ALBERT

Notes: Abstract A growing theoretical literature on optimal taxation predicts that governments will set the tax rates on money holdings and on more traditional tax bases to minimize the deadweight losses of collecting government revenue. Under the presumption that relative collection costs and tax bases have not changed significantly over time, the empirical time-series seigniorage literature has focused on the theory's tax smoothing implication, finding only weak support. We show that changes in collections costs and tax bases played an important role in the determination of tax composition and find stronger support for tax smoothing when this is taken into account.

Notes: Conclusion For the most part, economic studies have not exhibited even a rudimentary awareness of the elements of the consolidation record discussed in the previous section. This neglect stems from an implicit assumption that the institutional structure underlying consolidation has no impact on outcomes. Models have been constructed as if real-world consolidations require the consent of all parties. While not logically contradictory, we have argued that this approach has little to do with actual processes. Most successful consolidations simply do not provide all residents of the metropolitan community with the opportunity of choosing between the status quo and agglomeration. Redistributive processes are the norm. Future studies on consolidation must proceed from this basic observation.

Notes: Abstract This paper applies the theory of bureaucracy to the internal operation of the Federal Reserve System. According to the bureaucratic model, the Board of Governors will use budget allocations to the individual Reserve Banks as a means of controlling their provision of information. By examining the research publications of Reserve Banks, we find two Banks which began providing ‘adverse publicity’ for the System in recent years. The empirical analysis reveals that the budgets of these Banks have decreased following the changes in research agendas.

Notes: Concluding remarks While the final version of the 1986 Tax Reform Act retained budget funding of the IRS, the Senate proposal to finance spending from audit revenues represented a seriously debated alternative that continues to receive attention as a method for increasing enforcement. At one level, the ultimate failure of the proposal is puzzling. Given Congress's unwillingness to raise tax rates to cover spending, the budget deficits of the 1980s must be financed from some other source — inflation, borrowing or increased enforcement. Turning the IRS into a bounty-hunting agency would seem to be a straightforward way of producing extra revenue. This popular view reveals a basic confusion about the behavior of bounty-hunting agents. It implicitly assumes that a bounty-hunting agency would behave as a general revenue-maximizing Leviathan and automatically increase enforcement above the status quo budgetary level in pursuit of additional revenue. But bounty hunters want to maximize net audit revenues — not net general revenues. As I have emphasized, too thorough a hunt will reduce the bounty. The possibility of taxpayer adjustment implies a Laffer-like relationship between audit revenue and enforcement. Conceivably, enforcement activity could be raised to the point where taxpayers choose to report all of their income. Then audits would raise no revenue. Increased revenue and increased output in the form of enforcement do not necessarily go hand-in-hand. Allowing pure bounty hunters to spend all audit revenues could lead to either increased or decreased enforcement. Conceptually, Congress could obtain the enforcement outcome it desired by determining the representative individual's taxable base (Y), calculating the taxpayer response to changes in enforcement (dx*/dL), choosing the ‘correct’ model of the agency head's behavior and finally specifying in the financial proposal the appropriate fraction, m, of audit revenues that agents may spend. One potential problem with this procedure is its presumption that information is costlessly available to Congress. If Congress does not know how taxpayers respond to changes in enforcement, for instance, then it might miscalculate the appropriate revenue fraction. Indeed, my tentative prediction that the Senate's revenue multiple would have lowered enforcement suggests the difficulties in specifying the ‘correct’ multiple. But this problem would not appear to be insurmountable. In a static environment, where the taxpayer adjustment function does not change over time, Congress could discover the ‘correct’ revenue multiple through a trial and error process. Then, at a later date, Congress could amend the financing structure accordingly. A more serious problem confronts Congress once we acknowledge that the economic environment may change. Consider the possibility that taxpayer income changes within and across legislative periods. A revenue fraction that was appropriate initially would not necessarily be appropriate several years, or even months later. This helps explain why the Senate proposal ultimately failed even though it seemed ideally suited for the revenue ‘crisis’ of the 1980s. The failure of the 1986 proposal also suggests why legislators choose to budget finance the operation of most bureaus rather than place them on a type of performance payment schedule. Once government specifies a payment formula, it loses control of the bureau's output over the contract period. In contrast, the appropriation process allows government to use the budget as a tool to induce in-period and across-period adjustments in the output of a bureau when conditions outside legislators' control change unexpectedly.