温馨提示: 定期 清理浏览器缓存,可以获得最佳浏览体验。
鲁维洁 (武汉大学, luweijie@whu.edu.cn)
何美玲 (西南财经大学, 472374565@qq.com)
连玉君 (中山大学, arlionn@163.com)
目录
金融领域的研究学者常使用事件研究法 (Event Study) 的分析框架,研究某一特定事件发生对公司股票价格或收益率的影响,并以此检验金融市场对新信息披露的反应程度。按照事件影响持续时间的长短,文献中通常将事件研究法分为短期事件研究与长期事件研究 (Brown and Warner,1980; Fama,1991) 。在公司金融领域,短期事件研究法 (Dailey Event Study) 为衡量某一事件对公司股东财富的影响提供了一个良好的度量指标,即累计异常收益率 (Cumulative Abnormal Returns, CARs) 。
短期事件研究法依赖的三个基本假设:首先,根据有效市场假说 (Efficient Markets Hypothesis, EMH) , 金融市场是有效的, 即股票价格反映所有已知的公共信息;其次,所研究的事件是市场未预期到的,因此这种异常收益可以度量股价对事件发生或信息披露异常反应的程度;第三,在事件发生的窗口期间无其他事件的混合效应。
短期事件研究法主要关注事件公布日周围数天内的公告效应,这为投资者理解公司的分红配股、兼并收购等决策提供了相关证据 (e.g., Edmans, 2011; Deng et al., 2013) ; 此外,短期事件研究作为检验市场效率的一种方法,在资本市场研究中也起着重要的作用。除了金融经济学,短期事件研究也被广泛运用于其它相关领域,例如在会计学领域的文献中,公司的发布盈余公告对其股票价格的影响备受关注;在法律和经济领域中,短期事件研究被用于度量监管政策 (如中组部18号文的出台的出台对企业政治关联的影响) 的效果,以及评估法律责任案件中的损害赔偿等 (如D&O保险) 。
短期事件研究法的主要概念如下:
如果市场是有效的、某一事件是意料之外的,并且该事件的发生与市场中某些公司的价值相关,那么通过将事后公司股票的实际收益 (ARs) 减去统计模型估计的正常收益率便能得到股票的异常收益率。依据 MacKinlay (1997) 的估计步骤,在进行异常收益率估计之前,首先需要定义事件发生的窗口期。通常,包括事件发生日期当天与事件发生日的前后数天。接着定义估计窗口 ,并选择估计模型计算预期收益率。以市场模型 (Market Model) 估计预期收益率为例,异常收益率 (AR) 可表示为:
其中
在公司金融的研究文献中,通常可以将事件分为以下三种:
如2.1定义所述,事件窗口期 为包含事件日的一段时间。为检测该事件是否被预期或泄密,通常事件窗口不仅包含事件发生后的一段时间,也会包含事件发生前几天。
定义好事件发生的日期与事件窗口期后,我们需要进行短期事件研究的数据准备,一般包括公司个股与股票市场的历史收益率数据 (Stock return data) 以及事件发生的日期数据 (Event dates)。如果每个公司的一组股票收益率数据可以匹配到单个事件日期,那么短期事件研究将会简单得多。然而,在许多情况下,一个公司在研究样本期间可能匹配到多次事件。在这种情况下,我们需要研究与同一公司有关的所有事件对其股票收益率的影响,即需要为每个“单个公司股票收益率-事件发生日期”的配对组合创建一组重复的股票收益率数据。简言之,在进行短期事件研究之前,我们需要理清每个公司在样本期间内究竟有几次事件发生。
如前所述,我们可以使用四种估计模型计算正常收益率,以最常见的市场模型 (Market model) 为例,
得到估计系数
在使用市场调整模型 (Market-adjusted model) 估计正常收益率时,模型设定为
其中,
在此基础上,计算平均累积异常收益率,也即在特定时点
以上公式中,
计算出累积异常收益率之后,最后需要检验每只股票的累积异常收益是否在统计上异于零,以便判断事件的发生是否对股价产生了显著的影响。
estudy
应用 estudy
的简介与基本语法estudy
是由 LIUC Università Carlo Cattaneo 的三位作者贡献的 Stata 外部命令,它作为一个集成的事件研究法估计程序,简洁清晰,方便使用,主要用于分析已知发生日期的某一特定事件或公告消息对公司股价的影响。但是,也因为 estudy
程序的执行简单方便,导致了它只能分析单一确定事件对公司股价的影响而无法对不同时间段发生的多次事件 (如一年中公司发布两次甚至两次以上的并购公告) 进行分析。在运用 estudy
命令进行短期事件研究前,可通过在 Stata 命令对话框输入 findit estudy
查找到同名安装包与示例数据 data_estudy.dta
进行命令安装与数据试运行, estudy
命令的基本语法如下所示:
estudy varlist1 [(varlist2) ... (varlistN)], datevar(varname) ///
evdate(date) dateformat(string) lb1(#) ub1(#) ///
[options]
estudy
命令的简单解释:
varlist1 [ (varlist2) ... (varlistN) ]
:每个varlist中填入的是样本内某一公司的变量存储名,变量中存放的是该公司的历史股票收益率时间序列数据。 estudy
将区分不同的 varlist (公司变量名) ,分别汇报累积异常收益率 (CAR) 与平均累积异常收益率 (CAAR) 。datevar
: 定义日期变量 (date) ,设为时间格式。evdate
: 定义事件发生的日期,如07092015。dateformat
: 定义 evdate 中 “年月日” 的格式,比如 dateformat (MDY) 则表示事件的日期格式按照月份(M)、日(D)、年(Y)的顺序进行排列。lb1(#)
和 ub1(#)
分别表示事件窗口期的起点 lb1(-3)
与 ub1(2)
代表,从事件发生前三天起,到事件发生后两天止是事件窗口期,即计算 lb2(#)
和 ub2(#)
来完成设置。其他 options
设置:
eswlbound(#)
: 设定估计窗口的起始日期,如有缺省,则自动设定第一个交易日为估计窗口的起始日期。eswubound(#)
: 设定估计窗口的截止日期,如有缺省,则自动设定事件发生日前一个月的30日为估计窗口的截止日期。modtype
:设置估计股票正常收益率的估计模型。其中, modtype(SIM)
表示使用市场模型, modtype(MAM)
表示使用市场调整模型, modtype(MFM)
表示多因素模型, modtype(HMM)
表示历史平均模型。indexlist(varlist)
