arXiv subjects

Liugang Sheng

Publications and source records attributed to Liugang Sheng.

3 recordsLinked to original sources

arXiv2026

xtlp: A new Stata command for panel local projection

Panel local projection (LP) estimated with fixed effects (FEs) is widely used to evaluate the dynamic effects of economic shocks across countries or regions. However, FE estimation of panel LP models is subject to Nickell bias arising from the predictive structure of LP specifications, even when lagged dependent variables are not included in the regressors. This bias invalidates standard inference based on conventional t-statistics. We introduce xtlp, a new Stata command that implements panel LP with split-panel jackknife (SPJ) bias correction. It removes the Nickell bias and restores the validity of standard statistical inference. The command accommodates a broad range of specifications, including different transformations of the dependent variable, additional control variables, and flexible variance estimators. We illustrate xtlp through empirical applications that examine the effects of financial crises on output.

econ.EM↗

arXiv2026

Nickell Bias in Panel Local Projection: Financial Crises Are Worse Than You Think

Panel local projection (LP) with fixed-effects (FE) is widely adopted for evaluating the economic consequences of financial crises across countries. This paper highlights a fundamental methodological issue: the presence of the Nickell bias in the panel FE estimator due to inherent dynamic structures of predictive specifications, even if the regressors have no lagged dependent variables. The Nickell bias invalidates the standard inferential procedure based on the $t$-statistic. We propose a split-panel jackknife (SPJ) estimator as a simple, easy-to-implement, and yet effective solution to eliminate the bias and restore valid statistical inference. We revisit four influential empirical studies on the impact of financial crises, and find that the FE method underestimates the economic losses of financial crises relative to the SPJ estimates. Replication files are available at https://metricshilab.github.io/panel-lp-replication/, with links to R and Stata packages.

econ.EM↗

arXiv2025

AI as "Co-founder": GenAI for Entrepreneurship

This paper studies whether, how, and for whom generative artificial intelligence (GenAI) facilitates firm creation. Our identification strategy exploits the November 2022 release of ChatGPT as a global shock that lowered start-up costs and leverages variations across geo-coded grids with differential pre-existing AI-specific human capital. Using high-resolution and universal data on Chinese firm registrations by the end of 2024, we find that grids with stronger AI-specific human capital experienced a sharp surge in new firm formation$\unicode{x2013}$driven entirely by small firms, contributing to 6.0% of overall national firm entry. Large-firm entry declines, consistent with a shift toward leaner ventures. New firms are smaller in capital, shareholder number, and founding team size, especially among small firms. The effects are strongest among firms with potential AI applications, weaker financing needs, and among first-time entrepreneurs. Overall, our results highlight that GenAI serves as a pro-competitive force by disproportionately boosting small-firm entry.

econ.GN↗