By using VAR models, we relate five endogenous variables by industry sector to each other: Producer Price Index(P), wage(W), GDP(Y), Import Price Index(MP), employment(L). We use monthly data covering the period January 1970 through December 1998. The impulse response function describes the response of the employment(L) variable over time to shocks in MP and Y. It shows the response of employment variables by industry sector to a one-unit shock in Y. i.e., a one-period increase in. Over the next 6 or 7 months employment by machinary industry rises, but there is little effect on employment by food industry.
Another way of characterizing the dynamic behavior of the model is through a variance decomposition. This breaks down the variance of the forecast error for employment(L) variable into components that can be attributed to each of the endogenous variables. The greater the forecast horizon, the larger the proportion of forecast variance of employment variables that will be due to price variables(MP, P).
Finally, we try a conditional forecast, in which values for one or more explanatory variables are not known, so that forecasts must be used to produce forecasts of employment variables.