R2000 no observation error when estimating expected return

[Mai Hang Nguyen]

Dear Statalist,

I'd really appreciate any help members have to offer with my problem.

I'm running an event study on the princeton method (http://dss.princeton.edu/online_help...ventstudy.html). I've successfully merged my market data , with 316 events over 8 years.
My event window is (-5;-2) and estimation window is (-205;-6)
Everything went quite well until the step of calculating expected return

I used the following code and got r2000 no observation error. Therefore, I am unable to process with calculating CAR.

 forvalues i=1(1)284 {
  2. l id company_id if id==`i' & dif==0
  3. reg ret market_return if id==`i' & estimation_window==1
  4. predict p if id==`i'
  5. replace predicted_return = p if id==`i' & event_window==1
  6. drop p
  7. 
. }

Here an example of my data

* Example generated by -dataex-. To install: ssc install dataex
clear
input int(company_id date) str17 ret double market_return float set int event_date float(event_id group_id datenum td dif event_window count_event_obs estimation_window count_est_obs id)
1 18263 "226.06"             518.13 1 20670 1 1   1 1720 -1719 0 4 0 199 1
1 18266 "221.25"             517.05 1 20670 1 1   2 1720 -1718 0 4 0 199 1
1 18267 "222.22"             532.53 1 20670 1 1   3 1720 -1717 0 4 0 199 1
1 18268 "219.33"             534.46 1 20670 1 1   4 1720 -1716 0 4 0 199 1
1 18269 "219.33"             533.34 1 20670 1 1   5 1720 -1715 0 4 0 199 1
1 18270 "218.37"              520.9 1 20670 1 1   6 1720 -1714 0 4 0 199 1
1 18273 "225.1"              515.06 1 20670 1 1   7 1720 -1713 0 4 0 199 1
1 18274 "223.18"             498.31 1 20670 1 1   8 1720 -1712 0 4 0 199 1
1 18275 "223.18"             508.61 1 20670 1 1   9 1720 -1711 0 4 0 199 1
1 18276 "222.22"             512.33 1 20670 1 1  10 1720 -1710 0 4 0 199 1
1 18277 "219.33"             505.42 1 20670 1 1  11 1720 -1709 0 4 0 199 1
1 18280 "219.33"             487.13 1 20670 1 1  12 1720 -1708 0 4 0 199 1
1 18281 "217.41"             495.51 1 20670 1 1  13 1720 -1707 0 4 0 199 1
1 18282 "218.37"              489.5 1 20670 1 1  14 1720 -1706 0 4 0 199 1
1 18283 "217.41"             478.42 1 20670 1 1  15 1720 -1705 0 4 0 199 1
1 18284 "217.41"             477.59 1 20670 1 1  16 1720 -1704 0 4 0 199 1
1 18287 "214.52"             480.91 1 20670 1 1  17 1720 -1703 0 4 0 199 1
1 18288 "213.56"              497.9 1 20670 1 1  18 1720 -1702 0 4 0 199 1
1 18289 "213.56"             486.41 1 20670 1 1  19 1720 -1701 0 4 0 199 1
1 18290 "215.48"             481.76 1 20670 1 1  20 1720 -1700 0 4 0 199 1
1 18291 "213.56"             481.96 1 20670 1 1  21 1720 -1699 0 4 0 199 1
1 18294 "210.67"             486.95 1 20670 1 1  22 1720 -1698 0 4 0 199 1
1 18295 "208.75"             487.93 1 20670 1 1  23 1720 -1697 0 4 0 199 1
1 18296 "211.63"             495.48 1 20670 1 1  24 1720 -1696 0 4 0 199 1
