MNI INTERVIEW: Models Of Inflation Expectations Need Refining

Central banks should incorporate a more sophisticated model of inflation expectations into their projections, a former head of modelling at the Bank of England told MNI, noting how the distribution of expectations shifted to the upside well before the current surge in headline inflation, even as average expectations remained relatively subdued.

While some economists, including the Federal Reserve Board’s Jeremy Rudd, have questioned the usefulness of inflation expectations, whose failure to rise sharply after the Covid pandemic reinforced arguments against more rapid tightening of monetary policy, research by Francesca Monti, now a professor at Universite Catholique de Louvain, together with the IMF’s Roland Meeks has produced empirical evidence that variations in expectations feed through to headline inflation.

“After Covid, for example, the distribution of inflation expectations started moving upwards earlier than the consensus measure. The point of our paper is policymakers should pay attention to the whole distribution of inflation expectations. It does really matter,” Monti said in an interview.

In the UK, BOE Monetary Policy Committee member Catherine Mann, in a minority pushing for more aggressive early tightening, has placed great weight on inflation expectations. Monti was the chair when Mann delivered a policy speech on Sept 5, in which she argued for “fast and forceful” rate rises.

MODELS REQUIRE WORK

“I think central banks are looking at measures summarising the features of the cross-section of Inflation expectations in a more qualitative way. If you look at the speech Catherine Mann gave, it was really pointing to the fact that the MPC has a battery of measures that they look at,” Monti said. But, she added, central bank models need updating, though she acknowledged that the task of incorporating distribution analysis will take a lot of time and management effort.

“The model building process is long and involves a lot of testing and if you change a model that is already functioning well and is in use there are costs to that … Adoption of new models is a relatively slow process, but I am hopeful this is something that is going to be taken on board,” she added.

One long running question in the research has been the degree to which inflation expectations reported in surveys influence the economic decisions taken by firms and households. Monti says recent research, from, among others, Vellekoop and Wiederholt and Olivier Coibion and Yuriy Gorodnichenko, suggests that expectations do influence decisions.

“There are really interesting papers that are pushing the frontier on this, linking survey data on inflation expectations to administrative data on income and wealth. They find people’s expectations do indeed guide their decisions,” she said.

INFLATION MORE SALIENT

While in times of relatively low and stable price increases firms and households may have little awareness of the current inflation rate and are poor predictors of where it will go, one effect of the current surge is that the public have become much more knowledgeable. (See MNI INTERVIEW: Lagging Fed May Need 75BP Hikes To Catch Up)

“We lived in good times where we could forget about inflation. It was not salient, not on people’s minds. The situation has changed and I think people are going to have better forecasts,” she said.

The evidence that the distribution of inflation expectations matters adds to another puzzle, however. If the simple average of expectations is close to the inflation target but there is a large upside skew, would these expectations be “well-anchored”?

“Well-anchored is a bit fuzzy and difficult to define. For a long time, we considered well-anchored inflation expectations that were consistently above target and above the historical average of inflation outturns. What is different now is that expectations … are drifting further away from target and from their historical averages,” Monti said.