Response from Holland and Portes
This post was written by John McHale
Dawn and Jonathan have been kind enough to reply to the post below. Many thanks to regular commenter David O’Donnell for bringing the post to their attention.
We are grateful to John McHale (and others on this site ) for these very thoughtful comments.
First, we should emphasise that NIESR does not have the detailed specific expertise on the Irish economy of many of those commenting here: given that NiGEM is a global macroeconometric model, the Irish economy component is inevitably quite stylised, and can’t take account of some of the specific features of the Irish economy which John and other commenters have rightly highlighted. The fact that our calculations suggest that fiscal consolidation is less damaging in Ireland than in other countries reflects relatively low estimated multipliers, even in current circumstances. But the NiGEM multipliers may well overstate the extent of import leakages, given the specific structure of Irish trade, so actual multipliers might be higher than our model estimates suggest. We should also note that the fact that, in contrast to other countries, our estimates suggest fiscal consolidation has in fact reduced debt-GDP ratios in Ireland does not any measure imply that we think it was the optimal policy (see below).
There are two important general points John makes.
First, he points out that while our simulations show that debt-GDP ratios will be higher in 2013 (in all countries except Ireland!) than they would have been without fiscal consolidation, primary deficits will be lower, and in the long run the binding solvency constraint is the intertemporal government budget constraint. This is absolutely correct. Postponing fiscal consolidation doesn’t mean that it isn’t ultimately necessary in EU countries – almost all of them – that have significant structural primary deficits. The main point of the paper is that the negative impact on GDP, and hence on the amount of consolidation required, is much larger if you frontload consolidation during a period when multipliers are much larger. Later consolidation could have been both smaller and less damaging. We have not attempted to illustrate such an alternative path (and of course the “optimal” path depends on future economic developments) but certainly delaying would have been better. This is also true for Ireland.
Second, he notes that we omit any analysis or discussion of hysteresis. We agree entirely that this is a very important issue when considering the impact and timing of consolidation. In an earlier paper, Bagaria et al (see Vox here) we perform a similar, but more detailed, analysis for the UK, incorporating labour market hysteresis effects (but ignoring spillovers). In this (in contrast to Delong and Summers), we do exactly what John suggests, which is to assume that hysteresis effects matter but decay over time. Ideally we would indeed do the same for Ireland and other EU economies, but this is a somewhat more complex exercise.