Cluster policies are popular with policymakers, but we know surprisingly little about their effectiveness. This paper evaluates the causal impact of a flagship UK technology cluster programme that uses `light touch?, market-orientated interventions. I build a simple framework and identify effects using synthetic controls plus placebo tests to handle programme endogeneity concerns. I implement this design on rich workplace-level microdata from the Business Structure Database, alongside a range of other administrative sources. I explore mechanisms through further tests for timing, cross-space variation, scaling and churn. The policy grew and densified the cluster, but has much weaker, partial effects on tech firm performance. I also find most policy `effects? began before rollout, raising questions about the programme's added value. More careful policy design could make future interventions more effective. ? 2020

---

30 November 2022

---

Research Policy 51(9) 2022

---

DOI: 10.1016/j.respol.2020.104138

https://www.sciencedirect.com/science/article/abs/pii/S0048733320302134

This Journal article is published under the centre's Urban programme.