1 18297 "211.63"             503.83 1 20670 1 1  25 1720 -1695 0 4 0 199 1
1 18298 "208.75"             493.04 1 20670 1 1  26 1720 -1694 0 4 0 199 1
1 18301 "212.6"               491.2 1 20670 1 1  27 1720 -1693 0 4 0 199 1
1 18302 "208.75"             482.89 1 20670 1 1  28 1720 -1692 0 4 0 199 1
1 18303 "207.79"             490.91 1 20670 1 1  29 1720 -1691 0 4 0 199 1
1 18304 "208.75"             499.94 1 20670 1 1  30 1720 -1690 0 4 0 199 1
1 18305 "208.75"                507 1 20670 1 1  31 1720 -1689 0 4 0 199 1
1 18308 "208.75"             507.72 1 20670 1 1  32 1720 -1688 0 4 0 199 1
1 18309 "209.71" 508.44000000000005 1 20670 1 1  33 1720 -1687 0 4 0 199 1
1 18310 "211.63"  509.1600000000001 1 20670 1 1  34 1720 -1686 0 4 0 199 1
1 18311 "210.67"  509.8800000000001 1 20670 1 1  35 1720 -1685 0 4 0 199 1
1 18312 "213.56" 510.60000000000014 1 20670 1 1  36 1720 -1684 0 4 0 199 1
1 18315 "213.56"             510.02 1 20670 1 1  37 1720 -1683 0 4 0 199 1
1 18316 "217.41"             496.29 1 20670 1 1  38 1720 -1682 0 4 0 199 1
1 18317 "215.48"             494.59 1 20670 1 1  39 1720 -1681 0 4 0 199 1
1 18318 "217.41"             494.99 1 20670 1 1  40 1720 -1680 0 4 0 199 1
1 18319 "217.41"             496.91 1 20670 1 1  41 1720 -1679 0 4 0 199 1
1 18322 "217.41"             503.01 1 20670 1 1  42 1720 -1678 0 4 0 199 1
1 18323 "223.18"                501 1 20670 1 1  43 1720 -1677 0 4 0 199 1
1 18324 "223.18"             507.32 1 20670 1 1  44 1720 -1676 0 4 0 199 1
1 18325 "225.1"              511.91 1 20670 1 1  45 1720 -1675 0 4 0 199 1
1 18326 "224.14"             513.39 1 20670 1 1  46 1720 -1674 0 4 0 199 1
1 18329 "222.22"             521.12 1 20670 1 1  47 1720 -1673 0 4 0 199 1
1 18330 "221.25"             527.27 1 20670 1 1  48 1720 -1672 0 4 0 199 1
1 18331 "223.18"             525.67 1 20670 1 1  49 1720 -1671 0 4 0 199 1
1 18332 "216.44"             524.64 1 20670 1 1  50 1720 -1670 0 4 0 199 1
1 18333 "220.29"             531.51 1 20670 1 1  51 1720 -1669 0 4 0 199 1
1 18336 "220.29"             531.86 1 20670 1 1  52 1720 -1668 0 4 0 199 1
1 18337 "220.29"             516.08 1 20670 1 1  53 1720 -1667 0 4 0 199 1
1 18338 "221.25"             510.85 1 20670 1 1  54 1720 -1666 0 4 0 199 1
1 18339 "221.25"             519.73 1 20670 1 1  55 1720 -1665 0 4 0 199 1
1 18340 "221.25"             515.99 1 20670 1 1  56 1720 -1664 0 4 0 199 1
1 18343 "221.25"             511.58 1 20670 1 1  57 1720 -1663 0 4 0 199 1
1 18344 "221.25"             506.53 1 20670 1 1  58 1720 -1662 0 4 0 199 1
1 18345 "222.22"             512.12 1 20670 1 1  59 1720 -1661 0 4 0 199 1
1 18346 "222.22"             503.39 1 20670 1 1  60 1720 -1660 0 4 0 199 1
1 18347 "219.33"              505.3 1 20670 1 1  61 1720 -1659 0 4 0 199 1
1 18350 "218.37"             505.67 1 20670 1 1  62 1720 -1658 0 4 0 199 1
1 18351 "224.14"             500.72 1 20670 1 1  63 1720 -1657 0 4 0 199 1
1 18352 "221.25"             499.24 1 20670 1 1  64 1720 -1656 0 4 0 199 1
1 18353 "224.14"             508.18 1 20670 1 1  65 1720 -1655 0 4 0 199 1
1 18354 "224.14"             510.48 1 20670 1 1  66 1720 -1654 0 4 0 199 1
1 18357 "226.06"             514.79 1 20670 1 1  67 1720 -1653 0 4 0 199 1
1 18358 "226.06"             515.33 1 20670 1 1  68 1720 -1652 0 4 0 199 1
1 18359 "223.18"             515.82 1 20670 1 1  69 1720 -1651 0 4 0 199 1
1 18360 "218.37"             515.93 1 20670 1 1  70 1720 -1650 0 4 0 199 1
1 18361 "217.41"             517.42 1 20670 1 1  71 1720 -1649 0 4 0 199 1
1 18364 "213.56"             519.93 1 20670 1 1  72 1720 -1648 0 4 0 199 1
1 18365 "213.56"             517.98 1 20670 1 1  73 1720 -1647 0 4 0 199 1
1 18366 "213.56"             516.21 1 20670 1 1  74 1720 -1646 0 4 0 199 1
1 18367 "213.56"             521.49 1 20670 1 1  75 1720 -1645 0 4 0 199 1
1 18368 "208.75"             522.03 1 20670 1 1  76 1720 -1644 0 4 0 199 1
1 18371 "197.2"              516.74 1 20670 1 1  77 1720 -1643 0 4 0 199 1
1 18372 "206.82"              515.2 1 20670 1 1  78 1720 -1642 0 4 0 199 1
1 18373 "207.79"              519.7 1 20670 1 1  79 1720 -1641 0 4 0 199 1
1 18374 "208.75"             529.31 1 20670 1 1  80 1720 -1640 0 4 0 199 1
1 18375 "209.71"  530.7099999999999 1 20670 1 1  81 1720 -1639 0 4 0 199 1
1 18378 "208.75"             531.21 1 20670 1 1  82 1720 -1638 0 4 0 199 1
1 18379 "210.67"             535.72 1 20670 1 1  83 1720 -1637 0 4 0 199 1
1 18380 "203.94"              539.2 1 20670 1 1  84 1720 -1636 0 4 0 199 1
1 18381 "205.86"             542.37 1 20670 1 1  85 1720 -1635 0 4 0 199 1
1 18382 "206.82"             545.37 1 20670 1 1  86 1720 -1634 0 4 0 199 1
1 18385 "206.82"             545.51 1 20670 1 1  87 1720 -1633 0 4 0 199 1
1 18386 "208.75"             549.12 1 20670 1 1  88 1720 -1632 0 4 0 199 1
1 18387 "208.75"             548.01 1 20670 1 1  89 1720 -1631 0 4 0 199 1
1 18388 "206.82"             549.51 1 20670 1 1  90 1720 -1630 0 4 0 199 1
1 18389 "203.94"             542.03 1 20670 1 1  91 1720 -1629 0 4 0 199 1
1 18392 "204.9"              534.38 1 20670 1 1  92 1720 -1628 0 4 0 199 1
1 18393 "202.01"             534.36 1 20670 1 1  93 1720 -1627 0 4 0 199 1
1 18394 "203.94"              519.6 1 20670 1 1  94 1720 -1626 0 4 0 199 1
1 18395 "202.01"             518.93 1 20670 1 1  95 1720 -1625 0 4 0 199 1
1 18396 "202.01"             520.72 1 20670 1 1  96 1720 -1624 0 4 0 199 1
1 18399 "200.09"             511.27 1 20670 1 1  97 1720 -1623 0 4 0 199 1
1 18400 "202.01"              508.3 1 20670 1 1  98 1720 -1622 0 4 0 199 1
1 18401 "201.05"              494.9 1 20670 1 1  99 1720 -1621 0 4 0 199 1
1 18402 "201.05"             503.55 1 20670 1 1 100 1720 -1620 0 4 0 199 1
end
format %tdnn/dd/CCYY date
format %tdnn/dd/CCYY event_date

