Debunking Economics, Part XI: Predicting the Crisis

As many readers of this blog will know, Steve Keen is generally the economist credited with best foreseeing and warning about the 2008 financial crash. The 13th chapter of his book is dedicated to showing why his framework foresaw it, and what he did to warn of the coming crisis.

I have seen a few people saying that Keen didn’t really predict the crisis, and what predictions he did make were ‘chicken little’ predictions – repeating “there will be a crisis” until there was one. This is simply not true.

He certainly had the appropriate framework to foresee the financial crisis. His 1995 paper on Minsky and financial instability contains a model prone to endogenous fluctuations, and he concludes that any period of tranquility in a capitalist economy should not be accepted as anything other than a lull before the storm.

The key ingredient in Keen’s framework is, of course, private debt. Since banks create credit ‘out of nothing,’ new private debt adds to nominal aggregate demand. It follows from this that aggregate demand is current income plus the change in debt. I will quote Keen’s numerical example in full to explain why:

Consider an economy with a GDP of $1000 billion that is growing 10% per annum, where this is half due to inflation and half due to real growth, and which has a debt level of $1250 billion that is growing at 20% per annum. AD will therefore be $1250 billion: $1000 billion from GDP, and $250 billion from the increase in debt.

Imagine that the following year, GDP continues to grow at the same 10% rate, but debt growth slows down from 20% per annum to 10%. Demand from income will be $1,100 billion – 10% higher than the previous year – while demand from additional debt will be $150 billion.

Aggregate demand this year will therefore be $1250 billion – exactly the same as the year before. However, since inflation is running at 5%, that will mean a fall in output of 5% – a serious recession. So just a slowdown in the rate of growth of debt can be enough to trigger a recession.

For an economy to grow, either income must increase or private debt must increase at an increasing rate accelerate; this means that even a slowdown in the rate at which debt is decreasing can create a recovery (as with the US in 2010). The higher the level of private debt relative to income, the more dependent the economy becomes, and the more vulnerable it can be to even a mild slowdown in the rate of change of debt. Thus, in the mid 2000s, when Keen looked up the levels of private debt in developed economies, he was taken aback by the exponential increase:

At this point he went public – most of the evidence for his warning of a coming crisis is from the blog he started, and the monthly reports he released on there, tracking the level of private debt and explaining why it mattered. These reports first analysed the Australian, and then the US economy. He also spoke at a number of events, as well as a few TV and radio appearances which I cannot find online (although the media didn’t really start to take notice until the crisis began).

As a brief aside, I’ve seen a few people mention his failed prediction of the Australian housing crash, and his subsequent having to take a long hike. It is true that he got this one wrong, but there is quite an easy explanation: the government injected a large amount of money into the housing market in the form of first time buyer grants. Coupled with Australia’s resource boom, and the demand from China, this has kept their economy afloat so far.

So the charge that Keen did not predict the crisis, or simply shouted ‘there will be a crisis’ for 10 years until there was one, is false. He has a clear analytical framework that has performed incredibly strongly empirically, both before and throughout the crisis, and he got the dates approximately right (he said 2006). In my next post I will take a more in-depth look at his models and their implications for where we are now.


, , , ,

  1. #1 by Barry Ritholtz (@ritholtz) on October 3, 2012 - 6:32 pm

    Keen’s paper “Finance and economic breakdown” sounds like it owes an awful lot to Hyman Minsky’s work . . .

    • #2 by Unlearningecon on October 3, 2012 - 6:36 pm

      Absolutely, and I don’t think Keen would deny this (not that you’d necessarily imply that). He rarely claims originality for any of his work, but he does seem to be able to shout louder than the people whose work he builds on.

      In case you are interested, Minsky’s best and first book is the misleadingly named ‘John Maynard Keynes,’ where he states his instability hypothesis for the first time.

    • #3 by Roman P. on October 3, 2012 - 8:33 pm

      Well, Prof. Keen emphasized that his work is just a mathematical model of Minsky’s IFH…

      • #4 by pace on October 5, 2012 - 3:38 am

        I think Minsky stated the FIH before the JMK book. For example, in essays written in the 1960s.

