A playground closed during Melbourne's sixth lockdown (Image: AAP/James Ross)

“Get me a one-handed economist,” Harry Truman is said to have demanded. “All my economists keep saying ‘On one hand…’ then ‘but on the other…’”

The president’s frustration is a familiar one. Who doesn’t want an unequivocal explanation of a complex, adaptive and uncertain world? Making decisions can be hard. Trade-offs and value judgements are needed. Information is always imperfect.

If variety is the spice of life, uncertainty is the sauce.

Not for the cadre of newly minted economists, epidemiologists, virologists and psychiatrists, though, who seem to be quite certain how to manage the COVID-19 pandemic.

Everyone is permitted to an opinion, of course. But commenting on a virus that may have killed up to 15 million worldwide requires more than a feelpinion and might be better approached acknowledging the complexity, the stakes involved, and the diabolical choices faced by decision-makers. Here are three rules of engagement that might assist commentators in their thinking and their writing, publishers in their decisions on what to publish, and readers in appraising the merit of what they’re reading.

Think about conflicts of interest and perspective

This one seems simple at first. If your income is affected by closures and movement restrictions, it’s a good idea to say that up front (not just in the bio). It gives the reader the information about any real or perceived agenda as they’re reading.

But what about those who are not financially affected by a lockdown or not affected by lockdowns at all? This is me on both counts. My work and income, while certainly affected by the pandemic, are not at all disturbed by a lockdown. Moreover, I’m based in Perth which has been virtually COVID-free compared to the eastern states.

Does having very little skin in the game invalidate my take? Is it a conflict of perspective? In the eyes of some, perhaps.  So for the sake of transparency I suggest declaring one’s personal and professional circumstances (something I will endeavour to do from now on when writing on this topic).

Acknowledge the dynamic nature of the situation

One of the great mistakes of analysing complex systems is assuming things are linear — that patterns don’t change over time and that altering one aspect does not result in adaptation elsewhere. The pandemic is far from static. It’s a complex biological, social and political phenomenon. The virus evolves and adapts, as does individual and group behaviour in trying to live with it. This has several implications.

It’s generally thought that people largely follow instructions and orders on limiting the spread of the virus. It’s also understood that they restrict their movement and social contact regardless of official policy. But does this hold for the third wave? The fifth wave? Lockdowns within months of each other?

The impact on mental health is a strong argument against prolonged restrictions. It should be monitored. But then again, what about the impact of a much higher death toll? Equally, overloading the healthcare system will have deleterious consequences on people’s health.

Meanwhile, the virus is busy adapting. Based on early laboratory work, the emerging Mu variant may be more resistant to vaccines and natural immunity. If this is the case, it will significantly alter the debate on how to best manage the disease.

Use data (without cherry-picking)

Perhaps the most important rule, and it’s quite simple: examine the available data from reputable sources. If they support your take, present them while acknowledging that data is always imperfect and that new data may emerge. If they don’t support your take, perhaps change the take or stop writing.

The Swedish model was recently again favourably compared to Australia’s way of dealing with the pandemic: fewer deaths, more freedoms, lower impact on mental health.

There’s no questions that the human impact of restrictions is real. Children’s mental health, education and social development will be affected. Prolonged school closures may widen social and economic inequality. The true impact will not be known for some time and must be monitored. Suicide rates in Victoria, which has arguably seen the toughest lockdowns, thankfully do not appear to be increasing but they are a blunt metric.

But the first thing an objective appraisal of the available data shows is that Sweden’s cumulative death toll from COVID-19 dwarfs that of its Nordic neighbours and Australia. In fact, for every Australian death 36 Swedish residents died from COVID-19. There may have been very few deaths in Sweden recently, but this is also the case for its neighbours. It’s cumulative deaths that matter here, not least because the trauma to the community would be considerable.  

(Source: Johns Hopkins University)

Yes, deaths are a blunt metric. COVID-19 has been targeting older people and one can certainly argue that the figure would be less devastating if we counted lost life years. Whether this would outweigh the costs of lockdown (see below) can be debated. On the other hand (sorry, Harry) there’s the costs of hospitalisation, disability and time off work to consider.

In any case, please refrain from using quality-adjusted life years (QALYs). These are designed to assess the effectiveness of medical technologies that extend life and not the impact of a life-limiting disease. For that we have the Disability-Adjusted Life Year (DALY), and studies using this metric are still emerging.

Meanwhile, latest data suggests that Sweden’s economic output has been similar to its neighbours and Australia, which saw the greatest average quarterly growth over the period (0.37% compared to 0.13% for Sweden, 0.19% Denmark, 0.17% Norway, -0.02% Finland).

(Source: OECD statistics)

Yes, data can be interpreted in different ways and cherry-picked. The negative “deaths versus economic growth” correlation did the rounds last year. Typically, countries that didn’t fit the trend were excluded. Currently the situation looks very equivocal using an unbiased sample (OECD countries in the chart below). But remove seven data points (in orange) and one can start to see a pattern. Resist the temptation. (And by the way, such correlations rarely explain much in our complex world.)

(Source: OECD statistics and Johns Hopkins University)

And after all that, the Oxford University Stringency Index suggests that Swedes have, overall, not been as free from restrictions as has been reported, perhaps making the entire argument somewhat irrelevant. Then again, had the NSW government acted more decisively in June this year, or the Commonwealth moved to secure Pfizer doses earlier, we may also not be having this discussion (although, as I’ve argued previously on the latter, securing more vaccines when in such scarce supply could have merely transferred the toll elsewhere).

(Source: University of Oxford coronavirus government response tracker)

(The above index is based on nine metrics: school closures; workplace closures; cancellation of public events; restrictions on public gatherings; closures of public transport; stay-at-home requirements; public information campaigns; restrictions on internal movements; and international travel controls. A higher score indicates a stricter response.)

It’s important to debate managing a deadly pandemic. But the debate will benefit everybody if some basic rules are followed.