The key failing in Krugman’s response to Steve Keen’s response to Krugman’s paper (PDF) is here:
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.
Krugman assumes here that people have to save (spend less) in order for other people to borrow. It’s actually the fundamental assumption, the sine qua non, of his paper (and of Krugman’s beloved IS-LM — the linch-pin of “New” Keynesianism — created by Hicks to subsume Keynes into neoclassicism, and later disclaimed and discredited by Hicks as a “classroom gadget”; see my post, and Phillip Pilkington here).
But that’s not how things work (and it’s the very assumption that Keen is disputing). I tried to explain this in clear and simple terms here:
Think about it:
You get $100,000 in wages. Your employers’ bank account is debited, and yours is credited. Your bank can lend against your higher balance; your employer’s bank can’t. Net zero.*
You spend $75,000. It’s transferred from your account to other people’s/businesses’ bank accounts. Their banks can lend more, yours can lend less.
Is the total stock of loanable funds affected by whether the money is on deposit at your bank, your employer’s bank, or the banks of people you bought stuff from? No.
Meantime, you don’t spend $25,000. You “save” it. The money sits there in your checking account. If the action of spending — transferring money from one account to another — doesn’t change the total stock, how could not transferring money do so? Your bank still has the money, which it can lend out. Other banks still don’t, and can’t.
So here’s how the argument plays out:
Krugman assumes that people need to save in order for others to borrow.
Keen points out that they don’t.
Krugman explains that Keen is wrong by … assuming that people need to save in order for others to borrow.
And so the world goes round.
Cross-posted at Angry Bear.
Comments
27 responses to “The Central Flaw in Krugman’s Argument Against Keen”
and yet people do decide to save, and those savings do end up financing loans, don’t they?
and if somewhere somebody is deciding to save-not-spend, and somebody else is deciding to borrow-and-spend, what has happened on net to aggregate demand?
I bought some UK national savings certificates the other day – instead of spending that money on goods and services, I lent it to the government who will spend it on goods and services, or transfer it to people who will via welfare/pensions payments.
why do banks spend money trying to attract savers? what do they need them for? Why am I getting paid interest on my savings account?
I’ve spent no time on this at all.
But here’s a thought off the top. What if they’re both wrong, and both right in, their own important ways?
In theory, aggregate demand can be boosted using the catalyst of an acceleration in the velocity of existing bank money. That doesn’t require new bank loans or new money. So in that sense Krugman is right. But that’s a pretty limited idea in an empirical sense.
And it only really works for expenditure on consumer goods anyway. Any incremental net business investment requires the issuance of new financial claims. There must be a balance sheet representation of the new business wealth that’s been created. The intermediation circuit for that tends to start with new bank loans and progress to a substitution for that with debt and equity. So in that sense Keen is right in theory and empirically.
The fact that the US has progressed to the point of $ 60 trillion in household net worth has required the issuance of some financial claims, after all.
from the point of view of an individual bank, it make loans and these are its assets. On the other side of its balance sheet is what it has borrowed from people – some deposits taken, some other forms of borrowing, some equity – its liabilities.
when you deny that people have to save (spend less) in order for other people to borrow, that almost sounds to me like you are claiming banks could lend (create assets) without also having to borrow (create liabilities). But I don’t think that can be what you mean.
Individual banks do need people to save (including current account – checking – balances) in order to finance their loans, don’t they? sure, they can lend first and finance later – they don’t turn away borrowers because they haven’t first got the depositor’s funds to lend – but they must finance their lending somehow.
What about the system as a whole? What constrains lending there? Well in once sense there is an accounting relationship between the quantity of high powered money, bank reserves, lending, and the total money supply (as described by the text book money multiplier). But does more borrowing require more saving?
In your example, from the point of view of the system as a whole, it doesn’t matter how much of my $100,000 I decide to spend and how much I decide to spend, it’s all there in the banking system anyway, in somebody’s account somewhere.
So what happens if everybody in the economy decides to save zero? I get paid monthly, so let’s say the average balance of my account over the month is zero, say I have a positive balance for the first half of the month, and borrow for the second half. And say everybody else with any sort of income does the same. And say all firms do the same too – they collect revenues, pay wages and other inputs, and each runs an average bank balance of zero. None are net savers or borrowers.
Now, in this world, what happens if somebody wants to borrow a ton of money to finance a new investment project? Doesn’t somebody have to become a net saver before that can happen?
