Why China Had To Crash Part 1

Flattr this!

In this post I con­sider the econ­omy in gen­eral: I’ll cover asset mar­kets in par­tic­u­lar in the next col­umn, but you’ll need to under­stand today’s post to com­pre­hend the stock and prop­erty mar­ket dynam­ics at play. Hav­ing said that, the Shang­hai Index fell another 7.5% on Tues­day, after los­ing 8.5% on Mon­day, and is now down over 45% from its peak—so I’ll try to write the stock-mar­ket-spe­cific post by tomor­row. In this post I’ll show, very sim­ply, why a slow­down in the rate of growth of pri­vate debt will cause a cri­sis, if both the level and the rate of change of debt are high at the time of the slow­down.

Click here to read the rest of this post.

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.
  • Tim Ward

    …Given that we already have an esti­mate of the flow of new debt at a given point in time, the GDP fig­ure is a proxy for the turnover of exist­ing money at the same time. There is unavoid­ably some dou­ble-count­ing, but it will have to do: while dou­ble-count­ing will plague this analy­sis, the alter­na­tive of ignor­ing the impact of the change of debt on demand and income is a much big­ger mis­take.

    Given those caveats, we can treat total expen­di­ture in the econ­omy on both goods and ser­vices (GDP) and assets (prop­erty and shares) as GDP plus the change in debt. …”

    Just curi­ous if it is pos­si­ble to get a han­dle on some of the dou­ble count­ing.

    …A new home that is built dur­ing a given year is counted in that year’s GDP, while the pur­chase of a pre­vi­ously owned house has already been counted in the GDP of the year it was con­structed.” http://www.colorado.edu/economics/courses/econ2020/section6/GDP-components.html

    But there will be an addi­tion to the debt for that same year, the con­struc­tion financ­ing (then replaced by the mort­gage financ­ing).

    So this seems to be an exam­ple of the dou­ble count­ing. A new house will add to both GDP and the change in debt in that year. But we have a good idea of the amount of new homes, com­pared to exist­ing home sales. Very roughly 10% or some­thing like that.

    Apolo­gies if this isn’t a type of dou­ble count­ing, or if I have made some other error, or mis­com­pre­hend. E.&O.E. !

  • igmos00

    Steve I am I appre­ci­ate what you are try­ing to do. I saw that you have codes and you are try­ing to run sim­u­la­tions. Two issues. The dou­ble count­ing mat­ters because it low­ers the impact of the argu­ment, cor­rect? Sec­ond, why are you using 25% per­cent growth rates in debt in your table. Isn’t it more like slightly larger than 10% or are you say­ing that just in 2014 the growth rate in debt was say 20% and the gdp growth rate accord­ing to offi­cial fig­ures was less than 10%. The explo­sion in debt has been going on for sev­eral years. I guess I have no idea why you chose the num­bers you chose in the table. I would cer­tainly read a blog post that dis­cusses the def­i­n­i­tion and source of the num­bers that should go in that table. It seems to me that it could make a big dif­fer­ence on the chances that china will be able to lower the rate of growth of debt with­out a major crash.

  • igmos00

    I see the debt to gdp has almost dou­bled in less than ten years. I am not an econ­o­mist. A bit of back­ground would be nice. So you are say­ing for that to hap­pen debt must have been grow­ing a lot faster than gdp. Say 20% and 10%. Ok so it is still not clear where in the process of achiev­ing par­ity with the rate of growth of gdp is china at this point. The slow down in aggre­gate demand could have taken place in the past and could pos­si­bly result in yet another “japan” as you put it. It seems likely that you are right that china is in trou­ble but I am not sure where things stand at present based on what you have writ­ten. Any­way it is clear at least to me that there is rea­son to worry.

  • Steve,

    “So total expen­di­ture and income in our econ­omy is the sum of the turnover of exist­ing money, plus the change in pri­vate debt” 

    I under­stand that this will be true in the first time period, but won’t the change in the pri­vate debt of the first period pos­si­bly get added to the turnover of exist­ing money in the sec­ond period too? So even if the change in pri­vate debt is less in the sec­ond period there may still not be a defla­tion­ary effect.

