XTHETEROQUANT: Stata module to compute quantiles of heterogeneous individual-specific coefficients in panel data, with SQB and CDQB bootstrap inference (Galvao, Hounyo and Lin, 2026)

xtheteroquant implements the two-step estimator of Galvao, Hounyo and Lin (2026) for the tau-quantile of the cross-sectional distribution of heterogeneous individual-specific coefficients in the linear panel-data model y_it = a_i + x_it'b_i + e_it, i = 1,...,N, t = 1,...,T, where every unit i has its own intercept a_i and its own slope vector b_i. The procedure is: Step 1 (individual estimation). For each unit i, estimate the coefficient vector by OLS using only that unit's time series. Step 2 (quantile aggregation). For each coefficient, take the empirical tau-quantile of the N first-step estimates, theta_hat(tau) = argmin_theta (1/N) sum_i rho_tau(theta_hat_i - theta), where rho_tau is the quantile-regression check function. Unlike conventional fixed-effects quantile regression (qreg / panel QR), where tau indexes the conditional distribution of the outcome, here tau indexes the distribution of the structural coefficients across units: tau = .9 answers "what is the slope for the 90th-percentile most-responsive unit?", not "what is the effect at the 90th percentile of y?". Typical uses: heterogeneous treatment or policy sensitivities, the distribution of mutual-fund timing skill, firm-specific pass-through, country-specific elasticities.