Krugman on Say’s Law

Here:

http://krugman.blogs.nytimes.com/2013/02/10/still-says-law-after-all-these-years/

Still Say’s Law After All These Years

“When John Maynard Keynes wrote The General Theory, three generations ago, he structured his argument as a refutation of what he called “classical economics”, and in particular of Say’s Law, the proposition that income must be spent and hence that there can never be an overall deficiency of demand. Ever since, historians of thought have argued about whether this was a fair characterization of what the classical economists, or at any rate his own intellectual opponents, really believed.

Not being an intellectual historian myself, I won’t venture an opinion on that subject. What I will say, however, is that Say’s Law (Say’s false law? Say’s fallacy?) is something that opponents of Keynesian economics consistently invoke to this day, falling into exactly the same fallacies Keynes identified back in 1936.

In the past I’ve caught Brian Riedl and John Cochrane doing it; now Peter Dorman finds Tyler Cowen in their company.

Cowen can’t see why corporate hoarding is a problem. Like Riedl and Cochrane, he concedes that there might be some problem if corporations literally piled up stacks of green paper; but he argues that it’s completely different if they put the money in a bank, which will lend it out, or use it to buy securities, which can be used to finance someone else’s spending.

But of course there isn’t any difference. If you put money in a bank, the bank might just accumulate excess reserves. If you buy securities from someone else, the seller might put the cash in his mattress, or put it in a bank that just adds it to its reserves, etc., etc. The point is that buying goods and services is one thing, adding directly to aggregate demand; buying assets isn’t at all the same thing, especially when we’re at the zero lower bound.

What’s depressing about all this is that Say’s Law is a primitive fallacy – so primitive that Keynes has been accused of attacking a straw man. Yet this primitive fallacy, decisively refuted three quarters of a century ago, continues to play a central role in distorting economic discussion and crippling our policy response to depression.”

…………………………………

The underlying issue here at the present time seems to be the lack of investment mobilization of corporate cash hoards that presumably have resulted from corporate profits. To the degree this is correct, it concerns a certain tranche of macroeconomic saving – undistributed corporate profits or retained earnings.

Aggregate macroeconomic saving at any point in time is a recorded event that cannot have taken place unless there has been corresponding macroeconomic investment.

In addition to this macro level saving, well defined sector measures of saving can reflect dissaving in other sectors. This type of saving is not matched by investment. Sector dissaving cancels out sector saving at the macro level.

Both of these kinds of saving involve matching up of expenditure to income (or borrowed income), ex post.

Such measures are income statement measures, from a financial accounting perspective.

At any point in time, it is simply not possible for already measured macroeconomic saving to finance new investment. Such saving corresponds to macroeconomic investment that has already been made.

The saving that will correspond to investment that has yet to be made can only materialize in future accounting statements. All saving that has been made to date is already used.

And this is the case regardless of whatever pattern of corporate profits and cash hoarding that is already in place.

And that is really goes to what is wrong with Say’s Law at the most fundamental level.

Say’s Law denies that a propensity to save from income can cause underemployment of resources. In other words, it claims that a propensity to save will be matched by employment of resources. This is what Keynes attacked. Aggregate demand can fail as a process.

But saving from income doesn’t finance investment in ANY meaningful ex ante aggregate demand sense – let alone the way in which Say’s Law or Tyler Cowen seem to think it does. Macroeconomic saving cannot be deployed, ex ante, into new investment. That relationship has already happened. And so Say’s Law cannot hold simply due to the error in macroeconomic causality of saving and investment and the impossibility of attempting to force saving to “do something” in terms of an ex ante effect on the employment of resources.

Banking is part of the mix. This muddies the waters even further, and in complicated ways.

The income statement measures of consumption spending, investment, and saving are not directly related to the way in which the medium of exchange is sourced to spend on new investment – or the way in which it is used to provide inter-sector finance (e.g. bank consumer lending) that causes spending and related sector specific dissaving to generate sector specific saving in the future.

For example, there is no way that macroeconomic saving can be “put in the bank” that subsequently lends it out to new investment projects that “use” that same saving at the macro level. This is a sequence of conflated real and monetary causality and timing that is logically impossible.

