>

4.3. Related variety analysis

4.3. Related variety analysis
4.3. Related variety analysis logo

Related variety is a key concept in evolutionary economic geography that links knowledge spillovers to economic development, new growth paths and economic renewal (Asheim et al., 2011). It refers to the variety of industries within a region that are cognitively related (Frenken et al., 2007) and maximise the potential for learning opportunities and growth of existing industries as well as the local sources of growth for new industries (Boschma, 2014).

The concept of related variety has been developed in the recent literature as an attempt to respond empirically to the controversy known as MARS versus Jacobs, i.e. the theories of Marshall-Arrow-Romer on agglomeration externalities based on specialisation (localisation economies) that view spillovers to occur mainly within a single industry, versus Jacob’s theory of external economies being more evident in places where there is a variety of sectors (inter-industrial spillovers) (Glaeser et al., 1992). Localisation economies generally arise from labor market pooling in a given sector within a given area and subsequent increases in labour productivity in that specific sector which allow the emergence of knowledge spillovers. Jacob’s externalities, on the other hand, emerge due to savings in large-scale institutional operations (urbanisation economies) and the interactions among firms in different sectors that are in close geographical proximity to each other. Such interactions allow for the recombination of knowledge, ideas and practices among heterogeneous industries.

Frenken et al. (2007, 688) argued that spillovers within a region are expected to primarily occur “among related sectors, and only to a limited extent among unrelated sectors”. Jacob’s view of innovation is closely linked to Schumpeter’s concept of recombining pre-existing knowledge or artefacts in novel ways to create new products and services. In this sense, inter-industry spillovers are expected to be present mainly among sectors that draw on similar knowledge (either this refers to technology, to markets, institutional conditions, etc.) or have some degree of cognitive proximity (related variety) that will secure effective communication and interactive learning (Nooteboom, 2000).

The discourse on related variety also evolved around the relation of sectoral variety (unrelated variety) and regional economic resilience. Derived from business economics and portfolio theory, the existence of sectoral diversification reduces the risk of interdependent, sector-specific asymmetric shocks that trigger long term unemployment and economic decline (Boschma and Iammarino, 2009; Bristow et al., 2010).

In the recent literature, related variety associated with Jacobs-type externalities among related sectors appear to have a significant effect on regional employment growth and economic development (Frenken et al., 2007; Boschma and Iammarino, 2009; Neffke et al., 2011, Van Oort et al., 2013). Although empirical studies on unrelated variety provide mixed results (Content and Frenken, 2016), one should take into account not only quantitative approaches using indexes, but also qualitative studies on economic crisis and the resilience of regions.

Based on this discussion, the main question that arises is this: what is the best possible composition of industries within a region that will create most spillovers and will secure stability and economic growth?

Description of the method

Related/ unrelated variety is a relatively new concept to the scientific discourse and –so far- there is not a perfect way to measure it with the data that are already available at a cross-regional level. Existing measures of inter-industry relatedness include –but are not limited to- entropy indicators based on SIC-codes, co-occurrence of products within firms, input-output linkages (e.g. trade data) and the intensity of labour reallocation between industries (Boschma and Gianelli, 2014). Below, we provide a brief review of the main indicators used in empirical studies, along with some comments on their pros- and cons- and we continue with the selected indicator for the development of the online tool for the ONLINE S3 exercise.

Empirical measurements

At an empirical level, related variety has been mostly measured through country-level studies, and less in wider areas, mainly due to data limitations[9] (Van Oort et al., 2015). Frenken et al. (2007) in their seminal study measured related variety as the average entropy across employment in five-digit industries within each two-digit class using the Standard Industrial Classification, and unrelated variety as the entropy in employment across two-digit classes. Applying these measures to 40 Dutch NUTS3 regions they find that related variety positively affected employment growth and that unrelated variety was negatively related to unemployment growth. Other studies show the same results on other growth-related indicators such as value-added growth (Boschma et al., 2012) and labour productivity (Boschma and Iammarino, 2009).

Boschma and Iammarino (2009) estimate related variety by means of sectoral (3-digit) trade data in Italy by country of destination/origin, using data from ISTAT. They argue that extra-local linkages are essential in bringing new knowledge into the region, a fact which has been overlooked by the MARS vs. Jacobs literature. Their findings show that related variety (both at the regional level and through inter-regional trade linkages) positively affects regional growth although related and unrelated variety per se, do not affect regional growth.

