Conditional Asymmetries in Real GNP: A Seminonparametric Approach
使用半非参数方法，发现实际GNP增长存在条件不对称性，挑战了传统线性与高斯假设的合理性。
Two critical assumptions are often made in empirical research regarding the relationship between economic variables and economic disturbances—linearity and Gaussianity. Together, these two assumptions place strong restrictions on the time series behavior of a model. Most important, these restrictions imply conditional symmetry. Using seminonparametric (SNP) techniques, this article presents evidence that real gross national product growth displays conditional asymmetry. Although these results confirm related results of Brock and Sayers, Sichel, and Hamilton, the SNP approach is novel in that it emphasizes the relationship between common modeling assumptions and the restrictions that these assumptions place on data.