About AbuChowdhury

AbuChowdhury · ‎10-25-2014

Following is the sample of my dataset. I need to calculate lag of "asset" for each company_id. Then years 2003 and 2013 will have to be dropped. How can I do that? year company_id asset 2003 AN8068571086 3109000 2004 AN8068571086 2997000 2005 AN8068571086 3496000 2006 AN8068571086 2998873 2007 AN8068571086 3169033 2008 AN8068571086 3692000 2009 AN8068571086 4642000 2010 AN8068571086 5006000 2011 AN8068571086 4837000 2012 AN8068571086 6312000 2013 AN8068571086 8478000 2003 ANN6748L1027 26679 2004 ANN6748L1027 29661 2005 ANN6748L1027 65606 2006 ANN6748L1027 25151 2007 ANN6748L1027 28439 2008 ANN6748L1027 18375 2009 ANN6748L1027 17421 2010 ANN6748L1027 27269 2011 ANN6748L1027 62001 2012 ANN6748L1027 39696 2013 ANN6748L1027 39492 2003 BMG0129K1045 . 2004 BMG0129K1045 0 2005 BMG0129K1045 741997 2006 BMG0129K1045 58002 2007 BMG0129K1045 13546 2008 BMG0129K1045 80947 2009 BMG0129K1045 142666 2010 BMG0129K1045 239957 2011 BMG0129K1045 295522 2012 BMG0129K1045 618217 2013 BMG0129K1045 777386

AbuChowdhury · ‎06-17-2014

proc qlim is running and showing the results but the logfile also shows the following error message: WARNING: The Hessian matrix is singular. ERROR: QUANEW Optimization cannot be completed.

AbuChowdhury · ‎06-13-2014

Hi, Thanks for reply. Yes, the dependent variable will me missing many places because the dependent variable is cumulative abnormal return and independent variables are firm level variables such as book-to-market, market value, return on assets, dividend yield. There are observations for each firm each year of the independent variables but dependent variable, car5, does not have the observations for each firm each year. Yes, the syntax will be car5. In the log file, it tells that there is one cross cross section or time series. I need to run the regression with firm (companyISIN) and year fixed effects. So what will be the codes? should I continue to use proc panel? or should I use proc glm or proc mixed? Let me know if you need further information. Your help is highly appreciated.

AbuChowdhury · ‎06-12-2014

Hi All, I have the following variables in my dataset: companyISIN, year, car5, bm, mv, roa, dy. I also need to do the log of mv variable and divide roa by ression100. So the dependent variable will be car5 and independent variables will be bm, log of mv, roa/100, and dy. In the regression, there will be two fixed effects: firm (i.e. companyISIN) and year. How can I run the fixed effects regression using proc panel. proc glm, or proc mixed? I run the following codes but it does not work: proc panel data = x; model car = mv ta roa ni dy / fixone; id CompanyISIN year; run; I have included the data file so you can have a look. Thanks.

AbuChowdhury · ‎06-08-2014

Yes, they are. Thanks a lot Xia.

AbuChowdhury · ‎06-07-2014

Hi, thanks. Your codes work fine. But still there is a problem. There are no observations in any year for some variables. When I include those variables then it shows the error message: all variables in array list must be the same character i.e. character or numeric. Look at the complete excel file.

AbuChowdhury · ‎06-04-2014

Hi, Your codes also work fine.

AbuChowdhury · ‎06-04-2014

Hi Xia, Your codes work fine. Now for another xls file (attached now), I would like to sort the data in the same panel format. At first I used the following codes to import: proc import datafile='c:\temp\USA_mv.xls' out=temp dbms=excel replace;mixed=yes;getnames=yes;run; Then I used the following codes to sort the data in the panel format that I mentioned before: data want; set temp; length code $ 40; array a{*} an: bmg: bsp: ca: ch: gb: ie: il: kyg: lr: mhy: nl: pa: sg: us: vgg: ; do i=1 to dim(a); code=vname(a{i}); value=a{i}; output; end; keep year code value; run; These codes are not working. It shows the error message: ''all variables in the array list must be same type i.e. all character or numeric''. How to solve for this xls file?

AbuChowdhury · ‎06-04-2014

I used getnames when I imported, but still it did not show the years. It shows only the CompanyISIN as variable name, and instead of showing years (2013, 2012 .... ) it showed variable names as A, B, C .... etc.

AbuChowdhury · ‎06-04-2014

Hi, I can import to SAS but it does not show the year (2013, 2012 ets) in SAS dataset, Rather it shows (A, B, C) as column names. ''Running codes in sas'' -- after importing the xls file to SAS, when I was trying to sort the dataset according to the format that I mentioned before. Actually, I need to sort the dataset according to the way I mentioned before.

AbuChowdhury · ‎06-04-2014

Hi, I attach a xls file. I need to have a dataset in SAS and it should be in panel format like below: CompanyISIN Year Value US9311421039 2004 284310000 US9311421039 2005 312101000 US9311421039 2006 348368000 . . . CA123456789 2004 CA123456789 2005 .. . . How can I do that? After importing to SAS, when I am running codes in SAS, it shows the error message that all variables should be in the same format i.e. character or numeric. I also would like to change the n.a observations (in xls file) to blank observation in SAS. Abu Chowdhury

AbuChowdhury · ‎05-22-2014

Thanks a lot.

AbuChowdhury · ‎05-22-2014

I think the reason is as follows. When I wanted to calculate the CAR for (0,65) days window, I wanted to calculate actually for (0,3) months window. Since the dataset has approximately 260 observations per year (because, no data for weekends and holidays), so 3 months mean approximately 65 observations per year. But, according to calendar days, there are approximately 91 days in a 3-month period. Your sql proc chooses the calendar days. So in your proc sql when you wrote (…..between tradedate and tradedate+65), it took the 65 calendar days. So, since I want a 3-month window, I should write: …… between tradedate and tradedate+91 Please let me know whether I am correct or not.

AbuChowdhury · ‎05-22-2014

Hi, Something is going wrong. If you remember, you gave me the following proc sql for calculating cumulative abnormal returns. proc sql; create table want as select a.*,(select sum(value) from x where code=a.code and date between tradedate and tradedate+65) as _65 from trading_obs as a; quit; However, I ran this for 3 days ( ..... tradedate and tradedate+3) and manually checked the results. Surprisingly, it shows correct car in some event days but in many cases it shows wrong cars for respective event days (tradedate). What's the reason for that? I gave you the two datasets previously; please check and let me know.

AbuChowdhury · ‎05-21-2014

Thanks a lot.

Online Status	Offline
Date Last Visited	‎01-15-2021 09:49 AM

Re: keep specific number of time series observations for each cross se...

keep specific number of time series observations for each cross sectio...

Re: Merge Two Datasets using String Variable and Year

Re: Merge Two Datasets using String Variable and Year

Re: Merge Two Datasets using String Variable and Year

Re: Merge Two Datasets using String Variable and Year

Re: Merge Two Datasets using String Variable and Year

Re: Merge Two Datasets using String Variable and Year

Re: Merge Two Datasets using String Variable and Year

Merge Two Datasets using String Variable and Year

Re: How to remove digits from the observations?

GMM Estimation

calculating lag

Re: Regression

Re: Regression

Regression

Re: Sorting Data

Re: Sorting Data

Re: Sorting Data

Re: Sorting Data

Re: Sorting Data

Re: Sorting Data

Sorting Data

Re: Cumulative Abnormal Return Calculations

Re: Cumulative Abnormal Return Calculations

Re: Cumulative Abnormal Return Calculations

Re: Proc sql