Saturday, March 14, 2020

Benefit-Cost Analysis and the Coronavirus


We are in the middle of a flurry of decision-making on how to deal with COVID-19.  After much resistance, officials are now canceling public events, closing schools and discouraging other activities that put us in contact with each other.  Travel restrictions and possible shutdowns of workplaces, as we’ve seen in Italy, may be up next.

It’s interesting we haven’t heard anything about benefit-cost analysis in all this.  Nearly all economists profess to think that BCA is the single best decision method.  Almost every introductory economics textbook is built around benefit-cost thinking, and for decades federal regulations have mandated BCA for proposals with significant economic impacts.

But now we are facing immense choices—what could have a more drastic impact than shutting down most of the economy by fiat?—and BCA is nowhere to be found.

As a public service, here’s a quick and dirty.  Coronavirus policy is primarily about saving lives, right?  So, if you believe in this sort of thing, the official “value of a statistical life” (VSL) as determined by the Environmental Protection Agency is $7.4 million.  According to BCA wisdom, we should spend up to this amount to save the life of a currently unknown (statistical) person, but not a penny more.*

In order to get a first impression, suppose the more stringent measures proposed will shave 1% off US GDP for the year.  Based on last year’s figure, that would eliminate $214 billion in economic value.  Using the value of life metric, that means we shouldn’t do this unless we expect to save at least 28,919 lives.  If not, let’em die.

Actually, I think it’s likely that we will see even greater economic costs from stringent social distancing policies, especially taking into account that the economy would probably have grown by a couple percent or so this year had the virus not struck.  Maybe these actions would pass the BCA test, maybe not.

Personally, I think it would be madness to stick a monetary value on people’s lives and base our policy choices on whether the dollars on one side of the ledger outweigh those on the other.  I wrote a book about that a while ago.  The considerations that lead us to think primarily in public health terms at a time like this apply just as well to other issues, from food safety to climate change.  The economists who crank out monetary values of life and believe all choices should be based on benefit-cost thinking have not yet had the integrity to step forward and make their case.  This silence speaks volumes.

*On a technical point, economists at EPA and elsewhere argue that the VSL should vary based on the population at risk.  Since mortality from the novel coronavirus is higher for those in the highest age brackets, and death from the disease would eliminate fewer years of life for them than for younger people, their VSL should be lower.  It may also be the case that mortality will be lower among those with higher incomes, who can afford better preventive measures and treatment, further depressing the relevant VSL metric.  But these considerations would require even more lives saved to justify the expense.

26 comments:

Peter Pan said...

"The human toll here looks to be much worse than the economic toll, and we can be grateful for that." - Larry Kudlow on Japan disaster, in 2011.

Sandwichman said...

O.K., so ballparking a mortality rate of 2% and infection rate of 50% gives roughly 3 million deaths at (rounding) $7 million each, or $21 trillion. So there's that.

Peter Dorman said...

But S-man, what matters for BCA is the incremental effect of a set of measures. If those worst case-ish numbers can't be shifted by more than, say, 150,000 even if we do a 2008 on our economy, the recommendation is to do less.

Jerry Brown said...

So how do you argue with an economist if you don't believe a value in dollars can be put on a human's life? But I am flattered that my own life might be considered worth 7.4 million dollars by anyone else. Kind of makes me feel good for some reason...

Marian Ruccius said...

It also misses the point that the value of a statistical life would not be the only metric, not the only cost. So the anti-CBA argument is painfully simplistic. And in any competent BCA, the dollar value is just notional -- an easier computation than using, say, utils.

Anonymous said...

Barkley Rosser wrote about International Women's Day on March 8. There was an organized celebration of the day in Spain which evidently has sadly resulted in a significant number of COVID-19 infections. I am wondering why the celebration was allowed. Would there have been any cost to postponing the celebration? What does such a mistake mean?

Peter Dorman said...

JB, part of the argument is negative -- what's wrong with the BCA approach. The other part, as you say, is positive -- how should we do it instead?

This is a complicated topic, but the simple version, as I see it, is to think of a package that includes actions to slow down the progress of the epidemic, like shutting down large parts of the economy temporarily, and measures to meet the essential needs of the population during the interim. This includes paid leave, enhanced unemployment benefits, direct provision of credit to affected businesses at concessionary rates, publicly provided health care, child care and education, emergency housing, etc. You know you're going too far in shutting down the economy if you don't have the means to sustain essential living standards. Just maximizing a sum of GDP and deaths avoided * $7.4 million doesn't do this.

Anonymous said...

Terrific post, and I will read the book in coming weeks:

Markets and Mortality: Economics, Dangerous Work, and the Value of Human Life
by Peter Dorman

rosserjb@jmu.edu said...

A.,

Peter's book you mention is one of the best things he has done, really excellent.

