The Social Security Administration projects that its trust funds will be depleted by 2033-not an optimistic forecast. But it may be even bleaker than that.
New studies from Harvard and Dartmouth researchers find that the SSA's actuarial forecasts have been consistently overstating the financial health of the program's trust funds since 2000.
"These biases are getting bigger and they are substantial," said Gary King, co-author of the studies and director of Harvard's Institute for Quantitative Social Science. "[Social Security] is going to be insolvent before everyone thinks."
The Social Security and Medicare Trustees' 2014 report to Congress last year found trust fund reserves for both its combined retirement and disability programs will grow until 2019. Program costs are projected to exceed income in 2020 and the trust funds will be depleted by 2033 if Congress doesn't act. Once the trust funds are drained, annual revenues from payroll tax would be projected to cover only three-quarters of scheduled Social Security benefits through 2088.
Researchers examined forecasts published in the annual trustees' reports from 1978, when the reports began to consistently disclose projected financial indicators, until 2013. Then, they compared the forecasts the agency made on such variables as mortality and labor force participation rates to the actual observed data. Forecasts from trustees reports from 1978 to 2000 were roughly unbiased, researchers found. In that time, the administration made overestimates and underestimates, but the forecast errors appeared to be random in their direction.
"After 2000, forecast errors became increasingly biased, and in the same direction. Trustees Reports after 2000 all overestimated the assets in the program and overestimated solvency of the Trust Funds," wrote the researchers, who include Dartmouth professor Samir Soneji and Harvard doctoral candidate Konstantin Kashin.
The administration's Office of the Chief Actuary produces the forecasts. Stephen Goss, the chief actuary since 2001, was unavailable to comment and has not seen all of the new research, an administration spokesman said.
However, Goss published a 2013 note that outlined the SSA's projection methodologies. The note was written in response to 2012 research by King and Soneji that found the Social Security Administration underestimated how long Americans will live and projected that the trust funds would be depleted by 2031, two years earlier than the government has estimated.
"The projections developed by the Office of the Chief Actuary for the Trustees Reports are intended to reflect all aspects of future possible trends in demographic, economic, and programmatic factors, given current Social Security law," Goss and other SSA officials wrote. King and Soneji's projections "were within the range of reasonable uncertainty as specified in the Trustees Report, and therefore should cause no alarm."
But King argues that the agency should reveal more about how they produce their forecasts and publish an annual evaluation of its forecasting performance to avoid forecasting bias. Social security programs in other countries and other U.S. government agencies, such as the Congressional Budget Office, the Census Bureau and the Bureau of Labor Statistics, routinely publish self-evaluations of their forecasts, he said.
King and the studies' co-authors confidentially interviewed former and current Social Security Administration actuaries involved in the forecasting process as part of their research to figure out why the forecast bias happened since 2000. He noted that actuaries are in the center of a political firestorm over the future of Social Security as lawmakers debate whether to cut benefits, raise taxes or a combination of both. "Actuaries worked really hard at being unbaised," he said. "We find no evidence that they bend to political pressure."
But the agency's attempts to be unbiased have created a bunker mentality, he said, which lead them to ignore evidence that Americans life expectancy has risen more than they have projected and how that could hurt the future funding of Social Security system.
"The issue is not the people," King said. "If you swap out the people and kept the same procedures, you would very likely get the same result." He urges the administration to openly share their data and methods with the public so outside researchers are able to replicate their forecasts and contribute to their improvement.
The bigger problem with the Social Security Administration is not disclosure, it's accounting, said Laurence Kotlikoff, a Boston University professor of economics and co-author of "Get What's Yours," a New York Times best-seller about how to maximize claiming Social Security retirement benefits.
Kotlikoff applauds the efforts for more disclosure, but wants the agency to calculate its liabilities using fiscal gap accounting, which considers the difference between the government's projected financial obligations and the present value of all projected future tax and other revenue.
The Social Security Administration projects its unfunded obligations out 75 years in the annual trustees' reports. Kotlikoff wants the administration to calculate unfunded obligations using the "infinite horizon," which accounts for funding after 75 years. Under this accounting system, SSA's projected unfunded liabilities would be $24.9 trillion (instead of the $10.6 trillion projected in 2088).
Trustees noted in their 2014 report that "the degree of uncertainty associated with estimates increases substantially for years further in the future."
Yet 17 Nobel Prize-winning economists have endorsed Kotlikoff's push for the SSA and other government agencies to use the fiscal gap accounting method more broadly.
"We have a situation that is like Enron accounting," Kotlikoff said. "And the public doesn't want to hear about it."
While King and Kotlikoff differ on how the Social Security Administration can improve its reports, they agree that better information about Social Security's financial health will enhance the public debate and help lawmakers fix the program before it's too late.
"Fair, transparent and accurate forecasts give Congress more of a chance to consider of all the policy proposals to preserve the solvency of Social Security," King said. "And it's easier to make changes to Social Security now than in the future."
More From CNBC