Up and Down with the Pandemic; Why Pandemic Policies Do Not Last and How to Change That

post by Timothy_Liptrot · 2021-03-29T23:04:02.080Z · EA · GW · 2 comments


  Up and Down with the Pandemic
  What is to be done

Community control measures, such as closing nonessential indoor spaces has successfully suppressed past Covid-19 outbreaks and saved lives. For political reasons these policies have not lasted long enough to prevent Covid-19 infecting 20-40% of Americans. The current response relies on ephemeral public support. When deaths and hospitalizations recede, the public’s interest and cost-tolerance decline. Without strong public support, politicians face irresistible pressure to relax costly restrictions. This article uses the issue attention cycle and policy feedbacks to explain the fragility of restrictions to fight Covid-19 and to advocate for antigen testing as viable and durable alternative.

Through shrewd issue framing, the public health field has passed a suite of policies to reduce face-to-face contact during the pandemic, including stay-at-home (SAH) orders, closings of non-essential businesses, state of emergency declarations, school closures, and restrictions on gatherings. Combined with social distancing, the lockdowns significantly reduced first-wave Covid-19 transmission and saved lives.

But that response depended on unsustainable levels of public support. Most social distancing was not institutionalized at all. Economists Cronin and Evans found that state restrictions caused only 17 to 26 percent of the mobility decline for discretionary goods and services, rather than regulations. Those changes institutionalized proved politically unsustainable without support. As a result, America is caught in a cycle of resurgent Covid-19, a return to restrictions and compliance, a dropping compliance, and resurgent Covid-19. Few other issues move so cyclically. Seat belts, Ozone protection and underage drinking policies were all passed during spikes of public attention but have constant effects. Our intermittent response results from ignoring a well-known political dynamic: The issue attention cycle (IAC).

Up and Down with the Pandemic

In 1972 Anthony Downs proposed that public attention is not proportional to problem magnitude. In the IAC, interesting new problems suddenly leap into the center of attention, but the public soon becomes bored of weeks old news and moves onto new, more interesting stories. The IAC proceeds as follows; an undesirable social condition exists, alarming experts or interest groups. A dramatic event causes public awareness spikes, dominates the media cycle and leads to high public interest and willingness to sacrifice. As the response continues, the public realizes that solving the problem requires costs that they are hesitant to pay. Interest gradually declines because the public consumes news as entertainment. Each issue competes with other issues, sports, drama and horse-race politics for bandwidth. This pattern plagues many issues; racial inequality, environmental protection, and epidemics. Issues can avoid the IAC by reaching a majority of American communities. Knowing a victim creates salience without entertainment news, so cancer and the Vietnam War avoided received constant attention.

The pandemic response goes through a typical IAC except during major outbreaks. During spikes in deaths and hospitalizations, the public focuses on Covid-19 and tolerates high costs. People stop hosting parties, and some barely leave their homes. Patrons no longer attend bars and restaurants. State and local officials are rewarded for banning activities from indoor dining to outdoor playgrounds. Virus transmission declines but soon the issue no longer affecting most communities. Attention drops, people pay less attention to the CDC, and the cycle repeats. The cyclical pattern is directly observable from Apple maps requests. The lag between virus reproduction and deaths and hospitalization allows significant deaths in each cycle. Even worse, each cycle weakens the feedback mechanism, just as the public slowly bored of famines, Black Panther martyrs, and famous artists overdosing.

We cannot end the issue attention cycle. If editors print and broadcast only snazzy pandemic graphics, people will simply change the channel. Doing so would also restrict attention to climate change and global poverty. Cost-tolerance and attention are closely-linked. Most actors perform cost-benefit calculations simply by comparing the availability of the harms with costs. No one writes expected utility equations decide if they should eat out. High personal cost-tolerance cannot be maintained solely by elite messaging. To paraphrase Washington, depending on public attention is resting upon a broken staff.

Misinformation is a deceptively tempting explanation for people’s behavior. It offers simple solutions in elite message discipline (and a pleasing self-image to elites). But misinformation would not predict a cyclical pattern, as if people forget about exponential virus reproduction then suddenly remember during spikes. Furthermore, public willingness to sacrifice decreases on all such issues, even nuclear politics after Chernobyl. The public cost tolerance is simply tied to attention, which declines tragically but inevitably when victims are rare.

We must use attention peaks to build institutions that outlast public attention. Policies, NGOs, businesses, and norms can continue to operate without public attention. Ralph Nader reduced vehicle deaths through vehicle design regulation, not individual choices. German political parties remained opposed to nuclear power long after Chernobyl. Americans have forgotten what a chlorofluorocarbon is, but EPA continues to protect the ozone. But these new institutions must be politically durable.

##Why Lockdowns Were Not Durable Institutions

Those restrictions which were institutionalized did not last because they created self-undermining feedbacks. The restrictions have been lifted early across the country. Just as puzzlingly, cities have proven more willing to reopen bars and restaurants than even primary schools. The failure of the restrictions is a direct result of the policies characteristics, which create powerful self-undermining feedbacks. As the government acts on society through prohibitions and incentives, people affected enter the political arena. Government action also rearranges the resources actors have, such as money, status and activatable social networks. It is no coincidence that 2020 saw unprecedented expansion of government influence (mostly by states) and a 15 percentage point increase in voter turnout. New policies create new politics.

The most durable policies produce self-reinforcing feedbacks that prevent repeal. Our generous welfare policies for the elderly give them the time and resources to organize to defend the policy. But other policies undermine themselves. Policies are likely to self-undermine when they impose concentrated and enduring losses on powerful constituencies and when those losses can be easily traced to origin. Ambitious social engineering programs have unanticipated costs which grow over time, particularly when implemented under time pressure. The lockdowns are a textbook example of a self-undermining policy.

