DYNAMIC ECONOMETRIC MODELS Dr C Ertuna DEFINITION In

DYNAMIC ECONOMETRIC MODELS Dr. C. Ertuna

DEFINITION In Dynamic Econometric Models time plays a central role. Past (lagged) values of dependent or independent variables are introduced in the model to describe the underlying process. Dr. C. Ertuna

TYPES of DYNAMIC MODELS In general there are three types of dynamic models: a) Distributed Lag Models: Models that include lagged values of independent variables. i. The Koyck Transformation ii. The Almond Transformation b) Autoregressive Models: Models that include lagged values of dependent variables. i. The Partial Adjustment Model ii. The Adaptive Expectations Model c) Autoregressive Distributed Lag Models: Dr. C. Ertuna

REASONS for LAGGED VALUES 1. Habit (Psychological inertia) 2. Transition / Time to Adjust 3. Technical or Technological Reasons causing delay in change. 4. Institutional Reasons (such as contracts) Dr. C. Ertuna

USE of DISTRIBUTED LAG MODELS 1. Impact of Advertising (over several periods) on (current) Sales 2. Impact of Safety Training on Accidents. 3. Impact of Marketing Mix on Market Share. 4. Impact of Air Pollution on Mortality Rate. Dr. C. Ertuna

DISTRIBUTED LAG MODELS and NUMBER of LAGS Too many lags may cause multicollinearity and lost of degrees of freedom. In general there are two approaches to overcome those problems: a) Koyck Transformation b) Almond Transformation Dr. C. Ertuna

Koyck Transformation • Dr. C. Ertuna

DISTRIBUTED LAG MODELS and OLS REGRESSION To estimate Distributed Lag Models OLS Regression can be used. The results are BLUE as long as the residuals do not exhibit autocorrelation. Dr. C. Ertuna

END 9