Chapter 1 examines the relationship between marriage and the economy during the 1930s.The Great Depression provides an ideal setting to examine the impact of economic downturns and recoveries on marriage outcomes. Using microeconomic data, I find that during the Great Depression, a standard deviation decrease in retail sales per capita, my proxy for local GDP, lowered a woman's probability of marriage by 18 percent. During the first few years of the crisis, the effect of GDP on marriage rates operated largely through high male unemployment.

Chapter 2 explores the impact of housing on marriage in the postwar baby boom period. The U.S. experienced an unprecedented increase in fertility during the baby boom. The elevated birth rates from 1946 to 1964 were driven in part by a shift toward more universal marriage; marriage rates increased by 25 percent from 1930 to 1950 and the average age of marriage fell by two years. This chapter argues that growth in the supply of housing after World War II contributed to the expansion of marriage during this period. Specifically, the paper estimates the effect of additional building permits at the city level on individual marriage outcomes. An instrumental variable approach is used to address endogenous permit location. I find a standard deviation increase in permits to a city increased the probability of marriage in that city by 13 to 16 percent over a two-year period. The estimates suggest that the growth in housing supply in the late 1940s can explain about 33 percent of the difference in marriage rates between 1930 and 1950. Overall, the increase in housing supply can account for nearly ten percent of the baby boom.

Chapter 3 re-examines the Easterlin hypothesis with microeconomic data. Easterlin argues relative income was the primary driver of the post war baby boom. I test this hypothesis using state income data in conjunction with individual level data from the U.S. census. I find relative income did positively impact completed fertility. However, this effect largely operates through a cohort effect and relative income can explain at most fifty percent of the baby boom.

I document trends in black and white retirement rates and in living arrangements among retirees. I show that the retirement rates of both blacks and whites rose between 1900 and 1930 but that convergence in black and white rates and in living arrangements only occurred between 1930 and 1950. I examine whether rising income explains the rise in black retirement rates prior to 1930 and whether rising income and the institution of Social Security in 1935 led to convergence by looking at the impact of the first pension program available to both blacks and whites, that serving Union Army veterans. I find that blacks were 2 to 5 times as responsive as whites to income transfers in their retirement decisions and 6 to 8 times as responsive in their choice of independent living arrangements. The results suggest that income effects from the institution of Social Security explain up to half of the convergence in black-white retirement rates and in living arrangements.

The first chapter of my dissertation uses a competition framework to reexamine the U.S. railway expansion from 1870 to 1890. Do reductions in trade costs increase the income levels of trading regions? The answer of most theoretical trade models is yes, and these models are used by policy-makers to justify investments in new transportation infrastructure to reduce trade costs. I reexamine this question by investigating the impact of the U.S. railway expansion from 1870 to 1890 on county economic growth. My main insight is that while the construction of a new transportation network can benefit a location by increasing producers' access to markets, it also can hurt a location by giving other locations a competitive edge. I derive and test empirically testable implications of the joint effects of market access and competition on economic outcomes from a multisector trade model. I find that on average more than 50% of the benefit of market access is offset by the negative effects of competition. Overall, I find that the expansion of railways increased agricultural land values by 20% to 40%, thus implying that railways explain 1% to 2% of 1890 GNP. My estimates of the impact of railways on American economic growth are lower than those of Donaldson and Hornbeck (2016) and closer to those of Fogel (1964).

In the second chapter I study the effects of sector-specific productivity shocks on the distribution of production. The general-equilibrium trade model implies that the production in a certain location depends not only on its own productivity shocks but also on productivity shocks in competing locations. I derive several testable implications of the joint effects of one location's own productivity shock and productivity shocks in competing locations on the location's production. I test these implications by investigating the impact of the boll weevil in the early 20th century United States on the cotton belt's cotton acreage. I find (1) when few lands in the cotton belt were infested by the boll weevil, the presence of the boll weevil decreased the infested county's cotton acreage; (2) as the boll weevil expanded to more lands, the infested county's cotton acreage would gradually recover; (3) as the boll weevil expanded to more lands, cotton acreage in uninfested counties would increase. The movement of other economic outcomes which closely associated with the cotton production (such as corn acreage, total farm acreage, agricultural land value, population, and the tenancy system) mirrored the movement of the cotton acreage. My findings are in line with the argument that the boll weevil caused the internal shift in the cotton belt but rarely changed the whole cotton belt.

