Dude! Where’s My Recov­ery?

Flattr this!

I ini­tially planned to call this post “Eco­nomic Growth, Asset Mar­kets and the Credit Accel­er­a­tor”, but recent neg­a­tive data out of Amer­ica makes me think that this title is more in line with con­ver­sa­tions cur­rently tak­ing place in the White House.

(Click here for this post in PDF)

Accord­ing to the NBER, the “Great Reces­sion” is now two years behind us, but the recov­ery that nor­mally fol­lows a reces­sion has not occurred. While growth did rise for a while, it has been anaemic com­pared to the norm after a reces­sion, and it is already trend­ing down. Growth needs to exceed 3 per cent per annum to reduce unemployment—the rule of thumb known as Okun’s Law—and it needs to be sub­stan­tially higher than this to make seri­ous inroads into it. Instead, growth barely peeped its head above Okun’s level. It is now below it again, and trend­ing down.

Fig­ure 1

Unem­ploy­ment is there­fore ris­ing once more, and with it, Obama’s chances of re-elec­tion are rapidly fad­ing.

Fig­ure 2

Obama was assured by his advi­sors that this wouldn’t hap­pen. Right from the first Eco­nomic Report of the Pres­i­dent that he received from Bush’s out­go­ing Chair­man of the Coun­cil of Eco­nomic Advis­ers Ed Lazear in Jan­u­ary 2009, he was assured that “the deeper the down­turn, the stronger the recov­ery”. On the basis of the regres­sion shown in Chart 1–9 of that report (on page 54), I am sure that Obama was told that real growth would prob­a­bly exceed 5 per cent per annum—because this is what Ed Lazear told me after my ses­sion at the Aus­tralian Con­fer­ence of Econ­o­mists in Sep­tem­ber 2009.

Fig­ure 3

I dis­puted this analy­sis then (see “In the Dark on Cause and Effect, Debt­watch Octo­ber 2009”), and events have cer­tainly borne out my analy­sis rather than the con­ven­tional wis­dom. To give an idea of how wrong this guid­ance was, the peak to trough decline in the Great Recession—the x-axis in Lazear’s Chart—was over 6 per­cent. His regres­sion equa­tion there­fore pre­dicted that GDP growth in the 2 years after the reces­sion ended would have been over 12 per­cent. If this equa­tion had born fruit, US Real GDP would be $14.37 tril­lion in June 2011, ver­sus the recorded $13.44 tril­lion in March 2011.

Fig­ure 4

So why has the con­ven­tional wis­dom been so wrong? Largely because it has ignored the role of pri­vate debt—which brings me back to my orig­i­nal title.


Economic Growth, Asset Markets and the Credit Accelerator


Neo­clas­si­cal econ­o­mists ignore the level of pri­vate debt, on the basis of the a pri­ori argu­ment that “one man’s lia­bil­ity is another man’s asset”, so that the aggre­gate level of debt has no macro­eco­nomic impact. They rea­son that the increase in the debtor’s spend­ing power is off­set by the fall in the lender’s spend­ing power, and there is there­fore no change to aggre­gate demand.

Lest it be said that I’m par­o­dy­ing neo­clas­si­cal eco­nom­ics, or rely­ing on what lesser lights believe when the lead­ers of the pro­fes­sion know bet­ter, here are two appo­site quotes from Ben Bernanke and Paul Krug­man.

Bernanke in his Essays on the Great Depres­sion, explain­ing why neo­clas­si­cal econ­o­mists didn’t take Fisher’s Debt Defla­tion The­ory of Great Depres­sions (Irv­ing Fisher, 1933) seri­ously:

Fisher’s idea was less influ­en­tial in aca­d­e­mic cir­cles, though, because of the coun­ter­ar­gu­ment that debt-defla­tion rep­re­sented no more than a redis­tri­b­u­tion from one group (debtors) to another (cred­i­tors). Absent implau­si­bly large dif­fer­ences in mar­ginal spend­ing propen­si­ties among the groups, it was sug­gested, pure redis­tri­b­u­tions should have no sig­nif­i­cant macro-eco­nomic effects… (Ben S. Bernanke, 2000, p. 24)

Krug­man in his most recent draft aca­d­e­mic paper on the cri­sis:

