siat 3.10.132__py3-none-any.whl → 3.11.1__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- siat/__init__.py +0 -0
- siat/allin.py +8 -0
- siat/assets_liquidity.py +0 -0
- siat/beta_adjustment.py +0 -0
- siat/beta_adjustment_china.py +0 -0
- siat/blockchain.py +0 -0
- siat/bond.py +0 -0
- siat/bond_base.py +0 -0
- siat/bond_china.py +0 -0
- siat/bond_zh_sina.py +0 -0
- siat/capm_beta.py +0 -0
- siat/capm_beta2.py +4 -4
- siat/common.py +9 -6
- siat/compare_cross.py +0 -0
- siat/copyrights.py +0 -0
- siat/cryptocurrency.py +0 -0
- siat/economy.py +0 -0
- siat/economy2.py +0 -0
- siat/esg.py +0 -0
- siat/event_study.py +0 -0
- siat/exchange_bond_china.pickle +0 -0
- siat/fama_french.py +0 -0
- siat/fin_stmt2_yahoo.py +0 -0
- siat/financial_base.py +0 -0
- siat/financial_statements.py +0 -0
- siat/financials.py +0 -0
- siat/financials2.py +0 -0
- siat/financials_china.py +0 -0
- siat/financials_china2.py +0 -0
- siat/fund.py +0 -0
- siat/fund_china.pickle +0 -0
- siat/fund_china.py +0 -0
- siat/future_china.py +0 -0
- siat/google_authenticator.py +0 -0
- siat/grafix.py +55 -4
- siat/holding_risk.py +0 -0
- siat/luchy_draw.py +0 -0
- siat/market_china.py +0 -0
- siat/markowitz.py +0 -0
- siat/markowitz2.py +1 -0
- siat/markowitz2_20250704.py +0 -0
- siat/markowitz2_20250705.py +0 -0
- siat/markowitz_simple.py +0 -0
- siat/ml_cases.py +0 -0
- siat/ml_cases_example.py +0 -0
- siat/option_china.py +0 -0
- siat/option_pricing.py +0 -0
- siat/other_indexes.py +0 -0
- siat/risk_adjusted_return.py +0 -0
- siat/risk_adjusted_return2.py +8 -4
- siat/risk_evaluation.py +0 -0
- siat/risk_free_rate.py +0 -0
- siat/save2docx.py +345 -0
- siat/save2pdf.py +145 -0
- siat/sector_china.py +0 -0
- siat/security_price2.py +0 -0
- siat/security_prices.py +168 -6
- siat/security_trend.py +0 -0
- siat/security_trend2.py +2 -2
- siat/stock.py +11 -1
- siat/stock_advice_linear.py +0 -0
- siat/stock_base.py +0 -0
- siat/stock_china.py +0 -0
- siat/stock_info.pickle +0 -0
- siat/stock_prices_kneighbors.py +0 -0
- siat/stock_prices_linear.py +0 -0
- siat/stock_profile.py +0 -0
- siat/stock_technical.py +0 -0
- siat/stooq.py +0 -0
- siat/transaction.py +0 -0
- siat/translate.py +0 -0
- siat/valuation.py +0 -0
- siat/valuation_china.py +0 -0
- siat/var_model_validation.py +0 -0
- siat/yf_name.py +0 -0
- {siat-3.10.132.dist-info/licenses → siat-3.11.1.dist-info}/LICENSE +0 -0
- {siat-3.10.132.dist-info → siat-3.11.1.dist-info}/METADATA +234 -235
- siat-3.11.1.dist-info/RECORD +80 -0
- {siat-3.10.132.dist-info → siat-3.11.1.dist-info}/WHEEL +1 -1
- {siat-3.10.132.dist-info → siat-3.11.1.dist-info}/top_level.txt +0 -1
- build/lib/build/lib/siat/__init__.py +0 -75
- build/lib/build/lib/siat/allin.py +0 -137
- build/lib/build/lib/siat/assets_liquidity.py +0 -915
- build/lib/build/lib/siat/beta_adjustment.py +0 -1058
- build/lib/build/lib/siat/beta_adjustment_china.py +0 -548
- build/lib/build/lib/siat/blockchain.py +0 -143
- build/lib/build/lib/siat/bond.py +0 -2900
- build/lib/build/lib/siat/bond_base.py +0 -992
- build/lib/build/lib/siat/bond_china.py +0 -100
- build/lib/build/lib/siat/bond_zh_sina.py +0 -143
- build/lib/build/lib/siat/capm_beta.py +0 -783
- build/lib/build/lib/siat/capm_beta2.py +0 -887
- build/lib/build/lib/siat/common.py +0 -5360
- build/lib/build/lib/siat/compare_cross.py +0 -642
- build/lib/build/lib/siat/copyrights.py +0 -18
- build/lib/build/lib/siat/cryptocurrency.py +0 -667
- build/lib/build/lib/siat/economy.py +0 -1471
- build/lib/build/lib/siat/economy2.py +0 -1853
- build/lib/build/lib/siat/esg.py +0 -536
- build/lib/build/lib/siat/event_study.py +0 -815
- build/lib/build/lib/siat/fama_french.py +0 -1521
- build/lib/build/lib/siat/fin_stmt2_yahoo.py +0 -982
- build/lib/build/lib/siat/financial_base.py +0 -1160
- build/lib/build/lib/siat/financial_statements.py +0 -598
- build/lib/build/lib/siat/financials.py +0 -2339
- build/lib/build/lib/siat/financials2.py +0 -1278
- build/lib/build/lib/siat/financials_china.py +0 -4433
- build/lib/build/lib/siat/financials_china2.py +0 -2212
- build/lib/build/lib/siat/fund.py +0 -629
- build/lib/build/lib/siat/fund_china.py +0 -3307
- build/lib/build/lib/siat/future_china.py +0 -551
- build/lib/build/lib/siat/google_authenticator.py +0 -47
- build/lib/build/lib/siat/grafix.py +0 -3636
- build/lib/build/lib/siat/holding_risk.py +0 -867
- build/lib/build/lib/siat/luchy_draw.py +0 -638
- build/lib/build/lib/siat/market_china.py +0 -1168
- build/lib/build/lib/siat/markowitz.py +0 -2363
- build/lib/build/lib/siat/markowitz2.py +0 -3150
- build/lib/build/lib/siat/markowitz2_20250704.py +0 -2969
- build/lib/build/lib/siat/markowitz2_20250705.py +0 -3158
- build/lib/build/lib/siat/markowitz_simple.py +0 -373
- build/lib/build/lib/siat/ml_cases.py +0 -2291
- build/lib/build/lib/siat/ml_cases_example.py +0 -60
