VoxEU Column Competition Policy Industrial organisation Productivity and Innovation

Impact evaluation of cluster development programmes

Cluster development programmes (CDPs) aim to support industrial clusters of agglomerated firms to achieve higher productivity and sustainable development. Such programmes have been prominent in Latin America over the past decade, but there have been few impact evaluations. This column presents the findings from an evaluation of Latin American CDPs. Various case studies show positive medium-term effects of the programmes on employment, exports, and wages. CDPs are also found to have positive spillover effects on untreated firms, and to improve the network connectivity and technology-transfer ties between firms.

Several studies have shown that the agglomeration of firms in specific locations improves productivity and growth (Becattini 1989, Chatterji et al. 2013, Krugman 1991, Marshall 1890). Following this idea, many governments have increasingly developed and implemented programmes to support industrial clusters to achieve higher productivity, increased and better quality jobs, and sustainable development (Casaburi et al. 2014). These programmes have become part of the package of modern industrial policies, where public-private collaboration and coordination is essential in identifying missing inputs and public policies for economic development (Hausmann and Rodrik 2003).

Do such cluster development programmes (CDPs) work? Do they fundamentally encourage essential inter-firm networking and coordination? Does this lead to enterprise development, more employment, and export growth? What is the size of the impact and how long does it take for the programmes to produce an impact? Do other firms also benefit from these programmes? And finally, how do we measure all this? Over the last decade, throughout many countries in Latin America, CDPs have been pursued to help exploit the externalities emerging from geographical agglomeration and inter-firm coordination, in order to foster the development of industries in specific localities (Casaburi et al. 2014, Rodríguez-Clare 2007). Yet the impact evaluations of these policies have been very few and far between.

Our recent book, The impact evaluation of cluster development programs: Methods and practices (Maffioli et al. 2016), introduces new evidence and insights to answer these questions, taking stock of methodological and empirical research developed over the last decade. We contend that CDPs involve diverse and multi-dimensional processes that require a variety of instruments to be fully understood and assessed. Moreover, we reaffirm the idea that policy evaluation is crucial in ensuring the best use of public resources, the accountability of policymakers, and, most of all, in feeding the necessary learning to improve the design and implementation of public policies.

Direct effects

The book discusses the conceptual foundation of rigorous evaluation techniques, and reports on the main findings of applications to various examples in Latin America. One of the studies uses firm-level administrative data from 2002 to 2009 to assess the impact of cluster development policies on employment and exports in the Brazilian states of Minas Gerais and São Paulo. Using a combination of matching and differences-in-differences techniques, the study finds positive direct average effects on employment, total export values, and the likelihood of exporting. Interestingly, impacts grow years after the implementation of the policy, showing that these kinds of policies need time to produce effects. Similarly, another study uses panel data techniques to estimate the effect of a CDP designed to promote ICT in Cordoba, Argentina. The study finds significant effects on the performance of enterprises (sales, employment, wages, and the probability of exporting).

Indirect effects: Spillovers

These programmes are also of interest because of the indirect effect they may have through spillover effects on firms not directly participating in the programme (‘treated’ by the policy tool), but that are related to the firms that did take part in the programme. This may occur either because these firms were located in the same geographical area or because they employed workers previously employed by the firms in the programme, thereby carrying with them their ‘embodied’ competences and knowledge. In other words, the overall impact of these programmes appears to be a multiple of the direct impact on the firms taking part in the programme via the spillover effects generated. Indeed, in both cases studied in Argentina and Brazil, these indirect effects were substantial – an increase in the sales of non-participant firms by 0.9% in the case of Argentina, and an increase in the value of total exports and the likelihood of exporting in the case of Brazil. Moreover, these effects were sustained over time and were detectable up to six years after the programme inception.

