Back from Glasgow, where I’ve been at the University talking about entrepreneurial ecosystems. Many thanks to Ben Spigel and Fumi Kitagawa for inviting me, and for organising such a nice group of people.
It was a pretty striking event. Here are some thoughts while they’re still fresh.
(I’m new to this field, though as an economic geographer I’ve done some work on related issues. So apologies in advance for anything that’s wrong or missing in what follows.)
1/ This is the first field I’ve found where many of the key authors seem deeply, deeply sceptical about the core subject matter. That’s odd. To get a sense of this, read this piece by Ben, this one by Erik Stam, or this one by Ross Brown and Colin Mason.
Erik defines entrepreneurial ecosystems as the ‘set of interdependent actors and factors co-ordinated in such a way that they enable productive entrepreneurship’. But he then suggests that this approach ‘contains no new separate insights’ over existing work on clusters, innovation systems, or urban economies more broadly. Rather, the appeal is in bringing a range of perspectives together.
Ben suggests that ‘ecosystems represent more of a conceptual umbrella encompassing a variety of different perspectives on the geography of entrepreneurship rather than a coherent theory about the emergence of sustainable communities of technology entrepreneurs.’
Debates at the workshop were pretty similar.
2/ The origin story is fascinating. It’s Jamie Peck’s ‘fast policy’, but in reverse. Discussions of entrepreneurial ecosystems were largely led by policymakers and have travelled from there back into academia. This is a paper in itself.
If we think of the entrepreneurial ecosystem as a kind of economic imaginary for policymakers, that helps explain its appeal. As we discussed on the day, the term can sustain both a laissez-faire policy agenda, and a more interventionist one.
It’s a good metaphor (I’ve used it more than once). It’s less obviously robust as a basis for research designs or detailed policy decisions. There are echoes of the Triple Helix, another policy-driven concept that has — arguably — shed less light on innovative activity on the ground than we’d hoped.
3/ Some of the specific limitations of entrepreneurial ecosystems thinking, as it currently stands, are:
- Lack of clarity on scale, and on identifying boundaries. In the past, similar objections were levelled at Michael Porter’s iteration of cluster theory. Here, we can respond by recognising that different industries / activities will generate a range of cluster shapes, that spillovers dissolve boundaries, and that production milieux are often parts of larger, cross-national systems. The regional innovation systems literature also has a version of this problem, which usually gets solved empirically by looking at the geography of knowledge spillovers. But this response is harder to apply to ecosystems, precisely because the concept is broadly drawn. If ecosystems include the role of national institutions and policy regimes, is there a boundary?
- Unclear processes and mechanisms within ecosystems. Or rather, the concept contributes nothing on its own. All of the analytical tools used come from urban economics, economic geography and economic sociology, or closely related fields. The ecosystems approach only seems distinctive in the emphasis it places on individual entrepreneurs over area-level externalities, communities and institutions: but this is largely asserted rather than backed up by data (and goes against a wealth of evidence in the other direction).
- Who’s (not) considered. As Rhiannon Pugh pointed out, so far empirical studies have largely focused on a small set of high-tech sectors and high-growth firms, rather than wider bodies of less prestigious or visible entrepreneurial activities done at home, or by more marginalised groups such as women and minority ethnic communities. Similar issues came up at this 2017 session organised by Suzi Hall, Monder Ram and I.
- Overly narrow outcomes. Theory and evidence focus on a small set of economic variables, without considering either social / environmental outcomes, or the role of policy choices. Public policy and institutions are held to influence ecosystems; so there should be conceptual space both for these channels, and for different policy stances that might shift the relevant set of outcomes we as researchers should be looking at.
4/ So what’s the value of the ecosystems concept? Partly it feels like a case of ‘new to who’: it’s a spatial turn for entrepreneurship research, which historically has focused on individuals rather than places.
Erik suggests that ecosystem thinking has specific value for debates about industrial strategy, as it integrates market failure and co-ordination / network failure perspectives. But arguably, that analytical work has already been done by economists and political economists – see this 2004 piece by Dani Rodrik, this 1996 piece by Adam Jaffe, or this by Carlota Perez.
More broadly, though, there’s potentially great value in bringing together multiple disciplines under a single organising concept: ‘economists and geographers and economic sociologists and management scholars and innovation scholars’, as Fumi put it.
Many geographers already try to do this — after all, we look at where economic, social, cultural and political things happen, which inevitably involves interdisciplinarity and multiple perspectives. And there are various ways to do this. One is super-rich description: say, Stephen Graham. Another is analytical narrative: say, AnnaLee Saxenian or Michael Storper. Both of these approaches are good for single or small-n case studies, such as cross-city or cross-country comparison. A third option is to shift towards abstraction and model-based thinking, that is, towards spatial economics or social network analysis. We saw some promising examples of how newer datasets can contribute here: for example, Augusto Rocha’s work using Meetup data to map actor networks and clustering.
All of these are feasible ways forward for entrepreneurial ecosystems research in the future. The challenge will be to get the core concepts to deliver some insight that isn’t already present somewhere else.