Replication Data for: Replication Data for: Labor-Profit Dynamics and Structural Rents in Sector I: A Unified Gamma Mixture MLE Analysis (2005–2023)

# Labor-Profit Dynamics via Unified Gamma Mixture MLE (2005–2023) This repository contains the Python source code and computational framework developed to analyze the distributional characteristics and structural dependencies of the Industrial and Technology sectors. ## Overview The core of this project is a **Unified Gamma Mixture Maximum Likelihood Estimation (MLE)** framework. It is designed to: 1. Model the empirical distributions of EBIT, R&D, Fixed Capital (PP&E), and Personnel using Gamma mixtures. 2. Independently estimate cross-moments ($E[XY]$) to quantify structural dependencies (covariances) without assuming linear correlations or specific copula forms. 3. Validate model fit using Kolmogorov-Smirnov (KS) tests and Information Criteria (AIC/BIC). ## Repository Structure - `/data`: Contains the sectoral datasets (CSV format) for the 2005–2023 period. - `/scripts`: - `main_analysis.py`: The primary script to run the global MLE analysis. - `gamma_mixtures.py`: Core functions for the Expectation-Maximization (EM) algorithm and Gamma PDF fitting. - `plotting_tool.py`: Script used to generate the Black & White histograms (Figures 3 and 4 in the manuscript). - `requirements.txt`: List of necessary Python libraries. ## Requirements To run the scripts, you will need Python 3.8+ and the following libraries: - `numpy` - `scipy` - `pandas` - `matplotlib` You can install the dependencies using: ```bash pip install -r requirements.txt# 基于统一伽马混合极大似然估计的劳动与收益动态（2005—2023年） 本仓库包含用于分析工业与科技行业分布特征及结构依赖关系的Python源代码与计算框架。 ## 项目概述 本项目的核心为**统一伽马混合极大似然估计（Unified Gamma Mixture Maximum Likelihood Estimation, MLE）**框架。其设计目标如下： 1. 采用伽马混合模型对息税前利润（Earnings Before Interest and Taxes, EBIT）、研发（Research and Development, R&D）、固定资产（物业、厂房及设备，Property, Plant and Equipment, PP&E）以及从业人员数量的经验分布进行建模。 2. 无需假设线性相关关系或特定Copula函数形式，即可独立估计互联合矩（$E[XY]$）以量化结构依赖关系（协方差）。 3. 采用柯尔莫哥洛夫-斯米尔诺夫检验（Kolmogorov-Smirnov, KS）与信息准则（赤池信息准则AIC/贝叶斯信息准则BIC）对模型拟合效果进行验证。 ## 仓库结构 - `/data`：存放2005—2023年行业细分数据集（CSV格式）。 - `/scripts`： - `main_analysis.py`：用于执行全局极大似然估计分析的主脚本。 - `gamma_mixtures.py`：包含期望最大化（Expectation-Maximization, EM）算法与伽马概率密度函数（Probability Density Function, PDF）拟合核心函数的脚本。 - `plotting_tool.py`：用于生成黑白直方图（论文中图3与图4）的脚本。 - `requirements.txt`：所需Python库的依赖清单。 ## 运行依赖 若要运行本仓库脚本，需配置Python 3.8及以上版本环境，并安装以下Python库： - `numpy`、`scipy`、`pandas` 与 `matplotlib` 可通过以下命令安装所需依赖： bash pip install -r requirements.txt