: 用于存放用于估计股票正常收益率的各因子。diagnosticsstat(string)
: 设定检验显著性的方法。可选择的有参数法 diagnosticsstat(Norm)
、 diagnosticsstat(Patell)
、 diagnosticsstat(ADJPatell)
、 diagnosticsstat(BMP)
、 diagnosticsstat(KP)
、 diagnosticsstat(KP)
以及非参数法 diagnosticsstat(Wilcoxon)
、diagnosticsstat(GRANK)
。显著性检验方法具体说明可在Stata命令窗口输入 help estudy
了解更多。estudy
还有多个设置结果输出的选项,例如 suppress(string)
、 decimal(#)
等。suppress(group)表示报告每家公司的CAR, suppress(ind)表示报告整体CAAR。decimal可设置保留的小数点精确位数。estudy
命令实战在确认给 Stata 13.0 及以上版本的 Stata 安装好 estudy
安装包与示例数据 data_estudy.dta
之后,我们首先通过 cd
命令将当前工作路径所在的文件夹设置为保存示例数据的文件夹以方便调用 (当然,也可以通过菜单操作找到数据存放的文件夹直接打开数据) 。示例数据 data_estudy.dta
内储存的是 IBM 与可口可乐等公司的日度股票时间序列数据,并假设有且仅有一确定事件于2015年7月9日发生,研究这一事件对所有样本内公司股票收益率的影响。接下来,利用 estudy
命令,我们可以方便地按照以下步骤进行短期事件研究分析:
*-下载数据 https://gitee.com/arlionn/data/tree/master/data01
copy "https://gitee.com/arlionn/data/raw/master/data01/data_estudy.dta" "data_estudy.dta"
use "data_estudy.dta", clear
estudy boa ford boeing (apple netflix amazon facebook google), ///
datevar(date) evdate(07092015) dateformat(MDY) ///
lb1(-1) ub1(1) lb2(-3) ub2(3) ///
indexlist(mkt) decimal(4)
结果输出如下:
By default the upper bound of the estimation window has been set to (-30)
Event date: 09jul2015, with 2 event windows specified, under the Normality assumption
SECURITY CAAR[-1,1] CAAR[-3,3]
Bank of America Corporation 0.3647% -1.1494%
Ford Motor Company -2.1413% -1.8529%
The Boeing Company 0.9844% 3.4777%
Ptf CARs n 1 (3 securities) -0.2641% 0.1585%
CAAR group 1 (3 securities) -0.2641% 0.1585%
-------------------------------------------------------------------------
Apple Inc -1.9051% -2.1204%
Netflix Inc 2.8084% 2.9968%
Amazon com Inc 1.7837% 4.1678%
Facebook Inc 0.8082% 0.0043%
Alphabet Inc 1.1752% 5.3336%*
Ptf CARs n 2 (5 securities) 0.9341% 2.0764%
CAAR group 2 (5 securities) 0.9341% 2.0764%
-------------------------------------------------------------------------
*** p-value < .01, ** p-value <.05, * p-value <.1
estudy boa ford boeing (apple netflix amazon facebook google) (boa ford boeing apple netflix amazon facebook google), datevar(date) evdate(07092015) dateformat(MDY) lb1(-1) ub1(1) lb2(-3) ub2(3) lb3(-1) ub3(0) lb4(0) ub4(3) modtype(MFM) indexlist(mkt smb hml) diagnosticsstat(KP)
结果输出如下:
By default the upper bound of the estimation window has been set to (-30)
Event date: 09jul2015, with 4 event windows specified, using the Boehmer, Musumeci, Poulsen test, with the Kolari and Pynnonen adjustment
SECURITY CAAR[-1,1] CAAR[-3,3] CAAR[-1,0] CAAR[0,3]
Bank of America Corporation 0.95% 1.01% 0.41% 2.67%
Ford Motor Company -2.06% -1.31% -1.87% -0.13%
The Boeing Company 1.00% 3.40% 1.07% 0.87%
Ptf CARs n 1 (3 securities) -0.04% 1.04% -0.13% 1.14%
CAAR group 1 (3 securities) -0.04% 1.04% -0.13% 1.14%
-------------------------------------------------------------
Apple Inc -2.11% -2.93% -3.50%* -0.51%
Netflix Inc 2.30% 2.07% 2.81% 1.91%
Amazon com Inc 1.37% 2.97% 1.28% 3.40%
Facebook Inc -0.07% -2.04% -0.01% -0.93%
