We have no theory of inflation

Inflation is the biggest debate in macro. The models don't work.

Forecasting is hard. Forecasting inflation is especially tricky. And economists are, in the main, bad at it. This is, as one might expect, an especially unhappy problem for the economists working in central banks who are guiding policy which is itself targeting a given rate of inflation.

Back in December last year the Fed’s median forecast for PCE inflation in 2021 was 1.8%. In March that was nudged up to 2.4% and then in June up to 3.4%. It is now 4.2%. Over the same period their median forecast for 2022 has risen from 1.9% to 2.2%. The Bank of England’s and the ECB’s numbers have followed a similar path.

Of course, pandemic related disruptions and extreme supply chain turbulence explain much of the forecasting miss over the last year. But while 2021 may be an extreme case, inflation forecasting has been, at best, hit and miss since at least the financial crisis. In 2019 the ECB looked back over their own inflation forecasting of the previous decade. The track record they found is hardly great.

The projections made following the financial crisis significantly underestimated inflation, which was expected to fall more significantly in response to the additional economic slack during this period, a feature common across advanced economies at the time. In the ensuing years, particularly between 2014 and 2016, inflation surprised persistently on the downside.

Given the poor run of forecasting, it is comforting that the economics profession is beginning to rethink its approach.

The Fed recently published a paper from Jeremy Rudd, a senior official and inflation forecaster, snappily entitled “Why Do We Think That Inflation Expectations Matter for Inflation? (And Should We?)”. It’s a bracingly good read, full of what where previously seen as heretical thoughts.

The second footnote - “I leave aside the deeper concern that the primary role of mainstream economics in our society is to provide an apologetics for a criminally oppressive, unsustainable, and unjust social order” – is the most eye-catching line but, unless I am seriously underestimating the radicalism of the leadership of the Fed, probably not the most significant for policy.

Before turning to Rudd’s paper on why inflation expectations are not an especially useful guide for policy makers, it is worth understanding both the wider context and their importance to central banks in recent years.

At last week’s ECB get-together for the great and good of central banking in Sintra that context was neatly provided by Charles Goodhart, a professor at the LSE and former member of the Bank of England’s monetary policy committee.

Goodhart, as only the truly great and good of the profession can get away with, did not write a paper but instead riffed off a few slides.  The video of the relevant session is well worth a watch and Goodhart begins about 15 minutes into panel 1.

Goodhart opened with the remarkable statement that “the world at the moment is in a really a rather extraordinary state because we have no general theory of inflation”.

As Goodhart explained, once upon time there were once two competing theories: the Friedmanite monetary theory that inflation was the result of too much money chasing too few goods and the Philips Curve theory that postulated a relationship between inflation and unemployment.

Both theories have broken down empirically over the past three decades. For myself, both are nice examples of the notion that it is better to think in terms of mechanisms (and ask if those mechanisms are still working) rather than models.

The replacement, in central bank land, has been what Goodhart dubbed a “a bootstrap theory of inflation”: that as long as inflation expectations remain anchored, inflation itself will remain anchored. With a wonderful turn of phrase, Goodhart went on to call this “a very weak reed”. As he demonstrated with his first slide, inflation expectations are reasonably backward looking and people tend to extrapolate their recent experience into the future.

Rudd’s paper, which Goodhart went on to quote, picks up the story from this point.

It is easy to see why such an inflation model is appealing to central bankers. As he argues:

For a central bank with a price stability mandate, monitoring measures of inflation expectations can provide an important gauge of how well the monetary authority is meeting its goal, while attempts to shape the public’s inflation expectations through central bank communications and policy actions will represent time well spent.

In such a world the job of central banking would be quite straightforward: as long as policymakers can credibility commit to take the necessary steps to keep inflation at a given rate – say, 2% - over the medium to longer term then firms and households would set prices and bargain for wages based on such an expectation. Shocks (like, for example, a pandemic) might temporarily cause inflation to rise or fall but it should return to its expected path. Expectations become almost a self-fulfilling prophecy.

Rudd though, it seems, is not one for telling reassuring stories. After demonstrating that the theoretical case for expectations determining actual inflation was always weak (surely for example firms and workers care more about short term inflation when setting prices and bargaining for wages and yet the theoretical case, in policy circles, has usually rested on long term expectations) he goes on to demonstrate that the empirical case has also always been weak. He does not pull his punches.

…it nevertheless remains the case that we have nothing better than circumstantial evidence for a relationship between long-run expected inflation and inflation’s long run trend, and no evidence at all about what might be required to keep that trend fixed

Right now, the debate about how transitory or temporary the global spike in inflation will be is the hottest topic in macro. Between them, Goodhart and Rudd have done a good job of demonstrating that the best answer might be “we don’t know”.  None of the existing models provide a solid basis for forecasting. Many of the people claiming that inflation is definitely staying high have been saying much the same thing for a decade or more, while many of those insisting that it is transitory failed to spot the coming spike.

In the weeks ahead Value Added will try to step away from the models and think through the mechanisms which will determine the path of inflation.  

The first subscriber only posts on Value Added will be coming next week. If you’re enjoying this newsletter please do subscribe.