The paradox of thrift

Or, how come you used to say that if consumers don’t save more, it will wreck the economy, and now you say, if consumers do save more, it will wreck the economy?

For the record, I am certainly among those who had been suggesting that America’s low saving rate was a significant problem. Let me begin by reviewing why I said that. Recall that we can separate the various components of GDP (Y) in terms of goods and services purchased by consumers (C), government purchases (G), investment spending (I), and net exports (X):

Y = C + I + G + X

Subtracting C and G from both sides of the equation,

Y – C – G = I + X

The two terms on the right-hand side are the critical determinants of what kind of economic future we’ll have. Investment in plant and equipment is the single most important variable that will determine our future productivity and standard of living. And negative net exports, such as the U.S. has increasingly opted for over my lifetime, necessarily involves selling off our national assets and going further into debt to foreigners. The size of our current account deficit is large enough relative to GDP that, if this were any country other than the United States, I would worry that a currency crisis (a sudden flight from dollars) is a very real possibility. And even for the United States, it is something I for one do worry about.


From the equation above, if we want I + X to be bigger, we must want Y – C – G to be bigger as well. We can define private sector saving to be gross domestic income less consumption spending and net taxes paid:

private saving = Y – C – T.

Notice I’m using the same symbol Y for both GDP and GDI, since the two are conceptually the same– every dollar of production necessarily generates a dollar of income. There is a statistical discrepancy between the actual measures available for GDP and GDI, though these are not relevant for the longer run issues I’m discussing here. We likewise can define “public saving” to be the excess of the government’s receipts over its expenditures,

public saving = T – G,

and national saving to be the sum of private and public saving:

national saving = Y – C – T + T – G = Y – C – G.

In other words,

national saving = I + X

This equation is an accounting identity, as well as a condition that has to characterize equilibrium in any coherent macroeconomic model. Hence my longstanding advocacy of measures to raise the private saving rate or lower the federal deficit.

So then, aren’t I delighted that consumers are now, finally, saving more?


Well, no. It is one thing to identify a higher national saving rate as the long-term goal, and quite another thing to try to get there overnight in the form of a sudden drop in consumption spending. Here I am very much taking the side of Brad DeLong ([1],[2]) and Arnold Kling and against Eugene Fama ([1], [2]) and John Cochrane. The relevant question is whether, in response to an abrupt decrease in consumption spending such as we’re now experiencing, some of the other variables (most importantly, Y) might adjust in response as well. It is certainly true that in a very simple economic setting– for example, an economy that consists of a single farm producing only one good– the decision to save more of your income (leave some of your wheat unconsumed) is necessarily identical to the decision to invest more (save the wheat for later). And one can write down more complicated models in which economic actors and markets adjust in a way to see through the veil of production and exchange and make sure it is I + X that adjusts in response to a higher saving rate, and not Y.

But it’s also possible to write down models in which there are significant frictions that cause the adjusting to come in the form of lower Y in response to lower demand. The traditional such friction is the textbook Keynesian notion that wages and prices fail to adjust. In such models, responding to the lower C by increasing G may succeed in mitigating the loss in Y. Though here I must agree with Cochrane that those same models imply that if monetary policy could stimulate aggregate demand, that would achieve the same objective. I am definitely of the view that it is within the current power of the Federal Reserve to stimulate demand, and have urged the Fed to try to aim for a 3% inflation rate over the next several years.

But where I may disagree with some of my colleagues is in their presumption that wage or price rigidities are the core frictions that are responsible for producing the present situation. I have in my research instead stressed technological frictions. For example, when spending on cars abruptly falls, there is a physical, technological challenge with getting the specialized labor and capital formerly employed in manufacturing cars into some alternative activity. In my mind, it is a mistake to pretend that any federal program is capable of immediately re-employing those resources into an alternative, equally productive enterprise. More fundamentally, I have suggested that our present situation is as if someone had quite successfully sabotaged the basic functionality of our financial system. Until we once again have a financial sector that can successfully allocate credit to worthy projects, we’re not possibly going to be able to produce as much in the way or real goods and services, no matter what the level of aggregate demand or stimulus package might be. In terms of the textbook Keynesian models that people play with, I’m suggesting that “potential” GDP growth for 2009:Q1– that growth rate which, if we try to exceed it by stimulating aggregate demand, we primarily just get more inflation– is in fact a negative number. I do not accept the proposition that there is a level of government spending– however large a number you choose to suggest– that will prevent the unemployment rate from rising above 8%. But I do believe that if the government borrows a sufficiently large amount, we will have to worry in a very concrete way about what will sustain the foreign demand for U.S. assets.

What, then, do I propose? The first principle that’s quite clear to me is that drops in state and local government spending, or increases in state and local taxes, are a likely response to the current situation and are clearly counterproductive. Hence I’ve advocated ([1], [2]) additional federal borrowing in order to provide unrestricted block grants to states. That’s a simple, effective plan that could and should be immediately implemented, while still preserving complete flexibility in responding to our serious longer run challenges.

I also am very comfortable endorsing additional government investment in infrastructure for which the argument can be made that the facilities could make a significant contribution to future productivity. At the top of my personal list would be investments in the electricity transmission grid, mass transit, and basic scientific research.

And unquestionably the number 1 priority for the federal government should be to restore a functioning financial sector. That in my mind should be done in a way that maximizes the return on any taxpayer funds invested.

But rushing through new government spending plans, just for the sake of spending? Count me off of that bandwagon.

Originally published at Econbrowser and reproduced here with the author’s permission.