1- Allameh Tabataba’i University
2- Allameh Tabataba’i University , F.matoori@yahoo.com
Abstract: (62 Views)
Econometric approaches can be analyzed based on their choice of model evaluation criteria, including economic theory, empirical evidence, forecasting, and policy relevance. Historically, theory assumptions often fail to align with non-stationary economic data due to location shifts caused by structural changes. This leads to forecasting and policy simulations failure, and in some cases, the mathematics behind their derivations fails when such shifts occur. This paper thus reviews the challenges of these criteria, then summarizes the Autometrics approach, which nested 'theory-driven' and 'data-driven' approaches, hence often leading to more candidate variables, N, than observations, T. This approach incorporates other potentially relevant variables of theories, dynamic reactions, nonlinear functions, deterministic and stochastic trends, and indicator variables (IIS, SIS) for breaks and exogeneity besides theory-models' variables. So, theory-models’ parameter estimates are unaffected by selection despite searching and without imposing to the model. The unrestricted model is solved by the Autometrics multiple-path model selection algorithm, and diagnostic checks on the reduction process led to the congruent, parsimonious, encompassing model.
Type of Study:
Research |
Subject:
econometrics Received: Feb 28 2025 | Accepted: Nov 30 2024