Why central banks getting the job done is more important than accurate forecasting

Why central banks getting the job done is more important than accurate forecasting

Originally published on The Australian


By common judgement, central banks everywhere failed dismally in forecasting inflation during the post-Covid period.

Central banks didn’t foresee the sharp rise in inflation in 2021, reacted tardily and raised interest rates unexpectedly dramatically.

That said, none of this actually mattered much for monetary policy settings.

Even a prescient forecast would have led to much the same policy response – a sharp rise in interest rates when supply-side factors (Covid-19 followed by the Ukraine war) made a big rise in inflation inevitable.

But the forecasting errors did do significant damage to central bank reputations, which are the vital anchor for inflation targeting. The public must believe that the central bank has the competence and determination to bring inflation back to target.

The past decade of increasing transparency has left central banks hostage to the intrinsic uncertainty in the economy. Central banks should say less about what inflation will do in the short term. Unfortunately the current international debate is going in the opposite direction.

The Bank of England (BoE) – seen as a failure on all counts – is the first with a detailed post-mortem, carried out by former US Fed chairman Ben Bernanke. On his watch at the Fed, transparency moved from Alan Greenspan’s meticulously contrived obfuscation to the full-frontal revelations of the “dot-plot” guesses of individual Fed board members. This has mainly served to demonstrate the volatility of expert opinion.

Bernanke’s report has been described by the BoE Governor as a once-in-a-generation opportunity to overhaul the bank’s forecasting. But it is a damp squib, with “scenarios” to capture the uncertainty previously embodied in the fan chart – the forecast of inflation surrounded by probability bands.

A few model-based scenarios will not capture the inherent uncertainty of the policy process. The problem is two-fold.

First, the real-world economy is driven by people’s behaviour – variable, fickle, and dependent on Keynes’ “animal spirits” – too complex and volatile to be captured in a mechanical model.

Second, policy is playing out in an evolving environment, subject to unpredictable shocks. Not even the smartest modellers would have incorporated the Covid-19 shock into their forecasts in advance, nor the Ukraine war; they were “unknown unknowns”.

The sudden rise in the policy rate in 2021-2 was more painful because of a much earlier policy mistake, which had nothing to do with faulty inflation forecasts. The US Fed, guided by a model which said that policy should be loosened whenever inflation was below target, drove down the policy rate after the 2008 GFC and keep it near zero for most of a decade.

Other central banks had to follow, to avoid stronger exchange rates and loss of international competitiveness.

Thus when the Covid-19 supply-side shock arrived, not only did low rates require a large step/jump in order to normalise, but borrowers (and financial markets more widely) had adapted their balance sheets and financial engineering to cheap money. Ouch!

The Bank of England promises to do more and better model building, but this mindset may be part of the problem.

Rather than a couple of model-based scenarios, it would be better to use a narrative approach – wide-ranging storytelling using history and words to supplement the models, drawing out the unusual characteristics of the current conjuncture, overlaid by common sense.

In the post-GFC decade, when the model was pushing interest rates ever lower, a narrative would have explored the 1930s when the Fed chairman warned that policy was “pushing on a string”. Common sense might have warned against a zero neutral rate in the post-GFC decade.

In the post-Covid period, the narrative would have drawn on the post-WWII years and the Korean War period, when supply-side inflation passed through quite quickly without the need for a policy crunch.

Looking ahead, the narrative approach would change the tone of central bank communication – away from the current day-to-day fixation with inflation forecasts.

Central banks operate in a world of radical uncertainty. They need a credible policy structure to stabilise price expectations: this is the anchor which will return inflation to the target after a supply-side shock.

This structure should mimic the Mario Draghi approach to policy guidance – affirming that they will do “whatever it takes” to achieve the inflation target. Resolute, clear and unspecific as to detail. That is enough.

No need to provide forecasts of inflation or policy settings: exactly what will be done is, properly, unknowable in advance. To pretend otherwise just leaves central banks open to reputation-sapping errors.


Areas of expertise: Regional economic integration; Australia's economic relations with East Asia; international financial flows and the global financial architecture; financial sector development in East Asia