The Effects of Clean-Milk Regulations on Urban Infant Mortality, 1900-1910

It is well-documented in the economic history literature that high levels of mortality began to decline in the United States around the turn of the twentieth century and that the substantial “urban penalty” (excess mortality associated with living in urban versus rural areas) in U.S. cities disappeared by the 1940s.  Existing work documents these population-level declines, but few studies attempt to explain differences in mortality declines by population subgroup.  In this paper, I begin by documenting that city-level infant mortality rates declined in the largest U.S. cities by around 25% during the time period 1900-1910, and that this decline was larger in both absolute and percentage terms than the decline in the overall (all ages) mortality rate.  In the remainder of the paper, I explore one potential explanation for this observed phenomenon.  

Modified cow's milk (cow's milk diluted with the addition of sugar and water) was the main artificial infant food prior to the widespread availability of formula.  Importantly, cow's milk also had the potential to carry a host of infectious diseases that were detrimental to infants, including scarlet fever, typhoid, and diarrheal diseases.  Historical evidence suggests that cow’s milk was widely used for artificial infant feeding during 1900-1910.  This fact -- coupled with documented declines in the practice of wet-nursing, evidence on short weaning periods in the early 20^th century, and evidence on the dual-use of breast milk and artificial food for infants during this time period -- suggests that cow's milk played a central role in the infant diet during the time when infant mortality rates declined rapidly.  This same time period of 1900-1910 was also a time of expansion and action for municipal government.  Regulation of the food supply was one area in which city governments experimented widely.  Despite the coincident timing of city-level regulation of the food supply and the decline in infant mortality, the contribution of milk supply regulations to these observed infant mortality declines remains unexplored.  

In this paper, I analyze the causal impact of municipal “clean milk regulations” – i.e. city-level regulations aimed at improving the safety of cow’s milk, such as the inspection of dairy farms and the collection of milk for laboratory testing – on infant mortality between 1900-1910.  Exploiting plausibly exogenous variation in the introduction of clean milk regulations at the city level, I find that these interventions reduced infant mortality rates by 14-21%.  I find no evidence that clean milk regulations had any effect on overall mortality (all ages) or on the mortality of older children (ages 1-4), but this is consistent with evidence about the limited consumption of cow's milk for these age groups at the time.  My findings are robust to a number of different model specifications and are further reinforced by several falsification exercises.  The findings in this paper illustrate the importance of subgroup analysis in explaining population level phenomena.