When a tail event occurs, past patterns are a poor guide to future performance.  Supply shocks produced by the pandemic stressed inflation far more than anticipated.  An advancing myopia made it difficult to set effective monetary policy.  British inflation ultimately exceeded 11%, the highest peak seen in a developed country. 

This cost British consumers dearly in purchasing power.  The rapid rise in interest rates that was required to quell inflation brought economic growth to a virtual standstill.  Problems in forecasting created very real costs for the country.

In response, the BoE commissioned a review of its forecasting process and engaged Dr. Bernanke for the exercise.  Earlier this month, his team issued their findings.  Among them:

• The infrastructure supporting the forecasting process must be current, so that more advanced techniques (like AI) can be employed.  Special attention needs to be paid to the quality and organization of data needed to perform analysis.
• Active engagement between model operators and subject matter experts is essential to good results, especially during times of rapid change.  Economic rules of thumb like the Phillips Curve can lose their effectiveness over time, threatening the accuracy of projections. Keen observers of conditions can initiate changes to code in an effort to keep algorithms current.
• Forecast errors should be studied transparently and should guide potential changes to models.  Throughout 2020 and 2021, the BoE was underpredicting inflation, but projection approaches weren’t adjusted sufficiently to correct variances.

• The BoE was encouraged to drop their “fan chart” which illustrates the uncertainty of outcomes.  The range in the chart is so wide that it makes almost any outcome seem possible.
  
  Instead, the Bank was encouraged to consider alternative scenarios when setting policy, as opposed to focusing on a central case.  Monetary decisions may be informed by a desire to manage the risk of an impactful outcome that may be less likely.

In a statement, the Bank of England said that it “is committed to action on all…of the Review’s recommendations.”  In the meantime, it behooves all of us who run models to read the report and consider whether its observations might be worth following up on. In particular, the dynamics that drove economic outcomes through 2019 have changed considerably since then; models must adapt to the new reality.

I couldn’t imagine what it would be like to have Ben Bernanke review the models we run here at Northern Trust.  You talk about intimidating…I would much rather warble “Shake It Off” on stage during the Eras Tour.