US, EA, Japan: The Impact of Oil Prices on Inflation

TL;DR: The current movement in oil prices is likely going to add only 10-20bps to headline inflation. We show our estimates and discuss why the simple rules-of-thumb that most people have in mind are too simple to approximate the current dynamics.

One of the questions we have been receiving most frequently in recent days, given the rather sharp moves in oil prices, concerns the likely impact on headline inflation. The first point to recall is that the relationship between oil prices and gasoline prices, using the United States as an example, although the same logic broadly applies to the other economies we monitor, is a relationship in levels.

This means that, when an oil shock occurs, what matters is the dollar change in the price of oil, rather than its percentage change. In other words, the relevant question is how many dollars the price of oil rises or falls, not how large that movement is in percentage terms. This level relationship is clearly visible in the first figure, which shows the link between oil prices and retail gasoline prices.

 

It follows from the discussion above that a 10 dollar per barrel oil shock translates into an almost proportional increase in gasoline prices, typically in the range of 21 to 24 cents per gallon. The inflationary impact of this shock, however, varies substantially depending on the initial level of gasoline prices, or equivalently on the initial level of oil prices, because the same absolute increase in gasoline prices implies a very different percentage change depending on the starting point.

For this reason, the table below reports the estimated impact on inflation across a range of gasoline price levels, from 2 dollars per gallon to 5.4 dollars per gallon. As the table shows, when gasoline prices are very low, the impact on inflation is much larger and can be estimated up to five tenths in U.S. CPI space. By contrast, when gasoline prices are already much higher, the same oil shock can have an impact on headline inflation of only around one tenth (in PCE space).

For precisely this reason, it is not especially easy to generalize the relationship into a simple rule of thumb, unless one is willing to rely on an average across all possible elasticities (which is risky because they are very different!). This is why we constructed the visualization shown in the two figures below, which illustrate the impact of a 10 dollar per barrel oil shock on U.S. headline inflation in both CPI and PCE space, across the full range of possible initial combinations of oil prices and gasoline prices.

One important assumption should be made explicit. In this exercise, we assume that the shock is permanent, in the sense that it shifts the entire oil price curve upward for at least 12 months. Under current conditions, this implies that an additional 10 dollars relative to today’s level would raise headline inflation by around 24 basis points in CPI space and by roughly 17 basis points in PCE space.

That said, these estimates are very likely to represent upper-bound estimates, for the reason explained below.

What should we expect at this point, and why? As discussed above, the estimates reported so far should probably be interpreted as upper bound estimates. On the one hand, they have been produced by calculating only the direct effect of commodity prices, and more generally of energy goods, on headline inflation. On the other hand, the exercise assumes that the entire oil curve shifts upward by 10 dollars per barrel, whereas the actual market move has been more concentrated on the front end of the curve, with the curve remaining in backwardation.

For this reason, the actual impact has been smaller than the mechanical estimates would suggest. This is also what we observe in our models, which we do not report here for reasons of space. The models have indeed moved higher, but by less than one would infer from a mechanical application of the numbers above. Over the past few weeks, oil prices have increased from roughly 70 dollars to roughly 90 dollars per barrel, a move of around 20 dollars. Taken literally, the estimates above would imply an upward revision to the inflation forecasts of as much as 50 basis points. That is not what our models produce, mainly because the shock has not shifted the entire curve upward by 20 dollars per barrel and has instead been largely concentrated in the front end (on top of this, the average price over the month has increased by about 10US$ per barrel, not 20US$).

The broader point is that the rules of thumb often used to assess the inflationary impact of oil shocks are too crude relative to the actual mechanics of the problem. The impact on inflation depends on the nature of the shock, on the initial level of oil and gasoline prices, and on the movement of the full oil curve rather than only the marginal move of the front contract. Once we account for these elements, including direct and indirect effects and the movement of the entire curve, our models estimate that the oil shock observed over the past few weeks has added roughly 1 tenth (or 2 tenths) to headline inflation, depending on the inflation measure and the region considered.

This is, in our view, the more prudent and more defensible estimate. It can of course turn out to be a bit too low, but an estimate of 50 to 60 basis points, which we have heard from some clients, looks way too high because it relies on technical assumptions that do not match the actual configuration of the shock, as discussed in this note.

Contact

Want something more tailored?

We provide tailored consulting on ad-hoc projects.