    • #5 by Dave Holden. on October 4, 2012 - 12:17 pm

      Keen has named his software modelling package appropriately.. “ 😉

  2. #6 by Mathieu Dufresne on October 3, 2012 - 10:52 pm

    Just a little precision, private doesn’t necessarily have to increase in order to have a positive acceleration of debt. Take Keen’s quote you provide in reverse, if you’re deleveraging and the rate of deleveraging slows down, the acceleration of debt will turn positive, boosting aggregate demand. Therefore, the statement that debt need to increase at an increasing rate for the economy to grow is inexact, you can have growth while debt is still decreasing if the acceleration of debt turns positive.

  3. #8 by Pig Head on October 4, 2012 - 2:02 pm

    “AD will therefore be $1250 billion: $100 billion from GDP, and $250 billion from the increase in debt.”

    Is this just a basic math error (100 + 250 = 350, not 1250), a bad explanation, or indicative of bad thinking?

    To cut to the chase, how come an increase in debt creates AD that is somehow separate to GDP? Whatever value GDP is, it includes the demand created by any change in debt. “Demand from income” is not separate to “demand from additional debt” because spending funded by credit creates income.

    Does GDP not recognize all AD? If not, why not? If it does, then maybe Mr Keen should think of a better example – or maybe you should, as you seem to endorse what he’s saying.

    • #9 by Unlearningecon on October 4, 2012 - 2:15 pm

      It’s clearly a typo; the 100 is supposed to be 1000.

      And the disaggregation is just tool for understanding. You could also include ‘income from government borrowing’ and the like.

    • #10 by Mathieu Dufresne on October 4, 2012 - 2:56 pm

      @Pig Head

      Here’s a short presentation by Keen which addresses your point, it’s basically a confusion between ex-ante and ex-post.

  4. #11 by Pig Head on October 4, 2012 - 2:53 pm

    RE: your reply above. It’s not just a typo, it’s wrong however you interpret it e.g. if the change in debt has been 250 and the total is now 1250, then debt must have grown at 25%, not 20%.

    It’s also not disaggregation. He’s adding the change in debt to income – some of which was generated by the change in debt.

    The bottom line: if GDP is still growing by 10%, then it’s still growing by 10% regardless of any change in debt.

    • #12 by Unlearningecon on October 4, 2012 - 2:59 pm

      Pig Head: the change in debt is 250 from 1250. That is 20%, then in the following paragraph a 10% change is 150.

      And I don’t see the problem. Income grows at 10%; debt-financed income grows at a different rate. It’s just disaggregation (Mathieu has a good link above).

  5. #13 by Pig Head on October 4, 2012 - 3:50 pm

    I don’t want to argue the math for ever. My point is simply that it is a badly constructed example, which doesn’t add up. That is, he says:

    “Consider an economy with a GDP of $1000 billion that is growing 10% per annum… and which has a debt level of $1250 billion that is growing at 20% per annum. AD will therefore be $1250 billion…”

    So, at the end of the year when GDP has been 1000, debt is 1250. Now, I believe that he states AD to be 1250 for the same period. You, on the other hand, seem to believe he is talking about AD the year after, when debt grows from 1250 to 1500. Either way, he is wrong.

    In my interpretation, the growth rate of debt has been 25%, not 20% as explained above. In your interpretation, AD (by his formulation) should be 1100 + 250 = 1350, because if debt has grown 250 to 1500, as you claim, then GDP would have grown by 10% in the same period.

    Perhaps this illustrates why it is never a good idea to mix up stocks and flows in the same sentence. Anyway, it’s a bad example and explanation.

    Thanks for the link to the presentation. I will look at it.

    • #14 by Mathieu Dufresne on October 4, 2012 - 5:53 pm

      The example is accurate, aggregate demand for a given period will be income from the previous plus the change in debt during the current period. Keep in mind all this happens in continuous time and thinking in terms of periods can be confusing but it’s probably the best way to explain the concept to a wide audience since thinking in continuous time is too abstract and throwing integrals at people doesn’t help either. The goal is to explain how negative acceleration of debt can lead to a fall in aggregate demand when the change in debt is still positive.

      • #15 by Pig Head on October 4, 2012 - 7:25 pm

        Mathieu, thank you. You may be right, but if you are, then the author of this site was wrong to say that:

        “aggregate demand is current income plus the change in debt”

        By your account, he should have said “last period income” not current income. Taking last period income makes the numbers work (but makes less logical sense). Taking current period income is probably more logical, but make Keen’s numbers inconsistent.

