Mortgage Finance Association of Australia Talk

[video src="" width="500" height="400" ]

flattr this!

The screen cap­ture video of my talk at this sem­i­nar gives an overview of both my eco­nomic analy­sis and my views on the Aus­tralian hous­ing mar­ket. Sev­eral blog mem­bers have com­mented that it’s the best overview I’ve pro­vided, so I’ve put it on the essen­tial read­ings list.

I spoke at a MFAA Pro­fes­sional Devel­op­ment Day, fol­low­ing a speaker who pointed out that most deci­sions are made by the emo­tional com­po­nents of our brains–hence some of my ref­er­ences to using the CEO seg­ment of your brain instead.

Steve Keen’s Debt­watch Pod­cast 

| Open Player in New Window

Click here for the Pow­er­Point slides.

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.

43 Responses to Mortgage Finance Association of Australia Talk

  1. ak says:


    I fully agree with the ini­tial posi­tion pre­sented by VK.

    I did not want to wade into the dis­cus­sion about the non-existence of the “bank vault” and wrong cau­sa­tion in the model described in one of the pre­vi­ous posts as I didn’t think the time was right but this issue is too impor­tant to be left as it is.

    The mod­els dis­cussed now are mis­cal­i­brated and this explains why mod­er­ate fis­cal stim­u­la­tion was suc­cess­ful in the US in 2009 (pre­vented slid­ing the global econ­omy into another Great Depres­sion). The addi­tional debt increase con­tri­bu­tion to GDP is not 0% of d(debt)/dt as the neo­clas­si­cals would like us to believe but it is also not 100% of d(aggregate_debt_of_the_private_sector)/dt

    (Note. I am not talk­ing about the net posi­tion but about the sum of all the loans)

    The coef­fi­cient could be maybe 20–40% ( I did a rough esti­ma­tion a long time ago so this num­ber may be wrong). This cor­re­sponds exactly to the amount of money spent on the sce­nar­ios dis­cussed by VK com­pared with the total amount of money cre­ated by the bank­ing system.

    If we ini­tially the ignore sec­ondary effects like spend­ing mul­ti­plier we should look at the amount of money injected to the econ­omy in a period of time which is spent on buy­ing new prod­ucts and ser­vices. This flow (accounted for as a com­po­nent of “invest­ment”) passes through the cir­cuit and is even­tu­ally with­drawn in the form of sav­ings. The rest of the money used to buy prop­er­ties on the sec­ondary mar­ket is sim­ply saved as deposits, used to pur­chase other exist­ing assets or to repay loans. Only some of the money is used to finance con­sump­tion. There is also some stim­u­la­tory effect due to spend­ing some of the bor­rowed money on Real Estate ser­vices, addi­tional costs like removals, etc. Some money is with­drawn as taxes or may flow over­seas if there is a trade deficit.

    How­ever dur­ing the debt delever­ag­ing phase dif­fer­ent effects take place and the impact of d(Debt)/dt on GDP may me much higher. Peo­ple repay loans and all the credit money is sim­ply destroyed.

    Look­ing at the NIPA accounts sys­tem we may eas­ily be con­fused as:
    1. “sav­ing” includes an increase in bank deposits, bonds but also pur­chas­ing com­pany equity — this flow of money is recy­cled and deposits do not change when one buys shares
    2. buy­ing a house even for myself is con­sid­ered to be invest­ment rather than con­sump­tion what is counter-intuitive for non-economists
    3. not all the invest­ment is financed by tak­ing new loans — some is financed by sav­ing (see 1)
    4. there is an imputed income paid to one­self by a home owner

    All of these issues are dis­cussed in detail in “Mon­e­tary Eco­nom­ics” W. God­ley and M. Lavoie
    espe­cially in Chap­ter 2.

    Every­thing is explained there in detail and I do not want to repeat the analysis.

    AD(t) = GDP(t) by def­i­n­i­tion and this is not the for­mula one can play with.

