Where's the deflation? The supply side of banks

If banks go bust and firms can't get bank loans to finance their operations, that has supply side effects too. A firm that has plenty of customers, but can't get financing to produce enough goods to meet the demand, may raise prices. And if a firm closes down because it can't get financing, its competitors may raise their prices. And if the threat of entry is diminished because new firms can't get financing, incumbent firms may raise prices.

Perhaps we should think of banks, and financial intermediaries more generally, as like producers of an intermediate good. If producers of intermediate goods go bust, that's a real shock. The demand for factors of production like labour will fall. But the supply of final goods will fall too. It creates a bigger wedge between input prices and output prices.

Yes, bank failures will affect aggregate demand, unless monetary policy can and does ensure they don't. And yes, the recent recession should be understood as a fall in aggregate demand. And yes, the fall in aggregate demand may itself have caused bank failures. But that's not what this post is about. This post is about the supply side.

One of the things that has puzzled me about the recent recession has been the failure of inflation to fall more quickly. Reading some of Stephen's and my old posts from late 2008 and early 2009 (put "deflation" into the search box on the top right) reminds me of just how big a concern deflation was back then. Now the Canadian recession was less severe than previous recessions, so the failure of inflation to fall more in Canada is less of a puzzle. But if you had told me in late 2008 that many countries would still be seeing few signs of recovery in 2012, I would have predicted deflation in many countries. And I would have been wrong.

There are a number of possible explanations for the failure of inflation to fall more, despite the recession. Here are those that come to mind:

1. The Phillips Curve is flatter than it used to be in previous decades because a history of successful inflation targeting has made it flatter.

2. The Phillips Curve is flatter than it used to be in previous decades because inflation is closer to 0% than it used to be and there is absolute downward nominal wage rigidity.

3. Oil price rises, indirect tax increases, exchange rate depreciation, and other special factors have caused an upward shift in the short run Phillips Curve. (Note that some of these explanations only work for some countries. In particular, the exchange rate depreciation story cannot by definition work for all countries.)

4. Others I've forgotten.

I'm not saying those explanations are all wrong. There may be some truth to some of them. But I want to set them aside and consider a fifth explanation: the supply-side consequences of bank failures, and financial market problems more generally.

I am reasonably confident that it works theoretically. An exogenous shock that sent a rifle bullet directly to banks and financial markets would cause an adverse shift in the Phillips Curve. Any short run trade-off between inflation and unemployment would get worse. There would be more inflation for any level of unemployment, or more unemployment for any level of inflation. But, like any theory, it's not very useful if the effects aren't big enough to matter. And that's where I'm a lot less confident.

How could we test this theory, to see if the effects might be big enough to matter, and make it a useful theory?

My first thought is to look at cross-country comparisons. Did those countries that had the most bank failures and other financial market disruptions also seem to have the biggest adverse shift in their Phillips Curves (and so had higher inflation than one would have expected given the depth of their recession and their previous level of inflation)?

Doing that cross country test rigourously would not be straightforward. You would need some sort of index to measure the degree of bank failures (including banks that didn't fail but had to contract their operations to avoid failure) and other financial disruptions. Plus, since the depth of the recession would also be correlated with that index of financial stress (with causation running both ways) you would need to be able to distinguish between shifts in the Phillips Curve and movements along a given Phillips Curve, which requires some sort of judgement about the slope of the (short run) Phillips Curve.

My second thought is to look at historical comparisons. Comparing past recessions, did inflation fall more quickly, for a given depth of recession, in those recessions that were not associated with bank failures and other financial market disruptions? The Great Depression had bank failures, and deflation, but was also a very deep recession. So I'm not sure what lesson to draw there. And Canada's history isn't very useful as a test of the theory, simply because there haven't been enough bank failures.

So, does anyone have any better ideas about how to test or measure whether this effect is big enough to matter? Can it can explain (part of) the cross-country or historical experience?

