Revised October 26, 1999
Spending a Non-Existent Surplus
by James Horney
Several pieces of legislation now under serious consideration in Congress entail either substantial increases in expenditures or substantial reductions in revenues starting in fiscal year 2001. These pieces of legislation include: the health insurance tax cuts that the House attached to the patients' bill of rights legislation; the package of tax cuts the House Republican Leadership plans to add to minimum wage legislation; pending legislation to roll back some of the Medicare savings provisions included in the 1997 Balanced Budget Act; and legislation the House Ways and Means Committee has passed to extend a series of expiring tax credits.
Promoters of these pieces of legislation generally argue that these tax reductions and expenditure increases will be financed from the non-Social Security surplus. Such statements seem to reflect a belief that although Congress is encountering major difficulty in avoiding a non-Social Security deficit for fiscal year 2000, this problem is a one-year phenomenon and ample non-Social Security surpluses will appear starting in 2001.
This view is mistaken. Based on the levels of discretionary spending reflected in Congressional action to date on appropriations bills for the current fiscal year, the best prognosis now is that no significant non-Social Security surplus can be counted on until about fiscal year 2006. As a result, the pending legislation would likely result in non-Social Security deficits over the next five years.
This does not mean all such legislation should be summarily dismissed. As budget analysts have noted, modest non-Social Security deficits in the next few years would injure neither Social Security nor the economy.(1) But Congress should understand and be straight with the public about the fiscal consequences of actions it is taking. Members of Congress should not simply assert despite contrary evidence that these tax reductions and expenditure increases will readily be covered by non-Social Security surpluses.
There has been considerable controversy concerning claims by the Republican Leadership that its budget plan will produce balance in the non-Social Security part of the budget in fiscal year 2000. According to Congressional Budget Office estimates, the legislation the House and Senate have produced to date falls well short of this goal.(2)
To try to support this claim, the Leadership has issued directives that various appropriation bills be assumed to result in billions of dollars less in expenditures than CBO believes the bills actually will cost. Even with those unrealistic estimates, however, the spending provided in the House and Senate appropriation bills would produce a non-Social Security deficit. Thus, Congress now is considering a variety of devices to produce "savings" that will allow the Leadership to continue to maintain that it is balancing the non-Social Security part of the budget. Among the proposals that are under consideration or have been adopted are a delay in the earned-income tax credit refunds that low- and moderate-income working families are due; a slight delay in some payments to defense contractors; and, a three-day delay in the last fiscal year 2000 paycheck for military personnel so that a portion of their 2000 pay will count as expenditures in fiscal year 2001.
On top of this, lawmakers and Congressional committees seeking to move tax and entitlement legislation are resorting to a variety of other devices to keep the tax cuts and entitlement spending increases they are promoting from resulting in increased expenditures or reduced revenues in fiscal year 2000. For example, a bill that the House Ways and Means Committee has passed to extend expiring tax credits would offset the cost of the extensions in fiscal year 2000 by temporarily changing the rules that govern estimated tax payments from high-income taxpayers; this maneuver would shift $1.8 billion of these estimated tax payments from 2001 into 2000. In other cases, the date on which the provisions of a tax or entitlement bill become effective would be delayed to postpone costs until 2001. For instance, the package of health insurance-related tax breaks that the House recently tacked on to the patients' bill of rights would reduce revenues by about $9 billion over the 2001-2004 period but have no cost in 2000 because the tax cuts would not take effect before 2001.
One common element in these efforts to avoid both expenditure increases and revenue losses in 2000 is that, at best, they do nothing to improve the budget picture in 2001 and years after that. In fact, in most cases, these efforts make the budget situation more problematic in 2001 and subsequent years. This is seen, for example, in provisions to shift estimated tax payments scheduled for 2001 forward into 2000 and to shift pay dates from the end of 2000 into the first days of 2001.
In part, this pushing of problems into later years and this solving of an immediate fiscal problem at the expense of fiscal conditions in subsequent years reflect the normal human tendency to address today's problems today and worry later about tomorrow. But the growing momentum toward such schemes on Capitol Hill also reflects two assumptions that many policymakers appear to hold: 1) that under CBO's current projections, there will be a large non-Social Security surplus available that can readily cover increased spending or tax cuts in 2001 and years after that; and 2) that the budget projections which CBO will produce early next year will reflect substantially larger non-Social Security surpluses.
Unfortunately, the first of these assumptions is simply incorrect, while the second is highly speculative. As a result, the easy assumptions that various pieces of legislation can readily be financed from non-Social Security surpluses in years after 2000 lack any firm foundation.
Non-Social Security Surpluses Under Current Assumptions
CBO's most recent baseline, issued in July 1999, projects that the non-Social Security part of the budget will feature a surplus of $14 billion in fiscal year 2000 and that this surplus will grow to $38 billion in 2001 and larger amounts in subsequent years (see table on page 5). These projections, however, assume that Congress and the President will strictly comply with the discretionary caps in 2000 through 2002 and that discretionary spending will be held, after 2002, to the level of the cap for 2002 adjusted only for inflation. These CBO projections also assume there will be no emergency spending outside the caps in any year.