[n.b. please take these comments in the spirit of somebody trying to understand the topic, not a “opponent” trying to tell you you are wrong – I’ll be quite happy if you can explain to me where I’ve got it wrong]
JKH
I don’t understand your point about business investment, but on the topic of whether lending creates aggregate demand and both Keen and Krugman being correct in their own ways … doesn’t credit expansion amount to monetary expansion? And wouldn’t Krugman agree that monetary expansion boosts aggregate demand? It does in the IS/LM and New Keynesian models he uses, as far as I recall.
What I mean is … suppose you hold constant the base money supply, but expand the money multiplier – so the gross amount of borrowing/lending increases. That increases the money supply and hence increases AD. What it does to real output is another question, involving inflation and the supply side. Now suppose the base money supply increases in response to any increases in the gross amount of borrowing/lending … well then we have credit expansion increasing AD too.
So that sounds like Keen is correct – credit expansion increases AD. At the same time, within that, you have some people borrowing more and others saving more – as Krugman says – but this fact doesn’t mean there can be no increase in AD because the pie – the quantity of money to be lent and borrowed – has increased. How much has AD increased? If gross lending rises from £100bn one year to £200bn the next, does that mean AD has increased by £100bn? I can’t immediately see the answer. Is Keen claiming that increasing debt creates a one-for-one increase in AD, and is that what Krugman is disputing?
What I need to think about is whether you have have an increase in gross debt, without also having an increase in the money supply. If you can, then the pie needn’t grow and maybe somebody’s borrowing is offset, w.r.t. AD, by somebody else’s saving. OTOH, Krugman does talk about lending raising AD by transferring resources to people with a higher propensity to spend, and if that’s happening, but the money pie isn’t growing, then I guess we must have velocity of money increasing .. Krugman isn’t denying lending can increase AD.
@Luis Enrique “In your example, from the point of view of the system as a whole, it doesn’t matter how much of my $100,000 I decide to spend and how much I decide to spend, it’s all there in the banking system anyway, in somebody’s account somewhere.”
From the perspective of “loanable funds”, exactly right. From the perspective of *spending and income* — the true source of the residual, saving — it makes all the difference in the world.
IOW (see the subtitle of my post linked above), it’s the velocity, stupid. Or maybe better, it’s the income stupid, not the saving. (This is one thing I take away from the JKH Tautology.) More saving/less spending is *reduced* velocity, reduced GDP, reduced income. (This is distinct from saving as a *proportion* of income, and also from investment as a proportion of total spending.)
Do go back and check the linked post. It gets somewhat rambling at the end, but it puts across much of my current thinking on all this.
http://www.asymptosis.com/no-saving-does-not-increase-the-supply-of-loanable-funds.html
or maybe I’m wrong to think that increasing the money supply increases AD – and wrong that Krugman would think that (and that IS/LM or NK models say that) – after all he wrote “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.”
I don’t really understand what Krugman is getting at there. It seems to me that if the quantity of money increases b/c of credit expansion that involves higher AD.
@Luis Enrique “Is Keen claiming that increasing debt creates a one-for-one increase in AD, and is that what Krugman is disputing?”
Yeah that’s part of what’s going on. Keen does claim that, and Krugman dismisses the claim. But Keen also understands what I think Krugman doesn’t: that increased demand does not necessarily result in increased price or quantity of real goods (things that must be produced, and that are consumed). It can “leak” into increased prices for financial assets, which are not produced except in small quantities at the margins (with essentially zero production costs), and are not, cannot be, consumed.
That understanding puts paid to many of the notions built upon PV=MY.
@Asymptosis
thanks – I will have a read. so my little story asking how there would be any loanable funds in a world where everybody saved zero hasn’t moved you?
@Asymptosis
ok, so roughly speaking people have more money (because credit expansion has increased the money pie) but people don’t want to spend (all of) their money on real goods, so they buy financial assets whose prices rise …. but for everybody buying a financial asset isn’t there somebody selling it, and what do the sellers do with the money they receive if not cause prices or quantities of real goods to rise? I guess you have a bunch of financial assets swapping hands at ever increasing prices, but how does that somehow “use up” some of the AD and divert it away from real goods?
@Asymptosis
Yes definite misinterpretation of Keen by Krugman.
