By David Milliken
LONDON (Reuters) - When the Bank of England's chief economist was asked to explain why its forecasting models had failed to anticipate runaway inflation, he sought to manage expectations.
"All economic models are wrong, but some are useful," Huw Pill concluded in a letter to lawmakers last June that laid out the limitations of prediction methods.
Britain's central bank was nonetheless unable to escape the censure of economic experts in parliament who judged its "inadequate" projection models and a narrow outlook had frustrated its efforts to rein in rampant inflation in the wake of the COVID pandemic and Russian invasion of Ukraine.
The report in November turned a public microscope onto the arcane world of economic forecasting, a combination of science, art and guesswork that aims to divine the future state of the economy and guide central bankers in adjusting interest rates.
"We should really be thinking of economic forecasts in terms of probability distributions," said Stephen Millard, deputy director of Britain's National Institute of Economic and Social Research, who spent more than 26 years at the BoE.
"They are like weather forecasts - along the lines of 'there is a 20% chance of rain'."
The BoE has called upon a Nobel Prize winner, former Federal Reserve Chair Ben Bernanke, to review its methods. His report is expected in April and heralds what Pill said this month was a "once in a lifetime" chance to shake-up the central bank's forecasting and communication methods.
While Bernanke himself declined to be drawn about his review, Reuters interviewed eight leading economists, including current and former members of the BoE's interest rate-setting Monetary Policy Committee, who identified some of the main weaknesses of the Bank's approach and changes they envisioned.
Michael Saunders, who stepped down from the MPC in 2022, described a sometimes dysfunctional internal process where rate-setters disagreed with their central bank's own projections for key indicators like inflation and growth.
"The problem that needs solving is that the Bank publishes a forecast which many MPC members - often the bulk of MPC members - don't think is a realistic description of what the economy's likely to do," he said.
One radical option to resolve this would be a shift from the BoE producing single forecasts to a system where each of the nine MPC members anonymously gives their own projections, which are then collated into charts called "dot plots". Bernanke introduced this system at the Fed over a decade ago.
A more broadly supported reform among the economists interviewed was a move to publish a series of alternative scenarios alongside the main forecast.
Saunders said that if he were still on the MPC, he would want to consider scenarios around global shipping costs staying high for six months versus two years, and also for wage growth failing to slow as forecast.
There is also a communications element.
Current MPC member Jonathan Haskel backed a wider use of alternative scenarios, telling Reuters they could help people outside the bank understand how the BoE's modelling worked and the "reasonable parameters" for uncertainty.
Many of the economists stressed that the BoE's forecasting was on a par with other major central banks, including the Fed and European Central Bank.
The banks have all faced similar criticism for failing to anticipate that the end of COVID lockdowns followed by the Ukraine war would presage runaway inflation, and that they were too slow to raise rates.
HIT BY HISTORIC INFLATION
The BoE and its peers were not expected to predict the pandemic or war. They nonetheless faced criticism from politicians and investors that they failed to foresee the scale of the surge in inflation in 2022, or how slowly it would fall.
British inflation peaked at a 41-year high of 11.1% in October 2022, after Russia's invasion of Ukraine in February that year caused European natural gas prices to leap.
Yet inflation was already 6.2% in February 2022, triple the BoE's forecast just a year earlier, as central banks underestimated the scale of supply-chain difficulties and labour shortages after the pandemic.
Inflation was also slower to fall in Britain than in other countries, hovering around 10% in the second quarter of 2023 compared with a BoE forecast in May 2022 it would fall below 7%.
One major challenge that has faced the BoE and other central banks is that it has been decades since inflation last leapt as high as it did in 2022, and most economic models are not based on historic data going that far back.
While models can be recalculated to include this data, many aspects of Britain's economy have changed since the 1980s - such as union membership, energy sources and trading partners - making it hard to draw valid comparisons.
Furthermore, while the BoE's models did face a tough test during the financial crisis, for much of the past 30 years they have been forecasting in relatively calm conditions, only to come up against Brexit, COVID and Ukraine in quick succession.
Philip Shaw, chief UK economist at Investec and a long-time BoE watcher, said he had "absolutely no relative criticism at all" of its forecasting performance compared with other central banks and forecasters.
