Tuesday, October 31, 2017

Parsing Trump's Deregulation Plans

Many people have visceral but opposite reactions to the word "regulation." Some have an immediately positive reaction to almost any mention of regulation, in a belief that it is likely to be necessary corrective. Others have an immediately negative reaction, in the belief that it is likely to be a wasteful and perhaps even harmful overreaction. Me, I'm just a wishy-washy guy who thinks that some regulations can be useful, while others are misguided.  On the off chance that there are a few more like me out there, how should we be reacting to the Trump administration's deregulation agenda?

Ted Gayer, Robert Litan, Philip Wallach provide an overview in "Evaluating the Trump Administration’s Regulatory Reform Program" (published by the Center on Regulation and Markets at Brookings, October 2017). For a discussion of potential benefits from regulatory reform, one of the first reports from the Council of Economic Advisers in the Trump administration is "The Growth Potential of Deregulation" (October 2, 2017). I'll draw on both reports here.

Gayer, Litan, and Wallach start off this way: "In his first week in office, President Trump issued Executive Order 13771, which aims to `manage the costs associated with the governmental imposition of private expenditures required to comply with Federal regulations.' It requires that `for every new regulation issued, at least two prior regulations be identified for elimination, and that the cost of planned regulations be prudently managed and controlled through a budgeting process.'” To be clear, there are two separate ideas here: the two-for-one reform and the "budgeting process" reform. Versions of both of these proposals have been implemented in other countries, so one can look to that experience.

But before digging into those details, it's worth noting that this "regulatory reform" or "deregulation" effort is in some ways quite different from the two main types of policy changes that have previously gone under the name of "regulatory reform." As one example, the "deregulation" that happened in the late 1970s and into the 1980s. The "deregulation" of airlines, banking, trucking, railroads, and some other industries around that time involved removing rules that involved setting prices and limiting competition. Two-for-one and a budgeting process for regulations is not this kind of deregulation.

The other main kind of regulatory reform, which dates back to the 1970s but has been pursued by each president since then, has typically involved requirements that cost-benefit analysis be carried out before certain kinds of rules are implemented. This may seem like an obvious step, but as Gayer, Litan, and Wallach point out, "many regulatory statutes—those that authorize or compel agencies to issue rules in the first place—do not permit agencies to balance benefits against costs, or effectively limit their ability to do so." In contrast, "President Trump’s approach—both the “two-for-one” requirement and the regulatory budget—breaks from the historical emphasis on maximizing net benefits and improving the use of and commitment to benefit-cost analysis, and instead offers a blunt institutional reform to rein in regulatory costs (without attention to benefits). This is presented as necessary to counter the political impulses that may produce excessive or inefficient regulation, or regulation that could be better designed (for example by using market-like incentives rather than commands and controls)."

So what are the potential gains from a regulatory reform agenda? And how have the two-for-one and regulatory budget policies worked in other countries?

The Council of Economic Advisers report on "The Growth Potential of Deregulation" offers an advocate's case, but at least for me, the evidence is a mixed bag. As already noted, I'm certainly someone who is prepared to believe that misguided or overly severe regulation can impose costs in excess of benefits in some cases, and the report mentions a number of examples. Airline deregulation saves US consumers something like $18 billion per year. Finding ways to streamline the way in which the Food and Drug Administration evaluates new drugs can both save money for producers and also save lives of patients. State-level rules for occupational licensing can impose costs on workers and the economy, and local rules that limit housing supply can make housing less affordable and limit geographic mobility into certain urban markets. These kinds of examples are fairly bipartisan: for example, the Council of Economic Advisers under the Obama administration expressed concerns about the extent of occupational licensing and restrictions on housing supply, too. 

The CEA cites evidence from a study by the OECD which ranks countries by the extent of their product market regulations. A few decades ago, the US was usually close to the least-regulated, which is no longer true.

The federal government reports each year the number “economically significant” rules that have an annual effect of $100 million or more. "Under the Obama Administration, the government promulgated 494 new rules deemed `economically significant,' and under the W. Bush Administration, the government issued 358 such rules. Under the Clinton Administration, those agencies issued 361 such rules ..."

A basic eyeball test also suggests that certain kinds of regulation have become quite extensive. I sometimes note that from Pearl Harbor in December 1941 to the surrender of Germany and Japan in mid-1945 was about 3 1/2 years. In any major US city, it can in some cases take longer than that to go through the process of obtaining building permits for a single high-rise.

