Dataset for the analysis of stock price responses to African Swine Fever Outbreaks in China

The African Swine Fever (ASF) outbreaks in China since 2018 caused a more than 100 million decline in its hog inventory. Leveraging publicly available announcements of ASF outbreaks and daily stock prices data from 25 major publicly listed hog companies from China and eight major hog exporting countries, we use the event study method to estimate firm-level abnormal stock price responses to China's ASF outbreak announcements for both Chinese and foreign hog companies. This article describes the data used in the research article “A Fortune from misfortune: Evidence from hog firms’ stock price responses to China's African Swine Fever outbreaks” (Xiong et al., 2021). The daily stock price data in this article can be applied to other events that also occurred during the same sample period using a similar event study approach.


Specifications
Economics, Econometrics and Finance Specific subject area Statistical methods applied to daily stock price data to estimate the abnormal stock returns for Chinese and foreign hog firms from ASF outbreaks in China. Type of data Data Stata Codes Graph Figure  How the data were acquired All data are downloaded from publicly available sources (see Description of data collection for more). Data format Raw Analysed Description of data collection ASF announcements in China from August 2018 to September 2019 came from China Ministry of Agriculture and Rural Affairs ( http://www.moa.gov.cn/gk/yjgl _ 1/yqfb/ ). The ASF announcements detailed the release date, the county-level location and specific site (i.e., pig farm, slaughterhouse, or transport vehicle) of event detection, the number of hogs in inventory, and the number of infected and dead pigs. The release dates were used as event dates in the event study, while the county-level location, the number of hogs in inventory and the number of infected pigs were used as explanatory variables in the Pooled OLS model. Value of the Data • The daily firm-level stock price data contain important information that reflects not only to economic drivers, but often also the non-economic shocks. The stock price data are useful for evaluating firms' performance in the market in response to both economic and non-economic events. As was the case for African Swine Fever (ASF) outbreaks in China, Xiong, Zhang, and Chen [3] showed that announcements of ASF outbreaks resulted in statistically significant impacts on stock prices for both Chinese and international hog firms, where these hog firms profited from the reduction in Chinese hog inventories due to the epidemic. • Researchers interested in using stock price as the indicator for firm performance in the market can benefit from these data. The daily stock price data for Chinese and international hog firms allow others to study the effect of any economic and non-economic events that also occurred during the same sample period that might impact stock price movements. For example, researchers interested in the impact of U.S.-China trade war on stock price responses among these companies can use the daily stock price data in this article. • Both the stock price data for hog firms and the China ASF announcements data presented in this article can not only be reused to explore different aspects of the economic consequences as a result of the ASF outbreaks in China other than for the financial market, but they can also be used for empirical applications of new statistical approaches analysing stock price responses. Researchers that seek to develop new statistical methods to investigate stock price movements can use the presented stock price data in this article for model validation.

Data Description
Data presented in this article include raw data accessible from the respective public online sources, as well as analysed data that requires necessary statistical program to generate.
The 'Data' folder supplied with this article contains three sub-folders ('Chinese firms and market indices', 'Global firms and market indices', and 'Cumulative abnormal returns') and an EXCEL file. The 'Chinese firms and market indices' sub-folder contains ten Chinese hog firms' daily stock prices and four Chinese daily stock market indices. The 'Global firms and market indices' sub-folder has fifteen foreign hog firms' daily stock prices and seven global daily stock market indices. Both the market indices and stock prices were downloaded from Yahoo Finance ( https://finance.yahoo.com ). The 'Cumulative abnormal returns' sub-folder contains the cumulative abnormal returns (CARs), defined as the accumulated difference between the estimated market indices and observed stock returns since each event announcement, for both Chinese and foreign hog firms. The CARs are used for the pooled OLS regression analyses. 1 The data of ASF event announcements is stored as an EXCEL file under the subfolder 'Data'. ASF event announcements in China from August 2018 to September 2019 came from China Ministry of Agriculture and Rural Affairs [2] ( http://www.moa.gov.cn/gk/yjgl _ 1/yqfb/ ). Following each ASF outbreak, MARA announces the release date, the county-level location and specific site (i.e., pig farm, slaughterhouse, or transport vehicle) of event detection, the number of hogs in inventory, and the number of infected and dead pigs. MARA issued a total of 138 ASF announcements during our study period [2] . The release dates were used as event dates in the event study, while the county-level location, the number of hogs in inventory and the number of infected pigs were used as explanatory variables in the Pooled OLS model.
The definition of all variables which appear in the dataset is as follows. For the variables in the 'Chinese firms and market indices' sub-folder: where the ASF outbreaks were detected: hog farm, slaughterhouse, and transport. 7. Inventory: the hog inventory of the place where the ASF outbreaks were detected 8. Infected: the number of the infected hog at the place where the ASF outbreaks were detected 9. Death: the number of the dead hog at the place where the ASF outbreaks were detected

