How I learnt to stop worrying and love The Bank

• Tweet

• Tweet

Click here for this post in PDF

Hav­ing just read the World Eco­nom­ic Forum’s Report on sus­tain­able cred­it, I now realise that I was wrong to wor­ry about the growth in debt. After all, since 1932, the US’s debt to GDP ratio has actu­al­ly fall­en at a rate of 0.2% per year!

How could I ever have thought that the growth of cred­it could have caused the Great Reces­sion, when in fact the growth rate of debt has been neg­a­tive?

I am also chas­tened to realise that cred­it is only used for good pur­pos­es. As the report notes:

In the long run, the scale and dis­tri­b­u­tion of cred­it is only eco­nom­i­cal­ly sus­tain­able if it also meets soci­ety’s broad­er social objec­tives. Cred­it is linked to social objec­tives dur­ing all stages of a coun­try’s eco­nom­ic devel­op­ment. In ear­ly stages of devel­op­ment, cred­it is used to sup­port fam­i­ly-owned busi­ness­es; next, it sup­ports small and large cor­po­ra­tions; and final­ly it is used to smooth con­sump­tion. Muham­mad Yunus, founder of Grameen Bank, goes so far as to say that cred­it is a human right, and adds: “If we are look­ing for one sin­gle action which will enable the poor to over­come their pover­ty, I would focus on cred­it.” (p. 39)

Fool­ish me: here was I, think­ing that cred­it might also be used to fund Ponzi Schemes.

OK, enough with the irony. The WEF’s report is not all bad—there are some very good bits that I’ll get on to later—but it com­mits at least three fun­da­men­tal errors: it uses a ques­tion­able base year for its analy­sis, it omits a cru­cial vari­able, and it main­tains a whol­ly benign view of a fac­tor that expe­ri­ence indi­cates has both benign and malig­nant attrib­ut­es.

A Questionable Base Year: why 2000?

The ques­tion­able base year is 2000. The WEF team, work­ing with McK­in­sey & Com­pa­ny, have put togeth­er an impres­sive data­base on debt lev­els in 79 coun­tries, with debt dis­ag­gre­gat­ed into 3 sec­tors:

• Retail cred­it: All house­hold cred­it stock (loans out­stand­ing), includ­ing mort­gages and oth­er per­son­al loans such as cred­it cards, auto loans and oth­er unse­cured loans
• Whole­sale cred­it: All cor­po­rate and SME cred­it stock (loans and bonds out­stand­ing)
• Gov­ern­ment cred­it: All pub­lic sec­tor cred­it stock (includ­ing loans and bonds out­stand­ing) (p. 19)

But to start in 2000? That was just before the last big cred­it bust, when the Dot­Com fias­co came crash­ing down. Why not—at least for the coun­tries where debt data is read­i­ly available—go back a bit fur­ther? Debt data for the USA is read­i­ly avail­able to 1952 from the Flow of Funds, and his­tor­i­cal data for ear­li­er years can be derived from the US cen­sus. Aus­trali­a’s Reserve Bank pub­lish­es reli­able aggre­gate cred­it data till 1976, and ear­li­er data is avail­able to take it back to 1953. As a sole indi­vid­ual, I’ve been able to acquire data till 1920 for the USA and 1860 for Aus­tralia. Sure­ly the WEF and McK­in­sey and Co, with the resources they had to throw at this project, could have done bet­ter than 2000.

Why omit financial sector debt?

The report omits bor­row­ing by with­in the finan­cial sec­tor from its record of total debt, when this has been a major com­po­nent of the growth of debt (cer­tain­ly in the USA) in the last 60 years. I include finan­cial sec­tor debt in my analy­sis for two rea­sons:

• The ini­tial bor­row­ing by the shad­ow bank­ing sec­tor from the banks cre­ates both mon­ey and debt;
• The mon­ey onlent by the shad­ow bank­ing sec­tor to oth­er sec­tors of the econ­o­my cre­ates debt to the shad­ow bank­ing sec­tor, but not mon­ey

I fre­quent­ly get the argu­ment that debt with­in the finan­cial sec­tor can be net­ted out to zero, but I think this ignores those two fac­tors above: the cre­ation of addi­tion­al debt-backed mon­ey by the ini­tial loan, and the cre­ation of fur­ther debt to the finan­cial sector—most of which has been used to fund asset bub­bles rather than pro­duc­tive invest­ment.