        It beats me why, if he wants to take last period’s income, then he doesn’t take last period’s change in debt too. That would at least be consistent. Why take one number from one period and the other number from a different period?

        Keen’s site has been down so haven’t had a chance to look at your link, but I will now. Here’s hoping he explains it well. If not, I’ll be back to ask you for an explanation!

      • #16 by Mathieu Dufresne on October 4, 2012 - 7:44 pm

        It’s accurate to say current income in the sense that it’s the recorded income at time t (ex-post) plus the change in debt at time t. Here you see the difficulties to express a process happening in continuous time in term of periods. I think you may understand better why aggregate demand is current income (ex-post) plus the change in debt after you saw the presentation. In any case, the only way to be truely mathematically accurate is to work out the differential equations, periods are only mental representations that helps understanding the concept.

      • #17 by Unlearningecon on October 4, 2012 - 8:07 pm

        Yes, it is hard to express it verbally without losing some accuracy. With the differential equations you can have the rate of change of debt plus current income as a time path which makes things clearer.

  6. #18 by Pig Head on October 4, 2012 - 9:06 pm

    Okay, the video was interesting but went too quick for my tiny brain to really think thru the logic. I’ll go back when I’m less tired. However, regardless of the logic or illogic involved in that argument, the following claims that have been made are false:

    1. Unlearningecon:

    “For an economy to grow, private debt must accelerate”

    2. Minsky via Keen:

    “For real AD to be increasing … it is necessary that current spending plans … be greater than current received income”

    Neither claim is true. As has always been the case, an economy (and AD) can grow simply thru increases in productivity or workforce. Think of a simple barter economy. Doubling productivity would double real GDP and AD. Doubling up the workforce would do the same. Neither would require an increase in debt (let alone acceleration) or that spending plans exceed income.

    • #19 by Mathieu Dufresne on October 4, 2012 - 10:36 pm

      I’m not sure what you’re trying to argue, that increasing output doesn’t require increasing investements and that doubling the workforce won’t increase the wage bill?

      • #20 by Pig Head on October 4, 2012 - 11:37 pm

        Productivity gains may or may not require investment. But investment can be funded by saving.

        Doubling the workforce may involve increasing the wage bill, but there is no necessity for this to be planned, or indeed for any plans to involve an excess of spending over income.

        All you need to do is to consider a simple credit free, barter economy to see that both claims are false. In such an economy there is no debt and it is impossible to plan for an excess of spending over income, yet such an economy can still expand. Therefore, both claims are false.

      • #21 by Unlearningecon on October 5, 2012 - 8:48 am

        All you need to do is to consider a simple credit free, barter economy to see that both claims are false. In such an economy there is no debt and it is impossible to plan for an excess of spending over income, yet such an economy can still expand. Therefore, both claims are false.

        But that’s not the economy we live in!

        We are talking about nominal growth in a credit economy, not real growth in barter one (or, indeed, a credit one).

        Having said that, a lack of credit and deflation would certainly cause problems for real growth, too.

      • #22 by Mathieu Dufresne on October 5, 2012 - 12:21 am

        This have been thought through at lenght by Schumpeter btw. We live in a capitalist monetary economy and the growth rate of new guinean tribes barter economies is anemic.

        Profits and wages are spent on consumption and if you start retaining more profits to increase investements, demand for goods and services will need to fall prior to the increase in investement. In addition, raising output without increasing the cash flows from which this output is purchased will lead to deflation. If firms need to employ more workers, thus increasing the wage bill, and their revenues doesn’t rise they will end up investing more to lower their share of output.

      • #23 by Pig Head on October 5, 2012 - 8:50 am

        It doesn’t matter if the growth rate of a barter economy is less than a modern capitalist one or not. The point is that the example of the barter economy clearly shows that the 2 claims are false. I don’t see you disputing that logic.

        Having established that fact, we can then show that the same 2 effects that I discussed (productivity gains and an increase in workforce) apply in a modern economy too. So, the 2 claims are false in a modern economy too.

        A change in debt may add to, or subtract from demand, but the 2 claims, as stated, are false.

      • #24 by Unlearningecon on October 5, 2012 - 9:02 am

        AD can expand due to both an increase in income and an increase in the change in debt.

        Btw, I have clarified the post.

