![]() |
|
![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() ![]() |
|
Who Pays and When? An Assessment of Generational Accounting November 1995 |
![]() |
In analyzing these findings, the Congressional Budget
Office used a computer program and data provided by the authors of the
accounts. CBO cannot vouch for the data or program. There is no reason
to believe that they contain errors, although errors have appeared in past
versions (a common occurrence when developing a complex system).
Assessing the Evolution and Status of Generational Policy
According to the accounts, lifetime net tax rates of succeeding generations have risen steadily over this century. The rates rose from 24 percent for people born in 1900 to 37 percent for those born in 1990 (see Table 1). Those figures combine net taxes at all levels of government--federal, state, and local (see Box 4 ).
The lifetime net tax rate of future generations would have to rise to nearly 80 percent to settle the bill. That represents about twice the lifetime net tax rate of current newborns, or a difference in lifetime net tax rates of future generations and current newborns of more than 40 percentage points.
The bill for the future would be even higher if not for the Omnibus Budget Reconciliation Act of 1993 (OBRA-93). The version of the accounts that this study used predates OBRA-93. But an updated version that includes the provisions of OBRA-93 also includes new economic and technical assumptions that offset the effects of those provisions. As a result, using the updated version would not appreciably change any numbers reported here (see Box 5 on page 22).
Rapidly rising medical costs and aging of the population account for most of the difference in lifetime net tax rates of future generations and current newborns. Rising costs of medical services per recipient account for most of the rise in health spending; a rising number of Medicare recipients accounts for a much smaller part. The aging of the population means that baby boomers are scheduled to receive more in Social Security benefits when they retire than prevailing tax rates will provide.
Why Do Lifetime Net Tax Rates Seem So High?
The lifetime net tax rates shown in Table 1 may seem high when compared with the more familiar "current net tax rates" (current net taxes as a percentage of current market income). For example, the lifetime net tax rate of current newborns is 37 percent, whereas net taxes of the nation now stand at only 24 percent of national income.
But current net tax rates do not directly compare all
generations on the same basis because such rates vary widely with age.
The young and middle-aged typically earn most of market income and pay
most of total taxes. And the old typically receive more in transfers than
they pay in taxes.
Given that pattern, three factors explain why estimated
lifetime net tax rates are as high as they are. First, present value gives
more weight to taxes paid earlier than to transfers received later. Second,
people do not all live to old age. Therefore, the accounts give more weight
to prospective taxes than to prospective transfers because people are more
likely to pay the taxes than receive the transfers. Finally, lifetime net
tax rates are based on labor income rather than total income, which current
net tax rates are based on. Using the smaller measure as a base leads to
a higher lifetime net tax rate. For comparison, current net taxes are 30
percent of labor income, whereas they are only 24 percent of total income.
Table 1.
Estimated Lifetime Tax and Transfer Rates by Year of Birth (Average for males and females, in percent) |
|||
|
|
|
|
1900 |
|
|
|
1910 |
|
|
|
1920 |
|
|
|
1930 |
|
|
|
1940 |
|
|
|
1950 |
|
|
|
1960 |
|
|
|
1970 |
|
|
|
1980 |
|
|
|
1990 |
|
|
|
Future Generations |
|
|
|
SOURCE: Congressional Budget Office, using a computer program and data provided by the authors as described in Alan J. Auerbach, Jagadeesh Gokhale, and Laurence J. Kotlikoff, "Generational Accounts: A Meaningful Alternative to Deficit Accounting," in David Bradford, ed., Tax Policy and the Economy, vol. 5 (Cambridge, Mass.: MIT Press, 1991), pp. 55-110. | |||
NOTES:The
rates shown are for taxes and transfers at all levels of government combined--federal,
state, and local.
