Electoral institutions, which determine the manner in which citizens choose their representatives, and economic geography, which refers to the distribution of economic activities across space, together help to explain the variation in subsidy spending between democracies. Economic geography is measured here using disaggregated employment data that identify employees’ geographic location as well as their sector of employment. These data reveal that real world employment patterns rarely conform to the assumptions adopted in theoretical models of economic policy making. Relaxing these assumptions reveals new predictions about the policy effects of electoral institutions and novel quantitative evidence supports these predictions. Subsidies for the manufacturing sector constitute a larger share of government expenditures in plurality systems than in proportional systems when manufacturing employment is geographically concentrated. When manufacturing employment is geographically diffuse, governments in proportional systems assign relatively more of their budgets to manufacturing subsidies than governments in plurality systems, holding all else equal. This finding suggests economic geography can help to solve the debate over which electoral institutions generate the most particularistic economic policies.