Alphabet Inc 0.91% 4.53% 0.68% 4.53%**
Ptf CARs n 2 (5 securities) 0.48% 0.92% 0.25% 1.68%
CAAR group 2 (5 securities) 0.48% 0.92% 0.25% 1.68%
-------------------------------------------------------------
Bank of America Corporation 0.95% 1.01% 0.41% 2.67%
Ford Motor Company -2.06% -1.31% -1.87% -0.13%
The Boeing Company 1.00% 3.40% 1.07% 0.87%
Apple Inc -2.11% -2.93% -3.50%* -0.51%
Netflix Inc 2.30% 2.07% 2.81% 1.91%
Amazon com Inc 1.37% 2.97% 1.28% 3.40%
Facebook Inc -0.07% -2.04% -0.01% -0.93%
Alphabet Inc 0.91% 4.53% 0.68% 4.53%**
Ptf CARs n 3 (8 securities) 0.29% 0.96% 0.11% 1.47%
CAAR group 3 (8 securities) 0.29% 0.96% 0.11% 1.47%*
-------------------------------------------------------------
*** p-value < .01, ** p-value <.05, * p-value <.1
estudy boa ford boeing (apple netflix amazon facebook google) (boa ford boeing apple netflix amazon facebook google), datevar(date) evdate(09072015) dateformat(DMY) lb1(-1) ub1(1) lb2(-3) ub2(3) lb3(-1) ub3(0) modtype(HMM) diagnosticsstat(Wilcoxon) showpvalues nostar
结果输出如下:
By default the upper bound of the estimation window has been set to (-30)
Event date: 09jul2015, with 3 event windows specified, using the Generalised SIGN test by Wilcoxon
SECURITY CAAR[-1,1] CAAR[-3,3] CAAR[-1,0]
Bank of America Corporation -0.12% 0.16% -1.39%
(0.9609) (0.9662) (0.4940)
Ford Motor Company -2.61% -0.60% -3.62%
(0.2819) (0.8718) (0.0670)
The Boeing Company 0.61% 4.50% -0.46%
(0.7865) (0.1904) (0.8026)
Ptf CARs n 1 (3 securities) -0.71% 1.35% -1.82%
(0.6986) (0.6289) (0.2215)
CAAR group 1 (3 securities) -0.71% 1.35% -1.82%
(0.0000) (0.0000) (0.1088)
---------------------------------------------------------------------------------------
Apple Inc -2.21% -1.30% -4.76%
(0.4272) (0.7609) (0.0362)
Netflix Inc 2.33% 4.29% 1.09%
(0.6769) (0.6163) (0.8112)
Amazon com Inc 1.27% 5.54% -0.71%
(0.7082) (0.2879) (0.7969)
Facebook Inc 0.28% 1.42% -1.91%
(0.9450) (0.8219) (0.5680)
Alphabet Inc 0.76% 6.46% -0.97%
(0.7451) (0.0701) (0.6089)
Ptf CARs n 2 (5 securities) 0.49% 3.28% -1.45%
(0.8318) (0.3495) (0.4364)
CAAR group 2 (5 securities) 0.49% 3.28% -1.45%
(0.0000) (0.0000) (0.0000)
---------------------------------------------------------------------------------------
Bank of America Corporation -0.12% 0.16% -1.39%
(0.9609) (0.9662) (0.4940)
Ford Motor Company -2.61% -0.60% -3.62%
(0.2819) (0.8718) (0.0670)
The Boeing Company 0.61% 4.50% -0.46%
(0.7865) (0.1904) (0.8026)
Apple Inc -2.21% -1.30% -4.76%
(0.4272) (0.7609) (0.0362)
Netflix Inc 2.33% 4.29% 1.09%
(0.6769) (0.6163) (0.8112)
Amazon com Inc 1.27% 5.54% -0.71%
(0.7082) (0.2879) (0.7969)
Facebook Inc 0.28% 1.42% -1.91%
(0.9450) (0.8219) (0.5680)
Alphabet Inc 0.76% 6.46% -0.97%
(0.7451) (0.0701) (0.6089)
Ptf CARs n 3 (8 securities) 0.04% 2.56% -1.59%
(0.9834) (0.3705) (0.2954)
CAAR group 3 (8 securities) 0.04% 2.56% -1.59%
(0.0000) (0.0000) (0.0000)
---------------------------------------------------------------------------------------
