Value chain, regional institutions and firm growth in Europe

Abstract

We analyse whether and to what extent the quality of regional institutions has a differential effect on firms’ growth driven by heterogeneity in firm value chain positioning. We analyse turnover growth during the period 2010–2013 for a sample of manufacturing firms located in four European countries—France, Germany, Italy and Spain. We distinguish final firms serving end markets from suppliers serving other firms. Our instrumental variable estimates point to high-quality regional institutions enhancing the growth performance of only locally embedded suppliers with operations confined to the own regional market—that is, the ‘weakest’ node of the value chain.

1. Introduction

The relationship between institutions and firm performance has received great attention in the past years. There is a broad consensus that firms benefit from location in high-quality institutional contexts. The idea of this literature is that public institutions, ensuring efficient juridical systems, contract enforcement, market competition and high-quality public goods provision, may create a favourable business environment characterised by high stability and low transaction costs (e.g. Lasagni et al., 2015; Che et al., 2017; Ganau and Rodríguez-Pose, 2019; Rodríguez-Pose et al., 2021).

Most studies on firm performance have analysed the economic effects of national—that is, country-level—public institutions (e.g., Dollar et al., 2005; Bowen and De Clercq, 2008; Aidis et al., 2012; Dreher and Gassebner, 2013; LiPuma et al., 2013; Dutta and Sobel, 2016; Andrieu et al., 2018). By contrast, there is still scarce evidence on how sub-national regional institutions affect firms’ performance and, particularly, the economic performance of companies operating at different positions of the value chain.

We contribute to this debate by investigating the effect of regional institutional quality on firms’ economic performance by distinguishing companies according to their value chain position. In other words, we distinguish final firms serving end markets from suppliers serving other firms. We employ manufacturing firm-level data for four European countries—France, Germany, Italy and Spain¹—and exploit exogenous cross-regional variations in historical literacy rate to estimate the causal effect of regional institutional quality on firms’ turnover growth over the short-run period 2010–2013.

Our instrumental variable (IV) estimates suggest that location in high-quality regional institutional environments affects the growth performance of only locally embedded suppliers serving other firms in their own region. This implies that only suppliers whose economic performance is almost entirely determined by the productive and commercial relationships with other firms operating in the same region show a significant increase in their short-run turnover growth from improvements in local institutional quality. By contrast, we do not find any effect of regional institutional quality for final firms, as well as for suppliers serving both (non-regional) national and international markets.

Our article has interesting implications not only for the academic literature, but also for policymakers and practitioners working in regional institutions, local governments and regional agencies. First, and more generally, our results suggest that a better regional institutional quality, by improving the competitiveness of local suppliers, may have positive effects on the economic dynamism of the regions where these productive units are localised. Second, and more specifically, local institutions in many European regions provide public goods (e.g. infrastructures) and services (e.g. vocational training and education programmes, financial and fiscal incentives and ‘real’ services such as information on new markets and technologies and assistance to international trade fairs) to support local firms (Brusco et al., 1996). In this way, regional institutions and agencies can facilitate the ‘strategic coupling’ between the strategic needs of final firms, suppliers and customers participating in global value chains (GVCs) and global production networks (GPNs) and the (intangible and tangible) resources available in a region (MacKinnon, 2012). Our main policy implication is that high-quality regional institutions and, therefore, ‘good’ local public policies seem to be effective in terms of economic performance only for local suppliers: that is, for firms that cannot internalise the localisation diseconomies deriving from the lack of high-quality institutions. Regional institutions can compensate for these deficiencies, thus allowing an increase in the economic performance of this type of firm. Vertical relationships among firms operating in different sectors can facilitate the transfer of these benefits to units embedded in GVCs/GPNs, and this identifies a potential additional channel between regional institutions and (global) value chains.

Moreover, these implications are particularly relevant for firm- and local-level economic performance in times of crisis, such as the Great Recession period we cover in our analysis: that is, when high-quality institutions’ intervention can ‘absorb’ the costs of exogenous shocks, thus ‘protecting’ local firms and their regional productive systems. Indeed, as locally embedded suppliers represent the type of firm that may suffer the most from negative (global) shocks, them being the ‘weakest’ node of (global) value chains, high-quality regional institutions may provide such firms with additional support and compensate for their internal deficiencies, thus making them able to cope with adverse external events.

Our analysis is related to different literature streams. The first concerns studies analysing the heterogenous effects of regional institutional quality on the performance of European firms with different internal characteristics—such as size, productivity level and physical and human capital endowment (e.g. Ganau and Rodríguez-Pose, 2019; Agostino et al., 2020b; Rodríguez-Pose et al., 2021). The second stream is related to the role of contextual factors in influencing the performance of firms occupying different positions along value chains. To the best of our knowledge, only Cainelli et al. (2018) have investigated this empirical relationship focusing on differential effects of agglomeration forces. The third stream concerns the literature analysing the relationship between regional institutions and GVCs/GPNs (Coe et al., 2004; MacKinnon, 2012; Coe and Yeung, 2019; Rodríguez-Pose, 2021; Yeung, 2021; Boschma, 2022), and especially the role regional institutions play in affecting firms’ probability to participate in GVCs and GPNs (e.g. Dollar et al., 2016; Accetturo et al., 2017; Ge et al., 2020; Hong et al., 2020). Finally, our article is related to the literature on the relationship between institutional quality and trade (e.g. Barbero et al., 2021), and extends this type of analysis at the firm level by considering explicitly trade in intermediate goods along (local, national and international) value chains.

The rest of the article is structured as follows. Section 2 develops our theoretical framework by discussing the main mechanisms driving a potential heterogeneous effect of institutional quality on firms’ performance related to their value chain positioning. Section 3 outlines the empirical setting. Section 4 presents and discusses the results. Section 5 concludes the work by discussing our key findings and drawing some policy implications.

