Secular stagnation III –minus the irony

By Steve Keen

I’m sorry, I couldn’t help it: when Larry Summers first made his secular stagnation speech at the IMF, and the American economics tribe heralded it as if it were the greatest (and latest) thing since sliced bread, my irony gene went into overload—and that showed in my first post on the topic. The argument that the West has been suffering from secular stagnation, and that only a series of financial bubbles have kept the illusion of prosperity going, has been part of non-orthodox economics for over three decades.

Much of this has been in the real underworld of economics, emanating from the handful of avowedly Marxist economists that have survived in academic institutions and non-mainstream media. But issues such as the impact of transnational corporations relocating production to third world free trade zones have been hot topics amongst non-mainstream economists for decades (for instance, I wrote my first piece on this topic in 1979—here’s the somewhat dated Technical Appendix for wonks).

Back then. the reaction of mainstream economists to this perspective was at best a deafening silence, and at worst derision. Weren’t these the lefties who were always expecting the final crisis of capitalism, even after the fall of the Berlin Wall?

Of course, there was more than a grain of truth to that riposte. But rather than tackling the secular stagnation arguments directly, mainstream economists dismissed them with derision, in the sublime confidence that capitalism could never have crises. My favorite such statement was by Edward Prescott (who, with Finn Kydland got the Nobel Prize for inventing the “real business cycle, representative agent” approach to economics that still dominates the mainstream today). Entitled “Some obser­va­tions on the Great Depression” and writ­ten in 1999, it con­cluded with this rhetor­i­cal flourish:

 The Marx­ian view is that cap­i­tal­is­tic economies are inher­ently unsta­ble and that exces­sive accu­mu­la­tion of cap­i­tal will lead to increas­ingly severe eco­nomic crises. Growth the­ory, which has proved to be empir­i­cally suc­cess­ful, says this is not true. The cap­i­tal­is­tic econ­omy is sta­ble, and absent some change in tech­nol­ogy or the rules of the eco­nomic game, the econ­omy con­verges to a con­stant growth path with the stan­dard of liv­ing dou­bling every 40 years. (Prescott 1999)

Yeah, right. One decade later, that open­ing Marx­ist sen­tence began to sound a lot more real­is­tic than Prescott’s dis­missal of it.

Of course, main­stream econ­o­mists could never acknowl­edge the prior argu­ments of a rival intel­lec­tual tra­di­tion, but now that Larry (peace be upon him) Sum­mers has said it, sud­denly it’s OK to spout what the left­ies have been say­ing for 30 years (though with dif­fer­ent expla­na­tions, of course—it’s declin­ing pop­u­la­tion growth and falling lev­els of inno­va­tion, not nasty transna­tion­als relo­cat­ing pro­duc­tion, or finan­cial­iza­tion of the econ­omy white-anting the indus­trial sector).

Oh give me a break—hence the irony. But since Sum­mers’ take on sec­u­lar stag­na­tion looks like it will be the flavour of the month for Amer­i­can econ­o­mists for the fore­see­able future, I have to take it more seri­ously. Hence this more seri­ous post for my first col­umn in 2014, which builds on the last two on this topic—and with no irony whatsoever.

But I’ll com­mence with two caveats: yes sec­u­lar stag­na­tion is real, but it has far more to do with rea­sons the main­stream has always rejected than with any­thing it will dream up post-Summers; and the left­ies got there first.

That said, let’s get aca­d­e­mic over how “secular-stagnation-augmented Loan­able Funds” might allow Neo­clas­si­cals to take pri­vate debt seri­ously some of the time…

The cri­sis of 2007/08 has gen­er­ated many anom­alies for con­ven­tional eco­nomic the­ory, not the least that it hap­pened in the first place. Though main­stream eco­nomic thought has many chan­nels, the com­mon belief before this cri­sis was that either crises can­not occur (Edward C. Prescott, 1999), or that the odds of such events had either been reduced (Ben Bernanke, 2002) or elim­i­nated (Robert E. Lucas, Jr., 2003) cour­tesy of the sci­en­tific under­stand­ing of the econ­omy that main­stream the­ory had developed.