p-values in parentheses
my_output_tables.xlsx
和 my_ar_dataset.dta
文件中,可运行以下语句:estudy boa ford boeing (apple netflix amazon facebook google), ///
datevar(date) evdate(07092015) dateformat(MDY) ///
lb1(-1) ub1(1) lb2(-3) ub2(3) ///
indexlist(mkt) outputfile(my_output_tables)
结果输出如下:
By default the upper bound of the estimation window has been set to (-30)
Event date: 09jul2015, with 2 event windows specified, under the Normality assumption
SECURITY CAAR[-1,1] CAAR[-3,3]
Bank of America Corporation 0.36% -1.15%
Ford Motor Company -2.14% -1.85%
The Boeing Company 0.98% 3.48%
Ptf CARs n 1 (3 securities) -0.26% 0.16%
CAAR group 1 (3 securities) -0.26% 0.16%
-------------------------------------------------------------------------
Apple Inc -1.91% -2.12%
Netflix Inc 2.81% 3.00%
Amazon com Inc 1.78% 4.17%
Facebook Inc 0.81% 0.00%
Alphabet Inc 1.18% 5.33%*
Ptf CARs n 2 (5 securities) 0.93% 2.08%
CAAR group 2 (5 securities) 0.93% 2.08%
-------------------------------------------------------------------------
*** p-value < .01, ** p-value <.05, * p-value <.1
estudy boa ford boeing (apple netflix amazon facebook google), ///
datevar(date) evdate(07092015) dateformat(MDY) ///
lb1(-1) ub1(1) lb2(-3) ub2(3) ///
indexlist(mkt) mydataset(my_ar_dataset)
结果输出如下:
By default the upper bound of the estimation window has been set to (-30)
Event date: 09jul2015, with 2 event windows specified, under the Normality assumption
SECURITY CAAR[-1,1] CAAR[-3,3]
Bank of America Corporation 0.36% -1.15%
Ford Motor Company -2.14% -1.85%
The Boeing Company 0.98% 3.48%
Ptf CARs n 1 (3 securities) -0.26% 0.16%
CAAR group 1 (3 securities) -0.26% 0.16%
-------------------------------------------------------------------------
Apple Inc -1.91% -2.12%
Netflix Inc 2.81% 3.00%
Amazon com Inc 1.78% 4.17%
Facebook Inc 0.81% 0.00%
Alphabet Inc 1.18% 5.33%*
Ptf CARs n 2 (5 securities) 0.93% 2.08%
CAAR group 2 (5 securities) 0.93% 2.08%
-------------------------------------------------------------------------
*** p-value < .01, ** p-value <.05, * p-value <.1
(note: file my_ar_dataset.dta not found)
file my_ar_dataset.dta saved
运行以上命令之后,Stata会展示每个不同的事件窗口期的累积异常收益率与平均累积异常收益率的值及其显著性,通过正负符号及显著性分析,我们可以判断某一特定事件对不同公司价值的影响。
eventstudy2
应用 eventstudy2
的简介与基本语法eventstudy2
是由 Thomas Kaspereit (2019) 贡献的可用于进行事件研究的 Stata 外部命令,允许用户使用包括市场模型、 Fama 三因子模型等在内的估计模型。相对于 estudy
命令而言, eventstudy2
命令不仅让用户自由选择估计窗口和事件窗口的长度,而且能够同时计算超过10个事件窗口的累积(平均)异常收益率 CARs 和 CCARs ,并描绘图形。简言之,eventstudy2
命令是 Stata 执行事件研究中处理复杂的统计数据的程序。 eventstudy2
命令 的基本语法如下:
eventstudy2 security_id date using security_returns_file,
returns(security_returns)
[model(abnormal_return_model)
marketfile(file_with_market/factor_returns)
marketreturns(market_returns)
idmarket(market_id)
factor1(factor_return1) ... factor12(factor_return12)
riskfreerate(risk_free_rate)
prices(prices)
tradingvolume(trading_volume)
evwlb(event_window_lower_boundary)