2. Theoretical background

2.1. Institutions and economic performance

That institutions matter for economic performance is now a well-recognised empirical fact. Indeed, since the seminal work of North (1990), a growing number of studies has shown how differences in institutional quality contribute to explain cross-country differentials in technical efficiency, entrepreneurship and economic growth (e.g. Knack and Keefer, 1995; Hall and Jones, 1999; Acemoglu et al., 2001; Dollar and Kraay, 2003; Acemoglu et al., 2005; Méon and Weill, 2005; Sobel, 2008). The key idea of this literature is that institutions set the ‘rules of the economic game’ and shape the environment where economic transactions occur. For this reason, economic activity is expected to thrive in those socio-economic systems where governments are efficient and institutions are of high quality (Baumol, 1990; North, 1990).

The aggregate effect of ‘good’ institutions on economic performance is generally considered to be positive. Efficient, transparent and un-corrupted institutions, by guaranteeing economic freedom, property rights protection and crime-free market competition, stimulate entrepreneurship and, thus, an efficient allocation and use of the available resources (Baumol, 1990; Sobel, 2015; McCaffrey, 2018). Moreover, high-quality institutions promoting bureaucratic efficiency, as well as stability, trust and reciprocity in the business environment, favour a reduction of transaction costs. This creates the basis for repeated and stable production, commercial, technological and knowledge relationships among firms, leading to higher efficiency and economic growth (North, 2005).

More recent research on the aggregate economic effects of institutions has increased the geographical disaggregation of this kind of analysis moving from the national to the sub-national level. The rationale for this choice is two-fold. First, institutions vary substantially both across countries and within them (e.g. Charron et al., 2014) due to their ‘spatially bounded and historically embedded nature’ (Ganau and Rodríguez-Pose, 2019, 1636). Second, better knowledge of local specificities, and closer geographical proximity to economic actors, put local institutions in a better position than the national one for designing and implementing policies tailored to the local community (Rodríguez-Pose, 2013). Research at the regional level has especially spurred in the context of European countries (e.g. Akçomak and ter Weel, 2009; Tabellini, 2010; Rodríguez-Pose and Di Cataldo, 2015; Crescenzi et al., 2016; Ketterer and Rodríguez-Pose, 2018; Rodríguez-Pose and Ganau, 2022) in light of the increasing emphasis put by European Union (EU) policymakers on improvements in institutional quality as a means for reducing spatial inequality and promoting regional economic growth (e.g. Farole et al., 2011; Charron et al., 2014; Barbero et al., 2021).²

2.2. Regional institutions, firm heterogeneity and GVC participation

Since firms are heterogeneous, differential effects of regional institutions on firm performance can be assumed (e.g. Cainelli et al., 2018; Ganau and Rodríguez-Pose, 2019; Agostino et al., 2020b). Previous research on the EU has shown how smaller, less capital-endowed and less productive firms tend to benefit the most from high-quality institutions (Ganau and Rodríguez-Pose, 2019). Differences in regional institutional quality also help explaining differential effects of credit rationing on productivity levels among firms of different size (Rodríguez-Pose et al., 2021). Despite somehow scarce, some cross-country evidence now exists on how regional institutional quality can have a differential effect on the performance of firms with heterogenous internal characteristics (e.g. Percoco, 2012; Powell and Weber, 2013; Choi et al., 2015; Lasagni et al., 2015; Chakraborty, 2016; Che et al., 2017; Ganau and Rodríguez-Pose, 2018, 2019, 2022; Agostino et al., 2020b; Rodríguez-Pose et al., 2021).

By contrast, heterogeneity related to the relative position a firm occupies along the value chain has been somehow neglected by previous research on the effects of regional institutions (Eckhardt and Poletti, 2018). In fact, existing evidence is limited to the role regional institutions may have in influencing a firm’s probability to participate in GVCs. For example, Accetturo et al. (2017) find that high-quality local judiciary institutions increase Italian manufacturing firms’ probability to enter GVCs. For China, Dollar et al. (2016), Ge et al. (2020) and Hong et al. (2020) find that high-quality local institutions—proxied by government intervention, custom efficiency and contract enforcement—significantly increase manufacturing and services firms’ probability to participate in GVCs. Despite the variety of proxies used for capturing institutional quality, a clear empirical regularity emerges: high-quality institutions increase firms’ GVC participation. However, these studies have focused exclusively on value chain ‘participation’, thus neglecting completely the issue of firm heterogeneity in value chain ‘positioning’. Moreover, they all refer to GVCs, thus ignoring the national and sub-national dimensions of inter-firm relationships, which are, instead, quantitatively relevant.³

Nevertheless, analysing firm heterogeneity in value chain positioning needs to distinguish companies between final firms and suppliers. Final firms—that is, producers serving end markets—operate in the most profitable stages of the value chain. By contrast, suppliers—that is, producers selling to other firms—are, on average, smaller, less productive and mostly confined to local value chains (e.g. Kimura, 2002; Razzolini and Vannoni, 2011; Giunta et al., 2012; Veugelers et al., 2013; Agostino et al., 2020a). Lower trade barriers and transportation costs, together with the spread of information and communication technologies—with the consequent downgrading of the role of face-to-face contacts between suppliers and buyers—have contributed to reduce the contractual power of suppliers (Razzolini and Vannoni, 2011). Indeed, globalisation provides final firms with a much larger pool of potential suppliers to choose among based on cost- and location-seeking criteria. Consequently, local suppliers are increasingly facing a reduction of bargaining power due to growing international competition.⁴ In other words, suppliers tend to suffer from a ‘globalisation discount’ compared with final firms, which makes them the ‘weak’ node of the (international) division of labour (Cainelli et al., 2018). Yet, suppliers constitute the bulk of the productive and industrial system of many European countries—including those we consider in this article, namely France, Germany, Italy and Spain (Agostino et al., 2016).

2.3. Regional institutions, strategic coupling and GPNs/GVCs

High-quality regional institutions can positively affect the participation and, therefore, the economic performance of firms embedded in GPNs and GVCs. Local institutions generally operate not only for ensuring efficient juridical and bureaucratic systems, contract enforcement and market competition, but also for offering high-quality public goods and services—such as infrastructures, vocational training and education programmes, regional development programmes based on financial/fiscal incentives and ‘real’ services related to internationalisation and innovation activities (Brusco et al., 1996). The provision of these public goods and services is aimed at developing and reinforcing the assets of a region, that is, its endowment of (tangible and intangible) resources (MacKinnon, 2012). However, the effects associated with the presence in a region of high-quality institutions—capable of offering these public goods and services—can be heterogeneous depending on the value chain positioning of a firm.