    What we need to analyse is what con­sti­tutes the out­flows of money at the nodes of the model and what con­sti­tutes the inflows of money. One of the com­po­nents of the inflows to the house­hold sec­tor is d(Debt)/dt

    There is even a sec­tion lat­ter in the book (which I haven’t thor­oughly analysed yet) where a sce­nario with “an increase in the gross new loans to per­sonal income ratio” was sim­u­lated by W.G. and M.L. what led to very famil­iar look­ing curves. Yes there was a spike in GDP when new lend­ing was increased but then there was a hangover.

    (sec­tion 11.8.2 Fig­ure 11.8B — see below)

  2. Steve Keen says:

    I dis­agree AK, on the basis of the for­mula that vk sub­se­quently agreed to: AD(t) = GDP(t-tau) + dD(t)/dt. This is some­thing you will find in Schum­peter, Min­sky and Marx. I believe that the God­ley et al analy­sis on this falsely brings in a con­ser­va­tion law that does not apply in our credit dri­ven economy.

    Schum­peter put it best:

    THE fun­da­men­tal notion that the essence of eco­nomic de-velopment con­sists in a dif­fer­ent employ­ment of exist­ing ser­vices of labor and land leads us to the state­ment that the car­ry­ing out of new com­bi­na­tions takes place through the with­drawal of ser­vices of labor and land from their pre­vi­ous employ­ments… this again leads us to two here­sies: first to the heresy that money, and then to the sec­ond heresy that also other means of pay­ment, per­form an essen­tial func­tion, hence that processes in terms of means of pay­ment are not merely reflexes of processes in terms of goods. In every pos­si­ble strain, with rare una­nim­ity, even with impa­tience and moral and intel­lec­tual indig­na­tion, a very long line of the­o­rists have assured us of the opposite…

    From this it fol­lows, there­fore, that in real life total credit must be greater than it could be if there were only fully cov­ered credit. The credit struc­ture projects not only beyond the exist­ing gold basis, but also beyond the exist­ing com­mod­ity basis. (Schum­peter 1934, pp. 95, 101; empha­sis added)

  3. ak says:


    We must not con­fuse dif­fer­ence equa­tions and dis­crete time mod­el­ling with dif­fer­en­tial equa­tions and con­tin­u­ous time domain. There­fore aggregate_expenditure(t) must always be equal to aggregate_income(t) (con­tin­u­ous time) or aggregate_expenditure(t_n) = aggregate_income(t_n) (dis­crete time)

    It is not that AD(t) = GDP (t minus lag) + dD(t)/dt

    t — tau is used in dif­fer­ence equa­tions and it is t_n-1

    Nyquist-Shannon sam­pling the­o­rem binds together synchronous-digital (dis­crete) and ana­log (con­tin­u­ous time) sig­nal pro­cess­ing approaches. As long as sam­pling fre­quency is high enough both approaches are equiv­a­lent. W. God­ley was using dis­crete time domain but this doesn’t change any­thing as his mod­els can eas­ily be trans­posed to the con­tin­u­ous time domain.

    If we have pos­i­tive dD(t)/dt (new mort­gage loans cre­ation) this will be seen in NIPA as a tem­po­rary increase in both I (invest­ment) and S (sav­ings) as (by definition)

    Y = C + I + G + (X – M) = C + S + T

    Nobody will notice what’s wrong. Every­one is happy as the econ­omy is hum­ming, invest­ment and sav­ings have risen, etc… In fact C will also rise (due to the spend­ing mul­ti­plier — see below).

    It is a bit more tricky to see the impact of ris­ing con­sumer credit as it may man­i­fest itself as falling S (com­pen­sated by ris­ing T-G lead­ing to a bud­get sur­plus or/and falling X-M lead­ing to trade and cur­rent account deficits) or may not be vis­i­ble at all in NIPA — except for an unex­pected growth of C

    Let’s split house­holds into H1 (savers) and H2 (spenders) and assume that G=T and X=M.

