Rarely, if ever, has an Australian Prime Minister relied on statistical modelling as heavily as Scott Morrison. Modelling by the Doherty Institute is the sole piece of evidence on which the Prime Minister has formed the view that it is ‘safe’ to significantly reduce the social distancing measures that have helped Australia keep its death rate so much lower than that experienced in most countries around the world.

It is true that high vaccination rates lead to a significant reduction in the spread of COVID19, hospitalisations and deaths from the virus. But it is also true, as the Doherty modelling makes clear, that lifting restrictions on peoples movement, mixing and mingling when 80 per cent of adults are vaccinated will lead to up to 40,000 Australians per day becoming infected in the months after restrictions are lifted.

The Doherty modelling also makes clear that, even after 80 per cent of adults are vaccinated, lockdowns will still be a common feature in Australia, that large numbers of days will be lost to illness and quarantine and over 760 deaths are expected from the virus.

Perhaps most significantly, the Doherty modelling results are based on the assumption that the effectiveness of the Testing Tracing Isolation and Quarantine (TTIQ) system never deteriorates below the level experienced during Melbourne’s second wave infections that saw daily cases top 700 per day. As the Doherty modelling states:

TTIQ assumptions are based on the performance of the Victorian public health response at the height of the second wave in 2020 as our best estimate of achievable effectiveness at high case loads.

Given that NSW case numbers have already topped 1,000 per day and that the last time the NSW government publicly announced data on unlinked cases more than 800 of their daily cases were unlinked it would seem optimistic in the extreme for Scott Morrison to continue to base his national plan on the assumption made by Doherty, back in June 2021, that the effectiveness of TTIQ would not fall below that experienced during Melbourne’s second wave.

The fact that the Doherty Modelling is based on a wide range of simplifying assumptions is not a criticism of the model, it is a truism. All statistical models, be they economic, climate or epidemiological, are based on assumptions which help users of the model to focus on the linkages between variables. But it is also a truism that when the assumptions in a model diverge significantly from reality the results of the model will be a poor predictor of what will happen in the real world.

The job of the modeller is to make the assumptions they have made clear, and the job of the user of a model is to make the value judgment about the appropriateness of basing significant decisions on the basis of assumptions that might be unrealistic.

The purpose of this paper is to highlight some of the assumption and less commonly reported but important results of the Doherty Modelling. Different users will inevitably make different decisions about whether the simplifying assumptions made in the modelling are sufficient for realistic decision making purposes or not.