UC Irvine

In the first essay, ``Does Government Health Insurance Reduce Job Lock and Job Push?'', I estimate the extent that job mobility is affected by the link between health insurance and employment. Workers holding employment-contingent health insurance (ECHI) are often thought to stay in jobs that are otherwise inferior matches out of fear of losing their ECHI, while those without insurance may leave employment states that are otherwise good matches seeking access to ECHI. These two phenomena are known as job lock and job push, respectively. During the late 1980s and early 1990s, Medicaid expansions resulted in many working class households gaining Medicaid eligibility for one or more family members, an alternative source of health insurance that is not contingent on employment. Using this eligibility as a measure of variation in the dependence on ECHI for health insurance coverage, I find large estimates of job lock and job push for men. Medicaid eligibility for one household member results in an increase in the likelihood of a voluntary job exit for men over a four-month period by approximately 34%. Similarly, moves into jobs with ECHI fall by approximately 25% in response to Medicaid eligibility. For women, I do not find evidence consistent with job lock. For the case of job push, some of my estimates suggest large effects, though these estimates have interpretive difficulties.

The second essay, titled ``Does Regulation of Physicians Reduce Health Care Spending?'', examines the fear among physicians that legal liability increases health care spending. Theoretically, the effect of legal risk could be positive or negative on spending, and empirical evidence has supported both cases. Previous empirical work, however, has ignored that physicians face risk from centralized regulators -- industry oversight groups like medical boards -- in addition to civil litigation risk. This paper addresses this omission by incorporating previously unused data on punishments by oversight groups against physicians, known as adverse actions, along with malpractice payments data to study state-level health care spending. My analysis suggests that health care spending does not rise in response to higher levels of risk. An increase in adverse actions equal to 16, the mean year-to-year change within a state, is found to be associated with statistically significant average spending decreases of approximately 0.11% to 0.21%. Malpractice payments were generally estimated to have smaller, statistically insignificant effects.

The final essay, ``Preliminary Results On The Effect of Specialist Cost Information on Primary Care Physician Referral Patterns'', reports early results on a field experiment designed to test whether primary care physicians (PCPs) would use information on specialist costs in allocating their patient referrals between doctors within the specialty. The experiment was performed in partnership with a private-sector group of medical practices organized as an Independent Practice Association (IPA). Randomly chosen PCP practices within the IPA were provided with a report listing average cost information for Ophthalmology practices within the IPA. The response of the PCPs is compared to a control group of PCP practices within the IPA to see if the information influenced which Ophthalmology practices received PCP referrals. Analysis of experimental data so far available does not find any effects that are statistically significant at conventional levels. These results, however, are based on data from a very short post-period, and are not considered final. The experiment is ongoing at the time of the writing of this essay.