Forecasting Competitors’ Market Share Using Multivariate Linear Time Series Analysis

The paper presents an approach to forecast competitors' market share in the banking industry. Multivariate linear time series analysis was applied to a state space model with stochastic components. Three time series were interrelated as two exogenous variables and one endogenous variable. The interrelation was restricted to stochastic trends. Two dummy variables were added to the model to smoothen the exogenous variables. In the example, market shares development analysis was presented, two competitors correlate positively; that is, they have a nearly parallel development. Both competitors, however, correlate negatively with a third competitor; that is, the third competitor has an antagonistic relation to each of them. This type of analysis allows accurate and reliable forecasts which are important for competitor monitoring. Moreover, the forecasts can be used as an expected case in the market scenario. The paper is an important contribution to the exploration of the quantitative dynamics of competition systems that are generated by the business activities of competitors