    Or some it could be removed in tax­a­tion, or it could be saved. Then there would still pos­si­bly be a defla­tion­ary effect, but we’d need to know just how much was taxed away and how much was saved.

    It’s a gen­uine ques­tion. I’m not try­ing to be awk­ward!

    Thanks in advance.


  • Hi Peter,

    Think of it as water revolv­ing in a bucket (old Bill Phillips’s hydraulic model was a great insight!), with new water being added from a hose. The water in the bucket is the exist­ing money sup­ply. Its rota­tion in a year is the vol­ume of exist­ing-money-financed expen­di­ture and income. The water com­ing in from the hose is the new debt-cre­ated money (and demand and income).

    Next think of how we record flows like this: at a sin­gle point in time, we record the vol­ume and rate of flow of the water in the bucket in terms of mega­l­itres per year; at the same time, the water flow­ing in from the hose is recorded in mega­l­itres per year. At that point of record­ing, the two flows are dis­tinct, so you can add them with­out dou­ble-count­ing.

    I might use that explanation–if it makes sense to you–in the next col­umn!

  • I rushed that piece out given the time urgency right now; in the next one or two posts I’ll include the actual stats for China for the last 8 years. The rate of growth of debt has been as high as 35% of GDP per year; it’s cur­rently run­ning at 25% of GDP a year.

  • Com­ing in the next cou­ple of posts, as noted in the pre­vi­ous com­ment.

  • It’s more the poor qual­ity of GDP record­ing that I was focus­ing on than the con­cep­tual side of things. If the actual turnover of money was mea­sured, rather than GDP, then there would be no dou­ble-count­ing worry–as explained in the reply to Peter Mar­tin. It’s just that GDP mea­sure­ment is likely to mix some things that are debt-financed up with non-debt financed. But given that–unfortunately!!–most debt is taken on to finance asset pur­chases in the first instance, there won’t be as much dou­ble-count­ing as there could have been back in the 1950s when banks actu­ally financed busi­ness turnover.

  • John­Smith

    [Richard Vague] found that every [..eco­nomic crises across the globe over the last one and a half cen­turies..] occurred when the pri­vate debt to GDP ratio exceeded 1.5 times GDP, and when the ratio had risen by 17% or more over a 5 year period.”

    Is there any expla­na­tion why the actual level of debt mat­ters? Why 1.5*GDP? Is it the inter­est rates? Why does just the change of the ratio of debt to gdp mat­ter? Why 17%/5years?
    Up to now I found the cri­te­rion of Prof Keen about the start­ing point of a finan­cial crises (change in debt gets big­ger than income, i.e. the accel­er­a­tion(!) goes through zero) very con­vinc­ing.
    Does this cri­te­rion still apply in these cases?

    Another ques­tion: Does any­one know about a cri­te­rion on how long a phase of delever­ag­ing lasts?

  • Tim Ward

    The change in debt level not asso­ci­ated with GDP changes (ie not dou­ble counted) could pos­si­bly be approx­i­mated by:

    The sum of new debt for stock buy­backs & div­i­dends + new mar­gin debt + debt increases for exist­ing home sales and exist­ing com­mer­cial real estate. (and maybe + debt rollovers, where the debt is increased)

    These would all be debt increases asso­ci­ated with incomes. And they wouldn’t be in the GDP num­bers.

  • John­Smith

    That sounds very rea­son­able!
    Do you think (or bet­ter: Is it quan­tifi­able) if these kinds of debt not directly affect­ing GDP still indi­rectly affect the real econ­omy (like employ­ment, wages, pro­duc­tion level, etc.)?

  • Hi Steve,

    Thanks very much for your reply.

    I’ve been think­ing about this and I’ve made your water spin­ning in a bucket into water cir­cu­lat­ing around the moat of a cas­tle. We’ll assume the exis­tence of pumps to make it do that. Also assume zero evap­o­ra­tion. So we can add water from a pipe to rep­re­sent the cre­ated money, as you sug­gest, and add the two flows which can be regarded as sep­a­rate in the first instance.But, what hap­pens to the main flow as the level in the moat rises? As it will if we add water. If the water speed stays the same then even if the flow from the pipe drops, the total flow won’t decrease. But if the flow is gov­erned by the capac­ity of the pumps then more water will mean less speed and what you say will be exactly right!

