This paper examines investment in transport infrastructure in Swedish municipalities according to the three National Transport Infrastructure Plans of 2004, 2010 and 2014. The plans cover 12 years each. The test of a swing voter model, combined with variables relevant to the Keynesian model of demand side policies, supports the proposition that there is less investment in municipalities with highly partisan electorates. The model seems to work better for road than for rail investments. Municipalities with a high density of voters at the ideological cut-point (middle of the ideological distribution) got more investment in the 2010 plan but not in the other plans. The impact of the elasticity of output on public service provision raised investment in road projects in sub-plan period 1 compared to later sub-plan periods. The tax elasticity of output may influence the volume of investment downward. The Plan for 2010-2021 seems to be the most politically determined of the plans considered here.