My sense is that it might help explain the recent cross-country experience. For example, countries with bank problems like Greece, Spain, the UK, and the US have had much worse recessions than Canada, but they haven't had bigger falls in inflation. But I'm not sure.

Comments

[wh10]

Is (unusually) large amounts of fiscal stimulus not a possible reason in your mind?

[wh10]

(well I actually don't know what the Canadian economy did, just U.S.)

[Nick Rowe]

wh10: it is theoretically possible that an increase in government spending would cause an adverse shift in the short run Phillips Curve. For example, suppose all the new government spending is concentrated in one sector, and resources cannot easily move into that sector. That would increase the amount of structural unemployment associated with any given average level of unemployment. It would increase the natural rate of unemployment, in other words, at least temporarily.

Yep. Put that down under reason #4.

(I'm not sure how well it works to explain the facts, however.)

[Ralph Musgrave]

If and when it IS SHOWN that bank failures have an adverse supply side effect, I hope this won’t be taken as a reason to continue with the taxpayer support for banks (TBTF subsidy, etc).

The adverse effect probably exists, but it ought to be temporary. And by “temporary”, I mean “possibly several years”. That is, reduced availability of bank funding will ultimately induce firms to find alternative sources of funding (e.g. equity), and/or engage in less capital intensive forms of production. And nothing wrong with that.

Bank assets and liabilities in the UK have expanded TEN FOLD in the last 30 years relaltive to GDP, and to what benefit? Economic growth is no better than 30 years ago.

[Nick Rowe]

Ralph: for this question (where's the deflation?) a temporary effect (lasting several years) is all that's needed. Yes, eventually financial systems would get re-created. And if you are right about banks being an inefficient form of financial intermediation, that only exist because of explicit and implicit subsidies, the long run effect could go the other way. But the transition to that new long run might be painful and take a long time, under current circumstances. Though increasing AD/NGDP might speed it up.

[Kevin Donoghue]

I don't know how persuasive Irish experience is, but FWIW the deep recession of the 1980s involved no bank failures worth mentioning. Also the 6-month 1970 bank strike was a non-event, with no effect on the Phillips Curve AFAIK:

http://en.wikipedia.org/wiki/Irish_bank_strikes_1966-1976

[Nick Rowe]

Kevin: if the Irish 1980's recession was about as deep as the recent recession, but had a bigger drop in inflation, that would confirm my hypothesis.

I didn't know about the Irish bank strike. That does look to be a good natural experiment (unless the bank strike was in turn caused by something else that also caused the high inflation and unemployment of the time).

Link for the Wiki (which doesn't say much, but has good links).
Interesting how a parallel credit system started to get created.

Must figure out (again) how to try to get that Manchester School article.

[J.V. Dubois]

Excellent post: I would just add another point that is needed to analyse for impact of this: debt versus equity financing of businesses. The point is that in most developing countries private equity financing is not widely used and these countries rely on bank financing of businesses. There is a nice study of these pehenomena from ECB here: http://www.ecb.int/pub/pdf/scpops/ecbocp63.pdf

So the point is, that countries that have less developed equity markets tend to have their corporations financed by debt. This in turn makes banking sector and its health more important from the supply point of view as you describe.

[Nick Rowe]

OK. I figured out how to find and read Antoin Murphy's Manchester School article on the Irish bank strike via the Carleton Library. It's a fascinating article, but unfortunately for me it only shows the effects on (presumably nominal?) retail sales (he's a little sloppy in slipping between "statistical significance" and "economic significance"). He doesn't discuss its impacts on the Phillips Curve. It's an AD rather than an AS analysis. It's all about banks as providers of money as a medium of exchange, which is what I'm normally interested in. But the one time I'm interested in banks as providers of loans, it's not there! Irony!