Given both the promi­nence of debt in pop­u­lar dis­cus­sion of our cur­rent eco­nomic dif­fi­cul­ties and the long tra­di­tion of invok­ing debt as a key fac­tor in major eco­nomic con­trac­tions, one might have expected debt to be at the heart of most main­stream macro­eco­nomic models—especially the analy­sis of mon­e­tary and fis­cal pol­icy. Per­haps some­what sur­pris­ingly, how­ever, it is quite com­mon to abstract alto­gether from this fea­ture of the econ­omy. Even econ­o­mists try­ing to ana­lyze the prob­lems of mon­e­tary and fis­cal pol­icy at the zero lower bound—and yes, that includes the authors—have often adopted rep­re­sen­ta­tive-agent mod­els in which every­one is alike, and in which the shock that pushes the econ­omy into a sit­u­a­tion in which even a zero inter­est rate isn’t low enough takes the form of a shift in everyone’s pref­er­ences…

Ignor­ing the for­eign com­po­nent, or look­ing at the world as a whole, the over­all level of debt makes no dif­fer­ence to aggre­gate net worth — one person’s lia­bil­ity is another person’s asset. (Paul Krug­man and Gauti B. Eggerts­son, 2010, pp. 2–3; empha­sis added)

They are pro­foundly wrong on this point because neo­clas­si­cal econ­o­mists do not under­stand how money is cre­ated by the pri­vate bank­ing system—despite decades of empir­i­cal research to the con­trary, they con­tinue to cling to the text­book vision of banks as mere inter­me­di­aries between savers and bor­row­ers.

This is bizarre, since as long as 4 decades ago, the actual sit­u­a­tion was put very sim­ply by the then Senior Vice Pres­i­dent, Fed­eral Reserve Bank of New York, Alan Holmes. Holmes explained why the then fad­dish Mon­e­tarist pol­icy of con­trol­ling infla­tion by con­trol­ling the growth of Base Money had failed, say­ing that it suf­fered from “a naive assump­tion” that:

the bank­ing sys­tem only expands loans after the [Fed­eral Reserve] Sys­tem (or mar­ket fac­tors) have put reserves in the bank­ing sys­tem. In the real world, banks extend credit, cre­at­ing deposits in the process, and look for the reserves later. The ques­tion then becomes one of whether and how the Fed­eral Reserve will accom­mo­date the demand for reserves. In the very short run, the Fed­eral Reserve has lit­tle or no choice about accom­mo­dat­ing that demand; over time, its influ­ence can obvi­ously be felt. (Alan R. Holmes, 1969, p. 73; empha­sis added)

The empir­i­cal fact that “loans cre­ate deposits” means that the change in the level of pri­vate debt is matched by a change in the level of money, which boosts aggre­gate demand. The level of pri­vate debt there­fore can­not be ignored—and the fact that neo­clas­si­cal econ­o­mists did ignore it (and, with the likes of Greenspan run­ning the Fed, actively pro­moted its growth) is why this is no “gar­den vari­ety” down­turn.

In all the post-WWII reces­sions on which Lazear’s regres­sion was based, the down­turn ended when the growth of pri­vate debt turned pos­i­tive again and boosted aggre­gate demand. This of itself is not a bad thing: as Schum­peter argued decades ago, in a well-func­tion­ing cap­i­tal­ist sys­tem, the main recip­i­ents of credit are entre­pre­neurs who have an idea, but not the money needed to put it into action:

[I]n so far as credit can­not be given out of the results of past enter­prise … it can only con­sist of credit means of pay­ment cre­ated ad hoc, which can be backed nei­ther by money in the strict sense nor by prod­ucts already in exis­tence…

It pro­vides us with the con­nec­tion between lend­ing and credit means of pay­ment, and leads us to what I regard as the nature of the credit phe­nom­e­non… credit is essen­tially the cre­ation of pur­chas­ing power for the pur­pose of trans­fer­ring it to the entre­pre­neur, but not sim­ply the trans­fer of exist­ing pur­chas­ing power.” (Joseph Alois Schum­peter, 1934, pp. 106–107)

It becomes a bad thing when this addi­tional credit goes, not to entre­pre­neurs, but to Ponzi mer­chants in the finance sec­tor, who use it not to inno­vate or add to pro­duc­tive capac­ity, but to gam­ble on asset prices. This adds to debt lev­els with­out adding to the economy’s capac­ity to ser­vice them, lead­ing to a blowout in the ratio of pri­vate debt to GDP. Ulti­mately, this process leads to a cri­sis like the one we are now in, where so much debt has been taken on that the growth of debt comes to an end. The econ­omy then enters not a reces­sion, but a Depres­sion.

