Consistent specification testing via nonparametric series regression

Article Abstract:

Two consistent one-side specification tests for parametric regression models are presented. The first test involves an analysis of the sample variances between the residual from the parametric model and the discrepancy between the parametric and nonparametric fitted values. The other test is based on the difference in sums of squared residuals between the parametric and nonparametric models. Both tests expand at a rate faster than the parametric rate under misspecification but do not use weighting, sample-splitting and non-nested testing procedures.

Nonparametric statistics, Asymptotic distribution (Probability theory), Equations, Quadratic, Quadratic equations

---

High breakdown point conditional dispersion estimation with application to S&P 500 daily returns volatility

Article Abstract:

Two new estimation strategies were formulated to serve as efficient estimators of conditional dispersion models. Unlike the quasi-maximum likelihood (QML) estimators, the two-stage Hampel estimators and two-stage S-estimators were found to be effective in resisting the impact of outliers, such as market crashes and rallies. The two strategies, when applied to S&P 500 Cash Index series, were able to illustrate volatility dynamics during normal daily market activity.

Econometrics & Model Building, Stock price indexes, Estimation theory, Business models, Standard and Poor's 500-Stock Price Index (Index)

---

Consistent testing for serial correlation of unknown form

Article Abstract:

Three classes of one-sided tests are proposed for consistent serial correlation of unknown form. The test methods are asymptotically equivalent under a suitable class of local alternatives. However, their computing power can be maximized under both global and local alternatives by using the Daniell kernel. The proposed methods have better power than the Lagrange multiplier tests and the portmanteau tests.

Analysis, Testing, Correlation (Statistics)

Similar abstracts:

• Abstracts: Federal budget projections: a nonparametric assessment of bias and efficiency. Exact nonparametric orthogonality and random walk tests
• Abstracts: Unobserved components in ARCH models: an application to seasonal adjustment
• Abstracts: Strategy-proof exchange. Bayesian representation of stochastic processes under learning: De Finetti revisited
• Abstracts: Does school quality matter? Evidence from the national longitudinal survey of youth. Do school resources matter only for older workers?
• Abstracts: Fully modified least squares and vector autoregression. An asymptotic theory of Bayesian inference for time series

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