Clusters: Sexy but Mysterious and Elusive


Richard Dreyfuss looks out the window of his ascending airplane and sees the girl of his dreams, the girl he had spent a weekend chasing and never finding, driving off into the proverbial sunrise-never ever to be seen again. So ends American Graffiti. The frustration and disappointment which ends that story is the beginning of ours: is cluster based economic development the latest and greatest strategy to revitalize our tired and troubled communities? To suggest a response to this question we shall briefly summarize the “what” of cluster-based economic development and offer a few comments as to its value and promise as a key tool for local economic growth and revitalization?

It is my perception that cluster-based economic development is one of the very hottest, sexy concepts found in current local economic development. It certainly has been embraced by well-connected research institutes and think tanks and even a cursory review of academic research reveals literally a hundred different articles, books and studies on some facet of clustering. My own personal research on economic development organizations has revealed a small, but growing, use of something akin to cluster-based strategy. Intensifying this trend, is the relatively recent support from the Obama Administration which seems to have embraced the concept and while not smothering it in money, have certainly lent it credibility and a measure of sexiness (or at least a hopeful source of potential funds).

In such a troubled world as the present, one can easily believe that more local economic developers may be tempted to kick the tires and give it a try. So, let’s take a brief look under the hood and see what wonders lie underneath and make some judgments as to whether this economic revitalization vehicle is suitable for your community and EDO.

While clusters as a concept have been around for quite some time (often labeled agglomeration economics), it was Michael Porter who really launched this rocket and who remains its patron saint and most visible proponent. Accordingly our discussion should start with his concept and use it as a foundation and a definition of this approach (warning, his is not the only definition/conception of cluster based economic development). After wading through several Porter efforts, I would suggest interested readers start with his Economic Development Quarterly, February 2000 article “Location, Competition and Economic Development: Local Clusters in a Global Economy,”Vol. 14, Issue 1, pp 15-35.

For Porter “Clusters are geographic concentrations of interconnected companies, specialized suppliers, service providers, firms in related industries, and associated institutions … in a particular field that compete but also cooperate” (p.15). Clusters are a vital economic development force in that companies which exist in a cluster are more innovative and productive than those outside the cluster and these dynamics are the keystone for economic growth and enhancement of the area’s standard of living. A geographic cluster, in his mind, creates an advantage for its component firms which translate into enhanced competitiveness and profitability providing an edge against their competitors. Also, clusters “are a driving force in increasing exports and are magnets for attracting foreign investment” (p.15). A cluster is not a single sector/industry/NAIC code; rather it is the interconnected relationship of a core product/service/function with other firms within the cluster and which are outside the core industry/NAIC code of the cluster. To facilitate and grow the expansion/productivity/innovation of firms within the cluster is the objective of local economic development efforts. Any growth that clustering produces arises from the subtle, almost invisible, spillovers, innovation and efficiencies which result from geographic proximity and which is unique to those firms within the cluster. Acquiring firms to fit into the cluster is not the point of the cluster approach; it is enhancing those interrelationships to produce more growth and cluster efficiencies.

Acquiring firms to fit into the cluster is not the point of the cluster approach; it is enhancing those interrelationships to produce more growth and cluster efficiencies.

Cluster-based economic development is therefore MUCH MORE than targeting sectors (such as data storage or back offices) or rationally focusing an EDO’s attraction or retention program around export-intensive, innovative industries. Indeed, as we shall shortly discover, the cluster approach actually rejects sector targeting (picking winners) in favor of a more “generalized” program of economic development to which a whole host of different sectors within the cluster would benefit. Any firm within the cluster is potentially eligible for economic development programs and assistance and demonstrated inclusion in the cluster determines eligibility. Cluster-based economic development programs, therefore, are targeted to firms within the cluster rather than to a single firm or a SIC/NAIC code. Obviously, (and we shall return to this point), the scope and the design of local economic development cluster based programs will be vastly different from the currently traditional local economic development programs.

Porter believes some type of cluster exists in virtually any region. But this raises two issues (1) are all clusters equally deserving of local economic development attention, and (2) how does one identify a cluster in the first place. For Porter, the clusters deserving attention are those which are exporting, innovating and growing. Most desirable are those which are “knowledge-based” which disproportionately are composed of high paying, more educated occupations. Mature, low paying and/or declining clusters are not suitable candidates for public attention.

