Regulation shapes the American economy, but measuring the true cost and benefit of regulation is difficult. Patrick McLaughlin, a research fellow at Stanford’s Hoover Institution and visiting research fellow at the Pacific Legal Foundation, says that without clear, consistent data, the policy process is blind.
McLaughlin holds a doctorate in economics from Clemson University and has served in senior research roles at the Mercatus Center, the Federal Railroad Administration, and the Regulatory Studies Center. He is the creator of the RegData and QuantGov projects, which apply machine learning to translate the millions of words in statutes, regulations, and guidance documents into quantitative metrics. His methods have been used to compare regulatory burdens across states, track changes over time, and assess their economic impact.
“I was initially focused on environmental economics, but the studies I read were always about the design of regulation,” he says. “The thing that really mattered was: did it work or did it not? The variation in cost per life saved was all about the design of the regulation itself.” That insight led him to recognize “a big data gap” in the ability to measure regulation, which has defined his career.
Creating objective, repeatable metrics is central to his approach. “My method is to measure things in the same sense that we measure other economic variables like employment,” he says. “There’s hundreds of thousands of pages… We turn all that into numbers.” He counts restrictive terms like shall, must, and may not, matches them to affected industries, and produces annual scores showing how regulated each sector is.
McLaughlin says that the real economic weight lies in the accumulation of old rules rather than new ones. “It’s the old stuff that matters a lot more in terms of the economic effect,” he says. Yet most policy attention focuses on new “major” rules with estimated annual costs over $100 million, leaving 97 percent of rules without thorough cost-benefit review. “We do very little to look back at old rules and figure out what we want to keep.”
He sees examples worth emulating in states like Virginia, which recently announced a 25 percent cut in its regulatory burden. “They put together a very streamlined guide on how to do cost-benefit analysis,” he says. “It doesn’t require a PhD in economics. You lose some details, but you get consistency.” Virginia also reviewed third-party standards incorporated into state rules—such as international building codes—and carved out requirements that added cost without significant safety or public benefit.
McLaughlin’s research shows that regulatory accumulation slows GDP growth by nearly one percentage point per year. Holding the 1980 level of regulation constant, he estimates, “we would have had an economy 25 percent larger by 2012.” Regulations also contribute to consumer prices, with one study finding that a 10 percent increase in total regulation raises prices by about 1 percent—effects that fall hardest on low-income households.
According to McLaughlin, cost-benefit analysis itself can be manipulated. Changes to federal guidance under the previous administration allowed analysts to count benefits outside the United States, especially in climate rules, inflating benefit estimates. “Anyone who wants to make benefits look larger, they have a way of doing it,” he says. In his view, the question Congress should ask is simple: “Did the regulators accomplish the goal that Congress put in front of them? If they did, they should be done.”
McLaughlin sees artificial intelligence as a powerful tool for reform. “Regulation is kind of the perfect area for the application of AI,” he says, given the volume of text in statutes, regulations, guidance, and incorporated standards. Virginia is piloting AI programs that compare regulations to their authorizing statutes to flag possible “gold-plating” and highlight redundancies or obsolete provisions. “It’s not necessarily a decision rule. It’s just saying, here’s an area where the regulations seem to be going a lot farther… Humans need to take a look.”
With $100 million now allocated to the Office of Management and Budget for improving regulatory processes, McLaughlin says building such AI systems should be a priority. “That saves so much human work. And now humans can focus on the decision points,” he says.
