Concepts,Theory, and Techniques ROBUST METHODS: AN ALTERNATIVE APPROACH FOR ANALYZING DATA SETS CONTAINING INFLUENTIAL DATA POINTS

Recent advances in statistical estimation theory have resulted in the development of new procedures, called robust methods, that can be used to estimate the coefficients of a regression model. Because such methods take into account the impact of discrepant data points during the initial estimation process, they offer a number of advantages over ordinary least squares and other analytical procedures (such as the analysis of outliers or regression diagnostics). This paper describes the robust method of analysis and illustrates its potential usefulness by applying the technique to two data sets. The first application uses artificial data; the second uses a data set analyzed previously by Tufte [15] and, more recently, by Chatterjee and Wiseman [6].     

REGRESSION METHODOLOGY WITH GROSS OBSERVATION ERRORS IN THE EXPLANATORY VARIABLES***

The robustness of linear programming regression estimators is examined where the disturbance terms are normally distributed and there are observation errors in the explanatory variables. These errors are occasional gross biases between one set of observations and another. The simulation of short series data offers preliminary evidence that when these biases have a non-zero mean, MSAE estimation is more robust than least squares.     

L1 ESTIMATION IN SMALL SAMPLES WITH LAPLACE ERROR DISTRIBUTIONS

In this paper a Monte Carlo sampling study consisting of four experiments is described. Two error distributions were employed, the normal and the Laplace; and two small sample sizes (20 and 40) were tested. The question of simultaneous-equation bias called for two-stage estimators. The L¹, norm was employed as a means of comparing the performance of the L¹ or least squares estimators. A relatively new algorithm for computing the direct least absolute (DLA) and two-stage least absolute (TSLA) estimators was employed for the experiments. The results confirmed the hypotheses that for non-normal error distributions such as the Laplace the least absolute estimators were better.     

Optimization techniques for robust multivariate location and scatter estimation

Computation of typical statistical sample estimates such as the median or least squares fit usually require the solution of an unconstrained optimization problem with a convex objective function, that can be solved efficiently by various methods. The presence of outliers in the data dictates the computation of a robust estimate, which can be defined as the optimum statistical estimate for a subset that contains at least half of the observations. The resulting problem is now a combinatorial optimization problem which is often computationally intractable. Classical statistical methods for multivariate location \(\varvec{\mu }\) and scatter matrix \(\varvec{\varSigma }\) estimation are based on the sample mean vector and covariance matrix, which are very sensitive in the presence of outlier observations. We propose a new method for robust location and scatter estimation which is composed of two stages. In the first stage an unbiased multivariate \(L_{1}\)-median center for all the observations is attained by a novel procedure called the least trimmed Euclidean deviations estimator. This robust median defines a coverage set of observations which is used in the second stage to iteratively compute the set of outliers which violate the correlational structure of the data set. Extensive computational experiments indicate that the proposed method outperforms existing methods in accuracy, robustness and computational time.     

时变参数模型的最优滚动窗宽选择标准及应用

宏观经济领域中存在严重的结构突变性,模型估计量的优劣对估计样本规模是敏感的。本文针对时变参数模型,建立了滚动窗宽选择标准,通过最小化估计量的近似二次损失函数及最大化各子样本估计量间的曼哈顿距离选择窗宽大小,权衡了模型估计量的准确性和时变性两个相悖目标。蒙特卡罗模拟实验表明,本文所提出的方法在各种结构突变情形下均适用,能够应用于线性关系和非线性关系的时变参数模型中,且均具有稳健性。将该方法应用于我国金融网络的结构突变识别过程,显著改善了传统窗宽选择方法的结果。     

A Robust Approach to Risk Assessment Based on Species Sensitivity Distributions

The guidelines for setting environmental quality standards are increasingly based on probabilistic risk assessment due to a growing general awareness of the need for probabilistic procedures. One of the commonly used tools in probabilistic risk assessment is the species sensitivity distribution (SSD), which represents the proportion of species affected belonging to a biological assemblage as a function of exposure to a specific toxicant. Our focus is on the inverse use of the SSD curve with the aim of estimating the concentration, HCp, of a toxic compound that is hazardous to p% of the biological community under study. Toward this end, we propose the use of robust statistical methods in order to take into account the presence of outliers or apparent skew in the data, which may occur without any ecological basis. A robust approach exploits the full neighborhood of a parametric model, enabling the analyst to account for the typical real‐world deviations from ideal models. We examine two classic HCp estimation approaches and consider robust versions of these estimators. In addition, we also use data transformations in conjunction with robust estimation methods in case of heteroscedasticity. Different scenarios using real data sets as well as simulated data are presented in order to illustrate and compare the proposed approaches. These scenarios illustrate that the use of robust estimation methods enhances HCp estimation.     

