Bureaucratic politics is the favored explanation of those addressing the perversities of defense budgeting. But it is arguably devoid of politics, given its dependence on either aggregate top-down or horizontal models. I seek to redirect analysis. I disaggregate defense spending (by service and weapon type) and study budget sensitivity to program pressures in the buildups and builddowns of the Reagan-Bush eras. Applying a two-equation model to time-series cross-sectional data, the analysis shows weapon budgets increasing with program diversification and a commitment to defense spending. In turn, it shows programs diversifying to accomodate service objectives: when turning to missions, the services increased program varieties while concentrating program resources.
