摘要：Nevada gross taxable gaming revenues constitute an important revenue source for the state, just as revenues from lottery sales and taxable sales of service industries make major contributions to the general funds of other states. Typically, these data series are nonstationary and can exhibit periods of accelerating growth. The local linear, or stochastic, trend is proposed as an alternative to other time series methods such as the V AR approach. The inherent feature of the stochastic trend is that it provides a local approximation to a linear trend by allowing the level and slope to evolve over time according to a random walk mechanism. This type of structural time series model was used to forecast gaming revenues. It was found that forecasting performance exceeded that of a V AR model and that forecasts were adaptable to changing business conditions.