Estimation and Inference With Weak,Semi‐Strong,and Strong Identification

This paper analyzes the properties of standard estimators, tests, and confidence sets (CS's) for parameters that are unidentified or weakly identified in some parts of the parameter space. The paper also introduces methods to make the tests and CS's robust to such identification problems. The results apply to a class of extremum estimators and corresponding tests and CS's that are based on criterion functions that satisfy certain asymptotic stochastic quadratic expansions and that depend on the parameter that determines the strength of identification. This covers a class of models estimated using maximum likelihood (ML), least squares (LS), quantile, generalized method of moments, generalized empirical likelihood, minimum distance, and semi‐parametric estimators. The consistency/lack‐of‐consistency and asymptotic distributions of the estimators are established under a full range of drifting sequences of true distributions. The asymptotic sizes (in a uniform sense) of standard and identification‐robust tests and CS's are established. The results are applied to the ARMA(1, 1) time series model estimated by ML and to the nonlinear regression model estimated by LS. In companion papers, the results are applied to a number of other models.     

Robust Nonparametric Confidence Intervals for Regression‐Discontinuity Designs

In the regression‐discontinuity (RD) design, units are assigned to treatment based on whether their value of an observed covariate exceeds a known cutoff. In this design, local polynomial estimators are now routinely employed to construct confidence intervals for treatment effects. The performance of these confidence intervals in applications, however, may be seriously hampered by their sensitivity to the specific bandwidth employed. Available bandwidth selectors typically yield a “large” bandwidth, leading to data‐driven confidence intervals that may be biased, with empirical coverage well below their nominal target. We propose new theory‐based, more robust confidence interval estimators for average treatment effects at the cutoff in sharp RD, sharp kink RD, fuzzy RD, and fuzzy kink RD designs. Our proposed confidence intervals are constructed using a bias‐corrected RD estimator together with a novel standard error estimator. For practical implementation, we discuss mean squared error optimal bandwidths, which are by construction not valid for conventional confidence intervals but are valid with our robust approach, and consistent standard error estimators based on our new variance formulas. In a special case of practical interest, our procedure amounts to running a quadratic instead of a linear local regression. More generally, our results give a formal justification to simple inference procedures based on increasing the order of the local polynomial estimator employed. We find in a simulation study that our confidence intervals exhibit close‐to‐correct empirical coverage and good empirical interval length on average, remarkably improving upon the alternatives available in the literature. All results are readily available in R and STATA using our companion software packages described in Calonico, Cattaneo, and Titiunik (2014d, 2014b).     

LEAST SQUARES VERSUS MINIMUM ABSOLUTE DEVIATIONS ESTIMATION IN LINEAR MODELS

Previous research has indicated that minimum absolute deviations (MAD) estimators tend to be more efficient than ordinary least squares (OLS) estimators in the presence of large disturbances. Via Monte Carlo sampling this study investigates cases in which disturbances are normally distributed with constant variance except for one or more outliers whose disturbances are taken from a normal distribution with a much larger variance. It is found that MAD estimation retains its advantage over OLS through a wide range of conditions, including variations in outlier variance, number of regressors, number of observations, design matrix configuration, and number of outliers. When no outliers are present, the efficiency of MAD estimators relative to OLS exhibits remarkably slight variation.     

A Simple Minimum-Bias Percentile Estimator of the Location Parameter for the Gamma,Weibull, and Log-Normal Distributions

Estimating the unknown minimum (location) of a random variable has received some attention in the statistical literature, but not enough in the area of decision sciences. This is surprising, given that such estimation needs exist often in simulation and global optimization. This study explores the characteristics of two previously used simple percentile estimators of location. The study also identifies a new percentile estimator of the location parameter for the gamma, Weibull, and log-normal distributions with a smaller bias than the other two estimators. The performance of the new estimator, the minimum-bias percentile (MBP) estimator, and the other two percentile estimators are compared using Monte-Carlo simulation. The results indicate that, of the three estimators, the MBP estimator developed in this study provides, in most cases, the estimate with the lowest bias and smallest mean square error of the location for populations drawn from log-normal and gamma or Weibull (but not exponential) distributions. A decision diagram is provided for location estimator selection, based on the value of the coefficient of variation, when the statistical distribution is known or unknown.     

