« "Packed with Economists"? | Main | Shiller: Private Accounts a Bad Deal »

March 19, 2005

Yes, Bush Private Accounts Are a Bad Deal

Jonathan Weisman writes a pretty good article about Robert Shiller's calculations in "The Life-Cycle Personal Accounts Proposal for Social Security: An Evaluation". The 3% real interest rate on the clawback of contributions to private accounts is too high to make them a good deal. Shiller's right:

washingtonpost.com: Retirement Accounts Questioned: Nearly three-quarters of workers who opt for Social Security personal accounts under President Bush's 'default' investment option are likely to earn less in benefits than those who stay with the traditional Social Security system, a prominent finance economist has concluded. A new paper by Yale University economist Robert J. Shiller found that under Bush's default 'life-cycle accounts,'... a third of workers would bring in less in benefits than if they remained in the traditional system... [if future returns are like] historical rates of return in the United States. Using global rates of return, which... more closely track future conditions, life-cycle portfolios could be expected to fall short of the traditional system's returns 71 percent of the time.... Shiller used 91 computer simulations to analyze the past performance of stocks and bonds in a variety of portfolios. He measured the returns in 44-year increments, beginning in 1871, to approximate a worker's lifetime contributions to personal accounts.

The results 'showed a disappointing outlook for investors in the personal accounts relative to the rhetoric of their promoters,' concluded Shiller, a leading researcher in stock market volatility who gained fame in the late 1990s for his warnings of a stock market bubble. Shiller's paper... is adding to research that suggests the White House has been overly optimistic in its assumptions about personal investment accounts. A recent paper by Goldman Sachs economists said the White House's anticipated 4.6 percent rate of return above inflation could be nearly 2 percentage points too high.... Under the Bush proposal, workers would be better off choosing private accounts only if those accounts earned annual returns that exceed inflation by 3 percent.

'I'm one of these people who maintain the 3 percent rate is too high a trade-off,' said Jeremy J. Siegel, a finance professor at the University of Pennsylvania's Wharton School and a longtime advocate of stock investing. 'You can't get 3 percent in the market anymore.' Trent Duffy, a White House spokesman, said the administration is not contemplating changes to the proposal at this point. 'We're confident returns on the market will be well in excess of what we need to make the program work well for seniors,' he said....

[T]he 3 percent hurdle appears too high for many to clear, Shiller found, especially with the conservative strategy the administration has embraced. According to U.S. historical rates of return, the life-cycle portfolio fell short of the 3 percent threshold 32 percent of the time, meaning nearly a third of personal account holders would have been better off sticking with the traditional Social Security system. The median rate of return was 3.4 percent, barely better than the traditional system. Upon retirement, accounts would yield an annuity payment of about $1,000 a year, 'hardly a windfall,' Shiller said.

But he also adjusted for what he expects to be lower future rates of investment return by using historic rates of return from international stock and bond markets.... The life-cycle portfolio under these adjusted returns lost money compared with the traditional system 71 percent of the time, with a median rate of return of just 2.6 percent, $2,000 less in annual benefits than those of workers who stick with the traditional system. 'To say that there is a money machine in the stock market, that it can be tapped to yield great wealth without significant risk if one uses life-cycle investment methods, is a big mistake,' Shiller concluded...

I've been talking about this here, here, and here. But Shiller has real Monte-Carlo simulation results.

Posted by DeLong at March 19, 2005 11:55 AM