evwub(event_window_upper_boundary)
eswlb(estimation_window_lower_boundary)
eswub(estimation_window_upper_boundary)
minevw(minimum_observations_event_window)
minesw(minimum_observations_estimation_window)
aarfile(output_average_abn_ret_file)
carfile(output_cum_average_abn_ret_file)
arfile(output_abn_ret_file)
crossfile(output_cross_sec_file)
diagnosticsfile(output_diag_file)
graphfile(output_graph_file)
replace
logreturns
thin(thin_trading_threshold)
fill nokolari delweekend
datelinethreshold(dateline_threshold)
shift(maximum_event_date_shift)
garch archoption(arch_option)
garchoption(garch_option)
architerate(iterations)
parallel pclusters(clusters)
processors(processors_StataMP)
prapath(path_eventstudy2_parallel)
arfillevent arfillestimation
car1LB(CAR_window_1_lower_boundary) car1UB(CAR_window_1_upper_boundary)
car2LB(CAR_window_2_lower_boundary) car2UB(CAR_window_2_upper_boundary)
... ...
car10LB(CAR_window_10_lower_boundary) car10UB(CAR_window_10_upper_boundary)]
其中,
security_id
是数据中识别每只股票或每家公司的 id 代码,以美国 WRDs-Compustat/CRSP 为例, permno 和 GVKEY 通常是每只股票或每家公司的识别代码。 date
是事件数据中每件事件发生的日期。security_returns_file
是包含存储在工作目录中的所估计股票收益率的返回值(参见选项 returns(security_returns)
), security_returns_file
还必须包含变量 security_id
和date
, 而这里 date
指的是股票收益率观察值返回的日期,而不是事件日期。returns(security_returns)
用于指定 security_returns_file
中表示事件研究法中与所发生事件可能相关的公司股票收益率的变量名称。model(abnormal_return_model)
用于指定估计正常收益率的估计模型并计算异常收益率,估计模型包括市场模型 (MA) 、 Fama (FM) 三因子模型等。marketfile(file_with_market/factor_returns)
用于指定用于存放股票收益率的股票数据。marketreturns(market_returns)
用于指定存放市场收益率的数据。idmarket(market_id)
用于识别在内存中存储的事件列表(或者股票收益率文件)。factor1(factor_return1)
到 factor12(factor_return12)
是 Fama三因子模型中各种因子的名称,如 riskfreerate(risk_free_rate)
用于识别数据中无风险利率的存储名称。prices(prices)
在文件 file_with_market/factor_returns
中指定股票价格的变量名称。tradingvolume(trading_volume)
在文件中存放所估计股票的交易量。evwlb(event_window_lower_boundary)
与 evwub(event_window_upper_boundary)
分别表示事件窗口的下限与上限, Stata 默认分别等于 -20 与 20 。eswlb(estimation_window_lower_boundary)
与 eswub(estimation_window_upper_boundary)
分别表示估计窗口的下限与上限, Stata 默认分别等于 -270 与 -20 。minevw(minimum_observations_event_window)
与 minesw(minimum_observations_estimation_window)
分别表示事件窗口与估计窗口的最少样本数量。在 Stata 中前者的默认设置为 1, 即事件窗口中至少放置 1 天用于估计;后者对股票收益率数据的长度有要求,如果股票数据观测值的天数不够,那么该股票的收益率历史数据将不参与事件研究的模型估计,默认设置为 30 天。aarfile(output_average_abn_ret_file)
用于指定存放历史平均异常收益率的文件名称。carfile(output_cum_average_abn_ret_file)
用于指定存放累积异常收益率的文件名称。arfile(output_abn_ret_file)
用于指定存放异常收益率的文件名称。crossfile(output_cross_sec_file)
用于指定保存用于横截面分析的累积异常收益率dta文件的文件名。diagnosticsfile(output_diag_file)
用于存放统计显著性的文件。graphfile(output_graph_file)
用于指定存放异常收益率图形的位置。logreturns
用于表明数据中的股票收益率及市场收益率是否是连续复利。thin_trading_threshold
用于指定将股票收益率观察值分类为 “Arising from Thin Trading” (针对交易量少的股票命名) 的分界值(阈值)。fill
用于指定对于缺失的股票收益率历史数据是按照缺失值还是按照 0 代替缺失值进行处理。car1LB(CAR_window_1_lower_boundary)
至 car10LB(CAR_window_10_lower_boundary)
用于指定计算异常收益率估计窗口的下限值,可以同时指定 10 个,默认值是 -20 。car1UB(CAR_window_1_upper_boundary)
至 car10UB(CAR_window_10_upper_boundary)
用于指定计算异常收益率的估计窗口的上限值,也可以同时指定 10 个,默认值是 20 。其它命令部分的详情请参看 help eventstudy2
。
eventstudy2
命令实战在确认给 Stata 安装好 eventstudy2
、 moremata
、 nearmrg
、 distinct
、 _gprod
、 rmse
和 parallel
安装包 ( ssc install eventstudy2
与 ssc install moremata/nearmrg/distinct/_gprod/rmse/parallel
) 与示例数据 Earnings_surprises.dta
之后,我们首先通过 cd
命令将当前工作路径所在的文件夹设置为保存示例数据的文件夹以方便调用 (当然,也可以通过菜单操作找到数据存放的文件夹直接打开数据) 。示例数据 Earnings_surprises.dta
内储存的分别是 Date Earnings_surprise
、Market_reference
及 Security_id
三个变量。接下来,利用 eventstudy2
命令,我们可以方便地按照以下步骤进行短期事件研究分析:
*-下载数据 https://gitee.com/arlionn/data/tree/master/data01
copy "https://gitee.com/arlionn/data/raw/master/data01/earnings_surprises.dta" earnings_surprises.dta
use "Earnings_surprises.dta", clear
eventstudy2 Security_id Date using Security_returns ///
if Earnings_surprise>0.05, returns(Return)
结果输出如下:
Generating dateline ...
...succeeded
Preparation of event list ...
...succeeded
Preparation of security return data...
...succeeded
Merging event dates and stock market data...
...succeeded
Calculating abnormal returns...
1 out of 39 events completed.
2 out of 39 events completed.
(output omitted)
38 out of 39 events completed.
39 out of 39 events completed.
...succeeded
Assessing statistical significance of abnormal returns...
...succeeded
Diagnosing events that are excluded from the analysis...
...succeeded
-----------------------------------------------------
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
opened on: 23 Dec 2019, 15:55:16
Number of events in the event file: 45
-- thereof: Number of events for which security identifiers and event dates are available: 45
-- thereof: Number of events for which event dates are in the range of dates in the security file: 45
-- thereof: Number of events in the analysis (not deleted because of any insufficient data in the estimation or event period): 39
List of security identifiers for which no security market data was available: 101 102 103 104
ANALYSIS OF ESTIMATION PERIOD
Number of events with insufficient security return data: 1
ANALYSIS OF EVENT PERIOD
Number of events with insufficient security return data: 0
Events for which the IPO (deletion) date of the event firm is later (earlier) than the first (last) day of the event window: 1
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
closed on: 23 Dec 2019, 15:55:16
-----------------------------------------------------
-----------------------------------------------------
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
opened on: 23 Dec 2019, 15:55:16
t NoFirms AAR t_test CDA Patell PatellADJ Boehmer Kolari Corrado Zivney GenSign Wilcox
-20 38 .004729 ** *
-19 39 .0029191
-18 36 -.0023795
-17 36 .0049581 *
-16 37 .0019446
-15 38 -.0020521
-14 39 .0066838
-13 37 -.0059705
-12 36 .0101652
-11 36 .0002356
-10 38 -.0054185
-9 36 -.0025805
-8 36 .0045793
-7 35 -.0100787
-6 37 -.0019002
-5 37 .0071357
-4 38 .0000734
-3 37 .008936