  7. #25 by Pig Head on October 5, 2012 - 9:34 am

    Right, I’ve now looked at the video fresh and, to be honest, I’m shocked that Keen would put out this stuff. I respect him (and the other figures involved) but his “mathematical proof” is nothing of the kind.

    To be brief, at 3:54 in the video, he ASSUMES that “there are 2 sources of demand” where one of these is the change in debt, and then goes on to “prove” that AD = income plus change in debt. Of course he has proved nothing.

    Look, let’s get back to basics. If we assume that the economy is in equilibrium, then:

    AD = AS = GDP = Output = Income

    So for ANY time period, it is ALWAYS true that AD = Income.

    Therefore, for any period in which the change of debt is non zero, it is false to say that AD = income plus change in debt.

    The underlying reality is that a change in debt can ONLY affect demand if the increase in credit is spent. Then, the INSTANT that it is spent, it is recognized as income.

    So, while it may be true that a change in debt affects demand, it only does this as a result of the funds being spent, and the spending represents income. So it is false to claim that AD = income plus change in debt.

    • #26 by Unlearningecon on October 5, 2012 - 10:04 am

      Keen has plenty of stuff showing why AD = income plus the change in debt, for that video he assumes it.

      You berate him for assuming his conclusions then assume the economy is in equilibrium! But it never is.

      In any case the rest is semantics. As I said, you can disaggregate AD to try and highlight what’s going on. That’s all Keen does.

      • #27 by Pig Head on October 5, 2012 - 10:32 am

        My assumption that the economy is in equilibrium is a logical device. I show that Keen’s argument is false IF the economy is in equilibrium. You do not dispute that. If you do, where is your argument???

        If you then want to claim that Keen’s argument is true when the economy is not in equilibrium, then that is a different argument. But the key point is that is NOT Keen’s argument. If Keen wanted to assume that the economy is not in equilibrium, then he could not use the identities that he does (because they only apply in equilibrium), so his argument would again be false.

        If you had a decent argument you would set it out, or disprove mine. Clearly you don’t have any decent argument.

      • #28 by Unlearningecon on October 5, 2012 - 11:14 am

        Whether your argument is true in equilibrium or not is irrelevant because the economy isn’t in equilibrium. I am not interested in discussing how many angels can dance on a pinhead. Keen’s arguments do not assume equilibrium so I don’t know what you’re getting at – he uses differential equations =/= equilibrium.

        Here are the causal mechanics for you – it’s pretty simple, if you read Keen with an open mind instead of an eye for proving he is an idiot then you might agree.

        In Keen’s framework (and the real world) banks create debt out of nothing, i.e. a loan (asset) and deposit (liability). This is created as debt, and adds to purchasing power. Say £10 of debt is created one year. In order for AD to remain constant (assuming other income constant), £10 of debt must be created the next year, too. For AD to increase, the debt must increase by more. Hence, debt must increase by more and more to add to purchasing power.

        Nice story. If this were true, we’d expect to see robust correlations between the acceleration of private debt and growth, other variables etc. Guess what? We do!

      • #29 by Pig Head on October 5, 2012 - 10:37 am

        p.s. your link does not “show” that AD = income plus change in debt. How on earth can you claim that it does?

        In case you don’t understand this simple point, the fact that changes in debt can affect demand, is a very different point to claiming that AD = income plus change in debt.

      • #30 by Nathanael on October 21, 2012 - 8:42 am

        Pig Head, the simplest way to explain this is basic accounting. Aggregate demand is determined by cash flow, because only cash can be spent (this is the definition of cash for our purposes — something which can be spent in this fashion). This is equal to income, plus change in debt. (Or, minus change in savings, if the economy is massively balanced in the other drection — much the same thing.) I suppose the equation could be wrong if income was in an illiquid form — how often does that happen?

    • #31 by Mathieu Dufresne on October 5, 2012 - 11:35 am

      You’re still confusing ex-post and ex-ante. The mathematical proof doesn’t prove that the change in debt adds to demand, it shows there is no logical contradiction in saying that ex-ante AD = income plus the change in debt and that ex-post AD = income.

      If I go to a bank and ask for a loan in order to buy a house, my current spending plan is greater than my income. If that loan doesn’t come from someone else deposit but from an expansion of bank balance sheet, the change in debt necessarily need to add to AD. This will add to income once spent and recorded income will be equal to recorded AD ex-post. If you’re using dynamic modeling, you’re working ex-ante and that what’s relevant to understand the dynamics involved in the business cycle.