More generally, I suspect a major reason we have not heard anybody who strongly supports use of BCA (or CBA) in this situation has to do with the severity of our lack of knowledge about what is going on and what would be the outcomes of alternative policies. Even if one were to substantially accept the EPA estimates, whether age-adjusted or not, our lack of such basic knowledge regarding likely outcomes of particular policies would restrain such believers from saying much.

Finally, on the matter of income distribution effects, this is quite unclear. On the one hand it looks like a lot of poorer service workers are some who will not be able to work from home and thus will end up more exposed. OTOH, it seems that at least unti now some high income people who see lots of people (think Trump guests at Mar-a-Lago) have been exposing themselves to more people than others, and we are seeing some high profile people, including government ministers in several countries, who have become infected.

rosserjb@jmu.edu said...

BTW, noting the specific things Peter suggeests should be done immediately, some of this looks to be in the bill just passed by the House with actual support from Trump. But for whatever reason Mitch McMconnell and the Senate are off until late Monday twiddling their thumbs on this. They may not ideologically like some of it, like paid sick leave, but with even Trump on board and the obvious severity of the crisis, this is just ridiculous.

Of course even as Trump looks sort of reasonable in supporting the House bill, he has been at it again today suggesting he might "demote" Jay Powell at the Fed, even though it is quite unlikely he has the authority to do that, and it is likely to roil the markets again, which he actually managed to have going his way late Friday. But then some say this is just jawboning to get more expansive policy ou of the FOMC meeting this coming week. We shall see.

Anonymous said...

Astonishing:

https://www.nytimes.com/2020/03/14/opinion/coronavirus-pelosi-sick-leave.html

March 14, 2020

There’s a Giant Hole in Pelosi’s Coronavirus Bill
The legislation passed by the House Friday night doesn’t actually guarantee paid sick leave to most American workers.

Anonymous said...

Barkley Rosser:

Mitch Mcconnell and the Senate are off until late Monday twiddling their thumbs on this. They may not ideologically like some of it, like paid sick leave, but with even Trump on board and the obvious severity of the crisis, this is just ridiculous....

[ However, as the New York Times makes clear, there is almost no guarantee of paid sick leave, much to my astonishment and sadness. ]

Anonymous said...

My crude sense is that the unease or fear that has spread these last days will be remembered through the election and representatives beginning with the president but extending well beyond who have failed at being reassuring and calming will not be forgiven. Friends and family are afraid, though that is not mentioned. Cost was irrelevant then, what was needed was to be reassured.

Jerry Brown said...

Peter Dorman, I asked an unfair question of you in light of what is a very potential, probably actual, crisis. And it seems you were using these estimates to argue for a more robust response, which I agree with completely. And there is certainly nothing wrong with using every argument at hand to try to get the right things done especially in an emergency situation.

rosserjb@jmu.edu said...

Anonymous,

I have not looked at details of the House-pased bill. Reports I have seen claim it has paid sick leave in it, but maybe this is not the case or only weakly so. Was not able to use your link, so do not know what its argument is. That Trump so willing to support the bill after having previously opposed paid sick leave certainly opens possibility thst Mnuchin managed to get the bill to have enough it could be labeled as having done so while not really doing so.

I would like to think all this will damage Trump's chances or reelection, but if it ends sooner than most now think and economy is able to rebound (neither of these impossible), Tyler Cowen at Marginal Revolution has argued that Trump might yet even make this into something aiding his reelection, an event in which despite some stumbles at the start he led us through a troubled time and blah blah bllah. I think this is not particcularly likely, but I cannot rule it out as impossible.

Anonymous said...

Barkley Rosser:

https://www.nytimes.com/2020/03/14/opinion/coronavirus-pelosi-sick-leave.html

March 14, 2020

There’s a Giant Hole in Pelosi’s Coronavirus Bill
The legislation passed by the House Friday night doesn’t actually guarantee paid sick leave to most American workers.

House Speaker Nancy Pelosi on Friday night celebrated the coronavirus legislation that passed early Saturday as providing paid sick leave to American workers affected by the pandemic.

She neglected to mention the fine print.

In fact, the bill guarantees sick leave only to about 20 percent of workers. Big employers like McDonald’s and Amazon are not required to provide any paid sick leave, while companies with fewer than 50 employees can seek hardship exemptions from the Trump administration.

“If you are sick, stay home,” Vice President Mike Pence said at a news conference on Saturday afternoon. “You’re not going to miss a pay check.”

But that’s simply not true. Sick workers should stay home, but there is no guarantee in the emergency legislation that most of them will get paid....

Anonymous said...

Also, so far the British approach to the nearing epidemic has been remarkably "cost-benefit" rather than "benefit-cost," just as has the American approach. Please note the difference Peter Dorman and Sandwichman have made clear between "benefit-cost" and "cost-benefit." An editor of the Telegraph in supporting the government has gone so far as to horridly write about the economic advantage of "culling" the elderly.

https://news.cgtn.com/news/2020-03-15/UK-s-response-to-COVID-19-may-expose-significant-risks-to-citizens-OSf3rGVIQ0/index.html

March 15, 2020

UK's response to COVID-19 may expose significant risks to citizens
By Thom Brooks

Thom Brooks is the Dean of Durham Law School and a professor of law and government at Durham University.

