Skip Navigation

Chairman Kerrey, Chairman Portman, and Members of the Commission, it is a pleasure to appear before you today to provide an overview of the Congressional budget process as it affects the work of the Commission in considering alternatives for restructuring the Internal Revenue Service (IRS). My statement contains three parts. First, I will discuss the budgetary outlook for discretionary spending programs in general and how that outlook may affect appropriations for the IRS. Second, I will review the experience in linking increased appropriations for the IRS with possible increases in federal revenues. Third, I will describe the current budgetary scorekeeping conventions and how they may affect IRS appropriations.
 

THE OUTLOOK FOR DISCRETIONARY SPENDING PROGRAMS

Since 1991, discretionary spending programs have been subject to statutory limitations on the total amount that could be appropriated each year by the Congress. The statutory limit or caps were initially set for the five-year period from 1991 through 1995, but were extended in 1993 for another three years through 1998. In any plan to balance the budget by 2002, the caps are expected to be extended again for an additional four years.

In general, overall discretionary appropriations have been held constant in nominal terms since 1991. The overall spending limits for 1998 discretionary appropriations are at approximately the same levels as they were in 1991--$534 billion in budget authority and $546 billion in outlays. At this stage, it is not possible to say what limits might be placed on total discretionary spending beyond 1998. If the 1997 budget resolution is any guide, however, the prospect for any significant increases is not particularly bright. In fact, the 1997 budget resolution called for lower levels of discretionary spending in 1997 and 1998 than were set by the statutory caps, with nominal decreases through 2002.

The President's 1998 budget calls for increases in total discretionary spending levels, particularly for nondefense programs. The proposed spending levels for 1998 through 2002 are substantially above the 1997 budget resolution targets for that period and presumably will be an important focus of any budget negotiations between the Congress and the Administration.

Despite the overall constraint in total discretionary spending, discretionary appropriations for the IRS have increased significantly since 1990 (see Table 1). However, in the last few years, IRS appropriations have been reduced from the $7.5 billion level reached in 1995. The 1997 appropriation for IRS discretionary programs was $7 billion--$1 billion less than the President requested. The President's 1998 budget calls for an increase of $326 million for the IRS, which would bring discretionary appropriations back to nearly the 1995 level. In addition, the President is requesting $1 billion to be divided equally between 1998 and 1999 for investments in IRS information technology.
 

In a time of constrained spending limits, there is an obvious need to make a convincing case to the Congress for any increases in annual appropriations. In the past, the IRS has made the case that increased funding would more than pay off in terms of increased revenue collections, but that case has proven to be a difficult one to make convincingly. A factor contributing to that difficulty is that the increased revenue collections resulting from investment in IRS resources have been difficult to identify. A tendency to overpromise identifiable increases in revenue collections raised questions as to how effective the funded initiatives really were in increasing revenue collections. I will first explain the identification problem and then review links made in the past between IRS funding levels and revenue collections.
 

LINKS BETWEEN IRS FUNDING AND REVENUES

An important hurdle confronting investment in IRS enforcement initiatives is the difficulty of determining if such initiatives increase tax collections, even after the fact. Certainly, if a sufficient financial tracking apparatus is in place, some specific investments can have measurable outcomes--for example, investments in collecting past liabilities or in collecting taxes from identifiable groups of taxpayers. General Accounting Office reports have described financial management systems for tracking such investments and measuring their net effects on revenues.

Nonetheless, when examining total revenues collected by the IRS in a given year, the component of that total generated by new enforcement initiatives cannot be easily isolated. That problem arises because income and payroll tax revenues are affected by many influences. Projecting revenues in future years is a highly imperfect art. Even after those revenues have been collected and reported, some causes of the discrepancies between those projections and the actual revenue collections remain elusive. In addition to being affected by frequent changes in tax law, taxpayer behavior, and the timing of tax payments that are difficult to anticipate and measure, income and payroll tax receipts are very sensitive to the state of the economy. The effect of changes in the economy on tax receipts can be dramatic: the Congressional Budget Office (CBO) has estimated that economic growth of just 1 percentage point above that assumed in our macroeconomic forecast would increase income and payroll tax receipts by about $20 billion above our baseline budget projection for 1998 and by close to $40 billion above our projection for 1999.

Annual revisions to the revenue projections of CBO and the Department of Treasury are routinely of that scale and more, generated mostly by changes in the outlook for the economy. Last April, for example, both CBO and the Department of Treasury were surprised by the strength of final payments on 1995 income-tax liabilities. Those payments were close to $15 billion higher than the available data on personal income and tax collections had indicated. Because data on income lag far behind the cash receipts--for example, Statistics of Income data lag two years behind collections--revenue estimators are constantly playing a game of catch-up in their efforts to interpret developments in tax collections. In that context, increased revenues attributable to enforcement initiatives could not be singled out with precision within the $60 billion to $70 billion in additional income and payroll tax revenues projected for each of the next few years. Acknowledging that constraint on verifying the overall revenue return from a restructured IRS should help the Commission avoid overpromising a measurable return on such overall investment.

Now, what evidence do we have that increased IRS funding would reliably increase revenues? The evidence from past IRS initiatives is not solid. When links between IRS funding and increased revenues have been claimed in the past, they have not been convincingly documented, and their efficacy has not been verified.

