Testing for unit roots using forward and reverse Dickey-Fuller regressions

Article Abstract:

A simple approach is introduced for testing the null hypothesis that a process has a unit root against a stationary alternative. Since the method is based on the joint values of the standard Dickey-Fuller tests when used to both the forward and reverse data realizations, computation does not need complex techniques and can be done using available econometric software packages.

Methods, Regression analysis, Time-series analysis, Time series analysis

---

Specification testing of Markov switching models

Article Abstract:

A set of tests is presented to evaluate the appropriateness of Markov switching models, for both non-parametric and parametric encompassing and model consistency. Simulation is used in combination with conditional mean estimation and non-parametric density; results are discussed.

Product information, Usage, Evaluation, Testing, Comparative analysis, Test validity, Markov processes, Econometric models

---

Similar abstracts:

• Abstracts: Accounting for income inequality in rural China: a regression-based approach. Determinants of the performance of foreign invested enterprises in China
• Abstracts: Perpetual signalling with imperfectly correlated costs. Auction form preferences of risk-averse bid takers. Revenue-raising versus other approaches to environmental protection: the critical significance of preexisting tax distortions
• Abstracts: Receding horizon control of jump linear systems and a macroeconomic policy problem. Saddle-point calculation for constrained finite Markov chains
• Abstracts: Jamming a rival's learning. Ranking auctions with risk averse bidders. Bidding for the future: signaling in auctions with an aftermarket
• Abstracts: Public pension reform in Europe: policies, prospects and evaluation. Ageing, retirement and pension reforms

This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.