Created: 1999-07-10
Last Modified: 1999-07-10
Go to
Brad DeLong's Home Page

Teaching | Writing | Career | Politics | Book Reviews | Information Economy | Economists | Multimedia | Students | Fine Print | Other | My Jobs

Feedback is always very welcome...

Notes on the Mid-1999 Budget Outlook


July 10, 1999

J. Bradford DeLong

(510) 643-4027, (925) 283-2709
(510) 642-6615, (925) 283-9933 faxes

Official Forecasts

The most recent budget projections from both Dan Crippen's Congressional Budget Office [hereafter CBO] and Jack Lew's Office of Management and Budget [hereafter OMB] show substantial federal budget surpluses: rising from the $120 billion unified budget surplus in the about-to-be-completed fiscal 1999 to a $413 billion unified budget surplus in fiscal 2009, a decade from now. The cumulative projected budget surpluses over the next decade amount to nearly $3 trillion--approximately a quarter of a year's GDP.

The natural impulse--confronted with such projected surpluses in the short- and medium-run--is to use them: either for tax cuts, or for new federal spending programs, or to pay down the national debt to leave the federal government a decade from now in a better financial position.

But the fiscal position for the next ten years is rosy only under three assumptions:

  • That current economic forecasts are realistic and attainable.
  • That the long-term generational deficit in Social Security (and Medicare) will be solved without reference to the short- and medium- budget picture.
  • That the sharp reductions as shares of GDP in spending explicitly appropriated by Congress year-by-year--"discretionary" spending--during the 1990s are desirable and sustainable.


Are Current Economic Forecasts Reasonable?

The CBO forecasts a slow upward increase in the U.S. unemployment rate of 0.2 percentage points per year in each of the next six years to get the unemployment rate up from its current level of 4.3% to CBO's current 5.5% estimate of the long-run steady-inflation sustainable level of unemployment. CBO projects a steady unemployment rate thereafter.

Given the behavior of the economy in the 1980s and 1990s, a steady-inflation sustainable level of unemployment of 5.5% seems reasonable. Of course, reality will be different from the forecast: the next decade is at least as likely to see a steady-inflation sustainable level of unemployment averaging 4.5% (or below) or 6.5% (or above) than one in the range of 4.5% to 6.5%.

CBO forecasts a rate of growth of potential GDP--the level of production associated with unemployment at its steady-inflation rate--averaging 2.7% per year. Projected growth in real inflation-adjusted GDP over the next six years is 2.3% per year--slower than projected growth in the economy's productive potential because of the rising unemployment rate. Projected growth in real inflation-adjusted GDP over the rest of the ten-year period is equal to the projected rate of growth of potential output: 2.7% per year.

CBO's forecast of potential output growth is significantly faster than the experience of the 1990s by 0.7% per year. This is a big number: over a decade growth faster by 0.7% per year raises the projected level of (nominal) GDP in 2009 by $1 trillion, raises projected federal tax receipts in 2009 by approximately $300 billion. Thus faster potential output growth than seen in the 1990s accounts for two-thirds of the year 2009 projected unified surplus and for one and a half times the projected-on budget surplus.

CBO's forecast of potential output growth is higher for two reasons: first, the estimating agencies have undertaken and are planning revisions to the ways in which they calculate price indices to better take account of improvements in quality and choice; second, CBO is forecasting higher capital accumulation and faster total factor productivity growth.

Each of these two accounts for about half of the change in projected growth vis-à-vis the 1990s.

CBO's assumptions about revised procedures for calculating price indices are reasonable. CBO's assumptions about faster productivity growth and capital accumulation seem less so. A higher rate of investment and of productivity growth over the next decade would be welcome. But where is the evidence that it is coming to pass?

Revising down the projected rate of labor productivity growth over the next decade by 0.4% per year reduces the year-2009 projected surplus by $190 billion (and reduces projected surpluses in intermediate years by proportional amounts).

