skip to Main Content

The IMF fall into a loanable funds black hole … again

Household saving ratios (saving as a percentage of disposable income) have risen significantly in most countries since the onset of the recession. In many countries this has come after a period of increasing indebtedness as national governments pursued budget surpluses. As a result, the macroeconomic concept of the paradox of thrift has been resurrected in the popular press as a discussion point. There are fears that the end of the “consumer boom” will lead to stagnancy. A recently published IMF paper addresses this point but just cannot let themselves address the elephant in the room. They present a new way version of deficit hysteria.

A recently published IMF research paper entitled Frugality: Are We Fretting Too Much? Household Saving and Assets in the United States, argues that while the saving ratios will not continue to increase once asset losses are contained:

… there are reasons to believe that an era of “new frugality” may have begun: asset losses have been severe, the economic environment is less certain, public sector deficits loom ever larger, and many households have been forced to deleverage. These more extreme scenarios yield savings levels that are roughly equal to averages during previous eras: the 1990s, 1980s, and as most extreme, 1950s–70s.

I was amused by the notion that a return to saving ratios of yesteryear after some period of debt-fuelled consumer binging was appropriately labelled “extreme” or a “new frugality”. The fact is that the recent period has been atypical in our history. The typical behaviour over the period that we have had reasonable data series is for national governments to run budget deficits which have supported desired positive saving ratios from the household sector via sustaining adequate output levels.

To understand what follows we need to be absolutely clear what a paradox of thrift entails. It was one of the fallacies of composition that Keynes and others considered established a prima facie case for considering the study of macroeconomics as a separate discipline. Prior to that the neo-classical (the modern mainstream) had ignored the particular characteristics of the macro economy that require it to be studies separately.

They assumed you could just add up the microeconomic relations (individual consumers add to market segment add to industry add to economy). They overcame the unresolvable aggregation problem by fudging it and assuming that there was a representative firm or representative industry (that all sub-units behaved like). But in fact, because of this fudge, the mainstream had no aggregate theory anyway.

But as a consequence they just assumed that what held for an individual would hold for all individuals. This led the mainstream opponents to expose the most important error of the mainstream reasoning – they simply ignored the Fallacy of Composition that was endemic to analyses that tried to reason generally from the specific.

The paradox of thrift was one such example. Accordingly, if an individual tried to increase his/her individual saving (and saving ratio) they would probably succeed if they were disciplined enough. But if all individuals tried to do this at the same time, then the impact of lost consumption on aggregate demand (spending) would be such that the economy would plunge into a recession.

As a result, incomes would fall and individuals would be thwarted in their attempts to increase their savings in total (because saving was a function of income). So what works for one will not work for all. This was overlooked by the mainstream.

Paul Krugman captured it in this way in his New York Times blog When consumers capitulate:

… one of the high points of the semester, if you’re a teacher of introductory macroeconomics, comes when you explain how individual virtue can be public vice, how attempts by consumers to do the right thing by saving more can leave everyone worse off. The point is that if consumers cut their spending, and nothing else takes the place of that spending, the economy will slide into a recession, reducing everyone’’s income. In fact, consumers’ income may actually fall more than their spending, so that their attempt to save more backfires — a possibility known as the paradox of thrift.

The term paradox of thrift entered the nomenclature during the Great Depression as Keynes and others saw this particular problem as providing a prima facie case for government deficit spending (the something else that “takes the place of that spending”).

But recognising the importance of the fallacy of composition in mainstream economics at the time was a devastating blow to its credibility. How short our memories are? How the mainstream ideas that were discredited so comprehensively in that period have been able to reassert themselves as the dominant discourse is another story (of puzzling dimensions). I am working on a book about that.

The IMF article, however, puts a new spin on the current rise in saving ratios – asserting that it probably doesn’t apply much at all after a financial crisis (asset deflation) has occurred. But this doesn’t mean that recovery will be quick. It is in this light that the rising saving will be “good” for growth rather than hold it back.

Their main argument is motivated by another 2009 IMF article (Sustaining a Global Recovery by arch-New Keynesian Olivier Blanchard, who said:

The best guess (and there is little more to go on) is that the U.S. household saving rate will remain at least at its current (Q2 2009) level … a 5 percentage point decline in the ratio of consumption to disposable income relative to the pre-crisis period … Put simply, 3 percent more of U.S. aggregate demand will have to come from something other than consumption. Will it be from investment? This also seems unlikely … Less-efficient financial intermediation will affect not only the supply side, but also the demand side. Again, historical evidence from “creditless” recoveries suggests that investment will be weak for a long time.

