Influence analysis for Poisson inverse Gaussian regression models based on the EM algorithm

For Poisson inverse Gaussian regression models, it is very complicated to obtain the influence measures based on the traditional method, because the associated likelihood function involves intractable expressions, such as the modified Bessel function. In this paper, the EM algorithm is employed as a basis to derive diagnostic measures for the models by treating them as a mixed Poisson regression with the weights from the inverse Gaussian distributions. Several diagnostic measures are obtained in both case-deletion model and local influence analysis, based on the conditional expectation of the complete-data log-likelihood function in the EM algorithm. Two numerical examples are given to illustrate the results.     

信用认知对农户借贷还款行为的影响研究——基于河南省的实证分析

信用认知是影响农户拖欠借款的重要因素,并且认知程度越深,拖欠贷款的可能性越低.家庭外出务工人数、家庭公共费用缴纳情况、区位年平均家庭收入显著地影响农户拖欠借款的行为,而户主年龄及其受教育水平、纯收入及借贷利率对农户拖欠借款行为没有产生显著的影响.     

中国创业板上市公司盈利能力的因素分析——基于分位数回归的实证研究

影响上市公司盈利能力的因素有很多,不同的分位数水平下各因素对盈利能力的影响可能存在差异。传统的回归模型只能描述因变量的条件均值受自变量的影响,而分位数回归模型能够依据因变量的条件分位数对自变量进行回归,从而更加精确地描述自变量对因变量的影响。选取截至2013年6月30日的355家创业板上市公司的截面数据,从分位数回归的角度研究了不同分位数水平下各因素对公司盈利能力的影响,并且加入了地区虚拟变量,研究了区位因素与盈利能力的关系。     

基于地理加权回归的上海市房价空间分异及其影响因子研究

利用上海市外环以内2010年12月1014个小区的平均房价数据,通过构建地理加权回归模型,并与基于全局最小二乘法(OLS)进行比较,揭示上海小区房价的空间分异和不同影响因子的影响。研究发现,每增加或减少一个单位各影响因子对房价的影响大小依次为:建成时间,到CBD距离,绿化率,到公园距离,距地铁站距离,距超市距离和距学校距离。同时,地理加权回归分解成局部参数估计优于OLS提供的全局参数估计,它可以深刻的揭示出房价和空间影响因子之间复杂的关系,而且可视化的工具可以用地图的形式更详细的呈现出城市房价的整体景观,这些都是传统OLS无法比拟的。     

WHICH FISCAL MULTIPLIERS ARE REGIME‐DEPENDENT? A META‐REGRESSION ANALYSIS

We analyse whether estimated multiplier effects are systematically higher if the economy suffers a downturn. For that purpose, we conduct a meta‐regression analysis on a unique data set of 98 empirical studies with more than 1800 observations on multiplier effects and control for regime‐dependence of the multiplier. We find that spending multipliers are much higher (by about 0.7–0.9 units) during a downturn. Tax multipliers are not sensitive to the economic regime, and generally lower than spending multipliers. Finally, for all spending categories other than government consumption, the multiplier significantly exceeds one during downturns.     

Detecting nonlinearity in time series by model selection criteria

This article analyzes the use of model selection criteria for detecting nonlinearity in the residuals of a linear model. Model selection criteria are applied for finding the order of the best autoregressive model fitted to the squared residuals of the linear model. If the order selected is not zero, this is considered as an indication of nonlinear behavior. The BIC and AIC criteria are compared to some popular nonlinearity tests in three Monte Carlo experiments. We conclude that the BIC model selection criterion seems to offer a promising tool for detecting nonlinearity in time series. An example is shown to illustrate the performance of the tests considered and the relationship between nonlinearity and structural changes in time series.