Exploring Excess Death (part 3)

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

Right at the end of part 2 I suggested there is a better way to derive baselines. The ONS themselves are already on to this and had the cheek to publish this report whilst I was slap bang in the middle of typing up my investigations! If you’ll excuse me I’ll ignore these new-fangled developments and continue typing what I intended to type, and I shall start with the notion of exorcism…

Exorcising The Pandemic

What I am going to do here is reach for that all-time favourite spanner of mine for time series work: Autoregressive Integrated Moving Average (ARIMA). I am going to use the period 2010/w1 through to 2023/w52 and I am going to look at weekly all cause deaths in England & Wales. I am also going to switch on outlier detection in all guns blazing mode. Believe it or not there are no less than 7 outlier types defined within the ARIMA module of my stats package, and here they all are:

The idea here is to see what an ARIMA model throws up by way of outlying weekly counts, and in doing so …