Ricardo Strikes Back: The Net Effect of the Irish Credit Bubble on Cumulative GDP

A research colleague, Brian O’Kelly, and I have been looking at the impact of poor bank regulation in Ireland during 2003-2008, and the Irish credit bubble that this poor regulation fostered, on Irish GDP and other economic measures.  I am scheduled to talk about this research down in Kenmare this weekend.  I want to make some informal remarks in this blog, based on some simple calculations that are easy to follow.

Here is a simple question: what is the effect on cumulative GDP of the massive distortions in bank lending described in the Honohan report?  The good news is that the net cumulative effect, even after accounting for the €50 billion bank bailout costs, is not too far from zero.  The bad news is that the GDP benefits were all received during the 2003-2007 period when GDP was artificially increased by these extremely poor bank regulation policies, whereas the GDP costs are in the 2008-and-beyond period.   

Consider the Irish GDP real growth record since 2002.  During the 2002-2007 period (inclusive) Irish GDP had a very high real growth rate of 7.97% per annum compounded annually (this is aggregate, not per capita).  Then the Irish debt crisis hit and over the next three years, 2008-2010, the real growth rate (Eurostat estimates) was -5.21% per annum.  But this still leaves the real GDP growth rate over 2002-2010 at a very respectable 3.3% per annum. 

The Honohan report (Chart 2.10) notes that without the Irish debt crisis, facilitated by faulty bank regulation, the per annum decline in Irish GDP during 2008-2010 would have been cut by about three-quarters, that is, a decline of -5.21%/4 = -1.3%.  This smaller decline reflects the worldwide Great Recession and is not due to Irish bank regulation failures. What the Honohan report’s Chart does not show, however, is the effect of faulty bank regulation on the very high growth rate during the early 2002-2007 period.  This growth was substantially based on the massive foreign credit inflow facilitated by the banks (and poor bank regulation) rather than by true underlying productivity growth.  Net foreign borrowing by the domestic banking sector increased enormously during this period, rising to 87.5% of previous-year GDP in the third quarter of 2008 when the crisis broke.  All this overseas inter-bank borrowing, with the proceeds fed into property development spending, fuelled the big GDP increases of this period in an unsustainable way.  To quote from Kelly (2010) “The Irish Credit Bubble”:

“It is this more than tripling of bank lending that accounts for the Irish boom since 1997. Effectively the Irish economy segued from one driven by competitiveness in the 1990s to one driven by a credit fuelled bubble in the 2000s.”

O’Kelly and I estimate that perhaps one-half or even more of the growth during 2002-2007 is a credit-fuelled bubble, which depended upon poor bank regulation (and big immigration inflows to provide labour supporting the capital inflows).  If we assume instead good bank regulation and dilute the 2002-2007 growth by half, and the 2008-2010 decline by three-quarters, we actually wind up with 2010 GDP (before bank bailout costs and other deadweight costs) which is 7.7% lower in the absence of the credit-fuelled bubble.  This must be adjusted further for the interim GDP during 2003-2007 which was much higher on account of the bubble.  Even after adjusting for €50 billion bank bailout costs, and deadweight costs of fiscal adjustment, the cumulative GDP impact of the credit bubble is not that large.  It is just that the “good part” of the bubble experience is in the now-distant past (2002-2007) whereas the pain is in the present and future.  Note also this analysis is on a total basis not per capita, whereas the per capita figures are more difficult to interpret in some ways.

Needless to say, the Irish credit bubble of 2002-2010 was a terrible economic disaster, clearly caused to a large extent by grievously bad bank regulatory policy (as the Honohan report demonstrates).  This binge-based increase and calamitous decline in GDP during 2002-2010 has caused enormous social distress, unemployment and dislocation.  From the hard-headed perspective of economic analysis, though, it is interesting to note that the cumulative GDP impact is not that large over the entire period: on the limited metric of real aggregate GDP, the falsely-based gains in the early period roughly match the painful losses in the later period.  To quote Paul Krugman, “Economics is not a morality play.”  The cumulative aggregate real GDP impact of the Irish debt bubble and debt crisis does not have the “good-guys vs bad-guys” theme that we expect from a well-written morality play.       

28 replies on “Ricardo Strikes Back: The Net Effect of the Irish Credit Bubble on Cumulative GDP”

Do you have any idea how much of the fixed investment component of GDP has yet to be written down to realistic levels?