The estimates assume a real discount rate of 6 percent, a prospective annual rate of growth in productivity of 0.75 percent, and the midgrowth path of population used by the Social Security Administration in its 1993 annual report. The values in the table reflect the implication of generational accounts as constructed, not necessarily the views of the Congressional Budget Office. |
|||
a. A lifetime net tax
rate is the present value at birth of lifetime net taxes as a percentage
of the present value at birth of lifetime labor income. The net tax rate
equals the gross tax rate less the transfer rate (except for possible differences
because of rounding).
b. There are no unique values of gross tax and transfer rates for future generations. For the base case shown, any combination of lifetime tax and transfer rates that nets to 78 percent is feasible, at least arithmetically. |
What Do the Lifetime
Net Tax Rates of
Various Generations Imply?
Rising lifetime net tax rates did not necessarily make
successive generations worse off than their predecessors. First, most of
the increase has paid for a similar rise in the rate of government purchases
during the period from 1900 to about 1950. But the accounts do not assign
the benefits of purchases to specific generations. Second, estimated lifetime
incomes rose significantly during this century (aside from any benefits
from government purchases). The accounts estimate that the lifetime income,
after taxes and inflation, of the average person born in 1990 is about
three times that of the average person born in 1900.
Prevailing policy, however, implies that the lifetime
after-tax income of at least some future generations would fall below that
of current newborns (given the base assumptions of the accounts). That
would happen if, as the accounts assume, all future generations
pay lifetime net taxes at twice the rate of current newborns. But it would
also happen no matter how future generations might settle the bill.
|
|
It would be useful
to present generational accounts for the federal sector separately because
combined accounts for all levels of government mask the source of generational
policy. (The combined accounts, however, can isolate the effects
of changes in, rather than levels of, federal activity.)
For example, suppose that projected purchases by local governments reflect the need for relatively less spending on education. Thus, if local governments maintain budgets in approximate balance, they could reduce the rate of property taxes, which largely finance education. Even if local authorities kept the same property tax rate, as the accounts assume, it would not correct fiscal imbalance at the federal level. Instead, it would maintain the amount of property taxes deducted from the federal tax base. Combining the accounts also mixes apples and oranges. Typical federal activities usually have different implications for given age groups than typical nonfederal activities. For example, about half of federal excise taxes come from gasoline, tobacco, and alcohol. Taxes on those items tend to apply more narrowly to specific age groups than the more broadly based excise taxes of states and localities. And most of the purchases that state and local governments make are related to the age of the people, such as those for education, whereas most of the federal government's are unrelated to age, such as those for defense. Finally, particular taxes of state and local governments are more often tied to particular purchases--for example, property taxes and purchases to provide municipal services. Separating the accounts, therefore, would alleviate some problems of interpretation. |
Eliminating the Difference in Lifetime Net Tax Rates
of Future Generations and Current Newborns
What policy would equalize the lifetime net tax rates
of current newborns and future generations? An infinite number of policies
could do so. One way to pose the problem in an easily handled form is to
ask the hypothetical question: how much would a tax have to be raised or
a transfer reduced now if its total grew thereafter at its previously projected
rate?
Such an abstract approach does not consider realistic
options; it merely defines the size of the problem for further analysis
and debate. The approach does not address many issues that the Congress
would have to consider: the short- and long-term effect on the economy,
distribution by income, and so forth. In fact, the Congress may decide
to reduce lifetime net tax rates of future generations, but not to the
same rate as that of current newborns.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||
The results reported
in this study would not change significantly if the Congressional Budget
Office had used an updated version of generational accounts.
The version used reflects the 10-year economic and budget projections of
the Office of Management and Budget's (OMB's) 1992 midyear update, which
do not include the provisions of the Omnibus Budget Reconciliation Act
of 1993 (OBRA-93). The latest official version reflects the 10-year
projections of OMB's 1993 midyear update, which include the provisions
of OBRA-93, plus a new method of projecting the Medicaid expenditures of
state and local governments, and updated relative-age profiles for taxes
and transfers.
The results of those changes nearly cancel each other.
