In an interesting op‐​ed in the Wall Street Journal, Philippe Lemoine writes (with my emphasis):

The coronavirus lockdowns constitute the most extensive attacks on individual freedom in the West since World War II. Yet not a single government has published a cost‐​benefit analysis to justify lockdown policies—something policy makers are often required to do while making far less consequential decisions. If my arguments are wrong and lockdown policies are cost‐​effective, a government document should be able to demonstrate that. No government has produced such a document, perhaps because officials know what it would show.

There is a more charitable explanation for their reticence: doing a comprehensive cost‐​benefit analysis of a lockdown is actually incredibly difficult, especially before the lockdown is implemented. Libertarians, in particular, would appreciate all the uncertainties, knowledge problems, and difficulties in aggregating very subjective values at play.

To be clear: these are not reasons to avoid doing a cost‐​benefit analysis of a lockdown. I think given the consequential nature of the policies, an attempt, even with a bunch of uncertainties and caveats, would have been clarifying about the trade‐​offs incumbent in any major decision. But we should not pretend such a task was straightforward.

In my forthcoming book, Economics In One Virus, I try to delineate all the issues that would need to be addressed to do a cost‐​benefit analysis of a lockdown well. There are major challenges on both the benefit and cost sides, and indeed in using the results to inform policy.

On the benefits side, one would have to:

  • Place a value on the reductions of the statistical risk of dying from COVID-19. Economists commonly use the “value of a statistical life” for this, but these values tend to be derived from labor market studies assessing very minor risks facing working‐​age people at work. These are unlikely to be appropriate for older people facing much higher fatality risks from COVID-19, and whom are likely to have more varied preferences on how much they would pay to avoid death risks.
  • Estimate how many lives would be saved by lockdown policies. “Defining the counterfactual” here is hard. The alternative to lockdowns wasn’t normality, but quite a lot of voluntary social distancing that already saved lives. And as Lemoine writes in the op‐​ed, people’s behavior tends to tighten further when the disease prevalence is high, ending the exponential spread and resulting in waves of cases and deaths. The problem is that within countries there appears little consistency about when that behavioral tightening will occur. So assessing the difference between the disease patterns in lockdowns and without them in advance is near impossible. Of course, depending on when the lockdown was planned to take place, at least some of the COVID-19 “deaths avoided” would be deaths delayed until the next outbreak too.
  • Account for the value of the reduction in non‐​fatality risks. A lot of people value avoiding getting sick or the worry that their family members will become ill. It therefore stands to reason that there are economic welfare benefits from lockdowns reducing these non‐​fatality risks too. Some economists have even concluded these benefits may be of a similar aggregate value to that of reduced deaths, simply because the number of COVID-19 cases avoided from any lockdown restriction vastly exceeds the number of deaths avoided.

As if accounting for all that wasn’t difficult enough, the “costs” side is murkier still. One would have to:

  • Assess lost economic output relative to a realistic counterfactual. This would likewise have to take into account the voluntary social distancing we would otherwise see and its economic impact. It would also need to incorporate the dynamic effects of the possible faster spread of COVID-19 absent lockdowns bringing sharper retrenchments from economic activity later. Again, a lot of the impacts would be sensitive to when the lockdowns were applied too—we’d expect over time that more businesses would find ways to operate COVID‐​safely, for example. This would make lockdowns that ban such activity grow in cost with their duration.
  • Account for the lost value of eliminated non‐​market activity. Stay‐​at‐​home orders and other COVID-19 interventions ban a lot of non‐​market activity too, some of which would have extremely high value to individuals. How much would an individual value being able to attend a family member’s funeral? Or being able to see a close friend who is going through a difficult time? Or even the ability to travel across state to stay at another home with more space? Constrictions on these non‐​market actions have big economic welfare costs too. But their precise values are incredible subjective and difficult to assess or aggregate.
  • Consider longer‐​term scarring impacts. Who knows what the extent of school closures will be on the lifetime earning potential of kids? Or how much less human interaction during lockdowns will affect innovation and longer‐​term growth? Or the knock‐​on impact of lockdowns on other health outcomes? The uncertainties over what could be the most consequential “scarring” impacts are huge, especially when compared to the social distancing we’d see voluntarily.

So, totting up the costs and benefits of any given lockdown is, in reality, very difficult. But there are also limits to using such analysis to inform decisions.

First, given the huge uncertainties, policymakers would be looking at a range of potential outcomes. Inevitably the central scenarios would be reported, but it might be the tail risk that keeps policymakers up at night. If you’re told there’s a 10 percent chance your hospital system might be overwhelmed absent a lockdown, that might be enough to push you into action, even if on the most likely scenarios the costs and benefits of lockdowns look well‐​balanced, or even unfavorable.

Second, while a simple cost‐​benefit analysis of lockdowns can in theory inform us whether a set of policies should even be considered, this doesn’t tell us what is the “best” approach. In reality, lockdowns are a bundle of different regulations, some of which might pass a cost‐​benefit test on the margin, but some of which would not. Finding the “best” approach would in reality mean running very many cost‐​benefit analyses, including comparing lockdowns to completely different approaches, such as guidance, or no lockdown but mask mandates, or widespread testing, or fitting riskier places with ventilation equipment or a whole range or combination of other less intrusive measures.

Given all these difficulties (and others I’m sure I have missed), I suspect we will only fully appreciate either the wisdom or futility of lockdowns through careful retrospective analysis. In the meantime, all these difficulties highlight some reasons why we are ordinarily wise to oppose government control over our everyday lives and choices.

–

For more on the economics of COVID-19, you can pre‐​order my book, Economics In One Virus (U.S.) The UK pre‐​order site is here.