First, as noted, macro saving can’t create investment, ex ante. Second, the bank deposit that both Krugman and Cowen refer to in their examples already exists in the banking system. That money corresponds (by popular presumption) to the (presumed) cash result of retained earnings, which in turn have been generated by prior sales of goods and services. The monetary execution of those sales includes transfers of money from buyers’ banks to sellers’ banks.

That money cannot fund a new loan in the macro sense. Loans create deposits at the macro level – not vice versa. Existing loans already account for deposits originally created from them. In addition, the liability composition of banking is constantly swirling in mix such that deposits may be converted into other liability forms and vice versa. But all of that occurs within the accounting constraint of double entry bookkeeping, such that deposits that exist at a point in time cannot logically be linked to subsequent incremental lending at the macro level.

More generally, banking stocks and flows are quite separate and distinct from the macroeconomic measure of saving and investment. By even referencing corporate cash hoards in the same context as presumed corporate saving, the issue of Say’s Law has become commingled with the issue of accounting coherence.

The “primitive fallacy” of Say’s Law in conjunction with the existence of a monetary system mangles the required logical linkages of accounting across time and across the real and monetary subsets of accounting at a point in time. This mangling appears to be deeply embedded in mainstream economics. Banking transactions such as those referred to are generally captured in flow of funds accounting, while any expenditure and income effects that may be subsequently associated with that flow of funds are measured in income statement accounting. NIPA is an example of macro income accounting. The Fed Flow of Funds report is an example of macro flow of funds accounting. In this sense, roughly speaking, income statement accounting is aligned more closely to the mathematical measure of real economy output – while flow of funds accounting is aligned more closely to the mathematical measure of monetary economy activities that enable the liquidity to support that income generation. That bifurcation is not pure, but it is notable. These two types of accounting, while separate, are inextricably interconnected by pristine logic – which is what comprehensive financial accounting is about.

Krugman is obviously right on Say’s Law, and in his basic message about the macro dynamics of aggregate demand. But his writing seems to complicate the explanation of aggregate monetary dynamics by citing micro level examples that obscure substantive macro accounting issues. For example, there’s no need to introduce bank reserves into such a discussion. There is no need to talk about “putting money in the bank”, when the money is already in the bank at the macro level. This micro fall back tendency is in evidence in other discussions involving banking – such as the recent monetary base debate with Steve Waldman, or the big banking debate with Steve Keen a year ago. On banking, perhaps Krugman tends to go a little too micro, when he might stay macro. Perhaps there is a way of explaining banking system T accounts at the macro level that would be translatable to his NY Times readers, although this would be difficult. Apart from that, he seems to know and explain everything as well as anybody.

The thing about corporate cash hoarding is that there is not even a precise linkage between that and corporate saving. It is quite possible and even likely that a strong correlation does exist and perhaps has been excavated from the flow of funds reports – but it is also the case that corporate balance sheets reflect complex patterns in the flow of funds that overtake the assumption of a simple connection from retained earnings to cash in the bank. In any case, it is the future flow of funds that will end up determining the future deployment of those cash balances. And when that happens, the associated future investment and saving will be separate from the macro level saving that has already occurred.

About

View all posts by
Guest
Philip Diehl
3 years 5 months ago

Unfortunate that this is too long to post as a Comment on Krugman’s blog. Perhaps you could introduce it and provide a link within the 1500-character limit.

Guest
JKH
3 years 5 months ago

Hi Philip,

I’m not sure such a link would generate a whole lot of excitement for him.

However, I’m suddenly in the mood for some mild impertinence. I understand he lives in New Jersey, so I was thinking it would be nice if he were to go on a weekend camping trip to the Pine Barrens, for the sole purpose of reviewing banking system T accounts at the macro level.

I’ve watched one too many Sopranos episodes to be available for that discussion – but was thinking that Scott Fullwiler of MMT fame might want to volunteer, given his notable expertise in the area and the amount of frustration he’s experienced and the investment he has made in this very issue – and with Krugman in particular.

🙂

Guest
3 years 5 months ago

JKH,

I am not sure what you mean when you say

At any point in time, it is simply not possible for already measured macroeconomic saving to finance new investment. Such saving corresponds to macroeconomic investment that has already been made.

