As George W. Bush's second term begins, no item on his agenda is more controversial than Social Security privatization--that is, allowing Americans to divert at least some of their payroll taxes into personal accounts that they can invest in mutual funds or similar instruments. So far, Bush's own plan has been maddeningly vague, but it has opened up a serious debate about transforming a government program that was once so sacrosanct that it was called "the third rail of American politics."

Is privatization necessary? Preferable? Politically viable? In January, with Bush's second inaugural in the offing, Reason invited James K. Glassman, a fellow at the American Enterprise Institute and host of, to discuss and debate the ins and outs of Social Security reform with Tyler Cowen, the Holbert C. Harris professor of economics and director of the Mercatus Center at George Mason University.

Social Security's Fortuitous Crisis

James K. Glassman

This August marks the 70th anniversary of Social Security. If it ever had a reason to exist, it doesn't any more.

Social Security has three big problems, all of which can be solved by allowing Americans to invest--on their own --part of what they are now forced to send to Washington in payroll taxes.

First, Social Security is a blight on liberty. It extracts a big chunk of the pay of working people to benefit the retired. The vast majority of Americans are perfectly able to handle their own retirement savings, just as they buy their own food, clothes, and shelter. Even if we concede that government needs to force all Americans to save today so that taxpayers won't have to chip in to prevent the profligate from starving in their old age, why do we all have to buy the same annuity with the same terms from the same provider--that is, the government itself?

Second, Social Security creates a vast moral hazard. Since its implicit message is "Don't worry--Uncle Sam will protect you in your old age," it's a huge encouragement to spend rather than to save for your own or your family's needs.

Third, Social Security is a terrible retirement system, a Ponzi scheme that was inevitably going to collapse of its own weight. Instead of a conventional retirement account, with real assets accumulated over time, it was constructed as a pay-as-you-go plan: Current workers pay the bills of current retirees. (There's a small amount left over for a so-called trust fund, which predictably has become a piggy bank for the rest of the government.)

The first retirees in the system were big winners. For example, the very first recipient, Ida May Fuller, paid in $44 and collected benefits of $20,934. Times were different then. In the 1930s, 11 workers supported each retiree. The ratio is now 3.3 to 1. Soon it will be 2 to 1 and stay there. In 1929 life expectancy was 57; today it is nearly 80. Americans receive benefits far longer than they did before, and they can start earlier. Meanwhile, work force growth is slowing, and benefits are rising faster than inflation, since they are geared to wages, not prices.

As a result, by 2018, according to current predictions, retiree benefits will exceed workers' taxes paid into the system. Social Security will then present its IOUs to the Treasury for payment. The only way to get the cash will be to raise taxes, cut other government programs (fat chance), or borrow like crazy.

The system has been saved from insolvency in the past mainly by higher taxes. As a result, the actual returns that workers will receive from their contributions are minuscule: an estimated 1.5 percent annually after inflation for a typical person born in the last 30 years. Compare that with the average yearly return since 1926 from an account invested half in a stock index fund and half in Treasury bonds: 5 percent after inflation.

It's not surprising that young people especially have little taste for Social Security and little faith it will survive. The good news is that the impending disaster offers an opportunity to fix Social Security once and for all. The solution should:

� Guarantee the benefits of everyone now getting them, as well as others on the brink (say, those 55 and older).

� Gradually index future benefits to inflation, rather than wages, and increase the retirement age (currently 65 for those born before 1960, 67 for those born afterwards) by another year or two.

� Allow those under 55 to opt out of the current system by investing up to half of the retirement part of their payroll taxes (which totals roughly 10 percent of pay, including both the employee and employer contributions, for middle-income Americans) in an account with mandatory provisions that would restrict investment choices and require phased withdrawals starting perhaps at age 60 or when sufficient funds are acquired. Otherwise, ownership of the account would be unfettered and would belong to the worker and his heirs. It could be used to buy an annuity or simply provide needed income.