      As it has been explained by Shumpeter and Minsky, if the economy is to grow, income must be increasing, unless you assume productivity gains doesn’t require investement and labor is free. If income is to grow, it is neccessary that current spending plans is greater than current income. Yes, fluctuations in the turnover rate of the money stock will bring fluctuations in income and in economic growth but longer term trend will be relatively flat.

      • #32 by Pig Head on October 5, 2012 - 8:10 pm

        Mathieu, you do a good job of explaining your argument. Thank you for that. However, I don’t believe that I am confused in the way you suggest. I understand your distinction between ex-post and ex-ante.

        My point is that Keen offers no PROOF that even ex-ante AD = income plus change in debt (and in his “mathematical proof” makes no distinction between ex-ante and ex-post). As I said many times above, I have no problem accepting that a change in debt will impact demand, but this is very different from a claim that a specific equation is proven. Nothing has been proven, only assumed.

        Indeed it seems impossible for Keen to prove any ex-ante relationship, and if the full facts were known, we would likely find the relationship more complex than Keen suggests. For example, what about changes in the saving rate, or precautionary debt used as working capital?

        These and other factors make the ex-ante relationship very complex. All we know for sure is that ex-post

        AD = Income

        If Keen stuck to saying that the change in debt is one of many factors that impact AD, then I would agree with him. But he doesn’t say that.

      • #33 by Mathieu Dufresne on October 5, 2012 - 8:23 pm

        I completely agree that changing the saving rate will have an impact on AD. If you come to understand Keen’s dynamic model, you will see it it’s really easy to take that into account. For example, if you want to increase the saving rate of workers you only have to slow down the turnover rate of workers account.

        When working with empirical datas, the strong correlations between the credit accelerator and other datas strongly suggest that the change in debt is the main factor causing AD volatility. Moreover, if you accept endogenous money and you’re working from an ex-ante perspective, you have no choice but to consider the change in debt adds to AD.

      • #34 by Pig Head on October 5, 2012 - 10:18 pm

        But if you accept that changes in the saving rate will change AD, then you must accept that AD is not equal to Income plus change in debt, even on an ex-ante basis (a change in saving rate is not necessarily reflected in the change in debt).

        You can argue that this formula is a first approximation, but that is all. It’s certainly not something that is subject to proof.

      • #35 by Mathieu Dufresne on October 5, 2012 - 10:44 pm

        The proof is only proving that you’re not double counting if you add the change in debt to AD. I think it’s accurate to say that considering AD=income plus the change in debt ex-ante implicitly assume that the turnover rates are fixed. I don’t think this is a tautological identity (that’s why it’s useful) and a lot of other factors could be added to the equation. However, this is a high level of abstraction and I have never thought about that before so don’t hold me to it, I didn’t got my head around yet and probably won’t anytime soon.

  8. #36 by lyonwiss on October 5, 2012 - 10:49 am

    Keen saw from data the rapid growth of debt before the Great Depression and also noticed the rapid growth of debt in recent years and concluded another depression is imminent. He did the same sort of thing to predict a crash in Australian house prices, which has yet to happen. There is little more to it than that, an observation of correlation.

    The Minsky models had little to do with his public predictions. But most people make these connections (models and predictions) themselves and assume rigorous methods for his forecasts, without justification. Otherwise, ask him to show the modelling results produced BEFORE or at the time of his public predictions. If anything at all, it would be highly idealized and simplified models, without empirical calibration from data, showing the possibility (not high probability) of crashes. Only such evidence of predictive modelling output would convince me that there is anything more to it.

    The Revere award was merely a publicity stunt to dethrone mainstream economics, which deserves to be consigned to the dustbin of history. But please do not replace false prophets with others. Economics is still a long way from a respectable science.

    • #37 by Unlearningecon on October 5, 2012 - 11:06 am

      So he noted a strong correlation, had a broad but incomplete framework that suggested why it might matter, and when his predictions came true he started on developing a more rigorous model?

      That honestly doesn’t seem like a problem to me.

      • #38 by lyonwiss on October 5, 2012 - 11:32 pm

        It is only a problem if you think he has the right answer even before it has been proved.