- build/lib/build/lib/siat/option_china.py +0 -3069
- build/lib/build/lib/siat/option_pricing.py +0 -1925
- build/lib/build/lib/siat/other_indexes.py +0 -409
- build/lib/build/lib/siat/risk_adjusted_return.py +0 -1576
- build/lib/build/lib/siat/risk_adjusted_return2.py +0 -1900
- build/lib/build/lib/siat/risk_evaluation.py +0 -2218
- build/lib/build/lib/siat/risk_free_rate.py +0 -351
- build/lib/build/lib/siat/sector_china.py +0 -4140
- build/lib/build/lib/siat/security_price2.py +0 -727
- build/lib/build/lib/siat/security_prices.py +0 -3408
- build/lib/build/lib/siat/security_trend.py +0 -402
- build/lib/build/lib/siat/security_trend2.py +0 -646
- build/lib/build/lib/siat/stock.py +0 -4284
- build/lib/build/lib/siat/stock_advice_linear.py +0 -934
- build/lib/build/lib/siat/stock_base.py +0 -26
- build/lib/build/lib/siat/stock_china.py +0 -2095
- build/lib/build/lib/siat/stock_prices_kneighbors.py +0 -910
- build/lib/build/lib/siat/stock_prices_linear.py +0 -386
- build/lib/build/lib/siat/stock_profile.py +0 -707
- build/lib/build/lib/siat/stock_technical.py +0 -3305
- build/lib/build/lib/siat/stooq.py +0 -74
- build/lib/build/lib/siat/transaction.py +0 -347
- build/lib/build/lib/siat/translate.py +0 -5183
- build/lib/build/lib/siat/valuation.py +0 -1378
- build/lib/build/lib/siat/valuation_china.py +0 -2076
- build/lib/build/lib/siat/var_model_validation.py +0 -444
- build/lib/build/lib/siat/yf_name.py +0 -811
- build/lib/siat/__init__.py +0 -75
- build/lib/siat/allin.py +0 -137
- build/lib/siat/assets_liquidity.py +0 -915
- build/lib/siat/beta_adjustment.py +0 -1058
- build/lib/siat/beta_adjustment_china.py +0 -548
- build/lib/siat/blockchain.py +0 -143
- build/lib/siat/bond.py +0 -2900
- build/lib/siat/bond_base.py +0 -992
- build/lib/siat/bond_china.py +0 -100
- build/lib/siat/bond_zh_sina.py +0 -143
- build/lib/siat/capm_beta.py +0 -783
- build/lib/siat/capm_beta2.py +0 -887
- build/lib/siat/common.py +0 -5360
- build/lib/siat/compare_cross.py +0 -642
- build/lib/siat/copyrights.py +0 -18
- build/lib/siat/cryptocurrency.py +0 -667
- build/lib/siat/economy.py +0 -1471
- build/lib/siat/economy2.py +0 -1853
- build/lib/siat/esg.py +0 -536
- build/lib/siat/event_study.py +0 -815
- build/lib/siat/fama_french.py +0 -1521
- build/lib/siat/fin_stmt2_yahoo.py +0 -982
- build/lib/siat/financial_base.py +0 -1160
- build/lib/siat/financial_statements.py +0 -598
- build/lib/siat/financials.py +0 -2339
- build/lib/siat/financials2.py +0 -1278
- build/lib/siat/financials_china.py +0 -4433
- build/lib/siat/financials_china2.py +0 -2212
- build/lib/siat/fund.py +0 -629
- build/lib/siat/fund_china.py +0 -3307
- build/lib/siat/future_china.py +0 -551
- build/lib/siat/google_authenticator.py +0 -47
- build/lib/siat/grafix.py +0 -3636
- build/lib/siat/holding_risk.py +0 -867
- build/lib/siat/luchy_draw.py +0 -638
- build/lib/siat/market_china.py +0 -1168
- build/lib/siat/markowitz.py +0 -2363
- build/lib/siat/markowitz2.py +0 -3150
- build/lib/siat/markowitz2_20250704.py +0 -2969
- build/lib/siat/markowitz2_20250705.py +0 -3158
- build/lib/siat/markowitz_simple.py +0 -373
- build/lib/siat/ml_cases.py +0 -2291
- build/lib/siat/ml_cases_example.py +0 -60
- build/lib/siat/option_china.py +0 -3069
- build/lib/siat/option_pricing.py +0 -1925
- build/lib/siat/other_indexes.py +0 -409
- build/lib/siat/risk_adjusted_return.py +0 -1576
- build/lib/siat/risk_adjusted_return2.py +0 -1900
- build/lib/siat/risk_evaluation.py +0 -2218
- build/lib/siat/risk_free_rate.py +0 -351
- build/lib/siat/sector_china.py +0 -4140
- build/lib/siat/security_price2.py +0 -727
- build/lib/siat/security_prices.py +0 -3408
- build/lib/siat/security_trend.py +0 -402
- build/lib/siat/security_trend2.py +0 -646
- build/lib/siat/stock.py +0 -4284
- build/lib/siat/stock_advice_linear.py +0 -934
- build/lib/siat/stock_base.py +0 -26
- build/lib/siat/stock_china.py +0 -2095
- build/lib/siat/stock_prices_kneighbors.py +0 -910
- build/lib/siat/stock_prices_linear.py +0 -386
- build/lib/siat/stock_profile.py +0 -707
- build/lib/siat/stock_technical.py +0 -3305
- build/lib/siat/stooq.py +0 -74
- build/lib/siat/transaction.py +0 -347
- build/lib/siat/translate.py +0 -5183
- build/lib/siat/valuation.py +0 -1378
- build/lib/siat/valuation_china.py +0 -2076
- build/lib/siat/var_model_validation.py +0 -444
- build/lib/siat/yf_name.py +0 -811
- siat-3.10.132.dist-info/RECORD +0 -218
build/lib/siat/event_study.py
DELETED
@@ -1,815 +0,0 @@
|
|
1
|
-
# -*- coding: utf-8 -*-
|
2
|
-
"""
|
3
|
-
本模块功能:证券事件分析法
|
4
|
-
所属工具包:证券投资分析工具SIAT
|
5
|
-
SIAT:Security Investment Analysis Tool
|
6
|
-
创建日期:2024年11月14日
|
7
|
-
最新修订日期:
|
8
|
-
作者:王德宏 (WANG Dehong, Peter)
|
9
|
-
作者单位:北京外国语大学国际商学院
|
10
|
-
作者邮件:wdehong2000@163.com
|
11
|
-
版权所有:王德宏
|
12
|
-
用途限制:仅限研究与教学使用!
|
13
|
-
特别声明:作者不对使用本工具进行证券投资导致的任何损益负责!