It is my understanding (probably false!) that new building in a year adds to GDP based on the value of that building. If the value of the existing stock falls, there will be a negative impact on the existing stock value. As I say, that’s my understanding!

With commercial property prices (so existing commercial property values) and residential prices still falling, this adjustment has some way to go?

Thanks for any insight.

Greg,presumably if one includes risk aversion and, perhaps more importantly loss aversion, then one would be less sanguine about the effect of the policy failure? That is, a smoother and preferably monotone path to where we are now would have been better in welfare terms?

@ Gregory

Thanks for the heads up on your analysis. I will be interested to read the paper.

Are you familiar with Steve Keen’s recent work which has focused on the contribution of additional levels of debt to aggregate demand and GDP in the U.S. and now to how private sector deleveraging is affecting the same variables?
http://www.debtdeflation.com/blogs/2010/09/20/deleveraging-with-a-twist/

My sense is that, despite the similarities in the scale of the credit bubbles here and in the U.S., your analysis ascribes a lesser importance to debt than his does – the cumulative GDP impact of the credit bubble is not that large. This is surprising to me as the bubble here was larger than in the U.S. and in addition, he does not include bailout costs.

My sense is that it might be methodological differences in how you model the impact of increases / decreases in debt that are at the heart of this difference between your results. Do you have a perspective on this question?

Well, well. Some people finally figured it out. Aggregate growth (real) has to be netted against debt (virtual) growth. Real wealth has to be netted against virtual wealth. Well done lads!

@ hoganmahew: “With commercial property prices (so existing commercial property values) and residential prices still falling, this adjustment has some way to go? ”

Yep, a really long way! If property asset values incremented by 7% per annum from 1996, with a short dip in 1999 – 2002, and we had a virtual increment of 350% overall. That looks like a pretty steep drop! Kelly was wrong: He underestimated the collapse.

Brian P

Your excellent analysis makes nonsense of the claims by David McWilliams and Morgan Kelly that all the growth in the Irish economy since around 2000 was due to building houses, other construction, and credit.

How can it be?

Since new house building is at virtually zero in 2010, other construction is at rock bottom, and credit is at abysmal levels, if their claim were true, then GDP in 2010 would be back to its 2000 level, or lower (since there were quite a lot of new houses built in 2000). But, your figures show that even after the almost total removal of new house building and other construction from the economy by 2010, its average growth rate between 2002 and 2010 was a very respectable 3.3%. Even higher if we use the period 2000 to 2010. Eurostat gives the following figures for GDP in 2010 to base 2000=100.0.

Luxembourg 133.0
Ireland 128.1
Greece 127.5
Spain 122.4
Sweden 118.3
Finland 117.9
U. Sates 117.9
Austria 115.7
U. Kingdom 114.7
Netherlands 113.8
Belgium 113.7
France 112.4
Denmark 107.4
Germany 106.5
Japan 106.5
Portugal 105.9
Italy 102.3

Well, surprise, surprise, after all that’s been written.

Ireland coming in 2nd, Greece 3rd. France 12th. Germany 14th.

The reason for all this growth is that, contrary to all that’s been said, manufacturing and services exports did not dry up around 2000. They continued to grow strongly post 2000, dipped much less than in other countries during the global recession, and are growing strongly again now. Manufacturing figures for August published just today – manufacturing output up 10.4% in August over a year ago, with 17 out of 27 sectors showing y-o-y growth (so not down to viagra).

The decade 2000 to 2010 saw the worst global recession since the 1930s, and the complete collapse of the Irish construction industry by the end of it. Yet, GDP was still 28.1% higher in 2010 than in 2000, greater than in any other EU15/US/Japan, apart from tiny Luxembourg. It is unlikely that the next decade will see a global recession anywhere near as bad, maybe even no global recession at all. And, construction in Ireland is at rock bottom, so its only way in the next decade is UP. Taking these into account, over the next decade from 2010 to 2020, there is every prospect that GDP in Ireland will grow by much more than its 2000-2010 figure of 28.1%. I think we should be aiming for 50% over 2010-2020 at least.

@Kevin Denny

Absolutely correct – that is a big explanation for why social welfare is so much lowered by boom-bust patterns like this, even if the cumulative impact is not that big in terms of cumulative GDP over the boom-bust cycle.

By Post Author

Does anyone know what would have been the situation with employment/unemployment if we didn’t have a bubble during 2002-07. Becuase I think the quality of debate on this issue is lacking.