It is true investment creates saving but saving can finance new investment.

Say during a year I earn 100, pay taxes of 20 and consume 50.

My saving is 30.

I finance a new house purchase of 200 with the 30 and borrowing 170.

Of course this act of house purchase doesn’t change my saving and it is 100-20-50 = 30 still.

But the saving has partly financed an investment.

I can go one more step and think of another example – consuming all my disposable income and financing a new house by a mix of sale of existing financial assets and borrowing.

Tyler Cowen’s error seems to be that he is saying if a corporation purchases commercial paper, there is an increase in investment and thinking of the investment decision as a result of someone’s purchase of commercial paper – scarcity of funds etc.

Guest
JKH
3 years 5 months ago

Hi Ramanan,

In your example, “already measured macroeconomic saving” is 30.

The “point in time”, defined in sequence relative to what is “already measured”, is at the end of the year.

And the investment has already been made.

So saving equals (a portion of) investment, ex post.

So that saving can’t be available to finance “new investment” – which pertains to the time period starting at this “point in time” – i.e. the upcoming time period.

Cowen is making an error of the Say’s Law type, as well as an error in the relationship between income and flow of funds accounting.

Guest
Dan Kervick
3 years 5 months ago

JKH, isn’t this just a “saving” v “savings” issue? In Ramanan’s example, the income during Year One in 100, consumption is 50, taxes are 20 and saving is 30.

These are flow quantities. That saving of 30 adds to one’s stocks of savings. Suppose for the sake of argument that the pre-existing stocks are 200, so that at the end of the Year One accounting period they have grown to 230.

Now lets suppose that in Year Two, income is still 100, taxes are still 20, saving is zero and consumption is 300. In other words, the agent has used the 70 portion of current year income not taxed plus the 230 in pre-existing savings to finance 300 in consumption. In that case, wouldn’t we say Year One saving has helped finance the Year Two consumption?

No just change “consumption” to “investment” throughout and we have consumption financing .

So while it is certainly true that net saving measured over a given accounting period can’t finance investment during that same period, surely it can finance investment in a succeeding period, no?

Guest
Dan Kervick
3 years 5 months ago
That is a further complication, but what it really means is that somebody else has saved in the current period, in order to finance at the macro level the excess of your current period consumption over your current period income (which represents your disaving in the current period). Why is that true JKH? Again, it seems like you are only talking flows. If there are pre-existing stocks of savings, then why does excess current-period consumption or investment at the macro level have to be financed by current period saving? If a guy decides one year to eat all the canned beans he has stored up in his bomb shelter for ten years, no one in the economy has to save any additional beans that year for that to happen. The same is true if a carpenter decides one year to invest his entire stock of saved nails in the production of new furniture. It seems to me Say’s Law is bunk no matter what model we use. Say doesn’t even seem to recognize saving at all, or acknowledge that businesses can build up stocks of unused capital goods or put their money in a safe. He thinks that the depreciating value of both physical goods and money means every acts on an urgent need to exchange every bit of income right away. It’s crazy. It seems to me that the current Say-like reasoning is based on (i) the partially-correct view that at the level of the individual, household or firm, saving in the modern world always takes the form of purchasing a financial asset, and (ii) the false and deeply misleading idea that financial assets are all claims on flows of goods and services produced by business investment, or on the monies exchanged for those goods and services. The economist… Read more »
Guest
JKH
3 years 5 months ago
Dan, Yes, I am mostly talking flows (which BTW is unusual for me, but appropriate to the emphasis here). In fact, investment at the macro level equates to saving at the macro level – as a period flow. Any constituent micro interactions in the flow of funds within that period are subject to a macro level accounting constraint in terms of their additivity in the total effect. The fact that a corporation may have more cash on hand to spend on current period investment than earned in current period undistributed earnings can’t undo this macro level constraint. Things must add up. And because of that, what is properly recorded as a deficit of saving relative to investment within the corporation (the excess of investment over undistributed earnings) must be offset by current period saving elsewhere. I think Krugman does a pretty good job of debunking Say’s Law, apart from some fine tuning that might be desirable on the banking side. Saving doesn’t always take the form of purchasing a financial asset. Households can save and purchase new real estate. The macro accounting actually registers this as an asset swap for an investment that has already been made by the firm that constructed the house. But it is the production of a real asset that becomes part of GDP output in the current accounting period and a real investment that is offset by household saving. “The economist and ordinary folk then misleadingly classify the purchase of a financial asset as an “investment” – no matter what kind of asset it is, and conclude that all saving in the society is concurrently investment.” Yes, that’s definitely a confusion in colloquial terminology. I’ve never sweated this one myself (in the sense of walking and chewing gum) but I can definitely see how it… Read more »
Guest
Dan Kervick
3 years 5 months ago

