Exploring the Bernanke Review

What did Ben Bernanke have to say about the Bank’s forecasting record and process?

Before answering the question, I thought it worth noting that Ben Bernanke prefaces his report by thanking the people and organisations he spoke to while preparing the report. It was great to see that he ‘name checked’ the National Institute as one of the other UK forecasters whom he consulted and to know that we are considered to be among the experts in this area!

The report started with a general section on how central banks go about forecasting and what they are trying to get out of it before moving on to describe the current process at the Bank of England. The report noted that the Bank’s recent forecast performance had been poor, but that it was no worse than the of many other forecasters, both in central banks and the private sector. As I said in my recent article in Central Banking, no-one could have predicted the Russian invasion of Ukraine and the sharp increase in energy and food prices that followed it! That said, Bernanke emphasised the importance of understanding where forecast errors came from so that you could adjust policy as the economy evolved. That said, he noted that the ‘incremental’ approach to forecasting – where forecasters start from the judgements made in the previous forecast and make further adjustments rather than starting from a ‘clean’ forecast each time – had meant that policy had been slow to react to the shocks that affected the UK economy over recent years.

In terms of where improvements could be made, the report found deficiencies within the Bank’s forecasting infrastructure, that is, the data management and other software tools, as well as the models used in the forecast process, particularly the central forecasting model, COMPASS. The report noted that ‘these deficiencies in the framework, together with a variety of makeshift fixes over the years, have resulted in a complicated and unwieldy system that limits the capacity of the staff to undertake some useful analyses.’ The report made a series of recommendations to deal with these issues. It is interesting to note this lack of investment in infrastructure as it is emblematic of the more general lack of public investment in infrastructure in recent years that NIESR has long been stressing as part of the reason behind poor UK productivity growth.

The report also noted that the forecast process could better help policy makers to make their decisions if, rather than concentrating on the central forecast, the forecasters spent more time on scenario analysis. Part of this improvement also requires a rethink in terms of the use of personnel, with a need to increase experience levels among those working in ‘key substantive areas’.

What did the Review recommend?

The Review made 12 specific recommendations grouped around the issues of the infrastructure needed for forecasting and policy analysis, the link between the forecast process and monetary policy decisions, and the use of the forecast for communicating monetary policy.

Let me take these in turn.

In terms of the infrastructure, the report made four recommendations:

• The Bank should prioritise its ongoing updating and modernisation of software to manage and manipulate data
• The Bank should also prioritise model maintenance and development
• Over the longer term, the Bank should look to replace or, at a minimum, thoroughly revamp its main forecasting model, COMPASS
• The new forecasting framework should, at the very least, include a realistic representation of the monetary transmission mechanism, short and long-run inflation expectations, a wage-price system with gradual adjustment to shocks and wages affecting prices and vice versa, detailed models of the housing, energy and financial sectors, and greater attention to the supply side

There’s not much to argue with here but one question these recommendations throw up is what the new forecasting framework should look like? I would argue very strongly that a DSGE model should be at the centre of this, given that such models enable you to construct a consistent story around your forecast while also enabling you to consider the impact of possible shocks through scenario analysis, a key feature of later recommendations.

In terms of the link between the forecast process and monetary policy decisions, the report made three recommendations:

• More attention should be paid to forecast errors and what can be learnt from them so that policy can quickly and flexibly adjust to shocks as they affect the economy
• The Bank should consider whether it can deploy its staff in ways that improve the forecasting infrastructure and forecast quality
• The central forecast should be regularly augmented by alternative scenarios decided upon at an early stage of each forecast round by the MPC and staff

It could be argued that the failure to understand the reasons for, and respond to, the forecast errors that were building up in 2021 led to a delay in the MPC raising interest rates and, so, higher inflation than might otherwise have been the case. Hopefully, the first of these recommendations should ensure that this does not happen in the future. The need to increase experience levels – and arguably economic expertise – among the forecasters at the Bank is one that has been an issue for a long time and was also commented upon by Francesca Monti within a recent article in Central Banking. And, as I said in my recent article in Central Banking, it is clear to me that MPC communication could be greatly improved through scenario analysis as this would provide insight into how the MPC are likely to respond to different shocks:  the so-called ‘monetary policy reaction function’.

Finally, on the use of the forecast for communicating monetary policy, the report made five recommendations:

• The Bank should publish selected alternative scenarios in the Monetary Policy Report, along with the central forecast
• The MPC should de-emphasise the central forecast based and be particularly clear to let people know where the forecast conditioning assumptions are inconsistent with its view of the outlook
• The MPC should replace or cut back the detailed quantitative discussion of economic conditions in the Monetary Policy Summary in favour of a shorter and more qualitative description, following the practice of most peer central banks
• The fan charts should be eliminated
• The Bank should implement the changes proposed in the report in phases, first improving the forecasting infrastructure, while moving cautiously in adopting changes to policymaking and communications

In my Central Banking article, I argued that the MPC should go further than deemphasising the central forecast by handing it over to Bank staff. This would enable MPC members to concentrate on the scenario analysis in their communication of what they see as driving monetary policy decisions and what might cause them to change policy. I argued that the MPC could go further by publishing their view (or views a la the ‘dot plots’) on the likely path for interest rates as a means of communicating the monetary policy reaction function, but Bernanke stopped short of recommending this, given ‘that change would be highly consequential’, preferring to leave ‘decisions on this issue to future deliberations’. It seems that this issue is just too contentious for now! On eliminating the fan charts, I will be sad to see them go (having had a hand in their initial development). But it’s been clear that the fan chart has never really permeated into the general public’s conscious and maybe scenario analysis will help make it easier to explain the uncertainty around, and risks that could affect, the forecast.

How has the Bank of England responded to the Review?

Perhaps unsurprisingly given the political pressure to commission the report in the first place, the Bank welcomed the report and has committed to acting on all 12 of the Review’s recommendations. But we can expect the response to take time. After all, there are many recommendations, and they cover a broad range of issues. In their response, the Bank put forward a number of workstreams that will move forward in a ‘phased’ way, as per Recommendation 12 of the report. The Bank has promised to provide an update for an external audience on their progress on these workstreams by the end of the year. I shall very much look forward to seeing this!