This study analyzed the determinants of the explosion in the caseload of the United States federal district courts that commenced in 1960. First, the study sought to provide forecasts of future demands on the federal courts while reducing forecasting errors by taking account of the time series properties of the case data. The researchers constructed a comprehensive dataset based on annual aggregated civil and criminal case volumes of individual federal district courts spanning the period 1904-1998, for a total of 95 yearly observations. Secondly, the study specified and estimated multivariate econometric models of the determinants of civil case filings over time and across geographic space using panel data techniques. These empirical models were run on three alternative datasets consisting of observations on statewide, districtwide, and circuitwide United States civil, private civil, and total civil cases per capita, over the period 1960 to 1998. The empirical models included standard socioeconomic variables, such as income, population density, and race, along with variables that controlled for fixed effects associated with the courts' geographic location. The study also addressed the pressing issue of allocating judgeships across circuits and districts. Variables include total civil and criminal cases, percentage of minority population, unemployment rate, percentage of drug and immigration cases, annual unweighted and weighted total case filings per judge, and annual civil and criminal case filings per judge.