The timing of commercial breaks and music variety in the radio industry

Abstract

This thesis contains three empirical essays using a new panel dataset of airplay on contemporary music radio stations. The first and third essays examine the timing of commercial breaks. Stations tend to play their commercials at the same time but it is unclear whether this is due to stations wanting to coordinate on timing, to reduce the avoidance of commercials, rather than common factors which simply make some times better for commercials. The first essay models timing decisions as an imperfect information coordination game. The game has multiple equilibria, in which stations coordinate but on different times, if the incentive to coordinate is strong enough. The essay shows how the existence of multiple equilibria, both in the game and in the data, can help to identify the parameters of the game and how to test for multiple equilibria. There is evidence of multiple equilibria, allowing the incentive to coordinate to be identified, during drivetime hours. The implied degree of coordination in Nash equilibrium is relatively modest but, because of externalities in the coordination game, the estimates also imply that coordination would be almost perfect if stations maximized their joint payoffs. The third essay uses a set of simple models to show that, if stations want to choose either the same or different times for commercials, the degree to which commercials overlap should depend on market characteristics such as the number of stations and the degree of common station ownership. The majority of the evidence supports models in which, on average, stations want to choose the same time for commercials as other stations in their market.

(cont.) The second essay examines how changes in radio station ownership have affected music variety and finds that a common owner of stations in the same local radio market and the same music category significantly increases the degree of differentiation between these stations, consistent with a model in which common owners internalize "business stealing" effects. Common ownership of stations in the same music category but different local markets results in, at most, a small degree of playlist homogenization. Panel data on station listenership provides further support for the business stealing explanation, as when stations in a local market become commonly owned their audiences tend to increase.