    H2 draw new credit from the bank­ing sec­tor at a rate dL/dt = S2 and their con­tri­bu­tion to sav­ing is (-S2) They spend this amount of money on plasma TVs, vodka, gam­bling, overeat­ing, etc, what makes them instantly happy. This extra spend­ing increases C by S2 and the GDP increases (all things equal and assum­ing spare pro­duc­tive capac­i­ties) by S2 in the first period as well. The mech­a­nism is self-feeding as in the sec­ond period the econ­omy grows fur­ther. The length of the period (the time con­stant) is deter­mined by the aver­age production-sale period which is closely related to the money veloc­ity in the pro­duc­tive econ­omy (pro­vided that we defined the money cor­rectly). Can the econ­omy grow for­ever because of the extra load­ing? Assum­ing lin­ear tax­a­tion and zero mar­ginal propen­sity to import it is up to house­holds H1 to remove the addi­tional flow. Let’s assume that they save 20% of their extra income (mar­ginal propen­sity to save) which is equal to 0.5 of GDP. So they can remove 0.1 of the delta. Let’s assume zero mar­ginal sav­ing propen­sity of H2 The new equi­lib­rium will be reached at S1 = S2 but know­ing that delta S1 = 0.1*delta Y

    we will get delta Y = 10* dL/dt

    In this model as long as no inde­pen­dent invest­ment deci­sions are made, I=S

    Obvi­ously the stim­u­la­tory effect of con­sumer debt load­ing on the econ­omy may be throt­tled due to tax­a­tion, changes in G (“auto­matic sta­bilis­ers”) and non-zero mar­ginal propen­sity to import. How­ever I can guar­an­tee that no neo­clas­si­cal econ­o­mists will spot the real cause of the mys­te­ri­ous great mod­er­a­tion… that is the growth of the can­cer of debt and deposits in the bank­ing sec­tor gob­bling the whole econ­omy (con­tribut­ing to over 10% of GDP in Aus­tralia now — not by new credit but by income redis­tri­b­u­tion). We also need to con­sider sec­ondary effects of the invest­ment deci­sions where the accel­er­a­tor kicks in. This is the real dynam­ics which is dri­ven by credit binge and when debt load­ing goes into reverse, the mul­ti­plier also goes into reverse and the neg­a­tive effects of the cri­sis can only be mod­er­ated by increased gov­ern­ment deficits as long as this is polit­i­cally palatable.

    In the model out­lined above the quan­tity of money (credit) does not affect the GDP as the cau­sa­tion doesn’t go that way. Mon­e­tarism is incor­rect and the lag is not con­stant. There is no volt­age in eco­nom­ics and the value of GDP(2011) (flow) mainly depends on the value of GDP(2010) (another flow) and some other flows not on the stock of M1, M2 or M3 how­ever the stock of debt (and sav­ings) does influ­ence the flows due to inter­ests accrued on loans (again this is invis­i­ble in the highly aggre­gated GDP statistics).

    In this con­text to analyse the credit impulse one has to dis­ag­gre­gate the changes in debt into the com­po­nent related to repay­ing the old long-term loans — delever­ag­ing — and tak­ing new short-term pro­duc­tive loans (increas­ing the size of the revolv­ing fund used to finance pro­duc­tion). If the half-decay time of loans (“money veloc­ity”) is short­ened, there can be an increase in the eco­nomic activ­ity even in a pure credit econ­omy when the over­all level of debt decreases.

    I am fully aware that the analy­sis pre­sented above is extremely over­sim­pli­fied but my goal was to show how to tackle the mys­tery the impact of debt super­charg­ing on the econ­omy rather than to build a dynamic model since what’s really inter­est­ing is the behav­iour of the spec­u­la­tors (affected by the changes in the value of assets), the impact on invest­ment in pro­duc­tion (the accel­er­a­tor effect), changes in the infla­tion rate and unem­ploy­ment, sec­ondary effects caused by the changes in the inter­est rate etc… It will take years to get every­thing right but as long as the basic model is not sound I can guar­an­tee that the model sim­ply won’t work.