� Reduce, accordingly, the Social Security benefits of those opting out.

As the new system proves successful, gradually allow all former Social Security retirement deductions to go to personal investment accounts and broaden choices for those accounts.

Just before the last election, Investors Action, a new advocacy group that I chair, commissioned Public Opinion Strategies to poll likely voters on their willingness to switch to a system like the one I've described above. Among all voters under age 65, a slight plurality wanted to switch, 49 percent to 46 percent. But among those aged 35 to 44, the margin favoring switchers was significant: 61 percent to 33 percent. Among those under 35, some 64 percent wanted to switch.

Despite Social Security's reputation as the "third rail" of politics, intelligent discussion of the issue has been possible. Such congressional advocates as Rep. Jim Kolbe (R-Ariz.) have called for personal accounts repeatedly and have been elected repeatedly. President Bush said during the last campaign that he backed reform, implying that he supported some form of privatization, although he didn't spell out the details (and still hasn't as of this writing).

Bush evidently understands that Social Security reform could transform politics more broadly. Public Opinion Strategies, in another poll we commissioned after the election, found that investors (defined as people owning stocks or bonds, individually or through mutual funds) voted for Bush over Kerry, 52 percent to 46 percent, while noninvestors voted for Kerry over Bush, 54 percent to 45 percent.

What was striking was that the investor-Bush link remained strong at all income and demographic levels. For example, among those making less than $40,000 a year, a Democratic stronghold, investors were almost evenly split, with 47 percent voting Kerry and 46 percent Bush. But noninvestors voted for Kerry massively, 57 percent to 36 percent.

There's a good case to be made that becoming an investor increases a person's stake in free markets. Reforming Social Security using personal accounts would probably increase the proportion of Americans who are investors from about 50 percent today to about 80 percent in a few years.

Can Americans handle this much freedom? More than two-thirds of Americans (and 83 percent of married Americans) have bought a house--a far more difficult and risky investment than funding, over 40 years, a long-term account split roughly evenly, as I would advise, between stocks and bonds. Social Security treats Americans as children, keeping them in a kind of bondage, beholden to government for no good reason. The opportunity has come, at long last, to break the chains.

Cut the Benefits, Hold the Accounts

Tyler Cowen

We can improve Social Security, but let us opt for a sound transition. Unfortunately, some proposals would restrict liberty and responsibility rather than enhancing them.

I agree with James Glassman's first and second suggestions, namely that we should guarantee benefits for the current elderly and gradually index future benefits to inflation. The latter proposal would cause benefits to rise at a slower rate than otherwise, since current benefits are indexed to wages rather than to prices. It also would suffice to bring the system into future fiscal balance.

But here we begin to differ. Glassman proposes new private but government-regulated accounts. My alternative proposal is simple: Keep the limits on benefits but do not create the new system of accounts.

Of course we already have private accounts, which anyone can, if they wish, hold at Merrill Lynch, Bank of America, or Fidelity. Individuals can save in (relatively) unregulated fashion, and in some cases in tax-free or tax-reduced form (e.g., IRAs and 401(k) plans). It runs counter to both liberty and responsibility to set up a new system of "private" accounts with regulated contributions, investments, and withdrawals. We would end up adding to government involvement in the market economy, not diminishing it.

Whether we like it or not, government would come to be seen as offering an implicit guarantee for these accounts. If the stock market were to stay low for 10 or 15 years, retirees would demand that government supplement their returns. Government would feel pressure to cough up additional revenue exactly when the budget would be in the tightest squeeze. The result would be higher taxes in times of recession, exactly the opposite of what is appropriate. (It is true that government does not end up guaranteeing current IRAs, but that is only because we already have Social Security as a backstop.)

The private accounts would function as a kind of forced saving, as Glassman admits. But how much can government really force people to save through this kind of mechanism?

