1 of 2
1
IBES database use
Posted: 14 April 2016 06:41 PM   [ Ignore ]
Newbie
Rank
Total Posts:  13
Joined  2016-04-14

Hi,

I feel so confused about how to use IBES even if I read teh manual twice…

1) the manual says it contains up to 5 fiscal year and 8 quarters estimate…...but when I use the database…the date variable I can pick over 10 years..is that contradicted?

2) I need to get 2005-2014 all the firms’ analyst forecast EPS and announced actual EPS…...should I pick the forecast year end as the date variable or announce date as the date variable? I feel so dizzy….

3) what is FPI for the database? Why we need it?

The manual does not help that much unfortunately….. I really need some help.  Thank you!

Profile
 
 
Posted: 15 April 2016 08:46 AM   [ Ignore ]   [ # 1 ]
Administrator
Avatar
RankRankRankRank
Total Posts:  901
Joined  2011-09-19

hi Superman,

The best way to get started is to select the records of a single company and manually go through these. Then, it will make a lot more sense.

Best,

Joost

 Signature 

To reply/post new questions: Please use the group WRDS/SAS on Google Groups! http://groups.google.com/d/forum/wrdssas

Profile
 
 
Posted: 15 April 2016 09:36 AM   [ Ignore ]   [ # 2 ]
Newbie
Rank
Total Posts:  13
Joined  2016-04-14
Joost Impink - 15 April 2016 08:46 AM

hi Superman,

The best way to get started is to select the records of a single company and manually go through these. Then, it will make a lot more sense.

Best,

Joost[/quote

I did,and it makes me even more confusing…....I did for one firm and 1 year only…... see attached…

Why there is no actual value at the bottom for the same announce date 2015/01/29?  and the FPI confused me as well….‘cause it shows very different results…...

The manual does not help that much ..I did try many times myself…...so desperate

Profile
 
 
Posted: 15 April 2016 12:39 PM   [ Ignore ]   [ # 3 ]
Administrator
Avatar
RankRankRankRank
Total Posts:  901
Joined  2011-09-19

hi,

Important columns are missing in the data you have. Try again with all columns. For example, I don’t see the date the forecast was released. That is important, because forecasts change over time. Also, analysts can forecast multiple quarters/years, so you need the variables that indicate what period is being forecasted.

Hope this helps,

Joost

 Signature 

To reply/post new questions: Please use the group WRDS/SAS on Google Groups! http://groups.google.com/d/forum/wrdssas

Profile
 
 
Posted: 15 April 2016 05:59 PM   [ Ignore ]   [ # 4 ]
Newbie
Rank
Total Posts:  13
Joined  2016-04-14
Joost Impink - 15 April 2016 12:39 PM

hi,

Important columns are missing in the data you have. Try again with all columns. For example, I don’t see the date the forecast was released. That is important, because forecasts change over time. Also, analysts can forecast multiple quarters/years, so you need the variables that indicate what period is being forecasted.

Hope this helps,

Joost

I still have one important question still bothering me…1st step date range….what exactly does the date variable do? Idk if I should pick announce date or forecast period end date?  Why this is a variable? so confusing

Thank you

Profile
 
 
Posted: 15 April 2016 06:31 PM   [ Ignore ]   [ # 5 ]
Administrator
Avatar
RankRankRankRank
Total Posts:  901
Joined  2011-09-19

Think about the weather forecast. Every day I make forecasts for the next 10 days. Do you want to filter on the days that I made forecasts (1 step date range), or, do you want to have the dates that I made forecasts for (forecast period end date).

So there are 2 date variables:
- the date (or month) that analysts are doing their work
- the dates of the quarter/year that analysts make forecasts for

Best,

Joost

 Signature 

To reply/post new questions: Please use the group WRDS/SAS on Google Groups! http://groups.google.com/d/forum/wrdssas

Profile
 
 
Posted: 15 April 2016 07:41 PM   [ Ignore ]   [ # 6 ]
Newbie
Rank
Total Posts:  13
Joined  2016-04-14
Joost Impink - 15 April 2016 06:31 PM

Think about the weather forecast. Every day I make forecasts for the next 10 days. Do you want to filter on the days that I made forecasts (1 step date range), or, do you want to have the dates that I made forecasts for (forecast period end date).

So there are 2 date variables:
- the date (or month) that analysts are doing their work
- the dates of the quarter/year that analysts make forecasts for

Best,

Joost

That is a great explanation. I am trying to find the relationship between previous forecast error ( announced actual earnings-previous forecast) and the following forecast ( compared to previous forecast). My hypothesis is there is a negative relationship, which means you had a positive error, you might adjust the following forecast lower than previous forecast; same for negative error experience, you adjust the following forecast higher based on previous forecast. I want to use quarterly data as well.

I still feel a bit hesitated….should I choose the forecast period end date then?

Thank you!

