Lockdowns Are Not the Answer to Contain COVID-19; Mass Testing & Contract Tracing Are

It is clear that the US response to the coronavirus has failed. The number of cases per capita and deaths per capita in the country is one of the highest in the world, and the country is ranked above most developed nations in these metrics, showcasing its failure to contain the spread of COVID-19.

Currently, the US is 12th in both total cases and deaths from COVID per 100,000 people since the start of the outbreak in March. Many countries with fewer resources have managed to coordinate a better response to containing the virus.

In addition, the virus is surging across the US overall and in almost every state with confirmed cases and deaths surpassing their previous highs. The months ahead are going to be very difficult despite the positive news about a vaccine since it will take some time to execute massive rollouts that reach the required level for herd immunity across the population (take as evidence the time required to create adequate testing infrastructure, which is still failing many people).

Part of the problem in the US is the lack of a sensible strategy by the federal government at the top, which has forced states to each create their own patchwork response. The relative strategies across states have varied widely from doing nothing to complete economic shutdowns with orders to stay at home.

The Bay Area in California, which implemented one of the strictest measures in the US back in March, has reimplemented a shelter in place order that bans gatherings with other households, bans outdoor and indoor dining, closes most retail establishments (but weirdly allows shoppers indoors at limited capacity?), and many other activities with the addition of requiring mask wearing when outside the home.

While the state (and in particular San Francisco) have credited this strategy with combating the virus in the past, is it really the reason why they saw cases go down previously?

The answer is most likely no. The shelter in place is a very blunt strategy that tries to limit all in person contact to reduce the risk of transmitting the virus, but everything we know about COVID-19 suggests that different activities have varied risk profiles. The shelter in place order does nothing to accommodate the scientific evidence in making policy, especially when weighed against the economic harm (and deaths that will likely result from this damage).

All the research suggests that outdoor activities, especially when accompanied by distancing and mask wearing, have a very low risk. Indeed, outdoor dining with properly spaced tables, mask wearing, and cleaning between diners is quite safe. The real important things to have are mask wearing and social distancing. Why then would the state ban outdoor dining but allow indoor shopping?

The answer is clearly to try to allow some economic stimulus during the holiday shopping season, but the state is picking winners (malls and retail stores) and losers (restaurants and bars) while allow a riskier activity to try to contain COVID. That does not make any sense.

Furthermore, the state has implemented this policy without really enforcing it, meaning that people are unlikely to comply with the rules since they do not really make sense. The lack of a sensible policy means it is falling on deaf ears and that activity is likely to continue in risky ways.

It is also clear that more severe lockdowns do not actually correlate with reducing COVID cases and deaths. The evidence in that paper comes from cross country comparisons in Europe as of August 6.

We can both broaden the sample of countries to those globally and use more recent data on the spread of coronavirus to date to see if the findings hold up with additional data.

We will take the weekly time series of COVID-19 cases and deaths per country per 100,000 people (to normalize for country size) and regress it on a measure of lockdown severity created by the Blavatnik School of Government from Oxford University. We will lag the index by 21 days to essentially predict future COVID cases and deaths as a function of lockdown policies implemented in the past. The 21 day lag is due to the timeframe of onset of COVID symptoms and also what the paper above that we are replicating uses.

This should ideally solve any endogeneity issues, allowing us to actually understand whether stricter lockdown policies drive a decrease in COVID cases and deaths in the future.

The lockdown index measures the implementation of various rules like school and workplace closings, cancelling public events, stay at home requirements, and other things, merging them into a containment index. Higher scores in the index mean a country has a more severe lockdown policy in place.

This index is measured over time, so we can look at how the weekly containment index in a particular country (lagged to the past) correlates with future COVID cases and deaths to understand the relationship between lockdown policy and containing the virus.

The results are in the table below. We see that generally the containment index is positively correlated with both COVID cases and deaths across specifications (except for one case where it is negative but not statistically significant). This suggests that more severe lockdowns do not predict lower COVID cases and deaths in the future across countries (even when focusing on just high income, developed countries similar to the US).

We have little evidence here that severe lockdowns as a blanket policy work to stop the spread of COVID, likely due to the fact that the policies are not nuanced enough to both target and enforce the restriction of actions that research says are actually high risk for spreading the virus.

What then is the right strategy? It is a combination of mass testing and mass contact tracing.

Mass testing would allow us to actually identify the number of cases, providing people with the information needed to self-quarantine and reduce their risk of transmitting the virus to others.

While the US has increased its testing infrastructure to the point where it is one of the top countries in terms of tests per capita, with it surging in recent weeks, that is not sufficiently enough to support true mass testing. Widespread testing would involve daily tests across the population and testing before engaging in certain activities to reduce the risk of super spreader events while allowing limited social gatherings.

Combining mass testing with mass contract tracing would enable officials to both identify who has the virus but also trace it to the other people who are likely to be infected, allowing them to quarantine, and crucially identify the reasons why the virus spread.

One objection to this is implementing contract tracing in the US due to a mistrust of government and privacy concerns for individual freedom. But lockdowns and shelter in place are already infringing on individual freedoms in a more obstructive way than contract tracing would and futhermore, there are ways to implement anonymized contract tracing that mitigate these concerns, especially when leveraging things like cellular data (California just rolled out an iPhone contract tracing protocol, but it is quite late into the pandemic obviously).

With that information states and local governments would be able to define targeted policies that restrict risky activities actually leading to the spreading of COVID as opposed to applying blanket lockdown policies that do not appear effective and instead inflict massive economic pain (see the US unemployment rate and labor force participation rate vs. other countries during the pandemic).

We might say that with the vaccine this is no longer important, but it will take time for the vaccine to rollout, meaning that the upcoming months will likely see both massive spreading of the virus and massive economic damage. Furthermore, this is unlikely to be the only pandemic we face going forward, and in the future ensuring the US has a better policy to confront future viruses will be important to minimizing the damage.