You can make people lock up funds in an account, but they can respond by borrowing more on their credit cards, taking out a bigger mortgage, and in general investing less in their future. The net increase in savings will be much less than the mandated increase. So the government will control more of our lives but without making our old age much safer. In essence we would see substitution from privately initiated saving to government-controlled saving. Is this really a step in the direction of liberty and responsibility?

And what will happen to the efficiency of private capital markets? Glassman already has admitted that investment choices will be "restricted." By whom? According to what standards? Will the government use these regulations to punish unpopular companies? How about companies that (supposedly) break the law? How about companies that supported the other political party? The proposal invites a politicization of investment decisions.

The Glassman proposal also would raise your taxes. Keep in mind that Social Security, whether we like it or not, is structured as a pay-as-you-go system. That means current taxpayers fund current retirees. If we allow current taxpayers to divert funds into private accounts, how do we fund the (constant) benefits to current retirees? It has to be either taxes or borrowing, and the latter means future taxes. Depending on how many people wish to switch to the private accounts, the size of this tax increase could run to the trillions.

In fairness to Glassman's proposal, it can be argued that these taxes would have been coming anyway. When subsequent generations of the elderly retire, we must pay taxes to fund their benefits. So arguably the Glassman reform is simply taking an implicit future liability and making it explicit in the present.

Unfortunately, the reality is grimmer than such logic would suggest. In theory it would work if government raised our taxes today, and then cut our taxes tomorrow as future retirees relied more on their personal accounts. But is that how governments operate? How often do they cut taxes just out of generosity? More likely they would keep the higher levels of taxation (and/or borrowing) in place. The plan would lead to a permanently larger role for government.

Whatever the problems with Glassman's idea, you might object, isn't the current system even worse? The correct comparison is not with the status quo but rather with caps on benefit increases. Those caps would bring fiscal stability to the system without requiring tax increases. Over time Social Security taxes could be cut. Individual saving would be encouraged. We could reap the higher returns available on private investment. At the same time we would not be breaking promises to current or near-term retirees. Even long-term retirees would receive a level of benefits at least as high as today.

And consider how benefit caps would operate in the long term. As benefits are held constant in real terms and the economy grows, the relative size of the Social Security program will shrink. Over time Social Security will become very small, in real terms, relative to the economy as a whole. This requires some patience, but it is the best recipe for privatization we have available. And it does not require additional regulation of savings or investment.

Glassman asks, "Can Americans handle this much freedom?" I think they are ready to handle more liberty than he is proposing.

Proceed With Caution

James K. Glassman

Tyler Cowen would reform Social Security by capping benefits, thus forcing people to save more on their own through existing private mechanisms, including IRAs, 401(k)s and taxable mutual funds, stocks, and bonds.

I am not sure what Cowen means by "caps." The implication is that he would stop increasing benefits, period. If inflation runs 3 percent annually, that would mean that in 24 years retirees would have half the purchasing power from their Social Security payments as retirees today.

Slowing the rise in benefits is part of my own proposal as well. I want to boost payouts to retirees each year according to the consumer price index, not according to the annual rise in wages.

Cowen's notion is only one step removed from the plan Milton Friedman outlined in Free to Choose and earlier writings: repeal payroll taxes immediately, keep doling out benefits to people currently in the system, pay all younger participants off (either in an annuity or in government bonds) for the value of what they have contributed, and let everyone fund his own retirement. Friedman would fund the reform by issuing new federal debt.

"This transition," he writes, "does not add in any way to the true debt of the U.S. government. On the contrary, it reduces that debt by ending promises to future beneficiaries. It simply brings into the open obligations that are now hidden. It funds what is now unfunded."

Let me be clear: I love the Friedman strategy, but it has two drawbacks. First, it is wildly impractical. There is zero chance that Congress will either shut down Social Security or, in the Cowen version, bleed it to death. Reformers need to be more subtle. Slowing the increase in benefits would, in a less brutal and more gradual way, force Americans to save more and rely on Social Security less.

