When the Volkswagen diesel emissions scandal erupted recently—resulting in a precipitous drop in the company’s stock price, the resignation of its CEO, a $14.7 billion judgment against the company, and extensive environmental harm—we were reminded of an eerily similar environmental scandal from about 20 years ago. In 1998, the US Environmental Protection Agency (EPA) brought suit against the leading manufacturers of heavy-duty diesel engines for failure to reduce emissions as required under the law. The suit resulted in manufacturers agreeing to pay $83 million, the largest settlement of an enforcement action that EPA had ever collected at that time.

If those who fail to learn from the past are destined to repeat it, regulators should try to learn from their earlier dark moments of industry evasion. With our recent article, The Law of the Test: Performance-Based Regulation and Diesel Emissions Control, we offer an in-depth retrospective on the EPA’s regulation of heavy-duty diesel truck emissions. Our study not only calls attention to how the EPA could have readily anticipated the Volkswagen scandal, but it more generally reveals several unacknowledged but critical limitations of performance-based regulation, the type of regulation underlying both diesel emissions scandals and used in a variety of other areas of regulatory policy.

As its name suggests, performance-based regulation requires firms to attain a specified level of performance. Because it specifies the ends that firms must achieve, and leaves it up to them to select the means, performance-based regulation in theory results in more cost-effective outcomes. That same flexibility also promises to achieve regulatory goals without stifling innovation. For these reasons, regulatory scholars and analysts have widely touted the superiority of performance-based regulation.

But the success of performance-based regulation depends on the regulator knowing whether a firm’s performance meets its standards. As our study of EPA’s diesel emissions regulation shows, designing an effective way to test performance can prove much more challenging than it might appear—and can lead to much more problematic outcomes than the proliferation of policy recommendations in favor performance-based regulation would suggest. Part of the difficulty stems from the fact that, whatever test regulators come up with, firms face incentives to try to circumvent the required tests or to structure their actions in such a way as to meet the letter of the test but that better serves the firm’s interests than society’s.

For nearly seven years, before its deception was uncovered, Volkswagen followed the same playbook that manufacturers of heavy-duty diesel engines followed in the 1990s. In both cases, manufacturers found a way to program their engines so that they met the legally mandated performance test. But once the test was over, the engines automatically shifted into a different operating mode designed to maximize power and fuel efficiency, while polluting at higher levels. The Volkswagen engines met environmental performance requirements when hooked up to laboratory emissions monitors, but emitted pollution at levels more than 40 times the regulatory standard when driven on the road.

Two decades earlier, heavy-duty diesel engine manufacturers programmed their engines to meet EPA’s performance requirements when operating at specific speeds and torques; however, once on the open road for an extended time, these same trucks would spew more than a million excess tons of nitrogen oxides (NOx)—a key chemical in the creation of harmful ozone pollution. According to one estimate, the diesel truck emissions that escaped regulatory control equated to roughly the volume that would have been produced by adding 65 million more cars per year to the nation’s road. Ultimately, these excessive emissions may have contributed to more than 4,100 premature fatalities.

Performance-based regulation is clearly no regulatory panacea. Although this flexible form of regulation might sound less cumbersome, complex, and costly than those standards that prescribe specific means, EPA’s experience with diesel emissions control offers a much more circumspect picture. Performance-based rules can only be as effective as the tests that regulators devise to verify performance. They depend on vigilance by regulators who search for efforts by regulated firms to circumvent the letter or spirit of those tests. Only by seeking to anticipate alternative ways that firms may respond to testing protocols, and then engaging in ongoing efforts to search for signs of evasion, can regulators ensure that performance-based regulation will work in practice as it should in theory. And only by learning the lessons from past experiences can regulators avoid yet another diesel emissions scandal in the future—or similar failures in other areas of regulation.

Cary Coglianese is the Edward B. Shils Professor of Law at the University of Pennsylvania Law School and a Professor of Political Science at the University of Pennsylvania College of Arts and Sciences. Jennifer Nash is the Director of the Business and Environment Initiative at Harvard Business School.