|
14
|
-
"""
|
15
|
-
|
16
|
-
#==============================================================================
|
17
|
-
#关闭所有警告
|
18
|
-
import warnings; warnings.filterwarnings('ignore')
|
19
|
-
#==============================================================================
|
20
|
-
from siat.common import *
|
21
|
-
from siat.translate import *
|
22
|
-
#from siat.security_trend2 import *
|
23
|
-
|
24
|
-
from siat.stock import *
|
25
|
-
#from siat.security_prices import *
|
26
|
-
#from siat.security_price2 import *
|
27
|
-
#from siat.capm_beta2 import *
|
28
|
-
#from siat.risk_adjusted_return2 import *
|
29
|
-
#from siat.valuation import *
|
30
|
-
|
31
|
-
from siat.grafix import *
|
32
|
-
|
33
|
-
import pandas as pd; import numpy as np
|
34
|
-
|
35
|
-
import datetime as dt; stoday=str(dt.date.today())
|
36
|
-
#==============================================================================
|
37
|
-
#==============================================================================
|
38
|
-
if __name__=='__main__':
|
39
|
-
#测试组1
|
40
|
-
ticker='600519.SS'
|
41
|
-
|
42
|
-
event_date='2024-4-2' #贵州茅台2023年报于2024年4月2日晚披露
|
43
|
-
start='2024-3-1'; end='2024-4-30'
|
44
|
-
event_window=[1,1] #事件发生时股市已经收盘,故检测下一个交易日的股市反应
|
45
|
-
market_index='000001.SS' #贵州茅台在上交所上市,故使用上证综合指数
|
46
|
-
RF=0
|
47
|
-
|
48
|
-
#测试组1b
|
49
|
-
ticker='600519.SS'
|
50
|
-
|
51
|
-
event_date='2024-4-2' #贵州茅台2023年报于2024年4月2日晚披露
|
52
|
-
start='2024-3-1'; end='2024-4-30'
|
53
|
-
event_window=[0,2] #事件发生时股市已经收盘,故检测下一个交易日的股市反应
|
54
|
-
market_index='000001.SS' #贵州茅台在上交所上市,故使用上证综合指数
|
55
|
-
RF=0
|
56
|
-
|
57
|
-
#测试组2
|
58
|
-
ticker=['600519.SS','399997.SZ']
|
59
|
-
|
60
|
-
event_date='2024-3-15' #315晚会
|
61
|
-
start='2024-3-1'; end='2024-3-30'
|
62
|
-
event_window=[1,2]
|
63
|
-
market_index='000300.SS'
|
64
|
-
RF="market model"
|
65
|
-
|
66
|
-
#测试组3
|
67
|
-
ticker=['600519.SS','399997.SZ']
|
68
|
-
|
69
|
-
event_date='2024-4-2' #贵州茅台2023年报披露日
|
70
|
-
start='auto'; end='auto'
|
71
|
-
event_window=[0,1]
|
72
|
-
method='CAPM'
|
73
|
-
market_index='000001.SS'
|
74
|
-
RF="1YCNY.B"
|
75
|
-
|
76
|
-
#共同部分
|
77
|
-
post_event_days=7
|
78
|
-
method='CAPM'
|
79
|
-
early_response_days=-2
|
80
|
-
estimation_window_days=-365
|
81
|
-
|
82
|
-
ret_type="Daily Adj Ret%"
|
83
|
-
ticker_type='stock' #贵州茅台为股票
|
84
|
-
facecolor="whitesmoke"
|
85
|
-
show_AR=True
|
86
|
-
show_RF=True
|
87
|
-
show_BHAR=True
|
88
|
-
loc='best'
|
89
|
-
|
90
|
-
es=event_study("600519.SS",event_date="2024-4-2", \
|
91
|
-
start='2024-3-1',end='2024-4-30', \
|
92
|
-
event_window=[0,0],post_event_days=7, \
|
93
|
-
method='CAPM', \
|
94
|
-
market_index='000001.SS',RF=0.0143)
|
95
|
-
|
96
|
-
es=event_study("600519.SS",event_date="2024-4-2", \
|
97
|
-
start='2024-3-15',end='2024-4-20', \
|
98
|
-
event_window=[0,1],post_event_days=7, \
|
99
|
-
method='CAPM', \
|
100
|
-
market_index='000001.SS',RF=0.0143)
|
101
|
-
|
102
|
-
es=event_study("600519.SS",event_date="2024-4-2", \
|
103
|
-
start='2024-3-1',end='2024-4-30', \
|
104
|
-
event_window=[0,0],post_event_days=7, \
|
105
|
-
method='market',market_index='000001.SS')
|
106
|
-
|
107
|
-
es=event_study("600519.SS",event_date="2024-4-2", \
|
108
|
-
start='2024-3-1',end='2024-4-30', \
|
109
|
-
event_window=[0,0],post_event_days=7, \
|
110
|
-
method='random walk')
|
111
|
-
|
112
|
-
|
113
|
-
def event_study(ticker,event_date, \
|
114
|
-
start='auto',end='auto', \
|
115
|
-
event_window=[1,3], \
|
116
|
-
post_event_days=0, \
|
117
|
-
method='CAPM', \
|
118
|
-
early_response_days=-2, \
|
119
|
-
estimation_window_days=-365, \
|
120
|
-
market_index='000300.SS', \
|
121
|
-
RF="market index", \
|
122
|
-
ret_type="Daily Adj Ret%", \
|
123
|
-
ticker_type='auto', \
|
124
|
-
show_AR='auto',show_RF=False,show_BHAR=False, \
|
125
|
-
draw_CAR=True,draw_BHAR=False, \
|
126
|
-
facecolor="whitesmoke",loc='best'):
|
127
|
-
"""
|
128
|
-
===========================================================================
|
129
|
-
功能:展示事件研究法的累计异常收益率CAR。
|
130
|
-
参数:
|
131
|
-
ticker:证券代码,可为股票、债券、基金、指数、国债收益率等。可为单个或多个。
|
132
|
-
event_date:事件发生日(注意时区的影响),以此日期为基期0,注意该日期可能在周末或假日。
|
133
|
-
注意:允许标注多个事件日,但仅以第一个事件日计算相关日期。
|
134
|
-
start/end:展示事件影响的起止日期,至少需要将事件日、事件窗口和事件后窗口包括在内,主要用于绘图。
|
135
|
-
注意:如果不绘制AR仅绘制CAR,事件窗口前CAR均为空,start日期在绘图中将不起作用。
|
136
|
-
event_window:事件窗口的起止日期,为相对事件日的相对日期
|
137
|
-
默认[0,0],即事件当日一天。注意窗口期不宜过长,因为过长的期间中可能混杂其他事件的影响。
|
138
|
-
注意:事件窗口不一定包括事件日(适用于事件日在非交易日的情形,例如周末或假日,或者在当日闭市后发生)
|
139
|
-
如果事件日为非交易日,事件窗口需要后移至事件日后的第一个交易日。
|
140
|
-
如果怀疑市场提前对事件发生反应,可以考虑前移事件窗口的开始日期。
|
141
|
-
使用CAR时,事件窗口长度一般为数日;使用BHAR时可长达数月。
|
142
|
-
post_event_days:用于分析事件窗口后的漂移效应,取事件窗口后多少天。
|
143
|
-
默认不分析,取0天。可以指定天数,注意是否跨过非交易日情形,过长的窗口期也可能混杂其他事件的影响。
|
144
|
-
method:估计事件窗口以及事件后窗口收益率预期值的方法
|
145
|
-
默认为CAPM(主要用于ticker为股票等),即通常所说的市场模型法。
|
146
|
-
如果ticker为股票等,也可直接使用指数收益率为其预期收益率,此时method为Market或Index,即常说的市场调整模型。
|
147
|
-
如果ticker为指数,无法再借助指数,method只能使用Random Walk,即使用前一个收益率为预期收益率。
|
148
|
-
注意:不管多个ticker时里面的不同证券类型,仅按第一个ticker的类型判断,并使用同一种证券类型。
|
149
|
-
使用CAR时,对每日异常收益率相加,反映短期逐日异常收益的累积;使用BHAR时则为复利累积,反映长期异常收益。
|
150
|
-
early_response_days:默认为-2,即提前2天市场就开始有反应。
|
151
|
-
市场很可能对事件提前发生反应(因为泄密等原因),例如中国市场规定上市公司董事会开完后两天内必须披露。
|
152
|
-
很可能刚开完董事会,市场就得到了消息。为规避这种情况对估计窗口的影响,可以调节此参数。
|
153
|
-
estimation_window_days:当method使用CAPM时,用于估计贝塔系数和截距项,以便计算预期收益率。
|
154
|
-
默认在事件窗口开始日期+提前反应天数前的365个自然日(约250个交易日)。
|
155
|
-