A recent strand of research in economic geography (Coe et al., 2004; MacKinnon, 2012; Coe and Yeung, 2019; Yeung, 2021; Rodríguez-Pose, 2021; Boschma, 2022) has identified an interesting mechanism through which local institutions can exert a positive effect on these processes. This mechanism emphasises the role played by regional institutions in facilitating the ‘strategic coupling’ between the strategic needs of trans-local actors participating in GPNs/GVCs (final firms, suppliers and customers) and regional assets. These matching processes, based on intentional actions and active deliberations by the participants to GPNs and GVCs (MacKinnon, 2012), can occur in two ways (Coe et al., 2004). First, regional institutions can enhance ‘specific’ regional assets that, in turn, impact on: (i) the spatial concentration of knowledge, skills and technologies; (ii) the local variety of high value-added activities (Coe et al., 2004) and (iii) social and industrial relations (Boschma, 2022). Second, regional institutions can promote value-enhancement activities of firms participating in GPNs/GVCs by investing in local assets such as infrastructural endowment, human capital and ‘real’ services (Brusco et al., 1996). In this way, regional institutions can shape, transform and develop the assets of a region (Boschma, 2022) facilitating ‘strategic coupling’ processes, thus enhancing not only the participation and, therefore, the economic performance of those firms participating in GPNs and GVCs, but also the overall connectivity of the regional system with the global economy.⁵

As already mentioned, regional institutions can have heterogeneous effects on firms’ economic performance depending on their position along the value chain. Distinguishing firms according to their position along the value chain allows us to identify specific effects produced by regional institutions on the economic performance of different types of firms. In fact, high international competition and low bargaining power make local suppliers more dependent on the (tangible and intangible) assets of a region compared with both suppliers with operations encompassing a broader geographical scale and final firms. This entails that high-quality regional institutions should exert their major effects on local suppliers. In fact, these firms cannot internalise the localisation diseconomies deriving from the lack of high-quality institutions. Indeed, local suppliers are highly dependent on their own local business and institutional contexts. This is not surprising. For example, Cainelli et al. (2018), accounting for firm-level heterogeneity in GVC positioning, find that a local system ‘context factor’ such as spatial agglomeration produces positive effects only on the economic performance of domestic suppliers. High-quality regional institutions can compensate for these deficiencies of local suppliers, thus allowing a cost reduction and, therefore, an increase in their economic performance. Then, through vertical relationships among firms operating in different sectors, these benefits can be transferred to units embedded in GPNs and GVCs, and this process identifies a potential ‘channel’ through which regional institutions can affect (global) value chains.

By contrast, suppliers whose operations are not confined to the own local market are likely to face less risks due to market diversification, and if involved into GVCs, they tend to enjoy size and productivity premia compared with their local counterparts (e.g. Melitz, 2003; Helpman et al., 2004). Similarly, final firms enjoy a higher bargaining power and, by operating at the high value-added stages of the (global) value chain, tend to accrue most of the value generated through the division of labour. Moreover, and contrary to locally embedded suppliers, these types of firms are generally sufficiently endowed with ‘internal’ resources and tend to be less dependent on the local external environment. These companies may have the capacity to internalise external diseconomies coming from ‘weak’ regional institutions.

Therefore, location in a high-quality institutional environment can provide locally embedded suppliers with a competitive advantage (e.g. in terms of acquired know-how and specificities of demanded goods) against new (international) competitors for maintaining long-lasting production relationships with local buyers. For these reasons, our (theoretical) expectation is that high-quality regional institutions should enhance the economic performance of locally embedded suppliers with operations confined to the own regional market—that is, the ‘weakest’ node of the value chain. Moreover, we expect this general relationship to be even more relevant in times of crisis (e.g. the 2008 Great Recession), that is, when locally embedded suppliers—them being the ‘weakest’ node of value chains—need additional support from high-quality regional institutions to compensate for internal deficiencies and, thus, internalise the costs of exogenous shocks.

3. Empirical framework

3.1. The dataset

The main data source used in our empirical analysis is the EU-EFIGE/Bruegel-UniCredit dataset (EFIGE, henceforth), which provides quantitative and qualitative survey information—collected in 2010, and referring to the period 2007–2009—on ownership structure, employment, investments, innovation, internationalisation, finance and market for manufacturing firms with more than 10 employees operating in seven EU countries, namely Austria, France, Germany, Hungary, Italy, Spain and the United Kingdom (UK). The EFIGE dataset provides also balance sheet data of the surveyed firms drawn from the Amadeus database (Bureau Van Dijk), and referring to the period 2010–2013. The dataset, which is stratified by sector and firm size, covers about 3000 firms from France, Germany, Italy and Spain, about 2000 UK firms and about 500 firms from Austria and Hungary (Altomonte and Aquilante, 2012).

Our aim is to analyse whether the quality of regional institutions has a differential effect on firm-level performance—proxied by turnover growth over the period 2010–2013—related to firms’ heterogeneous positioning along the value chain—that is, whether a firm operates as a supplier of other firms or as a final firm, and whether it serves domestic or international markets. We thus cleaned the EFIGE dataset to maximise the sample size according to our empirical goals. First, we dropped firms without information on region of location, industrial sector and year of incorporation. We consider the geographical level 1 of the EU Nomenclature des Unités Territoriales Statistiques (NUTS) for Germany and the UK, while level 2 for the remaining countries. This choice is driven by the geographical level of aggregation characterising the available data on regional institutional quality. However, the same geographical levels have been considered in previous research on EU regions, having these administrative units the effective institutional powers to influence the economic performance of firms in each specific country (e.g. Ganau and Rodríguez-Pose, 2019; Rodríguez-Pose et al., 2021). Second, we removed firms with missing or incomplete information on produced-to-order goods—it being our key information to distinguish between final firms and suppliers—international activity, innovation and business group membership. Finally, we dropped firms with missing employment data for the year 2010, and turnover data for the years 2010 and 2013. The cleaning procedure left us with a final sample of 6599 firms reporting complete selected survey information, a positive number of employees for the year 2010 and positive turnover data over the period 2010–2013. The cleaned sample covers firms operating in France, Germany, Italy and Spain, while we decided to exclude entirely firms from Austria, Hungary and the UK due to the very limited number of observations left after the cleaning procedure.

