Blogotariat

Oz Blog News Commentary

Lockdown cost-benefit analysis for Australia by Martin Lally

April 28, 2021 - 19:19 -- Admin

Our most popular cartoons of the yearMartin Lally is a kiwi economist who late in 2020 decided to calculate for himself what his own country was losing by locking itself away from the world, coming to the conclusion that New Zealand was sacrificing something like 26 life-years in the future to ‘save’ 1 life-year. The way he arrived at that number was to essentially calculate how much less the government would have available to spend in the future as a direct result of the effects and costs of lockdowns, and then compare that reduction with how much governments historically had to pay to produce healthy years of life. The logic is that if you reduce government expenditures to that of Russia, then you will get the life expectancy and health of Russians.

Helpfully, he has now done the same type of calculation for Australia (THE COSTS AND BENEFITS OF A COVID LOCKDOWN-6).

You will not be surprised that his headline conclusion is that the lockdowns in Australia will cost around 20 times more than they save, though even that number is achieved by being extremely optimistic about the benefits of lockdowns and by ignoring many of its costs.

Let me point out some interesting aspects of Martin’s effort:

  1. Looking at the data around the world till the end of 2020, Martin estimates that on average, the more stringent the lockdowns, the higher the claimed subsequent covid death rates. In these estimations, Martin takes account of population density, date of first death, population size, GDP, the proportion over 65, the severity of the previous flu waves, the proportion of nursing home beds per head, and average household size. His estimates are repeated in many other publications, but it has really not yet hit the public consciousness: lockdowns do not merely destroy happiness and are extremely costly, but they actually are associated with more covid deaths as well. One likely mechanism is that lockdowns make people unhealthy and thus more susceptible to lots of health problems, including covid.
  2. In order to arrive at some ‘possible benefit’ of lockdowns to compare with costs, Martin essentially has to assume they help avoid covid deaths, against his own estimates. It is somewhat humorous to see that Martin thus essentially has to disregard his own best estimates so as not to have to dismiss the case for lockdowns at the outset. Many others doing cost-benefit analyses of lockdowns have struggled with this as well: if there is no benefit in ANY dimension, there is no cost versus benefits, only costs. Like Martin, in my own calculations I also just often presumed lockdowns to reduce covid-deaths by a huge amount, contrary to all the evidence, just so as to be able to say “even if….”.
  3. In some of his calculations, Martin appeals to estimates of the WELLBY literature in order to value the misery of unemployment.
  4. Martin uses Swedish data to estimate that the residual life-expectancy of the average covid-death is 5 more years, which he then uses for Australia to infer the number of saved years of life from lockdowns. What he does not do in that calculation is use the information about the high proportion of covid-deaths that are in nursing homes and care homes, which would reduce that number of years lost per covid-death further because we know that it is the relatively unhealthy going into those homes. Also, Martin does not adjust those 5 years by the (health) quality of those years. Had he done those things, he would have arrived at around 3 healthy years of life (QALY) lost per covid-deaths. So he is, once again, loading his calculations in favour of lockdowns. It’s a common refrain.
  5. To get at costs of lockdowns, Martin estimates the GDP effects and unemployment effects of lockdowns, using middle-of-the-road estimates from around the world. He leaves out of these calculations the huge effects of debts incurred during lockdowns and that provide an easier method of calculating the effects of lockdowns as governments themselves publish the changes and expected further changes in debt. Using those official estimates bypasses the need to do own calculations and hence to disagree with governments: by their own estimates one already gets huge effects of lockdowns. So once more, Martin loads the case for lockdowns by essentially ignoring most of the costs already incurred.
  6. Martin does not calculate many known negative direct effects of lockdowns, such as disruption of IVF services, the mental health costs, the human capital loss of disrupted schooling, the costs of disrupted health services, etc. Martin hence in a way does his very best to underestimate the costs of lockdowns and to overestimate the benefits, still arriving at a cost-benefit ratio of about 20 to 1.