This anom­aly remains unre­solved, but time has added another that is more press­ing: the fact that the down­turn has per­sisted for so long after the cri­sis. Recently Larry Sum­mers sug­gested a fea­si­ble expla­na­tion in a speech at the IMF. “Sec­u­lar stag­na­tion”, Sum­mers sug­gested, was the real expla­na­tion for the con­tin­u­ing slump, and it had been with us for long before this cri­sis began. Its vis­i­bil­ity was obscured by the Sub­prime Bub­ble, but once that burst, it was evident.

This hypoth­e­sis asserts, in effect, that the cri­sis itself was a second-order event: the main event was a ten­dency to inad­e­quate pri­vate sec­tor demand which may have existed for decades, and has only been masked by a sequence of bub­bles. The pol­icy impli­ca­tion of this hypoth­e­sis is that gen­er­at­ing ade­quate demand to ensure full employ­ment in the future may require a per­ma­nent stim­u­lus from the gov­ern­ment – mean­ing both the Con­gress and the Fed – and per­haps the reg­u­lar cre­ation of asset mar­ket bubbles.

What could be caus­ing the sec­u­lar stag­na­tion – if it exists? Krug­man (Paul Krug­man, 2013b) noted a cou­ple of fac­tors: a slow­down in pop­u­la­tion growth (which is obvi­ously hap­pen­ing: see Fig­ure 1); and “a Bob Gor­donesque decline in inno­va­tion” (which is rather more conjectural).

Though Sum­mers’ the­sis has its main­stream crit­ics, there’s a cho­rus of New Key­ne­sian sup­port for the “sec­u­lar stag­na­tion” argu­ment, which implies it will soon become the con­ven­tional expla­na­tion for the per­sis­tence of this slump long after the ini­tial finan­cial cri­sis has passed.

Krugman’s change of tune here is rep­re­sen­ta­tive. His most recent book-length foray into what caused the cri­sis – and what pol­icy would get us out of it – was enti­tled End This Depres­sion NOW!. The title, as well as the book’s con­tents, pro­claimed that this cri­sis could be ended “in the blink of an eye”. All it would take, Krug­man then pro­posed, was a suf­fi­ciently large fis­cal stim­u­lus to help us escape the “Zero Lower Bound”:

The sources of our suf­fer­ing are rel­a­tively triv­ial in the scheme of things, and could be fixed quickly and fairly eas­ily if enough peo­ple in posi­tions of power under­stood the realities…

One main theme of this book has been that in a deeply depressed econ­omy, in which the inter­est rates that the mon­e­tary author­i­ties can con­trol are near zero, we need more, not less, gov­ern­ment spend­ing. A burst of fed­eral spend­ing is what ended the Great Depres­sion, and we des­per­ately need some­thing sim­i­lar today. (Paul Krug­man, 2012, pp. 23, 231)

Fig­ure 1: Pop­u­la­tion growth rates are slowing

Post-Summers, Krug­man is sug­gest­ing that a short, sharp burst of gov­ern­ment spend­ing will not be enough to restore “the old nor­mal”. Instead, to achieve pre-crisis rates of growth in future – and pre-crisis lev­els of unem­ploy­ment – per­ma­nent gov­ern­ment deficits, and per­ma­nent Fed­eral Reserve spik­ing of the asset mar­ket punch via QE and the like, may be required.

Not only that, but past appar­ent growth suc­cesses – such as The Period Pre­vi­ously Known as The Great Mod­er­a­tion– may sim­ply have been above-stagnation rates of growth moti­vated by bubbles:

So how can you rec­on­cile repeated bub­bles with an econ­omy show­ing no sign of infla­tion­ary pres­sures? Summers’s answer is that we may be an econ­omy that needs bub­bles just to achieve some­thing near full employ­ment – that in the absence of bub­bles the econ­omy has a neg­a­tive nat­ural rate of inter­est. And this hasn’t just been true since the 2008 finan­cial cri­sis; it has arguably been true, although per­haps with increas­ing sever­ity, since the 1980s. (Paul Krug­man, 2013b)

This argu­ment ele­vates the “Zero Lower Bound” from being merely an expla­na­tion for the Great Reces­sion to a Gen­eral The­ory of Macro­eco­nom­ics: if the ZLB is a per­ma­nent state of affairs given sec­u­lar stag­na­tion, then per­ma­nent gov­ern­ment stim­u­lus and per­ma­nent bub­bles may be needed to over­come it:

One way to get there would be to recon­struct our whole mon­e­tary sys­tem – say, elim­i­nate paper money and pay neg­a­tive inter­est rates on deposits. Another way would be to take advan­tage of the next boom – whether it’s a bub­ble or dri­ven by expan­sion­ary fis­cal pol­icy – to push infla­tion sub­stan­tially higher, and keep it there. Or maybe, pos­si­bly, we could go the Krug­man 1998/Abe 2013 route of push­ing up infla­tion through the sheer power of self-fulfilling expec­ta­tions. (Paul Krug­man, 2013b)

So is sec­u­lar stag­na­tion the answer to the puz­zle of why the econ­omy hasn’t recov­ered post the cri­sis? And is per­ma­nently blow­ing bub­bles (as well as per­ma­nent fis­cal deficits) the solution?

Firstly there is ample evi­dence for a slow­down in the rate of eco­nomic growth over time – as well as its pre­cip­i­tate fall dur­ing and after the crisis.

Fig­ure 2: A sec­u­lar slow­down in growth caused by a sec­u­lar trend to stag­na­tion?

The growth rate was as high as 4.4% p.a. on aver­age from 1950–1970, but fell to about 3.2% p.a. from 1970–2000 and was only 2.7% in the Naugh­ties prior to the cri­sis – after which it has plunged to an aver­age of just 0.9% p.a. (see Table 1).

Table 1: US Real growth rates per annum by decade

 

So the sus­tained growth rate of the US econ­omy is lower now than it was in the 1950s–1970s, and the undoubted demo­graphic trend that Krug­man nom­i­nates is clearly one fac­tor in this decline.

Another fac­tor that Krug­man alludes to in his post is the rise in house­hold debt dur­ing 1980–2010 – which at first glance is incom­pat­i­ble with the “Loan­able Funds” model of lend­ing to which he sub­scribes. In the Loan­able Funds model, the aggre­gate level of debt (and changes in that level) are irrel­e­vant to macro­eco­nom­ics – only the dis­tri­b­u­tion of debt can have significance:

Ignor­ing the for­eign com­po­nent, or look­ing at the world as a whole, we see that the over­all level of debt makes no dif­fer­ence to aggre­gate net worth – one person’s lia­bil­ity is another person’s asset. It fol­lows that the level of debt mat­ters only if the dis­tri­b­u­tion of net worth mat­ters, if highly indebted play­ers face dif­fer­ent con­straints from play­ers with low debt. (Paul Krug­man,
2012a, p. 146)

Fur­ther­more, the dis­tri­b­u­tion of debt can only have macro­eco­nomic sig­nif­i­cance at pecu­liar times, when the mar­ket mech­a­nism is unable to func­tion because the “nat­ural rate of inter­est” – the real inter­est rate that will clear the mar­ket for Loan­able Funds, and lead to zero infla­tion with other mar­kets (includ­ing labor) in equi­lib­rium – is negative.

Prior to Sum­mers’ the­sis, Krug­man had argued that this pecu­liar period began in 2008 when the econ­omy entered a “Liq­uid­ity Trap”. Pri­vate debt mat­ters dur­ing a Liq­uid­ity Trap because lenders, wor­ried about the capac­ity of bor­row­ers to repay, impose a limit on debt that forces bor­row­ers to repay their debt and spend less. To main­tain the full-employment equi­lib­rium, peo­ple who were once lenders have to spend more to com­pen­sate for the fall in spend­ing by now debt-constrained borrowers.

But lenders are patient peo­ple, who by def­i­n­i­tion have a lower rate of time pref­er­ence than bor­row­ers, who are impa­tient people:

Now, if peo­ple are bor­row­ing, other peo­ple must be lend­ing. What induced the nec­es­sary lend­ing? Higher real inter­est rates, which encour­aged “patient” eco­nomic agents to spend less than their incomes while the impa­tient spent more. (Krug­man, “Delever­ag­ing and the Depres­sion Gang”)

The prob­lem in a Liq­uid­ity Trap is that rates can’t go low enough to encour­age patient agents to spend enough to com­pen­sate for the decline in spend­ing by now debt-constrained impa­tient agents.

