This is the
graph [see Graph 1] which is causing so much anxiety in Washington this
winter. It shows the assets held in the Old Age, Survivors, and
Disability Insurance Trust Fund as a percentage of annual cost.
Alternative # II is the Social Security Administration's intermediate
estimate for the future, showing a 300 percent surplus in the Trust Fund
as of 2005, rising to 450 percent in 2018, plummeting to zero in 2042, and
falling thereafter into deficit.
Graph 1: Long-Range OASDI Trust Fund Ratios Under
[Assets as a percentage of annual cost]
Table 1: Ultimate Values1
of Key Demographic and Economic Assumptions
Total fertility rate (children per woman)
Average annual percentage reduction in
total age-sex-adjusted death rates from
2028 to 2078
Annual net immigration (in thousands)
Annual percentage change in:
Productivity (total U.S. economy)
Average wage in covered employment
Consumer Price Index (CPI)
Real-wage differential (percent)
Unemployment rate (percent)
Annual trust fund interest rate (percent)
1Ultimate values are assumed to be reached
within 5 to 25 years.
It is important
to remember that these numbers represent projections by government actuaries
seventy-five years into the future [see Table 1].
these projections on eight variables: the total fertility rate (or the average
number of children born per woman over her reproductive lifespan); the average
annual change in age- and sex-adjusted death rates; annual net immigration;
annual productivity changes; shifts in the average wage; the consumer price
index; the unemployment rate; and the annual trust fund interest rate. Again, the assumptions used to produce
the claim of “bankruptcy” by 2042 are the “Intermediate” list found here. In making these assumptions, the
actuaries largely relied on recent experience and then more or less assumed
that what has recently been, will continue to be.
On the one
hand, this seems altogether reasonable.
On the other hand, the one thing I can guarantee is that this
“intermediate” projection will be wrong. It assumes stasis — no change.
Over seventy-five years, the odds of a “shock” — or unexpected disturbance
— in one or more of these variables are almost certain. The problem, of course, is that the
nature and timing of such “shocks” cannot be predicted. To choose one relevant example, during
the 1930's the architects of the Social Security system assumed that the
nation's total fertility rate, which had fallen to 2.2 in 1935, would remain at
that low level, just enough to replace the population. They completely failed to foresee the
Baby Boom, which commenced in the early 1940's and raised the Total Fertility
Rate to 3.8 by 1957.
when Congress passed the Social Security Amendments of 1972, immediately
raising pension benefits by 20 percent and indexing all future payments to
increases in the average wage, it built its projections on the Total Fertility
Rate experienced in 1960-63, the tail end of the Baby Boom, namely 3.6. In fact, the TFR had already fallen
below 2.1 in 1971; and Congress was aggressively pursuing new population
control measures — such as the Title X birth control program — designed to
reduce that number still further.
Indeed, Senator Paul Fannin of Arizona blasted his colleagues of the
time who promoted this “extravagant” pension “spending scheme” based on
assumptions of rapid growth, while also advocating “zero population growth and
even zero economic growth in the name of ecology.” All the same, this popular, bipartisan measure passed almost
Now it is
true that the Social Security Administration does build in possible
variations. Note on this chart the
so-called “Low Cost” and “High Cost” projections, involving only modest changes
from the Intermediate numbers.
Even so, the possible effects are large. For example, the “Low Cost” fertility projection of 2.2
children born per woman is just as close as the “Intermediate” number to recent
American experience, where the TFR has been as high as 2.08. And the U.S. Census Bureau actually
projects a TFR for 2010 of 2.123.
If just this 2.2 figure is used, about 15 percent of the looming Social
Security deficit disappears. If
somewhat more flexible variations are used, as recommended by the Congressional
Budget Office, a TFR of 2.31 would eliminate a third of the projected
shortfall. More broadly, if
all of the “Low Cost” variables occurred — a total fertility rate of 2.2
instead of 1.95; a slowdown in longevity increases; 1.3 million annual
immigrants instead of 900,000; slightly lower wage increases, inflation, and
unemployment — the result is spectacular.