So focusing on the capital investment side they accept the view that “an increase in capital investment will not immediately replace lower consumer demand”.

They point to high “unsold housing inventories” and low rates of “capital utilization” and the damaged financial system. In this case, they argue that the rising household savings, far from impeding growth via the paradox of thrift channel, will:

… stimulate capital expenditures in two other, less well-recognized ways. First, increases in household savings should help compensate for the recent drop in corporate savings. Second, savings flows may have even greater impact on investment as they accumulate. As savings flows accumulate, firms and banks recapitalize. This permits firms to fund more investment projects, either internally or through leveraged external finance … new household savings will now be even more essential to repair balance sheets (both household and corporate) and hence replenish financial resources. The impact of stronger balance sheets on investment is a gradual and cumulative process. Thus, in order to return to pre-crisis investment levels within a decade, we would need higher levels of savings than those predicted by our baseline model—starting today. If not, we should fret about too little frugality, rather than too much.

So the argument is that firms typically use retained earnings to finance their investment plans. But the financial crisis has damaged their balance sheets so much that the companies need to find other means of finance. So … come in household savings …. which if they continue to rise will provide the funds necessary to stimulate the levels of investment we saw before the plunge.

There are so many things wrong with this analysis that is it difficult to know where to start. So in the interests of brevity I will concentrate only on the most obvious errors in understanding.

First, capital expenditure is usually asymmetric over the business cycle. You might like to read a paper I wrote with Joan Muysken a few years ago (that went into our book) which shows how irreversibilities in investment lead to asymmetric labour market behaviour. That is, firms are cautious in the upturn and wait until they are sure the recession is over before they commit to new investments which will lock them into increased capacity.

Second, the paper represents a fundamental denial of how the actual economy operates, particularly with respect to the banking system. As I have noted several times, banks do not accumulate deposits (derived from household saving) before they are capable of lending to firms to finance investment. That is not how the banking system operates.

Firms who are credit worthy will always be able to get the funds they desire (at some price to reflect risk) from the banking system irrespective of the deposits that the banks hold. Loans create deposits.

Third, as saving is not required prior to investment spending the only constraint on the latter is the confidence that firms have that aggregate demand will improve in the future to justify their capacity augmentation. And once the entrepreneurs are confident and resume spending saving will rise to match the injection. Spending generates saving not the other way around.

The IMF paper commits a basic error of reasoning because it still thinks that saving and investment are determined in some loanable funds market and mediated by interest rate variations. We jettisoned those notions in the 1930s as Keynes showed them to be deeply flawed. The fact that these discredited notions have returned in the neo-liberal period shows how retrograde and backward this period of macroeconomic debate has been.

But then one hardly expects anything special from the IMF. They are the exemplars of retrograde thinking and analysis.

Fourth, the basic problem with the IMF paper is that they just cannot let themselves address the elephant in the room – it is a paper that is basically in denial of that elephant. Almost as if they would be giving the IMF game away if they did recognise the elephant. Instead they have to pose questions and instances that just ignore the main issue.

There is no problem for the economy in households increasing their desired saving. The solution does not rely on private investment improving any time soon, although a rise in capital formation would quicken the pace of recovery.

The leakage from the expenditure stream that occurs as household increase their saving just has to be filled by a rising “injection”. And it doesn’t involve rocket science to know where than injection is coming from – the rising public deficit.

As long as the government sector “finances” that rising saving behaviour from the households economic growth can continue and the paradox of thrift effect thwarted. The rising net spending promotes income and employment growth, which combine to generate the rising saving capacity desired by the households. It is a win-win.

So by denying or failing to recognise the simplicity of this option, the IMF paper misses the most important point about macro behaviour in a downturn and is largely irrelevant despite its elegant econometric analysis and fancy algebraic mode.

But it is hard to believe that professional economists in 2009 still publish this sort of erroneous loanable-funds logic?

Closer to home, another professional economist was revealing he has no idea about the way saving works in a macroeconomic setting. The relationship between aggregate saving and net public spending came up yesterday in the Senate Committee hearing in Australia which arose as a result of a Greens-motivated resolution.

As an aside, The Greens as a consequence of this resolution have demonstrated to me that they are categorically a non-progressive party. They are riddled with neo-liberal macroeconomic elements that will undermine any hope they have of pursuing a sophisticated social and environmental agenda. You might like to read the foundations of my criticism in this regard which was outlined in this blog – Neo-liberals invade The Greens!.