  4. ak says:


    Ques­tion­ing account­ing prin­ci­ples will not lead us any­where. I am com­ing from a per­spec­tive of a guy who first writes down num­bers on a piece of paper and then tries to draw any con­clu­sions. This is the only work­ing approach in engi­neer­ing even if analy­sis and design involve top-down stages — induc­tive think­ing is the key to get every­thing right. So I will first exam­ine bal­ance sheets of a firm, worker and a bank, then aggre­gate them and add the trea­sury and the cen­tral bank. Only at this stage I can start exam­in­ing the nature of rela­tion­ships between parts of the sys­tem and in the end — the nature of money.

    W. Godley’s mod­els are con­sis­tent with the real­ity, please find a sin­gle hole in his maths or any incon­sis­tency between his mod­els and the real­ity. If a model is incon­sis­tent with the account­ing iden­ti­ties then it is not a model of the finan­cial econ­omy. Do I need to quote every­thing here again? If after a pro­duc­tive cycle all the com­modi­ties are con­sumed (or invested as pro­duc­tive cap­i­tal) by the work­ers, cap­i­tal­ists and bankers the ini­tial loan can be repaid and credit/money destruc­tion can occur — but this obvi­ously doesn’t make sense in the real life so the next pro­duc­tive cycle starts imme­di­ately. It is not the quan­tity of money lent what deter­mines the out­come. It is the amount of real assets which can be used as a lien to secure a loan and the expected rev­enue, costs and prof­its. The only rel­e­vant lag is deter­mined by the pro­duc­tion period length what is not con­stant because some­times sell­ing pro­duc­tion takes longer than antic­i­pated (and varies depend­ing on the indus­try). After aggre­gat­ing mul­ti­ple firms the quan­tity of money (the size of the revolv­ing fund of credit) has its upper limit because it is deter­mined by the value of the real cap­i­tal goods (assets) but it is also an arte­fact of the busi­ness activ­ity — hence there is another fun­da­men­tal error in the cau­sa­tion of the mod­els stat­ing that the amount of credit deter­mines the pro­duc­tion level. Again — it is the value of the pro­duc­tive (real) cap­i­tal and the demand. The demand depends on the wages, changes in con­sumer credit and invest­ment deci­sions regard­ing the next cycle. That’s why QE doesn’t change any­thing — banks don’t lend despite rolling in liq­uid­ity and com­pa­nies have a lot of cash sit­ting on their accounts and do noth­ing until there is enough antic­i­pated demand to restart the pro­duc­tion of goods.

    I didn’t read the whole Cap­i­tal but from what I saw Marx also got the basic pro­duc­tion cir­cuit right — what he messed up was his value the­ory because as a philoso­pher he wanted to cre­ate the ulti­mate the­ory of every­thing. Hoard­ing money is not a goal of the cap­i­tal­ists (except for very spe­cial cir­cum­stances), they are after increas­ing prof­its and build­ing wealth (retain­ing some of the prof­its). They want to be rich — own the biggest cor­po­ra­tions, have huge man­sions, planes, yachts, land, other real assets, maybe some gov­ern­ment debt secu­ri­ties. Only a cer­tain small per­cent­age of their assets are deposits in the bank­ing sys­tem with­drawn from the circulation.

    The fun­da­men­tal mis­take is in my opin­ion related to assum­ing that RETAINED prof­its in the basic mon­e­tary cir­cuit are mon­e­tary — they are mostly in real assets. I have already sent a link to the paper explain­ing that. This could explain the fun­da­men­tal error lead­ing to a leak in the mod­els. The total value of sav­ings and bank equity in a pure credit econ­omy must be always be equal to the aggre­gate debt and there is no mys­te­ri­ous way of accu­mu­lat­ing prof­its based on bank vaults, notes or what­ever. It is not the act of print­ing bank notes what cre­ates money (tokens), it is the cre­ation of trans­fer­able IOUs which can be extin­guished in a sim­i­lar way they were cre­ate — like old debt tal­lies. Again this is so obvi­ous to any­one who ever worked with any account­ing data­base sys­tems. In the end the asset and lia­bil­i­ties sides on all the bal­ance sheets must off­set them­selves and the only way to get there is to acknowl­edge that there is a kind of “con­ser­va­tion” law (I’m not sure what you mean by it but it sounds about right to me). Only the gov­ern­ment can cre­ate money out of thin air.