-2 38 -.0109795 ** ** *** **
-1 38 .012602 ** ** ** *** **
0 39 .025763 *** *** *** *** *** ** ** ** * **
1 39 -.0095898
2 36 -.0033064
3 37 -.0040544
4 37 -.0083442
5 39 .001478
6 38 -.0091507
7 38 .0034253
8 37 -.0024559
9 36 .0053815
10 38 -.0003621 *
11 39 .0017001 *
12 39 .0045353
13 38 .0018392
14 38 .0018035
15 37 .0027843
16 37 .0107235
17 38 -.0145819 * * * * *
18 39 .0078064 * *** **
19 36 -.0002348
20 39 .0011819
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
closed on: 23 Dec 2019, 15:55:16
-----------------------------------------------------
-----------------------------------------------------
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
opened on: 23 Dec 2019, 15:55:16
t NoFirms CAAR t_test CDA Patell PatellADJ Boehmer Kolari Corrado_Cowan Zivney_Cowan GenSign GRANKT Wilcox
[-20;20] 37 .0313972 ** * *** *** **
[-20;20] 37 .0313972 ** * *** *** **
[-20;20] 37 .0313972 ** * *** *** **
[-20;20] 37 .0313972 ** * *** *** **
[-20;20] 37 .0313972 ** * *** *** **
[-20;20] 37 .0313972 ** * *** *** **
[-20;20] 37 .0313972 ** * *** *** **
[-20;20] 37 .0313972 ** * *** *** **
[-20;20] 37 .0313972 ** * *** *** **
[-20;20] 37 .0313972 ** * *** *** **
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
closed on: 23 Dec 2019, 15:55:16
-----------------------------------------------------
-----------------------------------------------------
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
opened on: 23 Dec 2019, 15:55:16
The following result files are available in the evenstata directory and are loaded into memory by clicking.
Graph of cumulative average abnormal returns: graphfile
Average abnormal returns (daily basis): aarfile
Cumulative average abnormal returns: carfile
Abnormal returns: arfile
Cumulative abnormal returns for cross-sectional analyses: crossfile
Diagnostic of events that are excluded: diagnosticsfile
Logfile: diagnosticsfile
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
closed on: 23 Dec 2019, 15:55:16
-----------------------------------------------------
eventstudy2不仅完成了累积超额收益的计算与显著性检验,并且将相关数据自动储存在当前读取的文件夹里,可直接点击 "graphfile" 、" aarfile" 等阅览相应文件。
eventstudy2 Security_id Date using Security_returns if Ea>0.05, ///
ret(Return) car1LB(-1) car1UB(1) mod(FM) ///
marketfile(Factor_returns) mar(MKT) ///
idmar(Market_reference) ///
factor1(SMB) factor2(HML) ///
risk(risk_free_rate)
结果输出如下:
Generating dateline ...
...succeeded
Preparation of event list ...
...succeeded
Preparation of security return data...
...succeeded
Preparation of market and/or factor return data...
...succeeded
Merging event dates and stock market data...
...succeeded
Calculating abnormal returns...
1 out of 38 events completed.
2 out of 38 events completed.
(output omitted)
37 out of 38 events completed.
38 out of 38 events completed.
...succeeded
Assessing statistical significance of abnormal returns...
...succeeded
Diagnosing events that are excluded from the analysis...
...succeeded
------------------------------------------------
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
opened on: 23 Dec 2019, 16:29:49
Number of events in the event file: 45
-- thereof: Number of events for which security identifiers and event dates are available: 45
-- thereof: Number of events for which event dates are in the range of dates in the security file: 44
-- thereof: Number of events in the analysis (not deleted because of any insufficient data in the estimation or event period): 38
List of security identifiers for which no security market data was available: 102 103 104
ANALYSIS OF ESTIMATION PERIOD
Number of events with insufficient security return data: 2
Number of events with insufficient market index/factor return data: 2
ANALYSIS OF EVENT PERIOD
Number of events with insufficient security return data: 0
Number of events with insufficient market/index factor return data: 0
Events for which the IPO (deletion) date of the event firm is later (earlier) than the first (last) day of the event window: 1
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
closed on: 23 Dec 2019, 16:29:49
---------------------------------------
---------------------------------------
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
opened on: 23 Dec 2019, 16:29:49
t NoFirms AAR t_test CDA Patell PatellADJ Boehmer Kolari Corrado Zivney GenSign Wilcox
-20 37 .0049708
-19 38 .0030984
-18 35 -.0038899
-17 35 -.000733
-16 36 -.0003419
-15 37 .00481 * **
-14 38 .0043478
-13 36 -.0074897 * * ** * *
-12 35 .0062511
-11 35 .0013786
-10 37 -.0112272 * *
-9 35 .002375
-8 35 .0068386 *
-7 34 -.0116341
-6 36 -.003284
-5 36 .0092179
-4 37 .0008256
-3 36 .006481
-2 37 -.0035439
-1 37 .0076635
0 38 .0293003 *** *** *** *** *** *** *** *** *** ***
1 38 -.0035208
2 36 -.0005523
3 36 -.0063949
4 36 -.0054967
5 38 -.0005514
6 37 -.0020611
7 37 -.0024591
8 36 -.0018874
9 35 -.0034669
10 37 -.0040193 *
11 38 -.0010819
12 38 .0036728
13 37 -.0023843
14 37 .0008085
15 36 .0022357
16 36 -.0015717
17 37 -.0131729 * * * ** ** * ** * ***
18 38 -.0011146
19 35 -.0051176
20 38 -.004766
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
closed on: 23 Dec 2019, 16:29:49
-----------------------------------------------------
-----------------------------------------------------
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
opened on: 23 Dec 2019, 16:29:49
t NoFirms CAAR t_test CDA Patell PatellADJ Boehmer Kolari Corrado_Cowan Zivney_Cowan GenSign GRANKT Wilcox
[-1;1] 37 .0360421 *** *** *** *** *** ** *** *** ** *** **
[-20;20] 36 -.0114423
[-20;20] 36 -.0114423
[-20;20] 36 -.0114423
[-20;20] 36 -.0114423
[-20;20] 36 -.0114423
[-20;20] 36 -.0114423
[-20;20] 36 -.0114423
[-20;20] 36 -.0114423
[-20;20] 36 -.0114423
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
closed on: 23 Dec 2019, 16:29:49
-----------------------------------------------------
-----------------------------------------------------
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
opened on: 23 Dec 2019, 16:29:49
The following result files are available in the evenstata directory and are loaded into memory by clicking.
Graph of cumulative average abnormal returns: graphfile
Average abnormal returns (daily basis): aarfile
Cumulative average abnormal returns: carfile
Abnormal returns: arfile
Cumulative abnormal returns for cross-sectional analyses: crossfile
Diagnostic of events that are excluded: diagnosticsfile
Logfile: diagnosticsfile
name: <unnamed>
log: E:\draft\diagnosticsfile.smcl
log type: smcl
closed on: 23 Dec 2019, 16:29:49
atuo.dta
模拟 Stata 范例使用这些命令,可以一次性完成基本的 Event Study 估计和检验工作,非常便捷。
需要事先用 ssc install cmdname, replace
下载最新版;亦可使用 findit
命令搜索相应的命令,以便查看完整的 Stata 范例数据和 dofile。
help eventstudy
//基本命令help eventstudy2
//基本命令help st0532_1
// Stata Journal 19-2,只能指定单一的事件日期,但可以分析特定公司的事件效果
连享会-直播课 上线了!
http://lianxh.duanshu.com
免费公开课:
直击面板数据模型 - 连玉君,时长:1小时40分钟 Stata 33 讲 - 连玉君, 每讲 15 分钟. 部分直播课 课程资料下载 (PPT,dofiles等)
支持回看,所有课程可以随时购买观看。
专题 | 嘉宾 | 直播/回看视频 |
---|---|---|
⭐ 最新专题 ⭐ | DSGE, 因果推断, 空间计量等 | |
⭕ Stata数据清洗 | 游万海 | 直播, 2 小时,已上线 |
研究设计 | 连玉君 | 我的特斯拉-实证研究设计,-幻灯片- |
面板模型 | 连玉君 | 动态面板模型,-幻灯片- |
面板模型 | 连玉君 | 直击面板数据模型 [免费公开课,2小时] |
Note: 部分课程的资料,PPT 等可以前往 连享会-直播课 主页查看,下载。
关于我们
课程, 直播, 视频, 客服, 模型设定, 研究设计, stata, plus, 绘图, 编程, 面板, 论文重现, 可视化, RDD, DID, PSM, 合成控制法
等
连享会小程序:扫一扫,看推文,看视频……
扫码加入连享会微信群,提问交流更方便
✏ 连享会学习群-常见问题解答汇总:
✨ https://gitee.com/arlionn/WD