All the attempts at budget reform described here have had some impact on the way in which the federal budget is constructed. Both program budgeting and zero-based budgeting were significant rationalistic efforts at reforming the budget process. Although both had a great deal to commend them, neither was particularly successful in changing the behavior of budget decision-makers. The less rationalistic methods, such as Gramm–Rudman–Hollings, have had a somewhat greater impact, in part because they did not attempt to change the basic format.

        Why does the traditional, line-item, incremental budget persist despite the real shortcomings of both the process and the outcomes? There appear to be several reasons. One is that the traditional budget gives the legislature an excellent means of controlling the executive branch. It allocates funds to identifiable organizations for identifiable purposes (personnel, equipment, etc.), not to nebulous programs or decision units. The political and administrative leaders who manage the real organizations to which the funds are allocated can then be held accountable for how they spend the money appropriated to them. Blunt instruments tacked onto the process, such as the Budget Enforcement Act, to enhance the control elements of the budget process without altering it fundamentally.

        More important, although the benefits promised by both planning, programming, and budgeting system and zero-based budgeting were significant, so too were the costs, in terms both of the calculations required to reach decisions and of the political turmoil created. Incremental budgeting provides ready guidelines for those who must make budget decisions, minimizing the necessity for them to engage in costly analysis and calculation. In addition, as most political interests are manifested through organizations, the absence in incremental budgeting of threats to those organizations means that political conflicts can be confined to marginal matters instead of repeated battles over the very existence of the organizations.

        In short, although incremental budgeting does nothing very well, neither does it do anything very poorly. Incrementalism is a convenient means of allocating resources for public purposes. It is not an optimal means of making policy, but it is a means that works. It is also a means of making policy in which policymakers themselves have great confidence. These factors are not in themselves sufficient to explain the perpetuation of the incremental budget process in the face of so many challenges by presumably superior systems. There are always proposals for change, but there is rarely sufficient agreement on which of the possible changes to make to move the system away from the status quo. There have been some improvements, but the system remains firmly incremental.