Fall 2016 - Xiye Yang and Yuan Liao
Spring 2017 - Roger Klein and Norman Swanson
- All seminars will be held in the New Jersey Hall 3rd Floor Library on Thursdays at 4:00 p.m., unless otherwise noted.
- Papers can be downloaded from the Department's website when available in advance.
- Schedule of Dates and Speakers may change.
February 9 POSTPONED rescheduled for February 16
Han Xiao, Rutgers University Statistics
"Simultaneous inference of covariances: time series and high dimensional statistics"
Abstract:This talk focuses on the maximum deviation of sample covariances. Three types of problems are considered, which are also intrinsically related. First, an omnibus test of serial correlation for a single time series, based on the maximum sample autocorrelation. Second, maximum deviation of sample covariances, which can be used to test the structure of high dimensional covariance matrix. Third, tests of pairwise independence of high dimensional time series, using maximum sample cross correlations. Under mild conditions on the dependence among the data, we establish the asymptotic distributions of the test statistics. A common feature of these tests is that the number of sample covariances involved in the tests are large. In particular, under the high-dimensional setting, our results allow the dimension to grow exponentially fast with the sample size. Bootstrap methods are employed to calibrate the sizes of the tests for finite samples. Some variants of the tests, as well extensions to nonstationary time series are also considered.
Marc Henry, Penn State University
"Single market nonparametric identification of multi attribute hedonic equilibrim models"
Abstract: This paper derives conditions under which preferences and technology are nonparametrically identified in hedonic equilibrium models, where products are differentiated along more than one dimension and agents are characterized by several dimensions of unobserved heterogeneity. With products differentiated along a quality index and agents characterized by scalar unobserved heterogeneity, single crossing conditions on preferences and technology provide identifying restrictions. We develop similar shape restrictions in the multi-attribute case. These shape restrictions, which are based on optimal transport theory and generalized convexity, allow us to identify preferences for goods differentiated along multiple dimensions, from the observation of a single market. We thereby extend identification results in Matzkin (2003) and Heckman, Matzkin, and Nesheim (2010) to accommodate multiple dimensions of unobserved heterogeneity. One of our results is a proof of absolute continuity of the distribution of endogenously traded qualities, which is of independent interest.
Nese Yildiz, University of Rochester
"Identification and Estimation of a Triangular Model with Multiple Endogenous Variables and Insufficiently Many Instrumental Variables"
Abstract: We develop a novel identification method for a partially linear model with multiple endogenous variables of interest but a single instrumental variable, which could even be binary. We present an easy-to-implement consistent estimator for the parametric part. This estimator retains √ n-convergence rate and asymptotic normality even though we have a generated regressor in our setup. The nonparametric part of the model is also identified. We also outline how our identification strategy can be extended to a fully non-parametric model. We use our methods to assess the impact of smoking during pregnancy on birth weight.
Konrad Menzel, New York University
"STRATEGIC NETWORK FORMATION WITH MANY AGENTS"
Abstract: We consider a random utility model of strategic network formation, where we derive a tractable approximation to the distribution of network links using many-player asymptotics. Our framework assumes that agents have heterogeneous tastes over links, and allows for anonymous and non-anonymous interaction effects among links. The observed network is assumed to be pairwise stable, and we impose no restrictions regarding selection among multiple stable outcomes. Our main results concern convergence of the link frequency distribution from finite pairwise stable networks to the (many-player) limiting distribution. The set of possible limiting distributions is shown to have a fairly simple form and is characterized through aggregate equilibrium conditions, which may permit multiple solutions. We analyze identification of link preferences and propose a method for estimation of preference parameters. We also derive an analytical expression for agents’ welfare (expected surplus) from the structure of the network.
October 5, 2017
Hiroaki Kaido, Boston University
Yacine Ait-Sahalia, Princeton University
A Hausman Test for the Presence of Market Microstructure Noise in High Frequency Data
Abstract: We develop tests that help assess whether a high frequency data sample can be treated as reasonably free of market microstructure noise at a given sampling frequency for the purpose of implementing high frequency volatility and other estimators. The tests are based on the Hausman principle of comparing two estimators, one that is efficient but not robust to the deviation being tested, and one that is robust but not as efficient. We investigate the asymptotic properties of the test statistic in a general nonparametric setting, and compare it with several alternatives that are also developed in the paper. Empirically, we find that improvements in stock market liquidity over the past decade have increased the frequency at which simple, uncorrected, volatility estimators can be safely employed.
Xi Zhang, Rutgers University (Ph.D. Graduate Student)
"Financial Stress Prediction: A Bayesian Approach"
Abstract: The objective of this paper is to assess which variables are more informative in predicting the financial stress up to one year interval under the Bayesian variable selection framework. The financial stress indicator is an ordinal variable constructed by taking into account of financial stress indices published by Federal Reserve Banks and periods identified by Serena Ng (2015) and Nicholas Bloom (2009). 16 variables belonging to three categories: interest rate, yield spread and market volatility are studied. Probit and multinomial Probit model are used. Under the Bayesian variable selection framework it shows that the first difference of interest rates and yield spreads rather than the market volatility variables are more informative in predicting the financial stress. Since the market volatility variables refer to the magnitude of the change of the financial market variables, the results also suggest that the volatility level matters in predicting the financial stress.