Assuming that we didn’t have large inflows of labour, where would all our construction workers/lawyers/accountants have been employed? It is likely that in the absence of our bubble we would have had higher unemployment than we did during that period.

What seems funny to me is that if we had a Financial Regulator in 2004/5/6 that got wind of a big investment project and the bank loans to fund it said STOP on the basis that it was bad banking policy and destabalising effects on the economy. We would have people ringing up Joe Duffy saying things like ‘My son, an electrican, was told he would have 6 months work on a new project but the someone from D4, the Financial Regulator said No, the project couldn’t go ahead. Some high fluteing language about stability of financial sector…..whatever that means Joe’. or Joe, my daugher just qualified from law school. She studied hard and got good grades, she was to be taken on by a law firm but she just got a call from the company and the job is gone. The firm was to get the contract overseeing some new property development, but someone called the Financial Regulator said it was unsustainable and risky to the economy in the long term’. Now Joe I’m an uneducated man, but I sent my daugher to college, why is she being denied a job because some overpaid fool from D4 is worried about risk’.

[reposting this as my attempt to post it got held up ‘for moderation’ due to the link]

@ Gregory

Thanks for the heads up on your analysis. I will be interested to read the paper.

Are you familiar with Steve Keen’s recent work which has focused on the contribution of additional levels of debt to aggregate demand and GDP in the U.S. and now to how private sector deleveraging is affecting the same variables?
[Google “Steve Keen deleveraging with a twist”]

Despite the similarities in the scale of the credit bubbles here and in the U.S., your analysis ascribes a lesser importance to debt than his does – the cumulative GDP impact of the credit bubble is not that large. This is surprising to me as the bubble here was larger than in the U.S. and in addition, he does not include bailout costs.

My sense is that it might be methodological differences in how you model the impact of increases / decreases in debt that are at the heart of this difference between your results. Do you have a perspective on this question?

By drawing a straight line between 2000 and 2010 are we ignoring the fact that we had a balanced budget in 2000, but we are adding a stimulus of 12% ish of GDP this year.

If we removed this 12% stimulus the GDP number would likely fall by this amount + some multiplier. Using a multiplier of 1.4 and JTO’s numbers about that would put us around 111.3 (2000 = 100).

I think the real problem with this kind of analysis is that the credit boom has caused greatest damage at the individual level.
The impact on GDP will be felt over the next 35 years and will increase with every rise in interest rates.
The nature of credit is to be limited in its upside and more exert a more protracted negative impact on the downside.
Ask the guys getting the boat!

Is it possible to estimate the extent of the total value of the fall in property values?

If it was possible, then it might be possible to estimate to what extent this was funded by borrowed money. & assuming that interest is paid on the part that was funded by borrowed money it might be possible to estimate how much extra interest is paid for literally no extra value…. A bit of drag…

@JTO
it’s always great to see the stats but what I don’t understand is: if everything is as good as you say why does this feel so bad?
Why does it not feel like manufacturing is up 10%. I think that although the heart is beating healthily we are bleeding capital through this bank bailout. And we’re slowly losing.
Even though we’re fundamentally healthy it’s the bleeding that will weaken us.

Gregory Connor’s analysis is based on the period 2002 to 2010. So, it is plausible.

Morgan Kelly’s analysis, however, is based on a longer period, starting in 1997. This is what Gregory Connor quotes Morgan Kelly as saying:

“It is this more than tripling of bank lending that accounts for the Irish boom since 1997. Effectively the Irish economy segued from one driven by competitiveness in the 1990s to one driven by a credit fuelled bubble in the 2000s.”

OK, let’s look at the figures (from the CSO website):

rise in GDP between 1997 Q1 and 2007 Q4 (peak of boom): +102.3%

fall in GDP between 2007 Q4 (peak of boom) and 2010 Q2: -13.4%

leaving:

rise in GDP between 1997 Q1 and 2010 Q2: +75.2%

So, GDP more than doubled between 1997 Q1 and the peak in 2007 Q4. If it was all due to bank lending and building houses, as Morgan Kelly claims, how come only a small proportion of that massive growth went after 2007 Q4? And, of course, there was a global recession during the period when that small proportion went. The economy is still +75.2% larger in the latest quarter than it was in 1997 Q1, which is miles ahead of any other developed country, and GDP is now rising again. If Morgan Kelly’s analysis is correct, and correct back as far as 1997, how is this possible? Gregory Connor’s analysis is plausible because he picks 2002 as the start date. Morgan Kelly renders his analysis daft by picking 1997 as the start date. The reason he did so is, of course, because that was the year FF/PD came to power.

