Replication Code and Data for Kogan, L., Papanikolaou, D., Seru, A. and Stoffman, N., QJE 2017

原仓库地址：https://github.com/KPSS2017/Technological-Innovation-Resource-Allocation-and-Growth-Replication-Kit

🍎 Note: 数据文件需要到原仓库逐个下载。批量下载整个仓库的方式无法正确下载数据文件。(Update: 2021/7/19 12:38)

This package provides the replication code for the main results in Kogan, L., Papanikolaou, D., Seru, A. and Stoffman, N., 2017. Technological innovation, resource allocation, and growth. Quarterly Journal of Economics, 132(2), pp. 665-712. The paper is available at https://academic.oup.com/qje/article/132/2/665/3076284.

The folder ./Code contains all programs, while the folder ./Data includes all needed input data files.

Code

The folder ./Code includes code files of two types of programs: Stata and Matlab.

Code files using Stata:

Prerequisite:

In order for the code to run correctly, you need the following packages installed in Stata:

• winsorizeJ: Winsorize variables using breakpoints that vary by year. (A winsorizeJ.ado file is provided in the folder ./Code)

• cluster2: 2D clustered standard errors. (A cluster2.ado file is provided in the folder ./Code)

• ivreg2: IV regression with more features than STATA's ivreg. (Install the package using command "ssc install ivreg2")

• reghdfe: Regression with multiple levels of fixed effects. (Install the package using command "ssc install reghdfe")

• estout: Output Stata tables in tex format. (Install the package using command "ssc install estout")

File description:

• FilterReturnsCreateFirmMeasures.do: Creates the firm-level innovation measure using patent and stock return data

• PatentValueCites.do: Descriptive statistics for the patent-level measure, including robustness for alternative distributions (Table 1 in the paper and Table A.6 in the Online Appendix); relates the estimated patent values to forward citations (Table 2 in the paper and Table A.7 in the Online Appendix)

• CreateFirmSample.do: Creates the data for the firm-level regressions

• FirmSummaryStats.do: Descriptive statistics for the firm-level measure (Table 3 in the paper)

• FirmProfitsRegressionSM.do: Firm Profits and innovation (Panel a of Table 4 in the paper)

  FirmOutputRegressionSM.do: Firm Output and innovation (Panel b of Table 4 in the paper)

  FirmReallocationRegressionSM.do: Firm Capital/Labor and innovation (Panels c and d of Table 4 in the paper )

  FirmTFPRegressionSM.do: Firm TFP and innovation (Panel e of Table 4 in the paper)

• FirmProfitsRegressionCW.do: Firm Profits and innovation using citation-weighted patents (Panel a of Table 5 in the paper)

  FirmOutputRegressionCW.do: Firm Output and innovation using citation-weighted patents (Panel b of Table 5 in the paper)

  FirmReallocationRegressionCW.do: Firm Capital/Labor and innovation using citation-weighted patents (Panels c and d of Table 5 in the paper)

  FirmTFPRegressionCW.do: Firm TFP and innovation using citation-weighted patents (Panel e of Table 5 in the paper)

• FirmProfitsRegressionSMCW.do: Firm Profits and innovation using both our measure and citation-weighted patents (Panel a of Table 6 in the paper)

  FirmOutputRegressionSMCW.do: Firm Output and innovation using both our measure and citation-weighted patents (Panel b of Table 6 in the paper)

  FirmReallocationRegressionSMCW.do: Firm Capital/Labor and innovation using both our measure and citation-weighted patents (Panels c and d of Table 6 in the paper)

  FirmTFPRegressionSMCW.do: Firm TFP and innovation using both our measure and citation-weighted patents (Panel e of Table 6 in the paper)

• PatentValueScatterPlot.do: Plots the cross-sectional relation between forward patent citations and patent market value (Figure 2 in the paper)

• TimeSeriesPlots.do: Produces and plots the aggregate measures of innovation (Figure 4 in the paper)