Indeed, the Social Security funding crisis disappears (as seen by line
#I on Graph 1): the Trust Fund is fat through 2080 and there is no need for
benefit cuts, higher taxes, or new debt.
However, for the same reason cited before, I can also guarantee we will
not achieve this result, either.
interesting question is the degree to which we might be able to influence one
or more of these variables.
Perhaps we are not helpless before these great material forces. Of still greater interest is the
possibility that the Social Security system itself influences these variables,
creating incentives that either strengthen or weaken the public pension system
as a whole.
is that fertility is just such a variable. Specifically, I will show that the New Deal architects of
Social Security crafted a system that favored a very traditional model of the
family, which in turn helped to create the conditions which spawned the
“Marriage Boom” and “Baby Boom” of the 1944-64 era. I will then explain how and why this system turned on its
head after 1965, generating incentives that discouraged the birth of
children. Third, I will assess
whether existing reform proposals would counter this systemic problem. And finally, I will suggest
alternatives that could make a difference.
A “Family-Friendly” System
New Deal still sparks political passions.
Conservatives tend to see it as creating a legacy of big government and
dependency. Liberals see the New
Deal as perhaps their finest hour.
Largely unappreciated today is the strong social conservatism found in
the New Deal.
political program was, in fact, decidedly pro-family. Some of this emphasis came from the Roman Catholic wing of
the old Democratic Party, represented by Father John Ryan. Author of the books, The Living Wage
and Social Reconstruction (1920), he served on the Advisory Board to the
Committee on Economic Security, which crafted the Social Security system. He argued for promotion and defense of
the breadwinner/homemaker/child-rich family, as outlined in the 1930 Papal
encyclical, Quadragesimo Anno.
Part of this pro-family influence also came from the New Deal women
called Maternalists: Secretary of Labor Frances Perkins; Grace Abbott and
Katherine Lenroot at the U.S. Children's Bureau; Mary Anderson at the Labor
Department's Women's Bureau; Molly Dewson on the Social Security Board; and
even Eleanor Roosevelt, herself.
These women all favored a family wage for fathers sufficient to support
a wife and children in dignity; opposed working mothers; favored mandatory
training in homemaking for girls; denounced day-care schemes as assaults on
childhood; favored mothers' pensions for widows; and repeatedly attacked and
thwarted the proposed Equal Rights Amendment.
assumptions shaped the whole New Deal domestic program, from the National
Industrial Recovery Act to the Works Progress Administration to the Social
Security Act of 1935. Regarding
the latter, Abraham Epstein — a member of the Committee on Economic Security —
laid out the new system's guiding assumptions regarding family:
American Standard assumes a normal family of man, wife, and two or three
children, with the father fully able to provide for them out of his own
income. This standard presupposes no supplementary earnings from
either the wife or young children.
before Congress on the proposed Social Security Act, Grace Abbott reported:
services are worth more in the home than they are in the outside labor
market and...she should be enabled to stay home and care for the
feminist historians frequently complain, women gained benefits under the new
Social Security Act primarily through their ability to conceive and bear
children. Title V provided funds
for prenatal, maternal, and infant care education. The new Aid to Dependent Children program provided mothers'
pensions to women who had lost a male breadwinner.
Even Old Age
Insurance, which seemed to be relatively individualistic and gender-neutral on
the surface, favored working men. The system covered only industrial workers,
overwhelmingly male. Most working
women in fact were in jobs then exempted from the system: clerical work; sales;
teaching; nursing; domestic service; farm labor; and charitable work.
Security Amendments of 1939 strongly reinforced this orientation toward the
breadwinner/ homemaker family.
