Revised data for CapitalSpectator.com’s coronavirus (Covid-19) forecast, based on an 8-model combination, continues to show that global reported cases are on track to rise further in the near term (using the median point forecast as proxy).
The main focus of this proprietary forecasting project is to identify when a downside turning point appears likely, based on the median forecast. For the moment, no sign of peaking on the horizon.
Note, too, that the total number of reported cases has exceeded the last two vintage point forecasts (Feb. 27 and Mar. 4), which implies that the trend remains skewed to the upside. In short, the Covid-19 crisis still looks set to get worse before it gets better.
Is Recession Risk Rising? Monitor the outlook with a subscription to:
The US Business Cycle Risk Report
Why are you forecasting linear growth? Epidemics normally grow exponentially until they run out of uninfected people. This one appears to have a doubling time between 3 and 5 days outside of China. I don’t think most of the world will be willing or able to shut down the way that China has, so I don’t think there is any chance of the doubling time going above 10 days until at least 1/3 of the world has been exposed. The official number outside of China is about 30k, but the real number is probably between 50k and 100k. 2 billion is still 12-15 doubling periods away, or about 2-3 months.
I’m not passing this off as the last word on estimating the future but, rather, a first step–a baseline–for tracking the trend. I, for one, want to track how the actual numbers stack up and how the near-term future compares. But I agree: exponential-based analytics are relevant too, and on that front I’m crunching data, along with looking at mortality rates. Lots of analysis to run. For the moment I’m only sharing the bare bones publicly.
–JP