Thank you in advance

[Jorrit Gosens]

Crucial error here is that your variable ret is string.

destring ret, replace

but make sure you understand why it is string

Apart from that, the selections you have in your loop also mean that no observations satisfy your conditions, i.e. in your example diff is never 0, and event_window is never 1. This may be different fro the rest of your dataset, but could be another reason why you would get 'no observations' error even after destringing ret

[Mai Hang Nguyen]

Crucial error here is that your variable ret is string.

destring ret, replace

but make sure you understand why it is string

Apart from that, the selections you have in your loop also mean that no observations satisfy your conditions, i.e. in your example diff is never 0, and event_window is never 1. This may be different fro the rest of your dataset, but could be another reason why you would get 'no observations' error even after destringing ret

Hi, thank you very much for your help

I tried the destring command as you suggested and got the following result

ret contains nonnumeric characters; no replace

For the second problem you mentioned, this is only a small example of my data, in the rest, there is 0 diff and 1 event_window
I really don't know what the problem is

[Jorrit Gosens]

The first problem to address is still destringing the variable ret.
do:

clonevar ret_copy = ret
destring ret, replace force

Your result "ret contains nonnumeric characters; no replace" means that ret has some text values for some observations. Stata therefore treats all values as text, you can notice this from them being displayed in red color rather than black in the data editor screen.
You can use destring to convert text to numeric, but the observations with text values would be lost when destringing, and Stata will not throw them away unless specifically asked to do so.

The above code first makes a copy of your ret variale, and then tells Stata to destring the ret variable, including setting observations with text (maybe n/a or ?) to missing.

Again, before you destring force (especially when using force option), you might want to investigate why the ret variable is a string and not numeric

[Mai Hang Nguyen]

The first problem to address is still destringing the variable ret.
do:

clonevar ret_copy = ret
destring ret, replace force

Your result "ret contains nonnumeric characters; no replace" means that ret has some text values for some observations. Stata therefore treats all values as text, you can notice this from them being displayed in red color rather than black in the data editor screen.
You can use destring to convert text to numeric, but the observations with text values would be lost when destringing, and Stata will not throw them away unless specifically asked to do so.

The above code first makes a copy of your ret variale, and then tells Stata to destring the ret variable, including setting observations with text (maybe n/a or ?) to missing.

Again, before you destring force (especially when using force option), you might want to investigate why the ret variable is a string and not numeric

Thanks so much for your help

I will try the code you suggested

[Mai Hang Nguyen]

The first problem to address is still destringing the variable ret.
do:

clonevar ret_copy = ret
destring ret, replace force

Your result "ret contains nonnumeric characters; no replace" means that ret has some text values for some observations. Stata therefore treats all values as text, you can notice this from them being displayed in red color rather than black in the data editor screen.
You can use destring to convert text to numeric, but the observations with text values would be lost when destringing, and Stata will not throw them away unless specifically asked to do so.

The above code first makes a copy of your ret variale, and then tells Stata to destring the ret variable, including setting observations with text (maybe n/a or ?) to missing.

Again, before you destring force (especially when using force option), you might want to investigate why the ret variable is a string and not numeric

Hi

I tried the code you suggested and it work perfectly fine destringing my ret
However, I used the following code to calculate the expected return, I encountered another problem

 forvalues i=1(1)284 {
  2. l id company_id if id==`i' & dif==0
  3. reg ret market_return if id==`i' & estimation_window==1
  4. predict p if id==`i'
  5. replace predicted_return = p if id==`i' & event_window==1
  6. drop p
  7. 
. }

the result I received is as following and stata only gives expected return for 1 out of 284 companies in my data set

 
variable predicted_return not found

[Nick Cox]

Means what it says. You try to replace predicted_return but you need to generate a variable before you can replace it.

[Mai Hang Nguyen]

Means what it says. You try to replace predicted_return but you need to generate a variable before you can replace it.

I already did, using the following code
But Stata only gave expected return for the first company and no results for the rest
Any suggestion to fix this?
Thank you very much in advance

gen predicted_return=.
egen id=group(company_id)

[Nick Cox]

That's not the code you showed in #6 or indeed in #1.

Further, if such a variable did exist, then the error message would not have occurred. We have not a hope of knowing for sure what you do not show us nor a reason not to make inferences based on what you do show us.

Yet further, your example data covers the first company only, so we are in the dark on what is happening for other companies.

I suggest that you try again with your data for company 2 and the complete code you are now using.