La décision publique, l'expertise et le droit III. Time and Virus

Public forums or deliberative debates with representative groups of citizens and other stakeholders take time to organize, and they work best if factual material exists that participants can study. The “facts”, however, may be statistical estimates with margins of error. Public choices depend not only on policy preferences for specific outcomes but also on social attitudes toward risk. Even if individuals are risk averse in their daily lives, they may advocate for public policies that are risk neutral that is, based on expected values calculated by multiplying each outcome by its probability of occurrence and adding them up. Conversely, others argue that states should be risk averse in making choices that can affect mortality and morbidity in the population. This distinction between average net benefits versus tolerance of social risk can make certain kinds of public choices especially fraught with controversy. Emergencies, such as the corona virus pandemic raise the issue of society’s tolerance toward risk in a particularly stark form. There may be a clear tension between individually rational behavior and social welfare.

Before COVID-19 hit in early 2020, most of us had not given serious thought to the appropriate social response to a pandemic involving difficult risk-benefit tradeoffs.  Under one view of the proper operation of a democracy, low-probability events with high costs should be part of the public debate in quiet times where research can increase our understanding, and there is time to learn and to reflect. The debate should not be left entirely to experts, although they clearly provide necessary inputs. One model is the United States notice-and-comment process for most executive rules with the force of law. The Administrative Procedure Act (APA) requires public notice, an open-ended hearing, and a published statement of reasons. Before the APA process begins, career bureaucrats and politically appointed officials seek help from experts inside and outside the government and seek to isolate politically feasible approaches. The experts provide estimate of costs and benefits, along with calculations of their degree of precision. The entire process of developing a rule, from initial scoping exercises to the rule’s publication in the Federal Register, can take several years, especially for rules with major consequences for society and politically powerful gainers and losers.

Emergencies, such as the corona-virus pandemic, test this model of democratic accountability. Suddenly, people are dying, and the disease spreads exponentially, both domestically and across national borders.  The corona virus, although similar to some past outbreaks, was new to science and is very contagious. Governments moved quickly both to support research on the virus and on diagnostic tests, vaccines, and anti-viral medications. Drastic decisions to shutdown regions and whole countries imposed severe costs on households and businesses, justified, at first, only by the imperfect history of past quarantines.  At the same time, governments with the financial ability quickly made cash transfers to households and businesses. For example, the legislative process in the US, generally mired in partisan debate, rapidly enacted two massive COVID aid packages under presidents Trump and Biden. Each act took only a few weeks from the bill’s introduction to its passage by the Congress. The first bill included a series of institutional checks to limit fraud and misappropriation of funds, but these controls took a back seat to the goal of producing a rapid stimulus package. There was little organized public consultation, and even experts did not have definitive answers.

Pandemics have special features. All emergencies and disasters demand a speedy response, but some disasters quickly dissipate and return society to the old normal except for the costs in life and property flowing from the disaster itself. Rebuilding can commence, and those who suffered harm can be cared for and compensated without the risk of a cascade developing. A hurricane does not itself trigger additional hurricanes, and a flood does not spawn other floods, although each may be evidence of an increased risk caused by other factors, such as climate change or drought. The disaster can spur society to prepare for future events, for example, by stockpiling emergency supplies, creating training programs for first responders, and, more fundamentally, moving people away from a floodplain or revising building codes to limit damage from natural disasters. Governments will have time between periods of heightened risk to reflect on how best to respond based on both science and the views of the public.

A pandemic is different in one important respect. The disease spreads from person to person so, absent a vaccine, inter-personal contacts need to be sharply limited and tests for the disease developed. Returning to normal life depends on control of the disease itself, not on repairing the damage from the past. Of course, a cholera epidemic may follows a flood or earthquake, but in such cases, scientific advances are unnecessary, and solutions are costly but straightforward.

In responding to the corona virus, the need for speed clashed with both the demands of scientific proof and the need for democratic accountability. Governments moved quickly in the US and Europe, and, indeed, around the word to limit the spread of the disease and to develop vaccines, treatments, and tests. In the US, such speed circumvented American administrative law requirements for open, participatory, and well-justified rulemaking. It also clashed with the Trump administration’s interest in slowing down regulatory activity and deregulating the economy. Even though the Occupational Health and Safety Act explicitly permits the Labor Department to issue emergency temporary standards (ETSs) outside of APA procedures, the administration did not act. With a new administration in place and the pandemic on the wane, OSHA did issue an ETS in June 2021, but only to cover the health care sector.

