Over the past two decades, urban and regional policy-makers have increasingly looked to the arts and culture as an economic panacea, especially for the older urban core. The arts' regional economic contribution is generally measured by totalling the revenue of larger arts organisations, associated expenditures by patrons and multiplier effects. This approach underestimates the contributions of creative artists to a regional economy, because of high rates of self-employment and direct export activity, because artists' work enhances the design, production and marketing of products and services in other sectors and because artists induce innovation on the part of suppliers. Artists create import-substituting entertainment options for regional consumers and spend large shares of their own incomes on local arts output. The paper takes a labour-centred view of the arts economy, hypothesising that many artists choose a locale in which to work, often without regard to particular employers but in response to a nurturing artistic and patron community, amenities and affordable cost of living. Because evidence on such economic impacts and location calculus is impossible to document directly, the distribution of artists across the largest US metropolitan areas is used as a proxy, using data from the PUMS for 1980, 1990 and 2000. It is found that artists sort themselves out among American cities in irregular fashion, not closely related to either size or growth rates. The paper further explores variations in the definition of artist, the relationship between artistic occupation and industry, and differentials in artists' self-employment rates and earnings across cities. It is concluded that artists comprise a relatively footloose group that can serve as a target of regional and local economic development policy; the components of such a policy are outlined.