The Keynesian model does not evaluate government expenditure
using the standard microeconomic concept of economic efficiency (cost-benefit
analysis). Instead, the model assumes policy should target increased GDP. This
sounds reasonable, and consistent with efficiency considerations, until one
examines government expenditure in detail.
This expenditure has two components: purchases of goods and
services (e.g., roads, education, research, and the military) and transfer
payments (e.g., unemployment insurance, welfare, food stamps, Medicaid,
Medicare, and Social Security).
The efficiency concern with government purchases is that the
National Income and Product Accounts value them as equal to the expenditure on
these items. This means that bridges-to-nowhere or a military buildup aimed at
an imaginary alien invasion are both desirable from the Keynesian perspective
because such expenditures increase measured GDP. Yet this expenditure is pure
waste.
More broadly, the fact that some government expenditure generates
benefits in excess of costs (the economy needs some roads) does not mean
additional expenditure generates value in excess of costs (the economy does not
need to re-pave its roads every year). At some point additional expenditures hit
diminishing returns and make no sense in cost-benefit terms.
Transfer payments are also problematic from an efficiency
perspective because they distort economic incentives. Unemployment insurance
discourages work effort. Social Security subsidizes early retirement for people
who are still able-bodied. Medicare and Medicaid create moral hazard, thereby
generating excess health costs. Thus even if transfers help stimulate consumer
spending, their net effect on the economy is unclear.
This implies that whether the sequester will harm or help the
economy depends on whether cost-benefit considerations can justify the existing
level of government expenditure. And on this question, the answer is clear.
Across all categories, federal expenditure is far greater than necessary to
achieve the legitimate goals of government intervention.
National defense is a classic public good — something that
everyone values but that free markets are unlikely to provide — so some
expenditure on national defense makes sense. Yet many national security
activities have large and certain costs but hard-to-measure and intangible
benefits, if any. Examples include the ongoing occupation of Afghanistan, the
provision of national security for Western Europe and other parts of the globe,
not to mention misguided weapons systems, redundant military bases, and
more.
Non-defense, non-entitlement spending is also rife with programs
that have no compelling justification and frequently do substantial harm: the
National Endowments for the Arts and Humanities, the Corporation for Public
Broadcasting, NASA, earmarks, the Post Office, Amtrak, foreign aid, agricultural
subsidies, the Small Business Administration, drug prohibition, and more. Even
for expenditure that is defensible in moderation (e.g., transportation and
education), many specific projects are wasteful (the Big Dig, high-speed rail,
No Child Left Behind).
Entitlement spending is also excessive relative to reasonable
cost-benefit considerations. A minimal government safety net that provides
insurance against the worst-case scenarios for unlucky individuals is one thing;
the current generosity of Medicare, Medicaid, and Social Security, which by
themselves threaten to bust the entire budget in coming decades, is another. The
introduction of higher deductibles, increased eligibility ages, and less
generous indexing are all desirable from a cost-benefit perspective and
essential if the United States is to avoid a fiscal meltdown.
While the recommendations espoused here may sound heretical,
considerable evidence suggests that policymakers should set government spending
based on hard-nosed cost-benefit considerations, not Keynesian stimulus
concerns.