The “Global Savings Glut” Is Conceptually Incoherent. “The Economy” Cannot “Save”

When you hear people talk about the Global Savings Glut, you can be quite sure they are talking about monetary “savings” — the global aggregate stock of money embodied in financial assets.

What they don’t seem to realize is that the net holdings of global financial assets minus liabilities — claims and counterclaims — is always exactly zero. By accounting identity.

Every financial asset is a claim against another party. For every asset (credit, claim) in the account of a financial-asset holder, there is a equal and opposite liability (debit, counterclaim) in the account(s) of some other party or parties.

Sum up all those assets and liabilities, and you get zero. The gross quantity of financial assets (and liabilities) can increase, but the net quantity always remains the same.

Here’s a physical example that might help make this clear:

Imagine an organism floating in space. It has one input (sunlight) and one output (heat). It’s able to convert the sunlight into stored mass and energy, with some loss in the form of heat. So it can grow, get larger, storing up mass and energy in various physical/chemical forms.

Now you could call that “saving,” but it’s probably better described as “growing.”

Now suppose that for whatever reasons, parts of that organism periodically run short of resources, while others are holding more than they currently need. The surplus holders can give resources to those who are short resources. The receiver notes down a new tally in a financial account book — a debt or liability. The giver records a promise, an asset. An IOU. A credit.

You’ve just created money.

An aside: it drives me crazy when people say that liabilities are money. They’ve got it exactly backward. When you’re holding someone’s IOU, you’re holding an asset, a credit. Think: a Target gift card. They’ve got an offsetting debit, a liability, recorded on their books. When you give the Target gift card to your brother-in-law, you’re not transferring the liability; it remains unchanged on Target’s books. You’re transferring a credit. Money is credit. I think that everyone will agree that in normal vernacular speech, money is an asset, not a liability.

Two new things are created by that transaction in space, ex nihilo: two tally entries on accounting books — the credit and the debit, the asset and the liability. (But no “real” things are created.) Together they increase the gross quantity of financial assets and liabilities (which are just tally entries, often represented by physical tokens), but have no effect on net aggregate “money savings.”

This transaction, of itself, has no effect on the organism’s creation or accumulation of new mass and energy from sunlight, or the loss via heat. It just changes different parts’ future claims on that mass and energy. So if you want to call that physical accumulation “saving” (rather than “growth”), you have to say that the transaction has no effect on saving.

And if you want to call the accumulation of accounting tallies “saving” (“money saving,” which is what most people think of when they hear the word), you have to say that the transaction also has no effect on net saving. There is no more net money in the world than there was before the transaction (though there is more gross money). That net number is still, and will always be, zero.

This is even true if a government prints, say, dollar bills and helicopter-drops them over the countryside. People pick up those credits, so it seems like there’s more “money.” And for the picker-uppers, there is — they have more credits against government liabilities (taxes and fees). But since they can use those credits to pay off liabilities to the government, the government has reduced its ability to demand the already-existing dollars out there through taxes.* While it’s increased the amount of credits out there, it’s reduced its power to demand (claim) existing credits back by an equal amount.

Once again: net zero.

So: there are massive quantities of financial assets out there. Does this mean there’s a Global Savings Glut? No: because there are equal and offsetting quantities of financial liabilities.

Should we call that a Global Liability Glut?

Net global money “savings” (financial assets minus financial liabilities) is always zero. If we’re talking about money savings, there can never be a Global Savings Glut. It’s an incoherent concept.

I’ll leave it to my readers to ponder what that means for the notion of “loanable funds” — a notion you’re invoking every time you use the term “Savings Glut.”

What we have instead of a Global Savings Glut is:

1. A Global Labor Glut: more human effort and ability available than is needed to provide goods that provide high aggregate marginal utility, and,

2. A global financial and political system that — despite the reality of #1 — fails to transform that abundance into maximum aggregate human utility via reasonable distribution of that abundance.

* Used to be, you had to give the lord of the manor one of your cows, or some of your corn, every year. When he issued money ex nihilo — probably to pay soldiers — and started accepting that credit instead for taxes, he sacrificed some of his claim on your cow. He exchanged one asset/credit/claim (for your cow/corn) for another (the solders’ services) .

Cross-posted at Angry Bear.








8 responses to “The “Global Savings Glut” Is Conceptually Incoherent. “The Economy” Cannot “Save””

  1. Ramanan Avatar

    I am not sure what you say.

    While true global assets and liabilities cancel out, but it is not that the distribution doesn’t matter.

    The error of the global saving glut theory is that it put the blame of the big U.S. trade deficit *entirely* on others and worse also puts the blame of the private sector deficit on other nations entirely.

    But it does have some decent element of truth in it for the trade deficit part. A nation with high current account surpluses would have high saving because national saving is gross fixed capital formation plus the current account balance of international payments. So such surplus nations have a “glut”. The solution is then to have them raise domestic demand and hence national income and imports which would reduce the current account balance and reduce or remove the “glut”.

    Of course there are other reasons too for the U.S. trade deficit.

  2. The Arthurian Avatar
    The Arthurian

    Thumbs up, Ramanan.

