Social Security: Reform, Refund or Opt-out? (Part 2)

Public Pension Concepts and Alternatives

Social Security reform combines three basic possibilities. It can raise contributions or it can cut benefits sometime in the present or future (pre-funding or pay-as-you-go (PAYG)). It can be private or public. It can be diversified or undiversified. The current system is public, has no diversification and is not pre-funded (PAYG). It features forced savings in that income cannot be spent before retirement. It pools social risk so that it can provide insurance against earnings loss, disability, inflation, and longevity. It redistributes income from high to low lifetime earners. The Social Security System is controlled and administered by the U.S. government and is a defined-benefit plan (DBP) .
“Pre-funding” is a plan to reduce the sum of the system’s implicit and explicit debt. As alluded to previously in the discussion of the vulgar error, the current system has inherited or legacy debt. This is something many politicians either willfully forget or ignorantly omit. Any analysis and reform must include provisions for the cost of inherited debt or it is a truly disingenuous and misleading discussion. Omitting taking care of this unfunded liability (which is estimated at around $10 trillion) is the vulgar error committed by the many politicians who compare returns from strictly switching to other assets. They do not mention deducting this debt from the alternate assets’ internal rates of returns (IRRs) to the IRRs of Social Security. The IRRs of the Social Security Trust fund for future generations account for the legacy debt.
The earliest cohorts received very high IRRs in real terms. An anecdotal and extreme example is that of my grandfather who worked for the Federal Reserve System which did not become part of the social security system until six months before his retirement in the mid 1960s. He actually deferred his retirement six months to buy into the social security system and qualify for benefits. He, and for awhile my grandmother, received nearly 15 years of monthly benefits for six months of his contributions. A truly amazing IRR and one for which his progeny will be paying for years to come.

Leimer (1994) has shown that workers born in 1876 received a real return of over 35 percent while worker born in 1900 receive a 12 percent real return. The inherited burden of payments to previous generations will transfer to whatever reformed program is selected. It will lower future IRRs by the amount already paid to that early cohort. Leimer’s data suggests that future cohorts will receive about a 1.5 real rate of return under the pure pay-as-you-go (PAYG) system that now exists. Since early cohorts received extreme high positive net transfers from Social Security, some present or future cohorts will receive negative net transfers unless benefits are reduced or taxes increased.
This legacy debt was further worsened when Social Security extended benefits beyond pension obligations to fund various survivor and disabilities benefits in the sixties during the “Great Society” schemes of the Kennedy-Johnson regime. Survivors insurance and disability insurance benefits accounted for $166 billion or one-third of Social Security benefits in 2004.
Some pre-funding adjustments were made in the social security reform initiated during the Reagan years when increased contributions and retirement ages were implemented. Despite the Reagan era tax increase being the largest in U.S. history, the changes were not nearly enough to retire the unfunded liability or perpetually sustain the system. Had it started and remained a fully funded system, all participants would receive market returns and there would be zero net transfers between generations. Since the early cohorts of this type of system would not get any positive transfers, the successes of early the Social Security program in poverty reduction among the elderly would most likely not occurred. However, that is a $10 trillion futile exercise at this point. So, now it should be understood that even if the fund goes completely private, the unfunded liability gets paid by some U.S. taxpayers, somewhere and somehow.
Privatization plans create individual accounts must create some type of ‘recognition bonds’ to cover the liability to the current recipients and payees. It would end all government involvement with retirement income. The ‘recognition bonds’ would be funded with new explicit debt and the new system would be set up with full funding so that when the last person under the current system dies, the last of the recognition bonds is retired. A private system can also be set up as PAYG. Pre-funding a privatized system would eventually raise the IRRS to future generations, but only at a cost to early cohorts so generational redistribution still occurs in private systems depending on their initial funding terms. This is a truly dismal outcome for those currently in the system. They are still responsible for the high level of benefits paid to previous generations under PAYG. Setting the new system up as pre-funded would transfer the eventual high IRRs to the future generations. They would be hit with the costs of both generational redistributions.
Besides pure privatization and pure public provision, there are blended models. One such is example is where some groups of people or all can ‘opt-out’ of the public system if covered by an at-least-as-good-as-the-public-system private account. There is still some examples of this feature in the current Social Security System which initially allowed workers covered by other government plans to opt out. Employees covered by state plans in Louisiana are ‘opted out’ and currently covered by a state plan. U.S. Congressional Representatives and Senators ‘opted out’ and have their own plan. ‘Opting-out’ is no longer an available option in the current system.
There are also suggestions of combined plans where individuals hold both private and public accounts with potentially different investment strategies. Again, the pre-funding challenge exists in both private and public systems. The choice there is to raise contributions or cut benefits to payoff the unfunded liability of the inherited debt and to determine what parts of the system will be fully funded or PAYG from the reform plan forward. Any blended program would most likely pull younger workers away from the current system and it has been estimated that it would drain about10 year of funds out of the current system. Additionally, those most likely to opt-out would be the rich who already have some type of diversified retirement savings. This leaves the current system to be funded by elderly and low income workers.