Just for com­par­i­son, here’s that same 1932 to 2010 com­par­i­son, but with the finan­cial sec­tor’s debt includ­ed:

Even with the same non­sense base year, and even with com­bin­ing pri­vate and pub­lic debt—factors that I believe should be kept separate—this paints a some­what dif­fer­ent pic­ture.

And a focus on total pri­vate sec­tor debt dur­ing and after the Great Depres­sion also con­veys a some­what dif­fer­ent per­spec­tive.

Which rais­es the third issue…

Why ignore Ponzi Schemes—and Minsky?

The report’s list­ing of the uses to which cred­it is put is so inno­cent as to make me won­der whether one of the author’s pri­ma­ry school chil­dren wrote the rel­e­vant para­graph:

In ear­ly stages of devel­op­ment, cred­it is used to sup­port fam­i­ly-owned busi­ness­es; next, it sup­ports small and large cor­po­ra­tions; and final­ly it is used to smooth con­sump­tion.

But maybe I’m being harsh: it could, after all, have been writ­ten by a neo­clas­si­cal econ­o­mist.

Please, let’s get real: yes cred­it can do all of those things, but it can also fund asset bub­bles and Ponzi Schemes, and that has been by far the dom­i­nant aspect of cred­it growth since the report’s base year of 2000, and arguably since the 1987 Stock Mar­ket Crash. To ignore this aspect of cred­it after the biggest finan­cial cri­sis since the Great Depres­sion is sim­ply puerile.

So is ignor­ing the one aca­d­e­m­ic who analysed the dynam­ics of cred­it long before it was fashionable—Hyman Min­sky. The report makes much of it aca­d­e­m­ic research:

Final­ly, the research was under­pinned by an exten­sive review of the key aca­d­e­m­ic and indus­try lit­er­a­ture. (p. 19)

Pri­or to 2008, such igno­rance was excus­able sim­ply because it was so wide­spread, as the dom­i­nant neo­clas­si­cal school sim­ply ignored dis­si­dents like Minksy. After the cri­sis, he is receiv­ing long over­due respect for focus­ing on the impor­tance of cred­it in a cap­i­tal­ist econ­o­my while neo­clas­si­cal econ­o­mists effec­tive­ly ignored it.

This is why Min­sky-ori­ent­ed researchers like myself, Michael Hud­son and the late Wynne God­ley were able to see the Great Reces­sion com­ing while neo­clas­si­cal econ­o­mists from Ben Bernanke down were deny­ing that any­thing unto­ward was unto­ward. Pub­lish­ing a major report on cred­it now, while ignor­ing the only sig­nif­i­cant research done into cred­it dynam­ics, is a sign of con­tin­ued igno­rance rather than wis­dom.

Its over­all conclusion—the only part of the report that is like­ly to get an air­ing in the gen­er­al media—should there­fore be tak­en with a truck­load of salt:

The rapid expan­sion of cred­it in recent decades has enabled unprece­dent­ed lev­els of eco­nom­ic devel­op­ment, busi­ness activ­i­ty, home own­er­ship and pub­lic sec­tor spend­ing. Yet, excess lend­ing in some mar­kets and sec­tors sparked a glob­al cri­sis which brought the entire finan­cial sys­tem to its knees. Not sur­pris­ing­ly, many com­men­ta­tors believe cred­it should be scaled back, even at the expense of eco­nom­ic growth.