In the third chapter, coauthored with Ming Gu, we discuss the long-run effects of the propaganda by the central government on people's preference formation. This link is taken as given but never proved in the previous literature. This paper examines the impact of government propaganda on people's attitudes towards gender equality in China, which imposes strict media censorship, advocates gender equality, and experiences growth in female labor participation. We find the high propaganda intensity would lead to progressive gender-related attitudes in female labor participation, but also conservative attitudes in housework share. It is consistent with the context of gender equality propaganda in China. Propaganda has larger positive effects on women than men. Women were affected in the age range between 0-25, while men were affected in the age range between 0-17. Propaganda also reduced boy preference for both men and women at the age range between 18-25, the marriage age and childbearing age of most Chinese.

This dissertation contributes toward our understanding of social change and socio-economic development. Chapter 1 introduces multiple key factors driving social change and leading to economic and social development. In Chapter 2 to Chapter 4, by analyzing historical natural experiments, I study the role of information, social interactions, and human capital in social changes, and their relationship with economic and social development.

Chapter 2, “Political Information, Social Interactions, and Protest in Late Imperial China”, investigates how information and social interactions drive social change by leading to collective actions (protests). A modern communication infrastructure allows information to reach people through the news media, while social interactions are needed to coordinate social movements. The postal system’s rapid construction in 1903-1910 in China let newspapers spread information directly to an increasing portion of the population. The change in information diffusion coincided with intensive media attention on revolutionary activities before the Revolution of 1911. I find that the construction of more post offices in a place led to more protests in the years with more reports about revolutionary activities in newspapers. I further disentangle the roles of direct information diffusion and social interactions. I define a village network based on the village’s location, the walking time between villages, and the village’s dialect group. I build and estimate a game-theoretical model based on the village network. As political information directly changed the villages which had post offices nearby and could receive information, I also find a strong peer effect: a village was affected by its expectation of its neighbors’ actions. The peer effect spread the direct impact of political information through social interactions.

Chapter 3, “Wartime Social Interactions and Veteran Migration in The Post-American Civil War Era”, explores the role of social interactions in social change and economic development after the American Civil War. I focus on temporary social interactions among African American veterans during the American Civil War (1861--1865), and examine the long-term impacts of temporary social interactions on veterans’ migration and income in 1870--1900. I find that wartime social networks (veterans from the same company) persistently affected veterans’ location choices in the post-Civil War period. By estimating discrete choice migration models, I quantified that the veterans were more likely to move to a county where men from their military company lived. I further show the long-term benefits of living together. Veterans earned higher incomes after the war if they lived in the same county with wartime friends who had higher incomes after the war.

Chapter 4, “Temple Destruction, School Construction, and Modern Human Capital in 20th Century China” (jointly written with Shaoda Wang), studies how modern human capital emerged in early 20th-century China and its impacts on economic development. We documented a historical episode known as the Temple Destruction Movement (TDM), during which Chinese local governments appropriated huge amounts of Buddhist and Taoist temple assets to support the modernization of the local schooling system. We found that before the TDM, the initial stock of temple assets was uncorrelated with the levels and trends of human capital development. However, after the TDM started, regions with higher initial stocks of temple assets constructed more modern schools, enrolled more students in modern educational programs, and produced more modern elites.

This dissertation contributes to our understanding of commodity markets, financial markets, and production during wartime.

Chapter 1 provides a short introduction to multiple channels through which wars affected an economy studied in Chapter 2 to Chapter 4.