J.V. Thanks! That makes sense to me. Not all countries are equally reliant on banks. (One of the fascinating things about the Murphy article is that it discusses the high level of personal trust in Ireland, where people knew each other, and so were prepared to accept and circulate cheques even though those cheques could not be deposited for months).

[Steve Roth]

I'm with Ralph Musgrave on the longer term effects:

"reduced availability of bank funding will ultimately induce firms to find alternative sources of funding (e.g. equity), and/or engage in less capital intensive forms of production."

I think even more important: it would induce them to retain profits and use those for funding instead of debt and equity. Not a bad thing either. Think Apple.

Evidence: NFIB business owners who say that financing and interest rates are a crucial constraint, since '86:

http://www.nfib.com/Portals/0/PDF/AllUsers/research/sbet/small-business-top-problem-II-201206.jpg

That item is always dead last on the business owners' list of constraints, was so even in the depths of the "credit crisis."

At risk of spamming with links to my posts (sorry Nick!), more support with links here:

http://www.asymptosis.com/the-sky-is-falling-business-lending-down-1-2-percent.html

http://www.asymptosis.com/no-kidding-loan-demand-not-credit-is-the-problem.html

It's not hard to tell a story in which lots of easy lending to businesses is detrimental to aggregate growth in the medium to long term.

[Diego Espinosa]

The shadow banking system was based on a technology of portfolio management. Shadow banks engaged in maturity transformation by buying safe assets long and funding ultra-short. The idea was that this activity needed little capital cushion based on certain analytics: Value at Risk, Guassian Cupolas, house price volatility, etc. These analytics were either flawed (they mis-gauged tail risk) or incorrectly applied (using too-small historical samples, for instance). The balance sheets contained more risk than expected, rendering shadow bank liabilities "unsafe". The inability of shadow banks to manufacture safe assets is arguably the structural cause of the "excess demand for safe assets". In short, a technology came into being that caused a one-time boost in AS (1998-2006) as shadow bank assets spiked; and a one-time adverse shock as those assets collapsed.

This is not about the housing sector. In the U.S., hosing collateral backed shadow bank assets; in Europe, it was sovereign bond collateral. The issue is not the type of collateral, but the broader effect of shadow bank asset growth/decline on the economy. For instance, in the U.S., home equity loans were a new, widely-used source of start-up capital for small businesses. That source has now dried up.

Its useful to note that 80% of a typical subprime deal was rated AAA. The technology to do this is now, arguably, defunct.

[Nick Rowe]

Steve: Yep, simply asking firms if finance is a problem (or more of a problem than in the past) would be one way to test the theory (not conclusive of course, but then no single test rarely is). From reading your posts, I get the sense that this theory doesn't seem to work well (i.e. any effect isn't big enough to matter much) for the UK and US, but it might matter more in some Eurozone countries where firms do report problems. Your quote from the Economist: "According to a Gallup poll published in September [2009?], small businesses in Greece and Croatia say that access to finance is their biggest problem. Credit concerns are high on the list for small firms in France, Hungary and Italy as well."

[123]

Nick, good post. Although a bit too narrow, as there are other financial intermediaries - central banks and government treasuries, that have AS impact too.

[Gregor Bush]

Very interesting Nick,
Your story would seem to be consistent with the dramatic widening of corporate margins that we've seen since the beginning of recession. Unit labour costs have been very weak (particularly in the US) while selling prices have seen much less downward pressure. This widening of margins appears to be a contributing factor to the surge in corporate profit/NGDP ratio. So perhaps you could look at retail & wholesale margins and compare those to shifts in the Phillips curves across countries.

Another interesting thing might be to look at bank lending standards across countries and see if they're related to shifts in the Phillips curve. Certainly Spain seems like a case where there was (and is) major financial sector problems and a substantial shift in the Phillips curve (the unemployment rate is up 17% with only modest declines in inflation).

