,
10 tweets,
3 min read
Read on Twitter
NEW PAPER THREAD: Our paper on fiscal multipliers has just been released by @PIIE. We create a dataset of large fiscal expansions for 17 #OECD countries between 1960 and 2007 piie.com/publications/w… 1/8 
2/10 The database is constructed by first finding when a country's cyclically adjusted deficit increased by a lot and then by reading OECD and IMF reports published in those years to check whether a fiscal stimulus did in fact happen at that time and why
3/10 Our first result is that it's crucial to complement the statistical approach to identifying large shift in fiscal policy with narrative evidence. Relying only on large shifts in the cyclically adj deficit to identify episodes of large fiscal expansions is a REALLY bad idea
4/10 In fact, the downward bias created by reverse causality is so large that one would erroneously conclude that fiscal expansions are contractionary. Reverse causality arises for 2 reasons: a) b/c of output declines whose impact on public finances is not fully captured
5/10 by the cyclical adjustment procedure. Think of Ireland during Great Recession which saw increase in its cyclically adjusted deficit despite doing austerity. And it arises b) b/c of countercylical fiscal policy - fiscal stimulus designed and implemented to fight a downturn
6/8 These identification issues, which were shown to be central behind the expansionary austerity claims, are ubiquitous for stimulus. Let's keep that mind when designing fiscal stimulus for the next downturn
7/10 This being said we do find a number of fiscal expansions that were not motivated by countercyclical considerations - e.g., tax cuts to increase efficiency/potential output, infrastructure spending to redress past underinvestment ... - and can thus be used for estimation
8/10 Using these exogenous fiscal expansions, we find that the fiscal multiplier for stimulus does not appear smaller than that for austerity. If anything, our estimates are consistent with those found in the more recent austerity literature that also rely on narrative evidence
9/10 Our results thus challenge the recent view that the effect of fiscal policy might be so asymmetric that we cannot learn from the recent experience with fiscal austerity. In this asymmetric view, yes austerity was bad but stimulus would not be useful
10/10 Another result is about state-dependence or whether fiscal policy is more effective in bad times than in good times. There is no consensus on this issue so far in the literature. We find that the effects of fiscal expansions are largest in slumps and smallest in booms
