In 1974, the U.S. Congress passed the Safe Drinking Water Act, with the intention of ensuring that all Americans had access to clean and uncontaminated drinking water. Since then, the government has spent more than $2 trillion enforcing the law. But there have still been enough disasters — most notoriously in Flint, Michigan, and Jackson, Mississippi — and enough violations that many people remain suspicious of the quality of their drinking water; in every Gallup poll since 1990, Americans have listed drinking water pollution as their top environmental concern.

And, in fact, almost every published study of water pollution over the past few decades has shown that the number of water systems found to be in violation of the SDWA has increased dramatically. But is that actually because water pollution has gotten that much worse?

A group of researchers at the National Bureau of Economic Research that included David Molitor, an associate professor of finance at Gies College of Business at the University of Illinois Urbana-Champaign, wanted to find out what was behind that trend.

“It's actually not clear why the number of violations has gone up,” says Molitor, who also serves as director of the Health Care Research Initiative at Gies Business. “Is that because pollution is getting worse, like there's more contamination in the water? That would be one possibility. But another possibility is just that drinking water standards have gotten tighter over time.”

But when Molitor and his colleagues — David A. Keiser (University of Massachusetts Amherst), Bhashkar Mazumder (University of California, Irvine), and Joseph S. Shapiro (University of California, Berkeley) — went looking for answers, they ran into a major problem: the data they needed didn’t exist.

There are nearly 150,000 public drinking water systems in the United States, Molitor explains. Those systems test their water and report to state regulators, and when they fail a pollution test, the state reports it to the federal government. But states generally only report the violation, not the underlying pollution measurements. More crucially, no one reports the numbers for a system that doesn’t have a violation.

“It's not free, you know, collecting data, measuring things well, reporting them,” Molitor says. “That data infrastructure is an investment. But I think it’s really important for public health to understand important issues like this. Like, is our drinking water clean?”

So the researchers set about assembling a data set themselves. It was a long process that involved scraping state websites, intense correspondence with state officials, and many, many FOIA requests. It took nearly a decade. In the end they were able to collect data from 48 states — South Dakota and Arkansas were the two holdouts. In addition, they collected data about the improvements made to various drinking water systems thanks to subsidized federal loans and data linking health outcomes of people over 65 to the quality of their drinking water. Overall, they compiled the most comprehensive set of records ever created about drinking water safety in the United States. They released the results in a working paper titled “Water Works: Causes and Consequences of Safe Drinking Water in America.”

And what did they find?

US drinking water pollution has declined

“This paper is a lot of good news,” Molitor says. In particular, there are three distinct pieces of good news.

First, it turned out that, in answer to the researchers’ original question, the increase in drinking water violations was driven, at least in part, by tighter standards. Another possibility is that reporting has become more consistent. Overall, pollution concentrations have fallen by about 50 percent.

Second, Molitor and his colleagues compared pollution levels in drinking water systems that had received federal loans to finance improvements to their infrastructure with drinking water systems that hadn’t. They found that the improved systems did, in fact, experience pollution reductions relative to systems that had received no help at all. And, crucially, the loans weren’t prohibitively expensive: if every American contributed about $46 annually, it would be possible to eliminate pollution violations.

Finally, they looked at the health of people over 65 who lived in communities that benefited from improvements to their water systems. The researchers chose this particular cohort because Medicare administrative data is much more precise than other data sets about pinpointing exactly where people live. This meant the researchers were able to tell where their drinking water came from.

Sure enough, Molitor says, “when we look at the older adult population for whom we have a much better matching of who lives where, we do see evidence of systems that receive a loan, that population health, as measured by annual mortality rates, begins to gradually improve. We saw a substantial improvement about eight years after the drinking water system was in place.”

When Molitor and his colleagues finally assembled their data and began digging in, they weren’t sure what they were going to find. “It could have gone either way,” Molitor says. But they were glad to find proof that drinking water has been improving over time.

“I think it's not terribly surprising given that people who manage drinking water systems know that it's important to the public to have clean water,” he says. “So I'm not surprised that people are investing and making improvements on this, since it's such a salient and important issue, even if the data themselves are not always salient and apparent.”