ABSTRACT OF THE DISSERTATION

The Impact of Political Fragmentation on Health Disparities and Local Health Expenditure

By Yonsu Kim

Doctor of Philosophy in Planning, Policy, and Design

University of California, Irvine, 2016

Associate Professor Tim-Allen Bruckner, Chair

During the 1970s and1980s in the U.S., population movement, urban sprawl and governance reform led to a proliferation of local, autonomous jurisdictions. Previous literature examines how this creation of local governments, referred to as political fragmentation, contributes to public spending and social inequality. I build on this literature by examining three potential consequences of political fragmentation: increased health disparities, reduced public investments in health, and reductions in public economic efficiency. I ground these lines of inquiry from diverse fields including political science, urban planning, and public health.

In my first analytic chapter, I test the hypothesis that the mortality disparity between whites and African Americans varies positively with fragmentation. I retrieved mortality data from the multiple cause-of-death file and measured fragmentation among 226 U.S. counties (population size ≤ 200,000) over four distinct periods (1972-73, 1977-78, 1982-83, and 1987-88). Consistent with my hypothesis, I find a positive relation between political fragmentation and the African American / white difference in all-cause mortality.

Next, to explore a pathway through which fragmentation may increase to health disparities, I examined whether fragmentation corresponds with a reduction in public health expenditures. Consistent with this proposition, I find that fragmentation varies negatively with per capita health expenditure (coef: -0.03, p<0.01), and health expenditure as a proportion of total expenditure (coef: -0.0084, P<0.001). In my last analytic chapter, I test the Leviathan hypothesis that fragmented local governments act inefficiently and lead to unexpected rises in overall local expenditures. Counter to this hypothesis, results indicate that general purpose government fragmentation constraints per capita expenditure (coef: =-0.953, p<0.001).

Taken together, my results suggest that political fragmentation may exacerbate health inequalities by reducing the per capita public outlay for health spending among disadvantaged groups. However, my findings remain inconclusive regarding whether, and to what extent, political fragmentation promotes economic efficiency in public management of urban locales. I conclude with recommendations of future areas of research.

Background: Economic downturns may increase stress due to income and job loss among the directly affected, as well as uncertainty, fear and anxiety among those who remain employed. For those who work during recessions, studies suggest a rise in working hours, reduction in pay and increased stress due to uncertainty of employment. During such times, populations may experience adverse mental health and greater psychiatric help-seeking manifested as higher number of emergency department (ED) visits. Conversely, recessions may also correspond with increase in healthy behavior as a response to financial uncertainty- a phenomenon also referred to as the “inhibition effect.” Low-income African Americans appear especially vulnerable to economic recessions. During the recent 2008 recession, African Americans lost more jobs, had a steeper reduction in income and were more likely to lose health insurance relative to whites. Societal responses to ambient stressors also put African Americans at greater risk of psychiatric hospital admissions. In addition, reduction in tolerance towards behavior deemed as ‘deviant’ or ‘threatening’ during economic contractions may result in greater reporting-- especially of African American males-- to law enforcement for involuntary psychiatric holds and inpatient admissions.

Objective: In this dissertation, I test whether economic downturns precede an increase in: the overall incidence of psychiatric ED visits (Chapter 2); psychiatric ED visits among African Americans more than whites (Chapter 3); and psychiatric inpatient admissions requested by law enforcement/courts among African American males relative to other race/ethnicity and gender groups (Chapter 4).

Methods: I operationalize exposure to recession ‘shocks’ as monthly declines in employment in a Metropolitan Statistical Area (MSA). I use data from the Statewide Emergency Department Database (SEDD) and the Statewide Inpatient Database (SID) for select US states (Arizona, California, North Carolina, New Jersey and New York) to retrieve, as my outcome, the census of all psychiatric ED visits and inpatient admissions for the time period of 2006 to 2011. I use monthly time resolution to establish temporal order such that the exposure (percentage change in monthly employment) precedes the outcome (monthly psychiatric visits). I specify brief exposure time lags of 0 to 3 months to estimate proximate responses to MSA-level aggregate macroeconomic decline. I utilize linear, logistic, negative binomial regression methods with inclusion of region, month, year fixed effects and linear time trends.

Results: In Chapter 2, I find that psychiatric ED visits decline immediately following MSA-level aggregate employment decline. This decline concentrates among those with private insurance, and among alcohol abuse-related emergencies. Concurrently, I also find an increase in psychiatric ED visits following MSA-level employment decline among publicly insured children. In Chapter 3, I find that working age (18 to 64 years old), publicly insured African Americans (versus whites) show increased odds of a psychiatric ED visit within 0 to 3 months of MSA-level employment decline. In Chapter 4, I find that psychiatric inpatient admissions requested by law enforcement/court orders increase within one month of aggregate employment decline among African American males but not among other race/ethnicity and gender groups.

Conclusion: The aggregate, population-level decline in psychiatric ED visits supports the ‘inhibition’ mechanism wherein economic uncertainty may correspond with reduction in unhealthy consumption. These results also align with prior research which posits that mental health care shows high price-elasticity in that reduction in income may elicit a greater decline in mental health care utilization relative to other types of health care. However, findings among publicly insured children indicate greater vulnerability among low-income groups that may differ in psychiatric response during economic contractions, relative to high-income groups. Increased odds of psychiatric ED visits among working age, publicly insured African Americans, relative to whites, also provide evidence of differential vulnerability of this group during ambient economic crises. Lastly, an increase in psychiatric inpatient admissions requested by law enforcement/court orders among African American males supports the ‘reduced tolerance’ hypothesis and highlights unique social responses that may affect mental health outcomes among African American men during recessions. To my knowledge, the research described in this dissertation presents the first evidence reconciling pro- and countercyclical trends in population-level psychiatric emergencies following economic downturns, and highlights the unique vulnerability of African Americans over a time period that includes the recent Great Recession.

Objective

Design

Setting

Primary outcome and exposure

Results

Conclusion