首页 > 通知公告 > 正文

计量经济学及其应用2017成都研讨会
2017-07-05 11:32:17   来源:   评论:0 点击:

一、研讨会议程研讨会将于2017年7月8日(星期六)在柳林校区格致楼918举行开幕式9:00-9:15经济学院院长易敏利致欢迎辞第一场(9:15-10:15):主题报告主持人:经济学院 邹红教授报告人:Li Tong(李彤), Uni...
一、研讨会议程
研讨会将于2017年7月8日(星期六)在柳林校区格致楼918举行
开幕式9:00-9:15经济学院院长易敏利致欢迎辞
第一场(9:15-10:15):主题报告
主持人:经济学院 邹红教授
报告人:Li Tong(李彤), University of Vanderbilt, USA
题目: A Partial Identification Subnetwork Approach to Discrete Games in Large Networks: An Application to Quantifying Peer Effects
茶歇时间 10:15—10:30
第二场(10:30-12:00):特邀报告
主持人:经济学院 徐舒教授
报告人:Bin Peng (彭彬),英国巴斯大学(University of Bath)经济学院
题目:To Lie or Not to Lie: Survey Mode Effects on the Validity of Self-Reported Substance Use Data
报告人:朱涣君,厦门大学 王亚楠经济研究院
题目:High Dimensional M-estimation for Large Panel Data Models
 
报告人:蔡必卿,华中科技大学 经济学院
题目:A Nonlinear Stock Return Predicting Model

第三场(2:00-3:30):特邀报告
主持人:经济学院 贾男教授
报告人:涂云东  北京大学  光华管理学院
题目: Sieve estimation and variable selection in high-dimensional single index models
报告人:程婷婷  南开大学 金融学院
题目:Multi-step predictive regression model for stock return prediction
报告人:Zen Lu (卢增华), University of Southern Australia, Australia
题目:MinP Score Tests with an Inequality Constrained Parameter Space
茶歇时间 3:30—3:45
第四场(3:45-5:15):特邀报告
主持人:经济学院 马双副教授
报告人:郭萌萌,西南财大 经济管理研究院
题目:Modeling Systemic Risk-- A DCC-Filtered Historical Simulation Approach
 
报告人:黄勔  西南财大 经济管理研究院
题目:Simulation Based Estimation of Multinomial Discrete Choice Model with Fixed Effects
 
报告人:干卓泂 西南财大 统计学院
题目:Factor-augmented Prediction with Idiosyncratic Factors
 
闭幕 5:15-5:30

二、报告人和报告内容简介
Tong Li(李彤),范德比尔特大学(Vanderbilt University)经济学院,Gertrude Conaway 范德比尔特经济学教授。1997年获得南加州大学经济学博士学位,1993年获得加州大学圣地亚哥分校数学博士学位。现担任或曾担任Journal of Econometrics, Journal of Applied Econometrics 和 Journal of Economic Behavior and Organization 等期刊的副编辑。其研究论文发表于Econometrica, JOE, GEB, IER, RES, JAE等等顶级经济学和计量经济学期刊。
 
报告内容摘要:This paper studies identification and estimation of discrete games in large networks, with an application to peer effects on smoking in friend networks. Due to the presence of multiple equilibria, the model is not point identified. We adopt the partial identification approach by constructing moment inequalities on choice probabilities of subnetworks. Doing so not only significantly reduces the computational cost, but also enables us to find consistent estimator of the moment conditions even when the network is large and the friendship relationship structure varies significantly among networks. Monte Carlo studies are conducted to evaluate the performance of the subnetwork approach. In the application using the Add Health data, we find significant and positive peer effects on smoking.
Bin Peng (彭彬),英国巴斯大学经济学院,澳大利亚莫纳什大学统计学博士,其研究论文发表于JOE, JHE等国际顶级计量经济和健康经济期刊。
 
报告内容摘要:We examine effects of survey mode on the validity of self-reported substance use data by exploiting quasi-experimental variations in survey methods in Australian National Drug Strategy Household Survey. A single-index zero inflated probit model is employed to estimate such effects. We find that survey mode has a large effect on the surveyee’s misreporting behaviour for all three substances (tobacco, marijuana, and speed) examined in this paper, but particularly so for marijuana users. Moreover, such effects vary significantly depending on the nature of the relevant substance and the surveyee’s individual characteristics (gender, age, etc.). Finally, we propose a simple behavioural model to understand our findings.
 
涂云东,北京大学光华管理学院,2012年加州大学滨河分校经济学博士。其研究论文发表于JOE, JBES, CSDA等国际顶级统计和计量经济期刊。
 
报告内容摘要:This paper considers sieve estimation in high-dimensional single index models. The use of Hermite polynomial in approximating the unknown link function provides a convenient framework to conduct both estimation and variable selection. The penalized estimation of the index parameter is formulated as solutions obtained from the routine penalized linear regression procedure. The resulting index parameter estimator is shown to be consistent and sparse. The asymptotic normality of the post-selection estimator of the index and that of the link function are established. Numerical results show that both the variable selection procedure and the associated estimators perform well in finite samples.
Zenghua Lu(卢增华),澳大利亚南澳大学经济学院,澳大利亚莫纳什大学统计学博士,其研究论文发表于JASA,CSDA等国际顶级统计学和计量经济学期刊。
 