What types of costs accrue to privatization on top of the unfunded liability of the inherited burden? There are costs that are unique to privatization. First, there is the loss of social risk pooling. With a large group of persons it is much easier to insure against shocks in earnings, longevity, or inflation. It is also easier to insure folks at lower costs because of the pooling of that risk and the size of the pool usually lead to economies of scale. Inflation-indexed bonds may have to be introduced as assets to hedge against inflation risk should privatization occur. This type of low default risk asset would alleviate the inflation risk.
Hedging for earnings shocks and longevity in private accounts would be much more complex. Allowing any type of opting out reduces the size of the benefits of social risk pooling. In fact, many talk show hosts on the right use the argument that demographics suggest that whites out live blacks. These hosts argue that blacks should supporting opting out so they could take advantage of potentially higher returns during their lives. Rather than focus on finding ways to increase black longevity, they suggest it is ‘unfair’ to ask shorter-lived blacks to subsidize longer-lived whites in the social security system. Much of the original system was in fact set-up to insure that long-lived widows would not spend their golden years in abject poverty given the longevity of women. Insurance for longevity and for disability does exist in the private sector. It is very expensive. Currently, there is an adverse selection problem, because those who expect to live long, become disabled, or have spouses that will outlive them are the ones that purchase private annuities. Folks that feel the coverage will not be necessary opt-out. This leaves a smaller, riskier pool with higher costs of coverage. The Social Security system currently groups every one into one huge risk pool including those who are likely to die young and those who are most likely to live long. So while it is true that some groups subsidize others, it is based on pooling for a number of risks that probably no one person may experience. Chances a household benefits from at least one of those pooled risks.
Second, there will be higher administrative costs with private accounts. U.S. Social Security administration costs are one-fourth of those for private pension systems (on an active contributor basis; Reid & Mitchell, 1995). There is also research suggesting mutual fund managers and other private pension account managers that derive income from fees will practice churning to increase fee income. Potential agency problems exist that are usually not discussed in the public policy forum. These problems would potentially increase with the increase in asset choice in private accounts.

The current public system could be diversified by taking overpayments and investing the proceeds in equities through the trust fund. This was much discussed in the 1990s when the equity markets were booming. It was given so much serious attention that then chairman of the Federal Reserve System Alan Greenspan felt compelled to testify to congress that allowing it “has very far-reaching potential dangers for the free American economy and a free American society”.
A private system could be diversified if it permits individual accounts to hold equities and bonds or it could be undiversified by requiring individual accounts to hold bonds. Either way, risk enters the equation. While the annually inflation adjusted means of returns on stocks represented by some diversified portfolio like the S&P 500 are much higher than any government bond, so are their standard deviations. A retiree has just as much chance to reach 65 at a year on the positive side of that deviation as on the negative. Geanakopolos et al (1998) calculate the annual inflation-adjusted mean return on the S&P 500 was 9.4 compared to an intermediate-term government bond mean return of 2.3. The standard deviations are just as diverse. It is 20.4 for the S&P 500 and 7.1 for the bond. A mean, however is a mean. The market is not Lake Woebegone where all the bonds are paid in full and all the equities’ returns are above average.
What other types of costs are involved with diversification? There could potentially be a social and/or personal cost or risk associated if an account permits workers to make uninformed, bad portfolio choices. This could happen under private personal accounts or personal accounts managed by the Trust Fund with diversification and choice. If you have personal funds, and you allow diversification, then you run into the distinct possibility that two people can either retire at the same time with completely different payments or one person retires slightly earlier than the other and because of some exogenous shock, one person’s retirement is funded well, and the other person’s retirement must be put off indefinitely until the market improves. A broad range of assets carry a broad range of risks and the market has both winners and losers. If differentiation exists, what will become of a loser? Does the U.S. put them into some other guaranteed income program or let them spend retirement on the street?

Next Up in Part 3: Lessons from the World