Yes, lockdowns are costly. But the alternatives are worsePatrick Abraham, The University of Melbourne; Laxman Bablani, The University of Melbourne; Natalie Carvalho, The University of Melbourne, and Tony Blakely, The University of Melbourne
Lockdowns are costly. They damage businesses and livelihoods.
Victoria’s recent lockdown cost about $100 million a day in lost economic activity, according to Victorian Treasury. NSW’s current lockdown will cost about $140 million a day, according to AMP. The total cost of current lockdowns affecting Sydney, Darwin, Brisbane and Perth will therefore be in the billions.
Is there another way?
There is broad consensus among epidemiologists that Australia’s strategy of elimination, with hard and early lockdowns, is the best response until the population is vaccinated. But some economists disagree.
“Imagine if lockdowns caused more harm than good,” mused The Australian’s economic correspondent Adam Creighton this week, citing US research that “fails to find evidence that lockdowns saved lives in net terms”. The study has also impressed University of NSW economist Gigi Foster. “We need to stop this madness,” she wrote in The Sydney Morning Herald.
We too have been considering the costs of lockdowns, but have come to a very different conclusion – that “living with the virus” would mean both higher health and economic costs than our strategy of elimination, achieved through border controls and sporadic lockdowns.
How we did our research
Our research (in press at an international peer reviewed journal, but available as a pre-print) has involved modelling four scenarios using data from Victoria’s experience.
Two of those scenarios are elimination strategies – aggressive or moderate. The aggressive approach means implementing a lockdown when COVID cases reach about eight a day, the moderate approach at 30 cases a day.
The other two scenarios are suppression strategies, limiting cases to a given threshold. The tight suppression scenario involves locking down when cases hit about 120 a day, while the loose scenario at about 700 cases a day.
All four scenarios involve some form of lockdown, just as these strategies have in the real world. In countries pursuing suppression, such as the US and Britain, lockdowns have been deployed to regain control of infection rates that have gotten so high that cases requiring hospitalisation threaten to overwhelm the health system.
As the experience of nations such as Britain have shown, getting a workable suppression strategy has been extremely difficult. Measures to beat back the virus have always been temporary. Once restrictions are relaxed the virus has bounced back, meaning more lockdowns.
It shouldn’t be surprising that this approach tends to cost more, as our modelling suggests.
We ran the model a hundred times for each of these scenarios, to capture some of the randomness inherent in the spread of the virus in real life as well as uncertainty about inputs like costs per week of lockdown.
The costs of treating COVID-19 in hospitals were always greater for our two suppression strategies than the two elimination strategies.
Economic costs – measured by effect on GDP – were less clear-cut. However, in 77% of model runs GDP losses were greatest for either of the two suppression strategies.
Other research supports elimination
Our findings are consistent with other new studies, both for Australia and globally.
In a study published last month, researchers from the University of Melbourne and ANU have calculated the total economic costs of unmitigated spread would have been about four to eight times larger than quashing the virus early.
Another study published last month, in the Lancet, compares health and economic outcomes for Australia and four other OECD countries opting for elimination (Iceland, Japan, New Zealand and South Korea) with the 35 OECD nations that have opted for suppression.
Though the authors acknowledge their analysis does not prove causal connection between response strategies and outcomes, all indicators favour elimination. The elimination nations have had a COVID-19 death rate (per million) 25 times lower than the suppression nations, and higher GDP growth through almost every weekly period through to early 2021.
Go hard, go early
So what of the study cited by Creighton and Foster as evidence that lockdowns are not only ineffective but actually may be causing more deaths?
This study measures changes in excess deaths following the implementation of stay-in-place policies in all US states and 42 other countries. It finds extending lockdowns by a week has been associated with a 2.7% increase in excess deaths.
However, since many of these countries implemented suppression strategies, lockdowns were implemented in the presence of high and increasing COVID-19 cases. These high cases flowed on to high mortality in coming weeks. Essentially, correlation does not imply causation.
Significantly, the study notes Australia and New Zealand, two countries that used early lockdowns to eliminate COVID-19, had fewer deaths (allowing for both SARS-CoV-2 and other causes). This is also what you will usually find at our COVID-19 Pandemic Tradeoffs tool, which examines health impacts of different strategies allowing for unintended health impacts of lockdowns.
To put it simply, the costs of lockdowns can’t be calculated in isolation from their role in the strategy chosen to control COVID-19. Both elimination and suppression have lockdowns, but elimination requires fewer lockdowns with better health and economic outcomes.
Australia is having less economic scarring, and a stronger recovery than any other OECD country apart from South Korea. We can thank, in part, high iron-ore prices, but also the relative success of the elimination strategy, which has allowed economic activity to recover strongly following lockdowns.
The lesson is “go hard, go early” – at least in 2020 and until we have higher vaccination coverage. But we’re still a long way from that. Until then the elimination strategy, including early, sharp lockdowns where necessary when contact tracing is unable to “do the job”, remain our best policy.
Patrick Abraham, Research Assistant - Health Economics, The University of Melbourne; Laxman Bablani, Research Fellow, Population Interventions Unit, Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, The University of Melbourne; Natalie Carvalho, Senior Research Fellow, The University of Melbourne, and Tony Blakely, Professor of Epidemiology, Population Interventions Unit, Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, The University of Melbourne