Nineteenth century state and local governments continued to invest in railroads and other internal improvement projects long after it was clear that these projects were financially very risky. This paper provides a motivation for public involvement in internal improvements by estimating the effect of railroad construction on property values from 1850 to 1910. Using Census data on true and assessed valuations, we find that the increase in property values associated with railroad construction, would, at typical levels of taxation, pay for a substantial share, if not all, of the construction costs solely on the basis of property tax revenues. The effect of construction on property values declined with mileage up to several thousand miles, which may explain why state governments typically were involved in construction of the initial systems. The effect, however, was nonlinear and increased at higher mileages, consistent with the persistent participation of county and municipal governments.