Keen is talking of an autonomous increase in demand via borrowing ceteris paribus. Krugman seems to dismiss this claim by thinking of a situation in which ceteris is not paribus! Krugman dismisses Keen by saying that an autonomous increase in demand created by borrowing can be coincident with a decrease in propensity to consume by someone else.
All this cannot be modelled without the banking system.
I saw Keen saying never has been misunderstood felt so good after getting a birthday gift from Krugman.
JKH :
Hey JKH. As so often, I’m feeling dull-witted, like I don’t understand some conceptual jumps you’re making, jumps that would be obvious to the more perspicacious. Your questions feel kind of orthogonal or tangential to my understandings. So please excuse me if my responses are likewise to yours.
This may put us in one of our snake pits or tubs full of tangled spaghetti, but that’s okay because what sometimes emerges from those is is one of those Aha moments like the one embodied in what I’m happy to coin the term for: The JKH Tautology. 😉 (Hey what a great title for a novel!)
Yes that could certainly be possible. But I question that symmetry because:
1. Krugman is not addressing Keen at the fundamental level that Keen is addressing Krugman.
2. I think Keen is right (though he doesn’t say this in so many words) that the loanable funds model doesn’t make any (useful) sense. (See #1.)
Yes. The ridiculous assumption of fixed velocity is the hugest failing of paleo-monetarism, IMO.
But key logical flaw: just because AD can/does change without new loans does not mean that new loans don’t affect AD. They very well might.
Keen’s shown some darned good empirical evidence (see his posts on the Credit Accelerator) to suggest that they do. (Though he’s also posted a great guest post by Lyonwiss showing that his correlations don’t demonstrate an absolute or steel-rod kind of relationship: http://www.debtdeflation.com/blogs/2011/07/01/credit-accelerator-leads-and-lags/) He also understands that increased demand can leak into higher financial asset prices (which may circulate/increase velocity to some extent via a second-order wealth effect) rather than driving up prices or quantity (supply) of real consumable goods. I don’t think Krugman grasps that, or the resulting conclusion that PV≠MY. He’s thinking inside the NIPAs — all as if the FOFs and the financial assets and markets represented therein did not exist, or as if they constitute a huge, neutral black box.
Limited at best, IMO. I think for the logical flaw stated above makes him wrong.
Really? Imagine undistributed business profits spent on fixed assets, with the firm value unchanged, both book and market. Increased investment, but no increase in equity claims by households. Fixed stock is increased (because presumably the purchases of fixed assets spurred increased production to replenish inventory of fixed-stock producers), and deposits have shifted between firms. Thassit.
That’s a different scenario. I would suggest that new business wealth is really about sources of funds — income/profits — not the uses of funds. It’s the income, stupid, not the residual, saving(s). This is actually where I go, conceptually, from the JKH Tautology. (Hoping you can follow my conceptual leaps better than I can follow yours…)
We’re into behavioral and causational assertions, here, I think, and I think the circuit goes from optimism (based mostly on good and promising current conditions) to everything else: spending, producing, borrowing, lending.
Another conceptual leap: I think that all results from increased income, which results from the surplus from trade: the sell/buy nexus where the value-added of production gets monetized.
Semi-aside: I’m also starting to think about Net Financial Assets (the portion of household net worth in excess of the value of fixed assets) as something like a border area, a zone where emerging real value (embodied in higher incomes) gets monetized/financialized. Or as a buffer, much like bank reserves, that provides the financial flexibility/slack for that monetization to happen. As opposed the leverage thinking we’ve discussed in the past. Still vague, but I’m cottoning to the notion.
I haven’t been posting much lately because I keep trying to synthesize all this into my own personal Grand Theory of Everything, and absent that I feel hesitant. A fool’s errand, I know, and excuse me for inflicting undigested dribs and drabs of it on you.
Btw, I think there are some definitions Keen invents which I guess caught Krugman’s attention. He has a strange habit of adding change in debt here and there.
@Asymptosis
Steve,
Sorry, I just dropped by without being able to put much thought or effort into this, due to other diversions lately. You’re probably right about the asymmetry. I see humorous parallels between Krugman and Sumner regarding the denial of the importance of banking here. ISLM and loanable funds remains an enigma. (I wonder if it works in a model where the time period is limited to an instant epsilon.) I’m not entirely comfortable with the Keen model, but that’s another story too. Something about the definition of aggregate demand that seems too crude for me. Re undistributed profits and investment – I’m thinking of increases in equity book value as being an increase in the value of the equity claim issued, which then gets valued at market in total. It’s equivalent to the reinvestment of a dividend. My intermediation circuit is slang for the propensity of businesses to start their finance with the most liquid form of liability – a bank loan and then to repay the bank loan later by issuing debt or equity (or undistributed profit).