"The bank comes under fire because inflation was higher than the UK's peers, but the UK has been more dependent on European gas markets," he added. "I wouldn't blame that on the Bank of England."
The ECB has blamed most of its forecasting errors since the pandemic on rocketing energy and food prices, with President Christine Lagarde saying in September the bank needed to do a better job of conveying the uncertainties around projections.
Fed Chair Jerome Powell said last month the delay in raising U.S. rates was largely because policymakers thought inflation was caused by post-pandemic bottlenecks that would resolve themselves.
CONNECTING THE DOTS
In essence, forecasting models either extrapolate recent trends into the near future, or flesh out relationships in economic theory - between unemployment and wage growth, for example - based on historic data.
Projections are underpinned by a series of "conditioning assumptions", such as global energy prices and currency exchange rates, which if altered can throw out different results.
Important assumptions used by the BoE that some experts believe may be up for review include assumptions that interest rates will move as forecast by financial markets - which may be different to what MPC members expect - and that there will be no change to government policy on taxation and spending.
In its November report, the economic committee of Britain's upper house of parliament criticised BoE culture for a lack of diverse views to challenge prevailing orthodoxy. Many of the bank's models, it added, led to mistaken assumptions about the "transitory" nature of high inflation in 2020 and 2021.
In response, BoE Governor Andrew Bailey said MPC discussions were "open, frank and forensic" and that Bernanke's review would recommend improvements.
At present, the MPC issues a collective forecast of inflation and other indicators, though this can mask internal differences of opinion. One possibility the Bernanke review is likely to consider, according to the economists interviewed, is that the BoE adopts something similar to the Fed's dot plots.
Since 2012, as part of a drive by then-Fed chief Bernanke for more transparency, rate-setters at the U.S. central bank have each given forecasts for rates, growth and inflation, based on their own preferred assumptions.
These forecasts for the next one to three years are anonymous and plotted as a group on a chart, with each forecast represented by a dot. This allows markets to see the consensus and range of preferences among Fed policymakers for interest rates, growth and inflation.
"I have to say I think it's worked quite well," said Saunders, the former MPC member.
"The advantage of the dot plots is it gives the individual members a chance to express a view on where they're thinking interest rates will go over time."
Megan Greene, a serving MPC member, said she "quite liked" dot plots for the same reason, but worried markets misread them as a commitment on rates.
Other experts are opposed to such a change.
A previous review of BoE forecasting during the 2008-09 financial crisis, published in 2012, recommended individual MPC forecasts, only for the idea to be shot down by the bank which said it did not want to be "highlighting the relatively small differences between individual members".
James Smith, research director at the Resolution Foundation think-tank and formerly a long-time BoE staffer, said the MPC's collective projections forced policymakers to engage with each others' views more than at other central banks.
"You would potentially lose something quite important," said Smith about a shift to individual forecasts, which he believed would not have helped during the recent inflation shock.
MODELLING MANY FUTURES
The approach of including several alternative scenarios also draws criticism.
The BoE's projections already have one alternate scenario built in, with forecasts for growth, inflation and unemployment if the central bank were to keep rates unchanged, rather than to adjust them in line with market expectations.
Deputy Governor Sarah Breeden said in a speech last month that she found it very helpful to consider two alternate economic scenarios - one of unexpectedly weak demand, another of unusually persistent inflation - when considering policy.
However alternate scenarios are time-consuming to produce, and Investec's Shaw said multiple possibilities could muddy the BoE's message: "There's a clear trade-off between a simple message and discussing the various risks."
Who within the central bank produces the official forecasts can also play an important role.
The Fed and the ECB's forecasts are produced by the general staff at those institutions, rather than by the rate-setters themselves, so play a less critical role in its communication.
As the British bank's forecasts are produced by the MPC, investors can view them as a policy signal. If the committee predicts inflation in two to three years' time will be far from its 2% target, this implies it does not believe the current market rate expectations are appropriate.
The BoE is attempting a juggling act, said Millard of the National Institute of Economic and Social Research, who would like Bernanke to recommend taking the task of producing forecasts out of the hands of rate-setters altogether.
"The forecast is trying to do three things at the same time. It's producing a forecast, it's helping the MPC to set policy, and it's helping the MPC to communicate to the market," he said.