Replication Package for: "The Traffic Curse: Structural Distinction between Endogenous and Exogenous Traffic and Its Empirical Tests"

Description

This dataset contains the Python code and embedded data for replicating all empirical analyses in the paper "The Traffic Curse: Structural Distinction between Endogenous and Exogenous Traffic and Its Empirical Tests." The analyses include: (1) existence tests using 30 U.S. listed firms and 796 Chinese A-share firms; (2) difference-in-differences causal identification using China's 2015 Advertising Law revision as a natural experiment; (3) direct mechanism tests including bootstrap mediation analysis; (4) formal resource allocation model derivation and robustness checks (placebo test, Heckman two-stage, alternative dependent variables); (5) platform survival analysis using Cox proportional hazards model and Kaplan-Meier estimation. All data come from publicly accessible sources (SEC EDGAR, East Money, public news and industry reports).

Steps to reproduce

1.Environment: Python 3.8+ with pandas, numpy, scipy, statsmodels, matplotlib, lifelines. 1.2. Install dependencies: pip install pandas numpy scipy statsmodels matplotlib lifelines 3. Run analysis scripts in any order (each is self-contained with embedded data): python 05_tc_figures.py # Figures 1-5 (existence + platform) python 06_did_analysis.py # Figures 6-8 (DID causal identification) python 07_measurement_improvement.py # Figure 10 (measurement validation) python 08_mechanism_tests.py # Mechanism tests (bootstrap mediation) python 09_theory_robustness.py # Figure 9 (formal model + robustness) 4. Expected output: Each script prints key regression coefficients and p-values to console, and saves figures as 300-dpi PNG files. All numerical results reported in the paper can be reproduced.

Institutions

• Anhui Agricultural University

  Anhui, Hefei

Categories

Licence