Specifically, these Amendments incorporated the family responsibilities
of men into the 1935 system. Molly
Dewson, a key architect of this measure, explained the guiding principle:
Men who can
afford it always consider it their first duty to provide insurance protection
for their wives and children. Survivor benefits extend the same kind of
protection to families who need it most and can afford it least.
the 1939 Amendments included introduction of the “Pay as You Go” principle of
financing. Other changes included:
Aged women married for at least five years to eligible men
would receive an extra pension equal to 50 percent of their husbands'
benefits. Neither work nor prior
contributions would be necessary and divorced women were excluded. This is the so-called
Widowed mothers with children in the home were moved off of ADC,
receiving instead a monthly survivors benefit equal to 75 percent of the
pension her husband would have received, so long as she earned no more than $15
per month and did not remarry.
Surviving children received a benefit equal to half that which their
father would have received.
Amendments were a bipartisan measure: Republicans were as enthusiastic as
Democrats. Historians agree that
these reforms were overwhelmingly popular, and that they saved the fledgling
Social Security system from repeal.
They grafted onto Social Security the core “family values” of the
American people at mid-century:
marriage; the “family wage;” the mother-at-home; a flock of
children. Deviations from these
norms — divorce, illegitimacy, working mothers, deliberate childlessness —
faced significant financial disincentives.
system, I would maintain, helped to create what came after. Starting in the late 1930's, both the
U.S. marriage and fertility rates began climbing again. By the late 1940's, America had entered
an extraordinary time of family renewal.
The average age of first marriage fell to 22 for men; 20 for women. By age 40, over 95 percent of adults
had been married. Following a
war-induced spike in 1946, the divorce rate fell steadily. As noted earlier, the total fertility
rate climbed from 2.2 children per woman in 1935 to 3.8 by 1957 and remained
high through 1964. A Social
Security system focused on “family protection” combined with a
“family-friendly” income tax and housing programs focused on young married
couples to create a remarkably family-centered climate.
is econometric evidence pointing to the role of Social Security in this
change. Writing in the Journal
of Population Economics, the German scholar Berthold Wigger marshalls evidence to
show that moderate-sized public pensions actually have a positive effect on
fertility. Indeed, until 1965, the
American system was modest in size and scope. Many American workers remained outside the system. In 1947, the maximum payroll tax was
only $30 per year, one percent of average income; and in 1965, only $174 per
year, about two percent of income.
As late as 1957, 52 percent of the elderly still reported receiving some
support from their children, compared to only 42 percent receiving some support
from Social Security. In 1960,
two-thirds of widowed women 75 years and older still lived with
relatives. As Wigger's
calculations show: “median-sized public pensions...may stimulate fertility,” a
conclusion reinforced by the American experience between 1940 and 1965.
The True Social Security Crisis
this healthy balance did not last.
One possible explanation is that American political leaders lacked the
self-discipline to maintain a median-sized public pension system. From 1940 until 1965, Congress did keep
increases in pension payments modest, largely in line with the
cost-of-living. Then came the
Great Society, when money seemed limitless and all things possible. And, as already noted, Congress tossed
reason and caution to the wind in 1972, in an open, bipartisan act of pandering
to the elderly vote.
possible explanation is that a combination of ideologies brought the system
down. The new feminism of the
1960's looked with disgust at the “breadwinner/homemaker” model enshrined in
the Social Security system. They
objected to the assumption of traditional gender roles, and particularly to the
homemaker's pension, which rewarded women who gave full-time care to their
children. The neo-Malthusians, or
population control advocates, also roared back with force in the 1960's,
appalled by public policies that favored families with more than one or two
children. And the Grey Panthers
were on the march, too, elderly Americans claiming that they were the victims
of “ageism,” and demanding more money and programs.
possible explanation is that our Social Security system actually contained the
seeds of its own destruction.
America received its first warning of this problem back in 1940. The Cassandra in question was Gunnar
Myrdal, a Social Demo-cratic economist from Sweden. The venue was the Godkin Lectures at Harvard
University. His title: Population:
A Problem for Democracy.
argued that all pay-as-you-go public pension systems rested on a fundamental
contradiction. Before the creation
of such systems, parents depended on their own children as their ultimate
“safety net,” their true insurance.