Networks and enterprise performance

The book also discusses a new and important element for the evaluation of CDPs – social network analysis (SNA). CDPs explicitly aim at strengthening enterprise networks as a way to foster their productivity and growth. But do they actually succeed in reaching this target? Are stronger networks responsible for better enterprise performance? The merit of social network analysis is that it allows the researcher and policymaker to portray the shape of the network and to highlight the role played by each actor in it, its relationships with the others, the distribution of power, and the strategic relevance of each link in the network.

The book discusses the potential application of this technique to impact evaluations, and shows how the cluster development programme in Córdoba, Argentina modified the nature of the network before and after the programme (2005 and 2012). The network became more centralised, with fewer selected firms becoming more central over time, while others became progressively more peripheral or isolated. Firms economised on the number of relationships they formed by selecting only partners believed to provide tangible benefits. The authors define such central firms as dominant players, visionary and motivated entrepreneurs who invested time and resources in network-enhancing initiatives, and in avoiding disrupting the network over time. These firms were vital in guaranteeing network connectivity and creating the link between treated and untreated firms. The programme also led to strengthening and creating new technology-transfer ties between the electronics firms in Córdoba and other local, provincial, and national institutions.

The details of the policy process

None-the-less, although solid quantitative impact evaluations are necessary for all the reasons mentioned above, they may fall short in giving the full picture of a policy process and its implementation. What institution should be in charge of delivering the programme? Should the programme be directed to all firms in an area, to all sectors, or only to a selection of them? What activities should be supported and subsidised and to what extent? Rigorous case studies can be of great help to answer these questions that are part of the daily activity of a policymaker. The case studies reported in the book study CDPs carried out in Argentina, Brazil, Chile, and Uruguay and offer remarkable insights. One of the notable lessons learned is that public and private actors need to be prepared and trained prior to these innovative interventions, and that such preparation is as relevant as the project design itself. Moreover, the studies reveal that these programmes need to be coordinated with other programs and policies to achieve greater impact. This means that cluster selection requires careful evaluation, not only of the present and future comparative advantage of an area, but also of the local institutional capacity to coordinate actions among private firms and the public sector to design and carry out the details of the program. Finally, CDPs are more likely to be successful when other ingredients of the policy pie exist (e.g. financing, support for human capital and for research and development, export promotion).


Impact evaluations of modern industrial policies have received little attention, despite the great benefits they can have on policy design and implementation. CDPs are no exception – despite their diffusion in developed and developing countries, rigorous evaluations are quite scarce. Our book seeks to make a two-fold contribution. First, it shows how newly developed techniques can help to overcome the methodological challenges that often impede the production of rigorous evaluation in this area. Second, it demonstrates how rigorous evaluations of CDPs in Latin America have produced overwhelming evidence on the (direct and indirect) effects of this instrument on the competitive performance of enterprises. In doing so, the book also provides new evidence on the positive externalities achieved through these programs, reinforcing the basis of their implementation.


Becattini G (1989) “Sectors and/or districts: Some remarks on the conceptual foundations of industrial economics”, in E Goodman and J Bamford, Small firms and industrial districts in Italy, London: Routledge.

Casaburi, G, A Maffioli and C Pietrobelli (2014) “More than the sum of its parts: Cluster-based policies“, in G Crespi, E Fernández-Arias and E Stein (eds) Rethinking productive development: Sound policies and institutions for economic transformation, London: Palgrave.

Chatterji, A, E Glaeser  and W Kerr (2013) “Clusters of entrepreneurship and innovation”, NBER, Working Paper No 19013. Cambridge, MA.

Hausmann, R and D Rodrik (2003) “Economic development as self-discovery”, Journal of Development Economics, 72(2): 603–33.

Krugman, P (1991) Geography and trade, Cambridge, MA: MIT Press.

Maffioli A, C Pietrobelli and R Stucchi (2016) The impact evaluation of cluster development programs: Methods and practices, Washington DC: Inter-American Development Bank.

Marshall, A (1890) Principles of economics, London and New York: Macmillan.

Rodríguez-Clare, A (2007) “Clusters and comparative advantage: Implications for industrial policy”, Journal of Development Economics, 82: 43–57.

1,050 Reads