Hello Statalist,
I have queries about how to remove non-trading days. I use Stata v14.0. For simplicity, I use panel data, which have 10 firms for period 2006-2014. All data is in daily. I would like to create Return as a new variable using this command:
gen return = ln(_price /lag_price)*100
My question is, in order to generate return, should I use the ordinary calendar date or use the business calendar date? If I use the business calendar date, there are lots of missing value.
After generated the return, I would like to drop the non-trading days. Here, the non-trading days defined as days on which 90% or more of the stocks listed have a return equal to zero. I have tried calculated manually but it is not an efficient way. Does anyone have ideas how to solve it?
The last question is, I would like to drop a stock if the number of zero-return days is more than 80% in a given month. I have tried use this command:
tab return if return ==0
However, this is not what I want because it described in daily. How can I jumped from the daily return to the 80% in a month, let say zero-return stocks >80% in January, February and so on?
I really need help from Statalist. Your help is highly appreciated.
Regards,
Rozita
I have queries about how to remove non-trading days. I use Stata v14.0. For simplicity, I use panel data, which have 10 firms for period 2006-2014. All data is in daily. I would like to create Return as a new variable using this command:
gen return = ln(_price /lag_price)*100
My question is, in order to generate return, should I use the ordinary calendar date or use the business calendar date? If I use the business calendar date, there are lots of missing value.
After generated the return, I would like to drop the non-trading days. Here, the non-trading days defined as days on which 90% or more of the stocks listed have a return equal to zero. I have tried calculated manually but it is not an efficient way. Does anyone have ideas how to solve it?
The last question is, I would like to drop a stock if the number of zero-return days is more than 80% in a given month. I have tried use this command:
tab return if return ==0
However, this is not what I want because it described in daily. How can I jumped from the daily return to the 80% in a month, let say zero-return stocks >80% in January, February and so on?
I really need help from Statalist. Your help is highly appreciated.
Regards,
Rozita
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input float idc str21 company float caldate double(_marketvalue _price _returnindex) 1 "ALUAR" 16803 4422 2.0100000000000002 6915.96 1 "ALUAR" 16804 4395.6 1.99 6874.67 1 "ALUAR" 16805 4461.6 2.02 6977.89 1 "ALUAR" 16806 4474.8 2.0300000000000002 6998.54 1 "ALUAR" 16807 4554 2.07 7122.41 1 "" 16808 . . . 1 "" 16809 . . . 1 "ALUAR" 16810 4593.59 2.08 7184.34 1 "ALUAR" 16811 4593.59 2.08 7184.34 1 "ALUAR" 16812 4593.59 2.08 7184.34 1 "ALUAR" 16813 4580.39 2.08 7163.7 1 "ALUAR" 16814 4540.8 2.06 7101.76 1 "" 16815 . . . 1 "" 16816 . . . 1 "ALUAR" 16817 4501.2 2.04 7039.83 end format %td caldate