Profile
 
 
Posted: 15 April 2016 07:59 PM   [ Ignore ]   [ # 7 ]
Administrator
Avatar
RankRankRankRank
Total Posts:  901
Joined  2011-09-19

hi,

Why not download everything to make the largest dataset possible (more statistical power)?

Best,

Joost

 Signature 

To reply/post new questions: Please use the group WRDS/SAS on Google Groups! http://groups.google.com/d/forum/wrdssas

Profile
 
 
Posted: 15 April 2016 08:03 PM   [ Ignore ]   [ # 8 ]
Newbie
Rank
Total Posts:  13
Joined  2016-04-14
Joost Impink - 15 April 2016 07:59 PM

hi,

Why not download everything to make the largest dataset possible (more statistical power)?

Best,

Joost

But for the 1st step, I can only choose 1 date variable…..I cannot pick both of them…

Profile
 
 
Posted: 15 April 2016 08:06 PM   [ Ignore ]   [ # 9 ]
Administrator
Avatar
RankRankRankRank
Total Posts:  901
Joined  2011-09-19

ok, I see; at the first step, select announcement date. This will not restrict the periods forecasted. For example, if you were to filter the announcement date to 2010, you will get forecasts made in 2010, for periods (quarters/years) ending in 2010, 2011, 2012, etc.

Best,

Joost

 Signature 

To reply/post new questions: Please use the group WRDS/SAS on Google Groups! http://groups.google.com/d/forum/wrdssas

Profile
 
 
Posted: 15 April 2016 08:10 PM   [ Ignore ]   [ # 10 ]
Newbie
Rank
Total Posts:  13
Joined  2016-04-14
Joost Impink - 15 April 2016 08:06 PM

ok, I see; at the first step, select announcement date. This will not restrict the periods forecasted. For example, if you were to filter the announcement date to 2010, you will get forecasts made in 2010, for periods (quarters/years) ending in 2010, 2011, 2012, etc.

Best,

Joost

Makes perfect sense!!!!!!!

Profile
 
 
Posted: 15 April 2016 08:12 PM   [ Ignore ]   [ # 11 ]
Newbie
Rank
Total Posts:  13
Joined  2016-04-14
Joost Impink - 15 April 2016 08:06 PM

ok, I see; at the first step, select announcement date. This will not restrict the periods forecasted. For example, if you were to filter the announcement date to 2010, you will get forecasts made in 2010, for periods (quarters/years) ending in 2010, 2011, 2012, etc.

Best,

Joost

For the FPI indicator…would you pick just quarter 1 or more? That is confusing as well…I know quarter 2 is the forecast for next quarter and same for the following quarter indicator…....I think maybe I should use all the quarter indicators?

Thank you again!

Profile
 
 
Posted: 15 April 2016 08:12 PM   [ Ignore ]   [ # 12 ]
Administrator
Avatar
RankRankRankRank
Total Posts:  901
Joined  2011-09-19

I’m glad it does! smile

Take care,

Joost

 Signature 

To reply/post new questions: Please use the group WRDS/SAS on Google Groups! http://groups.google.com/d/forum/wrdssas

Profile
 
 
Posted: 15 April 2016 08:19 PM   [ Ignore ]   [ # 13 ]
Administrator
Avatar
RankRankRankRank
Total Posts:  901
Joined  2011-09-19

About the FPI - that is indeed a question on how you design your empirical test—I would probably go for something that you can apply consistently across different firms (firms that are easier to forecast/more stable may have more quarters that have forecasts, and that could interfere with what you are testing)

Best,

Joost

 Signature 

To reply/post new questions: Please use the group WRDS/SAS on Google Groups! http://groups.google.com/d/forum/wrdssas

Profile
 
 
Posted: 15 April 2016 09:09 PM   [ Ignore ]   [ # 14 ]
Newbie
Rank
Total Posts:  13
Joined  2016-04-14

Thank you soooo much! I will keep you posted once I finish my paper

Profile
 
 
Posted: 16 April 2016 11:37 AM   [ Ignore ]   [ # 15 ]
Newbie
Rank
Total Posts:  13
Joined  2016-04-14
Joost Impink - 15 April 2016 08:19 PM

About the FPI - that is indeed a question on how you design your empirical test—I would probably go for something that you can apply consistently across different firms (firms that are easier to forecast/more stable may have more quarters that have forecasts, and that could interfere with what you are testing)

Best,

Joost

It is weird: 

                                                              ANNDATS_
OFTIC   TICKER     FPEDATS   MEASURE     ANNDATS   FPI   ANALYS     VALUE     ACTUAL       ACT

      AMPR     2010/09/30     EPS         2011/01/03   1   049945       0.00     0.000   2011/01/13
      DANG     2010/12/31     EPS         2011/01/18   1   059325       0.04     0.020   2011/03/09
      EPOC     2010/12/31     EPS         2011/02/10   1   127593       0.30     0.460   2011/03/29

Why forecast period end date is 2010 while announce date is 2011? It should be reversed…....

Profile
 
 
   
1 of 2
1