Deleveraging with a twist
The latest Flow of Funds release by the US Federal Reserve shows that the private sector is continuing to delever. However there are nuances in this process that to some extent explain why a recovery appeared feasible for a while.
The aggregate data is unambiguous: the US economy is delevering in a way that it hasn’t done since the Great Depression, from debt levels that are the highest in its history. The aggregate private debt to GDP ratio is now 267%, versus the peak level of 298% achieved back in February 2009–an absolute fall of 31 points and a percentage fall of 10.3% from the peak.
This dwarfs any previous post-WWII experience–even the steep recession of the mid-1970s.
| US Debt Downturns | Now | 1990s | 1970s |
| Duration in years | 1.3 | 2.3 | 1.3 |
| Peak Debt | 298 | 169 | 114 |
| Debt Trough | 268 | 163 | 106 |
| Fall in Debt | 31 | 7 | 8 |
| Percent Decline | 10.3% | 4.2% | 6.7% |
| Rate of Decline p.a. | 7.9% | 1.8% | 5.1% |
The aggregate level of private debt now towers over the economy, putting into sharp relief the obsession that politicians of all persuasions have had with the public debt. Rather like Nero fiddling as Rome burnt, politicians have focused on the lesser problem while the major one grew out of control. Now they are obsessing about a rise in the public debt, when in a very large measure that is occurring in response to the private sector’s deleveraging.
If they had paid attention to the level of private debt in the first place, then we wouldn’t be facing exploding public debt today.
However, though the decline in private debt is steep and continuing, the rate of decline has slowed. Because debt interacts with demand through its rate of change, this has given a stimulus of sorts to the economy in the midst of its deleveraging.
This is obvious when one considers aggregate demand as I define it: the sum of GDP plus the change in debt (where this demand is spread across both goods & services and the asset markets). Though debt levels are still falling, because they are falling less rapidly there has actually been a boost to aggregate demand from debt from the fact that debt is declining less rapidly in 2010 than in 2009:
This is doubly so when the contribution to demand from the public sector is included, as this shorter term graph shows more clearly.
However while recent data shows a positive contribution to demand from debt falling more slowly, on an annualised basis, the change in debt is still subtracting from aggregate demand–and more so than in the previous year. So total demand (across all markets–commodity and assets) had to fall, even though GDP itself grew. Obviously most of the fall in demand has been absorbed by the asset markets, which have not recovered to the same level of turnover as in the boom years–and nor should they.
The next table, which uses the aggregate debt figure (public and private debt combined) from the Flow of Funds, shows that aggregate demand fell across July 2008 to June 2009, even though debt was still rising, because the rate of growth of debt fell from $3.7 trillion to $1.4 trillion. Across July 2009 to June 2010, the decline in aggregate demand was less than the previous year (a 9.7% fall versus a 15.2% fall), even though the change in debt had turned negative.
| Variable\Year | 2006.5 | 2007.5 | 2008.5 | 2009.5 | 2010.5 |
| GDP | 13,347,800 | 14,008,200 | 14,471,800 | 14,034,500 | 14,575,000 |
| Change in Nominal GDP % | 6.6% | 4.9% | 3.3% | –3.0% | 3.9% |
| Change in Real GDP % | 3.0% | 1.8% | 1.2% | –4.1% | 3.0% |
| Inflation Rate % | 4.1% | 2.4% | 5.6% | –2.1% | N/A |
| Total Debt | 43,337,326 | 47,528,151 | 51,272,735 | 52,686,684 | 52,054,500 |
| Debt Growth Rate % | 10.0% | 9.7% | 7.9% | 2.8% | –1.2% |
| Change in Debt | 3,934,348 | 4,190,825 | 3,744,584 | 1,413,949 | –632,184 |
| GDP + Change in Debt | 17,282,148 | 18,199,025 | 18,216,384 | 15,448,449 | 13,942,816 |
| Change in Aggregate Demand % | 0.0% | 5.3% | 0.1% | –15.2% | –9.7% |
The rise in aggregate demand supported a recovery in employment, but the prospects of this continuing to the point at which economic activity booms once more are remote: with debt levels as high as they are, the potential for further deleveraging still exceeds the worst that the US experienced during the Great Depression.
I have recently become aware of some other economists using a similar concept to my measure of the debt contribution to aggregate demand, which they call the “credit impulse” (Biggs, Mayer et al., http://ssrn.com/paper=1595980). They define this as the change in the change in debt, divided by GDP.
