BoE economists criticise DSGE inflation modelling

Economists at the Bank of England have taken a fresh look at the factors driving inflation in the UK, finding some major economics models fail to capture the dynamics at play.

Monetary policy committee member Kristin Forbes and economists Lewis Kirkham and Konstantinos Theodoridis adopt a method that breaks time-series data into a slow-moving trend and deviations around this trend in a recent paper, A trendy approach to UK inflation dynamics.

They cast a sceptical eye over dynamic stochastic general equilibrium (DSGE) models, which require “a number of assumptions” and generally treat trend inflation as constant, saying their own approach generates “insights starkly different than those attained using the more standard DSGE framework”.

The paper’s publication on June 16 was well timed, coming three days after the UK’s latest inflation figure was published. At 2.9%, inflation is running above target and now exceeds the BoE’s central forecasts for the year. One major input into the BoE’s analysis of inflation is a DSGE model.

Forbes and her co-authors warn that economists ignore trend inflation at their peril. Their results imply some 95% of trend inflation is explained by one principal component and that trend inflation plays a “powerful role” in determining actual inflation outcomes – explaining about 39% of headline inflation and 46% of core inflation.

The results also imply that after controlling for the trend, “most macroeconomic variables used in the inflation literature add little or no explanatory power to regressions predicting inflation”. Measures of “slack”, commonly assumed to have an impact on inflation, fail to show this effect in the data; international prices seem far more important.

Domestic factors commonly associated with inflation do still seem to have some impact on the trend part of inflation, but the exchange rate appears “even more important”. A depreciation in sterling of 10% is correlated with an increase in trend inflation of about 0.4 percentage points. “This suggests movements in sterling can have significant, large and persistent effects on headline and core inflation,” the authors say.

The authors urge central bankers to keep these results in mind when modelling inflation: “Central banks should appreciate the limits of structural models of inflation dynamics that are centred on relationships of prices with inflation expectations and/or slack.”

Global impact

The results have particularly strong implications for the way the Bank of England conducts its monetary policy. If the authors are correct, much of the volatility of inflation can be explained by global factors, so even stability in the UK economy may not be enough to guarantee on-target inflation.

Furthermore, the results imply the BoE should be less willing to “look through” movements in sterling as it has done in the past.

“Given the magnitude of the impact of sterling on UK headline and core inflation, and the persistence of these effects, monetary policy cannot automatically dismiss the effects of sterling on inflation as temporary and may need to respond,” the authors say.