Without OBRA-93, the changes would have raised the lifetime net tax rates
of future generations from 78 percent to 94 percent, while barely changing
those of current generations (see table below). About half of this
difference arises because the new version of the accounts assumes that
Medicaid spending by state and local governments will grow through 2004
at the same rate that the Health Care Financing Administration projects
total Medicaid expenditures to grow rather than at the same rate as gross
domestic product. The rest of the difference comes from updated economic
projections and relative-age profiles for taxes and transfers--
Alternatively, OMB estimated that OBRA-93 would directly reduce the deficit from 1994 through 1998 by $429 billion by means of a combination of higher taxes, lower spending for mandatory programs, and new caps on discretionary spending from 1996 through 1998. Given the 1993 economic projections, those provisions lowered the lifetime net tax rates of future generations to 82 percent, but again barely changed those of current generations. The net result of the two effects is essentially a wash, and there is no reason to believe that using the latest version would significantly change any results reported in this study, either qualitatively or quantitatively. The budget resolution of 1995 would substantially change the results that are shown in the table. But the decisions necessary to carry out the resolution have not been made. Therefore, it would be premature to speculate about how the generational accounts for people of different ages would change. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
The budget item that is changed determines both the size
of the change that is needed and the way its costs are spread by age (see
Table 2). The size of the hypothetical deficit reduction could vary from
$109 billion for cuts in Social Security benefits to $227 billion for cuts
in government purchases.
Table 2.
Distribution of Costs of Hypothetical Policy Changes Needed in 1991 to Reach a Sustainable Policy |
|||||||
Alternative
Proportionate Tax Increases |
Alternative
Proportionate Spending Cuts |
||||||
|
|
|
|
|
|
|
|
|
|||||||
90 |
|
|
|
|
|
|
|
80 |
|
|
|
|
|
|
|
70 |
|
|
|
|
|
|
|
60 |
|
|
|
|
|
|
|
50 |
|
|
|
|
|
|
|
40 |
|
|
|
|
|
|
|
30 |
|
|
|
|
|
|
|
20 |
|
|
|
|
|
|
|
10 |
|
|
|
|
|
|
|
Newborns |
|
|
|
|
|
|
|
Future Generationsd |
|
|
|
|
|
|
|
Corresponding Budget Change in 1991 (In billions of 1991 dollars) | |||||||
n.a. |
|
|
|
|
|
|
|
SOURCE: Congressional Budget Office, using a computer program and data provided by the authors as described in Alan J. Auerbach, Jagadeesh Gokhale, and Laurence J. Kotlikoff, "Generational Accounts: A Meaningful Alternative to Deficit Accounting," in David Bradford, ed., Tax Policy and the Economy, vol. 5 (Cambridge, Mass.: MIT Press, 1991), pp. 55-110. | |||||||
NOTES:
The estimates assume a real discount rate of 6 percent, a prospective annual
rate of growth in productivity of 0.75 percent, and the midgrowth path
of population used by the Social Security Administration in its 1993 annual
report.
The values in the table reflect the implication of generational accounts as constructed, not necessarily the views of the Congressional Budget Office. A feasible policy is sustainable if lifetime net tax rates of future generations and current newborns are equal. n.a. = not applicable. |
|||||||
a. There are no entries for a cut in other transfers because the required cut would be about twice their total. | |||||||
b. A proportionate cut in prospective government purchases affects the net taxes of future, but not current, generations. Both receive fewer services from government purchases. Current generations continue to pay as much as if there had been no cut, so future generations can pay less than they would otherwise. | |||||||
c. Average for males and females. | |||||||
d. The figures for future generations apply to those born next year. The respective figures for successive future generations would grow at the rate of productivity. |
Two main factors account for the wide variation in the
size of the hypothetical deficit cut that is needed: the rate at which
the budget item grows and the number of people that the deficit cut affects.
How Fast Does the Budget Item
Grow?
A smaller deficit cut is needed when the tax or spending
is projected to grow faster. To see why, consider what happens when a shrinking
item in the budget gets cut by the same proportion every year. Suppose
spending is going to be $100 this year and $90 next year. It would take
a proportionate cut of 10 percent--$10 this year and $9 next year--to save
a total of $19. By contrast, consider a growing program. Suppose spending
is going to be $100 this year and $110 next year. Then a proportionate
cut of only 9 percent would be necessary--$9 this year and $10 next year--to
save a total of $19. Cutting a growing spending program gives a bigger
bang for the buck. The faster the item grows, the smaller the proportionate
change needed to save a given total.
