Statistical tests of distributional scaling properties for financial return series

Existing empirical evidence of distributional scaling in financial returns has helped motivate the use of multifractal processes for modelling return processes. However, this evidence has relied on informal tests that may be unable to reliably distinguish multifractal processes from other related classes. The current paper develops a formal statistical testing procedure for determining which class of fractal process is most consistent with the distributional scaling properties in a given sample of data. Our testing methodology consists of a set of test statistics, together with a model-based bootstrap resampling scheme to obtain sample p-values. We demonstrate in Monte Carlo exercises that the proposed testing methodology performs well in a wide range of testing environments relevant for financial applications. Finally, the methodology is applied to study the scaling properties of a data-set of intraday equity index and exchange rate returns. The empirical results suggest that the scaling properties of these return series may be inconsistent with purely multifractal processes.     

Nonparametric Varying-Coefficient Models for the Analysis of Longitudinal Data

Longitudinal methods have been widely used in biomedicine and epidemiology to study the patterns of time-varying variables, such as disease progression or trends of health status. Data sets of longitudinal studies usually involve repeatedly measured outcomes and covariates on a set of randomly chosen subjects over time. An important goal of statistical analyses is to evaluate the effects of the covariates, which may or may not depend on time, on the outcomes of interest. Because fully parametric models may be subject to model misspecification and completely unstructured nonparametric models may suffer from the drawbacks of "curse of dimensionality", the varying-coefficient models are a class of structural nonparametric models which are particularly useful in longitudinal analyses. In this article, we present several important nonparametric estimation and inference methods for this class of models, demonstrate the advantages, limitations and practical implementations of these methods in different longitudinal settings, and discuss some potential directions of further research in this area. Applications of these methods are illustrated through two epidemiological examples.     

Comparing mean efficiency and productivity scores from small samples: A bootstrap methodology

This paper provides a bootstrap methodology for constructing confidence intervals for means of DEA and econometrically estimated efficiency scores, Malmquist productivity indices, and other similar measures in small samples. The procedure is nonparametric since no distributional assumptions are required. An empirical example is provided.This research was performed while Wilson was under contract with the Management Science Group, US Department of Veterans Affairs, Bedford, MA 01730.     

Testing for two-regime threshold cointegration in vector error-correction models

This paper examines a two-regime vector error-correction model with a single cointegrating vector and a threshold effect in the error-correction term. We propose a relatively simple algorithm to obtain maximum likelihood estimation of the complete threshold cointegration model for the bivariate case. We propose a SupLM test for the presence of a threshold. We derive the null asymptotic distribution, show how to simulate asymptotic critical values, and present a bootstrap approximation. We investigate the performance of the test using Monte Carlo simulation, and find that the test works quite well. Applying our methods to the term structure model of interest rates, we find strong evidence for a threshold effect.     

出口引导经济增长假说在我国成立吗?——基于Bootstrap仿真方法的实证分析

本文采用最新的适用于小样本的Bootstrap仿真检验方法，对我国出口贸易与经济增长之间的关系进行了实证研究。结果发现，以1952-2003年为样本期的协整分析表明两者间不存在任何长期关系，仿真检验亦显示出口引导经济增长假说不成立，且经济增长也非促进出口贸易的因素。然而，以改革开放之后1978-2003年为样本期的分析结果支持出口引导经济增长假说，这一结论与大多数已有的研究存在显著的区别。     

Bootstrap methods for autocorrelation test with uncorrelated but not independent errors

By using bootstrap technique we investigate the properties of the Breusch [Breusch, T.S., 1978. Testing for autocorrelation in dynamic linear models. Australian Economic Papers 17, 334–355]–Godfrey [Godfrey, L.G., 1978. Testing for higher order serial correlation in regression equations when the regressors include lagged dependent variables. Econometrica 46, 1303–1310] autocorrelation tests in dynamic models with uncorrelated but not independent errors. In this paper we show that, under conditions when the errors are uncorrelated but not independent, even the best likelihood ratio test cannot achieve the asymptotic distribution under the null hypothesis of no autocorrelation. Standard bootstrap methods also fail to produce consistent results. To overcome this problem we applied several bootstrap testing methods for the same purpose and found the stationary bootstrap and Wild bootstrap with static model to perform adequately among the other bootstrap methods.     

