因论文要求，需要建构股价信息效率指标
参照外文，以Hou的方式建构
国内也有很多中文期刊学者加以引用这个做法建构股价信息效率
研究了一番
我理解的是国内学者大多用日数据，并且用滞后四个交易日市场收益率做回归
分组算平均 后得到以月为单位的股价信息效率数据
(文章下有附上我参考的文章，是引用李志生老师的文献)
目前我所做的delay1 delay2有些卡住了
用sum的结果是最小值出现了负号
但过往的文献都是正数
在此附上我的指令 希望老师们帮我看看问题
delay1
- gen mydate=date(tradeweek,"YMD")
- format mydate %tddmy
- gen m = month(mydate)
- gen w = week(mydate)
- rename m month
- rename w week
- order stkcd year month
- *滞后四期
- bys stkcd month : asreg Ri Rm L_Rm L2_RmL3_RmL4_Rm
- gen R2_δ4=_R2
- drop _*
- bys stkcd month: asreg Ri Rm
- gen R2 = _R2
- drop _*
- gen D1=1-(R2/R2_δ4)
- bysort stkcd year : egen d1=mean(D1)
- sum d1
delay2
- gen mydate=date(tradeweek,"YMD")
- format mydate %tddmy
- gen m = month(mydate)
- gen w = week(mydate)
- rename m month
- rename w week
- order stkcd year month
- bys stkcd month: asreg Ri Rm L_Rm L2_RmL3_RmL4_Rm
- statsby _b, by(stkcd) saving("b.dta", replace): reg Ri Rm L_Rm L2_RmL3_RmL4_Rm
- merge m:1 stkcd using "b.dta"
- gen δi = _b_L_Rm+_b_L2_Rm+_b_L3_Rm+_b_L4_Rm
- gen β =abs(_b_Rm)
- gen D2 = δi/(β+δi)
- winsor2 D2,replace cuts(1 99)
- bysort stkcd year : egen d2=mean(D2)
- sum d2
- * Example generated by -dataex-. To install: ssc install dataex
- clear
- input long stkcd str12 tradeweek str21 日个股流通市值 double Ri float(year Rm L_Rm) str6 证券代码 str9 市场类型 str3 行业代码 str12 交易状态 float(N id t L2_Rm L3_Rm L4_Rm) str3 ind
- 2 "2009-01-05" "62914458.27" .03876 2009.039851 -.009905 "000002" "深圳A" "K70" "正常交易" 242 11 . . . "K"
- 2 "2009-01-06" "64792501.8" .029851 2009.032594.039851 "000002" "深圳A" "K70" "正常交易" 242 12 . . . "K"
- 2 "2009-01-07" "64416893.09" -.005797 2009 -.002036.032594 "000002" "深圳A" "K70" "正常交易" 242 13.039851 . . "K"
- 2 "2009-01-08" "64792501.8" .005831 2009 -.022718 -.002036 "000002" "深圳A" "K70" "正常交易" 242 14.032594.039851 . "K"
- 2 "2009-01-09" "64698599.62" -.001449 2009.020454 -.022718 "000002" "深圳A" "K70" "正常交易" 242 15 -.002036.032594.039851 "K"
- 2 "2009-01-12" "63947382.21" -.011611 2009.004423.020454 "000002" "深圳A" "K70" "正常交易" 242 16 -.022718 -.002036.032594 "K"
- 2 "2009-01-13" "61693729.97" -.035242 2009 -.024805.004423 "000002" "深圳A" "K70" "正常交易" 242 17.020454 -.022718 -.002036 "K"
- 2 "2009-01-14" "64698599.62".048706 2009.042738 -.024805 "000002" "深圳A" "K70" "正常交易" 242 18.004423.020454 -.022718 "K"
- 2 "2009-01-15" "64322990.92" -.005806 2009 .00339.042738 "000002" "深圳A" "K70" "正常交易" 242 19 -.024805.004423.020454 "K"
- 2 "2009-01-16" "64980306.15".010219 2009.014589 .00339 "000002" "深圳A" "K70" "正常交易" 242 1 10.042738 -.024805.004423 "K"
- end
我的思路是搜集日数据(日个股收益率那些)
再将股票中以月为单位回归 以年为单位求平均
因为最后的话我需要得到每只股票对应每年的股价信息效率
有问题的部份烦请指教
熟悉论坛请点击新手指南
京公网安备 11010802022788号
论坛法律顾问：王进律师
知识产权保护声明
免责及隐私声明