The lockdowns impose concentrated costs on powerful, well-organized industries. Because political decisions are made collectively, groups of people (interest groups) have far greater influence than disorganized individuals. Durable policies disperse costs across many disorganized persons. Policies that concentrate costs on well-networked groups make powerful enemies. Community control measures, particularly those that close industries, concentrate those costs on ready-made interest groups. This makes clear why bars and restaurants are powerful. Most bars are locally owned by high-status, politically aware entrepreneurs. They are prominently located in well-travelled areas, often the symbolic center of a community. Their profits are low, but they pay many workers a living wage. The owners have personal relationships with their workers, neighbors, and patrons. These leaders, incentives and social networks are the building blocks of a strong coalition.

By contrast, Dr. Fauci regularly appears on MSNBC which reaches a few million Americans each night. Marc Lipsitch’s “Reopening Primary Schools during the Pandemic” received just 300,000 views and 21 media mentions, heroic for an academic work. These viewers and readers are scattered across the 50 states, many isolated in their homes. They may be highly motivated by reproduction numbers and lost learning years, but they are not an organized coalition which can defeat the bars, restaurants and teachers unions of America without popular support.

Secondly, the lockdowns were incredibly ambitious projects of social engineering, bound to have unintended consequences. DC’s stay-at-home-order in March of 2020 listed 5 allowed reasons to leave a residence. Listing essential activities requires a single person or institution to understand the economy in such detail as to classify each activity. The set of unintended consequences is too long to list, and an excellent review of studies is found in Gupta et al. Cronin and Evans show that stay-at-home orders actually effected essential retail and nonessential retail the same, despite the exemptions (likely because the informational effect is dominant) . Ravindran and Shah find stricter lockdown enforcement associated with increased domestic violence complaints by 131%. They increased multi-generational living as students and the unemployed return to live with their vulnerable older relatives. Alexander and Karger find consumer spending at small businesses reduced by 35% by SAH’s. Beck et al. find unemployment increased by 1.9%. Some of these unintended consequences affect only isolated (and therefore politically weak) constituencies, but many affect politically cohesive unions, businesses, and communities. The longer each restriction continues, the larger the anti-restriction coalition.

Dr. Fauci regularly appears on MSNBC which reaches a few million Americans each night. Marc Lipsitch’s “Reopening Primary Schools during the Pandemic” received just 300,000 views and 21 media mentions, heroic for an academic work. These viewers and readers are scattered across the 50 states, many isolated in their homes. They may be highly motivated by reproduction numbers and lost learning years, but they are not an organized coalition which can defeat the bars, restaurants, and teachers unions of America once the public looks away. Even worse, the supporters of restrictions are withdrawn from their existing social networks. Geographic, occupational, social and activist networks are all slowly eroded by the reduced face to face interaction. Without those networks the beneficiaries struggle to communicate their common interest and to act collectively to protect the restrictions.

What is to be done

We must use the current spike in Covid-19 awareness to build durable institutions. They must impose low, dispersed or untraceable costs to outlast the lockdowns. Harvard professor Michael Mina has proposed just such a solution: rapid testing.

Our current testing policy follows a clinical model, as decided by the FDA. Tests are optimized for sensitivity, so they can identify infected persons long after they cease to be infectious. However, they lack speed, cost-effectiveness and scalability. The standard clinical polymerase-chain-reaction (PCR) typically require transport to a centralized lab, which increases costs, decreases frequency and delays results for one or more days. As a result, most Americans have never received one. Even worse, the long turnaround time leaves infectious persons active for days before results are received. As a result, Americans cannot realistically know when they are infectious.

We can solve this problem by authorizing tests as public health tools, not just clinically. Lateral-flow antigen tests can detect the infectious people today. Antigen tests cost less than five dollars, are scalable and can be performed at home. Processing times are often just 15 minutes. False positives can be mitigated by taking a second confirmatory test or a PCR test. Yes, their sensitivity to low virus levels is limited, but they successfully identify who is infectious. These tests are accessible enough to test workers multiple times per week, and catch enough asymptomatic infectious cases to gradually strangling the virus week by week. Some statisticians complain antigen tests miss PCR-positive cases, mostly as viral RNA slowly leaves post-infectiousness. Catching these no-longer infectious cases entirely misses the point of public health testing.

A common argument against antigen testing is that we know lockdowns work, so why experiment. Ali Khan of the University of Nebraska commented “I’m not a fan of [antigen testing] because we have not yet taken the strategies that we know work in Australia, China, New Zealand and Taiwan. (…) Once we exhaust the existing public health tools, we can use other tools”. People who failed to social distance long enough will not act their antigen test.

Khans argument is incorrect, antigen testing will be more durable than social distancing mandates, not less. Isolating only infectious individuals is hundreds of times less costly for both businesses and individuals than closing completely, so the repeal pressure is lower. Secondly, those costs are dispersed across the country rather than concentrated on particular industries. Notice that mask mandates, which likewise impose low and dispersed costs, have proven durable. The gains from institutionalizing mass public health testing would help even the current cycle of Covid-19 response and resurgence.

Lockdowns would work if elites could compel compliance consistently, but they cannot. America must understand why lockdowns do not last to design a response that does.


Comments sorted by top scores.

comment by Larks · 2021-01-07T05:27:15.326Z · EA(p) · GW(p)

Very interesting article. Two minor pieces of housekeeping:

  • The paragraph beginning "Dr. Fauci regularly appears on MSNBC" appears twice
  • There are some sections marked with "<>" which appear to be placeholders for subsequent content.
comment by Timothy_Liptrot · 2021-01-07T05:25:30.346Z · EA(p) · GW(p)

I haven't ported the citations to this format yet, fyi