But while I'm certainly open to the idea that a lot of regulations should be reduced or eliminated, the top-line big-number estimates in the CEA report are on a weak footing. For example, the first line of the "Summary" begins: "Excessive regulation is a tax on the economy, costing the U.S. an
average of 0.8 percent of GDP growth per year since 1980." If true, this would be a truly enormous amount. It implies that the $18 trillion US economy would be about one-third larger this year--call it an extra $6 trillion in output this year and every year moving forward--without the regulatory burden.

But in the report, the evidence for this claim rests entirely on a single unpublished working paper by Bentley Coffey, Patrick A. McLaughlin,and Pietro Peretto called "The Cumulative Cost of Regulations" Mercatus Working Paper, April 2016).  The study measures the extent of US regulation in each  industry by using machine learning algorithms to classify chunks of text in the Code of Federal Regulations. They look at three kinds of investment in 22 different industries (out of the nearly 100 industries in this US government data). They acknowledge that it's hard to figure out cause-and-effect between regulation and investment: for example, some regulations might require new investment (say, in antipollution equipment) while others might reduce investment by making it less profitable. But they do their best to build up a model that relates investment and regulation. They then estimate what would have happened in the US economy if regulation had been frozen in place since 1980. 

This working paper seems to me like a worthy exercise to do as a working paper and a piece of academic research. But there's no way in the world that it should serve as the main justification for a national program of regulatory reform. The results depend heavily on how they measure regulation, on the structure of the model, and on what other potential confounding factors aren't taken into account. As the authors note, the study does not look at potential benefits of such regulation in terms of health or safety. 

Indeed, the CEA report itself mentions a 2004 government study from the Netherlands which suggests that cutting administrative costs of regulation by 25% can increase real GDP by a total of up to 1.7% in the long run. Without endorsing that particular study, I'll note that the claim that lower regulation can lead to a long-run increase that totals 1.7% is a LOT different from the claim that less regulation would have led to an economy that was 0.8% larger every single year for decades.

Another claim in the CEA report is that federal regulation imposes costs of $2.03 trillion per year. (This is of course quite a bit smaller than the $6 trillion in annual costs implied by the earlier study, but maybe they can be considered roughly the same.) The source for this claim is an appendix in a study by W. Mark and Nicole V. Crain. “The Cost of Federal Regulation to the U.S. Economy,
Manufacturing, and Small Business,” done for the National Association of Manufacturers (2014). The main study reports survey results from manufacturers on how they experience the burden of regulation. But back in Appendix C, there is a description of a calculation which measures the extent of regulation by using three components of the Global Competitiveness Index put together by the World Economic Forum. They then do a regression with per capita income as the dependent variable, and the measure of regulation as an explanatory variable, also including as control variables the trade/GDP, tax revenue/GDP, capital investment/GDP, and the dependency ratio. Some of these are lagged one year; some are in log values. I commented on a study by the same authors using this methodology a few years back in "Does Federal Regulation Impose Costs of $1.75 Trillion Per Year?" (June 6, 2011). My bottom line is that there's certainly nothing wrong with doing it as an illustrative exercise. But there's no reason to trust the results very much, either.

Thus, my overall sense of the CEA report is that it makes a soft case for that regulatory reform is worth considering, and may have real gains. However, when the report tries to make these gains look enormous, it ends up relying on shaky calculations from little-known working papers and reports.

What has been the experience of other countries with the Trump proposals? The Gayer, Litan, and Wallach report notes that both Canada and the UK have adopted versions of two-for-one and a regulatory budget. For example, here's an overview of happened in Canada (citations omitted):

"In 2001, the Canadian province of British Columbia committed to reducing the regulatory burden by one-third in three years. It required that each ministry establish a baseline of its existing inventory of “regulatory requirements,” defined as “an action or step that must be taken, or information that must be provided to access services, carry out business, or meet legal responsibilities under provincial legislation, regulation, policy, or forms”. . The initial count found over 330,000 such regulatory actions. In order to meet the three-year goal, each cabinet minister was required to match any new regulatory requirement with a plan to eliminate at least two offsetting requirements. In 2004, having surpassed the goal and achieving a 40 percent reduction in regulatory requirements, British Columbia imposed a regulatory cap mandating no net increase in regulatory requirements. This requirement has been extended three times, most recently to last until 2019, leading to a total reduction in regulatory requirements of 49 percent since 2001. Motivated by the success in British Columbia, in 2012 the Conservative-led Government of Canada released the Red Tape Reduction Action Plan, which required that for any new or amended regulation, regulators offset “an equal amount of administrative burden cost” from existing regulations. It also required at least one regulation be eliminated for every new one introduced. ...

