TCS Daily

Fighting Murphy

By Arnold Kling - June 30, 2003 12:00 AM

In his book Hard Heads, Soft Hearts, Alan Blinder formulated Murphy's Law of economic policy, which states that economists gain attention only when they take unreliable, controversial positions. When most economists agree on a well-supported idea, such as free trade, the media pays them little heed.

Social Security is an issue where the media debate is dominated by unreliable voices on both sides. The media serves as a megaphone for partisan think tanks, drowning out mainstream economics.

Social Security and the Baby Boom

Social Security is a system in which payroll taxes on young and middle-aged workers are used to fund benefits to retirees. This "pay as you go" model would not be an issue if population cohorts were constant. However, the Baby Boom and secular increases in longevity can play havoc with the system.

The table below presents a simple numerical illustration of the Baby Boom phenomenon. We have a population of seven people. There are three generations -- young workers, middle-aged workers, and retirees. In a steady state, there might have been just six people -- two in each generation. However, because of the Baby Boom, in the year 2000 there are now three middle-aged workers, and in the year 2025 there will be three retirees. Each worker produces $40,000 of output. Each retiree consumes $10,000 of output from Social Security transfers (he or she might consume more out of personal savings).

The Baby Boom and Social Security: An Illustration

As of year 2000 As of year 2025
Young Workers
2 2
Middle-Aged Workers
3 2
2 3
Total Workers
5 4
Total Output
$200,000 $160,000
Consumption of Retirees
$20,000 $30,000
Percent of Output Consumed by Retirees
10% 18%

The whole Social Security issue boils down to the last line of the table. The portion of total output consumed by retirees soars from 10% when the Boomers are working to 18% when the Boomers retire. Social Security has to find the money to pay those benefits. (For actual demographic data, Asymmetrical Information recommends a paper by Maureen Culhane.)

Don't Fret About Grandma

One conclusion that leaps out from the table is that grandma (someone who is retired in 2000) has nothing to worry about. Unless she lives another 30 years, the taxes needed to fund her Social Security are not onerous. The only reason to bring up the subject of cutting benefits to current retirees is to try to make a demagogic attack that consists of falsely accusing a political opponent of threatening to throw grandma out of her wheelchair.

What is up in the air is how the Baby Boomers' benefits will be paid for. Assuming that this consumption is paid for out of taxes on workers, we need a big honking tax increase in 2025. A concern that I have is that the tax increase required will be so big that it will discourage some people from working, which will lead to still higher tax rates, and still fewer workers, until the whole scheme implodes.

An Artificial Problem

If the Baby Boomers were to postpone their retirement, the problem would go away. There would be more workers producing output, and there would be fewer retirees to absorb Social Security taxes. In our numerical example, if only half of the Baby Boomers were retired in 2025 (think of them as working half time), then output would be $180,000 and Social Security payments to retirees would be $25,000. The proportion of output used to fund Social Security would be less than 14 percent, which would be much more manageable.

It is fair to conclude that the Social Security problem is largely an artifact of the retirement age. Raising the retirement age to something like 73 would make Social Security quite manageable. This increase in the retirement age does not need to affect anyone currently over the age of 50 (remember -- don't fret about grandma). In other words, the higher retirement age would not take effect for more than 15 years. By that time, continued improvements in health should make work a viable option for most Baby Boomers.

Of course, no one would be forced to work. People who are too unhealthy to work would still be eligible for disability. People who are healthy but want to retire before age 73 would have to finance their early retirement out of their own savings.

It is important to recognize that we are not necessarily doomed by demographics to suffer a reduced standard of living in 25 years. (It is even more important to recognize this in Japan and Europe, where, as Culhane's paper shows, populations are aging faster than in the United States.) A lower standard of living comes from demographics combined with the strict retirement age, which provides a strong disincentive for the elderly to work. There are mainstream economists on the left and on the right who agree that reducing the distortion caused by the retirement age is the single most powerful policy tool that we have available for addressing the Baby Boom Social Security issue.

Alternative Bailouts

Changing the retirement age would resolve the Social Security problem. There should be no need to look for alternative ways to bail out the system. However, we must examine alternative bailouts, because changing the retirement age is not yet on the table politically.

One possible bailout is productivity. Under a rather optimistic scenario in our numerical example, output per worker could double between 2000 and 2025, giving us $320,000 of output to distribute. As it happens, however, Social Security benefits are indexed to wages, which means that they too would double if productivity were to double. That would raise benefits to $60,000, which is still 18 percent of output. If instead Social Security benefits were indexed only to prices (meaning that benefits stay constant in inflation-adjusted dollars rather than increasing with productivity), then benefits would be $30,000, and the proportion of output going to retirees would actually fall under the optimistic productivity scenario.