Our sample provides a good statistical representativeness of the EFIGE dataset and covers approximatively 58.6% of the firms originally surveyed in the four countries considered (Supplementary Appendix Table A1). The sample covers all NUTS-1 German and NUTS-2 French and Italian regions, while only the Balearic Islands (Spain) are not covered due to no firms left after the cleaning procedure—having excluding à priori the five French Départements d’Outre-Mer, and the Spanish extraterritorial autonomous cities of Ceuta and Melilla (Supplementary Appendix Table A2). The sample includes firms belonging to all size classes covered in the EFIGE dataset, namely small (10–49 employees), medium (50–249 employees) and large (250 and more employees) firms.⁶ Small firms represent 77.8% of the sample, medium firms represent 16.1%, while large firms represent 6.1% and such distribution is reflected in all the four countries analysed (Supplementary Appendix Table A3). Moreover, we cover all manufacturing industries, except for the NACE Rev. 2 industry ‘CD—Manufacture of coke and refined petroleum products’ (Supplementary Appendix Table A4).

By contrast, one potential limitation of our sample concerns its regional coverage: indeed, 16 out of the 75 covered regions include less than 30 firms; 37 out of 75 regions include 30–100 firms; 18 out of 75 regions include 100–300 firms; and only 4 out of 75 regions include more than 300 firms (Supplementary Appendix Table A5). This bias is structural in the EFIGE dataset, as the original sample is not stratified by region. However, despite this potential limitation, the EFIGE dataset has been employed in the context of region-level analyses by previous studies (e.g. Cainelli et al., 2018; Agostino et al., 2020b). We are conscious that our analysis could suffer from a bias related to the under-/over-representation of firms by region, and, therefore, we will test the robustness of our results against this potential bias.

3.2. Defining firm positioning along the value chain

Following previous empirical works based on the EFIGE dataset (e.g. Veugelers et al., 2013; Accetturo and Giunta, 2016; Agostino et al., 2016; Cainelli et al., 2018), we identify a firm’s value chain positioning based on the available survey information on sales of produced-to-order goods.

We consider produced-to-order goods as the best available instrument to proxy for the highly targeted, vertical market relationships existing along value chains between suppliers and buyer firms. Specifically, firms were asked to indicate the average percentage of turnover made up by sales of produced-to-order goods, as well as the main customers of these goods. Based on this information, we classify a firm as final if it serves exclusively end markets—that is, it does not sell produced-to-order goods to other firms; by contrast, we classify a firm as a supplier if it sells produced-to-order goods to other firms. It is worth noting that 65.5% of suppliers are ‘purely suppliers’, that is, firms whose turnover is entirely made up by sales of produced-to-order goods (Supplementary Appendix Table A6). We will explore empirically this source of heterogeneity later in the article.

Furthermore, survey information allows us to identify whether firms are mainly involved into national or international (i.e., global) value chains. We classify a final firm as domestic if it does not serve foreign markets at all, while as international if it exports at least a part of its production to serve foreign end markets. We classify a supplier as domestic if produced-to-order goods are sold to firms operating in the same country, while as international if produced-to-order goods are sold also to foreign firms.

Suppliers represent 80.6% of the sample, as well as most firms in each single country—from 67.7% in Spain to 90.6% in France. Suppliers involved into GVCs represent most firms in all countries but Spain, where domestic suppliers represent 42.7% and international ones 25.0%. By contrast, international final firms are the majority in Spain and Italy, while German and French final firms are mostly involved into domestic value chains (Supplementary Appendix Table A7).

3.3. Empirical modelling

We analyse the differential effect played by value chain position heterogeneity in the relationship between regional institutional quality and firms’ turnover growth over the short-run period 2010–2013 by estimating the following Gibrat (1931)-type growth equation: where the dependent variable is defined as follows: and denotes the average yearly turnover growth of firm $i$⁠, operating in industry $s$ and located in region $r$ in country $c$ over the period 2010–2013, with $T$ denoting the time-span of the growth period.⁷ The right-hand side of Equation (1) includes the three variables of interest, namely: (i) the variable for regional institutional quality (⁠$Institutionsrc$⁠); (ii) the dummy variable capturing a firm’s positioning along the value chain (⁠$Supplierisrc$⁠), which takes a value of 1 if a firm is classified as a supplier, and a value of 0 if it is classified as a final firm; and (iii) the interaction term between the variables for regional institutional quality and firms’ supplier status, that allows us to assess whether a differential effect exists in the economic returns of regional institutional quality for suppliers versus final firms.

The term $Institutionsrc$ captures the level of institutional quality of region $r$ in country $c$ in 2009. We employ regional institutional quality data drawn from the European Quality of Government Index (EQGI) dataset, elaborated by the Quality of Government Institute at the University of Gothenburg. The dataset provides information derived from a survey conducted in 2010 on 34,000 citizens living in 172 EU regions. It focuses on individuals’ perception and experience with corruption, quality and impartiality of governance in their own region with respect to education, public health care and law enforcement (Charron et al., 2013, 2014).

