Nearly everyone involved in environmental regulations believes that compliance with environmental rules is pretty good and that it is enforcement’s job to take care of the rest. You hear this all the time — from regulators, companies, legislators, academics, and environmental advocates.

Both assumptions, that compliance overall is strong, and the work of ensuring compliance should be left to enforcement, are wrong. The data reveal that the rate of serious noncompliance — not just any noncompliance, but violations EPA defines as the most important — is typically 25 percent or more, according to the agency’s data of self-reported and government-identified violations. For many important rules with big health consequences, the serious noncompliance rates for large facilities are 50 percent to 70 percent or even higher. And those are just the ones we know about; for many rules EPA has no idea what the rates of noncompliance are because the regulations don’t include any way to figure that out.

We have also learned that the most important driver of compliance isn’t enforcement, but the design of the regulation. If a rule is structured to set compliance as the default, it can get impressive on-the-ground results without the need for much enforcement. Rules that instead include many opportunities to evade, obfuscate, or ignore will have dismal performance records that no amount of enforcement will ever fix. Robust enforcement is absolutely necessary for any strong compliance program, but enforcement alone will never close the compliance gap created by a poorly designed rule.

Next Generation Compliance, which I launched at EPA during the Obama administration, is a new paradigm for environmental rules. It argues that rules need to be tightly structured to make compliance the path of least resistance. Next Gen rule design acknowledges that in the messy real world where we actually live, equipment fails, people make mistakes, multiple priorities compete for attention and funding, and companies make close — and sometimes nowhere near close — calls in their own favor. And sometimes they just cheat. There is a mountain of evidence that rules only work if they find a way to align private incentives with the public good. These essential truths are the difference between a rule that is great in theory — and one that delivers in real life.

One common misconception about Next Gen is that it is about making rules enforceable. It isn’t. Yes, rules should be enforceable, because that’s a baseline condition that differentiates a rule from good advice. But Next Gen goes way beyond that. It is about creating a structure where the default setting is good compliance — where implementation is strong even if enforcement never comes knocking.

Compliance isn’t a nice-to-have regulatory extra. It’s the part that matters. That’s true for every rule. Standards are fine, but we only get public health benefits from regulations when the regulated companies do what the rules require. When they take steps to control pollution, or conduct the required monitoring, or implement process controls to reduce the risk of catastrophic releases, the standards in the rules translate to real protection. If facilities are doing what they are supposed to do, we have a good chance of achieving clean air and water and reducing our risk of exposure. If they aren’t, we don’t.

Rampant violations have consequences: millions of people living in areas of the country that are not achieving air pollution standards, impaired water quality for half of the nation’s rivers and streams, contaminated drinking water, public exposure to dangerous chemicals, and avoidable environmental catastrophes with health, ecological, and economic damages.

Serious violations aren’t limited to some rules or sectors or company sizes. Widespread noncompliance is the norm across the board: just about every large city has been in consistent and serious violation of Clean Water Act limits on discharge of raw sewage and contaminated stormwater, companies responsible for over 95 percent of the nation’s petroleum refining capacity and almost 50 percent of ethylene oxide manufacturers violated Clean Air Act pollution requirements, over 70 percent of the largest coal-fired power companies violated the obligation to upgrade pollution controls, and over 60 percent of phosphoric acid manufacturing sites were in serious violation of hazardous waste handling requirements.

For many other sectors, the full extent of violation isn’t known, but it doesn’t look good: oil and gas wells with excess emissions of benzene and volatile organic compounds, animal agriculture operations’ compliance with clean water limits on the handling of animal feces that is more than three times the sewage produced by the entire U.S. human population, widespread contamination of surface waters from stormwater runoff, agricultural workers exposed to pesticides through violations of the Worker Protection Standard, noncompliance by small-quantity generators of hazardous waste, and cars and trucks spewing pollution from aftermarket defeat devices. Some of the claimed-to-be-better compliance rates — like drinking water standards and stationary sources of air pollution — are based on data that are demonstrably wrong. For millions of facilities covered by rules about chemical safety, oil spill prevention, asbestos remediation, PCBs, or lead paint handling requirements, EPA has no idea how widespread serious violations are.

