Everything that needs to be said has already been said.
But since no one was listening, everything must be said again.

Andre Gide

Friday, May 8, 2015

Austerity in the UK

I just finished reading this piece by Brad DeLong: Optimal Control, Fiscal Austerity, and Monetary Policy. He begins in the following way:
I find myself perseverating over the awful macroeconomic policy record of the Conservative-Liberal Democrat government of the past five years in Britain, and the unconvincing excuses of those who claim that the austerity policies it implemented were not a disaster–and that the austerity policies it ran on would not have come close to or actually broken the back of the economy.
Then he presents us with the following picture:

The red line is supposed to show us how real GDP would have behaved in the UK if it wasn't for austerity. It must feel good to be so self-assured in one's diagnosis of an ailment. But Dr. Brad, forgive me, I do have some questions to ask of you.

First, I find it interesting that when you want to focus on the "depression" in the U.S., you point to the labor market performance there, not the GDP performance. You present pictures like this (source):

Which is fair enough, I think (I offer my own thoughts on your post here). But then, when you want to focus on the "depression" in the U.K., in and odd about-face, you ignore their labor market performance and focus solely on GDP. Why might this be? I think we should also take a look at the labor market data. If the effects of austerity are going to show up anywhere, I would think we should find them there (this is what our textbook models tell us, at least).

Here's the unemployment rate.  

Hmm...not what I would have expected, to be honest. Here's the employment to population ratio:

Yeah, I'm still not seeing it. How about participation rates?

Nope. Sorry. What if we focus on employment-to-population for prime-age workers?

Gee, the UK is actually looking pretty good against the US here. Remind me again, it was the U.S. that had the stimulus, right? How about the prime-age part rates?


Interesting, no? What about inflation? Here it is:


If anything, the UK had a higher rate of inflation over its period of austerity. Well. OK, let's go back to GDP.

Ah, so there it is. And in terms of real GDP per capita:

Ah  yes, there you have it. Austerity evidently killed GDP, but not the labor market. That's a very interesting hypothesis, but I'm wondering which textbook theory is consistent with it?

One story that makes sense is one in which austerity is somehow influences the productivity (or measured productivity) of workers. Classic Keynesian demand-stimulus measures are not typically thought to work through their effects on labor productivity (though, they could if government spending was in the form of infrastructure, for example.) The usual prediction is that an increase in G increases employment and reduces the average product of labor, where the effect on labor productivity is incidental (expansion of labor + diminishing returns to labor).

Interestingly, a neoclassical model with heterogeneous (high/low skill workers) could be consistent with these observations. Suppose that austerity takes the form of cuts to transfers (which may not have been the case). Then a standard wealth-effect motive induces workers to increase their labor supply (lower their reservation wages when searching for work), to make up for the loss in their after-tax wealth. To the extent that this wealth-effect is stronger for lower-skilled workers, the average quality of labor (measured labor productivity) declines. This is just a classic composition effect. I'm not suggesting that this is what happened. It's just interesting to note that it is theoretically possible from a neoclassical perspective and not so obvious from a Keynesian one.

Anyway, this post was meant more in the way of asking questions. In particular, I am not criticizing nor defending actual UK policy. I just want to know which textbook macro model (the model proposed by Krugman as all we need to know about macro) is consistent with DeLong's austerity story. Someone please do tell.

Postscript May 11, 2015
I asked my RA Michael Varley to provide some UK data on government spending, transfers and tax revenue. The data is plotted as a ratio of the working age population.


  1. Maybe Hagedorn, Mainovskii and Mitman have the answer. Labor markets reforms are optimal in recessions(i actually dont now if UK had any), https://www.dropbox.com/s/p6711tak9ly8r6e/UI_%26_U.pdf?dl=0

  2. DeLong may have simply failed to pinpoint which austerity has been problematic for the UK.

    Even if the UK has done all it could to solve their demand problems, the proximity to a broken Eurozone may hurt their GDP.

    In order to reach its full productivity potential, the UK may need to rely on trade with an economically healthy Europe. Even if stimulus was able to prevent people from going jobless in the UK, it probably can't prevent jobs from leaving what used to be higher productivity euro wide markets, now destroyed by eurozone austerity, and going to lower productivity local economies.

    What the UK government needs to do is pressure the ECB to stop depressing the western world's economies with their inadequate inflation targets and adopt a strategy that would lead to something like the UK has achieved: have near 4% inflation until investment markets resolve their gridlock and people have jobs.

  3. Maybe something like this theory could work? Another version I could see is one where product market matching frictions make it hard to switch between providing public and private sector goods, reducing TFP.

  4. You will know in a matter of months in UK but it may take a little longer here with Obama still in office. Austerity won there and some want more here even though we know the bad effects it had in the UK and US. Governments are not like household budgets and household budgets are like government but if I could print my own money way would it matter if I had to replace my roof after a storm or go to war with my neighbors. No we were not going to be Greece!

