Deploying municipally-run broadband Internet to deliver high-speed access to residents is an increasingly popular idea among local communities that are concerned that their current Internet offerings are substandard or otherwise uncompetitive. These public networks often operate alongside private incumbent networks and might have substantial effects on the competitive landscape. Many states have passed restrictions on how local communities can offer service. Colorado law prevents a city or county from offering service unless a referendum is passed. Since 2005, voters in over 140 local jurisdictions, including half of Colorado’s counties, have voted to approve bypassing the state prohibition. So far only one city has launched a network.
This research design exploits variation in the timing of a community’s approval of the referendum to see if incumbent private ISPs adjust their speed offerings (a proxy for quality) in response to the signal (referendum passing) that public entry is more likely. It is possible to utilize a difference-in-difference framework to see if, over the period 2012 to 2018, incumbent upgrades in quality are faster or slower depending on whether they face a greater public entry threat. Faster upgrades could suggest that incumbents do so to deter an entry threat or that potential competition could be quality-enhancing. Slower upgrades could suggest that public entry threats crowd out investment in quality. These effects might be amplified if they are in communities with an existing municipal electric utility, which arguably makes entry more likely.
Empirical studies have found it difficult to separately identify adverse selection from moral hazard since the individual effects tend to affect observable behavior in the same way. Using the state level dependent coverage mandates that were passed before the Affordable Care Act's dependent coverage mandate took effect, I am able to control for selection into insurance and more credibly identify moral hazard. More specifically, I use the variation in eligibility criteria and the timing of implementations of the mandates across states over time in order to discern among the individual effects of hidden information. Data from the Behavioral Risk Factor Surveillance System allows me to consider the potential role of moral hazard in health insurance markets and more generally allows me to consider the effects of dependent health insurance on health outcomes for young adults.
Willardsen presented on his upcoming article with the same title. The abstract from this paper is as follows:
Understanding the relative significance of adverse selection and moral hazard is important in determining effective policy for insurance markets. Separate identification of these two effects, empirically, is difficult. To overcome this limitation, this paper uses experimental methods to examine how adverse selection and moral hazard separately affect agent performance in a real-effort task. In particular, we explore how agent behavior (effort in the task) changes across a baseline with no insurance option, a treatment where individuals can choose to purchase insurance, and a third treatment where individuals must purchase insurance. We find that our platform can be helpful in studying issues that may arise in different insurance settings (e.g., workers compensation insurance or unemployment insurance).