There is considerable debate on whether the employment and earnings prospects are better for those on low pay or for the unemployed. Low-pay work tends to be undertaken more locally but no empirical analysis has focused on how local opportunities alter prospects. Using Understanding Society data for England matched with local unemployment rates, we estimate dynamic random effects panel models, which show robust evidence that the future unemployment risk is lower for those who are currently on low pay compared to those who are currently unemployed. The low-paid also have a higher chance than the unemployed of becoming higher-paid. These findings are most marked in neighbourhoods with high unemployment.