Elizabeth Tipton, Columbia University
"Small sample methods for cluster-robust variance estimation and hypothesis testing in fixed effects models"
Abstract: Data analysts commonly ‘cluster’ their standard errors to account for correlations arising from the sampling of aggregate units (e.g., states), each containing multiple observations. When the number of clusters is small to moderate, however, this approach can lead to biased standard errors and hypothesis tests with inflated Type I error. One solution that is receiving increased attention is the use of the bias-reduced linearization (BRL). In this paper, we extend the BRL approach to include an F-test that can be implemented in a wide range of applications. A simulation study reveals that that this test has Type I error close to nominal even with a very small number of clusters, and importantly, that it outperforms the usual estimator even when the number of clusters is moderate (e.g., 50 – 100).
Zhutong Gu, Rutgers University (Ph.D. Graduate Student)
"Testing Additive Separability with Excess Unobserved Heterogeneity: An Investigation of Hicksian-neutral Productivity in U.S. Manufacturing Industry 1990-2001"
Abstract: Additive separability between observables and unobservables is one of the essential properties in structural modeling of heterogeneity in the presence of endogeneity. In this paper, we propose a simple test based on quantile average differences of the average structural functions (ASF) generated by nonparametric nonseparable and separable models with unrestricted heterogeneity. Given identification, we establish conditions under which structural additivity is equivalent to the equality of ASFs derived from the two competing specifications commonly employed. We estimate the reduced form regressions by Nadaraya-Watson (NW) estimators and control the asymptotic bias by an iterative procedure proposed in Klein and Shen (2016). We show that the asymptotic test statistic follows a central chi-sq distribution under the null hypothesis and has power against a sequence of root-N local alternatives. Our proposed test statistic works reasonably well in a series of finite sample simulations with analytic variances, alleviating the computational burden often involved in bootstrapped inferences. We also show that the test can be straightforwardly extended to semiparametric models, panel data and triangular simultaneous equations frameworks. In the empirical application, we consider the context of production function estimation using firm-level data of U.S. manufacturing industry from 1990 to 2011. To control for the simultaneity bias, we generalize the proxy function approach to nonseparable models and then test for Hicksian-neutral technology in the presence of of multi-dimensional productivity shocks.
Yixiao (Ethan) Jiang, Rutgers University (Ph.D. Graduate Student)
"Did Shareholder Information Play a Role in Bond Ratings?"
Abstratc: This paper studies the role of common shareholders as efficient information intermediaries between the credit rating agencies (CRAs) and the bond issuers. The model allows the ratings to be determined not only by a issuer's spreadsheet variables, but also various soft information that the CRA might receive from its shareholders. The amount of soft information is aggregated by a index which is constructed from the CRA and its shareholders' shareholding data. Using the Mergent's Fixed Income Securities Database(FISD) from 2001-2008, I estimate and examine the pattern of quantile marginal effects (QME) and draw two conclusions: (1) the CRAs appear to be risk-adverse when the amount of soft information is small, but the information effect grows as information accumulates. (2) The impact of shareholder information is generally positive for investment-grade bonds but negative for high-yield bonds, which implies that the shareholder information has value.
Viktor Todorov, Kellogg School of Management, Northwestern University
"Unified Inference for Nonlinear Factor Models from Panels with Fixed and Large Time Span"
Abstract: We provide unifying inference theory for parametric nonlinear factor models based on a panel of noisy observations. The panel has a large cross-section and a time span that may be either small or large. Moreover, we incorporate an additional source of information provided by noisy observations on some known functions of the factor realizations. The estimation is carried out via penalized least squares, i.e., by minimizing the L_2 distance between observations from the panel and their model-implied counterparts, augmented by a penalty for the deviation of the extracted factors from the noisy signals for them. When the time dimension is fixed, the limit distribution of the parameter vector is mixed Gaussian with conditional variance depending on the path of the factor realizations. On the other hand, when the time span is large, the convergence rate is faster and the limit distribution is Gaussian with a constant variance. In this case, however, we incur an incidental parameter problem since, at each point in time, we need to recover the concurrent factor realizations. This leads to an asymptotic bias that is absent in the setting with a fixed time span. In either scenario, the limit distribution of the estimates for the factor realizations is mixed Gaussian, but is related to the limiting distribution of the parameter vector only in the scenario with a fixed time horizon. Although the limit behavior is very different for the small versus large time span, we develop a feasible inference theory that applies, without modification, in either case. Hence, the user need not take a stand on the relative size of the time dimension of the panel. Similarly, we propose a time-varying data-driven weighting of the penalty in the objective function, which enhances efficiency by adapting to the relative quality of the signal for the factor realizations.