"In January 2011, the Conservative and Liberal Democrat coalition government of the United Kingdom instituted a regulatory reform plan that included a “one-in, one-out” system in which each department must assess the “net cost to business” of complying with any proposed regulation, ensure that the cost estimate is validated by an independent committee of experts (known as the Regulatory Policy Committee), and find a deregulatory measure that offsets the cost of the new regulation. In January 2013, the requirement was increased to a “one-in, two-out” rule, which requires that the deregulatory measures must offset twice the cost of the new regulation, not merely eliminating two other regulations, as Canada has required and the Trump administration has just adopted. In March 2016, the United Kingdom ramped up its regulatory offset program again, to become “one-in, three-out,” again referring to costs, not the number of regulations. The “net cost to business” under the United Kingdom’s approach is computed as the “annualized direct net cost to business, incorporating direct recurring costs and transition costs, direct recurring benefits, and direct transitional benefits, spread out over the lifetime of the policy”. The “deregulatory” measures pursued as offsets in the U.K. system often do not actually remove any regulatory requirements, but rather make regulatory compliance less costly, for instance by streamlining paperwork processes so that businesses could make some filings without the need of a lawyer ... The United Kingdom’s regulatory initiative, however, does not use a social welfare yardstick, and thus does not seek to maximize the net benefits of its regulations to society as a whole."
In short, these general types of regulatory reforms have worked reasonably well in Canada and the UK. What about the Trump proposals in a US context? Here, Gayer, Litan, and Wallach are more cautious.

In US law, an existing regulation that has been duly created through a legislative and regulatory process cannot just be wiped out by the president. Instead, the two regulations that are supposed to be wiped out for each new regulation are will instead need to go through a process of comment and review. As the authors note: "Since some of the proposals to eliminate rules will undoubtedly invite legal challenges, there will be considerable uncertainty as to whether those rules really will end up being wiped from the Code of Federal Regulations. Any approach that attempted to bypass notice-and-comment procedures would likely run afoul of the APA [Administrative Procedure Act],  leading to defeats in courts and likely political backlash as well. Even when standard procedures are observed, there is no guarantee that attempts to roll back regulatory requirements will pass APA muster." 

The idea of a regulatory budget is that government will first set a certain total amount of regulatory cost that is acceptable. Then it will attempt to allocate those regulatory costs in the way that brings the greatest total benefit. Of course, measuring the total cost and benefit of a regulation is a tricky business. Gayer, Litan, and Wallach write: 
"It is fair to say that the Trump administration has launched the most ambitious regulatory budgeting program in human history—just a tremendous undertaking. Whereas Canada and the United Kingdom have managed to get their programs up and running with some success thanks to relying on relatively simple metrics of cost, in the United States the regulatory budget will attempt to get much closer to real social costs, at the expense of adding considerable complexity. That makes it potentially more meaningful and deep reaching, but also more likely to bog down and create a massive bureaucratic headache to go with those that already exist."
Again, I'm certainly open to the notion that lots of regulations have too little justification, and that regulatory reform can be socially beneficial. But the Trump proposals run a real risk that they will simply freeze all existing regulations in place: it will be too difficulty to remove old rules, and also too difficult to justify implementing new ones. As Gayer, Litan, and Wallach note: 
"But if all that the Trump administration’s regulatory budget turns out to be is an elaborate moratorium on new actions, that would represent a missed opportunity for would-be deregulators. The whole purpose of instituting a forcing mechanism is to confront the problem of accumulated and outdated regulatory requirements that burden U.S. businesses, thereby freeing Americans’ energies for productive purposes and unleashing economic growth. If this administration’s initiative ends up being nothing more than a pause in further accumulation—of both good and bad  prospective regulations—it would stand as a harsh judgment on the likelihood that existing regulation would ever be seriously reformed."