In fact, investment at the macro level equates to saving at the macro level – as a period flow.

This is what I’m challenging. Why is that a macro level constraint for any period?

Guest
JKH
3 years 5 months ago

Because it’s the way the national accounts are constructed.

And it’s straight out of Keynes – consistent with the sector financial balances derivation, also from Keynes:

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

S is private sector saving
(T – G) is government saving
(M – X) is foreign saving

S + (T – G) + (M – X) is total saving

So, I = total saving

Guest
Jose Guilherme
3 years 5 months ago

Of course, it’s never too much to stress once again that this should be understood as a mere accounting identity, not necessarily implying that “Investment is a function of Saving”.

In fact, the identity could be turned around to “prove” that saving is a function of investment.

Just like in the PK expression that “bank loans create deposits, instead of deposits creating loans”.

Guest
Dan Kervick
3 years 5 months ago

Garbled: “No just change “consumption” to “investment” throughout and we have consumption financing .”

Meant to write: “Now just change “consumption” to “investment” throughout and we have saving financing investment .”

Guest
3 years 5 months ago

JKH,

What I am saying is about the sources of financing. So saving is one source and borrowing is another. So investment is financed by both saving and borrowing. So although, investment is not limited by saving alone, there is no contradiction in saying that saving finances investment if it is understood that it is also financed by borrowing and understood that investment is not limited by saving or funds in existence.

Guest
JKH
3 years 5 months ago
Ramanan, I’m trying to address two intertwined issues here. The first is the issue of aggregate demand – in the context of Say’s Law, and the refutation of Say’s Law. That requires treatment of investment and saving at the macro level. If you look at the post closely, you’ll note the number of times I’ve prefaced various statements with the macro level qualification. The second is the issue of the relationship between that first issue and the banking system, or more generally the flow of funds. Here I’ve tried to distinguish aspects such as “money in the bank” or “borrowing” from the basic Say’s Law and related aggregate demand logic whereby expenditure must equal income, and investment must equal saving (e.g. global macro). The interesting thing about Krugman’s analyses is that he makes the right case with respect to the first issue (IMO), but in explaining this in banking system terms he tends to slip a bit. This appears to me to be a common theme – throughout his debates with MMT, Keen, and Waldman. With respect to your statement – it is a statement of micro de-consolidation of the flow of funds. And it is a statement of the ultimate de-consolidated financing configuration that corresponds to a macro level investment/saving equivalence. And I agree with it to a point. But it doesn’t negate the macro level ex post required equivalence of investment and saving. But I think one still needs to be careful in characterizing the difference between income accounting and flow of funds accounting at this micro level. As an example, if a household saves from income, and deploys that flow of funds into a bank deposit, and if a corporation has borrowed from that bank to invest, then the macro marginal effect is that saving will have… Read more »
Guest
3 years 5 months ago

Btw have you seen the “Kickstarter”?

Guest
JKH
3 years 5 months ago

no?

Guest
3 years 5 months ago

Yeah agree.

I didn’t quite follow his debate with Waldman.

Your points here and elsewhere about Keen are right. Krugman really had a point. But his description mixed so many things giving people a chance to ….

What about Krugman v MMT. Did he have any discussion on these matters with them?

Guest
JKH
3 years 5 months ago

Krugman’s latest on MMT here:

http://www.nbcnews.com/id/46979738/#50760434

@ 7:30

extended interview with Chris Hayes yesterday – excellent

1 4 5 6
wpDiscuz