Social Security, A Slightly Different Take

There has been quite a bit of wrangling lately about whether or not there is a Social Security crisis and if there is, exactly when it becomes a crisis. Instead of wading in to that again, I’m going to talk about how Social Security has changed, and how it could change again. (If you are interested in a good take on ‘crisis’ I suggest this post by Jane Galt.)

For purposes of this post, I will be talking about the old age portion of Social Security. I'm aware that there is also a disability portion. It is conceptually severable from the old age portion. I'm aware that it exists, but it isn't what I'm talking about.

At its inception, Social Security was politically positioned as a pension-like program. You would work, pay into it, and if you got lucky you might get a little bit out of it when you retired. The retirement age was set such that most men died before they would collect a single check and most of those who retired would collect checks for less than three years. This made many choices about how to set up the program look very easy. It made funding it in a pay-as-you-go fashion very attractive because there were dramatically more workers than there were pensioners. Making it a universal program despite rhetoric about it as a safety net was a cheap way to broaden political support because most people didn't live very long after retirement. In short, Social Security was designed around a number of assumptions which are no longer true. Talking about changing it is not an attack on The New Deal. It is a reevaluation of ideas which where enacted under a set of assumptions which do not apply to modern Social Security. The heart of Social Security is protecting people from poverty in their old age. The way it is set up now helps protect people from old-age poverty while paying tens of billions of dollars every year to retirees who are not poor.

Even if it needs changing, why would we change it now? If it isn't a crisis now, why not wait? The answer to that is simple. The system as currently set up, even if not immediately in crisis, represents a huge and growing portion of government expenditures. It is also something that people make their long term retirement plans around. If you are ever going to make changes to a program like this, it makes sense to phase them in over a long period of time so that those who have made plans under one set of assumptions are not damaged by the shift to another set of assumptions. Social Security is exactly the kind of program for which if it is going to be in crisis in 40 years, you want to start phasing in the changes now. The conceptual framework of workers paying into Social Security as we go along is not going to work 40 years from now to pay off all the benefits unless we have hugely unexpected amounts of long term economic growth. If that happens, we have no problem under almost any proposal. But if it does not, we will have more money going out of the program than coming in, a position which cannot continue indefinitely. There are many ways which we could correct that imbalance. We could dramatically raise the retirement age. We could almost double the amount of payroll taxes taken in. We could do something else. The thing to remember is that there is no free lunch. When more money is going out than is coming in, something has to give eventually.

We have allowed the mechanics of Social Security to get away from us by drift instead of intention. A plan which was designed to protect against old age poverty now give tens of billions of dollars a year to people who are not poor. A plan which was designed to avoid paying most people by making a high retirement age now has a retirement age which can leave huge numbers of people in the system for more than a decade. A plan which was designed with a 15-to-1 worker to retiree ratio faces a 3-to-1 worker to retiree ratio. These changes have creeped up slowly but have changed the underlying deal dramatically.

My proposal would be as follows. Please realize that the concepts are more important than the precise numbers. In other words the democratic process could negotiate the details. First, an old-age poverty program doesn't need to pay wealthy and upper middle class retirees. The cheap political cover isn't cheap any more. It now costs tens of billions of dollars per year. The mechanics of the current system (as influenced by the political pretense that it is a pension program) mean that the highest paid workers--those who are best able to plan and care for their retirement future--also get the highest benefits. There is a slight climbdown from this by making the highest benefits taxable, but this does not change the underlying fact that richer people are getting higher payouts. Much of this is tied into the conceit that this is a retirement or pension program. The analogy with a pension program explains why people who make higher incomes get higher benefits. It doesn't really make sense given the underlying justification for the program.

Retired Americans are typically much more wealthy than their peers. In each quintile of wealth, they are more wealthy (both with and without home equity) than all of their peer except those in the most wealth producing working years, 55-64. (See Table F in the Census Report on Net Worth and Asset Ownership of Households). They also have more wealth than those under 55 in the quintile above them. So when I talk about the wealthiest quarter of retirees I am talking about a population significantly wealthier than the top quarter of the general population. According to Table D, those in the the fourth quintile had an income of between $3,364 and $5,416 per month. That is a very liveable income, especially if you own your own house which a large percentage of these retirees do. Those in the fifth quintile of course have more income. (Note about census household statistics: elderly households are noticeably smaller than younger households so their per-capita wealth and income are understated by these statistics).

We could slowly phase out the Social Security benefits of the top quarter of income earners and come to a very large savings in Social Security outlays. I would suggest phasing benefits out at 3 1/3% per year over thirty years. Since the top quarter of Social Security beneficiaries earn more than the top quarter of benefits, this would save more than a quarter of old-age Social Security benefits. Gaming the system could be minimized by reducing the benfits on a less than 1 to 1 ratio compared with income. The income situation for those retiring more than 30 years from now should be better than just the current income situation minus social security benefits for the top quarter as upper middle class and wealthy people save more to compensate for the knowledge of reduced benefits.

What about privitization? I like the idea of the government encouraging privately controlled accounts in theory, but I am skeptical of it in practice. The major benefit of privitization plans is that it gets people used to actually being involved in decision-making regarding their retirement. I am skeptical about privitization plans for very conservative reasons--I think such plans would lead to nearly inevitable government attempts to tinker with market outcomes. Our financial system is not perfect, but it has a fair degree of efficiency. I am loathe to put that at risk. Even after a medium downturn of relatively short duration (say the market from 1999-2001) the pressure to meddle would become immense. At least one of the following two things would happen:

1. The government would intervene directly in the markets, ultimately making the markets much less efficient at exposing market weaknesses; and/or

2. The government would make up the shortfall, eliminating many if not all of the projected savings from privitization. Knowing this, some people may take especially risky investments, knowing that if they fail the government will bail them out.

I'm not interested in a scheme which I suspect will lead either to little savings or extreme government meddling with the financial market structure.

UPDATE: My initial link to the Census Report appears to be non-functional. The link can be found here.

Posted by Sebastian Holsclaw at January 5, 2005 3:55 AM