Blah!
@JKH “ISLM and loanable funds remains an enigma. (I wonder if it works in a model where the time period is limited to an instant epsilon.)”
Exactly. Keen has most of a chapter on this, how many (infinite) increments are not the same as a continuum, can be wildly different, arithmetically. Hence his obsession with dynamic modeling vs. comparative statics.
“Something about the definition of aggregate demand that seems too crude for me.”
Yeah it doesn’t ring completely true to me either. But I think that may have to do with the aforementioned “leakage” out of the real-goods market into financial asset prices.
“I’m thinking of increases in equity book value as being an increase in the value of the equity claim issued, which then gets valued at market in total.”
Right, I was just saying that investment does not necessarily, in and of itself, increase either book or market value *within the period,* either for individual firms or in aggregate. Just swapping dollars for drill presses.
“the propensity of businesses to start their finance with the most liquid form of liability – a bank loan and then to repay the bank loan later by issuing debt or equity (or undistributed profit)”
Right I totally get that transformation, but I think the most liquid form of finance companies have available is their own undistributed profits. This is thinking about companies like Apple, of course, as opposed to private-equity-owned shops that are leveraged to the limit (with all their borrowed cash sucked out buy squids), who can only invest by borrowing even more.
@Asymptosis
“I was just saying that investment does not necessarily, in and of itself, increase either book or market value *within the period,* either for individual firms or in aggregate. Just swapping dollars for drill presses.”
Ah – you stray from the asymptote of ideal illumination – OK for a dollar swap for new investment goods, but not for the original production of new investment goods, which produces income and saving. Net investment as a component of GDP definitely increases book value, and should logically increase market value, other things equal (ignoring the fact that things are rarely logical or equal).
🙂
@Luis Enrique
That is exaclty, what you get wrong.
In reality, when a bank makes a loan it creates simultaneously an asset – the loan itself and a liability – the entry on the deposit account of the borrower. For this step, it’s irrelevenat, whether it has other deposits or anything else. That is how banks create money “out of thin air”.
It’s important to understand however, that this created money isn’t central bank money, it’s private bank money and strictly speaking it isn’t legal tender. But all of us, except banks, use it as money. Banks on the other hand, need central bank money in order to be able to make payments to other banks on behalf of their clients. If they have not enough central bank money to do a payment, they go bust. They have three sources from which they can get this central bank money – peoples’s deposits, interbank market and the central bank itself. So, that is the answer to your question, why banks pay interest on savings.
@Alex Hummel
no, I don’t think I get this wrong, I haven’t mentioned it at all.
when a bank makes a loan it creates an asset – the debt owed to it by the borrower – and a liability – the entry on the deposit account of the borrower.
this is not interesting or important.
so a bank agrees to lend me £1000, and creates a deposit account for me with £1000 in it, a liability for the bank, and an asset “Luis Enrique owes me £1000”. But at that point nothing has happened. I owe the bank £1000 but I’ve also got £1000 on deposit with it. Our positions are net zero.
I could tell you Alex Hummel that you have £1000 on deposit with the bank of Luis Enrique, and I could tell myself that I have an asset, £1000 owed to me by Alex Hummel. But I don’t have the power to create money out of thin air. The moment you tried to withdraw that £1000, I’d have to come up with the money.
Same goes for a bank. The moment something actually happens – the money is actually lent, that is taken out of the borrowers deposit account with the bank, the moment where we “use it as money” to pay for something, the bank has to be able to finance the loan and that’s when, as you say, it needs a retail saver, a wholesale lender or the CB, to come up with the cash. Before that, it’s just uninteresting notional entries on an accounts ledger, and is rightly ignored by Krugman and other economists.
The answer to my question, why banks pay interest on savings, is that they do require savers to finance the loans they make (well, as one source of loanable funds – either savers, or wholesale lenders, or lending from the CB).
net borrowers require net savers.
people like the idea of stock-flow consistent models. I don’t think you can write down the balance sheets of individuals and banks, and the flows between them, and have some net borrowers with out some net savers. People do have to save so others can borrow.