We have computed the regional institutional quality variable in three steps. First, we have aggregated individual survey questions into three main region-specific institutional pillars capturing the dimensions of control of corruption, quality of local governance and impartiality of the local government in providing public goods and services. Second, we have standardised the three institutional indices to have zero mean and unitary standard deviation, and we have then averaged them to obtain a region-specific measure of institutional quality. Finally, following previous works (e.g. Ganau and Rodríguez-Pose, 2019; Cainelli et al., 2022), we have relied on a normalisation in the interval $0, 1$ to obtain the variable included in the regression model, such that the quality of the regional institutional environment increases as the index moves from 0 to 1. It is worth noting that the institutional variable proxies for the ‘quality’—rather than the ‘quantity’—of regional institutions, as it captures the capacity of regional governments to provide and administer public goods and services impartially, effectively and in a non-corrupt manner (Rothstein and Teorell, 2008; Charron et al., 2014).

Supplementary Appendix Figure A1 maps the spatial distribution of the institutional quality variable in the 75 regions covered in the sample, and suggests a certain degree of both cross-region and cross-country heterogeneity. A clear North–South divide characterises Italy, with Southern regions showing a very low level of institutional quality. By contrast, regions with high and low levels of institutional quality coexist in France and Spain without a clear spatial pattern. Germany shows a rather homogenous internal spatial structure, with almost all its regions showing high levels of institutional quality.

The right-hand side of Equation (1) includes also the vector $Xisrck$ of firm-specific control variables. We control for a series of factors that can affect a firm’s growth performance: (i) the log-transformed growth-initial turnover variable, to capture convergence in the growth process; (ii) a set of size dummy variables for small, medium and large firms; (iii) a measure of firm age, defined as the log-transformed difference between 2010 and the year of a firm’s incorporation; (iv) a set of dummy variables capturing the introduction of product innovations only, process innovations only or both product and process innovations over the period 2007–2009; (v) a dummy variable capturing whether a firm is member of a national business group—that is, a business group consisting of firms belonging all to the same country; (vi) a dummy variable capturing whether a firm is member of an international business group—that is, a business group consisting of firms belonging to different countries; and (vii) a dummy variable capturing whether a firm is an exporter. All these variables can affect the growth performance of a firm. Larger firms have more (physical, labour, financial) assets to leverage on for increasing their competitiveness and, thus, raising market shares and grow faster. Older firms may exploit competitive advantages related to learning mechanisms, accumulated experience and better knowledge of the market. Firms introducing process and, particularly, product innovations can translate this higher innovativeness into higher competitiveness and larger market shares, and, consequently, greater growth rates. Membership to a business group allows a firm to exploit the so-called ‘internal capital market’ of the group, with gains in terms of easier access to tangible (e.g. financial and human capital resources) and intangible (e.g. information about competitors and markets) assets contributing to grow more (Cainelli et al., 2022). Finally, exporters are traditionally found to be more efficient than purely domestic firms, and they are expected to grow more especially if they can diversify the risk between the domestic market and various foreign markets.

We have enriched Equation (1) also with a set of log-transformed region-specific variables (⁠$Xrcp$⁠), namely: (i) gross domestic product (GDP) per capita in 2010 (defined in million Euros, in Purchasing Power Standards) to capture the overall economic development level of a region; (ii) average yearly GDP per capita growth over the period 2010–2013 to capture region-specific demand shocks; (iii) population density, defined as population in 2010 per square kilometre, to proxy for agglomeration-related forces; (iv) human capital endowment, defined as the percentage of the population aged 15–64 years with tertiary education in 2010; and (v) GDP per capita in 1900 (defined in millions of 1990 International Dollars) to control for historical differentials in economic development level across regions, which may have affected subsequent development, urbanisation and education levels.⁸ Finally, Equation (1) includes the terms $μs$ and $νc$ denoting industry and country fixed effects, respectively, and the error term (⁠$εisrc$⁠).⁹

3.4. Identification strategy

We first estimate Equation (1) via Ordinary Least Squares (OLS). However, such estimates are potentially biased due to endogeneity of the regional institutional quality variable. First, shocks occurring at the regional level are likely to influence both the overall quality of the local government and the economic performance of local firms. Second, spatial sorting can be a problem if the most efficient firms tend to locate in—or re-locate towards—regions characterised by a better institutional environment. Third, it is not easy to measure the quality of regional institutions, and the institutional quality variable we employ is only a proxy for a highly complex phenomenon that is hard to capture through survey data. Therefore, we rely on an IV approach and estimate Equation (1) through a Two-Stage Least Squares (TSLS) estimator.

The identification strategy exploits exogenous regional variations in the 1870s literacy rate to identify the causal effect of regional institutional quality on firms’ growth performance.¹⁰ The validity of the proposed IV relies on the fact that historical educational levels are highly correlated with subsequent changes in institutional and political setting (e.g. Glaeser et al., 2004; Akçomak and ter Weel, 2009; Tabellini, 2010). Moreover, we expect our IV to be exogenous with respect to firms’ performance, as literacy rate in the 1870s represents a historical phenomenon which hardly could affect the current performance of individual firms.¹¹

We treat as endogenous both the variable for regional institutional quality, and the interaction term between the variables for regional institutional quality and firms’ supplier status, and we instrument the endogenous interaction term with the interaction between the excluded IV for historical literacy rate and the variable for firms’ supplier status.

We define the variable capturing historical literacy rate as the percentage of the population able to read and write. French data are drawn from Tabellini (2010) and refer to the percentage of literate population aged 6 years or more in 1872. German data are drawn from Cipolla (1969) and refer to the percentage of illiterate population aged 10 years or more in 1871.¹² Italian data are drawn from Flora (1983) and refer to the percentage of literate population (able to read only) aged 5 years or more in 1871. Finally, Spanish data are drawn from Núñez (1990) and refer to the percentage of literate population aged 10 years or more in 1877.

4. Empirical results

4.1. Main results

We first estimate a reduced version of Equation (1) by omitting the interaction term between the variables for regional institutional quality and firms’ supplier status. This exercise provides us with baseline results to check the overall validity of our empirical model, compare our estimates with previous studies on the link between regional institutional quality and firm performance across EU regions, and test whether suppliers—generally regarded as the ‘weak’ node of value chains—are in a worse position compared with final firms. We then estimate Equation (1) separately for final firms and suppliers to have preliminary evidences of potential differences in the turnover growth returns of regional institutional quality between these two types of firms. Finally, we estimate the full version of Equation (1), which we consider our main specification.