Of course, actual economic growth outcomes over the next decade are likely to be either much better or much worse than CBO is now proposing--only we cannot now tell which.

Which Budget Concept Is Relevant? The "Unified" or the "On-Budget" Surplus?

The United States has a social insurance system--Medicare, Social Security, and Medicaid--geared to the relatively fast-growth of the 1950s and 1960s. But since 1973 growth has been much, much slower: thus the funding sources of the U.S.'s social insurance system are not in balance with long-term projected levels of spending. It is possible to either (a) ignore this (believing that solving Social Security and long-run Medicare and Medicaid funding problems should be left to the decade of the 2010s and 2020s) and treat short- and medium-term projected budget surpluses as resources available for tax cuts and spending increases, (b) attempt now to look at the entire future spending and revenue projections as a unified whole (as, say, former Joint Tax Committee Deputy Staff Director Alan Auerbach advocates), or (c) strike a middle ground by looking at the on-budget balance only.

Possibility (b) is a lot of work--too much work for this memo--and makes it hard to communicate with journalists, politicians, and policymakers who do not want to be forced to take an intensive course in generational accounting. Possibility (c) is a halfway house that does not solve the long-run social insurance funding problems--they are still there, even with the Social Security surpluses of the next decade earmarked for paying down the national debt held by the public. But possibility (c) does give a somewhat more realistic picture of the federal government's free resources and commitments over the next decade.

The unified--"non-Social Security"--budget surplus with economic growth assumptions more in line with the experience of the 1980s and 1990s is projected to be effectively zero for the entire next decade.

Discretionary Spending.

The Budget Enforcement Act [BEA] of 1990 has proven remarkably effective in constraining spending by Congress through annual appropriations--the "discretionary" part of the budget. Discretionary spending that was 10% of GDP in 1986 and 9.1% of GDP in 1991 and is now 6.6% of GDP. CBO projects--under the assumption of the continuation of a BEA-like process--that total discretionary spending on both defense and non-defense will fall to 5% of GDP come 2009.

A decade of experience has shown that the BEA process is remarkably effective at constraining appropriations. But do we want to use a continuation of such a constraining process for our baseline? The baseline is supposed to give an idea of what spending and revenues will be if there is business-as-usual--if the government keeps on doing about what it has been doing in the ways that it has been doing it. But the assumption of a continuing long-run fall in discretionary spending as a share of GDP requires that the government abandon missions and programs.


A return of discretionary spending as a level of GDP at its level of fiscal 1998 would cause a further $200 billion swing in the fiscal 2009 budget.


The unified budget surplus projected by CBO for 2009 is $413 billion. Changing economic assumptions to make forecast real productivity growth during the 2000s conform to the 1990s, examining not the unified but the on-budget surplus, and maintaining discretionary spending at the same share of GDP that it is today leads on CBO's other assumptions to a projected budget deficit for 2009 or $214 billion.

One way to read these numbers is that there really isn't any headroom between the federal government's revenues and its current commitments over the next decade--that by 2009 the federal government is likely to have to come up with an extra $214 billion just to keep doing what it is doing (let along take positive steps to "save Social Security"). That is probably the most realistic way to think about policy.

Related Links

Sign up for Brad Delong's (general) mailing list

Add a Comment on This Website

Don't the problems with Social Security disappear given a more or less simple raising of the retirement age? I seem to remember seeing figures for Europe in which this was definitely the case. And given the increase in lifespans, it doesn't seem too unreasonable. As far as I know, it was a raise in the retirement age for women (and not anything else, no matter what the pensions industry says) which accounted for the UK having basically solved its "demographic timebomb".


Contributed by Daniel Davies (daniel.davies@flemings.com) on July 14, 1999.

Professor of Economics J. Bradford DeLong, 601 Evans Hall, #3880
University of California at Berkeley
Berkeley, CA 94720-3880
(510) 643-4027 phone (510) 642-6615 fax

This document: http://www.j-bradford-delong.net/Politics/Bradley/Budget_Notes_19970710.html

Search This Website