1. Office of Management and Budget, "Federal Government Finances: 1985 Budget Data," typescript, February 1984, p. 99. This is an unpublished but available to the public summary of historical statistics produced to supplement each year's budget. It has since been replaced by a published volume, Historical Statistics.

2. Ibid., pp. 97-99. Estimated as "relatively uncontrollable under present law: open-ended programs and fiscal costs."

3. Donald Axelrod, Budgeting for Modern Government (New York: St. Martin's Press, 1988).

4. No. 58, The Federalist Papers (New York: New American Library, 1961), p. 359.

5. In the cases the president cited, according to the report on the Agriculture appropriation, the research was in New Jersey, Michigan, Louisiana, and New Mexico.

6. The American revolutionaries insisted on the absolute nature of the power of the purse because they had seen how George III used his personal wealth and patronage to buy members of Parliament, thus corrupting and dominating the legislature. Modern members of Congress may not be aware of that history, but they are well aware of the uses of patronage in state and local political machines, such as Mayor Daley's Chicago. The power of the purse is therefore inextricably mixed with the everyday struggles of low politics. In a talk at the Brookings Institution in 1986, one leader of House liberals, David Obey (D-Wis.), quoted a leading conservative, Mickey Edwards (R-Okla.), on the nonpartisan subject of presidential discretion. "Imagine," Obey quoted Representative Edwards, "Lyndon Johnson with an item veto!" Obey then recounted an imaginary conversation: LBJ explained how thoroughly he understood Obey's desire for a community health clinic in a small northern Wisconsin town and how much that town needed the clinic but how badly he needed Obey's support for administration policy in Vietnam. Most legislators would like to avoid such situations, or at least have the upper hand.

7. The new Gramm-Rudman requires an earlier release, but the earlier release of the budget hasn't happened.

8. Frederick C. Mosher, A Tale of Two Agencies (Baton Rouge: Louisiana State University Press, 1984), p. 25.

9. Ibid., pp. 27-32; esp. p. 28.

10. The appropriations committees stayed within the president's totals partly because of a strong norm of economy—if they weren't there to make cuts, why have committees? Richard Fenno's The Power of the Purse (Boston: Little, Brown, 1966) is the classic work on appropriations norms. But one also must suspect more informal coordination than was visible—or so veteran staff and members report today.

11. Fenno, Power of the Purse. Joseph White's research on the House committee confirms this picture.

12. The foreign aid bill, Foreign Operations, is the exception that proves the rule of legislation passage: it often has to be packaged with something more popular. Nobody outside D.C. cares much about either the District of Columbia or legislative branch bills, but Congress, of course, does.

13. See Aaron Wildavsky, The Politics of the Budgetary Process, 1st ed. (Boston: Little, Brown, 1964); the revolution that has overtaken budgeting is described in a successor volume, The New Politics of the Budgetary Process (Glenview, Ill.: Scott, Foresman/Little, Brown, 1988).

14. "FY71." The federal government, like many businesses, operates on a fiscal year that differs from the calendar year. Until the Budget Act was passed, the fiscal year ran from July 1 through June 30; that is, fiscal 1971 began on July 1, 1970. The Budget Act began the year three months later, giving Congress and the president three more months to pass the necessary legislation before the year began. Thus, in 1988 they were considering the budget for FY89, beginning October 1, 1988.

15. See, in particular, Allen Schick, Congress and Money (Washington, D.C.: Urban Institute, 1980); and Gary Orfield, Congressional Power: Congress and Social Change (New York: Harcourt Brace Jovanovich, 1975).

16. Schick, Congress and Money, p. 48.

17. Ibid.

18. The new budget process worked as described here until the Supreme Court, in the Chadha decision, outlawed the legislative veto mechanism at its heart. If the president wished not to spend appropriated funds, he could propose a rescission; if he wished only to delay the spending, he had to propose a deferral. For a rescission to take effect, both houses had to pass a bill approving the change within forty-five legislative days of its proposal. If they did not do so, the money would be spent as appropriated. A deferral would take effect automatically, but it would last no longer than through the end of the fiscal year. Either house could reject the deferral with a vote specifically disapproving it. The procedure exploited the difficulty of congressional action. Rescission, the more serious policy change, was made highly unlikely by requiring positive action by both houses; deferral, the lesser change, was allowed to occur, but either house could choose by majority vote to enforce the appropriation legislation. Thus impoundment could be used when the need for spending had disappeared or been delayed, but the president could not use it to challenge Congress's priorities.

19. The first year of the new budget process, 1975 for FY76, formally a dry run, was taken surprisingly seriously.