You might think that the process would be sym­met­ric: debtors pay down their debt, while cred­i­tors are cor­re­spond­ingly induced to spend more by low real inter­est rates. And it would be sym­met­ric if the shock were small enough. In fact, how­ever, the delever­ag­ing shock has been so large that we’re hard up against the zero lower bound; inter­est rates can’t go low enough. And so we have a per­sis­tent excess of desired sav­ing over desired invest­ment, which is to say per­sis­tently inad­e­quate demand, which is to say a depres­sion. (Krug­man, “Delever­ag­ing and the Depres­sion Gang”)

After Sum­mers, Krug­man started to sur­mise that the econ­omy may have been expe­ri­enc­ing sec­u­lar stag­na­tion since 1985, and that only the rise in house­hold debt masked this phe­nom­e­non. Con­se­quently the level and rate of change of pri­vate debt could have been macro­eco­nom­i­cally sig­nif­i­cant not merely since 2008, but since as long ago as 1985.

Fig­ure 3: Ratio of house­hold debt to GDP

Com­ment­ing on the data (Fig­ure 3, sourced from the St Louis Fed’s excel­lent FRED data­base, is taken from Krugman’s post), Krug­man noted that per­haps the increase in debt from 1985 on masked the ten­dency to sec­u­lar stag­na­tion. Cru­cially, he pro­posed that the “nat­ural rate of inter­est” was neg­a­tive per­haps since 1985, and only the demand from bor­row­ers kept actual rates pos­i­tive. This in turn implied that, absent bub­bles in the stock and hous­ing mar­kets, the econ­omy would have been in a liq­uid­ity trap since 1985:

There was a sharp increase in the ratio after World War II, but from a low base, as fam­i­lies moved to the sub­urbs and all that. Then there were about 25 years of rough sta­bil­ity, from 1960 to around 1985. After that, how­ever, house­hold debt rose rapidly and inex­orably, until the cri­sis struck.

So with all that house­hold bor­row­ing, you might have expected the period 1985–2007 to be one of strong infla­tion­ary pres­sure, high inter­est rates, or both. In fact, you see nei­ther – this was the era of the Great Mod­er­a­tion, a time of low infla­tion and gen­er­ally low inter­est rates. With­out all that increase in house­hold debt, inter­est rates would pre­sum­ably have to have been con­sid­er­ably lower – maybe neg­a­tive. In other words, you can argue that our econ­omy has been try­ing to get into the liq­uid­ity trap for a num­ber of years, and that it only avoided the trap for a while thanks to suc­ces­sive bubbles.

In gen­eral, the Loan­able Funds model denies that pri­vate debt mat­ters macro­eco­nom­i­cally, as Krug­man put it emphat­i­cally in a series of blog posts in 2012:

Keen then goes on to assert that lend­ing is, by def­i­n­i­tion (at least as I under­stand it), an addi­tion to aggre­gate demand. I guess I don’t get that at all. If I decide to cut back on my spend­ing and stash the funds in a bank, which lends them out to some­one else, this doesn’t have to rep­re­sent a net increase in demand. Yes, in some (many) cases lend­ing is asso­ci­ated with higher demand, because resources are being trans­ferred to peo­ple with a higher propen­sity to spend; but Keen seems to be say­ing some­thing else, and I’m not sure what. I think it has some­thing to do with the notion that cre­at­ing money = cre­at­ing demand, but again that isn’t right in any model I under­stand. (Paul Krug­man, 2012b. Empha­sis added).

How­ever, the Sum­mers con­jec­ture pro­vides a means by which pri­vate debt could assume macro­eco­nomic sig­nif­i­cance since 1985 within the Loan­able Funds model. Once sec­u­lar stag­na­tion com­menced – dri­ven, in this con­jec­ture, by the actual drop in the rate of growth of pop­u­la­tion and a hypoth­e­sized decline in inno­va­tion – the econ­omy was effec­tively in a liq­uid­ity trap, and some­how ris­ing debt hid it from view.

