These cyclically adjusted measures are useful information and should not be discarded, but I don’t wish to use them as the sole or main or dominant source of information about the stance of fiscal policy. Here are a few reasons why:
1. The cyclically adjusted measure relies on a measure of potential output. That is not a big problem when potential output is clear, but in a lot of today’s cases the very debate is over the size of the output gap.
Let’s take the current UK. If the output problems all resulted from supply and productivity failures (not the case, just a hypothetical), the numbers would indicate that the current fiscal policy stance is quite expansionary with high budget deficits.
Alternatively, if you take a more Keynesian view, the potential output gap is much larger and, cyclically adjusted, the stance of fiscal policy looks much more contractionary by the cyclically adjusted measure.
Very often the key question is something like the following: how much are the current UK problems demand-side and how much supply-side? I see economists addressing this question by invoking cyclically-adjusted measures of the stance of fiscal policy.
This is has a significant element of circularity. If you assumed a big output gap to begin with, the demand-side crunch is measured as very large. If you assumed a very small output gap to begin with, the measure gives you a very small demand-side crunch in this case.
Maybe you have seven other reasons for believing it was a large demand-side crunch, but that doesn’t make this a good measure. Your results depend very much on the assumption — about potential output and thus demand — which you put into it.
Furthermore this point is far from transparent yet I don’t exactly see proponents of the cyclically adjusted measure tripping over themselves to remind us of it.
(As an aside, the IMF, before this issue became so politicized, expressed precisely this reservation about cyclically adjusted measures and in fact did not appear to favor them as a general approach, while admitting they do provide some interesting information.)
2. There is a big difference between two cases: “the government drove an AD collapse by massive net spending cuts” and “big problems happened and government didn’t fill the gap nearly enough.” These cases have different economic and political causes, may involve different remedies, and very likely will yield different multipliers, both for initial effects and when it comes to potential remedies. I want my metrics for the stance of fiscal policy, and my commentators on fiscal policy, to disaggregate them, rather than to blur them together. Cyclically adjusted measures blur them together. There simply is not one big multiplier for “austerity” as a general concept.
3.The cyclically adjusted measures are an abstraction, and furthermore an abstraction based on estimating a modal quantity, namely potential output.
To give an analogy, I get uneasy when I read sentences such as “inequality caused X.” “Inequality” didn’t cause anything. Inequality is a statistical residue of some other actual processes. It is better to say what caused X (say “the rage and poverty of inner city residents”) and, if relevant, connect this to inequality as well. Except that the cyclically adjusted deficit is an even more problematic causal concept than “inequality” because it relies on measurement of a modal, namely potential output.
The word “austerity” is a political concept and it does not belong in rigorous economics. Let’s just say what happened. Note the simple causal language in my point #2.
4. The mainstream literature on fiscal policy, starting at least with Blinder and Solow (1973), slots “G” and “Gdot” into the model as the measure of fiscal policy. Paul Krugman pieces do this too, as do hundreds of other economists. Models are there to give us some discipline, so let’s use that discipline. Why is it that so few of these classic models used the cyclically adjusted deficit measure to express the stance of fiscal policy? Because many of those economists know that, in the causal sense, that is a whacky measure and best used with caution. It simply is not “the standard measure of the stance of fiscal policy.”
5. These days there is a sudden near-moratorium on citing open economy models of fiscal policy in smallish open economies with floating exchange rates (it may not matter nearly as much as you think for AD). This should be incorporated into the discussion more, yet oddly it suffers from relative neglect. Note also that a) the UK hasn’t the whole time been at the zero bound, and b) you still can depreciate your currency at the zero bound.
6. Let me sum up where we (at least many of us) agree. British policy should have been much more countercyclical than it was and indeed I have thought this from the beginning. But I still see many commentators, writing on British fiscal policy, giving us an exaggerated sense of its potency, an exaggerated sense of the causes of the British downturn in the AD/fiscal policy direction, and cloaking all of this under non-transparent terminology, while keeping the various important qualifications rather silent.
I have further points to make about UK austerity in particular, and the Wren-Lewis blog post, but I’ll save those up for another day. On some details of the UK economy, Matt’s simple treatment makes a lot of sense and he is also (correctly) a major proponent of the relevance of AD analysis.