Simon Wren-Lewis on the British stagnation
Nick pointed me to a Simon Wren-Lewis post that does a great job of illustrating the British predicament in three graphs.
The first shows that British RGDP is lower than the levels of 5 years ago. The second shows that British labor productivity has fallen about 15% below trend since 2008. And the final graph shows a rise in unemployment, which is currently 7.9%. Wren-Lewis speculates that the elevated unemployment rate reflects a demand shortfall. I tend to agree, with reservations.
Unfortunately we don’t know the natural rate of unemployment in Britain, or even whether the natural rate is stable. If you take the longer view, it’s pretty obvious that the natural rate is not stable over time.
We also know that other European countries (notably France) have seen huge increases in their natural rate of unemployment since 1980. And we know that the size of the UK government expanded sharply after 2000, from about 37% of GDP to the high 40s. Note that the expansion was larger than the data suggests during the good years (2000-07) and smaller than the data suggests during the bad years (2008-12.) That’s because government spending should normally fall as a share of GDP during expansions, and rise during recessions. Thus the rise during 2000-07 is far more significant than many assume. This might have boosted the natural rate of unemployment, at least slightly. Note that unemployment rose 0.8% between 2005 and late 2006, a pattern that is highly unusual for an expansion, even given the slow RGDP growth.
Here I’d like to make the assumption that is most favorable to the Keynesian view; a stable natural rate that is lower than the actual unemployment rate during most years. Something in the 5.4% to 5.9% range. Anything much lower would imply that unemployment is consistently above the natural rate, which would of course be inconsistent with the natural rate model. Then the current unemployment rate is probably 2.0% to 2.5% above the natural rate, which suggests that high unemployment has depressed output by 4% to 5%.
To summarize, if we make the most Keynesian assumption that is plausible, a low natural rate of unemployment which has not been rising, and if we assume that 100% of unemployment is due to a demand shortfall, and if we assume that unemployed workers are just as productive as employed workers (no ZMPers), then it still appears that the recent British stagnation is 75% to 80% supply-side and 20% to 25% demand-side.
In other words, when Keynesians blame the slow British RGDP growth on the mythical “austerity” they are talking nonsense. This is partly because the unemployment numbers suggest that the British economy does not have a particularly large output gap, when compared to other developed economies, which means their massive budget deficit (trailing only Egypt and Japan) really does indicate a fairly expansionary fiscal stance. And second, because even if fiscal policy is quite austere, the slow RGDP growth is mostly due to supply-side factors such as declining North Sea oil output, less froth in high finance, and perhaps other factors such as labor shifting out of home-building. (I’m open to suggestions.)
Let me end on a positive note, by pointing to some areas where I agree with Wren-Lewis:
1. I agree with his claim that slow nominal wage growth suggests the elevated unemployment is mostly due to demand-side factors.
2. I agree that demand stimulus is appropriate, although unlike Wren-Lewis I’d rely 100% on the BoE. Ed Balls made things much harder for the BoE when he opposed changing their 2% inflation target. His comments were disgraceful, as one reason the government is reluctant to do more monetary stimulus is fear of being hammered by Labour on the “inflation” issue.
3. I agree with Wren-Lewis’s claim that NGDP targeting is not always optimal. I’ve argued (for instance) that the monetary authority could do a bit better by targeting NGDP net of indirect business taxes. Because these taxes rarely change significantly in the US, (which lacks a VAT) I tend to gloss over this problem. And there are a few other potential problems with NGDP targeting, especially in smaller and less diverse economies. But NGDPLT is a vast improvement over IT.
HT: Bill Woolsey