    What­ever can be added can also be sub­tracted. Every trans­ac­tion which can be done can also be undone. The only excep­tions to sym­me­try are the flow of time which goes in one direc­tion and the inter­est accrual (it is hard to imag­ine peo­ple vol­un­tar­ily allow­ing for neg­a­tive inter­est rates).

    If we ques­tion these basic account­ing rules and want to spec­u­late what money is what is not, I am afraid we will not be able to build a viable the­ory any bet­ter than these invented by Mil­ton Fried­man. It may be enough to get a Nobel price in eco­nom­ics but may not be enough to explain what’s going on.

  5. juk says:

    Thanks for the replies Steve and oth­ers. I think i got most of it, but i got a lit­tle lost in the end.

  6. hatless says:

    Hi Steve,

    Great talk as always.

    Just one ques­tion. Do you think there is a big dif­fer­ence in how every­thing plays out given that the RE bub­ble in Aus­tralia con­sisted mainly of bid­ding up exist­ing house prices, com­pared to the US and say Ire­land, which had a lot more addi­tional build­ing? I’d be inter­ested in your thoughts on this.



  7. Steve Keen says:

    Thanks Hat­less,

    And yes I do think that’s a sig­nif­i­cant dif­fer­ence in how things will play out here–though not as the spruik­ers see it of course.

    Their argu­ment has been that since we didn’t build as many houses here as in the US (true), there won’t be the same over­hang of unsold new prop­er­ties depress­ing prices (true). There­fore our prices won’t fall (false).

    What they’re omit­ting from their think­ing is that, given that Aus­tralians bor­rowed more money to gam­ble on hous­ing than even the Amer­i­cans did, our bub­ble was more of a purely spec­u­la­tive one than theirs was. They at least did some “invest­ment”, even if it was inap­pro­pri­ate to needs in the medium term. We did far less, so far more of our money went to gam­bling on house prices than increas­ing the quan­tity of housing.

    This is one rea­son our price bub­ble was more extreme than theirs, and there­fore poten­tially has much fur­ther to fall.

    It also will spread the pain of a price fall more broadly. Whereas we will have less losers among prop­erty devel­op­ers, we will have more losers amongst those who bought an exist­ing prop­erty for cap­i­tal gain as a retire­ment invest­ment. Some­thing like 30% of mar­ket demand came from “mum and dad investors” at the peak of the bub­ble. If a sig­nif­i­cant pro­por­tion of them think that the longer they hold a prop­erty, the less they’ll have for retire­ment, then–with a large lag–they could switch from the buy side to the sell side.

  8. ferb says:

    Love how stuff comes out in the wash when things are tight…

    Wont be long b4 it hap­pen here either.

  9. BrightSpark1 says:


    Nyquist-Shannon sam­pling the­o­rem binds together synchronous-digital (dis­crete) and ana­log (con­tin­u­ous time) sig­nal pro­cess­ing approaches. ”

    The sam­pling the­o­rem does no such bind­ing. It defines the min­i­mum sam­pling rate that will enable the dig­i­tal rep­re­sen­ta­tion to be used to recon­struct the orig­i­nal sig­nal — data stream. This is achieved as it avoids the effect of alias­ing. In any case most eco­nomic data is sam­pled at a far too low rate to meet the Nyquist-Shannon require­ment. Dif­fer­ence equa­tions are part of the dis­ci­pline of dis­crete math­e­mat­ics and are not valid in the descrip­tion of a dynamic sys­tem. They are not used for this pur­pose by engi­neers. Dif­fer­en­tial equa­tions are the only valid approach.

    The debt super­charg­ing is best illus­trated by the pos­i­tive feed­back loop of a dynamic sys­tem, too much loop gain and you don’t need any input (deposits) to get an out­put (credit). In days gone by before the fre­quency shift method was used to avoid the prob­lem every­one was famil­iar with this feed­back in a pub­lic address sys­tem — too much gain and it screeched! Now pos­i­tive feed­back is the warm and fuzzy type that you get from an ebay customer.