The President's annual budget proposal provides one means of articulating a multiyear plan for IRS funding, including investing in resources that would produce increased tax collections in future years. CBO's reestimate of the President's budget is the one occasion on which CBO goes on record evaluating proposed IRS enforcement initiatives. CBO's analysis of the President's budget is not a cost estimate for a piece of legislation. It is a reevaluation of the President's total budget package using CBO's macroeconomic assumptions and technical estimating methods.

In the late 1980s and in 1990, in reestimates of Presidents' budgets, CBO sometimes gave credit for potential revenues from proposed increases in IRS funding for additional enforcement staff and other revenue-related resources. Rules of thumb were applied to funding increases to calculate potential revenue increases. Those rules of thumb were based on information provided by the IRS. Over time, the rules of thumb were subject to increasing skepticism because of the IRS's inability to document that such initiatives had resulted in net increases in revenues so quickly after being put in place. The General Accounting Office reported that, contrary to the assumptions on which such rules of thumb had been based, some enforcement initiatives had actually reduced net revenues in the early years as experienced IRS personnel were temporarily reassigned to train new staff. That outcome was plausible even in cases of initiatives that were likely to boost revenues in the longer run.

President Bush's budget for fiscal year 1991 included an IRS management reform initiative that proposed to improve the management of IRS enforcement in a way that would generate greater revenues without increasing the resources expended on IRS activities. The Administration estimated the potential revenue increase at $4.1 billion over fiscal years 1991 through 1993. CBO, in the reestimate of that budget, attributed zero revenue increase to that initiative on the grounds that the effort was within the IRS's power to accomplish without any Congressional action; no observable event could be identified as generating the net additional revenue; and the IRS was unable to document how the reallocation plan would be implemented. Recent budgets have not explicitly included increases in revenues resulting from proposed enforcement initiatives.
 

ISSUES RELATED TO BUDGET SCOREKEEPING

In some situations, Congressional scorekeeping rules preclude CBO from assigning savings to proposals to improve the management of federal programs, such as tax administration. Even before the passage of the Congressional Budget Act and the creation of CBO, Congressional scorekeeping employed the principle that changes in discretionary appropriations for administrative activities do not produce scorable savings or costs in direct spending programs or tax receipts. That principle was given the force of law by the scorekeeping rules included in the conference report accompanying the Omnibus Budget Reconciliation Act of 1990 (OBRA-90).

Although somewhat arbitrary, the principle is consistent with the current structure of the budget process, which assumes a clear distinction between the budgetary effects of discretionary spending on the one hand and of mandatory spending and revenues on the other. The Balanced Budget Act, as amended by OBRA-90, controls mandatory spending and revenues by the pay-as-you-go process. Separate caps on budget authority and outlays limit discretionary spending.

Congressional controls on revenues and mandatory spending involve a five- or 10-year horizon, but the controls on discretionary spending apply only to the budget year. The costs of bills affecting revenues or mandatory spending are measured as deviations from current law, whereas appropriation bills are assigned their full cost. In addition, the current process lacks a mechanism for charging or crediting the Appropriations Committees with changes in revenues resulting from changes in funding levels. Moreover, if savings were scored for increases in administrative funding, costs would have to be shown if administrative funding was reduced. Such savings or costs might arise from even small increases or decreases in many budget accounts, thus significantly complicating the scoring of appropriation bills.

In three instances, special provisions have been made to facilitate initiatives to improve the administration of federal programs, but the basic scorekeeping rule has been maintained. OBRA-90 provided that the discretionary spending limits for 1991 through 1995 would be increased if appropriations for the IRS exceeded the CBO baseline. The Budget Committees and the Office of Management and Budget agreed to include an estimate of the resulting increase in revenues as part of the projected savings from the bipartisan budget agreement of 1990. However, the additional revenues were not attributed to a particular bill and were not reflected in the Congressional scorekeeping system.

The Contract with America Advancement Act of 1996 makes similar adjustments to the discretionary spending limits for 1996-2002 if additional appropriations are provided for continuing disability reviews by the Social Security Administration. Although increasing the number of reviews would lead to savings in the Social Security Disability Insurance and Supplemental Security Income programs, CBO's cost estimate for the bill did not include those savings because they depended on future Congressional action.

The 1995 budget resolution established procedures that would allow $405 million in spending for enhanced IRS compliance activities above the amounts in the President's 1995 budget request. The resolution provided for adjustments in the Appropriations Committees' allocations so that such additional funding could be accommodated without forcing reduction in other programs. That arrangement was justified on the grounds that the expenditures were expected to generate at least four times more in revenue collections than they cost. That procedure was possible only because the Congressional allocations for discretionary spending were below the statutory caps established in the Budget Enforcement Act.

President Clinton's next budget, for fiscal year 1996, proposed a second $405 million installment of added spending focused on hiring additional revenue officers to increase enforcement through more frequent face-to-face contacts with taxpayers. Although the continued funding of the 1995 IRS tax compliance initiative was endorsed by the Budget Committees, the special provision in the 1995 budget resolution was repealed and that funding add-on was not included in the 1996 appropriation (see Table 1). The President's budget for fiscal year 1997 included $359 million for a revenue protection initiative focused on further increasing early contacts with taxpayers, but that funding also was not included in the subsequent appropriation for the IRS.
 

CONCLUSION

Advocates of investing in a restructured IRS thus can take two approaches to obtaining additional funding. They can battle for an increasing share of a dwindling discretionary total, or they can seek legislation that would increase the limits on discretionary spending to accommodate additional funding for the IRS. Either way, they must make a convincing case that this investment would produce real returns in the long run.