The issue as to how an area discovers whether or not it has a cluster worth helping is relatively straight-forward. The existence of a cluster can be more or less measured either using consultant-based expertise or through existing local/state research institutions/departments. Porter’s own “Cluster Mapping Project” is probably the most sophisticated demarcation of large and small clusters currently existing. The methodologies (location quotients, interviews, surveys) are well understood, but are more useful for an existing, developed cluster than for uncovering a hoped-for emerging cluster. There is, however, a potential fly in the ointment of discovering a cluster. Clusters basically never really conform to a political jurisdiction such as a single county municipality, city, town etc. Indeed, a cluster may overlap several states and a cluster region can be quite large composed of many counties. Given that most EDOs are derived from jurisdictions, the existence of a cluster creates some obvious awkward complexities and hence would seem most appropriate to the state level or at least super region. One should assume that to embrace a cluster-based strategy one’s EDO had best be regional in scope, powers and membership.

Is it possible for a geographic area to build, attract or initiate a cluster on its own? Not likely! Clusters arise from natural economic, geographical and locational processes/advantages and are not created through public intervention.

Is it possible for a geographic area to build, attract or initiate a cluster on its own? Not likely! For Porter and virtually anyone connected to cluster based theory, clusters arise from natural economic, geographical and locational processes/advantages and are not created through public intervention. There is more ambiguity, however, as to whether one can discover a cluster in its infancy, and through various traditional economic development programs enhance the growth of this “emerging cluster”. Whether or not this can be done, there are very serious methodological issues which inhibit measuring and discovering such embryonic clusters and, at least at present, there is some skepticism regarding public intervention into some poorly defined and inadequately understood alleged interconnected companies. Any region or jurisdiction which possesses a number of companies/institutions which seem to offer a promising core for a potential “aspirational” cluster had best be wary of launching new program initiatives to build such an emerging cluster. They will find little, to no support, from the cluster literature.

CLUSTER DYNAMICS: How a cluster produces growth

To figure out how local economic development can positively impact an existing cluster, it is imperative we understand how clusters provide an advantage to their firms. Cluster-based economic development achieves its magic through two separate but related dynamics. First, are agglomeration advantages such as (1) intra-company efficiencies which result from co-location with other firms in the cluster (localization economies), but also through (2) “urbanization” economies which result from being surrounded by a ton of other firms and an infrastructure which is built to accommodate a large number of companies. Secondly, growth and innovation result from “knowledge spillovers” which arise from the informal, social networks that exist uniquely within the cluster. To accelerate and enhance these twin dynamic forces is the goal of public intervention.

Clusters potentially come in many different sizes, types and shapes. There is not one size fits all type of cluster. The typologies which seek to classify the various types of clusters are many and can be fine toothed or broadly construed. There does not appear to be any real consensus behind any one single typology. To provide some description as to the various types of existing clusters we can call attention to our favorite. Elegant and useful is Ann Markusen’s “Sticky Places in Slippery Space: A Typology of Industrial Districts” (Project on Regional & Industrial Economics, Rutgers University, New Brunswick, NJ). Dr Markusen suggests there are four types of clusters (industrial districts)

  • Marshallian/Italianate (locally-owned small and medium sized firms which are highly networked and dependent on its local base);
  • (b) Hub & Spoke (dominated by one or several large firms and surrounded by local suppliers and institutions);
  • (c) Satellite (dominated by branch facilities from multinational firms and their connected firms), and
  • (d) Government/Institutional Anchored (dominated by a public/non profit facility(s)
  • A useful, more recent addition to Markusen cluster typology is a functional cluster, such as the “headquarters” or logistics cluster.

Each of these types exhibit different dynamics, strengths, weaknesses, opportunities and each call for a distinctive set of public interventions to trigger growth/productivity/innovation. It isn’t sufficient to simply identify or discover a cluster; an EDO must understand its particular type and its characteristics and deficiencies in order to best devise a structured policy/program response.

