Skip Main Navigation   Site Search

BASIS OF SOCIAL SECURITY TAX RECEIPTS IN THE BUDGET

Most budget receipts equal cash collections. However, Social Security records the estimated tax liability incurred, adjusted for the estimated timing of tax payments and for reestimates of tax liability in previous years. Individual income taxes are a residual, after subtracting the estimated Social Security employment tax liabilities from total cash collections of income taxes and employment taxes.¹ Adjustments for reestimates of Social Security tax liability amounted to $5.6 billion in the first three quarters of 2001. In the past, this has not concerned decision makers, because it has no effect on the unified surplus. However, it reduces the apparent on-budget surplus for 2001 by $5.6 billion. When the fiscal policy goal is defined as avoiding a budget deficit excluding Social Security, the inaccuracies of the traditional estimation system can distort perceptions about the success or failure of the Nation's fiscal policy.

Receipts as Estimated Liability.—By law, the Social Security trust funds are credited with amounts equal to the tax liability incurred. The budget records receipts equal to the estimated tax liability, with the receipts spread over the months that taxes are expected to be paid. If actual tax payments are less than the estimated liability, the general fund implicitly makes up the difference. The initial estimate is based on the economic assumptions for the President's budget at the time the estimate is made. For example, the receipts reported in the March 2000 Monthly Treasury Statement were estimates of liability prepared in December 1999, using the economic assumptions for the 2001 budget. Beginning about one year after the initial estimate, as information about employers' reports of wages paid becomes available, adjustments are made to correct for differences between the estimated and actual tax liability incurred. For example, the first adjustment for January-March 2000 was made in March 2001. The adjustments can be increases or decreases. Because wage reports trickle in over many years, each adjustment reflects reestimates of tax liability for several years.

Budget Scoring of Adjustments.—(Reestimates of Social Security tax receipts are recorded as adjustments to current year receipts, not as revisions to the published estimates for the year in which the liability was incurred.)  ² About one-half of the June 2001 adjustment reflected actual reported wage data for 2000, one-third for 1999, and the remainder for all prior years. As a result, the amount of employment taxes recorded in the budget do not correspond to either the tax liability incurred or the taxes actually paid in that year. As explained above, the adjustments also affect the amounts reported as individual income tax receipts. If employment tax receipts are adjusted upwards, income tax receipts are automatically reduced by the same amount.

On-budget Surplus Effect.—These adjustments have no effect on the unified budget surplus, but they affect the Social Security surplus and the on-budget surplus by equal and opposite amounts. If Social Security tax receipts for the current year are increased by an adjustment for employment tax liability for previous years, then income tax receipts for the current year are decreased by the same amount. This artificially inflates the Social Security surplus and decreases the on-budget surplus. As a result, the reported on-budget surplus for the current year is distorted by the amount of the reestimates of employment tax liability in previous years. The impact can be large in comparison to the on-budget surplus. For the first three quarters of 2001, the adjustments raised the reported Social Security surplus and reduced the on-budget surplus by $5.6 billion.

This raises the following question for policy makers. If the fiscal policy target is set in terms of the on-budget surplus, should the success or failure of achieving the target be affected by a scoring convention that has nothing to do with actual tax collections in that year? Or should the targets exclude adjustments for events that occurred in prior years?