Dear statalists,
In my master thesis I am studying the impact of urban regeneration projects on a district's housing market values, within a single city, for projects realized between the years 2011 to 2021 (and data until 2023). I'm using a DiD, and as the treatment is non-binary and happens at different times for districts, I am using the did_multiplegt_dyn package.
As covariates, I have data available on income and population at a district level. My confusion stems from the issue of endogeneity: urban regeneration projects make a district more attractive, and therefore contribute to attracting more residents and, if house prices increases, higher-income residents. At the same time, income and population also impact housing costs: if more people live in the district, demand will rise, availability will decrease in price will go up; if for other reasons the neighborhood is gentrifying, the increase in average income will make prices increase.
My question is if I can still use them as covariates in this case? Is there any specific procedure I can use to reduce the endogeneity concerns?
Thank you everyone!
In my master thesis I am studying the impact of urban regeneration projects on a district's housing market values, within a single city, for projects realized between the years 2011 to 2021 (and data until 2023). I'm using a DiD, and as the treatment is non-binary and happens at different times for districts, I am using the did_multiplegt_dyn package.
As covariates, I have data available on income and population at a district level. My confusion stems from the issue of endogeneity: urban regeneration projects make a district more attractive, and therefore contribute to attracting more residents and, if house prices increases, higher-income residents. At the same time, income and population also impact housing costs: if more people live in the district, demand will rise, availability will decrease in price will go up; if for other reasons the neighborhood is gentrifying, the increase in average income will make prices increase.
My question is if I can still use them as covariates in this case? Is there any specific procedure I can use to reduce the endogeneity concerns?
Thank you everyone!