Looking at the actual Social Security System (as opposed to my numerical example), and looking at a far less optimistic productivity scenario, John F, Cogan and Olivia S. Mitchell argue that changing from wage indexing to price indexing would resolve the outlook for Social Security. Unfortunately, they do the analysis in terms of Social Security's "actuarial shortfall," which in my opinion is an artificial and understated measure of the Social Security problem.

Regardless of whether price indexing would completely eliminate the Social Security problem, it would permit productivity growth to go a long way toward bailing out Social Security. Replacing wage indexing with price indexing is a way of reducing benefits in a manner that is fair and gradual.

Another possible bailout, which recently received a lot of play on Asymmetrical Information and other blogs, was suggested by Michael Boskin. Under the Boskin Scenario, as the Boomers withdraw savings from tax-advantaged accounts, such as IRA's, the tax payments that they make will be sufficient to cover Social Security. He argues that this is a large tax increase that is overlooked in the forecasting models that are currently used to project government receipts in the long term. One might argue that this form of a big honking tax increase would not be so bad, because it would not fall on younger workers and it might not lead to major reductions in work, thrift, or other helpful activities that taxes usually discourage.

The Savings Bailout

The Boskin Scenario illustrates another way to bail out Social Security. If we save a lot now, then we can afford to pay for retirement later. The Boskin Scenario works because Boomers' savings accumulate, giving rise to tax payments that the government can use to pay for the Boomers' Social Security benefits.

As the Boskin Scenario exemplifies, it does not matter whether the private sector or the public sector does the saving. Confusion over this point accounts for a lot of spilled ink and distorted arguments, particularly over the effect of the Bush tax cuts on Social Security.

Think of the tax cuts as shifting revenues from government to the private sector. Which sector is more likely to save those revenues? Assuming that tax cuts go to "the already satiated," as Robert Solow put it, the marginal propensity to spend might be on the low side, say 1/2. That is, for every dollar that people get in tax cuts, they spend fifty cents. If the government's marginal propensity to spend is higher than that, then overall national saving -- and our ability to pay for Social Security benefits in 2025 -- will actually increase as a result of the Bush tax cuts. Of course, if we assume that the government has a marginal propensity to spend of zero and the tax payments would have been used for debt reduction, then the tax cuts almost surely serve to reduce national saving and to make it more difficult to pay for Social Security in the future.

Stock Market Scenarios

Bailouts that I find particularly misleading are based on what I refer to as stock market scenarios. The standard assumption in such scenarios is that the inflation-adjusted returns on the stock market will be 7 percent. Because this is far above the growth rate in the overall economy, which is likely to be more like 2 to 4 percent, it leads to highly distorted conclusions.

For example, if you assume that the government issues a large volume of debt and uses that debt to purchase shares in stock mutual funds, then with 7 percent returns the stock market generates more than enough money to repay the debt and fund Social Security. Obviously, the same thing happens if we substitute "privatization" for government buying of mutual funds. Thus, the stock market scenario allows think tank partisans, such as Peter Ferrara, to concoct privatization plans that appear to be a "free lunch," causing the Baby Boomer problem to disappear by magic. The stock market scenario makes an appearance in Cogan and Mitchell's discussion of privatization, and to a lesser extent in the Boskin Scenario.

There are several reasons why it is dangerous to assume that there will be real returns of 7 percent on massive amounts of new funds invested in the stock market. First of all, stock prices really cannot grow faster than the economy forever. As this careful analysis shows, it is not possible for stocks to continue to provide the extra-ordinary yields that they have returned over the past 75 years. Second, even if the existing portfolios of stocks could outperform the economy, it is highly unlikely that trillions of dollars in new investments could flood into the market and still achieve such spectacular returns.

I happen to like the idea of privatization. It appeals to my values of personal responsibility and "right-sized" government. But I am completely turned off by the severe demagoguery and counter-demagoguery on both sides of the privatization debate.

Fighting Murphy

Almost all aspects of the public discussion of Social Security appear to be dominated by Blinder's "Murphy's Law" phenomenon. I hope that the readers will print out this primer and keep it as a handy reference when the overheated partisans start blowing their usual smoke.

The reforms that mainstream economists believe would make the biggest difference, such as raising the future retirement age and changing from wage indexing to price indexing, need to receive political consideration. In contrast, the politically salient alternatives of repealing the Bush tax cuts or relying on a free lunch from the stock market are much less sound economically.

TCS Daily Archives