Estimation and Inference in Large Heterogeneous Panels with a Multifactor Error Structure

This paper presents a new approach to estimation and inference in panel data models with a general multifactor error structure. The unobserved factors and the individual‐specific errors are allowed to follow arbitrary stationary processes, and the number of unobserved factors need not be estimated. The basic idea is to filter the individual‐specific regressors by means of cross‐section averages such that asymptotically as the cross‐section dimension (N) tends to infinity, the differential effects of unobserved common factors are eliminated. The estimation procedure has the advantage that it can be computed by least squares applied to auxiliary regressions where the observed regressors are augmented with cross‐sectional averages of the dependent variable and the individual‐specific regressors. A number of estimators (referred to as common correlated effects (CCE) estimators) are proposed and their asymptotic distributions are derived. The small sample properties of mean group and pooled CCE estimators are investigated by Monte Carlo experiments, showing that the CCE estimators have satisfactory small sample properties even under a substantial degree of heterogeneity and dynamics, and for relatively small values of N and T.     

Estimation of Nonparametric Models With Simultaneity

We introduce methods for estimating nonparametric, nonadditive models with simultaneity. The methods are developed by directly connecting the elements of the structural system to be estimated with features of the density of the observable variables, such as ratios of derivatives or averages of products of derivatives of this density. The estimators are therefore easily computed functionals of a nonparametric estimator of the density of the observable variables. We consider in detail a model where to each structural equation there corresponds an exclusive regressor and a model with one equation of interest and one instrument that is included in a second equation. For both models, we provide new characterizations of observational equivalence on a set, in terms of the density of the observable variables and derivatives of the structural functions. Based on those characterizations, we develop two estimation methods. In the first method, the estimators of the structural derivatives are calculated by a simple matrix inversion and matrix multiplication, analogous to a standard least squares estimator, but with the elements of the matrices being averages of products of derivatives of nonparametric density estimators. In the second method, the estimators of the structural derivatives are calculated in two steps. In a first step, values of the instrument are found at which the density of the observable variables satisfies some properties. In the second step, the estimators are calculated directly from the values of derivatives of the density of the observable variables evaluated at the found values of the instrument. We show that both pointwise estimators are consistent and asymptotically normal.     

基于最小一乘准则的上证指数突变点研究

很多研究表明,上证指数序列既有结构突变的特征,也有厚尾的特征。但大部分现有的研究都没有考虑其厚尾特征对变点估计的影响。本文基于最小一乘准则提出了一个估计厚尾数据中变点的方法。模拟研究表明,当数据具有厚尾特征时,基于最小一乘准则的变点估计比基于最小二乘准则的估计有效。对上证指数的实证结果表明,基于最小一乘准则估计出的变点能更好地描述中国股票市场的结构突变特征。     

加权复合分位数回归方法在动态VaR风险度量中的应用

风险价值(VaR)因为简单直观,成为了当今国际上最主流的风险度量方法之一,而基于时间序列自回归(AR)模型来计算无条件风险度量值在实业界有广泛应用。本文基于分位数回归理论对AR模型提出了一个估计方法--加权复合分位数回归(WCQR)估计,该方法可以充分利用多个分位数信息提高参数估计的效率,并且对于不同的分位数回归赋予不同的权重,使得估计更加有效,文中给出了该估计的渐近正态性质。有限样本的数值模拟表明,当残差服从非正态分布时,WCQR估计的的统计性质接近于极大似然估计,而该估计是不需要知道残差分布的,因此,所提出的WCQR估计更加具有竞争力。此方法在预测资产收益的VaR动态风险时有较好的应用,我们将所提出的理论分析了我国九只封闭式基金,实证分析发现,结合WCQR方法求得的VaR风险与用非参数方法求得的VaR风险非常接近,而结合WCQR方法可以计算动态的VaR风险值和预测资产收益的VaR风险值。     

Constructing Optimal Instruments by First‐Stage Prediction Averaging

This paper considers model averaging as a way to construct optimal instruments for the two‐stage least squares (2SLS), limited information maximum likelihood (LIML), and Fuller estimators in the presence of many instruments. We propose averaging across least squares predictions of the endogenous variables obtained from many different choices of instruments and then use the average predicted value of the endogenous variables in the estimation stage. The weights for averaging are chosen to minimize the asymptotic mean squared error of the model averaging version of the 2SLS, LIML, or Fuller estimator. This can be done by solving a standard quadratic programming problem.     

