• rmacculloch

Covid Modeling Forecast Accuracy: An Update

Stuff News ran an article back in March with the headline, "Economic forecasting's spectacular Covid failure". It stated that, "Amid this disruption, economists rapidly slashed their forecasts for GDP and employment. But economic activity has bounced back from lockdown largely unscathed .. Why did forecasters get it so wrong?". Without getting into that debate, but on the topic of forecasting accuracy, how are the Covid Case Number Modelers doing with their predictions? They remain regularly on top of the news headlines and the media seem to hang on their every word.

Yes, these are the folks who forecast that the size of the current "outbreak" would range from 200 cases, which they called the "best case scenario" up to 1,000, being their worst case scenario. As virus numbers stated to soar, Shaun Hendy subsequently revised these forecasts. Toward the end of August, he said the current outbreak was "more comparable to the March-April cluster from 2020 - we could be looking at around 1,000 cases all up".

Today the total number of community cases arising from the outbreak stands at 3,046. If we economists were caught getting our forecasts wrong to this extent, I can't imagine how much fun the media would have at our expense. Yet strangely when it comes to Covid case number modeling they don't seem to mind a bit.