Chapter 2 investigates how financial integration impacts price integration in real sectors? Using hand-collected data from a domestic exchange market during the Chinese Civil War (1945-1949), I model the connection between these two integrations measured respectively by capital flow costs (or domestic exchange rates) and relative commodity prices across cities. I use battle shocks to a financial hub in the exchange network to identify the impact of exchange rates on price convergence between a city pair connected to the hub. I find that (1) city pairs with a direct domestic exchange link exhibited faster commodity price convergence than others; (2) battles around financial hubs tended to raise capital flow costs between a connected city pair, decelerating price convergence by 4% - 8%; (3) a weak form of purchasing power parity holds: a 1% depreciation in the domestic exchange rates was associated with a 0.2-0.3 percentage point reduction in inflation rate differentials; and (4) a higher inflation rate did not impede or strengthen the price convergence channel via domestic exchanges. These results imply that China's financial development was more sophisticated than expected relative to its status as an agricultural economy in the early 20th century.

Chapter 3 studies how bond traders' extreme beliefs of repayments led to bond market anomalies. The conventional valuation model states that the value of a bond is the sum of the present values of its future cash flows. However, the prices of the consolidated government bonds soared up to thousands of times their face values from 1946 to 1948 during the Chinese Civil War. I find that (1) the bond prices anchored the Shanghai wholesale price index amid hyperinflation; (2) the extreme price pattern was primarily explained by the bond traders’ evolving belief that the government would repay the bonds with an appropriate multiplier after implementing the Currency Reform; (3) the belief was significantly affected by relevant news about repayment. On a weekly average basis, one more positive news about the Currency Reform will lead to an 8.8 percentage points increase in bond price growth rate, while one additional negative news about the repayment with multipliers will decelerate the growth rate by 8.2 percentage points; and (4) at least before the eve of the Currency Reform, the Shanghai financial markets were not sensitive to political uncertainties resulting from the Civil War.

Chapter 4 builds a two-layer supply chain framework in a static centralized economy to investigate the propagation of a productivity shock across sectors. I show that a productivity shock can be transmitted downstream (from an intermediate goods sector to a final goods sector), upstream, or horizontally (from a final goods sector to another sector). A negative shock incurred by a sector may hurt another sector if they are linked directly, but will always stimulate the production in an indirectly linked sector. Using detailed bombing data and production indices across armament sectors during the Second World War, this chapter investigates whether a shock caused by the Allies' strategic bombing towards Nazi Germany propagated through the production network. The empirical results fit the theoretical model well.

This thesis consists of three chapters and covers topics in Energy Economics, Environmental Economics, and Labor Economics. The first chapter estimates the demand and supply functions of rooftop solar panels in California using a rebate program called the California Solar Initiative (CSI). Accurately estimating demand and supply is crucial for evaluating previous incentive programs and guiding future ones. I estimate the demand elasticity in California to be -3.284 and the supply elasticity to be 5.572. My contribution to the literature is a new method of using rebates as a source of exogeneous change to estimate both demand and supply functions simultaneously. I analyze disaggregated data at the Zipcode-month level, and I use a two-part model to incorporate large amounts of Zipcode-months with no solar panel installations. The second chapter is a joint work with Professor Edward Leamer and Jonathan Gu. We study the impact of Pasadena minimum wage on earnings, employment, and number of establishment. We use data from the individual zipcodes within and around Pasadena to conduct analysis. We find evidence of a positive impact of California/Pasadena minimum wages on the earnings of restaurant workers and of other low wage industries. Our model implies that a minimum wage increase of 10% would increase the average quarterly earnings per worker in limited-service restaurants by 8% and in full-service restaurants by 5%. Impact of minimum wage on employment and number of firms are less pronounced. The third chapter studies the impact of a Chinese environmental policy: Two-Control Zone(TCZ). I answer three questions in this study: One, has the TCZ policy been effective in reducing industrial emissions? Two, how has the TCZ policy affected economic activities? And three, how has the TCZ policy affected industry composition? To unpack these issues, I investigate whether a city designated as TCZ improved its environmental performance, city-level GDP growth, and level of foreign direct investment; had higher industrial output; and saw changes in industry composition. Using propensity-score matching to solve the possible endogeneity problem, I find evidence that TCZ cities had 5% lower ambient pollution from 2000 to 2012. During the same period, the TCZ policy caused 8% lower GDP growth and almost 3% less new foreign direct investment in TCZ cities compared to non-TCZ cities.

