Question

1. A researcher is interested in quantifying the effect of the number of broken windows in a block on property prices, and the results of a preliminary regression are: Hprice, 100 10(windows:) + Ei where windows, represents the number of broken windows counted on a block, i, and Hprice, is the average property value (in thousands of $) on that same block. a) What is the interpretation of the coefficient on windows,? b) What are the causes of endogeneity in this model (consider all three)? c) Use a DAG diagram (with signs for each side and overall) to make a case that this model over or under estimates the effect of broken windows on house prices. d) Another variable is included in the regression, emerg, which is a measure of the number of emergency services calls which were made from each block over a period of time. And the result of the regression is: Hprice, 100-3(windouwst) - 5(emerg:)+e

Answer 1

a) The coefficient on windows implies that an additional broken window in a block leads to the average property value decreasing by 10,000$ on that block.

b) Possible causes for endogeneity are when the explanatory variables are related to the error term itself. This could be because of omitted variables, measurement errors or unobserved heterogeneity.