The effect is most pronounced in cuts to health benefits and Social Security. Those programs are projected to grow much faster than other federal spending. Medical costs per recipient are assumed to grow faster than output per capita through 2030 in the version of the accounts that CBO used. Hence, a relatively small proportionate cut in health spending is required to reach a sustainable path. Similarly, retirees are expected to grow as a share of the population; thus, a relatively small proportionate cut in Social Security benefits is necessary to reach a sustainable path. By contrast, a big cut in purchases is necessary to achieve sustainability, mostly because defense purchases are assumed to grow slowly.
How Many People Does the
Deficit Cut Affect?
The size of the deficit cut that is needed also depends
on the number of people now alive who will pay for the cut with higher
net taxes. For instance, the old contribute little to taxes on labor income,
but a lot to taxes on capital income. As a result, taxes on labor income
would have to rise more than taxes on capital income to achieve sustainability.
Similarly, everyone now alive would contribute at some time if benefits
for Social Security or Medicare fell. Thus, those actions need a relatively
small proportionate deficit cut.
Excise taxes may have to rise by so much, according to
the accounts, simply because the accounts assign excise taxes to children.
But a newborn next year belongs to a future generation and will not contribute
to net taxes of current generations. By contrast, the average 15-year-old
does not contribute to payroll taxes this year, but will next year. Thus,
there are more current generations in the pipeline to pay tax at higher
rates on payroll than on sales.
To some extent, the size of the deficit cut that is needed
also depends on how it affects the lifetime net tax rate of current newborns.
The more the cut raises their lifetime net tax rate, the less that of future
generations has to fall to become equal. That effect is most pronounced
in the case of a cut in government purchases because such a cut does not
raise the net taxes of current newborns at all.
Assessing Past or Prospective Fiscal Policy
Both retrospective and prospective analyses show that
fiscal policy can head in a different direction than the deficit seems
to indicate. Looking back, most analysts agree that policies since World
War II shifted resources from the young to the old. The accounts, however,
put most of the blame on the policies of the 1950s, 1960s, and 1970s rather
than the high-deficit years of the 1980s. Looking ahead, the accounts show
that the way that fiscal policy shifts resources among generations can
be unrelated to the deficit.
Assessing Fiscal Policy
Since World War II
Generational accounts suggest that the direction of postwar
fiscal policy was quite different from what is commonly supposed. According
to an early version of the accounts, policy through the 1970s made the
old better off at the expense of then-young-and-future generations, although
deficits were low and debt was falling in relation to output.(2)
Perhaps more surprising, policy in the 1980s had little effect on the net
taxes of future generations, although deficits were high and debt was rising
in relation to output.
Two main factors account for the gains of the elderly
from the 1950s through the 1970s. First, benefits for Social Security and
Medicare were increased in each of those decades. Second, payroll taxes
rose to help pay for the higher benefits, while taxes on capital income
fell as a share of all taxes. The benefit increases helped the old more
than the young; the switch in tax bases helped the old and hurt the young;
and the mounting bill hurt future generations.
That pattern changed in the 1980s. The Social Security
Amendments of 1983 raised payroll taxes and reduced prospective benefits.
The law phased in an increase in the earliest age (from 65 to 67) at which
retirees could draw full benefits. The law also set a cap on the benefits
that people could receive in any year before they were subject to income
tax. Because the cap is set in nominal terms, inflation will expose a growing
portion of benefits to tax.
The amendments had different effects on different generations.
Net taxes changed little for those who were retired or about to retire,
but increased for the young and middle-aged. Moreover, the amendments lowered
the net taxes required of future generations by raising those of current
generations. According to the accounts, that change approximately offset
the effects of higher deficits in the 1980s on the net taxes of future
generations.
Those results, however, do not establish that policymakers
would have made different choices if they had used generational accounts
at the time. First, the accounts assign only the cash cost of net taxes,
not the benefits of government purchases. If the accounts could assign
those benefits, the pattern of distribution might look different. Purchases,
especially for defense, were raised in the 1950s. The benefits to the young
of more national security could have been enough to offset the prospect
of higher payroll taxes.