This dissertation identifies new healthcare policy mechanisms for promoting greater adherence among both physicians and patients to the HIV medication guidelines. The first chapter uses insurance claims data for the sample, compiled by Dr. Arleen Leibowitz, of HIV-infected patients who were insured by Medicare and Medicaid in California between 2007 and 2010 to analyze physicians’ adherence to the clinical care guidelines for prescribing HIV medications. The second and third chapters were written in collaboration with Dr. Sebastian Linnemayr, and describe the impact of the Rewarding Adherence Program (RAP) on promoting greater medication adherence among HIV-infected patients in Kampala, Uganda. The second chapter was originally published in AIDS and Behavior in May of 2015, and the third chapter was published in AIDS in March of 2017.

The first chapter examines the role of physician networks in promoting quality (more adherent) HIV care. Physicians’ non-adherence to clinical care guidelines has been observed for many health conditions, and has particularly damaging repercussions for both HIV-infected patients’ health and for policies to reduce the domestic HIV epidemic. I identify physician peers through shared patients and develop repeated observations of medication regimen and disease monitoring quality across physicians and patients. Using the structure of physician networks to create instrumental variables, I find heterogeneous effects across peer types. Generalist peers have no effect on medication decisions, but a one percent improvement in HIV specialist peers’ medication regimen quality increases generalists’ medication regimen quality by 0.15 percent. Simulations how that improving generalists’ network connections to specialists could provide adherent edication regimens to an additional 2,779 patients in California in 2010, reducing the annual number of new infections by 5 percent. These findings illustrate the potential for network connections to diffuse complex treatment protocols and suggest specific mechanisms for reducing the HIV epidemic, which is disproportionately burdening underrepresented demographic communities in the U.S.

The second chapter describes the influence of behavioral decision biases on patients’ medication adherence. Behavioral economic theory has been used to study a number of health behaviors such as smoking and drug use, but there is little knowledge of how these insights relate to HIV prevention and care. In this chapter, we present novel evidence on the prevalence of the common behavioral decision-making errors of present-bias, overoptimism, and information salience among 155 Ugandan HIV patients, and quantify their association with lower medication adherence. These findings indicate that behavioral economic tools may be used to screen for future adherence problems and to better design and target interventions addressing these behavioral biases and the associated suboptimal adherence.

The third chapter measures the impact of behavioral economic incentives on combating decision biases and improving patients’ HIV medication adherence in sub-Saharan Africa. 155 HIVpositive men and women in Kampala, Uganda aged 19-78 were randomized to 1 of 2 intervention groups or a control group receiving the usual standard of care. Participants in the first intervention group were eligible for prize drawings conditional on attending scheduled clinic appointments; eligibility in the second group was based on antiretroviral medication adherence measured by medical event monitoring system caps. Results from the first nine months of this intervention show statistically significant improvements in the percent of participants who maintain mean adherence rates of 90% or higher in both intervention groups relative to the control. Such behavioral incentives represent a highly cost-effective and scalable mechanism for improving adherence in this region.

We find that veterans of the Union Army who faced greater wartime stress (as measured by higher battlefield mortality rates) experienced higher mortality rates at older ages, but that men who were from more cohesive companies were statistically significantly less likely to be affected by wartime stress. Our results hold for overall mortality, mortality from ischemic heart disease and stroke, and new diagnoses of arteriosclerosis. Our findings represent one of the first long-run health follow-ups of the interaction between stress and social networks in a human population in which both stress and social networks are arguably exogenous.