Recent Updates

  • Brazil: PPI (Sep)
  • Israel: Credit Card Purchases (Sep)
  • Turkey: Foreign Trade (Sep)
  • France: Registered Unemployed & Job Vacancies (Q3)
  • France: INSEE Monthly HH Survey (Oct), Registered Unemployment and
  • more updates...

Economy in Brief

German PPI Surges to Largest Gain Since Early '70s Oil Shock
by Robert Brusca  September 20, 2021

Germany's headline PPI surged 1.6% month-to-month in August. From December 2008 through December 2020, the average year-over-year percentage PPI gain was half the percentage gain of this month. The PPI excluding energy is up by 1% per month or more for four months running. During this same four-month stretch, the CPI ex-energy averages monthly gains of 0.3% month-to-month.

The small table (below) shows the relationship between the PPI ex-energy and the CPI ex-energy as well as the two headlines. The ECB targets the HICP measure which is a stylized CPI measure on the same definition for all of the EMU. What this table demonstrates is that the correlation between the CPI ex-energy and PPI ex-energy is positive but not overly large (correlation = 0.43 which is also an R-Squared = 0.18).

The standard deviation shows that the PPI is nearly five times (4.7) more volatile than the CPI on an ex-energy basis. So, when the PPI ex-energy moves, there is not necessarily a presumption that the CPI ex-energy is going to move or to move anything like the PPI.

Looking at percentage changes over two years the correlation between the two ex-energy measures is negative and over three years it is a larger negative figure. This suggests to me that the PPI may have an outsized reaction to initial effects of various sorts that hit it with a high impact but then unwind in the coming period, possibly resulting in net negative price responses. Remember these correlations are taken from a period when inflation has been exceptionally low and well behaved. It has been a period in which, if inflation flared, it then went away. It is not exactly in line with the experience of most of the previous post-war period.

If we looked at correlations from 1972 to 1985, we would certainly find different results. I have done that, in fact, and I do find different results. But even in the earlier period, the correlation between the two ex-energy measures is only rises to 0.59 (from 0.43) for an R-squared of 0.34. The PPI and the CPI may have some common DNA but not enough to be of the same species. But the environment from which the correlation is extracted also matters. If we look at the two- and three-year correlations for the earlier period, they are about the same as for the one-year correlation and that is a huge difference with what the data are saying right now, when those correlations are negative.

Chart Germany's Inflation Trends puts the two ex-energy measures up against one another on one graph with one scale. The far greater PPI variability is evident without any statistical analysis and the fact that the PPI often moves and sometimes moves sharply without impact on the CPI is also clear. The ‘common' rise in the two measures since 2020 may be related less to a true statistical covariation that to the fact that each of them was struck with the same distortion: the arrival of Covid that depressed the price base for each at about the same time seems to have led to an excess rise in each at about the same time. That conclusion argues for seeing the progression of the CPI and the PPI as likely separate events even if they continue to show some similarity for a while because of their common experiences with virus distortion. That impact will fade with time.

close
large image