market_index:当method为CAPM时,用于计算市场收益率。默认中国市场采用000300.SS。
|
156
|
-
注意:需要根据不同市场采取不同的市场指数,例如香港市场为恒生指数,美国市场为标普500指数等。
|
157
|
-
RF:年化无风险收益率
|
158
|
-
默认使用市场模型"market index"自动计算,无需指定。
|
159
|
-
可直接指定具体数值。
|
160
|
-
也可指定特定指标替代,例如一年期中国国债收益率"1YCNY.B"或一年期美债收益率"1YUSY.B"等。
|
161
|
-
ticker_type:显式指明ticker的证券类型,当siat误判其类型(中国内地股票/债券/基金)时使用,默认'auto'。
|
162
|
-
show_RF:在使用市场模型或指定指标时是否显示计算出的RF均值,默认为False。
|
163
|
-
show_AR:是否绘图时绘制异常收益率AR
|
164
|
-
默认'auto'(单个ticker时绘制,多个时不绘制)。
|
165
|
-
也可指定True/False强行绘制/不绘制。
|
166
|
-
show_BHAR;是否显示BHAR数值,适用于长期窗口,默认否False。
|
167
|
-
draw_CAR:是否绘制CAR曲线,默认是True。
|
168
|
-
draw_BHAR:是否绘制BHAR曲线,默认否False。
|
169
|
-
注意:对于短期窗口,CAR曲线与BHAR曲线差异微小,可能基本重合,因此建议仅绘制其中之一。
|
170
|
-
facecolor:显式指定绘图背景颜色,默认"whitesmoke"。
|
171
|
-
|
172
|
-
|
173
|
-
示例:美的收购库卡事件对股价的影响
|
174
|
-
es=event_study(["000333.SZ"],
|
175
|
-
event_date="2021-11-24",
|
176
|
-
start='2021-11-20',end='2021-12-25',
|
177
|
-
event_window=[1,10],
|
178
|
-
post_event_days=15,
|
179
|
-
method='CAPM',
|
180
|
-
market_index='399001.SZ')
|
181
|
-
"""
|
182
|
-
|
183
|
-
DEBUG=False
|
184
|
-
DEBUG2=False
|
185
|
-
|
186
|
-
#=====事件研究各个日期的计算与调整===========================================
|
187
|
-
if isinstance(event_date,str):
|
188
|
-
event_date=[date_adjust(event_date,adjust=0)]
|
189
|
-
elif isinstance(event_date,list):
|
190
|
-
event_date=[date_adjust(ed,adjust=0) for ed in event_date]
|
191
|
-
else:
|
192
|
-
print(" #Warning(event_study): invalid date or list of dates {}".format(event_date))
|
193
|
-
return None
|
194
|
-
event_date.sort() #升序排序
|
195
|
-
|
196
|
-
#事件窗口日期:遇到周末需要调整,提前或顺延至最近的工作日
|
197
|
-
event_window_new=event_window.copy() #列表的普通赋值仅为指针,新列表的改动也会影响原列表
|
198
|
-
adjust_start=0
|
199
|
-
event_window_start=date_adjust(event_date[0],adjust=event_window[0])
|
200
|
-
if week_day(event_window_start) == 5: #周六
|
201
|
-
if event_window[0] >= 0:
|
202
|
-
adjust_start=2
|
203
|
-
else:
|
204
|
-
adjust_start=-1
|
205
|
-
elif week_day(event_window_start) == 6: #周日
|
206
|
-
if event_window[0] >= 0:
|
207
|
-
adjust_start=1
|
208
|
-
else:
|
209
|
-
adjust_start=-2
|
210
|
-
event_window_start=date_adjust(event_window_start,adjust=adjust_start)
|
211
|
-
event_window_new[0]=event_window[0]+adjust_start
|
212
|
-
|
213
|
-
adjust_end=0
|
214
|
-
event_window_end=date_adjust(event_window_start,adjust=event_window[1]-event_window[0])
|
215
|
-
if week_day(event_window_end) == 5: #周六
|
216
|
-
if event_window[1] >= 0:
|
217
|
-
adjust_end=2
|
218
|
-
else:
|
219
|
-
adjust_end=-1
|
220
|
-
elif week_day(event_window_end) == 6: #周日
|
221
|
-
if event_window[1] >= 0:
|
222
|
-
adjust_end=1
|
223
|
-
else:
|
224
|
-
adjust_end=-2
|
225
|
-
event_window_end=date_adjust(event_window_end,adjust=adjust_end)
|
226
|
-
event_window_new[1]=event_window[1]+adjust_start+adjust_end
|
227
|
-
|
228
|
-
if DEBUG:
|
229
|
-
print(" DEBUG: event window is between {0} to {1}".format(event_window_start,event_window_end))
|
230
|
-
|
231
|
-
if event_window_new != event_window:
|
232
|
-
print(" #Notice: event window adjusted from {0} to {1} because of weekend".format(event_window,event_window_new))
|
233
|
-
|
234
|
-
#事件后窗口日期
|
235
|
-
post_event_start=date_adjust(event_window_end,adjust=0)
|
236
|
-
if week_day(post_event_start) == 5: #周六
|
237
|
-
post_event_start=date_adjust(post_event_start,adjust=2)
|
238
|
-
elif week_day(post_event_start) == 6: #周日
|
239
|
-
post_event_start=date_adjust(post_event_start,adjust=1)
|
240
|
-
|
241
|
-
post_event_end=date_adjust(post_event_start,adjust=post_event_days)
|
242
|
-
if week_day(post_event_end) == 5: #周六
|
243
|
-
post_event_end=date_adjust(post_event_end,adjust=2)
|
244
|
-
elif week_day(post_event_end) == 6: #周日
|
245
|
-
post_event_end=date_adjust(post_event_end,adjust=1)
|
246
|
-
|
247
|
-
if post_event_end > stoday:
|
248
|
-
post_event_end = stoday
|
249
|
-
|
250
|
-
if DEBUG:
|
251
|
-
print(" DEBUG: post event window is between {0} to {1}".format(post_event_start,post_event_end))
|
252
|
-
|
253
|
-
|
254
|
-
#事件窗口前日期
|
255
|
-
event_eve_date=date_adjust(event_window_start,adjust=-1)
|
256
|
-
if week_day(event_eve_date) == 5: #周六
|
257
|
-
event_eve_date=date_adjust(event_eve_date,adjust=-1)
|
258
|
-
elif week_day(event_eve_date) == 6: #周日
|
259
|
-
event_eve_date=date_adjust(event_eve_date,adjust=-2)
|
260
|
-
|
261
|
-
if DEBUG:
|
262
|
-
print(" DEBUG: event eve is on {}".format(event_eve_date))
|
263
|
-
|
264
|
-
#提前反应日期
|
265
|
-
early_response_date=date_adjust(event_date[0],adjust=early_response_days)
|
266
|
-
if week_day(early_response_date) == 5: #周六
|
267
|
-
early_response_date=date_adjust(early_response_date,adjust=-1)
|
268
|
-
elif week_day(early_response_date) == 6: #周日
|
269
|
-
early_response_date=date_adjust(early_response_date,adjust=-2)
|
270
|
-
|
271
|
-
if DEBUG:
|
272
|
-
print(" DEBUG: early response started on {}".format(early_response_date))
|
273
|
-
|
274
|
-
#估计窗口日期的计算
|
275
|
-
est_window_end=date_adjust(early_response_date,adjust=-1)
|
276
|
-
est_window_start=date_adjust(est_window_end,adjust=estimation_window_days)
|
277
|
-
if DEBUG:
|
278
|
-
print(" DEBUG: regression period starts from {0} to {1}".format(est_window_start,est_window_end))
|
279
|
-
|
280
|
-
#处理绘图时显示的日期范围
|
281
|
-
if start=='auto':
|
282
|
-
start=date_adjust(early_response_date,adjust=-7)
|
283
|
-
if end=='auto':
|
284
|
-
if len(ticker) == 1 or show_AR:
|
285
|
-
end=date_adjust(post_event_end,adjust=7)
