I am working through my inflation book manuscript, and have not had much time to think about new articles. I added a subsection on producer price indices, and decided to go use that as my latest article here. I have added extra comments that is not in the manuscript since they are tentative.
There are multiple types of price indices, although the focus tends to be either on consumer prices indices (typically what is referred to when discussing “inflation”), or else the GDP deflator. One less popular type of price indices are producer price indices, which are based on prices of goods and services purchased by businesses and not finished goods purchased by households. Since these prices include many raw materials and lightly processed goods, price volatility is much higher than for consumer goods. The figure above demonstrates this by comparing the annual inflation rates of the headline consumer price index and the producer price index (PPI) for the United States.
One interesting theoretical wrinkle to producer price indices is the question why they would diverge from consumer price indices if we are to believe aggregated models where “expectations” are all important? Are producer prices the result of different inflation expectations? Why the gap, if the central bank is allegedly stabilising expectations? The fun thing about eyeballing the above chart is that deviations between the CPI and the PPI appear to have increased in the low inflation era of the 1990s, when central banks strained their elbows from patting themselves on the back with regards to stabilising inflation expectations.
Since producer prices are “upstream” from consumer prices, one might hope that there is leading information for the CPI from PPI data. In practice, the volatility of the PPI makes it hard to extract too much information that is already not included in commodity prices (and which will show up the food and energy components of the CPI). My feeling is that the producer price data is probably more of interest for industry analysis, as the sub-components of the PPI offer some information for sectoral pricing power.
In my first post-academic job, I spent a lot of timing creating charts and building indicators for economists that were doing market/economic commentary. While doing that, I spent a lot of timing looking at various time series. Although using PPI to get some leading information on the CPI sounded good, I eventually found that the noise dominated whatever signal the PPI contained. However, I did not chase after looking at sectoral modelling, which might have been more fertile ground.
Finally, one thing that I find surprising is that the inflation bugs do not highlight the PPI more. The eye-popping annual inflation rates should be useful for scaring members of the public. Of course, they would need to be selective – one would need to stop referring to the PPI when it has 10% drawdowns, like we see at the end of the figure.