An imp. Lesson for policy economists. Data driven predictive procedures that may( and I’m a known sceptic) that may work in BAU do not work now. 1/n. https://fivethirtyeight.com/features/coronavirus-case-counts-are-meaningless/

Inductive procedures: Analytical framework->Simulations->inductive validation or refutation->Policy action.
This takes time but negates preditctive error. Which is large in data driven deductible inferences when time and space correlations and covariances cannot be assumed.
2/n

That’s exactly what is happening in the Oxford/Imperial debate Both are using deductive procedures to make claims But policy analysts are trying to inductively validate these claims 3/n

Inductive skills are what doctors use. They require strong analytical frameworks and intuition but have atrophied. Cheap data easy computing tech incentivised theory and intuition light data driven analysis. @kaushikcbasu @_miller_mark @arvindsubraman @andymukherjee70 @maitreesh