|
286
|
-
else:
|
287
|
-
end=date_adjust(post_event_end,adjust=2)
|
288
|
-
|
289
|
-
#=====判断ticker是否为指数,调整预期收益率计算方法============================
|
290
|
-
if isinstance(ticker,str):
|
291
|
-
ticker=[ticker]
|
292
|
-
elif isinstance(ticker,list):
|
293
|
-
ticker=ticker
|
294
|
-
else:
|
295
|
-
print(" #Warning(event_study): unexpected type of ticker {}".format(ticker))
|
296
|
-
return None
|
297
|
-
|
298
|
-
if market_index in ticker:
|
299
|
-
print(" #Warning(event_study): market_index {0} duplicated in and removed from ticker {1}".format(market_index,ticker))
|
300
|
-
ticker.remove(market_index)
|
301
|
-
|
302
|
-
#tname=ticker_name(ticker[0],ticker_type)
|
303
|
-
#检查ticker是否为指数或国债收益率
|
304
|
-
"""
|
305
|
-
if ("指数" in tname or "index" in tname.lower()) or ("收益率" in tname or "yield" in tname.lower()):
|
306
|
-
if not ("random" in method.lower() or "walk" in method.lower()):
|
307
|
-
print(" #Notice: check the applicability of ticker {0}, method {1} with market index {2}".format(ticker[0],method,market_index))
|
308
|
-
"""
|
309
|
-
|
310
|
-
#=====获取证券价格和/或相关指数数据==========================================
|
311
|
-
#基于CAPM获取数据
|
312
|
-
if 'capm' in method.lower():
|
313
|
-
method_type="capm"
|
314
|
-
df_ret=compare_msecurity(tickers=ticker+[market_index],measure=ret_type, \
|
315
|
-
start=est_window_start,end=end, \
|
316
|
-
ticker_type=ticker_type, \
|
317
|
-
graph=False)
|
318
|
-
|
319
|
-
if isinstance(RF,int) or isinstance(RF,float):
|
320
|
-
#RF为具体数值
|
321
|
-
RF_type="value"
|
322
|
-
|
323
|
-
elif "market" in (str(RF)).lower() or "index" in (str(RF)).lower():
|
324
|
-
#RF通过市场模型计算,无需指定
|
325
|
-
RF_type="model"
|
326
|
-
else:
|
327
|
-
#指定RF代码,例如1YCNY.B,注意1:得到的是年化收益率%,注意2:中国的只有近一年的数据
|
328
|
-
RF_type="code"
|
329
|
-
|
330
|
-
if RF_type=="code":
|
331
|
-
df_rf=compare_msecurity(tickers=RF,measure='Close', \
|
332
|
-
start=est_window_start,end=end, \
|
333
|
-
graph=False)
|
334
|
-
RF=df_rf[list(df_rf)[0]].mean() / 100.0
|
335
|
-
|
336
|
-
#基于市场指数获取数据
|
337
|
-
elif 'market' in method.lower() or 'index' in method.lower():
|
338
|
-
method_type="market"
|
339
|
-
df_ret=compare_msecurity(tickers=ticker+[market_index],measure=ret_type, \
|
340
|
-
start=est_window_start,end=end, \
|
341
|
-
ticker_type=ticker_type, \
|
342
|
-
graph=False)
|
343
|
-
|
344
|
-
elif 'random' in method.lower() or 'walk' in method.lower():
|
345
|
-
method_type="random"
|
346
|
-
df_ret=compare_msecurity(tickers=ticker,measure=ret_type, \
|
347
|
-
start=est_window_start,end=end, \
|
348
|
-
ticker_type=ticker_type, \
|
349
|
-
graph=False)
|
350
|
-
for t in ticker_name(ticker,ticker_type):
|
351
|
-
try:
|
352
|
-
df_ret[t+"_predicted"]=df_ret[t].shift(1)
|
353
|
-
except:
|
354
|
-
#print(" #Warning(event_study): info not found for ticker {}".format(t))
|
355
|
-
continue
|
356
|
-
|
357
|
-
else:
|
358
|
-
print(" #Warning(event_study): unexpected type of AR method {}".format(method))
|
359
|
-
return None
|
360
|
-
|
361
|
-
#=====计算异常收益率AR=====
|
362
|
-
df_cols=list(df_ret)
|
363
|
-
if method_type=='market':
|
364
|
-
for t in ticker_name(ticker,ticker_type):
|
365
|
-
try:
|
366
|
-
df_ret[t+'_AR']=df_ret[t] - df_ret[ticker_name(market_index)]
|
367
|
-
except: continue
|
368
|
-
|
369
|
-
elif method_type=='random':
|
370
|
-
for t in ticker_name(ticker,ticker_type):
|
371
|
-
try:
|
372
|
-
df_ret[t+'_AR']=df_ret[t] - df_ret[t+"_predicted"]
|
373
|
-
except: continue
|
374
|
-
|
375
|
-
else: #按CAPM计算
|
376
|
-
#CAPM回归期间数据
|
377
|
-
est_window_startpd=pd.to_datetime(est_window_start)
|
378
|
-
est_window_endpd =pd.to_datetime(est_window_end)
|
379
|
-
df_reg=df_ret[(df_ret.index >=est_window_startpd) & (df_ret.index <=est_window_endpd)].copy()
|
380
|
-
|
381
|
-
#删除空缺值,否则回归会出错
|
382
|
-
df_reg=df_reg.replace([np.nan, None], np.nan).dropna()
|
383
|
-
|
384
|
-
import statsmodels.api as sm
|
385
|
-
if RF_type in ["value","code"]:
|
386
|
-
if not ("%" in ret_type): #注意:RF是年化收益率(需要转化为日收益率),这里不是百分比
|
387
|
-
X=df_reg[ticker_name(market_index)] - RF/365.0 #无截距项回归,指定RF具体数值
|
388
|
-
else:
|
389
|
-
X=df_reg[ticker_name(market_index)] - RF/365.0 * 100.0 #这里需要转化为日收益率百分比%
|
390
|
-
|
391
|
-
else: #RF_type=="model"
|
392
|
-
X=df_reg[ticker_name(market_index)]
|
393
|
-
X=sm.add_constant(X) #有截距项回归,基于市场模型
|
394
|
-
|
395
|
-
if DEBUG:
|
396
|
-
print(" DEBUG: method_type={0}, RF_type={1}, RF={2}".format(method_type,RF_type,RF))
|
397
|
-
|
398
|
-
#CAPM回归
|
399
|
-
beta_dict={}; intercept_dict={}; pvalue_dict={}; rf_dict={}
|
400
|
-
for t in ticker_name(ticker,ticker_type):
|
401
|
-
try:
|
402
|
-
if RF_type in ["value","code"]:
|
403
|
-
if not ("%" in ret_type): #注意:RF是年化收益率(需要转化为日收益率),不是百分比
|
404
|
-
y=df_reg[t] - RF/365.0
|
405
|
-
else:
|
406
|
-
y=df_reg[t] - RF/365.0 * 100.0
|
407
|
-
|
408
|
-
else: #RF_type=="model"
|
409
|
-
y=df_reg[t]
|
410
|
-
except: continue
|
411
|
-
|
412
|
-
model = sm.OLS(y,X) #定义回归模型y=X
|
413
|
-
results = model.fit() #进行OLS回归
|
414
|
-
|
415
|
-
if DEBUG2:
|
416
|
-
print(" DEBUG: RF_type={0}, results.params={1},results.pvalues={2}".format(RF_type,results.params,results.pvalues))
|
417
|
-
|
418
|
-
#提取回归系数,详细信息见results.summary()
|
419
|
-
if RF_type=="model":
|
420
|
-
intercept=results.params[0]
|
421
|
-
beta=results.params[1]; pvalue=results.pvalues[1]
|
422
|
-
try:
|
423
|
-
#此处回归得到的rf应该为日收益率,转为年化收益率。
|
424
|
-
#注意:不同证券回归出的结果可能差异较大,原因可能是混入了回归残差!
