stock. In so doing, however, we would also build up a liability to foreigners that we would have
to finance in the future; that is, less of our future GDP would be available for consumption by
U.S. residents.
413
THE AGE OF TURBULENCE
economists' most accurate forecasts, and because Social Security is a defined-
benefit program, so payments per beneficiary are reasonably predictable
in advance.
Medicare, by contrast, poses a much bigger problem. The trustees forecast
a seventy-five-year funding shortfall for Part A of Medicare, the Hospital
Insurance Fund, which could be closed with an immediate increase of
3.5 percentage points on top of the current 2.9 percent levy on taxable
payroll (to a total of 6.4 percent), or an immediate halving of benefits or
some combination of the two. But that's not the end of it. The costs of
Part B of Medicare, which pays doctors' bills and other outpatient expenses,
and the new Part D, which pays for access to prescription drug coverage,
are both projected to rise rapidly. But they are mandated to be met by general
tax revenues, rather than by payroll taxes. Although not as visible
as hospital insurance taxes, the claims on general tax revenues of future
Parts B and D benefits are of the same order of magnitude as those of
Part A.
The public trustees, coming at future Medicare costs from a different
perspective, commented in 2006 that "if the Trustees' projections prove a
reliable guide to the next few decades, absent an increase in earmarked
sources of revenue for the program, in just 15 years payment of currently
scheduled Medicare benefits would require General Fund transfers equal
to 25 percent of Federal income tax revenues ... —more than triple their
2005 fiscal burden—and less than 10 years later the General Fund transfer
would equal nearly 40 percent of Federal income tax revenues." But even
such numbers do not necessarily capture the full dimension of the problem,
because projections of Medicare benefits are highly uncertain.
Health spending has been growing faster than the economy for many
years, a growth fueled, in large part, by advances in technology. We know
very little about how rapidly medical technology will evolve or how those
innovations will translate into future spending. Technological innovations
can greatly improve the quality of medical care and can, in some instances,
reduce the costs of treatment and surely the costs of hospital administration.
But because technology expands the set of treatment possibilities, it also
has the potential to add to overall spending—in some cases by a great deal.
As encrypted private health records become sufficiently widespread,
414
THE WORLD RETIRES. BUT CAN IT AFFORD TO?
researchers will for the first time be able effectively to evaluate treatments
and outcomes for a broad spectrum of diseases. I assume a standard for national
best practices will emerge as a result. Eventually public ratings of
hospitals and physicians will come too; market competition can be expected
to follow. Yet none of this is likely to happen very fast. Medical
practice hinges on what traditionally has been a very private relationship
between physician and patient, and both are reluctant to risk a breach in
privacy.
Medical practice in the United States has evolved quite differently
from region to region. I presume that nationwide best practices, by eliminating
the worst treatments and practitioners, will enhance average outcomes.
But it is by no means clear whether a nationwide awareness of and
demand for best practices would increase or decrease medical expenditures.
We need to keep in mind that the uncertainties—especially our inability
to identify that upper bound of future demands for medical
care—warrant significant restraint in policymaking. The critical reason to
proceed cautiously is that new programs quickly develop constituencies
willing to resist any curtailment fiercely. As a consequence, our ability to
rein in deficit-expanding initiatives, should they later prove to have been
excessive or misguided, is quite limited.
Policymakers should err on the side of prudence when considering new
budget entitlement initiatives. Programs can always be expanded in the future
should the resources for them become available, but they cannot easily
be cut back if resources later fall short of commitments. This is why I believe
that moving forward with an unfunded prescription drug program in
2003, before the problems of the severely underfunded and out-of-balance
Medicare program as a whole were addressed, was a mistake, perhaps a
very large one.*
Having participated in a number of studies that simulated future Medicare
costs and benefits, I am struck by the breadth of the range of possible
outcomes for, say, the year 2030. As I noted, the range of possibilities for
*Fortunately (and unusually), the cost of Medicare Part D, the prescription drug program, has
to date come in below initial projections. Possibly the program has fostered the competition it
was structured to create. However, it is still an unfunded, and large, growing expense.
415
THE AGE OF TURBULENCE
Social Security is quite narrow. Demographers, of course, have the same
good handle on the number of future Medicare beneficiaries. But average
cost per beneficiary under current law depends not only on future technologies
but also on patient choices and a whole series of other variables.
The forecasting complexity is such that the trustees have had to fall
back on a simple algorithm—a projected rate of growth of benefits per
Medicare recipient relative to the projected rate of growth of per capita
GDP. Growth in real Medicare outlays per beneficiary has averaged approximately
4 percent a year over the past decade, about 2 percentage
points higher than growth in real per capita GDP. The public trustees go on
to note, "It seems reasonable to assume (per capita) health care and Medicare
expenditure growth will gradually slow to the rate of growth of
GDP—because there is presumably some upper limit to what share of
their growing incomes Americans will want to devote to health care."
That may in fact turn out to be true.
But, as anyone who knows the ways of Washington must realize, such
an assumption also involves quite a spectacular leap. It implies a degree of
restraint that is not written into current law. It assumes a fiscal victory in a
political battle that has yet to be fought. As the public trustees themselves
observe, "No such slowdown has materialized over the past half-century."
Thus, the projected tax burden of mounting Medicare spending presumably
would be even larger if based strictly on current entitlement law and
current trends in health care spending.