If other forms of savings and asset accumulation failed, the adult
children would be there to provide financial support, shelter, or care. This created a strong incentive for
public pay-as-you-go system reversed the incentives. Pensions were now provided by the government as a
right. Children were still needed
to make the system work in the future.
However, Social Security represented the socialization of the private
insurance value of children.
Pension benefits were now tied to work and salary; not family size. Childless workers received the same
pensions as workers who had spent much of their income to raise large
families. This created a perverse
incentive: the rational, income-maximizing individual would now have no
children at all. As he or she
would reason: Let others spend money on raising children who will grow up to be
taxed to pay for my retirement.
Economists label this the “free rider” problem. Myrdal called this contradiction
burning core” of the population problem.
Security-Fertility Hypothesis” has been repeatedly confirmed by other,
decidedly non-socialist researchers.
Looking at developing countries, B. Entwisle and C.R. Winegarden report
in Economic Development and Cultural Change that:
[Our] results suggest that increased
pension expenditures... lead to lower fertility levels five or so years
later. This lower fertility in
turn implies increased pension expenditure....”
experience of 81 nations, both developed and developing, Charles Hohm and his
research team report in The Social Science Journal:
...that after controlling for relevant
developmental effects, the level and scope of a country's social security program
is causally and inversely related to fertility levels.
the reverse hypothesis also proves true:
...reduced fertility levels result in
subsequent increases in social security expenditures.
researchers emphasize the “robust” nature of their results. From Australia
to Zambia, higher pension benefits result in fewer babies, while fewer babies
result in higher benefits, as each system more or less consumes itself.
Writing in Economic
Isaac Ehrlich and Francis Lui are more blunt:
any potential welfare implications of social security, our analysis shows that a
proportional tax, under the defined benefits, PAYG system operating in many
countries, will adversely affect fertility, savings, or investment in human
capital [that is, in children]....The specter of financial collapse...is an
inevitable outcome of persistent secular reductions in fertility, labor
productivity, and the aging of the population, given the mechanics of the PAYG
This is, I
suggest, the real Social Security crisis facing America. We can see this process at work in the
United States during the 1960's and 1970's with potential children consumed by
rising expenditures for social insurance directed primarily to the elderly (See
As % of GDP
Note how the
fall of the U.S. Total Fertility Rate by 50 percent occurred in inverse
relationship to an 85 percent increase in the scope of public expenditures for
social insurance, precisely as predicted by Dr. Myrdal. Moreover, this same time period saw the
near disappearance of family-centered solutions to old-age security. As example, the percentage of the
elderly receiving aid from their grown children fell from 52.5 percent in 1957
to only 4 percent by 1980.
On top of
this apparent innate dilemma of social insurance looms another problem with adverse
fertility effects: let us call it the “Generation X Syndrome.” A team of Dutch scholars at Tilburg
University has shown that a pay-as-you-go pension scheme has a positive
fertility effect on the first generation being taxed, possibly due both to the
low tax rates involved and the certainty of a new retirement benefit. In contrast, young adults of the early
21st Century — such as Generation X — already face almost unprecedented overall
taxes and — as surveys repeatedly show — express doubt that they will ever gain
much of a return from Social Security.
In the absence of high taxes, this cynicism toward “the generational
contract” might have led these young adults back to a family-centered solution:
having more children. But in the
real world context of already high taxes, the actual effect is for them to
invest more in private savings, rather than in children. In short, children
— living and
potential — are still seen primarily as already socialized “public” goods. Better to save than create a baby.
All of these
results point to one conclusion: fertility is at least partially a function of
public policy. The crisis of
Social Security that we now face is, in part, the result of perverse incentives
built into the very system. This
also suggests that fertility is a variable that we can effect. If past policy actions drove it down,
perhaps future policy actions might allow fertility to rise again.
