gretl / Feature Requests / #103 Document Nerlove's unbalanced method

Sven Schreiber - 2017-11-05

KTTK, would it perhaps be possible for you to write one or two paragraphs about this for inclusion in the docs? I could eventually look for the right place in the guide and maybe make the final edits, but some input would be very useful.
thanks,
sven

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- KTTK - 2017-11-06
  
  Sven, I don't really feel qualified for this. What I think should be covered about Nerlove's method is:
  1) original paper only treats the balanced case
  2) gretl implements "something" for the unbalanced case: document what it does (mean or weighted mean) and - if possible - why (natural extension? additional Monte Carlo simulations?).
  3) mention that the unbalanced case is not in the literature (AFAIK)
  4) maybe as a footenote: SAS also implements the unbalanced case (in the same way? their documentation, albeit giving formulae, is often wrong).
  
  Last edit: KTTK 2017-11-06
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Sven Schreiber - 2018-05-20

Could somebody confirm (or contradict) that it would basically be enough to mention Allin's gretl WP about that issue? (https://econpapers.repec.org/paper/ancwgretl/4.htm)
If so, a footnote in the docs would be enough.
thanks,
sven

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- KTTK - 2018-05-21
  
  Well, citing the paper and saying what gretl does and why (because neither the original paper nor the literatur provide an unbalanced version of Nerlove's method (best to my knowledge)). A sentence or a footnote will do.
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
  - Sven Schreiber - 2018-05-21
    
    Am 21.05.2018 um 13:20 schrieb KTTK:
    
    Well, citing the paper and saying what gretl does and why (because neither the original paper nor the literatur provide an unbalanced version of Nerlove's method (best to my knowledge)). A sentence or a footnote will do.
    
    Sorry, but that sounds self-contradictory. I don't see how "saying what
    and why" can fit in one sentence. My hope was that the problem can be
    outsourced to Allin's WP. Could you (re-) read Allin's paper and
    pinpoint what information is still missing? Or is the paper not really
    an answer to the question?
    
    thanks,
    sven
    
    If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
    - KTTK - 2018-05-21
      
      The paper contains all information. However, I would love to see an informative sentence and then a footnote to the paper. How about:
      
      "As neither the original paper by Nerlove (1971) nor general literature treat Nerlove's method in the unbalanced case, gretl uses the weighted average of the fixed effects as a natural extension of the method in case of an unbalanced data set." Footnote: ´"For details see [Allin's WP]."
      
      If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
      - Sven Schreiber - 2018-05-23
        
        [I had already replied from the mail program, but this time the quote formatting apparently wasn't recognized by sourceforge's tracker system, so it never appeared here. I wrote:]
        
        As neither the original paper by Nerlove (1971) nor general literature treat Nerlove's method in the unbalanced case, gretl uses the weighted average of the fixed effects as a natural extension of the method in case of an unbalanced data set." Footnote: ´"For details see [Allin's WP]."
        
        Thanks a lot for this constructive suggestion! I will integrate it soon. (Anybody who thinks the sentence is misleading, please speak up.)
        cheers,
        sven
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Sven Schreiber - 2018-07-26

status: open --> closed
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Sven Schreiber - 2018-07-26

Closing this as the documentation had already been added end of May.

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

KTTK - 2018-07-28

Really? I can't seem to find it. Maybe I am looking at the wrong file? Is this the path to the latest version of the manual? https://sourceforge.net/projects/gretl/files/manual/gretl-guide.pdf/download

Can you check, please?

If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
- Sven Schreiber - 2018-07-29
  
  Am 28.07.2018 um 15:01 schrieb KTTK:
  
  Really? I can't seem to find it. Maybe I am looking at the wrong file? Is this the path to the latest version of the manual? https://sourceforge.net/projects/gretl/files/manual/gretl-guide.pdf/download
  
  Can you check, please?
  The paragraph I'm referring to is at the bottom of p.168 of that file.
  It's not a verbatim quote of the sentence in the ticket, but AFAICT it's
  equivalent. If you want to add something or clarify some more, please
  tell us.
  
  cheers,
  sven
  
  alternate
  
  If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
  - KTTK - 2018-07-29
    
    Sven,
    That paragraph just speaks to the unbalanced version of the Swarmy-Arora estimator. It does not cover how gretl extended the original Nerlove estimator to the case of unbalanced panels. (This is probably why this ticket was not closed end of May?)
    
    So my suggestion is to place the paragraph that is currently there directly under formula (20.17), then continue with the text for the Nerlove estimator which is already there ("Alternatively, ...") and then plug in my sentence from this ticket (or some similar sentence).
    
    Last edit: KTTK 2018-07-29
    
    If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
    - Sven Schreiber - 2018-08-01
      
      Am 29.07.2018 um 16:21 schrieb KTTK:
      
      Sven,
      That paragraph just speaks to the unbalanced version of the Swarmy-Arora estimator. It does not cover how gretl extended the original Nerlove estimator to the case of unbalanced panels.
      
      So my suggestion is to place the paragraph that is currently there directly under formula (20.17), then continue with the text for the Nerlove estimator which is already there ("Alternatively, ...") and then plug in my sentence from this ticket (or some similar sentence).
      OK, I see what you mean. I still think it's better to discuss all the
      unbalancedness issues at the end, together. I'm currently away, so I
      can't do it right now.
      cheers,
      sven
      
      alternate
      
      If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
      - Sven Schreiber - 2018-08-10
        
        Am 01.08.2018 um 11:31 schrieb Sven S.:
        
        Am 29.07.2018 um 16:21 schrieb KTTK:
        
        So my suggestion is to place the paragraph that is currently there directly under formula (20.17), then continue with the text for the Nerlove estimator which is already there ("Alternatively, ...") and then plug in my sentence from this ticket (or some similar sentence).
        OK, I see what you mean. I still think it's better to discuss all the
        unbalancedness issues at the end, together. I'm currently away, so I
        can't do it right now.
        
        I have just pushed to git the following amendment, just in time for
        release I guess:
        "Unbalancedness also affects the \cite{nerlove71} estimator, but the
        econometric literature offers no guidance on the details. Gretl
        uses the weighted average of the fixed effects as a natural extension
        of the original method. Again, see \cite{cottrell17} for further details."
        
        Hope that settles things.
        thanks,
        sven
        
        If you would like to refer to this comment somewhere else in this project, copy and paste the following link:

Document Nerlove's unbalanced method

A cross-platform statistical package for econometric analysis

Group

Searches

Help

#103 Document Nerlove's unbalanced method

Discussion