Van Oort et al. (2015) extend the study of related variety to a pan-European level where they also distinguish between smaller and larger regions to account for differences in agglomeration forces. They conclude that variety has a positive effect on employment growth especially for small and medium sized urban regions.

Alternative measures of related variety include branching studies such as Hidalgo et al. (2007), which argues that the existence of exporting capabilities of a country/region in a specific group of related products increases the potential to export a related product that does not export yet. The difference here is that the focus is not on aggregate regional growth but on explaining the diversification into new products and industries. This increasing literature of studies (Klepper and Simons, 2000; Tanner, 2011; Neffke et al., 2011) showing that new industries branch out of existing related industries and that this branching process increases the possibility of new industries to survive has received attention. Finally, another method has recently been applied to measure relatedness (skill-relatedness) via the number of people changing jobs between two industries (Neffke et al., 2011, Neffke and Henning, 2013).

Although the abovementioned studies vary widely in their approach, for example, in terms of time period covered and spatial scales examined, as well as in terms of measures of relatedness and control variables, they seem to conclude that related variety has a positive effect on economic growth. There is also both quantitative and qualitative evidence collected through case studies, that technological competences accumulated through time affect diversification opportunities over time (Boschma and Gianelle, 2014).

Related/ Unrelated variety indicators

From the above exercise it is evident that the measurement of a region’s diversification over sectors is sensitive to the indicators applied. Existing measures of related variety present drawbacks as they lack systematic measurement in low levels of spatial analysis, or they over-emphasise industrial over service sectors (Frenken et al., 2007). However, following the majority of the empirical works, we will use an entropy measure, which can be decomposed at each sectoral digit level.

The index will compare 2-digit and 5-digit sector shares (%) and will estimate the entropy index for a regional / national entity. Assuming that all five-digit sectors i fall under a two-digit sector Sg, where g=1, …, G, we derive the two-digit shares Pg, by summing the five-digit shares pi:

and that ,

,

In this context, the method will allow for calculating the Related/Unrelated variety entropy indexes.

Related variety is concerned with the concentration of industries that present some form of similarity, It stimulates economic growth because it enhances effective interactive learning and innovation.

Unrelated variety concerns sectors that have no substantial complementary competences. They are beneficial for economic growth because they spread risks of sector specific shocks.

Usability and impact

Related variety has been used as a method to define regional diversification or the degree to which a region’s different industries have commonalities that allow knowledge exchange and spillovers to occur. RIS3 strategies are promoting regional diversification in that they assist in developing new areas of specialisation and new growth paths for the future (Boschma and Gianelle, 2014). Both concepts focus on relatedness in order to identify unused potentials and promising activities, while they also promote policies adjusted to the specific needs and available resources of the regions (Boschma, 2014).

In terms of policy making, the concept of related variety was used as an attempt to identify pathways to innovate and construct regional advantage in non-high tech regions which benefit less from policies that focus on R&D. Within the smart specialisation policy context, related variety is about learning and focusing on the context-specific intangible assets of a region, as existing specialisations and knowledge bases in a region provide the options for future diversification, while also bringing together industries and other areas of expertise (Boschma, 2014). At the implementation phase, related variety is a key method for assessing the potential industry branching towards new activities and niche markets.

Given that, ‘the theoretical notions of specialisation and variety seem too simplistic to capture the varied effects of an economy’s composition on its further development’ (Content et al., 2016), the degree of related variety should be considered along with a number of different tools that provide information in relation to the level of connectedness of a region, specialisation analysis, a detailed analysis of the economic history of a region etc.

Required data

Relevant data to measure industrial specialisation in terms of related variety is any system that classifies industries by –at least- a four-digit code. Such systems include:

  • the Standard Industrial Classification (SIC) code system that classifies industries by four-digits
  • The North American Industry Classification System (NAICS code) that classifies industries by six-digits

The International Standard Industrial Classification of all economic activities (ISIC), the international reference classification of productive activities into four digit classes

Relevant data sources

A number of existing databases and tools relevant for the application of the methods’ indicators are listed below:

  • National Statistical Offices
  • Geo-coded AMADEUS micro data (provided by Bureau van Dijk) on European firms aggregated into European regions
  • European Regional Database of Cambridge Econometrics
  • Eurostat
  • OECD database
  • LISA database on employment data
  • UN Comtrade

The analysis is based on the data used by existing empirical studies and does not necessarily reflect availability.

