Mind the gap! - New Zealand's experience

Earlier I wrote up a piece from the Bruegel think tank's blog about the fallibilities in the European Union's way of measuring the output gap, and the problems it causes in trying to estimate how much of a country's fiscal deficit might be structural or cyclical.

Bluntly, big problems with estimating the 'normal' state of the Eurozone economies, and what the 'structural' or 'underlying' fiscal balances look like in that 'normal' state, would make you very wary indeed of basing fiscal policy decisions on them.  And I wondered whether these issues tend to crop up elsewhere, and whether this exercise is a sensible goer anywhere.

As it happens, one of Treasury's officials had a close look at these issues in a New Zealand context. It was one of the papers presented at Treasury's 2011 conference New Zealand's Macroeconomic Imbalances – Causes and Remedies Policy Forum. Anne-Marie Brook's paper, Making Fiscal Policy More Stabilising in the Next Upturn: Challenges and Policy Options, was on the general topic of fiscal policy as a tool of macroeconomic stabilisation, and included both a literature review and empirical analysis of how fiscal policy has actually played out in New Zealand.

Here are two graphs that I thought especially interesting (from p14 and p18 of her paper).

The one above charts the fiscal impulse on the vertical axis. The fiscal impulse is the year to year change in the structural (cyclically adjusted) fiscal balance, i.e. how much of the change in the fiscal position is down to fiscal policy decisions as opposed to cyclical (or unusual one-off) factors. It is therefore a measure of whether fiscal policy is more expansionary or contractionary.

The horizontal axis charts the state of the economy - a negative number for the output gap means the economy is running below full potential, relatively weak in other words, and a positive number means it's running hotter.

You can see the logic of the four quadrants - for macroeconomic stabilisation purposes, you want to see the data points turning up in the countercyclical upper right and lower left quadrants, and not in the procyclical other two. On these numbers, in practice you see clear patterns: no instances of tightening in bad times (excellent), quite a few of tightening or loosening when you should have (jolly good), and a bunch of procyclical easings (not good at all), what Anne-Marie summarised (p14) as "a tendency towards asymmetric Keynesianism, in the sense that procyclicality is successfully avoided during downturns, but not so consistently during good times (too many outturns in the bottom right quadrant)".

At face value, this looks moderately encouraging for folks who might be inclined to estimate potential output, the output gap, structural fiscal balances, and the fiscal impulse, and use them for cyclical stabilisation purposes.

Except that Anne-Marie also provided one of the best graphs I've ever seen, which showed the difficulties in trying to do this exercise in real time. Here it is.

The graph shows what Treasury thought the output gap and fiscal impulse were a year before the Budget (green dots), around Budget time (red dots), and, crucially, what the situation actually was, as measured later with the benefit of hindsight. And it transpires that three of those procyclical fiscal stimulations were actually completely unintended: the economy was actually in better nick than Treasury realised at the time.

Not that Treasury ought to be hauled over the coals for it. "The magnitude of such forecast errors is not Treasury specific or New Zealand specific. It is well known that empirical estimates of the output gap are subject to significant and highly persistent revisions for all economies", Anne-Marie concluded (p17), and she recommended among other things that Treasury should therefore "expand the repertoire of indicators so that advice on the fiscal stance is less reliant on any single measure, with particular care taken to augment fiscal impulse measures with complementary measures" (p26).

Our experience would tend to confirm the international experience: estimating output gaps and structural fiscal deficits is iffy at the best of times (though the EU estimates take iffyness to a whole new level), and you wouldn't want to base cyclical fiscal policy solely or heavily on them.
All fair enough. But even if you conclude that the state of the art in cyclically adjusted fiscal deficit analysis isn't up to much heavy real-time or short-term stabilisation usage (or possibly not up to any real-time usage at all), the concepts are good ones. They still tell us important things about the profligacy or otherwise of the government's books over the longer haul.

Here's Ireland's recent story (data taken from the latest IMF World Economic Outlook database, accessible here if you ever feel like playing with it yourself).

We know, from the Bruegel piece and from other commentators, that the precise numbers may not be right. But they're not so bad that they can't tell us the broad high-level picture. Through the boom years (to 2007 or so) Ireland looked as if it was running a responsible fiscal ship, if you went by the headline numbers that got reported at Budget time. But cyclically adjusted, the government was running a decent sized structural deficit, one that would be exposed if the cyclical revenues dried up. As they did. More recently the graph is again telling us the fundamental truth: Ireland's run a massive fiscal restructuring exercise, which shows up in the hugely improved structural position, even if it's yet to be seen in the cyclically depressed headline balance.

So I still hold out some hope for intelligent use of output gaps and structural/cyclical splits, especially if there are more sanity checks (for example, from business opinion surveys of capacity utilisation rates) around the plausibility of the numbers.