    Yes there is a global savings glut AND a global liability glut. The liability glut increases costs in the productive sector. The savings glut increases income in the nonproductive (financial) sector.

    Look at the puzzle long-term and you see that these gluts arise periodically and peak in great depressions. And that the “minimum” between gluts is a time of economic vigor.

  3. Greg Avatar

    Ill say this about the “liabilities are money” part. In your example, its Targets ability and willingness to accept that card as payment for their stuff that makes it money. Target has the goods to back up the credit it issued you, so its two sided (of course) but it seems that its the guy who can promise something real that is creating the money, the guy with the credit just has a ticket. The guy who makes the promise is the money printer, the guy who redeems the promise is the money spender.

  4. Eric L Avatar

    Have to agree with the other commenters here. I’ve always understood the “global savings glut” to be about gross savings, not net. And I think it is a useful idea that we want more laymen to be aware of. “Everybody knows” there is too much debt in the economy. What isn’t intuitive to the layman is that you can’t have too much debt without also having too much saving. Making people more aware of the role saving plays in our economic dysfunction is a good thing in my book.

  5. JW Mason Avatar


    You are right that there cannot be a “savings glut” in the sense of an excess of realized savings over realized investment. I agree with you, there is a lot of sloppy language on this point.

    But I think this is one of those cases where, when you see people saying something that seems nonsensical, it is worth the effort to figure out if they are actually saying something sensible but expressing it poorly. Here’s how I would do that.

    Aggregate savings and aggregate investment must always be equal. But the individual savings and investment decisions are taken independently. So something else must be adjusting to keep them equal in the aggregate. That something else is the interest rate (understood broadly as the terms of credit contracts), and the level of income. Desired savings and investment both respond positively to income, savings more so than investment. Savings respond positively to the interest rate, and investment responds negatively. (Keynesians are often skeptical of the interest-elasticity of investment; that doesn’t matter here.) So the savings-investment equality condition requires that the economy always occupies only a certain subset of points in interest rate-output space. This is the original derivation of the familiar IS curve.

    In this sense, an excess of savings means that desired savings **at a given point in income-interest rate space** has increased relative to desired investment at that same point. Of course we don’t ever observe the implied excess of savings over investment, what we observe is a fall in interest rates and/or income. We call this excess a “glut” when we think that the fall in interest rates required to maintain the current level of income is either infeasible or undesirable. This is a perfectly coherent story, tho whether it applies to the current situation is of course a different story.

    On the financial side, similar reasoning applies. Yes, realized net acquisition of financial assets always equals realized net issuance of new liabilities. But that does not mean that *desired* net acquisition of assets equals desired new liability issuance for every possible value of whatever parameters we think influence these choices. For some sets of parameters, desired acquisition of financial assets would exceed desired issuance of new liabilities. Of course we don’t ever observe that — it’s impossible — what we observe instead is a change in the parameter values. When you think that such a change is due to an excess of desired asset acquisition over desired liability issuance at the old parameter values, it is perfectly reasonable to say that the old situation changed because a financial savings glut developed.

    i think this is important because I think your conclusion that unemployment must be a glut of labor is both wrong and politically destructive. By turning this into a labor-market problem, you are implicitly accepting Say’s law and rejecting the principle of aggregate demand. And you are wrong factually. All factors of production are in excess supply in a recession, not just labor.

  6. JW Mason Avatar

    What you are saying here is that changes in (realized or desired) balance sheet positions never affect the real economy. I’m sure you don’t think that’s what you’re saying, but it is.

  7. Eric L Avatar

    “Aggregate savings and aggregate investment must always be equal. But the individual savings and investment decisions are taken independently. So something else must be adjusting to keep them equal in the aggregate. That something else is the interest rate[…]”

    This is mostly tangential to your argument, but I think this part of it is poorly put. Desired savings and desired investment (really desired borrowing — credit card bills aren’t really investment, but I digress) are balanced in equilibrium by the interest rate. Aggregate savings and aggregate investment are not balanced by the interest rate. If they were they would only be equal in equilibrium. However, they are always equal, regardless of what happens to the interest rate, because there is no third thing that must adjust to make them equal.

    You seem to be confusing individual decisions with the individual effects of individual decisions. If I choose to save $5, the individual effect on me is that I have saved $5 but have not myself invested. But that is not the aggregate effect of my individual decision. To get that, we have to compare my decision to a counterfactual decision (I spend $5 on coffee) and see how the aggregates differ between the two scenarios if we hold everyone else’s spending decisions constant. And then we see my individual decision to save more also had the effects of the coffee shop saving less (because we are holding their spending decisions constant but they have less revenue) and investing in unused coffee beans. And the value of their negative saving and their investment will be such that the total effect of my decision on aggregate saving will be equal to its effect on aggregate investment, even if nothing else changes.

    The dirty little secret of S=I is that all the action is in I, which is actually on both sides of the equation. When a company spends on investment, that creates income that someone else can save, but its spending is not subtracted from its own saving as it would be if it were spending on consumption or intermediate goods.

  8. […] Ryan Avent’s excellent post at The Economist finally provides me the impetus to respond to Josh Mason’s comments on my recent post. […]