Work by the
Congressional Budget Office is helpful here. In its 1999 Report on Uncertainty in Social Security's
the CBO uses a far wider range of possible future fertility rates, varying from
a minimum TFR of 0.5 to a maximum of 3.5 in any given year, and a low of 1.0
and a high of 3.0 over time. The
Another way to view the uncertainty
about fertility is to look at other factors that may have caused fluctuations
over time. The Depression, World
War II, the great postwar economic expansion, the discovery of cheap and
effective birth control — all of those events had unpredictable and dramatic
effects on the fertility rate. By predicting uncertainty that is
consistent with past variation, CBO is implicitly assuming that such events
could happen again.
me as both reasonable, and hopeful.
then, should a pro-family advocate judge rival reform plans for Social
Security? I suggest four tests:
The anti-natalist incentives of a
pay-as-you-go pension system weaken and discourage families and the birth of
children. Does the proposal
reduce or reverse these incentives?
Even in 2005, there
remain some pro-family components of the social security system, including the
homemakers' pension (one of the few remaining public recognitions given to the
work of a parent at home) and survivors' insurance. Does the proposed reform protect or enhance these
The broad effect of social
insurance has been to substitute state programs for natural family bonds.
Does the proposal do anything to
strengthen inter-generational bonds within families? And...
reforms with new debt will only add to the already heavy future tax burden of
the young, and enhance the state-induced negative effects on fertility. Does the proposal avoid massive new
principals, let us consider various reform proposals. One approach is simply to defend the system as it is, tied
to some increase in the retirement age and/or a rise in or extension of the
payroll tax. This alternative
would preserve the homemaker's and survivors' benefits and avoid new debt. However, it would leave in place
existing incentives against the birth of children and would do nothing to
rebind the generations of a family.
The “true” crisis would remain.
the creation of individual accounts?
Here, under one popular scenario, younger workers could divert
two-thirds of their individual payroll tax payment into a mutual-fund-like
account. Alas, while I want to
believe in the magic of these accounts, I face certain difficulties here. Relative to the fertility issue, I
think economist John Mueller is right: “compulsory individual financial
accounts ... would worsen the demographic problem.” He quotes Martin Feldstein to the effect that the whole plan
rests on forcing “reduced consumption” on young adults while also pushing them
toward the accumulation of a “dedicated capital stock.” Translated from
“economese,” this means
less money for young adults to invest in children; and — as a probable result —
fewer children. Since individual
accounts would also be tied to reduced benefits, the homemaker's pension would
shrink, as would survivors' insurance.
On a more positive note, if funds in individual accounts were made fully
transferable to heirs at the insured's time of death, they could serve as a kind
of patrimony and help rebind the generations; however, if not fully
transferable, there would be no real gain here. Finally, most “individual account” plans assume borrowing up
to $2 trillion to finance the transition, a debt which would fall on young
adults and further discourage them from family formation and children.
Writing in The
last November, Mr. Mueller proposes an alternative. Allow young workers to cut their payroll taxes now, in
exchange for a reduced level of future benefits. Relative to fertility, this would let young adults keep more
of what they earn, which could be invested in children. However, this plan does not take into
account the “Generation X Syndrome,” and the probability that the money
retained would — if not immediately consumed — simply be put into another form
of savings, rather than invested in children. Moreover, the Mueller plan would reduce the homemaker's and
survivors' benefits, as well. On
the plus side, it apparently would not require direct new debt; merely the
early spending down of the existing Trust Fund surplus.
provocative idea comes from Phillip Longman, of the New American
Foundation. Writing this January
in The Washington Post, he proposes giving payroll tax relief to married parents
who successfully raise their children:
For example, have one child, and the
payroll tax you pay (and that your employer normally pays) drops by one
third. A second child would be
worth a two-thirds reduction in payroll taxes. Have three or more children and you wouldn't
have any payroll taxes again until your youngest child turned 18.
would still receive full benefits on retirement, provided that all their
children had graduated from high school.
