Exploring Excess Death (part 2)

I investigate alternative methods for estimating excess death using ONS weekly registration data for 2010 - 2023

In part 1 of this series I was forced to take an unconventional approach to calculating excess death; unconventional, that is, in that I didn’t adopt the usual ONS method¹ of subtracting prior 5-year means from observations on a matched week-by-week basis owing to the nature of the data in my pantry, being a short-span daily series.

The tastiest bit of this experimental baking consisted of deriving a year-long seasonal template from smoothed historic data and stitching this together in repeated fashion to form a baseline, thence to apply a modicum of ARIMA modelling. This got me to thinking about adopting a similar approach for the ONS weekly registered deaths series.

In my pantry I have weekly registered deaths for England & Wales going back to 1982, data for which was downloaded from the main ONS repository that goes by the main page title of All Data Related To Death - does what it says on the tin!

At the time of analysis the latest file for weekly registered deaths came in a rather di…