The harm from pervasive violations isn’t equally shared. It falls most heavily on already overburdened communities. Contrary to popular myth, it is almost never feasible to remedy through enforcement the ubiquitous violations that result from bad regulatory design. And, as we have recently observed to our dismay, some governments aren’t interested in enforcement anyway. Incorporating compliance drivers in environmental rules is one of the most important things we can do to protect environmental justice communities; it shields them from the harm caused by high rates of violation and is less dependent on the unreliable commitment of regulators.

There is no one-size-fits-all Next Gen strategy for regulations that ensure strong compliance. What works for sophisticated power plant operators isn’t likely to be effective for small and dispersed sources of stormwater runoff. Problems that are measurable and discrete, like emissions from stacks and discharges from pipes, are completely different from tough-to-spot violations of regulations to assure approved chemicals are safe, drinking water is clean, or pesticides are properly applied.

But there are some things we know. Exemptions and exceptions create confusion and off ramps that lead to more violations. Things that aren’t measured produce worse outcomes. The less visible violations are, the more there will be. Standards that require lots of fact-specific determinations or have a big gray zone of applicability provide lots of places to hide, and experience shows companies will use them.

In contrast, simple and clear rules — possible even when the underlying situation is complex — are more likely to be effective. Automatic consequences can work better than requiring government to ferret out problems and impose penalties. Monitoring, measurement, and targeted transparency are the single largest drivers of strong implementation. Innovative use of modern technologies and data analytics hold promise for leap-ahead compliance advances. The standard model in wide use today — creating complex requirements with multiple fact-specific exemptions and exceptions, allowing estimates rather than actual measurement or skipping measurement altogether, relying on trust rather than verification, and requiring government to find the violators and track them down one at a time — is why serious violations are widespread.

Sometimes Next Gen ideas can greatly improve outcomes without changing the overall regulatory approach. But sometimes a Next Gen analysis will make it obvious that the preferred regulatory strategy cannot work. In these cases, there is no plug-in solution; the near certainty of implementation collapse means that regulators have to find another way.

Both roles of Next Gen are illustrated in the following two examples for climate change: methane regulation of oil and gas, where Next Gen ideas could help fix big implementation problems, and energy efficiency as part of a clean energy standard, which Next Gen shows is likely to end up undermining the push for carbon reductions.

Methane released from oil and gas production is a huge source of climate-forcing emissions. Methane in the atmosphere traps over 80 times as much heat as carbon dioxide over its first 20 years, so it packs a big climate punch in the near term. The largest source of anthropogenic methane in the United States is fossil fuel production and its transportation, so any climate strategy needs to control those releases. The Obama EPA promulgated methane rules, the Trump EPA repealed them, and the Biden EPA is now set to move out quickly to address this troublesome problem.

Methane, the main component of natural gas, is brought to the surface during oil and gas production. The gas can vent into the air at the wellhead. It can be released by malfunctioning flares. It can leak from storage tanks, valves, and hatches left open. And it does; the amount of wasted methane released from oil and gas production is depressingly large.

The good news is that we know what to do to dramatically reduce methane releases and cut back wasteful flaring. The technology is available and in use today. The costs are reasonable. As climate challenges go, this is one of the easier ones.

While the technological solutions are comparatively simple, the compliance challenges are not. Oil and gas has all the indicia of a sector where compliance with rules to limit emissions is likely to be bad. There are over a million wells in the country, often in out of the way places. Once a well is completed there are no people routinely on site to keep an eye on failing or leaking equipment. Methane is invisible, so leaks can’t be spotted without specialized equipment. By far the biggest share of leaking methane comes from a comparatively small number of sites: at any given moment 90 percent of the emissions come from just 10 percent of emitters. That concentration of super emitters might normally make the compliance job easier, but not in this case ­— the worst emitters vary over time and are unpredictable. That’s the Next Gen nightmare scenario: huge numbers of sources in out of the way places, with violations that are unpredictable and hard to find. Violations are already common at oil and gas wells. That will get much worse as requirements for methane control are ramped up.