My definition emphasises aggregate demand and correlates this with the level of employment (or unemployment, as above), whereas theirs emphasises the change in aggregate demand and correlates with changes in the level of employment. The logic is identical, but has the advantage of being able to correlate the change in the change in debt with change in employment. It highlights an apparent paradox: the economy can receive a boost from debt, even though it is falling, if the rate of that decline slows.
The next few charts apply this concept using the recent Flow of Funds data, and shows why it is so important to consider the dynamics of debt when trying to understand why this downturn has been so severe—and why it also seems to have eased. Firstly, change in employment and change in real GDP are obviously correlated, and on this basis this downturn is bad, though not significantly worse than previous downturns in 1958, 1975 and 1983.
However when you consider the correlation between the “credit impulse” and the change in employment, this crisis has no precedent in the post-WWII period:
Furthermore, debt is the leading factor is this process. Though the correlation between changes in real GDP and changes in employment are higher than those for the acceleration in debt and changes in employment, the “credit impulse” leads changes in employment while GDP slightly lags changes in employment: credit, which is ignored by conventional “neoclassical” economics, is in the driving seat.
This is something that Keynes realized after writing the General Theory (Keynes 1936), but which never made its way into the textbook version of Keynes that conventional economists like Stiglitz and Krugman learnt as Keynesianism.
Planned investment—i.e. investment ex-ante—may have to secure its “financial provision” before the investment takes place; that is to say, before the corresponding saving has taken place. This service may be provided either by the new issue market or by the banks ;—which it is, makes no difference… let us call this advance provision of cash the ‘finance’ required by the current decisions to invest. Investment finance in this sense is, of course, only a special case of the finance required by any productive process; but since it is subject to special fluctuations of its own, I should (I now think) have done well to have emphasised it when I analysed the various sources of the demand for money. (Keynes 1937, pp. 246–247)
The good news in the latest Flow of Funds data is therefore that a slowdown in the rate of deleveraging can impart a positive impetus to employment. However the bad news is that the economy is now hostage to changes in the rate of deleveraging, from levels of debt that far exceed anything it has ever experienced beforehand. Since much of this debt was taken on to finance speculation on asset prices rather than genuine investment, it is highly likely that deleveraging will accelerate in the future, as speculators tire—literally as well as metaphorically—of carrying large debt loads that finance stagnant or declining asset prices.
Drilling down into the debt data, it’s apparent that the sector that caused the crisis—the finance sector—is the one that has delevered the most is also the one whose rate of delevering is slowing most rapidly.
This is not a good thing, nor is it likely to last. The finance sector exists to create debt, and the only way it can do that is by encouraging the rest of the economy to take it on. If they were funding productive investments with this money, there wouldn’t be a crisis in the first place—and debt levels would be much lower, compared to GDP, than they are today. Instead they have enticed us into debt to speculate on rising asset prices, and the only way they can expand debt again is to re-ignite bubbles in the share and property markets once more.
Here’s where the level of debt (when compared to income) matters, as opposed to its rate of change: reigniting these bubbles is easy when debt to GDP levels are low. But reigniting them when debt to income levels are astronomical is next to impossible. Speculators have to be encouraged to take on a level of debt whose servicing consumes a dangerously high proportion of their income, in the belief that rising asset prices will let them repay that debt with a profit in the near future.
With the debt to GDP levels for all non-government sectors of the American economy at unprecedented levels, the prospect that any sector can be enticed to take on yet more debt is remote. Deleveraging is America’s future.
Biggs, M., T. Mayer, et al. “Credit and Economic Recovery: Demystifying Phoenix Miracles.” SSRN eLibrary.
Keynes, J. M. (1936). The general theory of employment, interest and money. London, Macmillan.
Keynes, J. M. (1937). “Alternative theories of the rate of interest.” Economic Journal
47: 241–252.
“[Would the equity transfer you describe effect external debt? Were their any associated cashflows?]”
hi jason,
its the size of the deficit and on who’s balance sheet it ends up that determines whether the non bank sector of the economy improves nett worth.
and yes, i think if you have a currency monopolist(the government) issues too little currency so there isnt enough to pay some of that currency back in taxes, and spend, whether it be for investment or consumption, the private sector will have to find those funds from somewhere else, and thats from the banking system, which may source some of its funding overseas, increasing the foreign debt.
so yes a deficit would be associated with rising private sector saving. the question is would this additional saving by the private sector eventually be leveraged through the private banking system.
so we may be damned if we do and damned if we dont
one things for certain, budget surpluses do not help the private sector. if you look at the data, the last two recessions or near recessions, have been preceded by a series of budget surpluses.
