The transition costs of personal accounts

Bill Shipman and Peter Ferrara write in the Wall Street Journal that "private Social Security accounts are still a good idea." It's not a view I particularly disagree with, especially given that I have written an article called "Still a Good Idea" that makes some of the same points as the Shipman-Ferrara op-ed.

At the same time, though, I think one reason Social Security reform didn't fare well when President Bush attempted it in 2005 was that his supporters entered the debate with unrealistic expectations of what personal accounts can do. I discussed this issue in a recent article in National Review, but the Shipman-Ferrara piece offers another opportunity to hash out these questions.

Shipman and Ferrara provide a good example:

Suppose a senior citizen—let's call him "Joe the Plumber"—who retired at the end of 2009, at age 66, had been able to set up a personal account when he entered the work force in 1965, at the age of 21. Suppose that, paying into his personal account what he and his employer would have paid into Social Security, Joe was foolish enough to invest his entire portfolio in the stock market for all 45 years of his working career. How would he have fared in the recent financial crisis?

The answer, Shipman and Ferrara write, is that despite the ups and downs of the stock market, personal accounts "would still pay them about 75% more than Social Security would have."

To understand what's missing in this example, however, consider that it's mathematically and logically equivalent to the statement, "If we eliminated Social Security benefits for current retirees we could give working-age Americans a big tax cut, which they could spend or save as they wished."

In other words, the Shipman-Ferrara example ignores the fact that once individuals divert their payroll taxes to personal accounts, where they could indeed earn higher returns on their money, they leave a gap in Social Security's finances that would need to be made up.

How big a gap? Social Security's actuaries calculate that the value of Social Security benefits that have been earned but not yet paid out is around $20.2 trillion. If paid off over the next 100 years—roughly the period over which accrued benefits must be paid out—these liabilities are worth around 6.6 percent of payroll or 2.2 percent of gross domestic product. So, if we allowed workers to divert their 12.4 percent payroll tax to personal accounts, we could ensure that current benefits continue to flow by levying an additional 6.6 percent tax on their earnings, for which no additional benefits would be paid.

Obviously, once you factor this additional "transition cost" into the equation, much of the increased return that individuals would receive through personal accounts goes away. As an approximation, if the 6.6 percent "transition tax" were paid out of the current payroll tax, leaving only 5.8 percent of wages to invest, benefits for the new retiree in 2009 wouldn't exceed current law by 75 percent but fall short by around 18 percent. Likewise, if you deduct a little more to cover disability insurance, shortfall would increase. While we can quibble about the details, the point is that transition costs aren't ancillary to the personal accounts debate; in truth, we can say that transition costs are the personal accounts debate, since without transition costs we don't see the "transition benefits" of higher returns and higher retirement benefits in later years.

The rest of the increased return on personal accounts disappears if you adjust for the cost of market risk. Stocks have higher returns than bonds, but only as a compensation for higher risk. While stocks pay good returns most of the time, these returns aren't guaranteed—and investors are willing to give up a lot of upside to protect against potential downside.

I've supported personal accounts in the past and I have no problem with them today, but I also argued in National Review (and a follow-up blog post) that conservatives need to make some cost-benefit calculations regarding what policies to propose. Competing with personal accounts is the desire to hold the line against tax increases for Social Security. I don't think it's likely that Social Security reform with personal accounts and without tax increases is likely to pass. Even President Bush all but said he would increase the "cap" on Social Security taxes as a way to get his personal accounts plan passed. My judgment is that it's both easier and more important from a policy perspective to hold the line on Social Security tax increases than to introduce personal accounts to the program. It would be nice to have both, but I suspect conservatives are going to have to make a choice.