|
425
|
-
if not ("%" in ret_type):
|
426
|
-
rf=intercept / (1-beta) * 365.0
|
427
|
-
else:
|
428
|
-
rf=intercept / (1-beta) / 100.0 * 365.0
|
429
|
-
except: rf=0
|
430
|
-
|
431
|
-
else: #RF_type in ["value","code"]
|
432
|
-
intercept=0
|
433
|
-
beta=results.params[0]; pvalue=results.pvalues[0]
|
434
|
-
rf=RF
|
435
|
-
|
436
|
-
beta_dict[t] = beta; intercept_dict[t] = intercept; pvalue_dict[t] = pvalue; rf_dict[t]=rf
|
437
|
-
if DEBUG2:
|
438
|
-
print(" DEBUG: t={0}, intercept={1}, beta={2}, pvalue={3}, annualized rf={4}".format(t,round(intercept,4),round(beta,4),round(pvalue,4),round(rf,4)))
|
439
|
-
|
440
|
-
#计算收益率预期和AR
|
441
|
-
for t in ticker_name(ticker,ticker_type):
|
442
|
-
try:
|
443
|
-
if RF_type in ["value","code"]:
|
444
|
-
#CAPM模型:E(R) = RF + (Rm-RF)*beta
|
445
|
-
RF_text=str(round(RF*100.0,4))[:6]+'%'
|
446
|
-
if not ("%" in ret_type): #注意:RF是年化收益率,此处不是百分比
|
447
|
-
df_ret[t+"_predicted"]=(df_ret[ticker_name(market_index)] - RF/365.0)*beta_dict[t] + RF/365.0
|
448
|
-
else:
|
449
|
-
df_ret[t+"_predicted"]=(df_ret[ticker_name(market_index)] - RF*100.0/365.0)*beta_dict[t] + RF*100.0/365.0
|
450
|
-
|
451
|
-
else: #RF_type=="model"
|
452
|
-
#市场模型:E(R) = intercept + Rm*beta
|
453
|
-
RF_text="基于市场模型回归"
|
454
|
-
df_ret[t+"_predicted"]=df_ret[ticker_name(market_index)]*beta_dict[t] + intercept_dict[t]
|
455
|
-
|
456
|
-
df_ret[t+"_AR"]=df_ret[t] - df_ret[t+"_predicted"]
|
457
|
-
except: continue
|
458
|
-
|
459
|
-
if DEBUG2:
|
460
|
-
print(" DEBUG: RF_type={0}, RF_text={1}, rf_dict={2}".format(RF_type,RF_text, rf_dict))
|
461
|
-
|
462
|
-
#=====计算CAR和BHAR==============================================================
|
463
|
-
for t in ticker_name(ticker,ticker_type):
|
464
|
-
try:
|
465
|
-
df_ret[t+"_CAR"]=0
|
466
|
-
df_ret[t+"_BHAR"]=0
|
467
|
-
except: continue
|
468
|
-
|
469
|
-
event_window_startpd=pd.to_datetime(event_window_start)
|
470
|
-
event_window_endpd=pd.to_datetime(event_window_end)
|
471
|
-
post_event_endpd=pd.to_datetime(post_event_end)
|
472
|
-
startpd=pd.to_datetime(start); endpd=pd.to_datetime(end)
|
473
|
-
|
474
|
-
#计算CAR和BHAR
|
475
|
-
df_ret_event=df_ret[(df_ret.index >=event_window_startpd) & (df_ret.index <=endpd)]
|
476
|
-
for t in ticker_name(ticker,ticker_type):
|
477
|
-
try:
|
478
|
-
# CAR:单利累加求和(每日异常收益相加)
|
479
|
-
df_ret_event[t+'_CAR'] = df_ret_event[t+'_AR'].cumsum(skipna=True)
|
480
|
-
# BHAR:复利累积
|
481
|
-
df_ret_event[t+'_BHAR'] = ((1+df_ret_event[t+'_AR']/100).cumprod()-1)*100
|
482
|
-
except: continue
|
483
|
-
|
484
|
-
#合成事件前期间
|
485
|
-
df_ret_before_event=df_ret[(df_ret.index >=startpd) & (df_ret.index < event_window_startpd)]
|
486
|
-
for t in ticker_name(ticker,ticker_type):
|
487
|
-
try:
|
488
|
-
df_ret_before_event[t+'_CAR']=np.nan
|
489
|
-
df_ret_before_event[t+'_BHAR']=np.nan
|
490
|
-
except: continue
|
491
|
-
|
492
|
-
df_show=pd.concat([df_ret_before_event,df_ret_event])
|
493
|
-
|
494
|
-
#是否显示AR:默认单证券显示,多证券时不显示
|
495
|
-
df_show_cols=[]
|
496
|
-
for c in list(df_show):
|
497
|
-
if show_AR=='auto':
|
498
|
-
if len(ticker)==1:
|
499
|
-
if 'AR' in c or 'CAR' in c:
|
500
|
-
df_show_cols=df_show_cols+[c]
|
501
|
-
show_AR=True
|
502
|
-
else:
|
503
|
-
if 'CAR' in c:
|
504
|
-
df_show_cols=df_show_cols+[c]
|
505
|
-
show_AR=False
|
506
|
-
elif show_AR==True:
|
507
|
-
if 'AR' in c or 'CAR' in c:
|
508
|
-
df_show_cols=df_show_cols+[c]
|
509
|
-
else: #show_AR==False
|
510
|
-
if 'CAR' in c:
|
511
|
-
df_show_cols=df_show_cols+[c]
|
512
|
-
|
513
|
-
df_show2=df_show[df_show_cols]
|
514
|
-
|
515
|
-
#=====绘图=================================================================
|
516
|
-
#设置标签
|
517
|
-
df0=df_show2
|
518
|
-
|
519
|
-
y_label="收益率%"
|
520
|
-
|
521
|
-
#横轴注释
|
522
|
-
footnote1="首事件日{0},事件窗口{1},事件后窗口天数{2},市场提前反应天数{3}".format(event_date[0],event_window_new,post_event_days,early_response_days)
|
523
|
-
footnote2="收益率类型:"+ectranslate(ret_type)
|
524
|
-
|
525
|
-
if method_type == "market":
|
526
|
-
method_name="市场指数基准"
|
527
|
-
elif method_type == "random":
|
528
|
-
method_name="随机漫步模型"
|
529
|
-
else:
|
530
|
-
method_name="CAPM模型"
|
531
|
-
|
532
|
-
footnote3=",收益率预期方法:"+method_name
|
533
|
-
if not method_type == "random":
|
534
|
-
footnote4=',市场指数:'+ticker_name(market_index)
|
535
|
-
else:
|
536
|
-
footnote4=''
|
537
|
-
|
538
|
-
#显著性检验:异于零的t检验,事件窗口
|
539
|
-
df_event_window=df0[(df0.index >=event_window_start) & (df0.index <=event_window_end)]
|
540
|
-
#footnote5="事件窗口CAR(终值,p值):"
|
541
|
-
footnote5="事件窗口CAR(终值,均值,中位数,p值):"
|
542
|
-
for c in list(df_event_window):
|
543
|
-
if 'CAR' in c.upper():
|
544
|
-
c_name=c[:-4]
|
545
|
-
|
546
|
-
event_window_endpd=pd.to_datetime(event_window_end)
|
547
|
-
#car_value=df_event_window[df_event_window.index == event_window_endpd][c].values[0]
|
548
|
-
car_value=df_event_window[c][-1]
|
549
|
-
car_mean=df_event_window[c].mean()
|
550
|
-
car_median=df_event_window[c].median()
|
551
|
-
|
552
|
-
if car_value > 0:
|
553
|
-
car_value_str=str(round(car_value,4))[:6]
|
554
|
-
else:
|
555
|
-
car_value_str=str(round(car_value,4))[:7]
|
556
|
-
|
557
|
-
if car_mean > 0:
|
558
|
-
car_mean_str=str(round(car_mean,4))[:6]
|
559
|
-
else:
|
560
|
-
car_mean_str=str(round(car_mean,4))[:7]
|
561
|
-
|
562
|
-
if car_median > 0:
|
563
|
-
car_median_str=str(round(car_median,4))[:6]
|
564
|
-
else:
|
565
|
-
car_median_str=str(round(car_median,4))[:7]
|
566
|
-
|
567
|
-
if len(df_event_window[c])==1:
|
568
|
-
if abs(df_event_window[c].values[0]) > 0.01:
|
569
|
-
p_value=0.0