• AggregateOutput.do: Runs the aggregate OLS regressions and VARs that relate our two innovation indices to output and TFP (The regression results for Figure 5 in the paper and Figure A.3 in the Online Appendix)

Code files using Matlab:

Prerequisite:

In order for the code to run correctly, you need to install the following program as well:

• matlab2tikz: Converts MATLAB®/Octave figures to TikZ/pgfplots figures for smooth integration into LaTeX. (Please go to https://github.com/matlab2tikz/matlab2tikz for detailed installation information)

File description:

• plot_OLS_responses.m: Take the output of AggregateOutput.do and creates Figure 5 in the paper

• csvimport.m: Define the function called in plot_OLS_responses.m which helps import the .csv files

• jbfill.m: Define the function called in plot_OLS_responses.m which helps plot the confidence intervals

Additional Notes:

• There is a minor discrepancy between the value of average acceptance probabilty used in the codes and reported on page 677 (published version of the paper). This minor difference is due to a typo in the published version.

• There are some minor discrepancies between the output that the code generates and the results reported in Table 5 (published version of the paper). These minor differences are due to typos in the published version.

• There are copyright restrictions and file size limitations in posting daily data from CRSP. Please contact the authors below to obtain the code and data needed to replicate Figures 1 and 3 in the paper that use this data.

Contact

Please contact Dimitris Papanikolaou (d-papanikolaou@kellogg.northwestern.edu) or Amit Seru (aseru@stanford.edu) for any questions regarding the codes or data.

Please see the paper for more information on the codes and data. If you use these codes files or data, please CITE this paper as the source.

Stata连享会   主页 || 视频 || 推文 || 知乎 || Bilibili 站

1. 连享会课程

免费公开课：

• 直击面板数据模型 - 连玉君，时长：1小时40分钟，课程主页，B 站版
• Stata 33 讲 - 连玉君, 每讲 15 分钟. B 站版
• Stata 小白的取经之路 - 龙志能，时长：2 小时，课程主页
• 部分直播课 课程资料下载 (PPT，dofiles等)

⛳ 课程主页

⛳ 课程主页

2. 资源分享

视频公开课

• 连享会码云：100多个精选计量项目 | 新浪视频
• 五分钟 Markdown / Markdown 幻灯片 | 新浪视频
• 连老师给你的-听课建议
• 直击面板数据模型 - 连玉君，时长：1小时40分钟，B 站版
• Stata 33 讲 - 连玉君, 每讲 15 分钟. 课程主页，课件，B 站版
• Stata小白的取经之路，龙志能，上海财经大学，去听课

Stata

• 连享会推文 | 直播视频
• 计量专题课程: Stata暑期班/寒假班 | 专题课程
• Stata专栏：最新推文 | 知乎 | CSDN | Bilibili 站
• Books and Journal: 计量Books | SJ-PDF | Stata Journal-在线浏览
• Stata Guys：Ben Jann

Data

• CSMAR-国泰安 | Wind-万德 | Resset-锐思
• 常用数据库
• 人文社科开放数据库
• 徐现祥教授-IRE-官员交流、方言等
• 知乎-Data

Papers - 学术论文复现

• 论文重现网站
• Google学术 | 统一入口：虫部落学术搜索 | 微软学术
• iData - 期刊论文下载
• CNKI | 百度学术 | Google学术 | Sci-hub , 2, 3, 4
• Stata论文重现: Harvard dataverse | JFE | github | Yahoo-github
• 学者主页(提供了诸多论文的原始数据和 dofiles)：Angrist || Daron Acemoglu || Ross Levine || Esther Duflo || Imbens || Raj Chetty

关于我们

• Stata连享会 由中山大学连玉君老师团队创办，定期分享实证分析经验。直播间 有很多视频课程，可以随时观看。
• 连享会-主页 和 知乎专栏，500+ 推文，实证分析不再抓狂；Bilibili 站 有视频大餐。

⛳ lianxh 命令发布了：
随时搜索连享会推文、Stata 资源，安装命令如下：
  . ssc install lianxh
使用详情参见帮助文件 (有惊喜)：
  . help lianxh