Of course, your story also argues for easier monetary policy. If financial stress is endogenous to weak NGDP growth and also causes adverse shifts in the Phillips curve, central banks should aim for higher inflation/NGDP growth than would be the case if financial stress had no impact on the Phillips curve.

[Ritwik]

Diego

What you present as a technological shock is a re-assessment of wealth, in Cowen-speak. Or destruction of capital, in a Fischer Black-ian financial GE. An AS shock, yes.

I'd say that the best way to capture a conjoint AD/AS shock is from a Wicksellian natural rate perspective. Intertemporal evaluations have changed - you could see it form the perspective of demand, or supply.

[jt]

Although, not strictly what you are looking for, you could analyze by industrial sectors, where especially in cyclical industries, credit dries up extremely quickly which would be analogous to a bank failure. The advantage is you would have many more data points to work with.

[Bill Woolsey]

Ritwik:

This makes no sense to me.

Did labor supply decrease?

Did the existing capital stock decrease?

Was there a change in the composition of demand so that the existing capital goods were less appropriate?

Change in intertemporal evaluations? So?

[Ritwik]

Bill

1. The labour supply did not decrease, the the economy is not at the labour supply curve, it is at the labour demand curve. That is the demand-side formulation of the issue. Following Leijonhufvud (and Phelps, and indeed any other 'Keynesian') involuntary unemployment is isomorphic with the market for interest rates not clearing. This is an AD disequilibrium.

2. The capital stock did not decrease, if you're committed to the idea of capital as buildings + factories. If you understand capital as the present value of expected future earnings, capital stock decreases and increases instantaneously. AD economists typically try to model this as wealth effects on consumption, or q theory of investment, but following Fischer Black (Via Perry Mehrling - http://economics.barnard.edu/sites/default/files/inline/understanding_fischer_black.pdf - pgs 14 and 15, especialy), the idea goes much further, into the production function itself. In this finance-view of the world, a lot of capital stock has simply been destroyed. This is also isomorphic to a disequilibrium in the market for interest rates. This is, however, an AS disequilbrium.

You could argue for this or that, of both in differing magnitudes. I think of Tyler Cowen's (or indeed Raghuram Rajan's) partially AS view of the current economic scenario as being subsumable into the broader Fischer Black framework.

Diego's formulation of the collapse of securitization as a technology shock is also consistent with these frameworks. The natural real rate of interest is clearly negative. Securitization helped keep it positive for a few years longer than it would have been otherwise. A collapse of securitization is a collapse in technology, which is indistinguishable from a collapse in capital, as capital and technology are indistinguishable in the Fischer Black framework. It's the AS shock that Nick is searching for.

My apologies if I'm seriously misreading any of the brilliant economists I invoke in my arguments above. I don't think I am, but macro/money is hard. So I may well be.

[Alan T]

6. nominal wage rigidity

[Nick Rowe]

Alan T: that's #2.

jt. Hmmm. Maybe. But there might be an endogeneity problem. If those industries behave differently anyway, and have a different Phillips Curve, that might explain why credit dries up. In principle I could face the same endogeneity problem in a cross-country comparison. But at least in principle there might be exogenous differences in soundness of banks across countries that are independent of their Phillips Curves.

Gregor: thanks!

"Your story would seem to be consistent with the dramatic widening of corporate margins that we've seen since the beginning of recession."

Hmmm. Yes, I think it would. But I would have to work on the theory a bit more (I have only sketched it out above) to be more sure of that. But it does sound right.

[Simon van Norden]

Okay, I'll ask the ignorant question.

"If producers of intermediate goods go bust, that's a real shock. The demand for factors of production like labour will fall. But the supply of final goods will fall too. It creates a bigger wedge between input prices and output prices."

Um....does that last sentence means that the markup is rising? Is that what you meant to say? Do we think the markup rose during the crisis?

[Nick Rowe]

Simon: it can't be that ignorant, because I had to think about the answer.