The second problem is that in the United States we don't let people starve. Some people simply will not save for their own retirement--especially after 70 years of relying on the government. Even Hayek agreed that some form of forced savings is necessary to keep taxpayers from having ultimately to foot the bill for those who don't provide for their own future.

For this reason, and to meet political practicalities, mandatory elements are necessary in any reform plan--at least as a start. As I said in my first piece, these strictures can be phased out later as it becomes clear the plan is a success and as people get used to doing it themselves.

This is a good time to lay out the specific mandatory elements.

People who opt out of Social Security with, say, four percentage points (including the employer's contribution) of the 10 points that now go to the retirement part of payroll taxes must put that money into a personal retirement account (let's call it a PRA). The PRA would be privately managed and administered, but the investment choices would be limited--say, to six funds: two U.S. stock index funds (an S&P 500 fund and a broader, total-market fund), a corporate bond fund, a Treasury bond fund, a money market fund, an international stock fund, and a fund whose assets are split roughly 50-50 between stocks and bonds. A further requirement might be that persons over 50 would have to put all of their new PRA contributions into the last type of fund, to dampen volatility just before retirement. (A portfolio split 50-50 in this way has never lost money in any 10-year period since 1926.)

There would be restrictions on withdrawals from PRAs as well. We could say that, at age 60 (or at the time of the accumulation of at least $1 million in assets), a participant could either purchase an annuity or begin withdrawing 10 percent of the account each year.

I am not happy about this sort of engineering, but I think it is necessary to passing some kind of reform. Also, I agree with Cowen that these protections are not foolproof. Some people might borrow away all their savings.

The counterargument to this compromise plan is that Democrats, who understand that a nation of investors is the death knell for New Dealism (and thus the end of their party as we know it), will oppose Social Security reform in any form, so why not the best? Perhaps, but my strong impression is that many Republicans are squeamish about major changes in a popular system that dates back to FDR. Proceeding cautiously, putting a firm foot in the door, is a better approach.

I'm Not Milton Friedman

Tyler Cowen

We can and should reform Social Security without the mandatory elements postulated by James Glassman. Still, Glassman and I are in more agreement than he has noticed. He writes: "I am not sure what Cowen means by 'caps.' The implication is that he would stop increasing benefits, period." But I did note that "I agree with James Glassman...that we should guarantee benefits for the current elderly and gradually index future benefits to inflation."

So my plan would not create any new problem with starvation. Furthermore, rising real incomes over time will supplement these benefits, so each generation of the elderly will still be better off than the last, even without the government-managed private accounts Glassman calls for.

Glassman compares my ideas to Milton Friedman's plan, but the differences are significant. Most of Glassman's response thus really concerns Friedman, not my own notions of what we should do with Social Security. Some differences I have with Friedman: I don't think we can abolish the payroll tax, or that we can "buy people out" from the system.

To the benefit caps that we really agree on, Glassman wants to add government-regulated personal accounts. The economics here are straightforward, once unbundled from benefit caps. Our government would raise taxes (or borrow) to finance further private investment in equity markets. Furthermore those investments are to be regulated by the government. I expect that Reason readers, once they hear the plan described in these terms, will be suspicious.

Glassman makes two arguments about my proposal. First, he says it is too radical. But his points refer to Friedman's views, not mine. My own plan actually deviates from the status quo less than Glassman's does. We both agree on benefit caps. He wants to add an additional change--regulated personal accounts. I want to leave this choice in private hands. Of course, the mere fact that my plan is less radical does not mean it will be adopted. But a) it is a viable contender, and b) we should first put forward what is best to do.

Glassman's second argument is that my plan would lead to starvation. Again, this is confusing my plan (cap real benefit growth gradually over time) with Friedman's idea to shut down Social Security altogether. Furthermore, this emphasis on "starvation" is further reason to believe that if Glassman's regulated private accounts underperformed, government would step in to fill the gaps. We would end up with the worst of both worlds.