    Why do we have a minimum wage in congress? Why do women in congress make the same as man? Shouldn’t a fresh congressperson have a starting wage maybe half of what a 20-year veteran of congress receives? Why do government have pensions and everyone else has SS and 401Ks?

    Why is GE paying no taxes and getting a refund from the taxpayer? Do you think our system is balanced?

    We all know the answers… But some like it as it is, tilted!

  5. UK labor supply has actually increased since 2008. This is due to two things:

    1) immigration
    2) rising participation rate, particularly among older people (55+), single mums and the sick & disabled. The last two are due to cuts in welfare entitlements, while older people staying on in employment (or returning to it) is due to rising pension age, particularly for women, and low interest rates.

    At the other end of the scale, the participation rate of the young is falling, as more of them choose to stay on in education. However, the young employed & unemployed have also suffered cuts in welfare entitlements, while young people in education have higher tuition fees and lower maintenance support funding so are increasingly likely to be working as well as studying.

    So overall, we have had an increase in the absolute size of the working population, an increase in the participation rate, a decrease in the dependency ratio (just about) and a reduction in labor quality.

    In the light of this,the UK's strong employment performance, declining productivity and falling real wages suggest substitution of labor for capital accompanied by hysteresis in the general population.

    So your neoclassical model with heterogeneous workers sounds about right to me.

  6. It is worth looking at productivity in the US and UK in this regard as well.

    See: http://rweconomics.com/htm/USUKG7.htm

    The fact that productivity has risen steadily in the US and G7 countries throughout the crisis and fell in the UK from 2008 through 2010 and has subsequently risen very little suggests that the comparison of labor market statistics between the US and UK are misleading in attempting to understand the effects of austerity.

    The difference in the behavior of productivity between the two countries suggests that there are institutional differences in hiring and firing practices in the UK and US that are unrelated to austerity that account for the employment statistics.

  7. David,
    This graph on the composition of UK emp. growth (BoE) is surprising and does not, unfortunately, support your thesis. Despite the headline, it shows most employment growth during the recovery came from high-skilled labor (until last year).


  8. this is a very good post and it comes to the same conclusions i have independendly made own basicly. Though one thing to add here is that the UK actually has not relied more on austerity than the US in the time between 2010-2015 but less. The deficit in the US is down by 9.3% compared to 5.7% in the UK and the growth differences are simply not large enough to explain this by lack of revenue due to bad growth in the UK. The UK started out with slightly more austerity but in the last 2 years caught up with doing a lot less. The UK current account is also in a big deficit up from 1% to 5.5% in 2015, all this does not point towards weak demand.

  9. I think this deserves a careful look at the data before jumping to theorise. There is no good in models accounting for facts but failing miserably at the mechanisms. First, we need to ensure data is directly comparable and check the sources of measurement error. Second, take a look a subcomponents of public budget and labour market indicators. We are missing how budget consolidation was accomplished, or how are real wages behaving, perhaps they fell faster in Britain.

    1. I agree with you, David. Unfortunately, I don't have the time to do all the fact checking, but I hope others will. Btw, as George Blackford points out, UK labor productivity growth slowed down well before austerity. http://rweconomics.com/htm/USUKG7.htm

  10. More supporting evidence: in the UK at the moment "zero hours contracts" are a bit of an issue. There are (apparently) more and more people in jobs which do not guarantee any number of hours or income level. So it's possible that people are in work, but working fewer hours - though I don't know if this is taken into account in the collation of the statistics.

    1. If the labor data is collected by international convention, zero hour contracts would not be counted as employment. A standard labor force survey asks people "how many hours did you work in the past 4 weeks." If they answer any positive amount, they are labeled "employed."

    2. The recovery in UK average hours worked (per worker) was a little faster than the broader recovery in the labour market. Average hours per worker was roughly back to 2007 level by 2014, but the unemployment rate is only back to "normal" in 2015.

      As Frances says above, part of the supply shock is certainly the increase in older workers continuing to work part-time into retirement, so the recovery in average hours is all the more impressive.

      The most optimistic scenario is that the natural rate of unemployment has fallen, so even with unemployment back to "normal" for 2007, we still have a large amount of slack, which is more consistent with a demand-side explanation. Nobody is really sure, but official estimates of the output gap are currently small (<1%).

      The Bank of England's inflation report has a good summary of the issues here, see page 22 (Chapter 3) "Output and supply". http://www.bankofengland.co.uk/publications/Documents/inflationreport/2015/may.pdf

      As to the question of what model explains this: Keynesians typically rely on the assumption that the level of output is currently purely demand-determined, which is in turn currently determined by fiscal policy. From this assumption follows: shocks to productivity cannot affect the level of output, but they can (somehow) affect employment. e.g. here:


  11. This is the best paper I've seen on the subject; the evidence doesn't support a shift to low-skilled work, but a TFP problem concentrated in certain sectors: