Nickell Bias in Panel Local Projection

1. Empirical results -> applications.zip: The dataset in application comprises four empirical studies on macro-finance linkages, all focusing on the impact of financial crises on output, each emphasizing a different financial shock.

* Romer and Romer (2017) provide a semi-annual dataset consisting of an unbalanced panel for 24 OECD countries from 1967 to 2012, include the dependent variables log real GDP (LNGDP) and unemployment rate (UNEMP), along with the new regressor (CRISIS) and the corresponding lags and horizon h.

* Baron et al. (2021) cover bank equity returns for 46 advanced and emerging economies from 1870 to 2016, with the dependent variable being the log of real GDP and the change in credit-to-GDP, regressors including bank equity crash (R_B) and nonfinancial equity crash (R_N), and their respective lags and horizon h.

* Mian et al. (2017) assemble an unbalanced panel of 30 countries from 1960 to 2012, with the dependent variable log real GDP, regressors including household debt to GDP ratio (L0HHD_L1GDP) and nonfinancial firm debt to GDP ratio (L0NFD_L1GDP), and their respective lags and horizon h.

* Cerra and Saxena (2008) present panel data on currency crises for 175 countries from 1965 to 2000, including the dependent variable the change in the logarithm of real GDP (GRRT_WB), a dummy variable indicating a currency crisis (CRISIS), and the corresponding lags and horizon h.

2. Main-text simulations -> simulations.zip: Monte Carlo simulations for the main text (Figures 1-3). See simulations/README.md for details about the scripts and figures.

3. Appendix simulations -> simulations_appendix.zip: Simulation code for Appendix C (Figures C1-C12). The workflow is explained in simulations_appendix/readme.md for instructions.