For a while though, it looked like a recov­ery was afoot: growth did rebound from the depths of the Great Reces­sion, and very quickly com­pared to the Great Depres­sion (though slowly when com­pared to Post-WWII reces­sions).

Clearly the scale of gov­ern­ment spend­ing, and the enor­mous increase in Base Money by Bernanke, had some impact—but nowhere near as much as they were hop­ing for. How­ever the main fac­tor that caused the brief recovery—and will also cause the dreaded “dou­ble dip”—is the Credit Accel­er­a­tor.

I’ve pre­vi­ously called this the “Credit Impulse” (using the name bestowed by Michael Biggs et al., 2010), but I think “Credit Accel­er­a­tor” is both move evoca­tive and more accu­rate. The Credit Accel­er­a­tor at any point in time is the change in the change in debt over pre­vi­ous year, divided by the GDP fig­ure for that point in time. From first prin­ci­ples, here is why it mat­ters.

Firstly, and con­trary to the neo­clas­si­cal model, a cap­i­tal­ist econ­omy is char­ac­ter­ized by excess sup­ply at vir­tu­ally all times: there is nor­mally excess labor and excess pro­duc­tive capac­ity, even dur­ing booms. This is not per se a bad thing but merely an inher­ent char­ac­ter­is­tic of capitalism—and it is one of the rea­sons that cap­i­tal­ist economies gen­er­ate a much higher rate of inno­va­tion than did social­ist economies (Janos Kor­nai, 1980). The main con­straint fac­ing cap­i­tal­ist economies is there­fore not sup­ply, but demand.

Sec­ondly, all demand is mon­e­tary, and there are two sources of money: incomes, and the change in debt. The sec­ond fac­tor is ignored by neo­clas­si­cal eco­nom­ics, but is vital to under­stand­ing a cap­i­tal­ist econ­omy. Aggre­gate demand is there­fore equal to Aggre­gate Sup­ply plus the change in debt.

Thirdly, this Aggre­gate Demand is expended not merely on new goods and ser­vices, but also on net sales of exist­ing assets. Wal­ras’ Law, that main­stay of neo­clas­si­cal eco­nom­ics, is thus false in a credit-based economy—which hap­pens to be the type of econ­omy in which we live. Its replace­ment is the fol­low­ing expres­sion, where the left hand is mon­e­tary demand and the right hand is the mon­e­tary value of pro­duc­tion and asset sales:

Income + Change in Debt = Out­put + Net Asset Sales;

In sym­bols (where I’m using an arrow to indi­cate the direc­tion of cau­sa­tion rather than an equals sign), this is:


This means that it is impos­si­ble to sep­a­rate the study of “Finance”—largely, the behav­iour of asset markets—from the study of macro­eco­nom­ics. Income and new credit are expended on both newly pro­duced goods and ser­vices, and the two are as entwined as a scram­bled egg.

Net Asset Sales can be bro­ken down into three com­po­nents:

  • The asset price Level; times
  • The frac­tion of assets sold; times
  • The quan­tity of assets

Putting this in sym­bols:


That cov­ers the lev­els of aggre­gate demand, aggre­gate sup­ply and net asset sales. To con­sider eco­nomic growth—and asset price change—we have to look at the rate of change. That leads to the expres­sion:


There­fore the rate of change of asset prices is related to the accel­er­a­tion of debt. It’s not the only fac­tor obviously—change in incomes is also a fac­tor, and as Schum­peter argued, there will be a link between accel­er­at­ing debt and ris­ing income if that debt is used to finance entre­pre­neur­ial activ­ity. Our great mis­for­tune is that accel­er­at­ing debt hasn’t been pri­mar­ily used for that pur­pose, but has instead financed asset price bub­bles.

