Should we save more or spend more? An Economic expoloration

In the media, in think tank research papers and in Congressional hearings experts testify that the baby boom generation is ‘unprepared’ for retirement – has not saved enough and needs to save more for retirement. There are a variety of policy initiatives aimed at increasing retirement savings. These include, for instance, Senator Harkin’s (D-IA) USA Retirement Funds that would, he says, “drastically increase retirement savings through automatic enrollment” of the 75 million US workers currently not covered by a retirement plan. The Obama administration has backed a similar autoIRA plan. And there are a variety of state proposals along the same lines. There seems to be a bipartisan consensus amongst those in the retirement security business that increasing retirement savings is a good thing.

Amongst policymakers not concerned with retirement security, however, there has been concern that, to recover from the recent recession and to improve a weak recovery, we need to stimulate spending.

Obviously, you can’t spend what you have saved. In this article we consider the question: what does it mean for the broader economy to pursue policies that increase retirement savings?

The ‘macro’ point of view

What, for the economy as a whole, would it mean for, for instance, the 75 million workers currently not covered by a retirement plan to ‘save more?’ Oversimplifying, there are two general views of this question, which we will call the ‘classical’ and the ‘Keynesian.’ (There are in fact nearly as many views as there are economists who have thought about these issues, but these two labels may serve to illustrate two fundamentally different ways of thinking about the question.)

The classical view is that if everyone decides to save more (or there is a net increase in saving), money is diverted away from consumption and into investment, away from consumer goods and services (to take a common example, a meal at a restaurant) and towards capital goods. Slightly more technically, as the supply of savings goes up, interest rates go down, and investments that were not profitable at higher interest rates now become profitable and get funded.

Generally, in a complex economy (one with significant division of labor and specialization), there is a system of intermediation that makes this shift – from consumption to investment – work efficiently: banks, investment managers and joint stock companies. Depending on how efficiently this system, and the markets for goods and services and for labor generally, work, there may or may not be a significant dislocation as workers in the consumer sector (e.g., waiters) shift to the capital sector (e.g., machine tools).

The paradox of thrift

In contrast, the Keynesian analysis of the concept of ‘saving more’ begins with the ‘paradox of thrift.’ (Again, this is an oversimplification – there are ‘Old Keynesians,’ ‘New Keynesians,’ ‘neo-Keynesians’ and ‘post-Keynesians’ all with different views on this issue. The ‘paradox of thrift’ analysis is probably more characteristic of ‘Old Keynesians.’)

There are different versions of this concept, but they all begin with the fact that a saving by one person reduces the income of another:

The paradox of thrift represents this basic economic truth. Spending and income are two sides of the same activity. One cannot logically analyze higher desired saving, and therefore a reduction in desired spending, without at the same time recognizing the consequent destruction of income. An alternative way to see the same point is to recognize that no individual can raise aggregate saving by increasing her own saving. An individual act of higher saving without a corresponding increase of investment only redistributes saving among agents …. [Emphasis added.] (A Penny Saved May Not Be a Penny Earned: Thinking Hard About Saving and the Creation of Wealth, Fazzari, 2007.)

A benign outcome, in this view, is that the increased saving of one person results in the increased spending of another, with no net effect on the system as a whole: A saves $1; the $1 A saves is lent to B, who spends it.

In the worst case, our restaurant owner, her income reduced, cuts back on her spending with knock on effects on C and D and – some Keynesians believe – a general lowering of income and, conceivably, a decrease in net savings.

The best outcome would be that A’s $1 is invested in some project that will produce future wealth (investment spending). ‘Investment’ here does not mean buying stocks rather than leaving money in the bank. In this context ‘investment’ means the commitment of savings to projects that will increase future productivity – producing more means of production. Distinctive to Keynesianism is the belief that the process by which savings is turned into investment is far more problematic and uncertain than the classical economists claim. Simply put: turning savings into investment is more easily said than done.

Three options

On either view (classical or Keynesian), there are three options for what happens with increased savings:

  1. Increased savings by one group can be offset by increased spending by another. Under this option, there is no net increase in savings; some save more and some save less. All that changes is who is doing the saving and who is doing the spending.

  2. Increased savings turns into investment.

  3. The additional savings is exported – e.g., additional US savings is lent to a foreign country or to foreign spenders.

Option 3 depends on factors (including, e.g., the value of the dollar relative to other currencies) that are far outside the scope of this article. For the rest of this article we’ll focus on Options 1 and 2.

So – how might we reconcile retirement policy initiatives with this theoretical background?

Option 1: increased savings by Group A, reduced savings by Group B

Some retirement policy fits into this paradigm. Proposals such as the Auto-IRA (or, e.g., Senator Harkin’s USA Retirement Funds proposal), expansion of the Saver’s Credit and various tax credit proposals all target increased savings by, generally, low-income workers. The tax credit proposals and proposals to limit total tax-favored savings (the ”$3 million cap” (this year, $3.2 million) in the Administration’s budget) and to cap retirement savings tax preferences (included in the Administration’s budget but also included in House Ways and Means Chairman Camp’s (R-MI) tax reform proposal) discourages retirement savings by high-income workers.

