Valuation of Local Public Goods: Migration as Revealed Preference for Place

Abstract

We develop a residential sorting model based on a panel of county-to-county migration flows to estimate the marginal valuation of air pollution. Our approach exploits annual crosssectional variation in migration flows to estimate mean location utilities at the county level, while flexibly controlling for moving costs. The mean utilities provide a time-varying, county level index of residential attractiveness. We then use panel variation in county characteristics to decompose mean utility into observable and unobservable components using county fixed effects, which allows us to estimate the marginal value of local amenities. In our application to air pollution, we use an instrumental variables approach to provide robust evidence that the concentration of fine particulate matter (PM 2.5) is a disamenity that negatively impacts location decisions. In our preferred specification, we find that the median household is willing to forgo 2.65 percent of annual income for a 1 mg/m3 decrease in fine particulates.

Motivated by recent investigations over Google's practices in the smartphone industry, we study bundling in markets for devices that allow consumers to use applications. The presence of applications on a device increases demand for it, and application developers earn revenues by interacting with consumers. A firm that controls multiple applications can offer them to device manufacturers either individually or as a bundle. 
We present a novel mechanism through which anti-competitive bundling can be profitable: Bundling reduces rival application developers' willingness to pay manufacturers for inclusion on their devices, and allows a multi-application developer to capture a larger share of industry profit. Bundling can also strengthen competition between manufacturers and thereby increase consumer surplus, even if it leads to foreclosure of application developers and a loss in product variety.