However, non-parents would face a reduction in benefits to pay for the
change, “at least until birthrates rose sufficiently to increase the system's
tax base and avoid rapid population aging.” Well, this approach would probably solve the system's fertility
problem: babies would become a
very valuable tax shelter. It
would clearly preserve, and even enhance, the homemaker's and survivors'
benefits, for those with three children.
And it seems to avoid new debt.
However, its effects on family intergenerational bonds are unclear.
own preference would be for a modified version of this approach. There is still enough of the
“libertarian” in me to want to reduce the overall scope of the welfare state,
so that natural family bonds might recover. Specifically, for “Generation X” and younger, I propose
linking a 20 percent reduction in future OASDI benefits (perhaps achieved by
fixing future benefit increases half-way between the CPI and the wage index) to
a series of credits:
Taxpayers should be granted a credit of
25 percent against their total FICA tax (employer's and employee's portions)
for each child born or adopted, a credit to be continued until the child
reaches age thirteen. This would
mean that a family with four children, ages twelve and under, would pay no FICA
tax in that year (but would still receive all due employment credit).
Second, taxpayers should also be granted a 25 percent
credit against their total FICA tax for each elderly parent or grandparent
residing in the taxpayer's home.
Third, for each child born, a mother should receive three
years (or 12 quarters) of employment credits (calculated at the median fulltime
income) toward her future Social Security pension.
Fourth, a person should also receive one year's employment
credit toward Social Security, at the same median income level, if he or she
served as the primary care giver for an elderly relative residing in his or her
approach would also give a strong incentive toward bearing and rearing
children. It would strengthen the
pension claims of homemakers and parents-at-home, by preserving their net value
and making them more direct and visible.
It would encourage stronger inter-generational bonds within
families. It calls for no new
debt. This plan also builds on
recent econometric findings, namely that:
and child allowances [or in the American system, tax credits] are like
Siamese twins: you should see neither of them or both together, but
never one without the other.
revenue from the credits exceeds benefit reductions, I would favor either
raising the cap on taxable income or widening the definition of income subject
to the payroll tax.
So much for
my private policy fantasy. Is
there any way to rescue the idea of individual accounts, which is the real
innovation on the table this political season? Yes, there is.
The key here, as just hinted at, is to link any benefit reduction to a
substantial new child benefit. As
econometric analysis shows, this is the only way to prevent the change from
aggravating the fertility problem: “downsizing the PAYG-scheme can bring about
an...improvement [only] if it is combined with the introduction of a [new]
child allowance scheme.” In
America, this benefit could be delivered either through a new child credit
against payroll taxes, such as suggested by Mr. Longman or myself, or through a
substantive increase in the existing Child Tax Credit in the income tax: say,
to $2,000 per child; tied to an elimination of existing income limits on
eligibility and extended to dependent children through age 18. This would blunt, even reverse, the
system's incentives against children.
Any reduction in the homemaker benefit would be compensated for by this favorable
tax recognition of dependent children, particularly for larger families. If the credit is against the income
tax, how could we cover the lost revenue?
Perhaps a “family support” surtax of from 1 to 2 percent on incomes over
$150,000 (some of which would come back to the relatively well-off with
children through the expanded child tax credit). Only persons both well-off and childless would pay
significantly more tax under this approach. This group can afford it, and
— as “free riders” relative to
children — they owe it to their nation.
Some in this
audience, I suspect, were put off by my initial description of the Social
Security system of 1939: its commitment to a very traditional family order; its
assumption of rigid gender roles.
Even so, I urge you to consider that it was this vision of family life
that gave moral coherence and popular support to the program of social
insurance. It was something more
than a mere insurance plan; it embodied the values and virtues of the American
nation. And it helped produce a
new flowering of family life in mid-20th Century America.
In this new
century, I believe, any vision of Social Security — including one resting on
individual accounts — must build on a similar moral grounding. The details
need not be the same. However, I do predict that only a commitment to
children, to family living, and to a rebinding of the generations will be up to