The mismatch between the scope and scale of the compliance problem and government’s ability to either find or fix violations is all too obvious. A handful of regulators for the millions of potentially violating locations makes the standard assumptions that most will comply, and enforcers can take care of the rest, self-evidently untenable here. In this situation — a gigantic number of potential sources at which emissions are collectively huge but individually sporadic, unpredictable, and hard to spot — how can we ensure robust adoption of important strategies for cutting emissions?

One under-appreciated compliance powerhouse in the regulatory toolbox is simplicity. The more special conditions and fact-specific nuance the rule allows, the greater the opportunity to avoid or delay implementation. Repeated experience shows that compliance is less likely for rules with a wide band of compliance gray. Exempting low-producing wells from methane rules, as the owners of those wells propose, risks the same thing. Apart from the reality that low-producing wells are not for that reason less likely to be serious emitters, regulatory exemptions motivate companies to claim to be on the exempt side of the line. If determining the accuracy of such claims requires effort and investigation, a lot of violations — and their accompanying emissions — will slide under the bar.

Innovation is part of the answer for many complex compliance problems. Robust alternative monitoring strategies for oil and gas are being developed at a fast pace and could well be the answer in the long term. A rule can encourage that by motivating everyone to use them. One strategy that might provide an incentive is shifting the burden of proof. If government — or academic experts or NGOs — can provide credible evidence through remote monitoring that a site is a significant emitter, why shouldn’t the company now have to prove it isn’t? And take immediate action if it is? Nothing will motivate leak control more than knowing that an army of experts are looking.

The more automatic things are, the more likely it is that the desired action will happen. Hatches accidently left open are a big source of emissions; why not require hatches that automatically close? The same idea can work to motivate reliable emissions reporting. If the monitoring equipment is not working or a site visit is missed, how about requiring companies to assume that the results were bad, so firms, not the public, bear the burden of misfires? Penalties can likewise be automatic for key violations. The better job rule writers do of making the rule reliably self-implementing, the better the compliance record will be.

There are a host of other promising and low-cost ways to improve methane rule implementation in the real world. All of these ideas come to the fore once we abandon the fiction that compliance magically occurs because standards are written in a rule, or that rule writers can ignore obvious implementation disasters waiting to happen because compliance is someone else’s job.

The second example, a Next Gen analysis of energy efficiency as a part of a clean energy standard, is on one level discouraging, because it reveals that a popular idea for funding needed energy efficiency investments will lead to greater carbon emissions. But the good news is because we know that in advance, we can make another choice. Next Gen isn’t about saying no, it’s about understanding the strategies that won’t work, so we can design ones that will.

Electricity generation is one of the largest sources of climate-forcing pollution in the United States. Every strategy for tackling climate change depends on converting large portions of the economy to electric power, while reducing emissions from power generation. States have shown the way; renewable portfolio standards have been the motivating force behind a big share of the increase in clean generation. A national standard that pushes in the same direction can be the foundation for achieving President Biden’s drive toward 100 percent clean electricity by 2035.

The great news from a Next Gen perspective is that widespread compliance with the national equivalent of a renewable portfolio standard is readily achievable. We already accurately measure the amount of power generated by every source, there are a discrete and limited number of regulated entities, and they are all sophisticated in measurement and data. This situation presents close to ideal circumstances for regulations that achieve near universal compliance.

But here’s the rub: what else counts as “clean,” and will those alternatives actually achieve the promised emissions reductions? Next Gen doesn’t focus on the ideological sides in these debates. It asks just one question: will it work?

One of the most popular entrants in the clean energy sweepstakes is energy efficiency. It promises reduced demand for power by accomplishing the same thing with less power. It creates clean energy jobs. The issue isn’t the importance of energy efficiency. That’s clear. Energy efficiency is an essential part of our work to cut carbon emissions. We need as much of it as possible as fast as we can get it.

The Next Gen issue is the impact on power generation’s carbon emissions if energy efficiency is included in a clean energy standard or in any other regulatory program intended to reduce carbon in electricity generation. Design features can vary but the basic idea of such programs is limiting fossil-fired power generation to a fixed and declining amount of carbon emissions per unit of power. Utilities are allowed to comply with that limit by purchasing qualifying credits. When those credits are from solar or wind power, for example, we know exactly how much electricity utilities are buying and can be 100 percent confident that it is zero carbon.