That is the broad brush, but I expect that explain­ing this while remain­ing true to the Loan­able Funds model will not be an easy task—since, like a Liq­uid­ity Trap itself, the Loan­able Funds model is not sym­met­ric. Whereas Krug­man was able to explain how pri­vate debt causes aggre­gate demand to fall when debt is falling and remain true to the Loan­able Funds model (in which banks are mere inter­me­di­aries and both banks and money can be ignored – see Gauti B. Eggerts­son and Paul Krug­man, 2012), it will be much harder to explain how debt adds to aggre­gate demand when it is ris­ing. This case is eas­ily made in an Endoge­nous Money model in which banks cre­ate new spend­ing power, but it fun­da­men­tally clashes with Loan­able Funds in which lend­ing sim­ply redis­trib­utes exist­ing spend­ing power from lenders to bor­row­ers. Nonethe­less, Krug­man has made such a state­ment in a post-Summers blog:

Debt was ris­ing by around 2 per­cent of GDP annu­ally; that’s not going to hap­pen in future, which a naïve cal­cu­la­tion sug­gests means a reduc­tion in demand, other things equal, of around 2 per­cent of GDP. (Paul Krug­man, 2013a)

If he man­ages to pro­duce such a model, and if it still main­tains the Loan­able Funds frame­work, then the model will need to show that pri­vate debt affects aggre­gate demand only dur­ing a period of either sec­u­lar stag­na­tion or a liq­uid­ity slump – oth­er­wise the secular-stagnation-augmented Loan­able Funds model will be a capit­u­la­tion in all but name to the Endoge­nous Money camp (Nick Rowe, 2013). Assum­ing that this is what Krug­man will attempt, I want to con­sider the empir­i­cal evi­dence on the rel­e­vance of pri­vate debt to macro­eco­nom­ics. If it is indeed true that pri­vate debt only mat­tered post-1985, then this is com­pat­i­ble with a secular-stagnation-augmented Loan­able Funds model – what­ever that may turn out to be. But if pri­vate debt mat­ters before 1985, when sec­u­lar stag­na­tion was clearly not an issue, then this points in the direc­tion of Endoge­nous Money being the empir­i­cally cor­rect model.

I will con­sider two indi­ca­tors: the cor­re­la­tion between change in aggre­gate pri­vate non­fi­nan­cial sec­tor debt and unem­ploy­ment, and the cor­re­la­tion between the accel­er­a­tion of aggre­gate pri­vate non­fi­nan­cial sec­tor debt and the change in unem­ploy­ment. I am also using two much longer time series for debt and unem­ploy­ment. Fig­ure 4 extends Krugman’s FRED chart by includ­ing busi­ness sec­tor debt as well (click here to see how this data was com­piled – and a longer term esti­mate for US debt that extends back to 1834: the data is down­load­able from here). The unem­ploy­ment data shown in Fig­ure 5 is com­piled from BLS and NBER sta­tis­tics and Lebergott’s esti­mates (Stan­ley Leber­gott, 1986, 1954, Christina Romer, 1986) and extends back to 1890.

Fig­ure 4: Long term series on Amer­i­can pri­vate debt

Fig­ure 5: Cor­re­la­tion of change in aggre­gate pri­vate debt with unemployment

Cor­re­la­tion is not cau­sa­tion as the cliché goes, but a cor­re­la­tion coef­fi­cient of –0.57 over almost 125 years implies that the change in debt has macro­eco­nomic sig­nif­i­cance at all times – and not just dur­ing either sec­u­lar stag­na­tion or liq­uid­ity traps.

Table 2:
Cor­re­la­tion of change in aggre­gate pri­vate debt with unem­ploy­ment by decade

Shorter time spans empha­size the point that nei­ther sec­u­lar stag­na­tion nor liq­uid­ity traps can be invoked to explain why changes in the level of pri­vate debt have macro­eco­nomic sig­nif­i­cance. Sec­u­lar stag­na­tion surely didn’t apply between 1890 and 1930, yet the cor­re­la­tion is-0.6; nei­ther sec­u­lar stag­na­tion nor a liq­uid­ity trap applied in the period from 1950 till 1970, yet the cor­re­la­tion is sub­stan­tial in those years as well.

The cor­re­la­tion clearly jumps dra­mat­i­cally in the period after the Stock Mar­ket Crash of 1987, but that is more com­fort­ably con­sis­tent with the basic Endoge­nous Money case that I have been mak­ing – that new pri­vate debt cre­ated by the bank­ing sec­tor adds to aggre­gate demand – than it will be with any secular-stagnation-augmented Loan­able Funds model.

