how to improve official predictions – by asking the public

CPI Inflation Forecasts by the OBR for March 2022 from November 2017 - March 2022. Author's chart using OBR data^[5]

The first five forecasts from 2017 to March 2020 predicted a 2% inflation rate for March 2022, which is the Bank of England’s target rate^[6]. After that, the forecasts for November 2020 and March 2021 fell below this, but then started to climb in October 2021.

The final forecast, in March 2022, predicted a 7.4% inflation rate. This was below the actual March 2022 inflation rate of 8.5% – the final bar in the chart above. This reflects the fact that statistics are revised after they are initially published as new data is gathered. This means it takes time for the official rate of inflation to be finalised.

Read more: Is the UK in a recession? How central banks decide and why it's so hard to call it^[7]

The Bank of England uses data such as this forecast^[8] to set UK interest rates in order to keep inflation in check. It has been criticised^[9] for failing to raise rates fast enough to deal with recent inflation.

But the reason for this is apparent in the chart. The first seven forecasts were the same or lower than the Bank’s preferred rate of inflation, suggesting that the OBR’s model was of little use in helping the bank to predict inflation up to five years in advance. It was actually much higher. Even when the target month arrived, the prediction was still not accurate.

The policy implications

Given that interest rate changes affect inflation after long and variable lags, this means that the members of the Bank of England’s monetary policy committee (MPC^[10]) – whose job it is to set interest rates – are flying blind until the date of the forecast is quite close. This adds to the difficulty for the committee of working out what will happen to the economy if it raises rates^[11], something it has had to do a lot recently.

Much of this problem stems from the rather complex equation in the OBR model of the economy^[12] used to make inflation forecasts, which is heavily reliant on measuring inflation in the past. This means that the forecast is more of a projection, with the past dominating the picture. In reality, policy makers need to know what is going to happen in the future.

To be clear, the economists who draw up the forecasts are not trying to manipulate the figures. It is simply that their models are not up to the task of accurate forecasting beyond the very short term. The results are particularly skewed by unexpected events like the COVID-19 pandemic^[13] or the Russian war in Ukraine^[14].

The OBR acknowledges this problem in its discussion^[15] of the forecasting models:

… the model is a tool that we use, while the forecast is the result of the judgements and assumptions that we feed into it. Other forecasters using the same model will almost inevitably generate different forecasts.

This basically means the experts do their best to try to make up for the deficiencies in their models with ad hoc judgements.

An alternative approach

Is there any way of improving these forecasts in the future? It might help to take into account what has been described as the “wisdom of the crowd^[16]”.

This is the idea that large groups of people may do a better job of guessing the results of an uncertain event than a small group, even when the latter are experts in a particular field. People rely on this theory when they use forums like Quora or Reddit to crowd source answers to all kinds of questions and dilemmas – from parenting^[17] to partners^[18] and punctuation^[19].

Richard Curtin^[20], a behavioural economist from the Survey Research Center at the University of Michigan, has measured consumer expectations of inflation and other economic variables in the US since the 1970s. His approach is based on surveys of representative samples of the American public. He has looked closely at the relationship between these expectations and actual outcomes and found that the American public are pretty good at forecasting US inflation^[21].

This is true for respondents with different levels of education, according to Curtin^[22], even when the highly educated or those with more expertise (such as MPC members) might be expected to do better.

These findings suggest a way forward. Instead of relying on equations dominated by past observations, the OBR could also conduct regular surveys of UK consumers to find out what they think about inflation in the future.

Policy makers already monitor household reactions to rate changes and inflation news, believing inflation forecasts can be something of a self-fulfilling prophecy^[23] that influences consumer behaviour. But feeding consumer expectations into inflation forecasts could create stronger predictions for consumer price rises.

Read more https://theconversation.com/inflation-how-to-improve-official-predictions-by-asking-the-public-194666

• Prev
• Next