      • #39 by Unlearningecon on October 7, 2012 - 2:07 pm

        Well models cannot be proved, only falsified. I am unaware of substantial data that contradicts Keen’s theories.

  9. #40 by vimothy on October 5, 2012 - 12:13 pm

    I discovered this argument by Keen recently and I have to say that I don’t think it makes the least bit of sense.

    If you start with Keen’s example, forget about the change in growth rates, and just try to take it on its own terms you get,

    Nominal income in the next period = 1.1 * $1000bn = $1100bn
    Debt in the next period = 1.2 * $1250bn = $1500bn
    AD in the next period = GDP + Change in debt = $1100bn + $250bn = $1350bn

    But the economy itself only grew by 10% in nominal terms. That means aggregate demand was $250bn greater than nominal income. What is aggregate demand? Well, theoretically, it’s a function, but what’s realised is just aggregate nominal expenditure. And what’s that equal to by definition? Aggregate expenditure is equal to aggregate income. So $250bn has gone missing in the aethyr. It was used, contributed to aggregate demand, but no one received it as payment for anything.

    What Keen wants to say is that anyone has the following sources of funds: income and borrowings (or savings). Therefore, aggregate expenditure is aggregate income plus new borrowing. That’s fine for the individual but it’s not fine for the whole economy. (Keen is committing a fallacy of composition.) For the whole economy, expenditure can only exceed income if it is borrowing from the foreign sector, because borrowing and lending net out within the economy when you aggregate up. Consider it in terms of sectors. For the government sector, expenditure can only exceed income if it borrows from another sector (by running a deficit). Then, when we aggregate the government and the rest of the economy, we find that the government has expenditure in excess of income matched exactly by the non-government balance, which shows income in excess of expenditure. These two balances cancel out, so that income equals expenditure.

    • #41 by Unlearningecon on October 5, 2012 - 5:18 pm

      No he is not committing a fallacy of composition, he has an entirely different view of how the banking system works, one in which debt is created out of nowhere. This is the missing $250bn: banks create it out of nothing.

      Now you can define that as income but as Mathieu has been stressing above, it’s an issue of which periods we look at. It’s also much easier to see it from the perspective of differential equations.

      • #42 by vimothy on October 5, 2012 - 5:53 pm

        Entirely different view than what? When Apple issues bonds, those bonds don’t “come from somewhere”. They just appear out of nowhere. In the same way, when a bank issues liabilities, those liabilities don’t “come from somewhere”. The bank just issues them. They don’t exist before that event takes place.

        Banks have a balance sheet that has assets and liabilities. If there is lending on one side, there is saving on the other. That’s just the way it is on this planet.

        It’s a bit like thinking that you can become wealthy by lending money to yourself. Imagine that you lend yourself £1 million. Now you have £1 million in assets! Of course, you also have £1 million in liabilities, so you’re not actually better off. Still got that £1 million in assets though! Keen’s argument is similar. He thinks that the economy can spend its income and that it can also borrow money from itself. But of course, if it borrows from itself, then it lends to itself as well. Its net borrowing will be zero—unless it borrows from some other country. In that case, its expenditure will be higher than its income, because it borrowing the income of someone else.

      • #43 by Unlearningecon on October 5, 2012 - 5:57 pm

        Banks create assets and liabilities simultaneously when they make a loan, and M1 expands. The loan (asset) creates a deposit (liability) in the recipient’s account. When the loan is paid back the asset and liability both go to zero and the money is effectively ‘destroyed.’

        Banks need reserves but this is a secondary consideration and the quantity available does not constrain their lending decisions.

        Seriously, ask a banker, or financial lawyer, or accountant, or anyone who works in the real world. The ‘banks lend out savings’ view is completely misguided.

      • #44 by vimothy on October 5, 2012 - 6:02 pm

        In any case, the problem is not where the $250 billion comes from but where it goes to. If $1350bn is spent, then $1350bn should also be earnt by the recipients of that spending. Instead, $1100bn is earnt and $250bn vanishes into the air.

      • #45 by vimothy on October 5, 2012 - 6:05 pm

        I know that when a bank makes a loan it creates an asset and a liability simultaneously. This doesn’t make what Keen is claiming any more reasonable. How can a country spend more than it earns by lending money to itself?

