Research

Abstract: This paper considers a first-order autoregressive panel data model with individual-specific effects and heterogeneous autoregressive coefficients defined on the interval (-1,1], thus allowing for some of the individual processes to have unit roots. It proposes estimators for the moments of the cross-sectional distribution of the autoregressive (AR) coefficients, assuming a random coefficient model for the autoregressive coefficients without imposing any restrictions on the fixed effects. It is shown the standard generalized method of moments estimators obtained under homogeneous slopes are biased. Small sample properties of the proposed estimators are investigated by Monte Carlo experiments and compared with a number of alternatives, both under homogeneous and heterogeneous slopes. It is found that a simple moment estimator of the mean of heterogeneous AR coefficients performs very well even for moderate sample sizes, but to reliably estimate the variance of AR coefficients much larger samples are required. It is also required that the true value of this variance is not too close to zero. The utility of the heterogeneous approach is illustrated in the context of earnings dynamics.

JEL classifications: C22, C23, C36

Keywords: Earnings dynamics, heterogeneous dynamic panels, neglected heterogeneity bias, short T panels 

Abstract: The commonly used two-way fixed effects estimator is biased under correlated heterogeneity and can lead to misleading inference. The mean group estimator proposed by Pesaran and Smith (1995) is robust to correlated heterogeneity but requires the underlying individual estimates to have second-order moments that could fail if the number of estimated coefficients (k) is too close to the time dimension (T) of the panel. This paper focuses on panels

where k is close to T (including k=T), and proposes a trimmed mean group (TMG) estimator that shrinks individual estimates most likely to fail the second-order moment condition. The TMG estimator is shown to be $n^{(1-\alpha )/2}$-consistent and asymptotically normally distributed, where $\alpha$ is determined by the degree to which individual estimates might not have moments. The $\sqrt{n}$ convergence rate is achieved only if all individual estimates have second-order moments. Extensions to panels with time effects are provided, and a new Hausman test of correlated heterogeneity is proposed. Small sample properties of the TMG estimator (with and without time effects) are investigated by Monte Carlo experiments and shown to be satisfactory. The proposed test of correlated heterogeneity is also shown to have the correct size and satisfactory power. The utility of the TMG approach is illustrated with an empirical application.

JEL Classifications: C21, C23

Keywords: Correlated heterogeneity, irregular estimators, two-way fixed effects, mean group estimation, tests of

correlated heterogeneity, calorie demand

Abstract: Endogeneity is a primary concern when evaluating causal effects using observational panel data. While unit-specific intercepts absorb unobserved time-invariant confounders, dependence (i) between regressors and the current error term (regressor endogeneity) and/or (ii) between regressors and heterogeneous slopes (slope endogeneity) can introduce significant endogeneity bias. This paper proposes a two-stage, instrument-free copula-based endogeneity correction mean group estimator for panel models, simultaneously addressing both endogeneity concerns. We develop this control function estimation approach by introducing a semiparametric general-location Gaussian copula that effectively captures the panel structure. The heterogeneous slope coefficients are treated as unit-specific parameters without distributional assumptions, allowing for arbitrary dependence structure between slope coefficients and regressors. We further extend the estimator to dynamic panels, where intertemporal dependence in the outcome process can be suitably captured. We derive the estimator's asymptotic properties and an analytical variance formula for inference without bootstrapping. We demonstrate its usage by simulations and a marketing mix response application across 21 product categories accounting for both sources of endogeneity in store-sales panel data.

Keywords: Control function, correlated random coefficients, Gaussian copula, heterogeneity, panel data, regressor endogeneity, slope endogeneity

Abstract: This paper investigates how information presentation shapes the monetization of Intellectual Property (IP) through auctions by addressing three key questions: (1) Does improved information presentation enhance IP monetization? (2) What types of IP benefit more from enhanced information presentation? (3) Do textual descriptions featuring IP provide predictive signals for auction outcomes? To answer these questions, we analyze transaction data of 1,337 IPs sold through an auction house, leveraging a natural field experiment with variations in information presentation. Moreover, recognizing the domain-specific language and high commercialization uncertainty of IP assets, we employ topic models to identify key pro-words and passive words in auction catalogues that drive differences in IP monetization. Our findings offer actionable insights for both IP owners and auction managers on how to structure information to improve the marketability and perceived value of intangible assets.

Keywords: Information Presentation, intellectual properties, IP auctions, natural field experiment, valuation, text analysis, topic models

Abstract: This paper focuses on estimation and inference of the average effects in heterogeneous dynamic panel data models with weakly exogenous regressors when the number of cross-sectional units (n) is large and the number of time periods (T) is moderately short. We consider bias correction of mean group (MG) estimators by the split-panel Jackknife (JK). It is shown that the MG-JK estimator is root-n consistent as both n and T tend to infinity and n/T^{4} converges to zero once the first-order bias is eliminated. For the validity of the Jackknife in finite samples, a sufficient condition for the r-th moment existence of the MG estimator is derived for an ARX(1) model. Moreover, the paper revisits the long dispute over the disemployment effects of minimum wages. The MG-JK estimator addresses short-term dynamic heterogeneity in the outcome processes and treatment effects of continuous variables. In contrast to the two-way fixed effects estimator, it does not rely on the parallel trend assumption or a staggered treatment design. Using the MG-JK estimators and county-level data in the United States during 2002–2011, it is found that there is a close to zero effect of minimum wages on total employment but a substantial negative impact on teenage employment in the long run. 

JEL classifications: C13, C22, C23, J38, J88

Keywords: Dynamic panels, individual heterogeneity, incidental parameter problem, mean group, bias reduction, split-panel Jackknife, minimum wage policy