The debt accel­er­a­tion data (Michael Biggs and Thomas Mayer, 2010, Michael Biggs et al., 2010) ham­mers this point even fur­ther. Fig­ure 6 shows the accel­er­a­tion of aggre­gate pri­vate sec­tor debt and change in unem­ploy­ment from 1955 (three years after quar­terly data on debt first became avail­able) till now. The cor­re­la­tion between the two series is –0.69.

Fig­ure 6: Cor­re­la­tion of accel­er­a­tion in aggre­gate pri­vate debt with change in unemployment

As with the change in debt and unem­ploy­ment cor­re­la­tion, shorter time spans under­line the mes­sage that pri­vate debt mat­ters at all times. Though the cor­re­la­tion is strik­ingly higher since 1987 – a date I empha­size because I believe that Greenspan’s actions in res­cu­ing that bub­ble then led to the Ponzi econ­omy that Amer­ica has since become – it is high through­out, includ­ing in times when nei­ther “sec­u­lar stag­na­tion” nor a “liq­uid­ity trap” can be invoked.

Table 3:
Cor­re­la­tion of accel­er­a­tion in aggre­gate pri­vate debt with change in unem­ploy­ment by decade

I await the IS-LM or New Key­ne­sian DSGE model that Krug­man will pre­sum­ably pro­duce to pro­vide an expla­na­tion for the per­sis­tence of the cri­sis in terms that, how­ever tor­tured, emanate from con­ven­tional eco­nomic logic in which banks and money are ignored (though pri­vate debt is finally con­sid­ered), and in which every­thing hap­pens in equi­lib­rium. But how­ever clever it might be, it will not be con­sis­tent with the data.

Ref­er­ences

Bernanke, Ben. 2002. “Defla­tion: Mak­ing Sure “It” Doesn’t Hap­pen Here,” Wash­ing­ton: Fed­eral Reserve Board.

Biggs, Michael and Thomas Mayer. 2010. “The Out­put Gap Conun­drum.” Intereconomics/Review of Euro­pean Eco­nomic Pol­icy, 45(1), 11–16.

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

Eggerts­son, Gauti B. and Paul Krug­man. 2012. “Debt, Delever­ag­ing, and the Liq­uid­ity Trap: A Fisher-Minsky-Koo Approach.” Quar­terly Jour­nal of Eco­nom­ics, 127, 1469–513.

Krug­man, Paul. 2012a. End This Depres­sion Now! New York: W.W. Norton.

____. 2012b. “Min­sky and Method­ol­ogy (Wonk­ish),” The Con­science of a Lib­eral. New York: New York Times.

____. 2013a. “Sec­u­lar Stag­na­tion Arith­metic,” P. Krug­man, The Con­science of a Lib­eral. New York: New York Times.

____. 2013b. “Sec­u­lar Stag­na­tion, Coalmines, Bub­bles, and Larry Sum­mers,” P. Krug­man, The Con­science of a Lib­eral. New York: New York Times.

Leber­gott, Stan­ley. 1986. “Dis­cus­sion of Romer and Weir Papers.” The Jour­nal of Eco­nomic His­tory, 46(2), 367–71.

____. 1954. “Mea­sur­ing Unem­ploy­ment.” The Review of Eco­nom­ics and Sta­tis­tics, 36(4), 390–400.

Lucas, Robert E., Jr. 2003. “Macro­eco­nomic Pri­or­i­ties.” Amer­i­can Eco­nomic Review, 93(1), 1–14.

Prescott, Edward C. 1999. “Some Obser­va­tions on the Great Depres­sion.” Fed­eral Reserve Bank of Min­neapo­lis Quar­terly Review, 23(1), 25–31.

Romer, Christina. 1986. “Spu­ri­ous Volatil­ity in His­tor­i­cal Unem­ploy­ment Data.” Jour­nal of Polit­i­cal Econ­omy, 94(1), 1–37.

Rowe, Nick. 2013. “What Steve Keen Is Maybe Try­ing to Say,” N. Rowe, Worth­while Cana­dian Ini­tia­tive. Canada: Nick Rowe.