The analy­sis in this report sug­gests the oppo­site is true. There are major pock­ets of the world econ­o­my, par­tic­u­lar­ly in devel­op­ing mar­kets, whose growth has been held back by cred­it short­ages, even over the past 10 years. To unlock devel­op­ment in these areas, and to meet con­sen­sus fore­casts of world eco­nom­ic growth, cred­it lev­els must grow sub­stan­tial­ly over the next decade. At the same time, pub­lic and pri­vate deci­sion-mak­ers must avoid a repeat of the cred­it excess­es that have caused so much dam­age in recent years. (p. 19)

Like a Curate’s Egg

How­ev­er, despite its defi­cien­cies, the report is not all bad—but its good bits are too con­ser­v­a­tive. It pro­pos­es a num­ber of “rules of thumb” to indi­cate whether cred­it lev­els and cred­it growth are sus­tain­able or not:

These are relat­ed to the mea­sures that I have been putting for­ward on this site for years—the debt to GDP ratio, the rate of change of debt, and recent­ly the “Cred­it Impulse” as defined by Big­gs, May­er and Pick (see also this paper), the rate of change of the rate of change of debt, divid­ed by GDP. I’ll come back to these lat­er as alter­nate indi­ca­tors, but it’s worth not­ing the guid­ance these rules of thumb gave for where cred­it crises might occur in the near future. Their Exhib­it 15 showed a “local sus­tain­abil­i­ty” index for their sam­ple of coun­tries in 2006, with the sam­ple sort­ed by the aggre­gate lev­el of debt (hence Japan is at the top). The guide is laid out accord­ing to this leg­end:

And the coun­tries iden­ti­fied using 2006 data as poten­tial trou­ble spots were those with one or more red cells next to their names:

The same analy­sis on 2010 data yields the fol­low­ing table of sus­pects:

The rules of thumb them­selves are pret­ty good. The weak­ness­es with their analy­sis are (a) that they allow the thumbs to be much too large, and (b) that they under­play the role of the bank­ing sec­tor in caus­ing these prob­lems in the first place.

As reg­u­lar read­ers would know, I site respon­si­bil­i­ty for this cri­sis on the lenders them­selves, and not the bor­row­ers, on a num­ber of grounds (see my Rov­ing Cav­a­liers of Cred­it post if you haven’t seen these argu­ments before, and this paper for a more tech­ni­cal argu­ment). The finan­cial sec­tor makes mon­ey by cre­at­ing debt, and has fund­ed a series of spec­u­la­tive bub­bles since the ear­ly 1980s since that is the best way to encour­age bor­row­ers to take on more debt.

Min­sky him­self argued that the major objec­tive of eco­nom­ic man­age­ment should be to main­tain a “robust finan­cial struc­ture”:

, in order to do bet­ter than hith­er­to, we have to estab­lish and enforce a “good finan­cial soci­ety” in which the ten­den­cy by busi­ness and bankers to engage in spec­u­la­tive finance is con­strained.

The finan­cial insta­bil­i­ty hypoth­e­sis has pol­i­cy impli­ca­tions that go beyond the sim­ple rules for mon­e­tary and fis­cal pol­i­cy that are derived from the neo-clas­si­cal syn­the­sis. In par­tic­u­lar the hypoth­e­sis leads to the con­clu­sion that the main­te­nance of a robust finan­cial struc­ture is a pre­con­di­tion for effec­tive anti-infla­tion and full employ­ment poli­cies with­out a need to haz­ard deep depres­sions. This implies that poli­cies to con­trol and guide the evo­lu­tion of finance are nec­es­sary… (Min­sky 1982, pp. 69, 112)

He assert­ed that the US passed from such a struc­ture to a frag­ile one with the Penn State cri­sis in 1966.

The first twen­ty years after World War II were char­ac­ter­ized by finan­cial tran­quil­i­ty. No seri­ous threat of a finan­cial cri­sis or a debt-defla­tion process (such as Irv­ing Fish­er described15) took place. The decade since 1966 has been char­ac­ter­ized by finan­cial tur­moil. Three threats of finan­cial cri­sis occurred, dur­ing which Fed­er­al Rserve inter­ven­tions in mon­ey and finan­cial mar­kets were need­ed to abort the poten­tial crises.

The first post-World War II threat of a finan­cial cri­sis that required Fed­er­al Reserve spe­cial inter­ven­tion was the so-called “cred­it crunch” of 1966. This episode cen­tered around a “run” on bank-nego­tiable cer­tifi­cates of deposit. The sec­ond occurred in 1970, and the imme­di­ate focus of the dif­fi­cul­ties was a “run” on the com­mer­cial paper mar­ket fol­low­ing the fail­ure of the Penn-Cen­tral Rail­road. The third threat of a cri­sis in the decade occurred in 1974–75 and involved a large num­ber of over-extend­ed finan­cial posi­tions, but per­haps can be best iden­ti­fied as cen­ter­ing around the spec­u­la­tive activ­i­ties of the giant banks. In this third episode the Franklin Nation­al Bank of New York, with assets of $5 bil­lion as of Decem­ber 1973, failed after a “run” on its over­seas branch.

