Abstract

This paper estimates the impact of paved road in the during pandemic using household welfare using the Family Income and Expenditure Survey-Labor Force Survey. The difference-in-difference regression used accounted for the complex design of the survey data. The result shows that if the sampling design is ignored the standard errors are underestimated leading to an overestimated impact increasing the chance of type I error. If the complex sampling design is correctly accounted for, however, the impact of additional road is limited to urban areas.