Performance standards affect decisions and investments made by businesses. One firm might invest in R&D to improve energy efficiency and comply with an appliance standard, and another may finance a wind farm to meet a renewable portfolio standard, which requires a minimum share of electricity generation from renewable resources.

Business investments take time and come with a degree of risk. For example, it may take an auto manufacturer years, and cost tens or hundreds of millions of dollars, to succeed in research that drives markedly higher efficiency. Without long-term policy certainty, it might be too risky for the manufacturer to invest in that research. The manufacturer may believe there is a low chance of succeeding within the necessary timeframe or may not want to risk the upfront investment if they believe the policy has a high likelihood of changing before research investments pay off.

If a performance standard takes effect too soon after its adoption, businesses may be unable to meet the standard as cheaply or with products of as high a quality as would have been possible if they’d had a few more years to prepare.

Just as businesses should not be hit suddenly with tighter standards than they were expecting, performance standards should not be delayed or relaxed at the last minute. Delaying or weakening the standards has the effect of benefiting laggards (companies that did not sufficiently invest in R&D and other preparations to meet the standard) while hurting companies that made the necessary investments to meet the standards in good faith. Companies that worked to meet the new standards rely on the standards to get a return on their R&D investment, because they will be able to offer appealing products that meet the standards at lower cost than competitors. If they are unexpectedly forced to compete with competitors’ older, inefficient, but cheap products, they may not see the financial rewards they were expecting.

As a rule, in a given industry, standards should be known at least as many years ahead as it takes to complete a full product revision (that is, developing a new product, retooling factories to produce it, updating marketing materials, etc.). This timeframe varies by industry but generally is at least several years and may be as long as 10 years. Knowing standards even further ahead (say, two or three product revisions) might be even better.

Because of the uncertainty inherent in technological progress, it may not be possible to strictly specify what performance level is reasonable for a given technology to be required to meet 10 years in the future. There are two solutions to this problem.

The first solution is to base the new standard on a subset of the products available in the marketplace. For example, a standard might be set to the level achieved by the highest quintile of existing products, say, three years prior. This guarantees the standard can be met with available technology, because there existed commercialized products three years ago that met the standard.

The second option is to set a known schedule of when performance improvements will take effect, extending many years into the future, but the magnitude of the improvements is set closer to the date when they go into effect by a regulatory agency, with input from affected industries. This approach has the benefit of increased flexibility, but it reduces the certainty associated with a schedule of known improvements, and it makes the performance standard vulnerable to political interference and regulatory capture. Therefore, it is best used only when necessary to achieve consensus for tighter standards

A performance standard must have a mechanism for automatic tightening, so it does not become stagnant and ineffective. This might continue until the technology starts to approach fundamental limits (e.g., theoretical efficiency limits imposed by thermodynamics), is replaced by another technology entirely (e.g., gasoline cars may be replaced by electric cars), or begins to saturate the sector (e.g., renewable electricity as a share of total electricity generation).

It is valuable for the improvement mechanism to be built into the law or regulation promulgating the standard. This means that the law specifies a requirement for tightening the standard over time. Adjustments to the standard may happen according to a known schedule (say, every three or four years), and the law may specify the amount of the adjustment or, if uncertain, the manner in which adjustments are to be determined (see the preceding section for two ways to do this). Alternatively, a standard may simply require a fixed annual percentage improvement.

Building improvement into the standard helps prevent long periods of stagnation that can result when legislative bodies or even regulatory agencies must opt to tighten the standard with legislation or rulemaking. In the United States, the fuel economy standard for automobiles reached 27.5 miles per gallon in 1985 and did not rise above this level until 2011, costing the U.S. hundreds of billions of dollars.

Japan’s Top Runner program has an excellent mechanism for continuous improvement. The Top Runner energy efficiency standards cover many different categories of products, each reviewed on a known schedule. At the time of review, the efficiency of the most efficient product on the market is selected as the standard that must be met by all manufacturers within several years. The policy also accounts for additional potential technological improvements by slightly increasing the standard above the value of the most efficient product. To increase flexibility, manufacturers must meet the standard on the basis of weighted average of all their shipments of products in a given category. A manufacturer might comply by selling only products that meet the standard, or a manufacturer might offer a product that fails to meet the standard, as long as the manufacturer also sells enough products that exceed the standard to bring the weighted average efficiency of all shipments above the standard. As a result of the Top Runner program, passenger vehicles improved in efficiency by 49% from 1995 to 2010, refrigerators by 43% from 2005 to 2010, and TV sets by 61% from 2008 to 2012

Standards should be set based on desired performance outcomes (e.g., fuel efficiency, pollutant emissions) rather than mandating the use of specific technologies. For example, rather than mandate the use of a particular type of particulate filter on trucks, impose a maximum level of particulates that may be emitted by a truck per kilometer it travels.

Specifying performance outcomes rather than mandating specific technologies gives companies the maximum leeway to innovate and apply different solutions, helping to achieve the desired outcomes at least cost (and possibly using novel techniques or technologies that were not anticipated at the time the regulation was written). It also reduces the burden on policymakers to keep up with technology developments in all fields for which they have adopted standards.

There are exceptions to this principle. When a technology is not fully mature but is ready for commercialization and shows promise to achieve very good results in the future, it may be worthwhile to create a separate performance tier or carve-out for the promising technology. This may provide it space in a crowded market, allowing it to achieve economies of scale and to develop to the point where a separate tier is no longer necessary. A good example of this is a renewable portfolio standard (RPS).

A typical RPS is technology-finding: It specifies a fraction of electricity that must come from clean sources (and sometimes efficiency), and utilities determine the most cost-effective way to meet the standard. For instance, a utility may choose to build some wind turbines, build some solar panels, and start a demand-side efficiency program.

However, some RPS contain a carve-out for less-developed technology, such as offshore wind, in the form of a minimum percentage of energy that must come from that source. Even though it is not technology-finding, a carve-out may be justified in the short term, while offshore technology develops and prices come down. In time, a carve-out for offshore wind should become unnecessary

While some manufacturers will voluntarily create cutting-edge products that significantly exceed performance standards, others will do as little as possible while still complying with the law. This can lead them to seek out loopholes, ways of complying with the letter of the law while undermining its purpose.

If a standard is complex and intricate, and if it makes many distinctions between different types of equipment based on various design features or use cases, this leaves many openings where companies can potentially game the law. To insulate a standard against gaming and loopholes, it is helpful to write the standard to maximize simplicity and clarity and to state in broad terms the targets that must be achieved rather than making exceptions or different rules for equipment with different features.

For example, the U.S. Environmental Protection Agency defined different vehicle fuel economy standards for “cars” and “light trucks.” Light-duty vehicles could be classified as light trucks based on minor design features (such as a vehicle that is “available with special features enabling off-street or off-highway operation and use”). Manufacturers were able to exploit the two different standards by making the minimum number of design changes necessary to allow them to classify many of their cars as light trucks, even though these vehicles were marketed and usually used for on-road, personal transportation. This led to a boom in SUV sales (classified as trucks) in the 1990s, until the light truck fuel economy standard started increasing in 2005.

Another notorious loophole is found in sloppily designed testing standards. The Volkswagen diesel emission testing procedure was specifically designed to reduce pollution when a car was being tested, but not otherwise. Better testing, emulating real-world conditions, probably would have prevented this problem.