This sounds complicated! It can be, but, like most jargon laden theory when you strip away and simplify the verbiage (as one must if one is to actually implement and build support for this approach), a set of tools, programs, and policies become visible. Drawing from several research articles, but relying chiefly on (and restating)Edward Feser’s inventory taken from “On Building Clusters Versus Leveraging Synergies in the Design of Innovation Policy for Developing Economies” in Blien & Maier (eds.)., The Economics of Regional Clusters: Networks, Technology and Policy, Cheltenham, UK. Edward Elgar, pp. 185-207) the local economic developer can select from a fairly large, but not exhaustive menu of tools.

  • Create a trade or cluster-wide association if none exists
  • Recruit new members to enlarge the existing cluster (and perhaps fill in missing gaps)
  • Create focused networking (herding cats might be easier)
  • Key infrastructure enhancement (airport, e.g.)
  • Enhance the capacity of supporting public/non profit institutions (such as universities)
  • Incubators, venture capital and start up/small business counseling/financing for new company formation
  • Lobby for state and local business climate and regulatory adjustments
  • Entry level and skill upgrading worker/occupational training
  • Purchasing links and public sector procurement programs
  • Enhance and augment the supplier chain

Notice what is largely absent from this menu of program tools? Many of the traditional, conventional, firm-specific tools currently in use by local economic developer; tools, such as revolving loan funds, tax abatement, business retention and recruitment. These are typically single firm targeted and associated with more traditional firm-specific cost minimization strategies. This absence is not an accident; cluster-based economic development is meant to be an alternative approach to current economic development tools and practice. The recipient of a cluster based program should be a number of firms, not just a single eligible firm. In particular, left behind are the highly criticized and academically unpopular tax abatement and subsidy wars which inherently (1) come at the expense of the tax payer, and (2) provide subsidies and incentives to some private firms (who often do not require these subsidies) and not to other, often competing, firms.

Clusters represent a different way to approach firms in your region. Rather than through sectors and specific firms, cluster targets a group of firms which are interconnected to the area’s cluster and which, in their somewhat mysterious way, enhance the productivity, cost basis, and innovation of the end product/service. The public goal is to affect processes and supply chains so to achieve efficiencies without direct subsidization; the goal is to provide productive and innovative inputs, such as workers with appropriate skills, entrepreneurs, not tax abatements. Clusters are intended to move economic developers away from zero sum cost minimization programs. There is a very real agenda attached to cluster-based economic development and it should be no surprise that there is a strong ideological dimension (liberal/conservative) inherent in this approach. The benefits attached to the cluster approach typically flow less directly to the firms in the cluster, than to the institutions (often public/non profit) that provide the training and affect the processes (universities, workers and community colleges for instance).

A Few Questions and Comments about the Cluster Approach

Books have been written on clusters and our all too brief summary of clusters only captures some of what they are and what they offer. It is time, however, to discuss some implications which arise from cluster-based economic development, from which the local economic developer could benefit.

Clusters are a very sophisticated, complex strategy with considerable built-in subtleties. That frankly is one reason they are attractive to academics/think tanks and professional analysts/consultants. For local and even regional and state economic developers, seeking to build support, create structural/ programmatic responses, and to seek and defend resources needed for the strategy, cluster strategy possesses some very interesting issues. The Cluster approach is still largely theoretical and highly conceptual. Without math, sophisticated databases, advanced econometric techniques and reasonably sophisticated regression and correlation analysis, you wouldn’t always know it exists and what its boundaries are. You can’t even see one and you can actually work in one and not know it. This makes it hard to prove and to demonstrate to a larger audience its merits. Politicians and policy makers will have to assert that all these intricacies that exist within a cluster exist and woe be to them if someone actually asks to see one.

Initially, if this strategy were to be implemented, considerable capacity building or a sustained infusion of expertise must be imported to the EDO. A typical jurisdiction-funded EDO does not usually contain the mix of skills and program resources essential to the cluster approach. Most EDOs do not operate at the regional level. Existing Edo tools and its past decades of performance are now largely irrelevant. The key elements of the cluster approach are more employee/leader/process change internal to a wide variety of firms and workers; WIA and MEP EDOs as well as educational institutions, are more strategically placed than the traditional EDO. Proponents of clusters may label the possible tension as “turf-protection” but since the cluster based approach represents an alternative, if not rival, approach to traditional economic development practice, this charge would be very misleading.