The OLS and TSLS results of these exercises are reported in Table 1. As shown in the bottom part of the table, we find evidence that our IV capturing regional literacy rate in the 1870s has a good predictive power, as the first-stage F-statistic on the excluded IV is higher than the conservative cut-off value of 10 in all specifications.¹³

First, our baseline estimates confirm previous evidence as we find that regional institutional quality enhances firms’ growth performance. Considering Column (2), we estimate that a 1 percentage point increase in regional institutional quality leads to an increase in firm-level turnover growth of about 0.17 percentage points. Moreover, we do not find evidence of a turnover growth premium for final firms compared with suppliers, as the coefficient of the dummy variable capturing a firm’ supplier status is zero in magnitude.

Second, comparison of Columns (4) and (6) suggests a differential effect of regional institutional quality on the performance of final firms with respect to suppliers. We find that only firms operating as suppliers along the value chain benefit from high-quality regional institutions. We estimate that a 1 percentage point increase in regional institutional quality leads to an increase in suppliers’ turnover growth of 0.21 percentage points; we also find that the growth returns of regional institutional quality are about 4.1 times larger for suppliers than for final firms. Moreover, the difference in the estimated coefficients for the two types of firms is statistically significant (p-value equal to 0.000).¹⁴ These results confirm previous evidence suggesting how specificities of the local system where firms operate matter for suppliers only, as in the case of agglomeration economies and knowledge spillover effects (Cainelli et al., 2018).

This last result is confirmed when estimating the full version of Equation (1) based on an interaction strategy, through which we can exploit all observational units in our sample, especially in the light of a potential small sample size bias. The interaction term between the variables for regional institutional quality and firms’ supplier status is positive and statistically significant. From Column (8), we estimate that only suppliers benefit from improvements in regional institutional quality and that the growth returns of regional institutional quality are about 18.5 times larger for suppliers than for final firms.

Our results suggest a relatively large gain for suppliers, that is, an increase in short-run turnover growth of about 20.3%. Thus, improving regional institutional quality emerges as a key strategy for pushing firms’ economic performance, and this seems to be the case particularly for those regions where suppliers represent the bulk of the local productive system. It follows that any ad hoc regional policy and support strategy providing local suppliers with (tangible and intangible) assets targeted at increasing their competitiveness and efficiency (e.g. financial and fiscal incentives, training and education programmes, and support for internationalisation and innovation) would foster the economic performance not only of individual (targeted) suppliers, but also of the entire (regional, national and international) system through spillover effects and (global) ‘connectivity’ processes occurring along (regional, national and international) value chains.

4.2. Robustness analysis

We have performed several exercises to test the robustness of the main results presented in Table 1. These robustness tests are discussed in detail in Supplementary Appendix C, where we also present all the results of these analyses.

First, we have assessed the sensitivity of our analysis against potential biases related to sample selection, regional coverage and small sample size. Indeed, our sample includes a relatively small number of firms—all from Western EU countries—compared with previous firm-level studies on the effects of (regional) institutional quality. Moreover, and most importantly, our sample suffers from potential biases related to the under-/over-representation of firms by region. Second, we have assessed the turnover growth effects of changes—rather than levels—of regional institutional quality. Third, we have relied on alternative estimation and identification strategies to test the robustness of our IV estimates. Fourth, we have tested for the generalisation of our results—based on turnover growth over the crisis period 2010–2013—by considering both value added and labour productivity growth as alternative dependent variables, and turnover growth over a slightly longer time span (i.e. the period 2010–2015). Finally, we have accounted explicitly for firm size heterogeneity to both isolate the ‘true’ supplier effect from any small size-related effect—as 78.5% of suppliers are of small size—and assess potential threshold employment levels beyond which the estimated average positive effect of regional institutional quality on suppliers’ economic performance could vanish.

Our main results are fully confirmed. Indeed, we estimate that only suppliers benefit from improvements in regional institutional quality.

4.3. Heterogeneity between domestic and international destination market

We explore now heterogeneity related to the domestic and international destination market of final firms and suppliers. To this aim, we modify Equation (1) as follows: first, we replace the dummy variable capturing firms’ supplier status with a categorical variable which takes a value of 0 for final firms serving only the domestic market; a value of 1 for final firms serving also foreign markets; a value of 2 for suppliers serving other firms only within the national boundaries; a value of 3 for suppliers serving also foreign firms; second, we interact the categorical variable capturing simultaneously the value chain position of a firm and its destination market with the variable for regional institutional quality. For the sake of completeness, we also estimate this modified version of Equation (1) for each type of firm separately.

The TSLS results of this exercise are reported in Table 2 and suggest two interesting facts. First, as expected, we find evidence of a turnover growth premium for internationalised firms compared with domestic ones—see Column (1). Second, the results point to positive and statistically significant returns of regional institutional quality on the turnover growth of domestic suppliers only—that is, suppliers serving only other national firms. By contrast, we do not find evidence of (domestic and international) final firms and international suppliers benefitting from high-quality institutions. This result is confirmed both when relying on the interaction-based strategy—Column (2)—and when estimating Equation (1) separately for each firm type.

4.4. Heterogeneity in suppliers’ domestic destination market

We now focus on heterogeneity concerning the destination market of domestic suppliers. In fact, EFIGE survey information allows us to identify whether domestic suppliers serve through produced-to-order goods mainly firms located in their own region, firms located only in other regions of the own country or both types of firms. We thus modify Equation (1) by including a categorical variable that captures the main national destination market of a supplier and by interacting it with the variable for regional institutional quality. In this case, we estimate Equation (1) on the sub-sample of domestic suppliers, and, as before, we also provide evidence by splitting the sub-sample of domestic suppliers according to the three typologies of national destination markets.

The TSLS results of this exercise are reported in Table 3, and point to a positive and statistically significant effect of regional institutional quality for the growth performance of only suppliers serving firms operating in their own region.