I *think* it means that for any given level of output and employment, the markup would need to be higher. But output and employment won't be given, especially if there's an AD shock at the same time. And, in general, real wages (for example) can vary either procyclically, or countercyclically, depending on whether wages or prices are stickier, and whether MPL and elasticity of demand is decreasing or increasing in aggregate output.

I would need to do a proper model to get my head around it clearly. But I think we could say that markups would be higher in a recession than they otherwise would be in the same recession without bank problems.

Gregor Bush above says that corporate margins did widen dramatically (in the US?).

[Gregor Bush]

Simon,
"does that last sentence means that the markup is rising? Is that what you meant to say? Do we think the markup rose during the crisis?"

Doesn't it look like the markup has risen, particularly in the US? The price deflator/ULC ratio surged during/following the recession. And it looks to have contributed to the rise in the corporate profit share. Changes in the markup are strongly correlated with changes in the profit share historically.

[Luis]

Very good post.
I've thought that in a hard recession (or Depression) both demand And supply suffer contractive effects, but not so in prices.. Everything day I see less quality at higher prices in the good and services in te market. The first reaction of producer is looking for their margin. Lay off and reducing quality and quantity is the priority, before lowering price. So you get a much more rigidity In prices than cost. The monetary and fiscal policy support this strategy.
I Spain obviously it is te fiscal policy (of high growth in current expenditures) te main support for price rigidity and no sign -until now- of Deflation.
perhaps next rounds of evenements..

[rsj]

Perhaps we should think of banks, and financial intermediaries more generally, as like producers of an intermediate good. If producers of intermediate goods go bust, that's a real shock. The demand for factors of production like labour will fall. But the supply of final goods will fall too. It creates a bigger wedge between input prices and output prices.

Of course it is not just banks, but credit markets more generally. And it is not that funding is not available, but only on different terms.

In other words, this is a complicated way and excessively round about way of saying the interest rate is too high. All the other consequences -- prices not falling enough, unemployment -- are natural results of the interest rate being too high.

Why is the interest rate too high? Because the central bank does not set the rate. Say the short term rate is 2% (nominal). Then I am indifferent between buying a capital good that will decline in value by 4% over the next period and pay me a 6% yield, or I can buy a capital good that will appreciate in value by 1% and pay me a 1% yield.

So that nominal rate of 2% as set by the central bank is perfectly consistent with capital rental yields of 6%, or capital rental yields of 1%.

Perhaps we are now in a situation in which capital rental yields are high but the interest as set by the CB is low -- both are consistent with each other.

[bamboo investments]

Very interesting post. Take the UK as well. Back in recession, real wages have fallen by about 10% over the last few years, yet there's been inflation of 3% or higher for 30 straight months.

[Mark A. Sadowski]

On the subject of iflation and markups, disinflation during the eighties and the nineties was accompanied by a significant rise in the profit share of national income in most OECD countries. This suggests that changes in the rate of inflation are non-neutral with respect to the distribution of factor income. The mechanism by which this comes about is fairly simple. Disinflation is correlated to rising unemployment rates, rising unemployment rates lead to lower labor bargaining power, and lower labor bargaining power is correlated with higher markups. Furthermore, lower inflation rates create less price dispersion leading to less competition among producers to limit markups. This hypothesis was tested with a panel of 15 OECD countries over the period from 1960 to 2000 and a robust negative relationship between disinflation and the labor share of factor income was obtained:

http://pareto.uab.es/wp/2000/46000.pdf

On the subject of the supply side of banks, oddly, I've had a similar intuition, although I'm not sure it is acting through the credit channel. Firms are sitting on tons of cash precisely due to the increased markups. However, I don't have an alternate theory just yet.

