One strand of research has focused on the possibility that multipliers vary at different stages of the economic cycle. Summers and De Long (2012) look at the effect of fiscal policy in the depressed US economy, and assume a multiplier of one in their analysis – although as a small open economy we might expect the multiplier for the UK to be less than for the US.

Auerbach (2012), Parker (2011) and the IMF (2012) have all produced analysis where the size of the multiplier varies depending on whether the economy is expanding or contracting, or on whether there is a positive or negative output gap. These studies tend to conclude that multipliers are larger in recessions than at other times. But the same IMF analysis suggested that for the UK the multipliers were smaller than the OBR assumptions, and only significant for spending.

Barrell et al (2012) published new multiplier estimates for OECD countries using the same National Institute Global Econometric Model it has used in previous studies. The results also suggested generally lower multipliers than those we have used. In contrast, Cloyne (2011) estimates significantly bigger revenue multipliers for the UK, around twice the size of those assumed by the OBR. This work is the first application of a ‘narrative’ approach for UK data, where changes in tax rates (the fiscal impulse) are determined from budget announcements.

Meta-analysis by Gechert and Will (2012), which brings together the results of a host of studies, highlights the impact that the identification of fiscal impulse, model choice and control variables can have on the result. Given many of these uncertainties, the interim OBR’s chosen multipliers continue to be broadly in the middle of a very wide range of views.