Anonymous said...

https://metro.co.uk/2020/03/11/telegraph-journalist-says-coronavirus-cull-elderly-benefit-economy-12383907/

Telegraph journalist says coronavirus 'cull' of elderly could benefit economy.

[ A Labour-Jeremy Corbyn response to the nearing epidemic in Britain could not have considered what the Conservative government evidently considered but I can only hope has dispensed with already. ]

Anonymous said...

Britain has thankfully changed the approach, so much for "cost-benefit" conservatism or what is actually nihilism:

https://www.cgtn.com/special/Battling-the-novel-coronavirus-What-we-know-so-far-.html

March 15, 2020

UK to move to shielding over-70s in coming weeks: Health secretary

The British government has called for a national effort similar to the one which helped the country survive during the Second World War as it prepares to curb the spread of coronavirus.

Health Minister Matt Hancock said measures the government planned included isolating people aged over 70 for up to four months.

He said the announcement would come "certainly in the coming weeks, absolutely."

Anonymous said...

Peter Dorman:

The economists who crank out monetary values of life and believe all choices should be based on benefit-cost thinking have not yet had the integrity to step forward and make their case.

[ Devastating statement. ]

rosserjb@jmu.edu said...

I do not think there is any difference between "benefit-cost analysis" and "cost-benefit analysis." I have been studying this matter as a public policymaker and also published on it as an avadmic for over 40 years. Whether one puts the cost first or the benefit first, the actual formuli for calculating the relevwnt ratios are identical.

Too bad this bill seem to have such weak paid leave coverage.

Anonymous said...

Barkley Rosser

I do not think there is any difference between "benefit-cost analysis" and "cost-benefit analysis."

[ Sandwichman strongly made the case on this blog that there was a difference that developed after the New Deal years in willingness to spend on social programs. At first, social benefits were considered with costs considered in what would be needed to reach the benefits. Later costs were considered while social benefits were secondary to how affordable they were considered.

This is Sandwichman's argument on this blog, and I found this convincing from the time Sandwichman wrote.

Possibly though Barkley Rosser is correct and whether benefits or costs are first considered in thinking through social programs does not matter. I hope Sandwichman will respond on the matter and Peter Dorman as well. ]

Anonymous said...

Barkley Rosser

Too bad this bill seem to have such weak paid leave coverage.

[ This president and Congressional Republicans in general truly frighten me. Beyond being intent on smashing other peoples and building ever more destructive armaments, elected Republicans appear to have abandoned the concept of working for the public good. A terrible crisis has been slowly developing and the president has been incapable of reassuring and responding.

I am very frightened. I wish I were not frightened, but I am.

I am grateful for being able to express my thought here. ]

Jerry Brown said...

We will get through this Anneonymous. I'm scared too, but we will get through it.

rosserjb@jmu.edu said...

Anonymous,

Sorry you are frightened and glad we can be here as an outlet. This is indeed a time when it is completely rational to be afraid.

I wanted to clarify this matter of BCE vs CBA, which I think is simply a matter of people calling the same thing by slightly different names. Others can correct me if they disagree, but for me both refer to a technical procedure that has been used in the US government since the 1930s and has since then spread all over. It was first developed by the US Army Corps of Engineers to rank competing flood control projects in terms of priority.

What is involved is estimating the present value of an expected future stream of benefits from something (measured in monay) and then dividing that by the expected value of the expected future stream of costs (in miney terms). This then gives one a benefit-cost ratio, and these can be compared across competing projects to rank them.

In the beginning before we started worrying about benefits and costs not easily measured in monetary terms and also projects or activities not involving overly complicated time patterns of these, this was pretty straightforward. So in the 1930s when they did not worry about environmental effects such as loss of species habitat from building a dam and other such things, the future stream of benefits amounted to the value of the expected future stream of flood damage avoided by building a dam. There might even be some other benefits, such as recreation on a lake that might be created by the dam or electric power generated from building the dam. Those future benefits need to be discounted by an interest rate, a controversial matter, with outcomes highly sensitive to what is chosen. On the cost side in these old dam projects, the costs were generally upfront and were just the costs of building the dam, which would include payments to property owners whose property would end up under water in the lake created by the dam.

There are lots and lots of issues involved with this, some of which have been discussed in this post and associated commments. But the technicalities of these estimates do not vary between whether one calls it "benefit-cost analysis" or "cost-benefit analysis" because in effect both of them are simultaneously estimated and then compared. It is not a matter of estimating one first or somehow favoring or emphasizing one over the other. st least within the framework of this widely used technique.

Anonymous said...

Thank you.