The assumption that discretionary spending will stay within the caps is not realistic.
- Actions taken so far on appropriation bills for fiscal year 2000 demolish the assumption that discretionary spending in 2000 will be held to the level of the discretionary spending cap. According to CBO's estimates (and excluding the effects of the proposed delay in earned income tax credit refunds), the outlays flowing from appropriation bills the House has adopted exceed the cap level assumed by CBO for fiscal year 2000 by $36 billion.(3)
- Excluding the emergency funding provided in the appropriation bill for the Department of Agriculture, total discretionary spending in fiscal year 2000 under the House bills is just about equal to the amount CBO estimated would be expended if Congress provided the same level of non-emergency appropriations for fiscal year 2000 as was appropriated for fiscal year 1999, adjusted for inflation. The Senate-passed bills are not substantially different in total.
In addition, the $8 billion in expenditures that would result from the emergency appropriation for farmers included in the agriculture appropriation bill is roughly equal to the average amount of emergency spending provided from 1991 through 1998.(4)
- In short, despite the caps, Congress has not been willing or able to cut discretionary spending for fiscal year 2000 below the 1999 level as adjusted for inflation. It is unrealistic and unreasonable to assume that Congress will somehow make the even-larger cuts in discretionary programs in 2001 or 2002 that would be required to comply with the caps for those years.
- It is increasingly evident that a much more prudent and realistic assumption to make is that non-emergency discretionary spending in fiscal year 2001 and years thereafter will remain at the 1999 level, adjusted for inflation, and that emergency spending will continue to consume at least its $8 billion-a-year average level. Under these far more plausible assumptions (and CBO's economic and technical assumptions), there will be a $13 billion non-Social Security deficit in 2001 (see table on next page).
Furthermore, while there would be a $7 billion surplus in 2002, the non-Social Security budget would dip back into deficit in 2003. Not until 2006 would there be consistent non-Social Security surpluses.
These more realistic projections assume no changes in the laws governing entitlements or taxes. As noted above, various pieces of legislation before Congress would change tax and entitlement laws in ways that would cost more money. Those bills would add to non-Social Security deficits in 2001 and years after that.
For example, H.R. 2923, the bill the House Ways and Means Committee approved to extend the expiring tax credits, would reduce revenues by $7 billion in 2001 and $23 billion over the next five years. Tax cuts included in a bill (H.R. 3081) to increase the minimum wage that the House may consider would reduce revenues by $30 billion over the next five years and by $95 billion over the next 10 years. The health-insurance tax cuts tacked on to the patients' bill of rights also would reduce revenues in 2001 and subsequent years, while legislation under consideration in the House and Senate to scal back some of the Medicare savings provisions of the 1997 Balanced Budget Act would increase entitlement costs in these years. Under CBO's current forecast, the non-Social Security part of the budget would be in deficit every year until 2006 if these various pieces of legislation become law.
Will The Budget Outlook Be Brighter in CBO's Next Baseline?
Some policymakers understand that current CBO projections show no significant non-Social Security surpluses for a number of years when realistic discretionary spending assumptions are used, but do not worry about this matter because they assume the revised budget projections that CBO will issue in January 2000 will eliminate this problem. These policymakers are counting on the new CBO projections to show non-Social Security surpluses large enough to absorb the costs of the tax extenders, Medicare relief, health insurance tax cuts, and the tax cuts slated for inclusion in the minimum wage bill, while also accommodating realistic discretionary spending levels.
On-budget Surpluses Assuming Realistic Discretionary Spending and Proposed Tax and Entitlement Legislation
(by fiscal year, in billions of dollars)
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 10-Year CBO July 1999 Baseline On-budget Surplus1 14 38 82 75 85 92 129 146 157 178 996 Realistic Discretionary Spending2 Appropriations at 1999 inflation-adjusted level -25 -39 -61 -62 -62 -67 -66 -66 -72 -75 -595 Emergencies at 1991-1998 average level -8 -8 -8 -8 -8 -8 -8 -8 -8 -8 -80 Subtotal -33 -47 -69 -70 -70 -75 -74 -74 -80 -83 -675 Debt service -1 -3 -6 -10 -14 -19 -23 -28 -34 -40 -178 Total surplus effect -34 -50 -75 -80 -84 -93 -97 -102 -114 -123 -853 On-budget Deficit (-) /Surplus -20 -13 7 -4 0 -2 32 44 43 55 143 MEMORANDUM: Cost of Proposed Tax and Entitlement Legislation Ways & Means extension of expiring tax credits3 0 -7 -5 -6 -6 -5 -4 -5 -6 -7 -51 Tax provisions in House HMO legislation4 0 -1 -2 -3 -3 ? ? ? ? ? ? Medicare Balanced Budget Act revisions5 -1 -6 -2 -2 -1 -1 -1 -1 -1 -1 -16 Tax provisions in minimum wage bill6 0 -2 -8 -9 -10 -11 -12 -13 -14 -15 -95 NOTES:
1. Assumes that discretionary spending will equal the statutory caps in 2000 through 2002 and will grow at the rate of inflation thereafter.
