Thread by @MikeOtsuka

What's the most defensible inflation index? For pay negotiations (as opposed to tuition fees), @UCEA1 affirms CPIH as the right measure of inflation. @ucu affirms RPI. We can see why from this graph. RPI is highest (blue line); CPIH lowest (purple line). 1/

@UCEA1 @ucu The 'right' index, for the purposes of pay negotiations, however, is neither CPIH nor RPI. No CPI-based index is appropriate for this, since CPI measures the aggregate rise in prices across the economy as opposed to the rise in the cost of living for members of a household. 2/

@UCEA1 @ucu RPI is more sensitive than CPI to the cost of living for a typical individual. RPI, however, is widely regarded as statistically unsound, on account of the 'formula effect' I discuss in this blog post👇. 3/
medium.com/@mikeotsuka/ho…

@UCEA1 @ucu There is, however, a version of RPI, once known as RPIJ, which strips out the formula effect that renders RPI unsound in a manner that overstates inflation. RPIJ is represented by the green line on the graph in 1/👆. 4/

@UCEA1 @ucu The ONS maintains that RPI has problems beyond the formula effect. For that reason, it is developing a new set of Household Cost Indices (HCIs) which are meant to best capture the increase in cost of living for members of typical households. 5/

@UCEA1 @ucu HCI is still under development. But a refined set of experimental results has been released 👇, which I represent by the rust line in the above graph👆. Note that HCI runs closer to RPIJ (green line) than to either RPI (blue line) or CPIH (purple line). 6/
ons.gov.uk/economy/inflat…

@UCEA1 @ucu Here I graph how much university pay has declined in real terms (red line), as measured by HCI, since 2005👇. 7/

@UCEA1 @ucu HCI is still under development, not yet updated monthly, and extends only to Dec 2018. But it can be used to indicate how much our pay would need to rise in order to 'catch up' by recouping the erosion of our pay from the 2008 financial crisis to August 2018. 8/

@UCEA1 @ucu By this most relevant HCI measure of inflation, employers would need to raise our pay by 5.5% in order to restore our pay as of Aug 2018 in real terms to its level in Jan 08, which is a baseline that cannot be rejected as cherry-picked. That's the 'catch up' element. /9

@UCEA1 @ucu To 'keep up', we need to ensure that, on top of the 5.5% 'catch up', the employer's implemented 1.8% increase from August 2019 has kept up with the rise in inflation from August 2018 to August 2019. /10

@UCEA1 @ucu By the best existing measure of inflation from August 2018 to August 2019 – RPIJ – inflation rose by 2.1% during that period. So a further 0.3% (2.1%-1.8%) is needed to 'keep up'. 11/

@UCEA1 @ucu Bottom line: by the best measures of inflation, against a non-cherry-picked baseline, 5.8% (5.5% + 0.3%) would be necessary & sufficient to catch up & keep up with inflation. 5.8% is in fact slightly higher than @ucu's pay demand of RPI (Aug 2019 = 2.6%) + 3% = 5.6%. 12/12

@UCEA1 @ucu PS: Link👇to my spreadsheets with underlying data & graphs:
drive.google.com/open?id=1Nl5ZV…

@UCEA1 @ucu PPS: This linked Royal Statistical Society👇submission to a DWP consultation on DB mentions problems w/ RPI & CPI along lines I note👆& endorses the HCI measure I advocate.
rss.org.uk/Images/PDF/inf…

@UCEA1 @ucu Re👇, 5.8% is the *rise* necessary to cancel the fall in salaries against inflation. But, mathematically, the *fall* is lower than the rise necessary to cancel it. The fall of salaries against inflation is roughly 5.5%.

https://twitter.com/MikeOtsuka/status/1184424247348940806