The rapid response also stands in contrast to growing efforts in Europe to increase the public’s involvement in legislative and executive policymaking. If the infected can infect others in a widening spiral, governments will try multiple, overlapping strategies that could slow transmission and reduce illness and death.  They can consult the best scientists, but controlled experiments cannot be carried out in a spreading pandemic. Furthermore, organized public consultations are not feasible. The model of the French Citizens’ Assembly on Climate Change 2019-2020 in which a stratified random sample of the population deliberated over several months and produced 149 policy recommendations, would not have been realistic. Climate change is an existential threat, but there is time for both better scientific analysis and more in-depth public involvement. Advocates for quick and substantial action against climate change set near-term goals and point to tipping points that could prompt disastrous spirals. Nevertheless, even the most alarmist argue that there is time to take account of both scientific data and public concern. They contrast President Macron’s abrupt introduction of a climate-friendly tax on gasoline with the subsequent Citizens’ Assembly.  The gasoline tax led to the “yellow vest” protest movement that brought opponents onto the streets in France in 2019. The protesters objected both to the tax itself and to its high-handed introduction. In that case, there was no scientific imperative to rush the tax into place without public input.

A deliberate, balanced approach has not been possible in the response to COVID. Policy is made based on educated guesses rather than rigorous scientific testing.  Margins of error are large. The knowledge base was weak at the start of the pandemic, yet citizens, businesses, public officials, and politicians looked to experts for advice. As the knowledge base grew, the recommendations changed over time. Recall the shifting advice on facemasks and sanitizing surfaces as COVID spread. Some interpreted these changing recommendations as a reason to distrust experts, but that misunderstands the nature of scientific truth. Research ought to challenge existing findings with new data, theory, and techniques. However, these reassessments of the situation were occurring, not in laboratory experiments, but in real time as people became infected and crowded hospital emergency rooms.  There was no time to sort out the marginal effects of different types of behavior changes. Only research on vaccines was carried out through controlled experiments in efforts to get approval from national drug approval agencies, especially the U. S. Food and Drug Administration.

Experts may recommend that nothing be done because controlled experiments have not sorted out the marginal value of alternative strategies. Overriding such professional norms, some U.S. efforts espoused a “Swiss cheese” approach that arranges the options to eliminate holes from the resulting stack. If you do not know how to overcome the limitations of any particular policy, use multiple approaches that compensate for the weaknesses in the others.  During the corona virus pandemic, policymakers as well as scientists lacked confidence about what would work best on the margin. Whatever steps they took would make a segment of the population angry and resentful about restrictions on their freedom. Some used the conflicting advice of experts to argue against restrictions of any kind.

Of course, even the Swiss cheese model has limitations if the experts come from only a few disciplines, say, epidemiology and public health. Specialists may fail to consider issues outside of their own areas of expertise. The clearest example here is the widespread closure of schools that limited the spread of the virus but appears to have imposed severe costs on many children, especially those with limited internet access or teachers unable to adjust to on-line learning. There are risks in consulting only a limited range of experts and avoiding deep engagement with those most affected by the policy. If the impact will be felt for years in the future, however, no one today may be able adequately to assess the future effects of lost or inadequate schooling. These decisions involve statistical estimates of average costs and benefits, calculations of the distribution of harms across different segments of the population, and the public’s tolerance for the risks inherent in government policy choices.

The corona-virus pandemic illustrates the limits of both expertise and public participation when the pressure of time forces governments to act in the absence both of solid expert advice and of public input. Everyone understands that time is of the essence, but neither experts nor the public can provide definitive answers. We are back to Ed Lindblom’s view of policymaking as “muddling through”.  Progress toward ending the pandemic may be better achieved by acknowledging the tentative nature of factual claims and initial policy responses and by efforts to bring members of the public into the policy debate not just as objects but rather as participants.

‍

[1] For an introduction to the US law of rulemaking in comparison withFrance, the UK and Germany see: Susan Rose-Ackerman, Democracy and Executive Power: Policymaking Accountability in the US,the UK, France and Germany, (New Haven CT: Yale University Press, 2021forthcoming).

[2] OSHA, Subpart U--Corona Virus Emergency Temporary Standard, June11, 2021. https://www.osha.gov/sites/default/files/covid-19-healthcare-ets-reg-text.pdf

[3] Michael Lewis, ThePremonition: A Pandemic Story, New York: W. W. Norton, 2021.

[4] Charles E. Lindblom, “The Science of ‘Muddling Through’” Public Administration Review 19: 79-88

(1969).

‍