Essays in Health and Economic Development

Differences in mortality across space and time are \phenomena worthy of serious attention by economists" (Cutler, Deaton and Lleras-Muney, 2006, p.7). This thesis consists of three self-contained chapters all motivated by this observation and sharing the common themes of health and economic development. The thesis sets out to explore some of the causes of the decline in mortality in Europe and the United States in the period before the advent of modern medicine in the 1940s.

The first two chapters evolve around public health interventions and the decline in tuberculosis (TB) in Denmark and the United States. According to the World Health Organization (2015) TB is a major global health problem ranking alongside HIV as a leading cause of death. In 2014 there were 9.6 million new TB incidences, and even though the disease has been treatable by antibiotics since the 1940s, drug resistant strains now exist. However, while TB is mainly confined to developing countries at present, European countries and the United States also suffered from high rates of TB mortality before the advent of modern medicine. Yet, TB mortality was declining even though treatment was not possible. The TB mortality rate in the United States fell from above 200 in 1900 to circa 60 per 100,000 in 1935 (Cutler and Meara, 2004), and a similar decline was observed in many European countries (Daniels, 1949). Daniels (1949: p.1066) points out that \the most striking fall was in Denmark; the rate there was one of the highest recorded in Europe in 1885, with a mortality of nearly 300, and in 1935 it was below 50". Schelde Møller (1950) ascribes this decline to the policies pursued in Denmark, which were instigated by the National Association for the Fight Against Tuberculosis. Among other public health measures, the association established TB dispensaries locally, which were rolled out differentially across time and space. Chapter 1 sets out to quantitatively evaluate the implications of establishing this particular TB institution in Denmark. The TB dispensaries were designed to prevent the spread of the disease by personalized information provision. A dispensary was little more than a room used for linen, towels, disinfectants, and spittons (Schelde Møller, 1950). Local doctors would refer TB infected patients to the dispensaries, which among other, would instruct TB patients on how they could avoid spreading the disease. Moreover, nurses employed at the TB dispensaries would make home visits of families with an infected member. To evaluate the effect of the TB dispensaries, the chapter use annual difference-in-differences (DD) models. The results reveal that the rollout of the dispensaries across Danish cities led to a 19 percent decline in the TB mortality rate, but no significant impacts on other diseases when performing placebo regressions. Exploiting the fact that the dispensaries were targeting only TB, a second empirical strategy follows a triple-differences setup, and assigns other causes of death as controls. This model compares the development of TB mortality to the control diseases before and after the introduction of a dispensary in a given city. Reassuringly, both strategies result in similar estimates. As for the mechanism, the evidence highlights the dispensaries' preventive actions, such as information provision. At an estimated cost of 68 dollars per saved life-year, this particular public-health institution was extraordinarily cost effective. In addition, small positive effects of the dispensaries on productivity, as measured by annual income per taxpayer at the city level, digitized from historical tax-assessment records, are found.

Chapter 2 studies the immediate and long-run mortality effects of the first community-based health intervention in the world, which had a particular focus on controlling tuberculosis; the so-called \Framingham Health and Tuberculosis Demonstration" in Massachusetts, USA. Funded by the Metropolitan Life Insurance Company, the (Framingham) Demonstration was carried out by the National Tuberculosis Association and has ever since been widely regarded as successful in combating TB at the time. The official evaluation (Monograph No. 10, 1924) used seven pre-selected control communities in Massachusetts to measure whether the Demonstration reduced TB mortality in Framingham during the Demonstration period (1917-1923). This chapter digitizes official vital statistics for Massachusetts cities for the period 1901-1934, extending the number of control cities to evaluate whether the Demonstration reduced mortality during the intervention and in the long run. These data are supplemented by digitized city level total mortality rates and pulmonary TB deaths for cities throughout the United States for the same period. With these mortality data sets the Demonstration is studied more systematically with difference-in-differences estimations, and with methods that take into account that only one unit is treated using the method for inference with few treated units of Conley and Taber (2011), and the synthetic control method pioneered by Abadie and Gardeazabal (2003) and Abadie et al. (2010). Overall, the findings generally question the very positive conclusions in Monograph No. 10 (1924) and others. In particular, the simple DD strategy suggests that while TB mortality decreased significantly during the Demonstration period, these improvements were more than reversed in the 10 years following the Demonstration, with no evidence of sustained reductions in TB mortality as a result of the Demonstration. These numbers are even more dismal when compared to cities across the United States. Yet, when methods used to take into account that only one unit was treated, most evidence suggests that there is no discernible effect on TB mortality. The findings have at least two plausible interpretations: First, the Demonstration might have had only temporary positive effects (during the Demonstration period), which were subsequently completely reversed. Second, the pattern in the data is also consistent with the Demonstration having no significant effect on TB mortality. Importantly, both interpretations imply that the original conclusion regarding the effect of the Demonstration on TB was incorrect. However, the analysis does confirm the observation that the Demonstration had a negative effect on infant mortality during the Demonstration. This negative effect is unlikely to be observed by chance and it even persisted, after the Demonstration ended, from 1924 to 1934.

Chapter 3 is motivated by a crucial question within the literature of health and development; whether health improvements are best achieved through general economic development or specific targeted interventions. This debate goes back to McKeown (1976) arguing that the secular decline in mortality before the advent of modern medicine in the 1940s was largely unrelated to medical innovations or public health interventions. Instead McKeown emphasized improved standards of living, and especially improved diet and nutrition, generated by economic growth and development since the late 18th century. This view has gained support by some, but remains controversial. Contributing to this debate Chapter 3 studies the relationship between nutrition and mortality and the role played by improved nutrition and food supply for the mortality transition. It has proven difficult to find a context in which it is possible to study the relationship between nutrition and mortality econometrically during the historical decline in mortality. This chapter begins to fill this gap in the literature by studying this relationship by exploiting newly digitized data on spending on inmates rations and mortality in United States penal institutions from the mid 19th century to the early 20th century. The main analysis estimates the relationship between rations and the mortality rate using a fixed effects panel model and shows a significantly negative effect of rations on mortality. In the baseline estimate an elasticity of rations on mortality of around -0.43
percent is found, and a conservative estimate suggests that increased rations can explain 26.4 percent of the decline in mortality in the prisons. Further, similar negative effects on TB mortality of increased rations is found, but the results are not driven by TB alone. Using prison level data have several advantages, first and foremost, the convicts have little control over the food available for consumption, thereby excluding one possible source of endogeneity. The results suggest, that although nutrition accounted for a significant part of the decline in mortality in the prisons, the bulk of the decline in mortality remains unexplained by improved nutrition. This suggest that the prison mortality environment was improving for other reasons than the nutritional status of the prisoners.

As evident from the above, the three chapter all have a focus on (TB) mortality and contribute new datasets. They also suggest some joint conclusions. First, Chapter 1 grants some role to public health policies for the historical decline of TB mortality. Yet, Chapter 2 demonstrates that not all efforts were successful. Second, Chapter 1 shows that the mortality reductions led to higher productivity and income, whereas Chapter 3 is consistent with the view that general economic development via its effect on increased nutrition is also a factor in the mortality transition. This strongly suggests that public health interventions also could work not only through their direct preventive effects, but also through increased nutrition.