In this view, low-income workers are Group A and high-income workers are Group B. One can question the effectiveness of any of these policies: Regardless of tax incentives, do low-income workers really have additional disposable income that can be saved? And won’t, as some argue, high-income workers ‘save anyway,’ even if they don’t get tax incentives? But the policy itself seems coherent – rather than a net increase in savings it merely aims at tilting the distribution of savings away from Group B and towards Group A.

We began this article by raising the question: doesn’t the call for increased retirement savings raise questions that go beyond retirement policy, e.g., questions about the impact of more savings on the economy? Under Option 1, the answer is, more or less, no. Because under Option 1 the objective is not to raise savings overall but simply to change who saves.

Option 2: increased savings, increased investment

On the other hand if, as many policymakers say (including many who, e.g., advocate for Auto-IRA proposals), the objective is to increase total savings, questions about the effect on the economy as a whole are raised. We are, in effect, in Option 2, and based on our brief and summary review of the thinking on this issue, such a net increase in saving will either:

  1. ‘naturally’ (in the classical view) result in an increase in investment, or

  2. result (under a Keynesian view) in either a reduction in national income or an increase in investment, depending on how well the economy’s mechanisms for translating savings into investment work.

If we understand it correctly, the Keynesian challenge to those who advocate increased net savings is: are you sure this will result in increased investment? And some, at least, would advocate focusing on the challenge of increasing investment rather than (or at least before) increasing savings.

What about monetary policy?

We should note that, until recently at least, encouraging savings by anyone was a policy at odds with the more or less explicit policy of the Federal Reserve discouraging savings.

Treasury Inflation Protected Securities (TIPS) are Treasury securities indexed to the Consumer Price Index (CPI). Setting aside arguments about the accuracy/reliability of the CPI, TIPS can be thought of as a proxy for the ‘real’ rate of return – the interest rate minus inflation. At the low point – in the late summer of 2012 – the real return on 5-year TIPS was -1.67%. A saver who bought a 5-year TIPS was ‘paid’ a negative 167 basis points.

The Fed, through interest rate policy and Quantitative Easing programs, has pursued a policy of keeping interest rates low. Interest rates are what savers earn. But with interest rates at historical lows, some have characterized the current situation for savers as one of ‘return-free risk.’ Interest rates are also what drive annuity prices – lower interest rates make annuitization more expensive.


Finally we want to raise the question of demographics. In a demographically smooth world, it is theoretically conceivable that current workers savings could be used to finance the retirement spending of retired workers. This theory, which more or less underlies PAYGO retirement systems, is a version of Option 1. Here the members of Group A (the savers) are current workers and the members of Group B (the spenders) are retirees. In this sort of system, an increase of ‘saving’ (e.g., an increase in FICA taxes) would (or might) translate into an increase in retirement benefits and in spending by retirees. This system depends on a trans-generational promise that every currently working generation will have its retirement spending financed by some future working generation.

In a demographically ‘lumpy’ world, this picture is less neat. As the baby-boom generation retires, the supply of spenders (retirees) will go up relative to the supply of savers (current workers). This challenge is distorting almost all debates about fiscal policy. According to the latest CBO report, by 2024 “federal outlays for Social Security and for Medicare, Medicaid, and other major health care programs … will represent more than half of the federal budget and 12.4 percent of GDP.” Dealing with this challenge is proving to be problematic for both Democrats and Republicans.

One theory (voiced more ‘in the street’ than by policymakers) underlying some of the ‘save more now’ rhetoric is: Getting younger workers to save more is necessary because the amount they will get from Social Security/Medicare will be less than the current generation is getting, or, indeed, Social Security/Medicare may not be there for them at all. In effect, current workers, and particularly younger current workers, are being asked to both pay for the retiring baby-boom generation’s retirement and pre-fund their own. This theory implies a partial or total reneging on that trans-generational promise.

Obviously, this view/theory of the current situation is not held by everyone or even most, but it is, in a casual way, voiced by many: “Social Security won’t be there when I retire.” It’s hard to see this outcome as anything but a reduction in someone’s income – either of baby-boom retirees (through a reduction in benefits) or of current workers (through an increase in FICA taxes or a need to save to replace lost future benefits) or both. The only other alternative would seem to be a burst of extraordinarily productive investment – something close to a windfall.

* * *

We began this article with the question: what does it mean for the broader economy to pursue policies that increase retirement savings? Our (very tentative) conclusions are:

  1. If what is meant by ‘increasing saving’ is simply that one group should save more and another group should save less, then there may be no larger economic question.

  2. If what is meant is increasing net saving, then such an increase, generally, must (or, if you are a classical economist, will) be accompanied by an increase in investment (defined as ‘producing more means of production’). In this regard, some will argue that rather than focusing on increasing savings, the focus should be explicitly on increasing investment.

  3. The current demographic challenge – financing the baby-boom retirement – is disrupting retirement policy (along with tax and fiscal policy generally). Increased savings by current workers may ‘solve’ this problem but, perhaps, at the cost of a reduction in income.