Stella D, his data comes from the Flow of Funds report, Table L.1, “Credit Market Debt Outstanding,” so he’s using levels, not flows. You must compute “private debt.” Here is one way to do it, using quarterly data back to the early 1950s from St. Louis Federal Reserve FRED time series database at the follow site
http://research.stlouisfed.org/fred2/
Line 1, Total credit market debt = Series TCMDO
Less:
Line 7, State and Local Governments = Series SLGTCMDODNS
Less:
Line 8, Federal Government = Series FGTCMDODNS
Less:
Line 9, Rest of The World = Series WCMITCMFODNS
Equals:
Private Debt
The GDP data is St Louis Fed FRED time series GDP
You can now compute the ratio, but be aware the credit market data is not seasonally adjusted. Also, the GDP data is seasonally adjusted, but not “real.”
Having answered your question, perhaps Professor Keen would answer mine:
Professor, how do you compute your private debt/GDP ratio “per cent per annum growth rate?” I’ve used most of the methods I know and have given up.
Thanks in advance for your reply.
Jason,
It’s getting boring now. I believe it is the ability of the Australian Government to convince its citizens and residents (including immigrants)to take on more and more debt which is the primary cause of our current economic buoyancy. I also believe that overseas factors that you allude have less of an economic impact than increasing local debt levels. You think it is the other way around. Let’s agree to disagree.
Also Brett @ 97 made a pertinent point.
“This is exactly what Steve believes. He believes that too much debt to GDP means that an economy will eventually start deleveraging – and enter a depression (or long term recession). He regurlary states that nothing can stop the deleveraging (including external influences or increasing government debt). He’s tempered his statements recently saying that these things can have a short term influence, but eventually deleveraging must resume. I think we need greater clarity on what the “eventually” time period is.”
Alan Gresley @ 105,
Mate, I’ll try and sum up your long post succinctly from what I believe.
1. There has and always will be a geo-political dance.
2. Sometimes people have an overrealiance on mathematics.
3. History does not tell us the probability of future financial things happening.
4. In relation to life generally, people waste too much time focusing on what is unimportant and/or unknowable.
5. I have made mistakes along my journey in life. Some of my mistakes however have set me up for substantially greater future benefit. However, I would rather learn the lessons from other people’s mistakes.
6. I have no idea what the USD and gold price will do in the future. Nor do I still have any idea when Steve’s forecasts will come true.
7. Whilst I can’t give too much away, I am a fund manager who looks after OPM as well as my own.
8. There were numerous red flags with Lehman’s business model well before it collapsed.
9. Rightly or wrongly, one of the most prized assets in today’s society is the ability to earn an income. The pursuit of knowledge alone does not always pay the bills.
Citydoc, it is obvious your only interest here is to poor crap on Steve for reasons only known to yourself.
If you are bored don’t bother popping by.
I think its entirely reasonable to suggest that prediction of future economic outcomes cannot be credible unless they incorporate the crucial geo-political, demographic and social mood factors.
Steve has made some very key contributions (for me at least) to undertsanding how credit dynamics affect future outcomes — what we need now to complete the picture is individuals of similar capability and far-sightedness in the realms of geo politics, demographics and social trend analysis.
In an age where money is in fact a kind of social credit to be expanded and contracted at (at leastin theory) our collective will, these latter factors are necessarily key.
Steve,
I think this is excellent work. I am still troubled by the lack of attention your analysis is getting in the US. For instance, I would like to see you and Paul Krugman engaging in a blog “dialogue”. He claims to hold Minsky in high regard and your work is in part an extension of Minsky’s.
Do you think your work touches on “taboos” that make it “outside the pale”. If so, what makes it extra-controversial for critical economists?
That being said, I think the idea that the financial sector exists to “create debt” is an oversimplification. It also exists for other purposes but apparently, at least in the endogenous theory, creating debt is an important step on the way to the creation of money. I believe you’re collapsing ontology into teleology in that statement.
Jason @ 110,
Please re-read message 108. Thanks.
Thanks for that explanation Conjure Bag.
I calculate the percent per annum growth rate very simply–it’s the change in debt over a year divided by GDP–but I am using two interpolated rolling time series to do so; I also have a second measure (the debt contribution to aggregate demand) which is somewhat different, and that might be the reason for some confusion.
Firstly you’re right that I’m working from the level tables to derive the flow, rather than working from the flow data itself–this is just for reasons of simplicity, since I’m doing all this without research assistance.
Secondly I interpolate monthly figures for both credit and GDP in America (Australia’s RBA actually publishes monthly credit figures, so I use raw level data in Australia but interpolated monthly level data in the USA).
Then I take the change in debt over 12 months, and divide that by the relevant monthly figure at each step.