This result adds a further relevant ingredient to the policy debate on the role local institutions may play for enhancing firm performance and, consequently, aggregate (regional) economic growth. The fact that only the fraction of suppliers serving other firms in the own regional productive system benefits from improvements in local institutional quality should not be read as diminishing the role of local institutions and the importance of enhancing the quality and efficiency of regional governments. On the contrary, this evidence suggests two interesting insights. First, only the ‘weakest’ firms operating along the value chain are effectively influenced by the own regional environment and institutions. In fact, when regional institutions shape and transform the assets of a region (Boschma, 2022), these processes seem to affect only the economic performance of suppliers. An obvious consequence of this finding is that those regions where the bulk of firms operates within local value chains would be the territories gaining the most from improvements in the quality of local institutions. Second, locally embedded suppliers mostly involved into regional value chains are a widespread phenomenon across European countries, especially manufacturing ones such as Italy and Spain (e.g. Cainelli et al., 2018). With specific reference to our sample, we find this type of firm in 71 out of 75 regions, the only exceptions being the German regions of Bremen and Mecklenburg-Vorpommern, and the Italian regions of Aosta Valley and Liguria. However, these four regions are low manufacturing ones, as the contribution of the manufacturing industry to total regional gross value added ranges from 6.4% in Aosta Valley to 20.3% in Bremen as average value over the period 2010–2018.¹⁵ This implies that improving the quality of regional institutions remains a key factor for pushing firms’ economic performance and the overall growth capacity of EU territories, especially those industrial regions specialised in traditional and low value-added productions, and which are poorly connected internationally through GVCs.

4.5. Heterogeneity across suppliers with different degree of involvement into value chains

As previously highlighted, our sub-sample of suppliers consists of firms characterised by different percentages of turnover made up by sales of produced-to-order goods. Despite firms whose turnover is made up entirely by sales of produced-to-order goods represent 65.5% of the supplier category, the remaining 34.5% is characterised by a percentage ranging from 1% to 99%. We now exploit this source of heterogeneity across suppliers to assess whether the growth returns of regional institutional quality differ for firms with a different degree of involvement into value chains.

We consider only the sub-sample of suppliers, and modify Equation (1) by: (i) replacing the dummy variable capturing firms’ supplier status with a variable capturing a supplier’s degree of involvement into value chains (⁠$Degree supplierisrc$⁠) measured as the percentage of turnover made up by sales of produced-to-order goods ranging in the interval $1, 100$⁠; and (ii) including the interaction term between the regional institutional quality variable and the variable $Degree supplierisrc$ to explore heterogeneity in the growth returns of regional institutional quality across suppliers. We estimate this modified version of Equation (1) via TSLS on the sub-samples of international suppliers and domestic suppliers, also accounting for heterogeneity related to domestic destination markets.

At a first glance, the results of this exercise—which are reported in Table 4—suggest that suppliers’ degree of involvement into value chains does not matter in the way regional institutional quality affects their economic performance: indeed, the coefficients of the interaction term are negligible from a statistical viewpoint.

However, a closer inspection at the results provides more insights. As shown in Figure 1—which plots the estimated marginal effects of regional institutional quality on suppliers’ turnover growth evaluated at the different percentages of turnover made up by sales of produced-to-order goods—two key evidences emerge. First, we confirm the results previously presented in Tables 2 and 3: improvements in regional institutional quality enhance the growth performance of only domestic suppliers serving other firms in their own region. Second, we find that only locally embedded suppliers for which sales of produced-to-order goods account for approximatively 75% of turnover benefit from improvements in regional institutional quality. We estimate a statistically significant increase in turnover growth of 0.20 percentage points due to a 1 percentage point increase in regional institutional quality for firms with a 75% threshold value of turnover made up by sales of produced-to-order goods, while the gain from improved regional institutional quality raises up to 0.35 percentage points when considering locally embedded suppliers whose turnover is entirely made up by sales of produced-to-order goods.

These results reinforce the evidence that only locally embedded suppliers whose performance is almost entirely dependent on the economic decisions and activities of neighbouring firms, and, thus, representing the ‘weakest’ node of the value chain, gain substantially from improvements in local institutional settings.

5. Conclusions

This article has provided a novel and original contribution to the existing literature and ongoing policy debate on the role regional institutions can play as a growth-enhancing factor. We have contributed to four different literature streams. First, the literature analysing the nexus between the quality of regional institutions and firm-level performance. Second, the literature on firm heterogeneity related to GVC participation and positioning. Third, the literature on the role of regional institutions in explaining this firm-level heterogeneity. Fourth, the literature on the relationship between institutional quality and trade. Specifically, we have provided for the very first time new cross-country evidence from Western EU countries on the extent to which improvements in regional institutional quality can foster the growth performance of firms that are heterogeneous with respect to their positioning along the value chain.

Our results suggest that high-quality regional institutions have positive effects on the economic performance of only locally embedded suppliers that are almost entirely specialised in providing highly targeted produced-to-order goods to other firms located in their own region. By contrast, we find statistically negligible effects of regional institutional quality for final firms, as well as for suppliers serving both (non-regional) national and international markets.

While our analysis confirm previous studies finding a general positive relationship between high-quality regional institutions and firm-level performance in a single or a cross-section of EU countries (e.g. Lasagni et al., 2015; Ganau and Rodríguez-Pose, 2019; Rodríguez-Pose et al., 2021), as well as previous evidence on the different ways final firms and suppliers interact with the local environment—for example, in terms of agglomeration forces (Cainelli et al., 2018)—, it adds novel insights on the specific type of firm gaining the most from improvements in the quality of regional institutions in the era of international production fragmentation. We thus contribute to a relatively scarce cross-country literature studying the economic returns of regional institutions on heterogenous firms by going beyond firms’ characteristics such as size, capital endowment and productivity levels (e.g. Ganau and Rodríguez-Pose, 2019; Agostino et al., 2020b).

The fact that the quality of a firm’s regional institutional environment is a relevant growth-enhancing factor for only suppliers with operations confined to the local market is not surprising. This type of firm is locally embedded, such that its business opportunities and economic performance depend largely on the institutional and contextual conditions of the region in which it operates. Therefore, any improvement in the quality of regional institutions is likely to favour its growth performance through mechanisms such as higher bureaucratic transparency, higher government efficiency, lower corruption, fairer competition and the provision of high-quality public goods and services.