Anyway, an economics professor from RMIT University (Steven Kates) demonstrated why his students are unlikely to learn anything about the way the macroeconomy works from his classes. Pity them. He told the enquiry (via the ABC PM program (the transcripts are not yet available):

You should not have this blanket expenditure as a stimulus, four per cent of GDP which is an unbelievable amount of money.

That will not create growth and in fact wastes resources so comprehensively in ways that will they are destroying our savings.

They are going to push up interest rates, they are going to push up taxation in future and may yet push up our inflation rate.

So prospective economics students – don’t go to RMIT University in Melbourne unless you want to be served up this rubbish. For existing RMIT University economics students – leave immediately.

But where did Kates get these ideas from? If you open any mainstream textbook you will find this view. However, if you understand the way the operations of the banking system (interaction between the central bank and the commercial banks interact) in a fiat monetary system and the way that income adjustments response to aggregate demand then you would never make these connections.

Increasing net spending by government adds to bank reserves and if nothing else happens the overnight interest rate will be driven down by competition in the interbank market as the commercial banks try, in vain, to eliminate the excess reserves. The operational reality, ground in the underlying national accounts, is that the banks cannot eliminate a system-wide excess of reserves. All they can do is shuffle the excess around between each other.

So budget deficits put downward pressure on interest rates across the term structure.

There is no finite pool of saving that different borrowers compete for and thus drive interest rates up when borrowing demands increase. That view was discredited in the 1930s by Keynes (and others). It is based on the loanable funds doctrine which was the mainstay of the neo-classical marginalists. It assumes saving is a function of interest rates rather than income. It assumes that investment (or other sources of spending that relies on borrowing) is constrained by the available pool of saving.

We understand that none of those assumptions or propositions are even slightly correct. Saving is a function of income which, in turn, is a function of aggregate demand (given available aggregate supply). Bank lending is not reserve-constrained (loans create deposits) and so investment funds can be created for any credit-worthy customer at the stroke of a pen. Further, we understand that investment brings forth its own saving as income rises and induced consumption is less than 100 cents in the dollar.

The final part of the story which Kates seemingly cannot understand is that far from “destroying saving”, the rise in deficits has actually financed the increased them. As the household’s desire to save increased in response to the crisis the decline in aggregate demand would have instigated sharp contractions in output and hence income generation. The households might have increased their saving ratio but the overall level of saving would have fallen as the economy contracted.

By increasingly filling the spending gap, the rising budget deficits has prevented some of the contraction from occuring which has “propped up” household saving higher than it otherwise would have been.


It amazes me that the mainstream researchers and professors still trot the discredited doctrines out day after day. There is such a lack of scrutiny of their ideas from the professional economics media and current affairs interviewers that I guess they can get away with it.

Sadly, I don’t suspect we will get closer to the truth as a result of this downturn. I thought that the downturn would have given breathing space for the progressive macroeconomic theorists who not only predicted the crisis but also ground their explanations in the operational realities of the system we work within, rather than some stylised fairy land that the mainstream pontificate over.

Later in the week Steve Keen and I are forming a partnership to write something together to see if we can resolve some issues. It should be constructive.

Spread the word ...
    This Post Has 13 Comments
    1. Nice post.

      “Firms who are credit worthy will always be able to get the funds they desire (at some price to reflect risk) from the banking system irrespective of the deposits that the banks hold. Loans create deposits.”

      Another small point, but I’d be a bit careful about this. Certainly those with a scintilla of knowledge about how the banking system actually works will agree totally that banks are not reserve constrained in lending. But banks can be capital constrained – meaning only that they require unallocated capital (in the bank definition of capital) to lend. That doesn’t mean they don’t already have or can’t get the capital they need to lend, but it also doesn’t mean they always have it, or if that they don’t that they’ll always be willing to go and get that capital externally, or if they do attempt to get it externally that such capital will always be available. Yes, everything’s available at a price I suppose, but there can be points where the market for bank capital ceases to be continuous and liquid in that sense. The Minsky moments can apply to bank capital. The financial crisis delivered a huge hit to actual bank capital, perceived bank capital, the availability of bank capital, and the willingness of banks to raise new capital at highly dilutive stock prices. All of this is why governments stepped in with capital. This of course is also the foundation for the debate about whether governments should have stepped in with ultimate debtor assistance directly rather than recapitalizing the banks. I believe the PK’s would have preferred the former. That’s a legitimate debate. I’m just making a point about the dependence of banks on capital without that additional step. I’ve probably made this point before, but I’m not sure how familiar economists are with the details of the process by which banks allocate capital internally. If more of them were familiar with it, I suspect far more would begin to understand how almost completely irrelevant the issue of central bank reserves is to bank lending.