    F= G/(1-GH)

    Also what is this about destruc­tion of debt/money, is this a legal require­ment? is this a bank pol­icy? How is it done? Is it legal?


  10. ak says:


    1. Dif­fer­ence vs dif­fer­en­tial equations:

    The dif­fer­ence equa­tion (dis­crete time) approach doesn’t dif­fer from the dif­fer­en­tial equa­tion (con­tin­u­ous time) approach because these two approaches are (vir­tu­ally — I don’t want to get into detail) equiv­a­lent for all the com­po­nents of the sig­nal which have fre­quency lower than the half of the sam­pling fre­quency. This is what I wanted to say by men­tion­ing Nyquist-Shannon.

    In the fre­quency domain (assum­ing that the sys­tem is lin­ear) we can use a con­for­mal map­ping such as
    to trans­late continuous-time and discrete-time sig­nal rep­re­sen­ta­tions and trans­fer functions.

    Please notice that I was using this method­ol­ogy in dig­i­tal sig­nal pro­cess­ing so let’s assume that I know what I am talk­ing about.

    There are some lim­i­ta­tions related to the prob­lem of the unique­ness of the solu­tion of the dif­fer­ence equa­tions but this is not very rel­e­vant to our discussion.

    NIPA data is sam­pled either monthly, every quar­ter or annu­ally. We won’t get bet­ter time res­o­lu­tion any­way unless we want to model a stock mar­ket crash (trans­ac­tions hap­pen in a sub-millisecond timescale).

    Kalecki / God­ley were con­cerned about processes hap­pen­ing in the time scale of years so his sam­pling fre­quency (usu­ally 1 year) was ade­quate. Per­son­ally I would pre­fer to use the con­tin­u­ous time approach because to me it is eas­ier to think in terms of df(t)/dt rather than (f(t) — f(t-1))/ sampling_period (the Godley’s book is full of indices what makes it par­tic­u­larly hard to read) but we have to keep in mind that any ODE solver inter­nally runs Runge-Kutta so in the end we have a set of dif­fer­ence equa­tions anyway.

    2. The issue of money destruction:

    When you go to a bank and take a loan for L dol­lars, they record L on the lia­bil­i­ties side (this is your new deposit) and L on the assets side (this is your new debt). Obvi­ously they will start accru­ing inter­ests by adding r_L* L to your debt and r_D*L to your deposit every month where r_L is greater than r_D. But imag­ine that you come to the bank next day after tak­ing a loan and sim­ply change your mind and repay it, L is then sub­tracted from both the asset and lia­bil­i­ties sides of the bank’s bal­ance sheet (these trans­ac­tions are recorded in the elec­tronic ledger which is a data­base system).

    The sys­tem reverts back to the same state it had had before you took the loan (plus-minus bank fees but this is not the point).

    This is exactly an exam­ple of the destruc­tion of the credit money. If we want to say that banks can cre­ate money by extend­ing credit we also must say that debtors can destroy money by repay­ing the loans.

    Once we acknowl­edge this obvi­ous fact, we don’t need the non-existing “bank vault” which has been invented to store the money after repay­ing the debt. Please have a look at the pro-forma bal­ance sheet of Wes­pac on page 43:

    So where is the vault in West­pac? We can see Assets, Lia­bil­i­ties and Equity and by def­i­n­i­tion Assets = Lia­bil­i­ties + Equity but there is no trace of any “vault”. It is as fake as util­ity max­i­miza­tion by ratio­nal agents in neo­clas­si­cal eco­nom­ics I would say.

    3. Feed­back loops:

    I agree that there are var­i­ous feed­back loops in the sys­tem but in order to model them we must first iden­tify the sig­nal path and the error sig­nal sub­trac­tion node. The feed­back loop related to the bubble-burst cycle must include asset val­u­a­tion and credit creation/destruction what is cou­pled with the pro­duc­tive econ­omy and can affect the level of demand and as a con­se­quence, the level of production/consumption in the economy.