Implementation roadmap

Implementing a tool on related variety means to create and use of a tool that estimates regional specialisation in technologically related sectors. The outcome should not only provide numerical answer on the fields and levels of specialisation among related sectors but should give a schematic representation of the connections among industrial sectors. The process for implementing such a tool is the following:

Step 1. Selection of an area of analysis

Select areas of interest and level of analysis as related specialisation in lower spatial levels provide different results.

Step2. Automatic data collection and data normalisation

Examine non-normality of variables and perform the necessary corrections and transformations (log-transform or correct outliers). The step is conducted automatically by the tool without intervention of the user.

Step 3. Calculation of related/unrelated variety indexes

Calculate the indexes for the selected area based on the available data.

Step 4. Visualise results

Demonstrate results in terms of statistical information, network graphs and figures.

References
  • [9] For a detailed review of the empirical studies see Content and Frenken (2016) and Content et al. (2016)
  • Asheim, B., Boschma, R., & Cooke, P. (2011). Constructing Regional Advantage: Platform Policies Based on Related Variety and Differentiated Knowledge Bases. Regional Studies, 45(7), 893-904. http://dx.doi.org/10.1080/00343404.2010.543126
  • Boschma, R. (2014) Constructing regional advantage and smart specialization: Comparisons of two European policy concepts, Italian Journal of Regional Science (Scienze Regionali) 13 (1), 51-68.
  • Boschma, R. & Iammarino, S. (2009). Related Variety, Trade Linkages, and Regional Growth in Italy. Economic Geography, 85(3), 289-311. http://dx.doi.org/10.1111/j.1944-8287.2009.01034.x
  • Boschma, R. and Giannelle, C. (2014) Regional Branching and Smart Specialisation Policy, JRC Technical Reports, S3 Policy Briefings No. 06/2014, European Commission
  • Bristow, G. I., Healy, A., Norris, L., Wink, R., Kafkalas, G., Kakderi, C. Espenberg, K., Varblane, U., Sepp, V., Sagan, I., Masik, G., Sensier, M., Harding, A., Swash, A. and Carey, H. (2014). ECR2. Economic crisis: Regional economic resilience
  • Content, J. and Frenken, K. (2016) Related variety and economic development: a literature review, Papers in Evolutionary Economic Geography, 16.21
  • Content, J., Frenken, K., Economidou, C. (2016) Indicators and Growth effects of Related Variety at the national and regional level in the EU, D3.1 A review paper from task 3.1 on indicators and growth effects of related variety at the national and regional level in the EU, FIRES project, GA 649378
  • Frenken, K., Van Oort, F., & Verburg, T. (2007). Related Variety, Unrelated Variety and Regional Economic Growth. Regional Studies, 41(5), 685-697. http://dx.doi.org/10.1080/00343400601120296
  • Glaeser, E., Kallal, H., Scheinkman, J., & Shleifer, A. (1992). Growth in Cities. Journal Of Political Economy, 100(6), 1126-1152. http://dx.doi.org/10.1086/261856
  • Hidalgo, C., Klinger, B., Barabasi, A., & Hausmann, R. (2007). The Product Space Conditions the Development of Nations. Science, 317(5837), 482-487. http://dx.doi.org/10.1126/science.1144581
  • Neffke, F. (2011). Regional skill bases as a policy focus. Regional Insights, 2(1), 4-4. http://dx.doi.org/10.1080/20429843.2011.9677975
  • Neffke, F. & Henning, M. (2012). Skill relatedness and firm diversification. Strategic Management Journal, 34(3), 297-316. http://dx.doi.org/10.1002/smj.2014
  • Neffke, F., Henning, M., & Boschma, R. (2011). How Do Regions Diversify over Time? Industry Relatedness and the Development of New Growth Paths in Regions. Economic Geography, 87(3), 237-265. http://dx.doi.org/10.1111/j.1944-8287.2011.01121.x
  • Nooteboom B. (2000) Learning and Innovation in Organizations and Economies. Oxford University Press, Oxford.
  • Tanner, A. (2015). The emergence of new technology-based industries: the case of fuel cells and its technological relatedness to regional knowledge bases. Journal Of Economic Geography, 16(3), 611-635. http://dx.doi.org/10.1093/jeg/lbv011
  • van Oort, F., de Geus, S., & Dogaru, T. (2014). Related Variety and Regional Economic Growth in a Cross-Section of European Urban Regions. European Planning Studies, 23(6), 1110-1127. http://dx.doi.org/10.1080/09654313.2014.905003