As stated earlier, the cluster based approach is very dependent upon relatively sophisticated mathematical/modeling techniques. Arguably the most effective methodology for both indentifying a cluster and understanding its interconnectedness is an input-output table which (1) seldom exists for most geographies and (2) is very expensive and (3) the knowledge/information needed is proprietary and internal to the firms within the cluster which itself is scattered across a great number of jurisdictions and even states. This lack of readily available information and feedback regarding cluster programs creates some serious accountability issues.

It is very hard to separate an efficiency, innovation and spillover which results primarily from a cluster-related advantage from one created by other factors and forces. Worse, factors outside a cluster, including business climate, population and demographic changes, immigration or emigration, natural events and even congested infrastructure can mightily affect a cluster’s performance and even viability. We won’t even mention global transformations and permutations beyond the control of any cluster region. At the very least cluster theory is as vulnerable to external forces as any current approach to local economic development.

Also, it would appear extremely difficult to evaluate and measure the performance, even the appropriateness, of most cluster program and policy initiatives-except through more macro-economic statistical methodologies such as correlation analysis. These statistical techniques invite a level of analysis issue (that means if one observes change for a group of companies (cluster), one cannot assume that any one company reflects that change. Indeed cluster theory itself admits that two forces affect the individual firm (convergence which is negative and agglomeration which is positive). In the end, performance evaluation is likely to be antidotal and require a leap of faith. Fair share and does my community benefit from the cluster approach is going to be hard to prove.

Current cluster theory, as I understand it, cannot answer with any commonly understood data any of the following obvious questions which sometime will inevitably arise:

  1. Will the success and growth of the cluster lead to growth in the overall regional economy? Or can the cluster grow and the overall regional economies continue to decline or stagnate?
  2. What should we do with the companies and workers who are not in the cluster?
  3. Can a cluster mature (Yes?) Decline (Yes!)? If so, then what do we do? Must we suffer sometime in the future the plight of those regions presently in the old automotive cluster?
  4. If clusters are such growth machines, will the national landscape be transformed into growth and declining geographies, a reality which differs not at all from what exists at present?
  5. If I live and work in a region with a cluster, will my region grow 10%, 5% or even 1% more than my cousin who lives in a region without a cluster?

Even some cluster advocates question how much bang for the buck does a cluster actually deliver?

This last issue is a real thorn in the side of the cluster approach. Even some cluster advocates question how much bang for the buck does a cluster actually deliver? All of the elusive and somewhat mysterious benefits may well occur, but how does that translate into overall regional job growth, higher wages, greater prosperity and more taxes? Porter himself in a recent study, “Clusters, Convergence and Economic Performance” (Discussion Papers, U.S. Census Bureau, Center for Economic Studies, CES 10-34, October 2010) finds all sorts of performance advantages for clusters but on (p.26) begrudgingly admits “that the theoretical arguments for the role of clusters in productivity and innovation growth may be different from the ones that explain employment growth”.

This situation may be made worse by the fact Cluster advocates consciously select “gazelles” or the growing, dynamic clusters and firms within these clusters for public attention and involvement. Much of the cluster literature emphasize working with those clusters which are high paying, high skilled knowledge-based and innovation-prone with highest multipliers. With rare exceptions, however, most firms within regions and jurisdictions are not gazelles. These firms are at best secondary for cluster proponents; in many cluster models my sense is they are entirely forgotten.

Unemployed, low skilled, poorly educated workers cannot easily fit into these knowledge based innovation focused clusters. So what should we do, export them? There are examples of cluster-regions recruiting gazelle firms and highly educated workers into a region, superimposing them upon an existing layer of unemployed, marginally employed current residents. In these situations a supplemental strategy which focuses on providing assistance to less dynamic employers who can potentially hire existing residents may be desirable. Some less charitable than I might mumble words such as elitism, class conflict, or even anti populist as appropriate to a description of the cluster approach.

Closing Thought

Clusters are not a magic bullet, however, sexy and popular they might currently be. For most communities and EDOs, a cluster-based economic development strategy, if they are to see it at all, will be imposed upon them from above: by the state, regional EDO, or even the federal government. It is not yet a realistic or even feasible local economic development strategy. A community might be best served to wait for Version 2.0 of cluster theory.

Richard Dreyfuss went on to live his life and find new loves and opportunities (and so did Suzanne Somers his unknown lost love). At least as regards to cluster economic development, so can most communities.


No comments yet. You should be kind and add one!

The comments are closed.