  11. BrightSpark1 says:


    Thanks for your comments.

    Agreed you can use dif­fer­ence equa­tions to eval­u­ate an expres­sion which con­tains inte­gral and dif­fer­en­tial terms. Agreed this is done in Dig­i­tal Sig­nal Pro­cess­ing (DSP).

    What you can­not do is find the solu­tions of a dif­fer­en­tial equa­tion. Hav­ing found found the solu­tions using appro­pri­ate method­ol­ogy the solu­tion equa­tions can then be eval­u­ated using dis­crete math­e­mat­i­cal method­ol­ogy. Any set of dif­fer­ence equa­tions would yield a dif­fer­ent result with the error (rang­ing from small to infi­nite) depend­ing on the orig­i­nal equa­tions. Dis­crete method­ol­ogy sim­ply can­not cor­rectly model feed­back with the error increas­ing as open loop gain increases.

    On the money destruc­tion ques­tion we will just have to agree to disagree.

  12. ak says:


    Thank you for your response.

    I obvi­ously agree with you in regards to the issue of errors in dif­fer­ence equa­tions used to sim­u­late feed­back sys­tems with a closed loop. We effec­tively dis­re­gard all the fre­quency com­po­nents above f_sampling/2. But in economics/econometrics some of the GDP data (not only in China) is col­lected with accu­racy of sev­eral % any­way and so if we analyse the short-term trends we do not make a seri­ous mis­take by using dif­fer­ence equations.

    The U.S. cable reported that Li, who is now a vice pre­mier, focused on just three data points to eval­u­ate Liaoning’s econ­omy: elec­tric­ity con­sump­tion, rail cargo vol­ume and bank lending.

    By look­ing at these three fig­ures, Li said he can mea­sure with rel­a­tive accu­racy the speed of eco­nomic growth. All other fig­ures, espe­cially GDP sta­tis­tics, are ‘for ref­er­ence only,’ he said smil­ing,” the cable added.

    Li is widely expected to suc­ceed Wen Jiabao as pre­mier in early 2013, a posi­tion that will put him in charge of pol­icy mak­ing in the world’s second-biggest economy.”

    I also agree that dif­fer­en­tial equa­tions are much eas­ier to under­stand than matri­ces pol­luted with hun­dreds of indices. But my point was only that Kalecki-Kaldor-Godley approach is not invalid even if the meth­ods used there can be traced back to the 1921–24 period when Kalecki was study­ing at Danzig Poly­tech­nic (Gdansk Uni­ver­sity of Technology).

    How­ever the two remain­ing issues are much more seri­ous. The prob­lem of money destruc­tion is impor­tant but is sec­ondary. The most impor­tant issue is whether flows depend in prin­ci­ple on flows with some stock depen­dency (as in Kalecki-Godley) or whether flows depend mainly on stocks due to time con­stants deter­min­ing spend­ing / pro­duc­tion process lag (what is pure mon­e­tarism as in Fried­man). So far the mod­els based on the mon­e­tary cir­cuit are seem to be mon­e­tarist to me.

    N.B. pro­duc­tion lag for a sin­gle com­pany should be mod­elled as flow_out = flow_in(t-lag) what leads to rather inter­est­ing con­se­quences when we want to describe the process in terms of dif­fer­en­tial equa­tions. This is basi­cally a delay line which may need to be imple­mented as a Bessel fil­ter, good luck with that…

  13. Pingback: House prices: Examining inflation and supply/demand factors | Financial Insights

  14. Pingback: So much for the ‘conservative’ Canadian consumer: Another look at Canada’s credit bubble | Financial Insights

  15. Pingback: Which websites measures Steven Keen's Credit Impulse?

  16. SMATS GROUP says:

    The analy­sis pre­sented is oversimplified,but there are some errors and draw­backs which have been dis­cussed and is of great use.

    Aus­tralia Taxation

  17. Pingback: Sold Australian Dollar « Reasonably Thinking

  18. Pingback: Brian Zucker

Leave a Reply