|
570
|
-
else:
|
571
|
-
p_value=1.0
|
572
|
-
else:
|
573
|
-
p_value=ttest(df_event_window[c],0)
|
574
|
-
if p_value > 0:
|
575
|
-
p_value_str=str(round(p_value,4))[:6]
|
576
|
-
else:
|
577
|
-
p_value_str=str(round(p_value,4))[:7]
|
578
|
-
#footnote5=footnote5+c_name+p_value_str+","
|
579
|
-
#footnote5=footnote5+"{0}({1}, {2}), ".format(c_name,car_value_str,p_value_str)
|
580
|
-
footnote5=footnote5+"{0}({1}, {2}, {3}, {4}), ".format(c_name,car_value_str,car_mean_str, \
|
581
|
-
car_median_str,p_value_str)
|
582
|
-
|
583
|
-
if 'BHAR' in c.upper():
|
584
|
-
bhar_value=df_event_window[c][-1]
|
585
|
-
if bhar_value > 0:
|
586
|
-
bhar_value_str=str(round(bhar_value,4))[:6]
|
587
|
-
else:
|
588
|
-
bhar_value_str=str(round(bhar_value,4))[:7]
|
589
|
-
|
590
|
-
if show_BHAR:
|
591
|
-
footnote5=footnote5+"BHAR终值: {0}; ".format(bhar_value_str)
|
592
|
-
|
593
|
-
footnote5=footnote5.strip(", "); footnote5=footnote5.strip("; ")
|
594
|
-
|
595
|
-
#显著性检验:异于零的t检验,事件后窗口
|
596
|
-
df_post_event_window=df0[(df0.index >event_window_end) & (df0.index <=post_event_end)]
|
597
|
-
if len(df_post_event_window) == 0:
|
598
|
-
footnote6=''
|
599
|
-
elif len(df_post_event_window) == 0:
|
600
|
-
footnote6=''
|
601
|
-
else:
|
602
|
-
#footnote6="事件后窗口CAR(终值,p值):"
|
603
|
-
footnote6="事件后窗口CAR(终值,均值,中位数,p值):"
|
604
|
-
for c in list(df_post_event_window):
|
605
|
-
if 'CAR' in c.upper():
|
606
|
-
c_name=c[:-4]
|
607
|
-
post_event_endpd=pd.to_datetime(post_event_end)
|
608
|
-
if DEBUG2:
|
609
|
-
print(" DEBUG: c={0},post_event_end={1},df_post_event_window={2}".format(c,post_event_end,df_post_event_window))
|
610
|
-
#car_value=df_post_event_window[df_post_event_window.index == post_event_endpd][c].values[0]
|
611
|
-
car_value=df_post_event_window[c][-1]
|
612
|
-
car_mean=df_post_event_window[c].mean()
|
613
|
-
car_median=df_post_event_window[c].median()
|
614
|
-
|
615
|
-
if car_value > 0:
|
616
|
-
car_value_str=str(round(car_value,4))[:6]
|
617
|
-
else:
|
618
|
-
car_value_str=str(round(car_value,4))[:7]
|
619
|
-
|
620
|
-
if car_mean > 0:
|
621
|
-
car_mean_str=str(round(car_mean,4))[:6]
|
622
|
-
else:
|
623
|
-
car_mean_str=str(round(car_mean,4))[:7]
|
624
|
-
|
625
|
-
if car_median > 0:
|
626
|
-
car_median_str=str(round(car_median,4))[:6]
|
627
|
-
else:
|
628
|
-
car_median_str=str(round(car_median,4))[:7]
|
629
|
-
|
630
|
-
if len(df_post_event_window[c])==1:
|
631
|
-
if abs(df_post_event_window[c].values[0]) > 0.01:
|
632
|
-
p_value=0.0
|
633
|
-
else:
|
634
|
-
p_value=1.0
|
635
|
-
else:
|
636
|
-
p_value=ttest(df_post_event_window[c],0)
|
637
|
-
if p_value > 0:
|
638
|
-
p_value_str=str(round(p_value,4))[:6]
|
639
|
-
else:
|
640
|
-
p_value_str=str(round(p_value,4))[:7]
|
641
|
-
|
642
|
-
#footnote6=footnote6+c[:-4]+str(p_value)[:6]+","
|
643
|
-
footnote6=footnote6+"{0}({1}, {2}, {3}, {4}), ".format(c_name,car_value_str,car_mean_str,car_median_str,p_value_str)
|
644
|
-
|
645
|
-
if 'BHAR' in c.upper():
|
646
|
-
bhar_value=df_post_event_window[c][-1]
|
647
|
-
if bhar_value > 0:
|
648
|
-
bhar_value_str=str(round(bhar_value,4))[:6]
|
649
|
-
else:
|
650
|
-
bhar_value_str=str(round(bhar_value,4))[:7]
|
651
|
-
|
652
|
-
if show_BHAR:
|
653
|
-
footnote6=footnote6+"BHAR终值: {0}; ".format(bhar_value_str)
|
654
|
-
|
655
|
-
footnote6=footnote6.strip(", "); footnote6=footnote6.strip("; ")
|
656
|
-
|
657
|
-
footnote7="数据来源:Sina/EM/Yahoo/Stooq/SWHY,"+stoday
|
658
|
-
|
659
|
-
#x_label=footnote1+'\n'+footnote2+footnote3+footnote4+'\n'+footnote5+'\n'+footnote6+'\n'+footnote7
|
660
|
-
x_label=footnote1+'\n'+footnote2+footnote3+footnote4+'\n'+footnote7
|
661
|
-
|
662
|
-
|
663
|
-
axhline_value=0
|
664
|
-
axhline_label="零线"
|
665
|
-
title_txt="事件影响分析:"
|
666
|
-
for t in ticker_name(ticker,ticker_type):
|
667
|
-
title_txt=title_txt+t+','
|
668
|
-
title_txt=title_txt.strip(",")
|
669
|
-
|
670
|
-
#判断最新可获得日期
|
671
|
-
last_date=df0.index[-1].strftime("%Y-%m-%d")
|
672
|
-
if DEBUG:
|
673
|
-
print(" DEBUG: last_date={}".format(last_date))
|
674
|
-
if post_event_end > last_date:
|
675
|
-
post_event_end = last_date
|
676
|
-
|
677
|
-
if event_window_new[0] != event_window_new[1]:
|
678
|
-
attention_point_area=[event_window_start,event_window_end]
|
679
|
-
else:
|
680
|
-
attention_point_area=[event_window_start,post_event_end]
|
681
|
-
|
682
|
-
#去掉重复日期项标注且不改变顺序
|
683
|
-
event_date_new=[]
|
684
|
-
for d in event_date:
|
685
|
-
d_new=date_adjust(d,adjust=0)
|
686
|
-
event_date_new=event_date_new+[d_new]
|
687
|
-
|
688
|
-
attention_point=[event_eve_date,event_window_start,event_window_end,post_event_end]+event_date_new
|
689
|
-
if not show_AR:
|
690
|
-
period_days=calculate_days(event_eve_date,post_event_end)
|
691
|
-
if DEBUG:
|
692
|
-
print(" DEBUG: period_days={}".format(period_days))
|
693
|
-
|
694
|
-
if period_days< 6:
|
695
|
-
#绘图时横轴若少于6天会出现时间刻度,易误导需避免
|
696
|
-
draw_start_date=date_adjust(event_eve_date,adjust=period_days-6)
|
697
|
-
attention_point=[draw_start_date,event_window_start,event_window_end,post_event_end]+event_date_new
|
698
|
-
"""
|
699
|
-
if show_AR:
|
700
|
-
attention_point=[event_eve_date,event_window_start,event_window_end,post_event_end]+event_date_new
|
701
|
-
else:
|
702
|
-
attention_point=[event_eve_date,event_window_start,event_window_end,post_event_end]+event_date_new
|
703
|
-
df0=df0[(df0.index >= start) & (df0.index <=post_event_end)]
|
704
|
-
"""
|
705
|
-
attention_point.sort(reverse=False)
|
706
|
-
attention_point=list({}.fromkeys(attention_point).keys())