Any chance of Morgan Kelly, or one of his disciples, answering that question?

I’m not holding my breath.

It’s too late for these reactions to be dependable, but one immediate reaction is to wonder whether aggregate GDP = wealth, or not. If we borrowed all that money and used it to dig a great big hole in the ground, then the “investment” would count as GDP every year. Unfortunately, at the end of the project we’d have a big hole in the asset column of our balance sheet and a big debt in the liability column. The year by year GDP data would be nice, but largely irrelevant to the ultimate situation….which situation would be somehow similar to our current position.

Then, I’d also wonder how much of today’s GDP is govt spending. Govt borrowing has replaced a lot of private borrowing and even if you exclude transfer payments there’s a lot of govt spending left in the GDP number, and spending is a LOT higher than 2002. If we went back to 2002 spending, how much would GDP be today? Then, if I looked back at the wealth situation, I’d wonder for how long I could keep borrowing my GDP to try to keep it higher than 2002.

Gregory
Well done!
Simple is indeed what is required. Truth is always simple. Economics specializes in complexity. Well done, in arriving at a very strong truth! The morality, Krugman is an Ass, is that we should/not repay the money borrowed and wasted! Those who lent enabled the bubble.

However, the estimable analysis neglects the actual effects of the borrowing, understandably! I poresume that you agree that the bubble was wasted money, malinvestment that has enoprmous costs and is in noooooooo way neutral?

Brian Woods

+1 +10! As usual.

Mick Costigan

Steve Keen’s work may be seminal but is, again, too late for utility in this crisis! If a depression lasting three decades, for Japan at least, can be called a point in time=crisis. The others will eventually accept it or denigrate it as inimical to credit: such a necessary ingredient in modern kleptocracies!

Eureka
Hugh Sheehy

Read mine above for why it is so bad. We tend to glide over as you say, the actual effects of the bubble! Especially those who have slush funds that will not be exposed to tax. The effects are quantifiable and will last for much longer than the disease, the accumulation and waste of credit.

JTO

So where are you buying property these days? Or is Kelly in fact being modest?

http://www.rawstory.com/rs/2010/10/sp-60-countries-bankrupt-50-years/

S&P doing such a good job! It merely shows how much dubious credit was created in the squeeze period. Banks, shadow banks etc all joined in with Government help in Ireland and USA.

Jubilee looks likelier than ever, but what about the debit side? Each of these reckless loans represents somone’s asset! Who is going to take the loss? If we decide that all of this is a likely scenario, then gaming it makes sense, even if it is not moral? Should we then borrow as much as we can get? And ensure that what gets written off is the highest coupon? Whose assets are we damaging? Who would be so stupid as to lend to us and others who are likely to default?

S&P: what is their agenda?

Does this mean we didn’t “blow the boom”???

It is of course an uncomfortable fact that the bubble didn’t leave us empty handed and that light touch regulation (as opposed to wholly incompetent/non regulation) ws a valid policy at a particular stage.

As Micheal Martin has said, one of the big questions is how to prick a bubble, especially when Ireland doesn’t set its own interest rates.

I don’t want to make excuses. It has to be admitted that if we had tightened up home-loan credit substantially from 2004/2005 onwards then we could have avoided some of the worst losses we are now seeing in our banking sector.

I also think the fact, as aluded to by Gregory Connor, that this is not a per capita study is very important. It is also does not deal with median per capita production. The real pain in our society is being felt by the unemployed.

However, the vilification of the previous Government’s policies is over-done. Light-touch regulation brought us a lot of benefits and was not a secret.

Similarly, the Greenspan theory that bubbles should be allowed develop and burst had some truth in it, even if it did turn out to be wrong.

One of our problems in Ireland is that we did not have an eye on the global situation. We took no reponsibility for preparing for external factors despite designing ourselves as a small open economy.

We thought we could havea bubble, suffer its bursting and could recover thereafter. Unfortunately, our bubble was on top of another bubble which burst underneath it. While people did start to warn of this in 2006, our goose was effectively cooked at that stage.

I think Fianna Fail will have to take its medecine and it won’t do Fianna Fail or the country any harm when that happens. However, the fact is that while they made serious mistakes their over-arching policies were founded on the consensus modern economic thinking of the day and were designed to be anything but reckless.