2. Increases in spending are shown with a negative sign because they reduce the surplus.
3. H.R. 2923, as reported by the House Ways and Means Committee on September 28, 1999. The Senate Finance Committee has approved a bill extending expiring credits that would have no net effect on the surplus over 10 years but would reduce the surplus by $4 billion in 2001.
4. Estimates of the revenue losses in years beyond 2004 are not available.
5. As proposed by Senator Roth on October 13, 1999. Increases in spending are shown with a negative sign because they reduce the surplus.
6. Preliminary estimate of H.R. 3081, introduced by Representative Lazio.
It is, of course, possible that the budget outlook will seem significantly brighter to CBO in January. But a strong dose of caution is in order here. It is far from certain that the January CBO forecast will show any such change.
For the last few years, the budget projections that CBO has issued have improved with each new baseline, and this apparently has led some policymakers to assume this pattern will continue. Unfortunately, such an assumption is supported neither by history nor by budget and economic developments since CBO issued its last baseline in July.
History shows that CBO's projections do not consistently become more optimistic. As recently as the early 1990s, there were extended periods when unfavorable economic or budget developments led CBO to produce budget projections that became progressively bleaker with each new baseline.
Furthermore, and of particular importance, budget outcomes in fiscal year 1999, as well as economic developments since CBO issued its most recent forecast, do not indicate that the January baseline projection will be significantly brighter.
- The primary reason for the improvements in the baseline projections over the last few years has been an unanticipated surge in revenues. Although part of this increase was due to the economy performing better than CBO had anticipated, a substantial part of it was due to so-called "technical" factors not directly related to the economy outperforming the CBO forecast. (It should be noted that some of these technical factors such as an increase in capital gains realizations are related to economic developments even though they are not part of CBO's forecast of basic economic variables.) In 1998, for example, actual revenues turned out to be $56 billion higher than CBO had projected in January 1998, nine months before the end of the fiscal year. CBO estimates that only about $7 billion of this increase was attributable to better-than-anticipated performance of the economy. Nearly $50 billion was attributable to the "technical" factors. Largely on the basis of this increase in fiscal year 1998 revenues attributable to technical factors, CBO acted between January 1998 and January 1999 to increase its projection of revenues for 1999 by nearly $86 billion. (About $16 billion of this $86 billion adjustment reflected improvements in CBO's basic economic forecast.)
- The pattern of large, unexpected increases in revenues due to "technical" factors did not, however, continue in 1999. In January 1999, CBO projected that revenues in fiscal year 1999 would be $1,815 billion. In July, a slightly more optimistic economic forecast led CBO to increase its revenue projection for 2000 slightly to $1,821 (and to increase the projections for each of the following several years by somewhat more than $6 billion). Based on preliminary end-of-year data for fiscal year 1999, CBO now estimates that 1999 revenues were $1,826 billion, an increase of $11 billion above the January projection and $5 billion above the July number. These are very small increases just over one-half of one percent since January and suggest that CBO's July baseline has largely incorporated the effects of the various technical factors that have driven up revenues over the last few years. As a result, further substantial increases in CBO's revenue projections seem unlikely.
- Furthermore, the performance of the economy over the past three months has not differed significantly from what CBO forecast in July. Some news has been slightly more favorable (e.g., income growth has been a little faster than anticipated), but other news higher-than-expected interest rates, for example has been less favorable. Nor is there any clear evidence that long-term economic trends are more favorable than was forecast in July.
- Finally, there is no indication that expenditure projections for fiscal year 2001 and subsequent years will be significantly lower than CBO's current projections assume. Based on CBO's latest estimates, actual expenditures in fiscal year 1999 appear to have been only $5 billion lower than the estimate that CBO issued last January. Moreover, the current estimate of actual 1999 expenditures is $1 billion higher than the estimate CBO issued in July. And there is no indication of a change in the economic outlook that would reduce expenditure projections for 2001 and years after that.
The claim that a variety of policies that would reduce revenues or increase expenditures after 2000 can readily be financed by large non-Social Security surpluses is based on mistaken assumptions. Using CBO's current projections and assuming realistic levels of discretionary spending (i.e., the 1999 level, adjusted for inflation) yields estimates showing no significant non-Social Security surpluses until fiscal year 2006. Nor is there reason to assume that the budget projections which CBO is scheduled to release in January will be significantly more favorable. In short, legislation that would increase spending or reduce revenues in 2001 and subsequent years should not be predicated on convenient but unfounded assumptions that substantial non-Social Security surpluses will be available to finance such legislation in those years.
1. See James Horney and Robert Greenstein, "A Small Non-Social Security Deficit In Fiscal Year 2000 Would Not Adversely Affect Social Security," Center on Budget and Policy Priorities., September 17, 1999.
2. See James Horney, "Would Congressional Spending Plans Avert a Non-Social Security Deficit?" Center on Budget and Policy Priorities, October 6, 1999.
3. In the case of the appropriation bill for the Departments of Labor, Health and Human Services, and Education, the bill approved by the House Appropriations Committee is used in these calculations, since the full House has not considered the bill.
4. This $8 billion on average for fiscal years 1991 through1998 excludes emergency expenditures in the early 1990s for Operation Desert Storm.