If they buy an energy efficiency credit, on the other hand, we actually don’t know how much electricity savings, and therefore carbon reduction, utilities are getting. That’s because the nature of energy efficiency and the structural incentives of efficiency programs make determining how much energy is saved extremely difficult. What we do know is that far less energy is being saved than current estimates predict. That’s why including energy efficiency credits in a clean energy standard results in more carbon. The fossil-fuel-fired power plant emits more actual we-know-it-is-happening carbon — in exchange for the hoped for but most likely far smaller carbon savings promised by energy efficiency. Why is energy efficiency such a wild card in carbon accounting?

First is that the impact of energy efficiency is inherently uncertain. The theory of energy efficiency is that the same activity, like heating or lighting a home, is accomplished using less power. How much energy was saved? That’s calculated by comparing what actually happens to the hypothetical world of what would have happened without the efficiency project. If a utility pays me to add two inches of insulation to my attic, what’s the energy savings benefit? The answer isn’t as simple as my energy use before and after. There are hosts of variables that make the comparison highly uncertain: the weather is different; I might turn up the heat because I have more insulation; maybe I also bought an electric car or an energy sucking TV; maybe I would have put that insulation in anyway, without the incentive payment. Actual energy use can be measured, but the hypothetical alternative universe cannot. Even with unlimited measurement resources and the best of intentions, this is irreducibly complex, and it isn’t possible to be certain.

Second, the evidence suggests that the estimates we use to calculate energy efficiency savings are way off the mark. Just about everyone today uses estimates of the benefits of energy efficiency called deemed savings. Such metrics provide a guide for estimating how much energy is saved from installing, say, weatherization measures. Rigorously designed studies have found that actual energy savings fall substantially short of the deemed estimate, in some cases possibly delivering only 25 percent of the promised savings. As is true in so many programs, careful measurement reveals the sometimes gross error of estimates.

Third is another problem that is ubiquitous in Next Gen analysis: the incentive structure for energy efficiency encourages overclaiming of benefits while making it nearly impossible to figure out the truth. Utilities that get more money for programs with greater energy reductions have a built-in motivation to overstate the value of efficiency projects. And they do; a 2012 in-depth study of California utilities found that actual savings were 30 percent to 40 percent less than had been projected and that utilities were systematically overstating the savings. Nearly every participant in energy efficiency has an incentive to overclaim benefits.

And that’s before we even get to the fraud that is inevitable when implementation occurs at millions of locations, companies can make money by cutting corners, and government has virtually no visibility into what’s actually happening.

All of these factors combine to tell us that an energy efficiency credit is both highly uncertain and very likely to greatly overstate its value. So what? Energy efficiency is good, right? Who cares if we can’t be certain about exactly how much energy it saves? We care because by including efficiency credits in a program to cut carbon from electricity generation we set ourselves up for more carbon. We allow a ton of real we-are-certain carbon from a fossil fuel utility in exchange for less than a ton — possibly a lot less — of efficiency offsets. And the more energy efficiency credits utilities buy, the greater their actual net carbon emissions will be. That’s not what we are trying to do.

Lots of market-type ideas for climate suffer from the same implementation shortcoming: by allowing shaky offset credits that will not achieve the desired results in the real world, they undermine the integrity of the emissions-reduction goal. That doesn’t mean we shouldn’t do these projects; it means that we shouldn’t fund them through offset credits that will end up increasing carbon emissions. Other strategies — like an energy efficiency resource standard to prompt investment — can promote the desired funding without undermining carbon reduction.

Regulations must be designed to produce better results in the real world, which is the only place that counts. Next Gen is particularly essential for climate rules, where we cannot afford to fall substantially short of the goal because of widespread, and entirely predictable, implementation fails. For climate, that can make the difference between we have a chance, or we don’t. There are many exceedingly difficult problems to tackle in climate change; we can’t be fumbling on the comparatively simple ones like cutting climate-forcing emissions from oil and gas operations and electric-power generation. We know there are ways to get to a far better outcome. We just have to decide to use them. TEF

The ideas in this article are drawn from the author’s in-depth series “Next Generation Compliance: Environmental Regulation for the Modern Era” posted on the Harvard EELP website — C.G.