|
707
|
-
|
708
|
-
# 是否绘制CAR或BHAR曲线:对于短期窗口,CAR曲线和BHAR曲线很可能基本重合,建议仅绘制其中之一!
|
709
|
-
df0draw=df0.copy()
|
710
|
-
for c in list(df0draw):
|
711
|
-
if not draw_CAR and 'CAR' in c:
|
712
|
-
del df0draw[c]
|
713
|
-
if not draw_BHAR and 'BHAR' in c:
|
714
|
-
del df0draw[c]
|
715
|
-
|
716
|
-
#绘图
|
717
|
-
draw_lines(df0draw,y_label,x_label,axhline_value,axhline_label,title_txt, \
|
718
|
-
data_label=False, \
|
719
|
-
loc=loc,resample_freq='D',smooth=False, \
|
720
|
-
annotate=True,annotate_value=False, \
|
721
|
-
attention_point=attention_point, \
|
722
|
-
attention_point_area=attention_point_area, \
|
723
|
-
ticker_type=ticker_type,facecolor=facecolor)
|
724
|
-
|
725
|
-
#=====输出AR和/或CAR或BHAR表格====================================================
|
726
|
-
df1=df0.copy()
|
727
|
-
#df1=df1.replace([np.nan, None], np.nan).dropna()
|
728
|
-
df1=df1.replace([np.nan, None],'-')
|
729
|
-
df1["日期"]=df1.index
|
730
|
-
df1["日期"]=df1["日期"].apply(lambda x: x.strftime("%Y-%m-%d"))
|
731
|
-
|
732
|
-
df1=df1[(df1["日期"] >= event_date[0]) & (df1["日期"] <= post_event_end)]
|
733
|
-
df1["星期"]=df1["日期"].apply(lambda x: week_day(x)+1)
|
734
|
-
|
735
|
-
df1["事件标记"]=''
|
736
|
-
for d in event_date_new:
|
737
|
-
if len(event_date_new)==1:
|
738
|
-
event_text="事件日"
|
739
|
-
else:
|
740
|
-
pos=event_date_new.index(d)
|
741
|
-
if pos==1:
|
742
|
-
event_text="首事件日"
|
743
|
-
else:
|
744
|
-
event_text="事件日"+str(pos+1)
|
745
|
-
df1["事件标记"]=df1.apply(lambda x: event_text if x["日期"]==d else x["事件标记"],axis=1)
|
746
|
-
|
747
|
-
#event_text=",事件窗口开始日"
|
748
|
-
event_text="\n事件窗开始"
|
749
|
-
df1["事件标记"]=df1.apply(lambda x: x["事件标记"]+event_text if x["日期"]==event_window_start else x["事件标记"],axis=1)
|
750
|
-
#event_text=",事件窗口结束日"
|
751
|
-
event_text="\n事件窗结束"
|
752
|
-
df1["事件标记"]=df1.apply(lambda x: x["事件标记"]+event_text if x["日期"]==event_window_end else x["事件标记"],axis=1)
|
753
|
-
|
754
|
-
#event_text=",事件后窗口结束日"
|
755
|
-
if post_event_end > event_window_end:
|
756
|
-
event_text="\n事件后窗结束"
|
757
|
-
df1["事件标记"]=df1.apply(lambda x: x["事件标记"]+event_text if x["日期"]==post_event_end else x["事件标记"],axis=1)
|
758
|
-
|
759
|
-
event_text="\n事件窗"
|
760
|
-
df1["事件标记"]=df1.apply(lambda x: x["事件标记"]+event_text if (x["日期"] > event_window_start) and (x["日期"] < event_window_end) else x["事件标记"],axis=1)
|
761
|
-
|
762
|
-
event_text="\n事件后窗"
|
763
|
-
df1["事件标记"]=df1.apply(lambda x: x["事件标记"]+event_text if (x["日期"] > event_window_end) and (x["日期"] < post_event_end) else x["事件标记"],axis=1)
|
764
|
-
|
765
|
-
df1["事件标记"]=df1["事件标记"].apply(lambda x: x.strip('\n'))
|
766
|
-
|
767
|
-
#显示表格
|
768
|
-
df0_list=list(df0)
|
769
|
-
df1_list=["事件标记","日期","星期"]+df0_list
|
770
|
-
df1=df1[df1_list]
|
771
|
-
#title_txt=title_txt+",窗口收益率"
|
772
|
-
|
773
|
-
if "CAPM" in method.upper():
|
774
|
-
footnotex="CAPM回归期间:{0}至{1},无风险收益率{2}".format(est_window_start,est_window_end,RF_text)
|
775
|
-
footnotey="CAPM贝塔系数:"
|
776
|
-
for k in beta_dict:
|
777
|
-
footnotey=footnotey+k+str(round(beta_dict[k],4))[:6]+","
|
778
|
-
footnotey=footnotey.strip(",")
|
779
|
-
|
780
|
-
if show_RF:
|
781
|
-
footnotez="无风险收益率均值:"
|
782
|
-
for r in rf_dict:
|
783
|
-
footnotez=footnotez+r+str(round(rf_dict[r]*100.0,4))[:6]+"%, "
|
784
|
-
|
785
|
-
footnotez=footnotez.strip(", ")
|
786
|
-
footnote=footnote2+footnote3+footnote4+'\n'+footnotex+'\n'+footnotey+'\n'+footnotez+'\n'+footnote5+'\n'+footnote6
|
787
|
-
else:
|
788
|
-
footnote=footnote2+footnote3+footnote4+'\n'+footnotex+'\n'+footnotey+'\n'+footnote5+'\n'+footnote6
|
789
|
-
else:
|
790
|
-
footnote=footnote2+footnote3+footnote4+'\n'+footnote5+'\n'+footnote6
|
791
|
-
|
792
|
-
for c in list(df1):
|
793
|
-
if not show_BHAR and 'BHAR' in c:
|
794
|
-
del df1[c]
|
795
|
-
|
796
|
-
#显示结果表格
|
797
|
-
df_display_CSS(df1,titletxt=title_txt,footnote=footnote,facecolor=facecolor,decimals=4, \
|
798
|
-
first_col_align='left',second_col_align='left', \
|
799
|
-
last_col_align='center',other_col_align='center')
|
800
|
-
|
801
|
-
|
802
|
-
return df_show2
|
803
|
-
|
804
|
-
|
805
|
-
|
806
|
-
|
807
|
-
|
808
|
-
|
809
|
-
|
810
|
-
|
811
|
-
|
812
|
-
|
813
|
-
|
814
|
-
|
815
|
-
|