Each data series is stored as a two column vector–the first containing the date (in decimal format), the second the data. Series to be worked on are aligned on the first column and then operations applied to the matching rows.
The actual operation that derives that data in Mathcad is expressed as:
ToPercent(Divide(Change.Period(USA.DebtPrivate,12),USA.GDP)).
If I can intercede in the Jason-CityDoc debate, CityDoc’s surmise that “it is the ability of the Australian Government to convince its citizens and residents (including immigrants)to take on more and more debt which is the primary cause of our current economic buoyancy. I also believe that overseas factors that you allude have less of an economic impact than increasing local debt levels.” is the one that I find is supported by the data.
When I write that “Where did I go wrong?” post, that’s one of the major points I’ll be making: empirically, we avoided a downturn by re-igniting private sector leveraging. Empirically speaking, the Chinese factor only became dominant in March of this year, well after the manifestation of the crisis had been avoided here.
I won’t have time for that post for quite some time–I have a ridiculous level of work right now till early December–but I will post it before the end of the year.
Thanks Michael,
The reason that Krugman and co don’t follow up on my work is that they don’t recognise anything outside the neoclassical school as economics. There’s no difficulty in them spotting works by me and others in the Post Keynesian tradition–the online database Econlit lists my works as agnostically as it does Krugman’s–but they would all be utterly unfamiliar to him (both in the method of analysis and the authors) and he wouldn’t bother reading them.
And yes there is a bit of an ontology-teleology overload there–though as you note it’s not possible for the financial sector to create money without also creating debt in a pure credit economy.
I will regard the matter as interceded!
Let me propose this as I take the gloves off and respectively walk to the pavilion
:
A term in the function of the ability to “convince” is the existence of preset perceptions in the mind of the “convincee” upon which to leverage stimulus provided by the “convincer”
That is not controversial.
For example [without epressing a view either way on the intiatives merits] with regards the recent resource tax the government punted that “Greedy billionaires have no right to all the cash” had plenty. The miners view was that “China’s desire for our Coal and Iron Ore is the source of our wealth and this will disturb that” was more potent.
It’s plain for all to see in that most recent national test of macro perceptions of the Australian community what was proven to hold true.
The second point I would make is that taking out a loan to buy a house is an efficient proxy for “debt growth” in the context of our consideration.
Let me reflect on personal experience. If I run through in my mind the people in my circle who have purchased a house during the period being considered there would be 9 people, and of those 9 there were 2 that had access to the FHOG. The major drivers I observed for all 9 [though to varying degrees] was:
1. A sense of urgency of opportunity related to perceptions around continued price increase, or a sense of no risk to future price growth.
2. A sense of urgency around perceptions of not having a house to live in full stop if they don’t secure a house now.
3. Assessment of particular areas being, and/or about to be with great certainty, majorly influenced by particular and general population demographics, and/or majorly influenced by particular current and medium term infrastructure projects.
The question becomes is 1 and 2 driven by the convincing of government or by the “they are jumping all over us to get a piece of us and this is going to roll on and on” message coming from elsewhere.
[In the 2 who took up the FHOG one was an afterthought (wasn’t aware that they were eligible) and the other was relevant only to the determination of the price limit set for their purchase (though I accept that leverage into total debt argument of the FHOG and certainly the FHOG must have fed into the urgency component on the national level to some extent)].
The third point I would make is that the exchange rate is a very rough broad proxy for perceptions of others about whether those same others want a piece of us into the future.
A look at the exchange rate 1 jan 99 to now:
http://au.finance.yahoo.com/echarts?s=AUDUSD=X#chart4:symbol=AUDUSD=X;range=19990101,20100920;compare=;indicator=volume;charttype=area;crosshair=on;ohlcvalues=0;logscale=off;source=
shows that those perceptions were established strongly in the build up to the GFC, strong enough to bridge the crevasse of July 2008 to Nov 2009.
In conclusion, the government has tested it’s ability to convince in the face of “China, Coal, and Iron Ore” recently and failed. Are we so sure it’s convincing power in the face of that was different with this issue?
One can only hope for a general collapse in wages and prices.
Otherwise, it’s the great plunge into the abyss.
There are interesting things coming out of the Bank of England currently:
http://bit.ly/d7AC0Q
http://t.co/Y49DSc7
But also note:
http://www.bbc.co.uk/news/business-11418652
It would be great if we could find a formula to make the private sector DELEVER and DELIVER at the same time. If the proposals presented here can be implemented in practice this may actually happen. More austerity leading to even greater leveraging will have the opposite effect.