      • #46 by Unlearningecon on October 5, 2012 - 6:20 pm

        As I have said before, ultimately the new purchasing power is sourced from the CB, though only in name. printing new money = new nominal demand, though it is privately expanded and contracted.

      • #47 by Mathieu Dufresne on October 5, 2012 - 6:35 pm

        If you start with a simple Schumpetarian model, income and the change in debt is spent only on goods and services and investments. In that case, all spending will add to income. When you add Minsky’s extension of Schumpeter’s work, income and the change in debt is also spent on existing assets. It’s empirically obvious that all expenditures are not accounted for in GDP when you consider expenditure on existing assets, for example, the amount of transactions on the stock market in one year greatly exceed GDP. Nevertheless, this demand need to be considered since it has an impact on asset prices. That’s why Keen comes up with AD = income plus the change in debt and AS = output plus net turnover of assets. Consequently, if the change in debt is spent on existing assets, it doesn’t necessarily turns up in GDP but will be absorbed by the asset markets.

      • #48 by vimothy on October 5, 2012 - 6:37 pm

        Look, it’s fallacy of composition. For you, as an individual, your spending will be your income plus your borrowing. That’s fine. For the whole economy, income and expenditure are equal. That’s because income is earnt by selling goods to consumers, the government and firms. What I spend becomes someone else’s income. It doesn’t matter how banks fund their loan books. There’s no way that money can be spent without someone receiving that spending.

      • #49 by Unlearningecon on October 7, 2012 - 1:25 pm

        That’s absolutely, true, and as we’ve been saying it’s an issue of disaggregation and time periods. We could similarly talk about how much government borrowing adds to income, and perhaps it might tell us something useful.

      • #50 by vimothy on October 6, 2012 - 9:15 am


        All of these terms have specific definitions. Expenditure is outlays on durable and non-durable final consumption goods and investment. Buying stocks and shares in Microsoft is not part of the income-expenditure flow. It’s portfolio allocation. The same framework says that for any entity or sector, saving plus borrowing equals asset acquisition.

        If you want to reinvent the wheel here, you’re eiher going to come up with the wheel, which actually works, in which case you might as well just use the definitions from the national accounts like everyone else, or you’re going to come up with a square.

  10. #51 by Derek R on October 5, 2012 - 7:54 pm

    Vimothy, there are two ways for you as an individual to get money. The first is that someone else gives it to you in exchange for something. The second is for you to counterfeit new notes or coins. The first method is legal for an individual and the second method is not.

    But that is not the interesting thing. The really interesting point is that in both cases the total amount of spending in the economy matches the total amount of income, since every note that you received as income was someone else’s spending and every note that you spent (including the new ones that you made) became someone else’s income. So in the aggregate total income will equal total spending for both cases. However in the latter case more money was added to the economy, and the aggregate demand is higher, since you were able to buy more in the illegal case than you could in the legal case. This extra money did not come from foreign sources: it came from you.

    Now just remember that while it is illegal for you to make your own money, it is not illegal for the government to do so, or for the banks within limits, and you may have a better chance of seeing how more money can be added to the economy even though aggregate income = aggregate spending at all times.

    I’ve just showed how it can be done by printing money. Steve Keen shows how it can be down by creating loan agreements.

  11. #52 by Eric L on October 6, 2012 - 6:43 am

    Debt isn’t the only way for nominal gdp to grow. Any increase in saving is subtracted from aggregate demand and any drawdown in savings is added to it. But it’s weirder than this, because there must necessarily always exist savings that can be drawn down to cause growth, and there need not be a net drawdown in savings to cause growth. Having cash in your pocket gives you the ability to plan to spend in excess of what you expect to make without using debt. But spending that cash does not put a limit on how anyone else may spend their income or savings, and it also does not result in a net draw down in savings — no matter how much is spent, just as much is available in the economy, continuing to provide the ability for someone to spend in excess of their income.

    It’s weird, Keen has been vocal about the fact that economists need to be more dynamic in their modeling of the economy, yet here he is making a strong erroneous claim based on static thinking. Growth is possible, period. Debt is *a* way that it can happen, not *the* way.

    • #53 by Unlearningecon on October 7, 2012 - 1:30 pm

      I don’t think Keen claims debt is the only way growth can happen. Any confusion about that may be because of my poor wording.

  1. 10 Mid-Week PM Reads | The Big Picture