There isn’t a one-to-one link between accel­er­at­ing debt and asset price rises: some of the bor­rowed money dri­ves up pro­duc­tion (think SUVs dur­ing the boom), con­sumer prices, the frac­tion of exist­ing assets sold, and the pro­duc­tion of new assets (think McMan­sions dur­ing the boom). But the more the econ­omy becomes a dis­guised Ponzi Scheme, the more the accel­er­a­tion of debt turns up in ris­ing asset prices.

As Schumpeter’s analy­sis shows, accel­er­at­ing debt should lead change in out­put in a well-func­tion­ing econ­omy; we unfor­tu­nately live in a Ponzi econ­omy where accel­er­at­ing debt leads to asset price bub­bles.

In a well-func­tion­ing econ­omy, peri­ods of accel­er­a­tion of debt would be fol­lowed by peri­ods of decel­er­a­tion, so that the ratio of debt to GDP cycled but did not rise over time. In a Ponzi econ­omy, the accel­er­a­tion of debt remains pos­i­tive most of the time, lead­ing not merely to cycles in the debt to GDP ratio, but a sec­u­lar trend towards ris­ing debt. When that trend exhausts itself, a Depres­sion ensues—which is where we are now. Delever­ag­ing replaces ris­ing debt, the debt to GDP ratio falls, and debt starts to reduce aggre­gate demand rather than increase it as hap­pens dur­ing a boom.

Even in that sit­u­a­tion, how­ever, the accel­er­a­tion of debt can still give the econ­omy a tem­po­rary boost—as Biggs, Meyer and Pick pointed out. A slow­down in the rate of decline of debt means that debt is accel­er­at­ing: there­fore even when aggre­gate pri­vate debt is falling—as it has since 2009—a slow­down in that rate of decline can give the econ­omy a boost.

That’s the major fac­tor that gen­er­ated the appar­ent recov­ery from the Great Reces­sion: a slow­down in the rate of decline of pri­vate debt gave the econ­omy a tem­po­rary boost. The same force caused the appar­ent boom of the Great Mod­er­a­tion: it wasn’t “improved mon­e­tary pol­icy” that caused the Great Mod­er­a­tion, as Bernanke once argued (Ben S. Bernanke, 2004), but bad mon­e­tary pol­icy that wrongly ignored the impact of ris­ing pri­vate debt upon the econ­omy.

Fig­ure 5

The fac­tor that makes the recent recov­ery phase dif­fer­ent to all pre­vi­ous ones—save the Great Depres­sion itself—is that this strong boost from the Credit Accel­er­a­tor has occurred while the change in pri­vate debt is still mas­sively neg­a­tive. I return to this point later when con­sid­er­ing why the recov­ery is now peter­ing out.

The last 20 years of eco­nomic data shows the impact that the Credit Accel­er­a­tor has on the econ­omy. The recent recov­ery in unem­ploy­ment was largely caused by the dra­matic rever­sal of the Credit Accelerator—from strongly neg­a­tive to strongly positive—since late 2009:

Fig­ure 6

The Credit Accel­er­a­tor also caused the tem­po­rary recov­ery in house prices:

Fig­ure 7

And it was the pri­mary fac­tor dri­ving the Bear Mar­ket rally in the stock mar­ket:

Fig­ure 8


Leads and Lags


I use the change in the change in debt over a year because the monthly and quar­terly data is sim­ply too volatile; the annual change data smooths out much of the noise. Con­se­quently the data shown for change in unem­ploy­ment, house prices and the stock mar­ket are also for the change the pre­vi­ous year.

How­ever the change in the change in debt oper­ates can impact rapidly on some markets—notably the Stock Mar­ket. So though the cor­re­la­tions in the above graphs are already high, they are higher still when we con­sider the causal role of the debt accel­er­a­tor in chang­ing the level of aggre­gate demand by lag­ging the data.

This shows that the annual Credit Accel­er­a­tor leads annual changes in unem­ploy­ment by roughly 5 months, and its max­i­mum cor­re­la­tion is a stag­ger­ing –0.85 (neg­a­tive because an accel­er­a­tion in debt causes a fall in unem­ploy­ment by boost­ing aggre­gate demand, while a decel­er­a­tion in debt causes a rise in unem­ploy­ment by reduc­ing aggre­gate demand).