What got me thinking about this is the following. I was curious how the various major currency areas compared in terms of the pre and post crisis growth and inflation trends. 2008Q3 seems to be the break in trend for all of the big four. Here are the results:

Average annual rate of change (%) 1998Q3-2008Q3

Currency Area-NGDP–RGDP-GDPDeflator-CPI-Difference
Euro-17——-4.09–2.04—2.02——2.26—0.24
US————4.99–2.48—2.45——2.97—0.52
Japan——-(-0.28)-1.09-(-1.35)—–0.00—1.35
UK————4.93–2.61—2.25——1.86-(-0.39)

Average annual rate of change (%) 2008Q3-2011Q4

Currency Area-NGDP—-RGDP–GDPDeflator-CPI-Difference
Euro-17——-0.57–(-0.38)—0.94——1.57—0.63
US————1.93—-0.56—-1.38——1.12-(-0.26)
Japan——-(-1.74)-(-0.41)-(-1.33)—(-0.87)–0.46
UK————2.03–(-0.30)—2.34——3.15—0.81

Even if one ignores CPI and focuses purely on the GDP deflator (which of course measures output inflation) doesn't it seem like the UK has undergone some kind of negative AS shock? Inflation has been higher although real growth is lower. I've been scratching my brains trying to come up with a reasonable explanation and one of the first things that came to my mind was the size of the UK financial sector. Of course the US also has a large financial sector, but I suppose it all depends on how you measure things.

In any case it's interesting to find someone else has also considered the possibility of a negative financial sector AS shock.

[Nick Rowe]

Mark: I'm going to disagree with you on one point (I agree with the rest):

"The mechanism by which this comes about is fairly simple. Disinflation is correlated to rising unemployment rates, rising unemployment rates lead to lower labor bargaining power, and lower labor bargaining power is correlated with higher markups."

I don't think it's that simple. Disinfaltion is also correlated with an excess supply of output, not just labour, which should lower firms' bargaining power vis a vis their customers, which should result in lower markups. There's downward pressure on wages relative to prices, but also downward pressure on prices relative to wages. So the question is: which falls fastest: wages or prices? And I think that comes back to: which is stickier?

[Jacques René Giguère]

Nick:"And I think that comes back to: which is stickier?". That is, who will starve first? At this game, most workers will lose to most businesses.

[The Keystone Garter]

Nick, I'm pretty sure you need the model which might mean an actual paper. I forget what the models are called that measure correlations.
Banks lend to firms but also get bailed out of mistakes. This suggests to me banks should lend in weakness much more. There are a variety of ways of doing it. You could raise finance corporate taxes (or just enact a FTT) in GDP growth and lower in recession. You could force loan totals to be higher when unemployment is high and force lending cutbacks (interest rates rise) when employment is high...At least if banks are making "bad" loans in recession it is creating employment that might trickle down to "good" actors. I'd prefer to live in a society where the money the banks get bailed out with, makes me live longer or gives me more opportunities. The chip insertion instead of sideways swipe, notwithstanding.

[Steve Roth]

Just continuing on the importance of easily available lending to firms, well worth reading JW Mason, as always:

"the collapse was largely limited to financial commercial paper. Nonfinancial borrowers did not lose access to credit in the way that banks and shadow banks did. The gap between the financial and nonfinancial commercial paper markets wasn't discussed, I believe, because of the way the crisis was seen entirely through the eyes of finance."

http://slackwire.blogspot.com/2012/06/in-which-i-dare-to-correct-felix-salmon.html

[Britmouse]

Mark, did you strip out indirect tax changes from your UK deflator/CPI calculations? It should make a difference, albeit a small one.

Very interesting post, Nick. From a UK perspective (sorry, I have a one-track mind); your thesis seems to match very closely what Adam Posen has been saying at the BoE, e.g. here; Posen has particular concerns with new firm entry, and capital reallocation across sectors.

There is survey data (no link handy, sorry) from UK SMEs which suggests their constraints are more demand-side than credit availability. For larger companies, it seems hard to believe that the UK capital markets cannot easily substitute for bank lending if required. But that's all mostly anecdotal; the UK supply side is a real puzzle.