THE IMPACT OF EXTREME OBSERVATIONS ON SIMPLE FORECASTING METHODS*

In general linear modeling, an alternative to the method of least squares (LS) is the least absolute deviations (LAD) procedure. Although LS is more widely used, the LAD approach yields better estimates in the presence of outliers. In this paper, we examine the performance of LAD estimators for the parameters of the first-order autoregressive model in the presence of outliers. A simulation study compared these estimates with those given by LS. The general conclusion is that LAD does not deal successfully with additive outliers. A simple procedure is proposed which allows exception reporting when outliers occur.     

多因素数据重心预测方法及应用研究

本文基于数据重心概念,通过大量的研究和推导,提出数据重心参数估计理论,并利用数据重心法估计多项式回归预测模型的参数,应用于我国钢材消费量的预测。从应用的结果看,增加了我国目前钢材预测的方法。本方法能最大限度地平滑预测模型的误差,而且应用条件比最小二乘法宽松;也不会因为个别残差较大的异常点而对预测结果产生不稳定性,从而提高了拟合和预测的稳健度,计算更简捷。     

Consistent Estimation with a Large Number of Weak Instruments

This paper analyzes the conditions under which consistent estimation can be achieved in instrumental variables (IV) regression when the available instruments are weak and the number of instruments, K_n, goes to infinity with the sample size. We show that consistent estimation depends importantly on the strength of the instruments as measured by r_n, the rate of growth of the so‐called concentration parameter, and also on K_n. In particular, when K_n→∞, the concentration parameter can grow, even if each individual instrument is only weakly correlated with the endogenous explanatory variables, and consistency of certain estimators can be established under weaker conditions than have previously been assumed in the literature. Hence, the use of many weak instruments may actually improve the performance of certain point estimators. More specifically, we find that the limited information maximum likelihood (LIML) estimator and the bias‐corrected two‐stage least squares (B2SLS) estimator are consistent when , while the two‐stage least squares (2SLS) estimator is consistent only if K_n/r_n→0 as n→∞. These consistency results suggest that LIML and B2SLS are more robust to instrument weakness than 2SLS.     

Efficient Semiparametric Estimation via Moment Restrictions

Conditional moment restrictions can be combined through GMM estimation to construct more efficient semiparametric estimators. This paper is about attainable efficiency for such estimators. We define and use a moment tangent set, the directions of departure from the truth allowed by the moments, to characterize when the semiparametric efficiency bound can be attained. The efficiency condition is that the moment tangent set equals the model tangent set. We apply these results to transformed, censored, and truncated regression models, e.g., finding that the conditional moment restrictions from Powell's (1986) censored regression quantile estimators can be combined to approximate efficiency when the disturbance is independent of regressors.     

Estimation of Nonlinear Models with Measurement Error

This paper presents a solution to an important econometric problem, namely the root n consistent estimation of nonlinear models with measurement errors in the explanatory variables, when one repeated observation of each mismeasured regressor is available. While a root n consistent estimator has been derived for polynomial specifications (see Hausman, Ichimura, Newey, and Powell (1991)), such an estimator for general nonlinear specifications has so far not been available. Using the additional information provided by the repeated observation, the suggested estimator separates the measurement error from the “true” value of the regressors thanks to a useful property of the Fourier transform: The Fourier transform converts the integral equations that relate the distribution of the unobserved “true” variables to the observed variables measured with error into algebraic equations. The solution to these equations yields enough information to identify arbitrary moments of the “true,” unobserved variables. The value of these moments can then be used to construct any estimator that can be written in terms of moments, including traditional linear and nonlinear least squares estimators, or general extremum estimators. The proposed estimator is shown to admit a representation in terms of an influence function, thus establishing its root n consistency and asymptotic normality. Monte Carlo evidence and an application to Engel curve estimation illustrate the usefulness of this new approach.