Experimental Design, Materials and Methods
'Codes' sub-folder contains the codes and a demo for event studies, the main research method in the supporting article, and codes for the pooled OLS regressions. Specifically, we use the 'eventstudy2' module to perform event study in this paper by Kaspereit [1] . 'eventstudy2' is not an official Stata command. This module should be installed from within Stata by typing "ssc install eventstudy2". It should be noted that we should first install other six user-written modules by typing "ssc install moremata", "ssc install nearmrg", "ssc install distinct", "ssc install _gprod", "ssc install rmse", "ssc install parallel" before installing 'eventstudy2' module. For more information about 'eventstudy2' module, please type "help eventstudy2" in Stata.
The data and variables used in this DIB file mainly come from two sources: China Ministry of Agriculture and Rural Affairs (MARA), China ( http://www.moa.gov.cn/gk/yjgl _ 1/yqfb/ ) (in Chinese) and Yahoo Finance, United States ( https://finance.yahoo.com ).
The Chinese MARA data source includes the nationally published events information for China's African Swine Fever outbreaks, and this information includes the specific date of a particular outbreak (asf_date), whether the outbreak was detected in a large pig farm (de-tect_largefarm), the place where the outbreak was detected (Place_type), the number of the   Note: * * * p < 0.01, * * p < 0.05, * p < 0.1. In all regressions, the dependent variables are cumulative abnormal returns (CAR). CAR values in all regressions are calculated using a 15-day event window. Day dummies indicate days since each ASF event, from day 1 to day 15. Kao test for panel cointegration uses the Augmented Dickey-Fuller t statistics with 1 lag, following the specifications in Equation (12); Hadri LM test for panel unit roots on residuals also includes 1 lag. Inference is based on clustered-robust standard errors (in parenthesis).
hog inventory of the place, the number of infected pigs (infected), the number of dead pigs (infected). All variables which appear in both 'Chinese firms and market indices' and 'Global firms and market indices' sub-folders are taken from Yahoo Finance, and none of the variables are created by the authors.
The variable "car" denoted the cumulative abnormal returns in the 'Cumulative abnormal returns' sub-folder. For firm i over a time interval τ = [ τ 1 , τ 2 ] , we calculate CAR as

AR it
where T 2 + 1 ≤ τ 1 ≤ τ 2 ≤ T 3 and AR it is calculated from the event study. Please refer to Section 3 of the paper (Xiong et al., 2021).    (12); Hadri LM test for panel unit roots on residuals also includes 1 lag. Inference is based on clustered-robust standard errors (in parenthesis).
The primary function of each file in this sub-folder is briefly introduced as follows.
[1] Codes_eventstudy.do: It's the main function that performs event studies.
[2] Event_id.dta: It's a STATA data file that contains the event's information.
[3] Marketfile.dta: It's a STATA data file that contains the stock market's information.
[4] Security_returns.dta: It's a STATA data file that contains the hog firms' information.
[5] Codes_pooledols.do: It replicates the pooled OLS regression results in Tables 1 and 2 , using STATA data files from the 'Cumulative abnormal returns' sub-folder.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data Availability
Replication files for "A Fortune from misfortune: Evidence from hog firms' stock price responses to China's African Swine Fever outbreaks" (Original data) (Mendeley Data).