This paper considers a largely unknown public-health practice in the United States: the use of lead pipes to distribute household tap water. Municipalities first installed lead pipes during the late nineteenth century. In 1897, about half of all American municipalities used lead water pipes. Using data from 1900 Massachusetts, this paper compares infant death rates and stillbirth rates in cities that used lead water pipes to rates in cities that used non-lead pipes. In the average town in 1900,

more »

... the use of lead pipes increased infant mortality and stillbirth rates by 25 to 50 percent. However, the effects of lead water lines varied across cities, and depended on the age of the pipe and the corrosiveness of the associated water supplies. Age of pipe influenced lead content because, over time, oxidation formed a protective coating on the interior of pipes. As for corrosiveness, acidic water removed more lead from the interior of pipes than did non-acidic water. Consequently, infant death rates and stillbirth rates in Massachusetts towns employing old lead lines, and non-acidic water supplies, were no higher than in towns employing non-lead pipes. But in cities using new pipes and distributing acidic water, lead pipes increased infant mortality rates and stillbirth rates three-to fourfold. This paper considers an important, but largely unknown, public-health practice in the United States: the use of lead service lines to distribute household tap water. Municipalities first installed lead service lines during the late nineteenth and early twentieth century, and in some places, the original lead pipes are still in use. In 1897, about half of all municipalities had water systems based exclusively, or in part, on lead service lines; the remaining cities used iron pipes or metal pipes lined with cement. The use of lead pipes was positively correlated with city size. In 1900, the country's five largest cities-New York, Chicago, Philadelphia, Saint Louis, and Boston-all used lead pipes to varying degrees. For cities with populations greater than 30,000, more than 70 percent used lead lines exclusively or in combination with some other type of metal. 1 Although it is now well known that ingesting or inhaling even small amounts of lead can have severe health effects, particularly for young children, 2 little is known about the consequences of these lead water lines. 3 Accordingly, this paper exploits an unusually rich set of data sources to identify the health effects of lead water pipes in turn-of-the-century Massachusetts. With these data, it is possible to compare infant death rates and stillbirth rates in cities that used lead water pipes to rates in cities that used non-lead pipes. The central findings of the paper are as follows. In the average Massachusetts town in 1900, the use of lead pipes increased infant mortality and stillbirth rates by 25 to 50 percent. However, the effects of lead water lines varied across cities, and depended on the age of the pipe and the 4 One how age of pipe and corrosiveness of water influences the amount of lead in drinking water, see Howard (1923) ; Quam and Klein (1936) ; United States Environmental Protection Agency (2000) ; Wisconsin Department of Natural Resources (1993). 2 corrosiveness of the associated water supplies. Age of pipe influenced water-related lead ingestion because, over time, oxidation formed a protective coating on the interior of pipes and limited the amount of lead that dissolved into the water. As for corrosiveness, acidic water removed more lead from the interior of pipes than did non-acidic water. 4 Consequently, infant death rates and stillbirth rates in Massachusetts towns employing old lead lines, and non-acidic water supplies, were no higher than in towns employing iron or cement water lines. But in cities using new pipes and distributing water derived from acidic sources, lead pipes had enormous effects: they increased infant mortality rates and stillbirth rates three-to fourfold. Such estimates might sound high, but as will be made clear below, they are quite plausible when one considers just how much lead could be dissolved into household tap water as a result of lead service lines. For example, samples of household tap water in Lowell-a city that used lead pipes-contained average lead levels nearly 200 times greater than those currently allowed by the Environmental Protection Agency. One potential concern with the estimates presented here is that they could overstate the impact of lead water mains on the probability long-term survival if lead water mains killed off only the weakest and most vulnerable of all infants. Water-related lead exposure, in other words, might have been fatal only for infants who would have inevitably succumbed to some other environmental insult or childhood disease, such as scarlet fever, diphtheria, measles, or whooping cough. Although there is evidence that water-related lead poisoning affected the most vulnerable infants disproportionately, and that the use of lead water lines actually reduced death rates for some childhood diseases, these effects were small. After controlling for this selection issue, lead water lines still appear to have had a substantial effect on infant mortality in towns with acidic water 5 Of course, this observation cuts both ways and might also raise doubts about the estimated effects reported here. 6 On exposure pathways emphasized in recent studies, see Charney, Sayre and Coulter (1980); Lanphear and Rogham (1997) ; Markowitz and Rosner (2000) ; Trepka et al. (1997); Wolpaw Reyes (2003) ; and Xintaras (1992). 3 supplies, or with relatively new lead pipes. These results are important in at least four ways. First, their magnitude is significant: a 25 to 50 percent increase in infant and fetal mortality rates is a large number. Second, the results are important because they are surprising. Noone in 1900, not even the handful of public health officials who first spoke out against the use of lead pipes, thought the adverse health effects of lead water lines were as large as the findings here suggest. 5 As late as 1917, the engineers who designed urban water systems were dismissing claims that lead services could cause widespread lead poisoning, and continued to advocate the use of lead pipes because of their attractive mechanical qualities. And today, while the dangers of lead are widely appreciated, nearly all health scholars and regulators focus on other pathways of exposure, such as paint, household dust, dirt, leaded gasoline, and industrial pollution. 6 Third, during the early twentieth century, the federal government, as well as many state governments, started enacting laws prohibiting and/or severely limiting the use of lead service lines and lead based solders in household plumbing. The results here suggest that the gross social benefit of such laws was very high. Although it is not possible to precisely estimate the net benefit of such laws because there are no data on the costs of enforcement and compliance, it will be made clear below, that there were low-cost substitutes for lead pipes available in 1900, notably iron and cementlined pipes. While not as malleable or as durable as lead, iron and cement-lined pipes also did not poison the water they carried. Regulations governing the use of lead in plumbing fixtures are, as 7 See, for example, the analyses of Buchanan and Tullock (1962) and Downs (1957). 8 Relevant citations are provided in the concluding section of the paper. 4 regulations go, fairly obscure. There is a tendency among students of public choice to think that the more obscure the regulation, the less likely it is reflect the broader public interest, and the more likely it is to represent the lobbying of narrow special interests. 7 Laws governing the use of lead pipes certainly appear to be a counter-example to this general way of thinking. Finally and most important in terms of motivation, the early-twentieth-century United States witnessed dramatic improvements in life expectancy, particularly for urban-dwelling populations (Haines 2001) . It is widely accepted that most of these improvements were driven by reductions in infant mortality, which in turn, were promoted by better nutrition and improved public health services that protected infants and young children from infectious diseases, particularly deaths from infantile diarrhea. 8 This paper suggests that the latter part of this story needs to be revised: reductions in environmental lead exposure appear to have also played a significant role in promoting healthier and more robust children, who were better able to withstand the bacterial and viral insults so common in early twentieth century American cities. Of studies that look at infant mortality and mortality transitions in the United States, Preston and Haines (1991) is probably the best known and most important. In their analysis of Census data from the turn of the century, Preston and Haines emphasize the importance of infectious diseases; they, like most authorities, scarcely mention lead poisoning as a cause of infant mortality (see, generally, Preston and Haines 1991, pp. 3-48).