Replication Data for: Global Financial Cycle and the Chinese Economy: Asset Price Transmission and Stage-Dependent Monetary and Macroprudential Policies

Description
This repository contains the replication data and code for the paper regarding Global Financial Cycle and the Chinese Economy: Asset Price Transmission and Stage-Dependent Monetary and Macroprudential Policies. The package includes the necessary datasets and scripts to reproduce all empirical results.
The replication package is organized into two main software environments: Stata (for descriptive statistics and unit root tests) and MATLAB (for SVAR, Sign-Restricted SVAR, and TVP-SV-VAR modeling).
1. Software Requirements
Stata: Version 19 or later (Standard StataMP is sufficient; no specialized community packages required).
MATLAB: R2022b or later.
Toolbox Required: Statistics and Machine Learning Toolbox.
External Package: The TVP-VAR package (Nakajima, 2011) must be added to the MATLAB path for the time-varying parameter analysis.
2. Repository Structure and Functionality
The codebase is divided into four distinct modules corresponding to different sections of the paper:
A. Preliminary Analysis (Stata)
Script: replication_stata_appendix.do
Data: data_appendix.dta
Function: Generates Table 1 (Descriptive Statistics) and Appendix Tables A1–A4 and Figures A1–A2 (ADF unit root tests, lag-order selection, and inverse roots checks).
B. Baseline SVAR and Counterfactual Analysis (MATLAB)
Master Script: SVAR_main_reproduce.m
Key Data: China_Macro_Data.xlsx, data_structral_creditspread.xlsx, China_Macro_Data_robust.xlsx, data_quarter.xlsx.
Function: Implements the recursive identification SVAR. By adjusting the Run_Case parameter (1–4) within the script, users can reproduce:
Figure 2 & 3: Baseline IRFs and Counterfactual Analysis (Asset-price channels).
Figure 4 & 5: Subsample analysis (Pre/Post 2015) and SOE-Private Credit Spread extension.
Figure A13 & A14: Robustness checks involving Consumption/Investment and quarterly data.
C. Time-Varying Parameter (TVP-SV-VAR) Analysis (MATLAB)
Master Script: main_reproduce.m
Key Data: data_gfc_mp_rr_2.xlsx, data_gfc_mp_ltv_2.xlsx, data_gfc_mp_ccb_2.xlsx (and robust variants).
Function: Estimates the TVP-VAR model with stochastic volatility. By adjusting the Run_Case parameter (1–7), users can reproduce:
Figures 6–10: Time-varying impulse responses for Quantity vs. Price rules and different macroprudential tools (RR, LTV, CCB).
Appendix Figures: All impulse response robustness checks (VIX substitution, Ordering changes) and MCMC diagnostics (Table A5, Figure A3).
D. Sign-Restricted SVAR (MATLAB)
Script: SignRes_SVAR_reproduce.m
Data: China_Macro_Data.xlsx
Function: Reproduces Figure A15.
3. Data Sources
The datasets provided are processed and analysis-ready. Primary raw data sources include the CEIC Database, CSMAR, and the National Bureau of Statistics of China.
4. Usage Instructions
Detailed step-by-step instructions for running each module are provided in the specific README.txt files included within the downloaded package.

### 数据集说明
本仓库包含关于**全球金融周期与中国经济：资产价格传导与阶段差异化货币政策及宏观审慎政策**的论文的复现数据与代码。本数据包包含复现全部实证结果所需的数据集与脚本文件。

本复现数据包基于两大软件环境搭建：Stata（用于描述性统计与单位根检验）与MATLAB（用于结构向量自回归（SVAR）、符号约束结构向量自回归（Sign-Restricted SVAR）及时变参数随机波动率向量自回归（TVP-SV-VAR）建模）。

1. 软件要求
Stata：19.0及以上版本（StataMP标准版即可满足需求，无需额外专用社区插件）。
MATLAB：R2022b及以上版本，需配备统计学与机器学习工具箱（Statistics and Machine Learning Toolbox）。此外，需将TVP-VAR工具包（Nakajima, 2011）添加至MATLAB路径，以完成时变参数分析。

2. 仓库结构与功能
本代码库分为四个独立模块，分别对应论文的不同章节：
A. 预分析（Stata）
脚本文件：replication_stata_appendix.do
数据文件：data_appendix.dta
功能：生成表1（描述性统计）以及附录表A1–A4、附录图A1–A2（ADF单位根检验、滞后阶数选择与逆根检验结果）。

B. 基准SVAR与反事实分析（MATLAB）
主脚本：SVAR_main_reproduce.m
核心数据集：China_Macro_Data.xlsx、data_structral_creditspread.xlsx、China_Macro_Data_robust.xlsx、data_quarter.xlsx
功能：实现递归识别结构向量自回归（SVAR）模型。通过调整脚本内的Run_Case参数（1–4），用户可复现以下结果：
- 图2与图3：基准脉冲响应函数与资产价格渠道反事实分析
- 图4与图5：分样本分析（2015年前后子样本与国企-民企信贷利差拓展分析）
- 图A13与图A14：涉及消费/投资与季度数据的稳健性检验

C. 时变参数（TVP-SV-VAR）分析（MATLAB）
主脚本：main_reproduce.m
核心数据集：data_gfc_mp_rr_2.xlsx、data_gfc_mp_ltv_2.xlsx、data_gfc_mp_ccb_2.xlsx（及稳健性检验变体）
功能：估计带随机波动率的时变参数向量自回归（TVP-SV-VAR）模型。通过调整脚本内的Run_Case参数（1–7），用户可复现以下结果：
- 图6–10：数量型与价格型货币政策规则、不同宏观审慎工具（存款准备金率RR、贷款价值比LTV、逆周期资本缓冲CCB）的时变脉冲响应
- 附录图表：全部脉冲响应稳健性检验（VIX替代、排序变更）与马尔可夫链蒙特卡洛（MCMC）诊断结果（表A5、图A3）

D. 符号约束SVAR分析（MATLAB）
脚本文件：SignRes_SVAR_reproduce.m
数据文件：China_Macro_Data.xlsx
功能：复现图A15

3. 数据来源
本仓库提供的数据集均已完成预处理，可直接用于分析。原始数据主要来源于CEIC数据库、CSMAR数据库与中华人民共和国国家统计局。

4. 使用说明
下载的数据包内附带各模块对应的README.txt文件，其中包含运行各模块的详细分步操作指南。