Inflation, Energy and Logarithms
(… brief educational moment on the presentation of data …)
Using the Pink Sheet and the World Development Indicators I built this graph that compares the (global) energy price index with the Italian consumer price index:
It is noted that the energy price index is more volatile than the consumer price index and the correlation between the two indices is relatively weak. Since the two series both have an increasing trend, their correlation must be filtered on trend-free series (taking the first differences or growth rates), otherwise a spurious correlation would be obtained (the two series would appear to be related simply because they are both increasing, but this correlation could be illusory, as in these cases ). The correlation between the growth rates is 0.33.
There is a problem, however: the energy price index is in dollars, but the Italian consumer price index is in the national currency, which has never been the dollar! We can convert the energy price index to the national currency by multiplying it by the "Italian currency/dollar" exchange rate (in LCU per US$, local currency units per dollar, the cost in Italian currency of one dollar). Since the prices are expressed as indices, we can rebase the exchange rate to be 1 in the base year, so that the indices continue to be 100 in the base year. The result is:
and the correlation (calculated on growth rates) increases to 0.41. Note that having expressed energy prices in national currency, rather than in dollars, does not alter the profile of the series so dramatically, contrary to the narrative according to which when there was the lira it took wheelbarrows of banknotes to get a barrel of oil (if this were the case, in this graph the cost of energy should shoot up during the years of the lira, and then fall, but instead its profile is very similar to that of the cost expressed in dollars).
In any case, the energy price situation looks much more dire towards the end of the graph than towards the beginning. This is also an optical illusion: it depends on the fact that 1 is 10% of 10 but only 1% of 100: at the beginning of the graph, starting from low values, even imperceptible variations were actually percentage-relevant. This is remedied by taking the logarithmic scale, which transforms the data so that the slope of the curve corresponds to the percentage growth rate of the curve itself:
Framed (correctly) in this way, we see that the oil shocks of the 1970s were much more devastating phenomena than the last energy crisis that caused us so much suffering.
Moreover, one could also see this by taking the growth rates of the variables:
Here we note two things: that the variations in energy prices in the 70s reached peaks of 250%, while the most recent episode sees an increase of around 75%: let's not complain too much! The second interesting thing is that the inflation response to the supply shock is substantially proportional over time. With a peak growth of energy prices around 200% there was a peak of inflation around 20% and with a peak growth of energy prices around 70% there was a peak of inflation around 7%. The pass-through from global energy prices to consumer prices has remained the same, around 10%. What has changed, evidently, is the persistence of the inflation response to the shock: in the 80s the decline in inflation was slower, it took a decade, despite the oil countershock of 1986. Nowadays it has been almost immediate, as the news also says.
The greater persistence is attributed to the operation of wage indexation mechanisms. What strikes me (I had already told you about it), however, is that the pass-through has remained substantially identical. In short: if today we were to incur a supply shock like those of the 1970s, that is, if the price of energy tripled, we would again have double-digit inflation at 20%.
You just need to know.
(… I sleep on it …)
This is a machine translation of a post (in Italian) written by Alberto Bagnai and published on Goofynomics at the URL https://goofynomics.blogspot.com/2025/06/inflazione-energia-e-logaritmi.html on Thu, 05 Jun 2025 18:58:00 +0000. Some rights reserved under CC BY-NC-ND 3.0 license.