages of college graduates. He attributes this factor to a rising polarisation in jobs driven by high-skilled jobs that create demand for local non-tradable (low skill) services.

← 23. See for example IOM 2015; Walker and Leitner, 2011; Rhys et al., 2013; Caponio et al., 2010.

← 24. This analysis was based on the United States.

← 25. In this paper, there is no specific cut-off value for lagging regions, but the degree to which a region is “lagging” is defined by its gap relative to its country’s frontier in terms of productivity (i.e. GDP per worker).

← 26. See OECD (2013b) p.19 for additional discussions on the relationship between public investment and growth.

← 27. See OECD (2014a), based on communications with Dexia (July, 2012).

← 28. Data for 2014, per the OECD Metropolitan Database (OECD, 2016c).

← 29. Based on the share of respondents from 37 cities in the European Urban Audit Perception Survey who state that they are either “satisfied” or “very satisfied” with the public transport provision in their city. The difference between the two groups is statistically significant at the 95% confidence level.

← 30. See Boulant, Brezzi and Veneri (2016) for details. Most other studies of segregation are conducted at the neighbourhood level.