There is a great deal of work to be done to set Medicare right. It should
be apparent that to cover future Social Security and Medicare funding
shortfalls wholly by raising taxes is economically infeasible. Doing so would
imply unprecedented peacetime tax rates. At some point, tax rate increases
become self-defeating: by absorbing purchasing power and reducing work
and investment incentives, they reduce the economy's growth rate. Hence,
the growth of the tax base slows and the projected additional tax revenues
fail to materialize fully.
We are left with a most daunting reality: resolving the funding shortfall
for federal social insurance is going to require benefit cuts. Government has
a moral obligation to make these cuts sooner rather than later, to afford fu
416
THE WORLD RETIRES. BUT CAN IT AFFORD TO?
ture retirees as much time as possible to adjust their plans for work, saving,
and retirement spending. Failure to give Americans adequate warning that
the retirement income upon which they have planned will be reduced
could threaten major disruptions in people's lives.
Once the level of benefits that Social Security and Medicare can reasonably
promise is determined, how can public policy ensure that the real resources
will become available to fulfill those promises? Focusing on financial
solvency within the Social Security and Medicare systems without regard
to the broader macroeconomic picture does not ensure that the resources
will be there. Without additional net saving, the real resources required to
produce future benefits will not be produced. Thus, in addressing the imbalances
of Social Security and Medicare, we need to ensure that measures
taken now to finance future benefit commitments represent real additions
to national saving and the productive assets they fund.
We need, in effect, to make real the phantom "lockboxes" of a few
years back, which were supposed to contain the cash to fund future Social
Security benefits. The lockboxes were so real to many Americans that former
Speaker of the House Tom Foley tells the story of his mother's berating
him for trying to disabuse her of that belief. "Mr. Foley," she said, "I hope
you will not be offended at how surprised and shocked I am to find that the
majority leader of the House of Representatives knows nothing about
Social Security." At that time, the lockbox proposals would have required
not only putting the social insurance trust funds (currently in surplus) off
budget for accounting purposes, but also having Congress mandate that the
remainder of the budget be balanced. For a brief period in the flush fiscal-
surplus years at the turn of the millennium, a bipartisan commitment
emerged to do just that. But, regrettably, the commitment collapsed when
it became apparent that, in light of a less favorable economic environment,
maintaining balance in the budget excluding Social Security and Medicare
would require significantly lower spending or higher taxes.
Failure to address the imbalances between promises to future retirees
and the economy's ability to meet those promises could have severe consequences
for individual retirees and the economy as a whole. In the end,
I expect the Medicare funding imbalance to be resolved by rescinding
417
THE AGE OF TURBULENCE
the benefits of the more affluent.* The frenzy of politics and the so-farintractable
continued increase in income inequality, in my judgment, leaves
no other credible political alternative. Restored balance could occur through
the development of private accounts (which I support) or through legislation
requiring Medicare to be means-tested (as is Medicaid). Rationing is
the only other realistic possibility, and that has little support in the United
States. Most future Medicare benefits will surely be concentrated in the
middle- and lower-income groups. Medical service for upper-income recipients
will have to be funded by unsubsidized private medical insurance or
out of pocket, probably in the form of copayments approaching 100 percent.
Many will recoil from the concept of Medicare as welfare, as means-
tested programs tend to be seen, but the arithmetic of twenty-first-century
demographics in a highly competitive global economy necessitates it.
While I favor a liberal immigration policy, I do not do so as a means of
increasing the working population in order to raise social insurance taxes to
help address the Social Security/Medicare funding shortfall. Nor, for reasons
I will discuss later, can we count on a fortuitous increase in productivity;
the long-term ceiling for increases in output per hour in the United States
appears to be 3 percent a year, with 2 percent being the most likely outcome.
In brief, we likely won't have enough people working, nor will we
likely have a sufficient increase in the amount each worker on average can
produce, to cover the enormous shortfall from entitlements under current
law. It may not even be close.
With so many unknowns, I fear that given our demographics and the
limited upside potential of productivity growth, we may already have committed
to a higher level of real medical resources for baby-boomer retirees
than our government can realistically deliver. As previously noted, Congress
can enact an entitlement, but that in itself does not produce the economic
resources required to provide the hospitals, physicians, nurses, and
*How much a cut in benefits would reduce outlays on medical services is uncertain. Several
years ago, I requested the Federal Reserve Board staff to simulate the level of medical service
outlays through 2004, assuming that Medicare and Medicaid entitlements had never been enacted.
The staff concluded that outlays would have been only modestly lower. Market efficiencies,
however, could have been quite considerable.
418
THE WORLD RETIRES. BUT CAN IT AFFORD TO?
pharmaceutical companies that will be essential in 2030 to meet the letter
of current law. The size in 2030 of the transfer of real resources from
worker-producers to retirees may be too large for the former to accept. The
claims on the nation's output, because of an unfunded expansion of entitlements,
may far exceed the output produced by a workforce only marginally
larger than exists today. In short, the promises may have to be broken, or,
perhaps better said, they may have to be "clarified."
The significant uncertainties about the availability of future real resources
are reflected in uncertainties in retiree income replacement rates.
Given today's expected yawning gap between retirement needs and even
current large entitlement promises, private pension and insurance benefits
are going to have to play an increasingly greater role. At the end of 2006,
private pension funds in the United States had $5.6 trillion in assets: $2.3
trillion in the traditional defined-benefit programs and $3.3 trillion in
defined-contribution plans, largely 401(k)s.* Private pension and profit-
sharing funds paid out $344 billion in benefits in 2005.+ By comparison,
Social Security and Medicare paid out $845 billion. The former is bound to