Fig­ure 9

The cor­re­la­tion between the annual Credit Accel­er­a­tor and annual change in real house prices peaks at about 0.7 roughly 9 months ahead:

Fig­ure 10

And the Stock Mar­ket is also a crea­ture of the Credit Impulse, where the lead is about 10 months and the cor­re­la­tion peaks at just under 0.6:

Fig­ure 11

The causal rela­tion­ship between the accel­er­a­tion of debt and change in stock prices is more obvi­ous when the 10 month lag is taken into account:

Fig­ure 12

These cor­re­la­tions, which con­firm the causal argu­ment made between the accel­er­a­tion of debt and the change in asset prices, expose the dan­ger­ous pos­i­tive feed­back loop in which the econ­omy has been trapped. This is sim­i­lar to what George Soros calls a reflex­ive process: we bor­row money to gam­ble on ris­ing asset prices, and the accel­er­a­tion of debt causes asset prices to rise.

This is the basis of a Ponzi Scheme, and it is also why the Scheme must even­tu­ally fail. Because it relies not merely on grow­ing debt, but accel­er­at­ing debt, ulti­mately that accel­er­a­tion must end—because oth­er­wise debt would become infi­nite. When the accel­er­a­tion of debt ceases, asset prices col­lapse.

The annual Credit Accel­er­a­tor is still very strong right now—so why is unem­ploy­ment ris­ing and both hous­ing and stocks falling? Here we have to look at the more recent quar­terly changes in the Credit Accelerator—even though there is too much noise in the data to use it as a decent indi­ca­tor (the quar­terly lev­els show in Fig­ure 13 are from month to month—so that the bar for March 2011 indi­cates the accel­er­a­tion of debt between Jan­u­ary and March 2011). It’s appar­ent that the strong accel­er­a­tion of debt in mid to late 2010 is peter­ing out. Another quar­ter of that low a rate of accel­er­a­tion in debt—or a return to more deceleration—will drive the annual Credit Accel­er­a­tor down or even neg­a­tive again. The lead between the annual Credit Accel­er­a­tor and the annu­al­ized rates of change of unem­ploy­ment and asset prices means that this dimin­ished stim­u­lus from accel­er­at­ing debt is turn­ing up in the data now.

Fig­ure 13

This ten­dency for the Credit Accel­er­a­tor to turn neg­a­tive after a brief bout of being pos­i­tive is likely to be with us for some time. In a well-func­tion­ing econ­omy, the Credit Accel­er­a­tor would fluc­tu­ate around slightly above zero. It would be above zero when a Schum­peter­ian boom was in progress, below dur­ing a slump, and tend to exceed zero slightly over time because pos­i­tive credit growth is needed to sus­tain eco­nomic growth. This would result in a pri­vate debt to GDP level that fluc­tu­ated around a pos­i­tive level, as out­put grew cycli­cally in pro­por­tion to the ris­ing debt.

Instead, it has been kept pos­i­tive over an unprece­dented period by a Ponzi-ori­ented finan­cial sec­tor, which was allowed to get away with it by naïve neo­clas­si­cal econ­o­mists in posi­tions of author­ity. The con­se­quence was a sec­u­lar ten­dency for the debt to GDP ratio to rise. This was the dan­ger Min­sky tried to raise aware­ness of in his Finan­cial Insta­bil­ity Hypoth­e­sis (Hyman P. Min­sky, 1972)—which neo­clas­si­cal econ­o­mists like Bernanke ignored.

The false pros­per­ity this accel­er­at­ing debt caused led to the fan­tasy of “The Great Mod­er­a­tion” tak­ing hold amongst neo­clas­si­cal econ­o­mists. Ulti­mately, in 2008, this fan­tasy came crash­ing down when the impos­si­bil­ity of main­tain­ing a pos­i­tive accel­er­a­tion in debt for­ever hit—and the Great Reces­sion began.