报告内容摘要:This paper proposes new one-sided score tests for testing the model parameter that is subject to inequality constraints. We construct our score test statistic as the minimum p-values of existing score tests and some individual score tests that are designed for individual testing of elements of the parameter of interest. The proposed score tests are shown to have a more balanced power than existing score tests between the cases of the parameter lying in the central area and the boundary area of the restricted parameter space. Furthermore, we show that our score tests allow for simultaneous model testing within nested models under certain conditions. This feature is appealing that one can identify a model through testing without estimating candidate models thanks to the advantage of score tests. The optimality of the proposed score tests in terms of admissibility are studied. Finally, we illustrate two applications with one for GARCH model testing and the other for random coefficient model testing.
程婷婷,南开大学金融学院,2015年澳大利亚莫纳什大学计量经济与统计学博士。其研究论文发表于 JBES,ER等国际顶级统计和主流计量经济期刊。
 
报告内容摘要:In this paper, we propose three new predictive models: the multi-step nonparametric predictive regression model (NPR) and the multi-step additive predictive regression model (APR), in which the predictive variables are locally stationary time series; and the multi-step time-varying coefficient predictive regression model (TVCPR), in which the predictive variables are stochastically nonstationary. We also establish the estimation theory and asymptotic properties for these models. To evaluate the effectiveness of these models, we investigate their capability of stock return prediction. The empirical results show that all of these models can substantially outperform the traditional linear predictive regression model in terms of both in-sample and out-of-sample performance. In addition, we find that these models can always beat the historical mean model in terms of in-sample fitting, and also for some cases in terms of the out-of-sample forecasting.
朱涣君,厦门大学王亚楠经济研究院,2016年澳大利亚莫纳什大学计量经济与统计学博士。
 
报告内容摘要:This paper studies M-estimation for large panel data model with high dimensional parameters. We investigate the influences of various dimensions on the statistical inference for M-estimation, including cross-sectional dimension (N), time series length (T) and parameter dimension (p). The rate of convergence and asymptotic distribution of the provided M-estimator are established. Moreover, for easy practice, a good estimator for the asymptotic variance in the asymptotic distribution is constructed and its consistency is discussed in theoretical form. All kinds of simulations illustrate the validity of M-estimation for heavy-tailed data and data with outliers. Empirical application on stock returns is also provided.
 
蔡必卿,华中科技大学经济学院,2012年厦门大学王亚楠经济研究院经济学博士。博士后研究员:2012-2013 Monash University Australia; 2013-2016 Bergen University, Norway. 其研究论文发表于 JBES,Econometrics等国际顶级统计和计量经济期刊。
 
报告内容摘要:This paper discusses estimating the market stock return predicting function using Hermite series. The out-of-sample evaluation results suggest   the dividend yield has nonlinear predictive power for stock returns while the book-to-market ratio and earning-price ratio have little predictive power.
 
郭萌萌 西南财大经济管理研究院,德国柏林洪堡大学 经济学博士,在AStA Advances in Statistical Analysis, Statistics and Computing and Journal of forecasting 等主流统计和计量杂志发表论文。
报告内容摘要:This paper proposes a DCC-filtered historical simulation approach to modeling systemic risk. We explain our method in terms of CoVaR
and CoES (Adrian and Brunnermeie 2016), and show its applicability to Marginal Expected Shortfall (MES) and SRISK. Empirical applications to U.S. and Europe financial markets highlight the merits of our method.
 
黄勔 西南财大经济管理研究院,西班牙马德里卡洛斯三世大学经济学博士
报告内容摘要:Multinomial discrete choice models, including binary choice models as special examples, form a class of widely used nonlinear models. Including unobservable heterogeneities in such models is necessary in many applications due to the unobservable individual preference, attributes or technologies. Unfortunately this causes the issue of point identification if the fixed effects approach is used. Following the set identification proposed by Honoré and Tamer (2006), I further develop the estimation method by Chernozhukov et al. (2013a) for both static and dynamic multinomial discrete choice models with fixed effects, whose parameters and conditional average partial effects on choice probabilities can be only set identified. Although
there has been estimation method for binary choice models of closed-form,
what I study in this paper provides extension in two ways. First, specification for multiple choices is studied for both static and dynamic models. Second, it allows components with open-forms due to multiple choices and assumption on errors other than GEV distributions. For these models, conditional probabilities for each alternatives and partial effects can be found by Monte Carlo method. Therefore I propose a simulation based estimation for all the set identified quantities and show that this estimation is consistent under some general conditions and a perturbed bootstrap method can be used to implement its inference. Numeric examples with generated data are included in order to compare the simulation based estimation with the extant method when closed-forms are available and show its behaviour for model with only open-form components. I find that the estimated bounds of partial effects work well since they are close to the results of extant estimation for
closed-form model and cover true effects in most of the cases for both models with and without closed-forms.
 
干卓泂  西南财大统计学院 蒂尔堡大学 经济学博士 研究论文发表在 Journal of Statistical Planning and Inference 和 Journal of Mathematical Economics 等国际统计和经济数学主流期刊。
 
报告内容摘要:We consider a prediction problem with a large number of predictors. The predictors are assumed to have an approximate factor structure with latent factors. The most common approach for prediction - see Bai and Ng (2006) - is to extract common factors, and use them as predictors in the second step. Such methods may not be efficient because they disregard the fact that some variables are better predictors than others, yet their variation is not well-represented by the common factors.
 
In this paper, we allow some variables to have idiosyncratic factors that are relevant for prediction. We minimize the prediction error via a new model selection criterion and show that our criterion is consistent in terms of the selection of the relevant idiosyncratic factors, as the number of predictors (N) and time periods (T) gets large. The procedure is computationally efficient, and our simulation study shows that our method delivers a lower mean-squared prediction error than Bai and Ng (2006) for a wide range of N-T pairs. We further illustrate the performance of our method for predicting inflation.
 

相关热词搜索:

上一篇:2017年成都理论经济学研讨会
下一篇:最后一页

电话:028-87092197 028-87092156| 四川成都温江柳台大道555号格致楼9楼 | 邮编:611130 | 旧版