Rational inattention at the core of central bank strategy

The origins of inflation are not always well understood. While some economists had warned of impending inflation as early as last year, few had anticipated the dramatic shift in spending from services to goods and the effects of such a shift. The fact that inflation expectations remain subdued is likely due to rational inattention. In the absence of monetary tightening, we believe that inflation expectations would inevitably get de-anchored.

The return of inflation has become the central economic issue. Public policies and financial markets depend on its level and expected trend.

Yet its origin is not always well understood. The most striking fact since the beginning of the year is the forecasting error made by professional economists. In recent months, they have repeatedly revised their inflation forecasts upwards and their economic growth forecasts downwards, first for 2022 and more recently for 2023.

While some economists had warned of impending inflation as early as last year, few had anticipated the dramatic shift in spending from services to goods and the effects of such a shift in an economy where labour shortages in some sectors have led to persistent supply constraints. Recall that in the spring of 2021 the Fed was forecasting peak inflation at 2.4% and that in Europe inflation was not even a concern last autumn.

The combination of supply and demand shocks since the Covid-19 crisis, as well as policy mixes put in place, have rendered econometric models inoperative. The experience of stagflation in the 1970s had already led to a reconsideration of some of the assumptions used in the models at that time. Robert Lucas’ critique (1976) suggests that the use of parameters based on past experience does not allow for an assessment of the effects of changes in macroeconomic policies. If these policies are changed, then the way in which expectations are formed may change, so that the forecast made using a model calibrated with different policies is inaccurate.

The relative importance of aggregate demand and supply shocks differs across countries. In the recent period, aggregate demand factors have played a more important role in explaining US inflation, while negative supply shocks – supply bottlenecks and energy price shocks – have tended to play a dominant role in Europe, at least up to now. In Europe, one of the first sources of forecasting error was the trend of energy prices, especially natural gas prices, which is an exogenous (unpredictable) factor for economists. Initially, the price increase was therefore seen as a temporary supply shock that would dissipate as energy prices stabilised and supply chains normalised.

As wages were not indexed to inflation, there was no fear of a self-sustaining inflationary mechanism at the beginning of the year.

Energy prices are the trees that hides the forest. While increased consumer spending has brought production back to pre-pandemic levels, this rebound has coincided with supply chain problems. The result has been a persistent mismatch between demand and supply, which ultimately leads to more persistent, widespread, and higher inflation than was anticipated until the spring. As the months went by, it became increasingly clear that demand factors were predominant in the US, where inflation had already started to materialise last year, with a very tight labour market and wage increases. Excess demand was largely due to the expansionary policy mix during the Covid-19 crisis. The US case should have been a wake-up call earlier.

Against this backdrop, it is striking to observe that the major central banks are paying sustained attention to the concept of rational inattention¹, explicitly put forward by Jerome Powell at Jackson Hole. This concept allows us to understand how real inflation influences (or not) expectations.

In a nutshell, rational inattention is explained by the fact that economic agents have a limited capacity to understand interactions between economic variables. Their expectations evolve only gradually and insufficiently due to uncertainty about upward signals. In other words, inflation expectations are a lagging variable!

Economic models have been enriched over the last 10 years by taking into account this inertia of expectations. This literature is taken into account by central banks. Less accommodative monetary conditions (real rates are still very negative) are necessary to anchor long-term inflation expectations, even at the cost of a recession.

_{1 The ECB devoted a working paper to a review of the literature on this subject in June 2021.}