Questioning the empirical basis of the environmental Kuznets curve for CO2: New evidence from a panel stationarity test robust to multiple breaks and cross-dependence

This paper investigates the time series properties of per capita CO₂ emissions and per capita GDP levels for a sample of 86 countries over the period 1960-2000. For that purpose, we employ a state-of-the-art panel stationarity test which incorporates multiple shifts in level and slope, thereby controlling for cross-sectional dependence through bootstrap methods. Our analysis renders clear-cut evidence that per capita GDP levels are nonstationary for the world as a whole while per capita CO₂ is found to be regime-wise trend stationary. The analysis of country-groups shows that for Africa and Asia, per capita CO₂ is best described as nonstationary, while per capita GDP appears stationary around a broken trend. In addition, we find evidence of regime-wise trend stationarity in both variables for the country-groups consisting of America, Europe and Oceania. The results of our analysis carry important implications for the statistical modelling of the Environmental Kuznets curve for CO₂, since the differing order of integration in both variables for the world as a whole and for Africa and Asia calls into question the validity of panel cointegration techniques which assume that both variables are nonstationary and cointegrated with one another. Cointegration techniques would not be appropriate either for the case of America, Europe and Oceania which are characterised by per capita GDP and CO₂ emissions being stationary around a broken trend. Similar conclusions are reached when we analyse country-groups based on levels of development. Failure to properly characterise the time series properties of the data by not controlling for an unknown number of structural breaks and for cross-sectional dependence could be responsible for the fragility and lack of robustness surrounding the estimation of environmental Kuznets curves.     

Testing the value of directional forecasts in the presence of serial correlation

Common approaches to testing the economic value of directional forecasts are based on the classical χ²${\chi }^2$-test for independence, Fisher’s exact test or the Pesaran and Timmermann test for market timing. These tests are asymptotically valid for serially independent observations, but in the presence of serial correlation they are markedly oversized, as has been confirmed in a simulation study. We therefore summarize robust test procedures for serial correlation and propose a bootstrap approach, the relative merits of which we illustrate by means of a Monte Carlo study. Our evaluations of directional predictions of stock returns and changes in Euribor rates demonstrate the importance of accounting for serial correlation in economic time series when making such predictions.     

Impact of conversion to Critical Access Hospital status on hospital efficiency

This study examines the impact of Critical Access Hospital (CAH) Program on hospital efficiency using a two-stage approach, where data envelopment analysis is used in the first stage to estimate cost, technical, and allocative efficiency scores of a sample of rural hospitals. Densities of efficiency scores of CAHs and prospectively paid rural hospitals are estimated and compared using a nonparametric kernel density estimator and a bootstrap-based test. In the second stage, efficiency scores are regressed on environmental variables using bootstrapped truncated regressions. Density analysis and results from bootstrapped truncated regressions show that CAHs are less cost and allocatively efficient compared to prospectively paid rural hospitals, without being less technically efficient. Relative to their pre-conversion selves, CAHs appear to be slightly less allocatively efficient, while they are slightly more technically efficient and no less cost efficient. Overall, our results suggest that the CAH Program may have decreased the allocative and cost efficiencies of those rural hospitals that converted to CAH status relative to prospectively paid rural hospitals, without significantly increasing their technical efficiency.     

外资溢出效应的分位数回归研究

将中国30个省市分化为利用外资高水平区域、利用外资中水平区域、利用外资低水平区域,然后运用分位数回归模型研究了外资对中国工业企业的溢出效应。进行100次自举抽样的实证结果表明,就全国而言,外资具有显著的正向溢出效应,但贡献力度要低于内资。分区域而言,利用外资水平越高的地区,外资的溢出效应最为显著;利用外资高水平区域的外资对产出的贡献度要大于内资的贡献度,而其他两个区域则有相反的趋势;工业企业资产规模越大,外资溢出效应的证据越充分。