The negative effects of the credit bubble are not over yet.

Moreover, the effects of the bubble would almost surely have been national bankruptcy in the absence of the backstop provided by our membership of the EU/Eurozone.

@christy
+1
I think there is some wilful misinterpretation going on of Mr. Connor and Mr. O’Kelly’s work:
“The bad news is that the GDP benefits were all received during the 2003-2007 period when GDP was artificially increased by these extremely poor bank regulation policies, whereas the GDP costs are in the 2008-and-beyond period.”

When the public finances are in balance and the bank bailout costs have been factored in and we see the effects of 100+% debt:GDP, then we will have some idea of the full negative impact. We can then calculate whether, on balance, there was a boon or bane to GDP from the credit bubble.

Central banks cannot fight asset bubbles.

The federal reserve could do nothing about the property bubble in the US due to the existence and actions of Freddie and Fannie. Risk was underpriced due to their existence and due to their targeted underpricing of risk for building and selling homes it was impossible for the Federal Reserve to use the broader weapon of raising interest rates without causing significant collateral damage to the rest of the economy.

The ECB could do nothing about the Irish property bubble as the bubble was locally reinforced by tax credits and possibly hidden from some due to poor availability of market data.

It is too soon to calculate the final cost of the Irish property bubble. When the banking system has been cleansed, then it might be possible to start.

I read in a newspaper in Sweden that one third of Irish household wealth has disappeared since the peak. If that is true, I’d say that number is a better approximation of the total cost.

@Zhou,

+1

A characteristically balanced assessment. The rot started when Greenspan, Rubin, Levitt and Summers convinced Clinton to sign-off on the bonfire of financial and banking regulation in 1998 which was stoked by the Bush administrations. Brown and Blair bought into this unreservedly. Located between the UK and US and seeking to develop the IFSC, there was little Ireland could do except play this game.

The bubble concealed serious and economically damaging domestic political and economic dysfunction. This is now being revealed, but unlike the UK and the US, the political cycle is out of kilter and the political will does not exist to address this dysfunction. We are entirely reliant on the EC and the ECB to guide and support an excessively painful recovery.

That is a very nice ‘mathematical’ analysis but does it not beg the question of whether the gains and losses arose symmetrically across the population?

I am reminded of the joke about the hunting economists – one shot to the left of the target, one to the right and the third exclaimed ‘On average, we hit it!’

I have yet to see an analysis that disproves the generally perceived notion that we have seen over the said decade a massive privatisation of gains and socialisation of losses.

@hogan. +1
Again, the key point is that GDP is at some level merely a measure of activity. Our wealth position is entirely different.

Essentially Ireland borrowed a big chunk of GDP from 2002 to 2007, and wasted a lot of it.

Ireland is now borrowing a big chunk GDP because that waste started to become apparent and people stopped spending money quite so wastefully…which meant that people lost their jobs, taxes dropped and spending rose.

Only when the borrowing caused by the waste and by the result of the waste stopping have flushed through the system will we be able to measure either “real” GDP or the actual net wealth impact…which I suspect will be hugely negative.

A credit bubble would have been a lot more economically interesting if it had pushed the speculation into broadband, nanotechnology, patents, medical research or genetic engineering.

Loose lending – speculator greed – mania – bust – long-term benefit remains.

Housing is primarily a utility. Seeing it through the prism of speculative asset classes misses how blowing a credit bubble in housing becomes deeply destructive to the long-term health and wealth of an economy.

Nice analysis….
My feedback would be that a) it assumes that the crisis is over and that may not be the case. On that basis we may fyind that the cumulative damage in the 1997 to 2012 period id substantially greater than the 1997 to 2010. Also, it assumes that the investments made in the bubble phase (2005 to early 2008) have some residual value and that may not be the case.

Debt generally carries a 5 year maturity. So the debt contracted in the 2005 to 2008 period is likely to mature in the 2010 to 2013 period (using this average 5 year tenure). If the asset is not worth the value of the debt assigned to it, then it won’t be refinanced and must be repaid or defaulted upon. Since the government has decreed that, in the event that the original debtor defaults, then the taxpayer will assume payment, that means that the interest costs on debt contracted to build useless assets (at least compared top the debt) will drag GNP for some time.

@Mick Costigan and others

The paper is available at the following “link” (but take out the blank spaces before and after the // to make the link activate)

http: // economics.nuim.ie/research/workingpapers/documents/N214-10.pdf

Comments are closed.