1. Why is core inflation so low in the #Euro area despite a strong recovery and dynamic wages?
Thread on our #BdfEco piece with @diev_pavel and A. Lalliard
publications.banque-france.fr/en/why-have-st…
Thread on our #BdfEco piece with @diev_pavel and A. Lalliard
publications.banque-france.fr/en/why-have-st…
2. With my colleagues, we have been defending a plain old conservative #PhillipsCurve for several years.
3. But there is no way you can understand recent core inflation with a #PhillipsCurve: it’s been perfectly flat, while unemployment has dropped quite substantially, hence the #PhillipsCurveWar between for instance @farmerrf and @ojblanchard1 
4. And yet, there’s a clear Phillips curve for wages, as I’ve noted on several occasions. Average compensation accelerated as unemployment decreased. True even in Germany.
5. As @dandolfa would say, this is just a statistical correlation that could be consistent with different theories, but we view it as providing support to the conventional view that slack, measured by unemployment, causally affects nominal wage inflation.
6. So, why didn’t these accelerating wages lead to accelerating core prices?
That’s what the paper is about.
We propose an algebraic decomposition of core inflation to understand why it diverges from wages.
That’s what the paper is about.
We propose an algebraic decomposition of core inflation to understand why it diverges from wages.
7. Core inflation can be broken down into 4 main components:
(i) unit labor cost (= wages adjusted for productivity)
(ii) margins
(iii) *core* terms of trade
(iv) and the relative deflator of consumption wrt other demand components
(i) unit labor cost (= wages adjusted for productivity)
(ii) margins
(iii) *core* terms of trade
(iv) and the relative deflator of consumption wrt other demand components
8. So what happened in the last few quarters?
Unit labor cost inflation increased, both b/c of wages and b/c productivity slowed down.
Unit labor cost inflation increased, both b/c of wages and b/c productivity slowed down.
9. But corporations compressed their margins, which adjusted to slower productivity growth. This is usual along the productivity cycle (see chart). Unit labor cost did not fully pass thru to domestic prices (in other words, inflation of the GDP deflator was somewhat dampened).
10. But there’s more. Core inflation is also partly imported, either directly or thru input/output linkages.
Because the € appreciated in 2017-18, this imported part was weak. In our breakdown, core terms of trade improved.
Because the € appreciated in 2017-18, this imported part was weak. In our breakdown, core terms of trade improved.
11. The last source of divergence b/w labor cost & core inflation is that wages concern all industries while core inflation is only made up of consumption goods.
12. But the price of construction investment has been much more dynamic than that of consumption.
This is mainly driven by the construction boom in Germany.
This is mainly driven by the construction boom in Germany.
13. So, some of the domestic inflation you don’t see in consumption prices is actually to be found in construction prices!
14. Some numbers for 2019Q1:
2.6% of unit labor cost vs 1.0% core inflation, a 1.7% difference
1.1pp comes from lower margins
0.4pp comes from core terms of trade
0.2pp comes from the increase in the relative construction investment deflator
2.6% of unit labor cost vs 1.0% core inflation, a 1.7% difference
1.1pp comes from lower margins
0.4pp comes from core terms of trade
0.2pp comes from the increase in the relative construction investment deflator
15. Bonus: we apply the decomposition to German inflation since 1999, where there’s no visible link b/w core inflation & unemployment.
You might think it could be due to labor market reform, but it’s hard to reconcile with the nice wage/unemployment correlation in the data.
You might think it could be due to labor market reform, but it’s hard to reconcile with the nice wage/unemployment correlation in the data.
16. Our decomposition sheds light on what was going on.
Before the crisis, core inflation seems to have been driven by margins more than by labor cost.
And price-setting behavior seems to have changed after the crisis, with a come-back of labor cost.
Before the crisis, core inflation seems to have been driven by margins more than by labor cost.
And price-setting behavior seems to have changed after the crisis, with a come-back of labor cost.
17. (Kind of important to have in mind if you estimate a model over the whole period: there’s likely to be a structural change around 2008).
