Essays in health and development economics

Abstract

This thesis is a compilation of three empirical studies exploring significant but underexamined health and development challenges of the late 20th and early 21st centuries in South Asia. Chapter One investigates the effects of the expansion of ultrasound technology throughout India in the 1980s on the childbearing decisions of parents and the marriage market dynamics of exposed children. While ample work has documented the relationship between access to sex selection technology and heavily male-skewed child sex ratios, we know little about how such exposure translates into later life marriage market outcomes of children in highly sex-skewed regions, nor about how parental choice regarding sex selection is affected by such shifts in their children's marital prospects. I build on a theory proposed by Edlund (1999) that, in environments where hypergamy is practiced and parents derive utility from married children, a male-skewed sex ratio can generate a permanent female underclass. By examining the relationship between the child sex ratio of couples of childbearing age and that of their contemporaneous marriage market, I offer evidence that parents do indeed internalize the marriage prospects of their unborn children and adjust their use of sex selection technology accordingly. Importantly, this adjustment occurs significantly more amongst poor families than wealthy families. By exploiting spatial and temporal variation in exposure to ultrasound technology, I then examine the implications of such socioeconomically skewed ultrasound use on the marital outcomes of children in regions with high ultrasound access. I find that, relative to her unexposed counterpart, the average exposed married female has significantly poorer health and less education; there exists a wider marriage and education gap between herself and her husband; and she reports lower autonomy, less decision making power, and more abuse, among other bargaining outcomes. While existing literature suggests that scarcity of females in a marriage market should increase their bargaining power, I offer evidence to the contrary in this nationwide setting of endogenous and socioeconomically stratified sex selection. This exercise underscores the intergenerational welfare consequences of poorly regulated access to sex selection technology: not only upon the millions of 'missing women' lost to sex selection, but upon surviving females as well. Chapter Two explores the impact of a 1999 public health campaign in Bangladesh, which sought to protect millions of individuals from exposure to arsenic-contaminated water, on infant and child mortality. The study was motivated by the dearth of literature on the effects of arsenic exposure on children (whereas its effects on adults, often manifested in the cancer arsenicosis, are well known). It quickly evolved into an examination of the unintended consequences of a highly influential but poorly planned public health campaign. Exploiting the local random nature of arsenic contamination of groundwater in Bangladesh, paired with the timing of child births and thus exposure to such contaminated water, we find that households in which children were exposed to arsenic for a shorter duration (because the household responded to the health campaign by switching away from arsenic-contaminated groundwater sources) in fact experience significantly higher rates of infant and child mortality relative to their counterparts. We present evidence that this unanticipated rise in mortality is due to the quality of alternatives that a switching household faced: households had to choose between arsenic-laden but easily accessible shallow tubewell water, which was protected from fecal bacteria; arsenic-free and easily accessible surface water, which was heavily exposed to fecal bacteria; or distant and inconvenient potable water, which was more likely to be exposed to bacteria at the point-of-use. As bacterial contamination is a leading cause of infant and child death in Bangladesh, we argue through a series of exercises that this is a likely driver of the rise in mortality rates amongst young children whose families switched away from arsenic-contaminated tubewells. In determining their water source, households were essentially trading off arsenic exposure and the resulting rise in old-age mortality with bacterial exposure and the resulting rise in the mortality of their young. The study motivates caution in the execution of large-scale public health and behavioral change campaigns when alternatives to the discouraged behavior are poorly understood. While my first two chapters investigate household health behavior, a demand-side component of the healthcare market, the next chapter explores a critical player on the supply side. Chapter Three studies the impact of a nine-month generalized training program on the knowledge and performance of private informal healthcare providers in West Bengal, India. These providers, colloquially referred to as "quacks" and described here as "informal providers" (IPs), constitute nearly 80% of the Indian healthcare provider market. However, none possess medical degrees and few have any formal certification to practice medicine. They have been the focus of considerable debate in recent years, with many pushing for their elimination while others propose their integration into the public healthcare system. To inform the debate, it is important to understand whether the quality of healthcare provided by IPs can be improved sufficiently for effective and welfare-increasing integration. The training program examined in this study was the first of its kind to be rigorously evaluated for its impact on IP knowledge and quality of care. We employ a randomized controlled trial (RCT) design, in which we randomly assigned 152 IPs to treatment and 152 IPs to control. Treatment IPs were invited to attend the program, which was taught by certified doctors and consisted of two two-hour classes per week over nine months. Endline data was collected twelve to fourteen months after the start of training. Standardized patient data, corroborated by clinical observations, demonstrate that those IPs offered the program spent significantly more time with their patients, completed a more thorough set of history questions and examinations, and provided more effective treatments. However, we see no shift in the frequency with which they practiced polypharmacy nor the dispensation of unnecessary antibiotics, two harmful practices which plague both the private and public healthcare system. We conclude that training offers a low cost, highly effective method to improve the quality of care delivered by IPs, but that deeper knowledge failures or misaligned incentives may be driving practices such as polypharmacy, for which training may not be a sufficiently powerful intervention.