Forecasting the federal budget with time-series analysis

This paper demonstrates a simple, yet interesting application of a multivariate autoregressive-moving-average model to forecast the federal budget using time-series data. Multivariate models are generally considered to be more useful than most univariate models when explaining causal relationships among budget and related variables, or identifying some leading indicators of the series to be forecast in order to improve the predictions attained by a univariate model. The model used here is a special case of a multivariate method, called vector autoregressive-moving-average, because of its ability to deal with structural equations involving time-series data much better than conventional econometric methods.