Second, the exercise used 20/20 hindsight about the growth
of output and population. But analysts in the 1950s and 1960s did not have
20/20 foresight and made inaccurate economic and demographic projections.
Analysts at the time assumed that medical costs would rise more slowly
than they actually did, that the death rate would fall more slowly, and
that productivity would grow more rapidly. Those errors would also have
appeared in generational accounts if they had been prepared then.
In a sense, from the 1950s through the 1970s, the nation
won one gamble but lost another. During that period, the interest rate
on government debt was lower than the growth rate of the economy. That
made it possible to win a gamble that the ratio of debt to output would
fall without having to raise taxes to pay interest on the debt. But the
country lost the gamble that such conditions would persist; government
undertook implicit obligations for Social Security and Medicare that it
cannot meet as they are currently scheduled.
Assessing Prospective Policy Changes
As an experiment, CBO used the accounts to analyze some
hypothetical policy changes. In some ways the results are what the deficit
would indicate: for example, a deficit increase would worsen the lot of
future generations. The accounts, however, also show that changes in policy
can shift resources among generations without changing the deficit at all.
Similarly, different ways of changing the deficit by a given amount can
have very different effects on any given generation.
A Policy That Increases the Deficit. The Congressional
Budget Office considered a policy that would cut the income tax by 20 percent
for five years without changing spending. That would raise the bill left
for the future. The hypothetical policy would then raise the income tax
by enough to make the bill grow at the same rate that it would have if
the tax had not been cut.
Such a policy would benefit the middle-aged and old at
the expense of young and future generations (see Table 3). Those about
50 years old would gain the most. Those older than 50 would not gain as
much because their labor income has already peaked.
And those younger than 50 would not gain as much because,
after five years, they would have to pay higher net taxes for the rest
of their lives. In fact,those younger than 30 would pay more under the
policy. They would gain little or nothing from the tax cut, but would pay
higher net taxes for most of their lives. Furthermore, future generations
would lose because the higher bill would require higher net taxes to pay
the extra interest. The policy would raise the present value of net taxes
of current newborns and future generations by about the same amount.
Three Deficit-Neutral Policies. CBO
also considered three policies that would change the mix of taxes and transfers
without changing the deficit. The policies qualitatively resemble changes
in mix that have occurred slowly since World War II.
Table 3.
Alternative Policies That Would Change the Timing or Mix of Taxes and Transfers (Change in present value of net taxes, in thousands of 1991 dollars) |
||||
|
|
|||
Age in 1991 |
|
|
|
|
90 |
|
|
|
|
80 |
|
|
|
|
70 |
|
|
|
|
60 |
|
|
|
|
50 |
|
|
|
|
40 |
|
|
|
|
30 |
|
|
|
|
20 |
|
|
|
|
10 |
|
|
|
|
Newborn |
|
|
|
|
Future Generationsa |
|
|
|
|
SOURCE: Congressional
Budget Office, using a computer program and data provided by the authors
as described in Alan J. Auerbach, Jagadeesh Gokhale, and Laurence
J. Kotlikoff, "Generational Accounts: A Meaningful Alternative to
Deficit Accounting," in David Bradford, ed., Tax Policy and the Economy,
vol. 5 (Cambridge, Mass.: MIT Press, 1991), pp. 55-110.
|
||||
NOTES: The estimates
assume a real discount rate of 6 percent, a prospective annual rate of
growth in productivity of 0.75 percent, and the midgrowth path of population
used by the Social Security Administration in its 1993 annual report.
The values in the tables reflect the implication of generational accounts as constructed, not necessarily the views of the Congressional Budget Office. |
||||
a. The figures for future generations apply to those born next year. The respective figures for successive future generations would grow at the rate of productivity. |
The first policy would increase Social Security benefits by 20 percent and raise the payroll tax to payfor them. This policy would typically cause greater changes on the basis of age than would the temporary tax cut, even though the policy would not change the deficit. For example, the accounts estimate that the present value of net taxes of 60-year-olds would fall by about six times as much as under the temporary tax cut.