    So does this anal­ogy really answer our ques­tion? It depends on how the main pumps oper­ate. If their pump­ing capac­ity also dri­ves the water in the pipe etc.

    To stay on your hydraulic anal­ogy, I was think­ing we could rep­re­sent gov­ern­ment spend­ing as the release of water from a dam which for all intents and pur­poses was hold­ing a lake of infi­nite size. So gov­ern­ment spend­ing can be what­ever govt wants it to be. The released water forms a river which flows down­hill into the sea, as rivers do, and the water enter­ing the sea can rep­re­sent tax­a­tion. Again we’ll assume no evap­o­ra­tion. So if there’s no inter­fer­ence with the river, spend­ing must equal tax­a­tion rev­enue. But if the river flows too quickly we could have infla­tion. Too slowly means reces­sion.

    But what hap­pens if the inhab­i­tants of our hypo­thet­i­cal land decide to pump some water out of our river and save it in their own smaller finite sized reser­voirs? The amount of water enter­ing the sea will be less than enters the river. ie Govt will be in deficit. If they then drain their reser­voirs ie spend their sav­ings , there will be more water enter­ing the sea than enters from the govts dam and they govt will be in sur­plus.

    We can now pos­tu­late the exis­tence of a bank which also has the capac­ity to lend water to the inhab­i­tants. But unlike govt it wants back what it has lent out. So if it is net lend­ing to the inhab­i­tants the water flow in the river (which of course rep­re­sents total demand) will increase and the govt may even be in sur­plus as more water enters the sea than leaves the govt dam. But if net lend­ing slows or is even reversed then the water flow in the river will fall lead­ing to lower demand and so increas­ing a ten­dency towards a reces­sion. The govt can of course fix the prob­lem by let­ting more water out of its own dam to com­pen­sate for the fall in lend­ing. Govt just needs to learn to not fret about run­ning out of water which it can’t do because its own dam is essen­tially infi­nite in size.

    So I now think you’re right. But it has taken a bit of thought to see that. 

    Thanks again and Best Wishes


  • Tim Ward

    Do you think (or bet­ter: Is it quan­tifi­able) if these kinds of debt not directly affect­ing GDP still indi­rectly affect the real econ­omy (like employ­ment, wages, pro­duc­tion level, etc.)?”

    Yes, the other kinds of pri­vate debt have to affect the real econ­omy, because increases in pri­vate sec­tor debt decrease fur­ther cred­it­wor­thi­ness, and they have to be ser­viced. It’s some­thing like the pri­vate sec­tor econ­omy hav­ing a time vari­ant cred­it­wor­thi­ness, the total amount of pos­si­ble ser­vice­able debt. This total cred­it­wor­thi­ness changes over time. It also depends on inter­est rates. But as you get close to the bound­ary, ser­vice­able credit growth has to slow. Which means you run into a brick wall. So this speaks to the sig­nif­i­cance of the debt to GDP ratio, when it gets big enough, the creditworthiness/serviceability runs out. “All loaned up.” Empir­i­cally, the ratios are as Prof. Keen shows.

    I haven’t thought about it, but its per­haps some­thing like the heat death of the uni­verse. The cure is the debt jubilee.

  • John­Smith

    I was won­der­ing about it, because of Prof. Keens plot, where he got the strong cor­re­la­tion (~-0.9) between unem­ploy­men­trate and change of pri­vate debt to gdp ratio. And there he did not adjust for the usage of the debt. So it seemed that it did not mat­ter how the debt was spent to get the effect on unem­ploy­ment.
    So could a big com­pany just take a huge loan and buy what­ever they want (like stocks, bonds, or noth­ing) and there will be (indi­rectly) jobs cre­ated by it?

