Working paper No. 91
By Nicoletta Batini and Andrew G Haldane
This paper evaluates a class of simple policy rules that feed back from expected values of future inflation — inflation forecast-based rules. These rules are simple, and so are analogous to Taylor rule specifications. Because they are forecast-based, the rules are meant to mimic, albeit imperfectly, monetary policy behaviour among inflation-targeting central banks in practice.
In the paper, inflation forecast-based rules are assessed by evaluating how well they perform when the economy — a small rational expectations macro-model with sticky inflation and forward-looking agents — is buffeted by a combination of shocks, whose distribution is drawn from the Bank of England’s forecasting model.
The paper shows that inflation forecast-based rules confer some real benefits: they embody explicitly monetary transmission lags (lag-encompassing); they potentially embody all information useful for predicting future inflation (information-encompassing); and, suitably designed, they can achieve a high degree of output smoothing (output-smoothing). In fact, these rules prove more efficient at minimising inflation and output variability than standard Taylor rule specifications, and almost as efficient as fully optimal rules. These results seem robust across different model specifications.