      P.S. could the “loanable funds” lingo be corrected by referring to them as “loaned funds”?

    2. I’m gradually gaining more of an appreciation for how you develop two inextricably linked themes in parallel: loans create deposits; and spending generates saving. I think it’s pretty powerful that these are two distinct ideas, but joined at the hip through parallel causality. If you agree, it would be interesting to see a post where you explore this particular relationship in more detail.

    3. JKH – Hence Keynes’ emphasis on “loan expenditures” in the GT…which need not be restricted to government spending and commercial bank purchases of Treasury debt, as bank financed tangible investment by firms or households fits the bill just as well.

    4. Dear JKH

      Good points. The issue of bank capital is interesting and I might write something about that.

      On your PS – the loaned funds terminology would still not rescue their underlying misconception about how saving is generated via income growth.

      best wishes

    5. Dear JKH

      This is one of the starting points of modern monetary theory after you get the basic sectoral accounting and stock-flow consistency right.

      I am unsure whether I can say much more than I have already said about the two ideas. What specifically is left unsaid or is unclear?

      best wishes

    6. Hi Bill, I along with the rest of steve’s readers look fowards to your collaboration. I think it’s an excelelnt idea.

      On this matter I have just been reading a very good article by Willem Buiter in which he explores the issues of public bufget constraints, RE etc and the scope for further fiscal stimulus. When I say it is a good article, I mean it is well written and clear, and sets out his views well, not necessarily that I agree with it all.

      He seems positioned some half way between yourself and the neoliberal consensus. I hope through your post with steve that I’ll be better able to understand exactly where you may differ with buiter (I realise he makes some equilibrium based assumptions BTW).

      “Is there a case for a further co-ordinated global fiscal stimulus?”

    7. JKH

      Well put regarding bank capital. A little understanding of the “sustainable growth rate” for a firm applied to bank capital would go a long way. Also really liked your point on another thread regarding effects on bank cost of capital . . . I had been thinking the same and was glad to see you confirm my suspicion.

      And correct characterization of our opinion regardin recapitalizing vs. debtor assistance, though, as you probably know, the latter would have been a bit indirect (if our policy views were followed) since the preference is for general tax relief, revenue sharing with states, and a job guarantee, rather than direct help to particular debtors.


    8. Scepticus

      FYI, I did a paper for which the working paper version is published at (“interest rates and fiscal sustainability”) that gets at the chartalist operational view of much of this.

      Bill’s also done a lot on this topic here already. Check out his 3 part series on deficits 101 and the series on fiscal sustainability 101.


    9. Thanks scott. I’ll begin processing that paper. I see it addresses the issue of future demographics, which is encouraging.

      Interestingly buiter argues in part 4 (final) of his series:

      “The third is to shift the political equilibrium of the country to boost fiscal spare capacity. If President Obama can shift the destructive, polarised US political equilibrium, where Republicans veto all future tax increases and Democrats veto all future public spending cuts, there could be room for an additional fiscal stimulus that would not spook the financial markets.”

      Now that is in interesting statement, and completely relevant to what you guys are trying to achieve.

    10. JKH,

      Are you suggesting the following explanations and sequence of events ?

      1. Banks needed to be capitalized for market confidence. TARP bailed out banks by capitalizing them – however, it was fiscal and got added to the deficit.

      2. Someone thought something is not correct and selling government bonds will again cause banks to be undercapitalized

      3. Some of QE was forced. Plus Bernanke’s neoclassical philosophy got him to reach a figure of $2T of QE.

    11. Ramanan,

      Not sure I follow your points in total.

      But here’s something. The division of responsibility regarding banking system support as between the Fed and Treasury is interesting. It’s done now roughly according to the guideline that the Fed is responsible for bank liquidity support and the Treasury is responsible for bank solvency support. That’s more or less the difference in objectives between secured lending by the Fed and equity capital injections by Treasury.

      Another way of doing it would be to divide responsibility according to financial asset acquisition of any type by the central bank, and fiscal expenditure as reflected in the budget record on treasury’s books. This split could then be reflected consistently in more logical budget deficit accounting. As a result, the budget deficit would accurately reflect the corresponding net saving shift to the non government sectors. Financial asset acquisition would be isolated away from budget deficit accounting, which is appropriate since it does not alter net saving positions. It just changes the risk profile of gross saving positions. Only the income from financial asset intermediation changes the true budget deficit, depending on whether it generates a profit or loss.

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    To prove you're a person (not a spam script), type the answer to the math equation shown in the picture.
    Anti-spam equation

    This site uses Akismet to reduce spam. Learn how your comment data is processed.

    Back To Top