Fig­ure 14

From now on, unless we do the sen­si­ble thing of abol­ish­ing debt that should never have been cre­ated in the first place, we are likely to be sub­ject to wild gyra­tions in the Credit Accel­er­a­tor, and a gen­eral ten­dency for it to be neg­a­tive rather than pos­i­tive. With debt still at lev­els that dwarf pre­vi­ous spec­u­la­tive peaks, the pos­i­tive feed­back between accel­er­at­ing debt and ris­ing asset prices can only last for a short time, since it if were to per­sist, debt lev­els would ulti­mately have to rise once more. Instead, what is likely to hap­pen is a a period of strong accel­er­a­tion in debt (caused by a slow­down in the rate of decline of debt) and ris­ing asset prices—followed by a decline in the accel­er­a­tion as the veloc­ity of debt approaches zero.

Fig­ure 15

Here Soros’s reflex­iv­ity starts to work in reverse. With the Credit Accel­er­a­tor going into reverse, asset prices plunge—which fur­ther reduces the public’s will­ing­ness to take on debt, which causes asset prices to fall even fur­ther.

The process even­tu­ally exhausts itself as the debt to GDP ratio falls. But given that the cur­rent pri­vate debt level is per­haps 170% of GDP above where it should be (the level that finances entre­pre­neur­ial invest­ment rather than Ponzi Schemes), the end game here will be many years in the future. The only sure road to recov­ery is debt abolition—but that will require defeat­ing the polit­i­cal power of the finance sec­tor, and end­ing the influ­ence of neo­clas­si­cal econ­o­mists on eco­nomic pol­icy. That day is still a long way off.

Fig­ure 16

Bernanke, Ben S. 2000. Essays on the Great Depres­sion. Prince­ton: Prince­ton Uni­ver­sity Press.

____. 2004. “The Great Mod­er­a­tion: Remarks by Gov­er­nor Ben S. Bernanke at the Meet­ings of the East­ern Eco­nomic Asso­ci­a­tion, Wash­ing­ton, Dc Feb­ru­ary 20, 2004,” East­ern Eco­nomic Asso­ci­a­tion. Wash­ing­ton, DC: Fed­eral Reserve Board,

Biggs, Michael; Thomas Mayer and Andreas Pick. 2010. “Credit and Eco­nomic Recov­ery: Demys­ti­fy­ing Phoenix Mir­a­cles.” SSRN eLi­brary.

Fisher, Irv­ing. 1933. “The Debt-Defla­tion The­ory of Great Depres­sions.” Econo­met­rica, 1(4), 337–57.

Holmes, Alan R. 1969. “Oper­a­tional Con­straints on the Sta­bi­liza­tion of Money Sup­ply Growth,” F. E. Mor­ris, Con­trol­ling Mon­e­tary Aggre­gates. Nan­tucket Island: The Fed­eral Reserve Bank of Boston, 65–77.

Kor­nai, Janos. 1980. “‘Hard’ and ‘Soft’ Bud­get Con­straint.” Acta Oeco­nom­ica, 25(3–4), 231–45.

Krug­man, Paul and Gauti B. Eggerts­son. 2010. “Debt, Delever­ag­ing, and the Liq­uid­ity Trap: A Fisher-Min­sky-Koo Approach [2nd Draft 2/14/2011],” New York: Fed­eral Reserve Bank of New York & Prince­ton Uni­ver­sity,

Min­sky, Hyman P. 1972. “Finan­cial Insta­bil­ity Revis­ited: The Eco­nom­ics of Dis­as­ter,” Board of Gov­er­nors of the Fed­eral Reserve Sys­tem, Reap­praisal of the Fed­eral Reserve Dis­count Mech­a­nism. Wash­ing­ton, D.C.: Board of Gov­er­nors of the Fed­eral Reserve Sys­tem,

Schum­peter, Joseph Alois. 1934. The The­ory of Eco­nomic Devel­op­ment : An Inquiry into Prof­its, Cap­i­tal, Credit, Inter­est and the Busi­ness Cycle. Cam­bridge, Mass­a­chu­setts: Har­vard Uni­ver­sity Press.



About Steve Keen

I am Professor of Economics and Head of Economics, History and Politics at Kingston University London, and a long time critic of conventional economic thought. As well as attacking mainstream thought in Debunking Economics, I am also developing an alternative dynamic approach to economic modelling. The key issue I am tackling here is the prospect for a debt-deflation on the back of the enormous private debts accumulated globally, and our very low rate of inflation.
Bookmark the permalink.