Mapping inflation – estimating the direction of a cloud of a myriad of individual prices

Macro-economics is a science of totals – not of averages. One of these ‘totals’ is the ‘price level’. Beate Reszat reminds us of the complicated nature of this aggregate: a fuzzy maze of a myriad of individual prices wich are interrelated via as well the demand as the supply side of the economy but which also change because of technological and product changes. Which leads among other things to the question if inflation, as we measure it, is the right metric to investigate if the total cloud of interconnected prices goes up or down.

I’m less concerned (not: ‘unconcerned’) than Beate is about the measurement of individual prices. Tough I do think that, for instance, the people of the Dutch Centraal Bureau voor de Statistiek should publish a kind of manual about this it does seem that they are doing a good job when it comes to dealing with, for instance, changes in the ‘quality’ of products. A whole array of formal and tacit methods are used to do this (phoned them about this some time ago). But Beate is totally right that we should not just look at the change of the ‘average’ (i.e. headline inflation, a ‘measure of centre’ in statistical phraseology) but also at individual prices, with some kind of ‘measure of spread’. Eurostat data enable us to do this in a crude way. See graph 1 and 2.

.

The Eurostat datbase contains data on 96 three and four digit ‘COICOP’ groups, i.e. subgroups of items included in headline inflation, like ‘fruit’, CP0116, or ‘materials for the maintenance and repair of the dwelling’, CP0431. Graph 1 shows the spread of the year-on-year price increases of these groups in February 2013 and February 2014 (mind that this graph does not rank price increases according to COICOP group but according to the magnitude of increase/decrease). Graph 2 shows the ‘change of the change’ in prices of individual COICOP groups between February 2013 and February 2014 (i.e. the YOY change in February 2014 minus the change in February 2013).Both graphs show that disinflation is a very real phenomenon. But especially the second graph also shows that despite this disinflation there are also some products which in February 2014 show higher prices increases (or lower decreases) than in February 2013. The cloud is coming down  going up less fast. The graphs also show, imo, that price increases are interconnected. There is something like an aggreagate  price level! Beate is right to state that we do not understand this interconnectedness very well. See however this graph which suggests that average hourly wages play a key role – as long as the Euro does not depreciate while Eurozone wage increases are as low as they are and productivity keeps increasing as fast as it does  (+1%, year on year) Eurozone inflation will be below the ECB target for a long time to come, barring things like a Crimean War II.

This exercise should be done for every Eurozone country and with even more detailed prices (leading to a trimmed mean inflation rate or something like that), I leave this to Eurostat.