By 
Janet Novack, Forbes Staff
Over the last 15 years, the Social Security Administration’s Office of 
the Chief Actuary has consistently underestimated retirees’ life 
expectancy and made other errors that make the finances of the 
retirement system look significantly better than they are, a new study 
by two Harvard and one Dartmouth academics concludes. The report, being 
published today by the Journal of Economic Perspectives,
 is the first, the authors say, to compare the government agency’s past 
demographic and financial forecasts with actual results.
In a second paper appearing today in 
Political Analysis,
 the three researchers offer their theory of why the Actuary Office’s 
predictions have apparently grown less reliable since 2000: The civil 
servants who run it have responded to increased political polarization 
surrounding Social Security “by hunkering down” and resisting outside 
pressures—not only from the politicians, but also from outside technical
 experts. 
“While they’re insulating themselves from the politics,
 they also insulate themselves from the data and this big change in the 
world –people started living longer lives,’’ coauthor 
Gary King, a leading political scientist and director of Harvard’s Institute for Quantitative Social 
Science, said in an interview Thursday. “They need to take that into account and change the forecast as a result of that.”
In its annual report last July, Social Security predicted its old age
 and disability trust funds, combined, would be exhausted in 2033 and 
that after that point the government will have enough payroll tax 
revenues coming in to pay only about three quarters of  promised 
benefits. King said his team hasn’t estimated how much sooner the fund 
might run out, but described it as in “significantly worse shape” than 
official forecasts indicate.
In addition to underestimating recent
 declines in mortality (i.e. increases in life expectancy) for those 65 
and older, the Actuary has overestimated the birth rate—meaning the 
number of new workers who will be available to pay baby boomers their 
benefits 20 years from now , the researchers assert.
Before 2000, the 
Actuary also made errors, but they went in both directions and the 
Actuary was readier to adjust the forecasts from year to year as new 
evidence came in, King said. Since 2000, he added, the errors “all are 
biased in the direction of making the system seem healthier than it 
really is.’’
A Social Security spokesman said today that Chief Actuary Stephen 
Goss couldn’t comment on the papers because he wasn’t provided them in 
advance and is tied up today in meeting with the Social Security 
Advisory Board Technical Panel. But he pointed to an 
Actuarial Note Goss and three colleagues published in 2013 in response to a 
New York Times op-ed by King and one of his current coauthors,  
Samir Soneji, an assistant professor at Dartmouth’s Institute for 
HealthPolicy
 & Clinical Practice.
In that op-ed, they attacked the Actuary’s 
methods of projecting mortality rates and predicted the trust fund would
 be depleted two years earlier than predicted. In their response, Goss 
and his colleagues called Kind and Soneji’s methods of predicting death 
rates “highly questionable” and noted that the Actuary’s methods have 
been audited since 2006 by an independent accounting firm and received 
unqualified opinions.
The dust-up might be ignored as bickering by the pointy heads, if it weren’t so consequential.  In a 
recent Gallup survey,
 36% of workers said they were counting on Social Security as a major 
source of retirement income. Differences over the estimates are 
important, King observed, because they affect “basically half of the 
spending of the U.S. government,’’ including Medicare.  Moreover, the 
forecasting assumptions affect the projected impact of any proposed 
changes to the program.
In their political paper, King, Soneji and
 Konstantin Kashin, a PhD candidate at King’s institute, recount how 
partisan fighting over Social Security intensified in the late 1990s, 
when conservatives began arguing the program was unsustainable and 
should be partially privatized, with younger workers offered individual 
savings accounts. In 2001, newly elected President George W. Bush 
appointed a commission intended to support such a change, but he put the
 issue aside after the September 11 terrorist attacks. After his 
reelection in 2005, however, Bush started pushing for changes in a 
series of town halls and speeches that, the paper notes, put the Social 
Security actuaries under “an extreme form of political pressure.’’ 
Democrats
 and news reports pointed to changes in the language used by the Social 
Security Administration that seemed (in line with White House policy) to
 emphasize that the program was not financially sustainable. Goss openly
 clashed with a Republican Social Security Commissioner.
Bush’s 
privatization push flopped and during recent elections Republicans have 
attempted to cast themselves as the protectors of Social Security, which
 enjoys strong support from voters across the political spectrum. In 
2013, after President Obama proposed a deficit reduction deal that, 
along with raising taxes on the rich, would have chipped away at 
inflation adjustments in Social Security, the idea was attacked by 
politicians from both parties.
But the problem of how to solve 
the system’s long term funding deficit has hardly gone away and the 
partisan divide seems to be widening again. 
Democrats have slammed
 a provision adopted by the new Republican Congress that they would 
block a transfer of money from the Social Security old age fund to the 
Social Security disability fund, which will be depleted next year. They 
say such transfers have been routine in the past and that it is a ploy 
by Republicans to force cuts tor retirement program too. Last month, 
Republican New Jersey Gov. Chris Christie, a possible Presidential 
candidate, proposed that the age for receiving full Social Security 
benefits be raised gradually to 69 and that benefits be limited for 
individuals with more than $80,000 in other income and ended completely 
for those earning more than $200,000.
King emphasized that there 
is “no evidence whatsoever,” that Goss and his actuaries are bending to 
political pressure from either Democrats or Republicans.  On the 
contrary, he said, while resisting such pressure, they’ve put too high a
 value on remaining consistent in their forecasts, in part because they 
don’t want to “panic” the public.  “They’re trying to show the numbers 
don’t change because they think it will inspire confidence. Maybe in the
 very short run it will inspire confidence by not changing the numbers. 
But having the numbers be wrong doesn’t  inspire confidence at all,’’ 
King said.
The political paper asserts that  Goss has resisted 
changes in forecasting assumptions suggested by the  Social Security 
Advisory Board’s Technical Panel on Assumptions and Methods—a panel of 
actuaries and economists that meets once every four years and is in 
session now. In some cases, the paper claims, the Actuary has made 
some  suggested change in an assumption, but then changed another, 
unrelated assumption in the opposite direction “to counterbalance the 
first and keep the ultimate solvency forecasts largely unchanged.” 
In
 their 2013 Actuarial Note, however, Goss and his colleagues say that 
while the 2011 Panel did push for faster changes in mortality 
assumptions, the panel’s recommendations, if adopted in full, would have
 actually resulted in a projection that the Social Security trust funds 
would run out a year later.
King, who presented his own findings 
to the Technical Panel yesterday, is pushing for one big change in the 
Actuary’s practices that he says the Panel has also favored: making all 
the Actuary’s data and methods open for scrutiny by others.  
“This
 is a period of big data. When you let other people have access to data,
 things like Money Ball happen,’’ King said. In addition to new 
algorithms, he said, the government actuaries need to take note of 
recent findings about unconscious bias by researchers and apply new 
methods social scientists have developed to guard against such bias. 
“Four
 hundred years ago you had people sitting in a monastery and thinking 
they thought great thoughts and that was their entire life,’’ King said.
 “Now we check on each other. If they would leave things open they’d 
have so much help and they’d be better off politically because their 
forecasts would be better.”