I maintain my suspicion that Glassman's proposed transition will lead to a long-run increase in taxes and the size of government. Yes, we are just trading in implicit for explicit liabilities. But as I argued, taxes and borrowing will go up in the short run under Glassman's plan. I worry about this outcome. I do not trust government to lower future taxes simply because it will not need the money to pay future Social Security benefits.

The First-Best Approach

James K. Glassman

Now that Tyler Cowen has clarified his position, he makes a good argument for a second- or third-best approach to Social Security reform: Slowly squeeze recipients through lower benefits, pressuring them to save more--but do nothing to set up special tax-advantaged retirement accounts. I worry, however, that, rather than promoting more personal responsibility for retirement saving, the Cowen plan would create a powerful constituency for reinstating the current benefit levels and even raising them.

We need to get Americans out of the habit of expecting government will provide them with most of their retirement income (the role Social Security plays for two-thirds of the nation) and instead demonstrate clearly that there is an alternative. My plan, which is probably close to what the Bush administration will propose, also calls for reducing benefits, by indexing them to inflation rather than to wages. But I believe that America's workers should have the choice of putting, to start, up to half of the retirement portion of their payroll taxes (about 5 percent of a middle-income salary) into a personal retirement account. That account, of political and social necessity, would have some restrictions on investment choices and on the timing and amount of payouts. Cowen makes meritorious libertarian arguments against such restrictions, but I am convinced a Social Security reform consisting simply of benefit cuts has no chance of passage--nor does the more sweeping Friedmanesque reform I described.

I reject Cowen's contentions that government would "step in to fill the gaps" if private accounts underperform, and that a scheme like mine will "lead to a long-run increase in taxes." First, government has never stepped in to make shareholders whole in the case of other "regulated" (as Cowen calls them) retirement accounts such as IRAs, 401(k)s, and the Thrift Savings Plan for federal workers, and the chances that such accounts will perform worse than Social Security itself over time are essentially nil, if we judge from history. Second, Cowen's own plan is just as likely to lead to tax increases. Almost certainly, as in the rescue in the 1980s, the formula for keeping the leaky Social Security system afloat will be a combination of higher taxes and lower benefits.

Finally, the establishment of private retirement accounts--even with restrictions--will be a blessing for tens of millions of low- and middle-income Americans who have little or no savings. As we've seen in Chile, when people are introduced to investing in stocks and bonds, they like the experience and want more. The problem today is that payroll taxes are so high that young people especially don't have the cash to save and invest. Reforming Social Security would change that.

Reform also should have a profound effect on the political system, which is why the left so fiercely opposes personal accounts. (The New York Times is almost hysterical on its editorial page.) In the last election, investors across the board preferred Bush over Kerry compared with non-investors. About half of American families hold stocks and bonds, and that ownership appears to change their perspective on capitalism and markets for the better. Imagine if 80 percent of U.S. families owned stocks and bonds. That's what Social Security reform would do.

Yes, Let's Choose the First Best

Tyler Cowen

I've already discussed most of the substantive points Glassman offers. I would like to note two points of terminology:

1) It is unusual to write, "Now that Tyler Cowen has clarified his position..." when I simply repeated what I had already said in my first contribution to this exchange.

2) Second, Glassman still misrepresents my opinion. He writes that my idea would: "slowly squeeze recipients through lower benefits, pressuring them to save more."

But I wrote: "I agree with James Glassman's first and second suggestions, namely that we should guarantee benefits for the current elderly and gradually index future benefits to inflation. The latter proposal would cause benefits to rise at a slower rate than otherwise."

More generally, I'll repeat my core characterization of the Glassman (Bush?) proposal. It suggests that the government should tax/borrow more money so that some American citizens can invest more in stocks. I'll vote no on this idea and hope that something better comes along.