The lifetime net taxes of future generations would rise
by about three times as much as they would under the temporary tax cut.
The extra taxes needed to make extra payments would pass from generation
to generation. Those who are now young would pay their extra taxes to those
now old, leaving the deficit unaffected. But when those who are now young
became old, future generations--those not yet born--would have to pay the
extra taxes.
The second policy would raise payroll taxes by 30 percent
to pay for a cut in excise (or consumption) taxes.
That policy would cost 20- to 40-year-olds the most. Unlike the other policies,
it would reduce the net taxes of the very young because the accounts assign
excise taxes according to share of family consumption. The policy would
also reduce the net taxes of future generations; there are more middle-aged
people to pay higher payroll taxes than old people to pay lower excise
taxes.
The third policy would raise payroll taxes by 30 percent
to pay for a cut in income taxes. The switch in taxes would benefit the
old because more of their prospective income comes from capital than from
labor. But the policy would make smaller changes by age than would the
other deficit-neutral policies because about 80 percent of national income
comes from labor. The source is the same as the payroll tax, so the switch
in taxes would mainly shuffle the legal base of the tax. That switch would
raise net taxes of future generations because it would reduce net taxes
of the old and middle-aged more than it would raise those of the young.
Table 4.
Alternative Policies That Would Cut the Deficit by an Equal Amount (Change in present value of net taxes, in thousands of 1991 dollars) |
||||
Age in 1991 |
|
|
|
|
90 |
|
|
|
|
80 |
|
|
|
|
70 |
|
|
|
|
60 |
|
|
|
|
50 |
|
|
|
|
40 |
|
|
|
|
30 |
|
|
|
|
20 |
|
|
|
|
10 |
|
|
|
|
Newborn |
|
|
|
|
Future Generationsa |
|
|
|
|
SOURCE: Congressional Budget Office, using a computer program and data provided by the authors as described in Alan J. Auerbach, Jagadeesh Gokhale, and Laurence J. Kotlikoff, "Generational Accounts: A Meaningful Alternative to Deficit Accounting," in David Bradford, ed., Tax Policy and the Economy, vol. 5 (Cambridge, Mass.: MIT Press, 1991), pp. 55-110. | ||||
NOTES:
The estimates assume a real discount rate of 6 percent, a prospective annual
rate of growth in productivity of 0.75 percent, and the midgrowth path
of population used by the Social Security Administration in its 1993 annual
report.
The values in the tables reflect the implication of generational accounts as constructed, not necessarily the views of the Congressional Budget Office. |
||||
a. The figures for future generations apply to those born next year. The respective figures for successive future generations would grow at the rate of productivity. |
Policies That Reduce the Deficit. Finally, CBO also examined four policies that would cut the deficit by a given amount (about $29 billion in the first year). The first policy would permanently cut Social Security benefits by 10 percent; the others would raise capital income, payroll, or excise taxes to lower the deficit by the same amount as the cut in benefits.
Those policies would have different effects on people
who are now alive, even though they would have the same effect on the deficit
(see Table 4). The cuts in benefits would cost those who are about 60 or
70 years old the most because they are at or near the end of their working
lives and have most of their retirement ahead. (A phased-in reduction could
spread the costs more evenly.) The tax increases, however, would cost those
who are about 20 or 30 years old the most. They are still far from retirement
and will pay taxes for many years.
The policies would have slightly different effects on
future generations even though they would cut the deficit by the same amount.
That result occurs because the policies would change the timing of taxes
and transfers that pass from generation to generation. For example, cuts
in Social Security benefits and increases in capital income taxes fall
more heavily on the old, so those policies would eventually make all living
generations pay more toward the bill than would raising payroll or excise
taxes.
Such exercises show how little information the deficit
gives about the effects of policy on different generations. Generational
accounts keep track of such effects, some of which may not be apparent
without the tools that the accounts provide. The next question is, how
accurate and reliable are the accounts?