Since this recent finan­cial insta­bil­i­ty is a recur­rence of phe­nom­e­na that reg­u­lar­ly char­ac­ter­ized our econ­o­my before World War II, it is rea­son­able to view finan­cial crises as sys­temic, rather than acci­den­tal, events. From this per­spec­tive, the anom­aly is the twen­ty years after World War II dur­ing which finan­cial crises were absent, which can be explained by the extreme­ly robust finan­cial struc­ture that result­ed from a Great War fol­low­ing hard upon a deep depres­sion. Since the mid­dle six­ties the his­toric cri­sis-prone behav­ior of an econ­o­my with cap­i­tal­ist finan­cial insti­tu­tions has reassert­ed itself. The past decade dif­fers from the era before World War II in that embry­on­ic finan­cial crises have been abort­ed by a com­bi­na­tion of sup­port oper­a­tions by the Fed­er­al Reserve and the income, employ­ment, and finan­cial effects that flow from an immense­ly larg­er gov­ern­ment sec­tor. This suc­cess has had a side effect, how­ev­er; accel­er­at­ing infla­tion has fol­lowed each suc­cess in abort­ing a finan­cial cri­sis. (Min­sky 1982, pp. 62–63)

On that basis, the cor­rect time peri­od from which to derive rules as to how big the “sore thumbs” of finance should be is the 1960s, and not 2000. That implies alter­na­tive fig­ures for the WEF’s indi­ca­tors that would have most of its indi­ca­tor table in red—certainly the first col­umn for house­hold sec­tor debt.

My three indi­ca­tors are the deb t to GDP ratio (which tells you how many years it would take to repay debt and is a mea­sure of the degree of pres­sure debt is exert­ing on the econ­o­my), the rate of change of debt as a per­cent­age of GDP plus the change in debt (which tells you how much of aggre­gate demand is debt-financed, and there­fore whether you are in dan­ger ter­ri­to­ry for a finan­cial cri­sis), and the cred­it impulse—the rate of change of the rate of change of debt as a frac­tion of GDP, which tells you whether a cri­sis is immi­nent and how deep it is when it strikes.

My thumb size rules, based on these indi­ca­tors and for the USA only, are the fol­low­ing:

The US was finan­cial­ly robust when the total pri­vate sec­tor debt to GDP ratio was below 100%; it was approach­ing poten­tial Depres­sion-lev­el fragili­ty when this ratio exceed­ed 175% of GDP.

When debt-financed demand accounts for more than 10 per­cent of aggre­gate demand, trou­ble is afoot (again for the USA—other coun­tries may have dif­fer­ent thresh­olds):

Final­ly, and some­what ten­ta­tive­ly, I’d see dan­ger com­ing when the Cred­it Impulse exceeds 3% in either direc­tion: 3% plus indi­cates a bub­ble, and 3% minus indi­cates that you are in a bust. On that met­ric, this is the biggest bust of all (the 1945 fig­ure was an aber­ra­tion as we moved from a war econ­o­my to a peace­time one).

On all three indi­ca­tors, the USA has been in a finan­cial­ly frag­ile state since the ear­ly 1970s—a con­clu­sion that accords with Min­sky’s deci­sion to date the tran­si­tion to a frag­ile finan­cial struc­ture in 1966—rather than the year 2000. The WEF’s report, while it does per­form a use­ful ser­vice in final­ly rec­om­mend­ing that cred­it and cred­it growth be tak­en seri­ous­ly in eco­nom­ic man­age­ment, will ulti­mate­ly be seen as an indi­ca­tor of just how seri­ous­ly econ­o­mists under­es­ti­mat­ed the role of cred­it in caus­ing eco­nom­ic crises, even when they were in one.

Min­sky, H. P. (1982). Can “it” hap­pen again? : essays on insta­bil­i­ty and finance. Armonk, N.Y., M.E. Sharpe.