@JKH “OK for a dollar swap for new investment goods, but not for the original production of new investment goods, which produces income and saving. Net investment as a component of GDP definitely increases book value, and should logically increase market value, other things equal (ignoring the fact that things are rarely logical or equal).”
Right, I’ve been pondering a lot about investment vs investment spending, consumption vs consumption spending. I think this makes sense, and shows my error:
Drill-press producer has inventory on their books at cost (say $5K). When I buy it they get price ($10K), not cost. So my book value is unchanged: -$10K cash, +$10K drill press. But theirs goes up by $5K.
But isn’t this just saying that there’s surplus from trade, that investment spending (or consumption spending for that matter) monetizes the value-add from production?
@Asymptosis
Good point.
The purchase from the producer pays the final factor of production – equity capital.
That “writes up” the value of final output, income, and saving.
But the other factors have been/are being paid (labor; debt capital) while its still in the form of inventory investment?
@JKH “But the other factors have been/are being paid (labor; debt capital) while its still in the form of inventory investment?”
Right. That’s the real “investment,” no? (I’m not proposing new usage; just clarifying conceptually, perhaps…) But we *measure* investment by investment *spending.* With inventory as the buffer, arithmetically ensuring that S=I. Gets kind of heady with Apple-style production, where they don’t produce/ship until I’ve paid them (though they have produced or bought all the components already.)
It all balances out within the period, as long as one understands the definition of “investment spending” — purchases of goods (including inventory [including consumption goods inventory]) that won’t be consumed until future periods.
Enter, stage left, consumption of fixed assets.
Which brings me back to SavingS, which you’ve I think talked about as cumulative Saving. But: shouldn’t SavingS be cumulative Saving (which includes gross investment) minus cumulative Consumption of Fixed assets? IOW Cumulative Net Saving (“net saving” in the SNA sense). But maybe I’m remembering your words wrong.
@Asymptosis Edit:
“which includes gross investment”
should read:
“(= gross investment)”
@Asymptosis
“That’s the real “investment,†no?”
I would think the correctly measured investment flow includes the change in inventory stock, as valued as inventory, plus what has been sold from inventory in the current period, valued at full factor cost including the cost of equity capital. (I think; I’d have to double check/think that – but too lazy/time constrained right now.) That should cover all the factor costs and income generated by investment for the current period.
Ironic you mention cumulative saving. I’ve been sloppy about the depreciation part of that, as has been pointed out elsewhere and as you suspect here. My phrasing was always intended to refer to the balance sheet result of cumulative saving, which by necessity nets out depreciation from both asset value and equity. But I should be more explicit about that I guess, because a few people believe I totally overlooked it (not you necessarily). So I should start getting more precise about gross and net saving and what I mean by cumulative. Unfortunately, MMT uses “net save” in a completely different way. So …
@JKH
This brings me back to something I’ve asked you about before: shouldn’t C include consumption of fixed capital? It doesn’t; only includes purchases of goods which are consumed within the period.
Don’t worry: I am (thanks largely to you) beyond thinking the accounting needs to be changed. But the terminology attached to the accounting has to be used very carefully and clearly. IOW I think people should be very clear in *understanding* the distinction between consumption spending (C) and consumption (C plus depreciation). FWIW…
Steve,
Can I just say something? I haven’t read the comments, so I don’t know if anyone has made this point already, and I’ve said this in other forums repeatedly, but this,
“It’s actually the fundamental assumption, the sine qua non, of his paper (and of Krugman’s beloved IS-LM — the linch-pin of “New†Keynesianism.”
Is wrong. Wrong, wrong, wrong.
Another point that might be worth making is that loanable funds are ultimately real resources. In order for someone to “borrow” real resources from someone else, that someone else has to have some “saved”. This has nothing to do with the correct model of banking and everything to do with resources not disappearing into or out of the aether.
As ever, this debate is characterised by two groups of people talking past one another: the orthodox guys don’t know what the heterodox guys are talking about and the heterodox guys don’t know what the orthodox guys are talking about. More light than heat.
@vimothy
Hey I was just re-reading this thread and realize I never answered this. Am wondering which is wrong — the main clause or the parenthetical or both — and why.
[…] we saw quite clearly in The Great Keen-Krugman Debate, Paul is a firm believer in that model (and obviously Hempton is […]