  • Tim Ward

    The cat­e­gory of debt hav­ing to do with stock buy backs, mar­gin debt etc could be respon­si­ble for the cor­re­la­tion coef­fi­cient being less than the very high cor­re­la­tions, say ~0.97 — 0.995.

    LBOs, lever­aged buy outs, in which com­pa­nies are loaded up with debt, none of which goes into real cap­i­tal, pro­duc­tive equip­ment, research and devel­op­ment, prod­uct devel­op­ment, new hires, etc, often wind up going bank­rupt. Wip­ing out employ­ment.

    (**Addi­tional new debt for LBOs, M&A, and the like also looks like the type of debt that would be in the non-GDP, or non-dou­ble counted debt.**) 

    Typ­i­cally, high lev­els of M&A activ­ity are a late cycle sign. And we’ve seen this.

    The new debt for the stock buy backs and div­i­dend pay­outs looks some­thing like a com­pany is LBO-ing itself. Doom. Since the bor­rowed money has not gone into pro­duc­tive cap­i­tal, prod­uct devel­op­ment etc, how has the com­pany improved it’s abil­ity to earn money to pay back the loan/service the bonds? What it has done is crip­ple it’s revenue/income, by increas­ing ser­vic­ing costs, and decreas­ing prof­itabil­ity. And how can it then pay back the bonds/loans when the money has gone/been wasted? Only when the money goes into some­thing use­ful does the com­pany improve it’s abil­ity to pay back the debt. 

    So the surge in those kinds of debt looks like a very bad sign to me. Some­thing like ~$2 tril­lion in bonds for buy­backs, other junk bonds, lever­aged loans, and CLOs has been cre­ated in recent years in the US, last 3 years or some­thing. Debt with highly prob­lem­atic ser­vice­abil­ity. Declines in the under­ly­ing col­lat­eral will be dis­as­ter­ous. So when enough of this debt accu­mu­lates, it looks like a wipe out of employ­ment is on the way. The high dol­lar exac­er­bates this, as demand for US prod­ucts declines. So the slow­down in China and deval­u­a­tion, and the com­mod­ity crash, with com­mod­ity pro­ducer cur­ren­cies crash­ing, with high dol­lar, declin­ing trade vol­umes, and large junk debt accu­mu­la­tions, and the now large debt to GDP in China, looks like very bad.

  • John­Smith

    The cat­e­gory of debt hav­ing to do with stock buy backs, mar­gin debt etc could be respon­si­ble for the cor­re­la­tion coef­fi­cient being less than the very high cor­re­la­tions, say ~0.97 — 0.995.”

    This is an inter­est­ing and testable (!) hypoth­e­sis! Do you know where to get the data (debt/loans by usage)? So we could run some regres­sion?

  • Tim Ward

    NYSE mar­gin debt; http://www.nyxdata.com/nysedata/asp/factbook/viewer_edition.asp?mode=table&key=3153&category=8

    Some of the other data might be more dif­fi­cult to get, but some­times bank research comes out with data for them. Could try search­ing Bloomberg, or the St. Louis Fed. 


    Bloomberg men­tions Sun­dial Cap­i­tal Research Inc., Jason Goepfert, Sen­ti­men­Trader, as a source.

  • Tim Ward

    Another idea to get at an esti­mate of cor­po­rate debt ‘wasted’ in the US, could be to sub­tract the FRED data for non­fi­nan­cial cor­po­rate busi­ness cap­i­tal in some form, from total non­fi­nan­cial cor­po­rate busi­ness debt.

    But I don’t know the best series to use. eg. https://research.stlouisfed.org/fred2/series/NCBDBIQ027S
    sub­tract the cap­i­tal from that, could give the ‘wasted’ debt.

  • John­Smith

    Thank you for the links!
    When i find time i will look into it. Some of the data seem to need some pre-pro­cess­ing or smooth­ing first.

  • Pingback: What Is the Post-Post-Davos Model of the World? | technology market()

  • Pingback: What Is the Post-Post-Davos Model of the World? – pulsebell()

  • Pingback: What Is the Post-Post-Davos Model of the World? | Daily Blue Planet()

  • Pingback: What is the Post-Post Davos Model of the World? – Le Point International()