Suppliers embedded in the regional productive system, and especially those almost entirely depending on the demand of neighbouring local firms, represent the ‘weakest’ node of the value chain, as they have little bargaining power and are also highly vulnerable to both local shocks and shocks affecting them indirectly via (local) production networks. Therefore, these firms can thrive only if the institutional setting characterising their local system performs efficiently enough to guarantee them certainty and stability, protection against exogenous shocks and support strategies through the provision of public goods and services.

By contrast, final firms and international suppliers, as well as suppliers not exclusively embedded in the local market, are less dependent on the quality of their own regional institutional environment. A possible explanation for this evidence is that these firms are sufficiently endowed with ‘internal’ resources and, therefore, have the ‘capacity’ to internalise the external diseconomies arising from low-quality local institutions. On the one hand, and on average, they are larger and belong to a business group in a higher proportion than locally embedded domestic suppliers, thus having more resources—either internal to the firm or available through the ‘internal capital market’ of the business group—to support the costs of institutional failures (Cainelli et al., 2022).¹⁶ On the other hand, their operations are not confined to the own local market. This has two potential implications. First, these types of firms can exploit advantages related to risk diversification, especially if serving end markets and/or other firms geographically differentiated. Second, we can expect their performance to depend more on the conditions—among which, the quality of the institutional setting—of the locations where they sell.

But how can our results inform policymakers? First, locally embedded domestic suppliers represent a significant share of the firms operating in many European manufacturing systems. The lack of market diversification in general, and international operations in particular, signals a ‘productivity issue’ that can be exacerbated by the low level of regional institutional quality. This is an important result in the globalisation era, because the relative low quality of local institutions, combined with the small size and the high idiosyncratic risks of locally embedded domestic suppliers, could increase their probability of being crowded out and exiting the market. This implies also that improving the quality of regional institutions represents a key strategy not only for shaping the performance of individual firms, but also for promoting the overall economic dynamism and success of those regions which are poorly involved into GPNs and GVCs, and where production is still confined to low value-added activities.

There is substantial room for action by European and national policymakers, especially in Western EU countries—such as Italy, Greece or Spain—and Eastern ones—such as Bulgaria, Poland or Romania—where two problems coexist, namely the low quality of local institutions (Rodríguez-Pose and Ganau, 2022) and little economic dynamism ascribable to both a lack of international openness and an excessive specialisation in labour-intensive, low value-added industries (Ganau and Kilroy, 2023). Therefore, improving regional institutional quality becomes crucial when taking policy measures to improve the competitiveness of local suppliers, such as place-based policies to lower the sunk costs of internationalisation, subsidy programmes to increase the internationalisation capacity of small locally embedded firms or ‘strategic coupling’ strategies to develop and reinforce the assets of a region.

Our article has some limitations. First, our empirical analysis has been conducted on a small number of Western European countries—that is, France, Germany, Italy and Spain. Yet, these countries belong to the core of the EU in terms of economic development level and quality of institutions, even if two countries—Italy and Spain—are characterised by strong internal disparities along these two dimensions. It would be interesting to extend our analysis to other European countries, such as post-2004 EU Member States where the quality of institutions is generally recognised as very low. Second, the EFIGE dataset, despite being rich in firm-level information, consists of a relatively small number of firms, and this could lead to potential biases related to small sample size and region-level statistical representativeness. However, as shown in Supplementary Appendix C, our main empirical findings are robust to such potential biases. Third, our analysis focuses on the period 2010–2013, which could be potentially problematic, it being the aftermath of the 2008 Great Recession, and, therefore, could limit the generalisation of our results to non-crisis periods. However, despite the four countries we consider have been hit heterogeneously and with different degrees of severity by the crisis (e.g. in terms of sovereign debt crisis), they all experienced a similar GDP path around the crisis period: they all recorded a negative GDP growth rate between the second and the third quarter of 2008 after a relatively long period of positive growth rates, and they all started to recover—recording a positive GDP growth rate—between the fourth quarter of 2009 and the first quarter of 2010 after the crisis period.¹⁷ In addition, as shown in Supplementary Appendix C, our results are robust to the exclusion of individual countries from the estimation sample, as well as when considering a relatively longer turnover growth time span (i.e. the period 2010–2015). Anyway, future research should overcome this limitation by considering firms’ economic performance over a considerably longer period away from crisis shocks. Finally, our investigation does not consider potential spatial effects among regions. This does not generate particular problems when we consider the quality of regional institutions, since the majority of regional policies and interventions are concentrated within the boundaries of a region and target local firms. By contrast, spatial spillovers can occur when other variables are accounted for—for example, the productivity level of firms located in neighbouring regions that can affect the performance of firms located in a reference region. There is no doubt that these phenomena would be very interesting to be analysed; this may be subject for future developments of this research line, provided that the (relative) lack of good-quality micro-level data will be overcome in the near future.

Footnotes

Supplementary material

Supplementary data for this paper are available at Journal of Economic Geography online.

Acknowledgments

We are grateful to Simona Iammarino, the editor in charge, and to the three anonymous reviewers for their stimulating suggestions for improvements of previous drafts. We would also like to thank participants at the “European Trade Policy and Global Value Chain” International Workshop (Roma Tre University, June 2022), the SASE Conference (University of Amsterdam, July 2022) and the “Local Economic Development: Opportunities and Constraint” Conference (University of Campania Luigi Vanvitelli, October 2022). This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. All errors are our own.

Conflict of interest statement

None of the authors of the manuscript have any affiliations with or involvement in any organization or entity with any financial or non-financial interest in the subject matter or materials discussed in this manuscript.

Data availability

The data that support the findings of this study are available from Bruegel (EFIGE dataset) and Bureau van Dijk (Amadeus database). Restrictions apply to the availability of these data, which were used under license for this study. Data are available from the authors upon request and with the permission of Bruegel and Bureau van Dijk.

References