Introduction For policy makers, the allure of ethnic-minority entrepreneurship is multifaceted. Numerical significance is undoubtedly a key factor: there are more than a quarter of a million ethnic-minority businesses (EMBs) in the UK, which contribute at least »15 billion to the UK economy per year. In England alone, in 2004 EMBs made up 5.8% of small and medium-sized enterprises (SMEs) (Mascarenhas-Keyes, 2006). Furthermore, the growth amongst EMB start-ups is estimated to be double that of total business start-ups (Barclays Bank, 2005). Policy interest in EMBs has been boosted by a twin preoccupation of promoting`enterprise' and of combating`social exclusion' (Blackburn and Ram, 2006). Publicly funded initiatives to support EMBs have been a feature of the small-firm policy agenda since the Brixton disturbances in the early 1980s. Although commitment to this policy objective has waxed and waned, there has nonetheless been a steady stream of initiatives over the last couple of decades (Ram and Smallbone, 2002; 2003). Contemporary concerns over the`integration' of new and increasingly diverse migrant communities (Castles et al, 2002; Vertovec, 2001) also fuel interest in the capacity of self-employment to promote social mobility (Lyon et al, 2006). This paper assesses key academic and policy developments in the UK, and presents an agenda for future research on ethnic-minority enterprise. Its objectives are threefold. The first objective is to outline the`conceptual base' that underpins policy initiatives towards ethnic-minority entrepreneurship. We argue that an`ethnic-resources' model has held undue influence upon policy and academic discourses in this field. Key drawbacks include the tendency towards cultural determinism and a neglect of the context that shapes ethnic-minority entrepreneurship. A`mixed-embeddedness' perspective, which recognises the economic and social context of EMBs, holds greater promise. The second objective is to assess the extent to which policy initiatives have kept pace with developments in the evidence base. The`crowded platform' that Storey (1994)