[Mark A. Sadowski]

Britmouse,
Actually I made no attempt to correct CPI for other effects intentionally.

On a more important note, I was not aware that there are deflator series that strip out indirect tax changes. I can't find such series at Eurostat and the UK's ONS isn't very user friendly (to say the least).

[Britmouse]

Mark, the ONS do not produce any such deflator series, sadly. There was a significant increase in VAT (15% to 20% over 25 months) which is a quite a distortion.

I have used recently the difference between (growth of) NGDP at basic prices (ONS time series ABML) and real GDP to derive such a deflator measure, which is crude but I think correct.

[Sergei]

This is similar to arguing that Basel3 with its capital and liquidity requirements puts pressure on supply side. Well, it does so but only if central bank fails miserably on its monetary policy. All that Basel3 achieves is a "fiscal"-like redistribution of income from the real sector to financial sector (digging regulatory holes and filling them again) plus the standard redistribution between creditors and borrowers. Basel3 simply leads to lower interest rates.

The reason solvent and profitable Greek firms are failing is because ECB is miserably failing in its job. And banks always fail only because central bank fails to do what it has to do, i.e. supervision. Capital requirements there are not just for fun and enforcement of those is the primary task of central bank.

[Mark A. Sadowski]

@Nick,
"I don't think it's that simple. Disinflation is also correlated with an excess supply of output, not just labour, which should lower firms' bargaining power vis a vis their customers, which should result in lower markups. There's downward pressure on wages relative to prices, but also downward pressure on prices relative to wages. So the question is: which falls fastest: wages or prices? And I think that comes back to: which is stickier?"

But we don't even have to wonder about this, as we have the rather robust results for the OECD over 1960-2000. Markups evidently go up during disinflation and down during accelerating inflation. So you've seemingly answered your own question.

@Britmouse,
It never occured to me to look for such data.

The only problem I see with such a computation would be if there were some other conflicting policy that occured at the same time (a change in subsidies?). But I can't recall anything like that.

I've redone the calculations using nominal GDP at basic prices for the UK:

Average annual rate of change (%) 1998Q3-2008Q3

Currency Area-NGDP–RGDP-GDPDeflator-CPI-Difference
Euro-17——-4.09–2.04—2.02——2.26—0.24
US————4.99–2.48—2.45——2.97—0.52
Japan——-(-0.28)-1.09-(-1.35)—–0.00—1.35
UK————5.01–2.61—2.33——1.86-(-0.47)

Average annual rate of change (%) 2008Q3-2011Q4

Currency Area-NGDP—-RGDP–GDPDeflator-CPI-Difference
Euro-17——-0.57–(-0.38)—0.94——1.57—0.63
US————1.93—-0.56—-1.38——1.12-(-0.26)
Japan——-(-1.74)-(-0.41)-(-1.33)—(-0.87)–0.46
UK————1.51–(-0.30)—1.82——3.15—1.33

So the anomalous AS shock seems to disappear. The UK is clearly having a harder time adjusting to lower NGDP growth than the US (greater nominal rigidities?) but we no longer see accelerating inflation coupled with decelerating real growth. (Why isn't this getting more attention?)

I tried replicating your method with the other major currency zones. It appears that the US and Japan do not maintain equivalent series. The eurozone calls it "Gross value added (at basic prices)." Eurostat starts computing such data for the eurozone-17 in 2000Q1. The differences in growth rates between nominal GDP and GVA over 2000Q1/2008Q3 and 2008Q3/2011Q4 are not important (less than 0.05 points.)

I think you definitely have something. This no longer appears to be a mystery at all.

[Britmouse]

Mark, good to hear you confirm that. I am not aware of any changes to subsidies either but I will have a look through the budgets when I get a chance - I cannot think what subsidies exist in the UK offhand.

("GVA at basic prices" is the name used by the ONS too.)

I agree this data should get more attention; I've tried. :)