Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Survival analysis help, parametric

    Hi there,

    I need help with my research on the duration of a strategy adopted by private equity firms. The dataset is attached as an excel file. I'm really new to survival analyses, so I could use all the help. I have looked at the manual that is provided by Stata, but still I don't know how to proceed with this. I am using an other paper as an example, which uses the accelerated failure time models, and a logistic parametrisation. But I really don't know how to determine what model fits my data the best, and how to proceed with the duration analysis afterwards.
    The first part of my research concerns explaining the impact of certain complexity measures on the duration of the whole strategy. The second part of my research I would like to verify the impact of the complexity on the duration and the corresponding type of exit (IPO, Trade Sale, Financial Buyout, Dissolution, or still active).

    Some explanations regarding my variabels in the dataset.
    -Strategy number
    -P_BVDID, the id of the firm, this is irrelevant for the analysis
    -P_Date, the starting date of the strategy
    -E_Date, the ending date of the strategy, the right censored date is 04/08/2019, the end of the observation period
    -Duration, this is the dependent variable, measured as E_Date - P_Date, or last observed date (04/08/2019), measured in years

    Dependent variables for the second part of the research:
    -IPO, dummy variable
    -Trade_sale, dummy variable
    -Financial_Buyout, dummy variable
    -Dissolution, dummy variable

    --------
    Complexity measures, independent variables:
    -Multiple versus single, a dummy variable indicating whether the strategy involves multiple (1) acquisitions or just one (0)
    -Number of add-ons, the total amount of acquisitions per strategy
    -Domestic, percentage showing the amount of acquisitions that are domestic of all the acquisitions in that strategy
    -Cross_border, percentage showing the amount of acquisitions that are cross-border of all the acquisitions in that strategy
    -Industry penetrating, percentage showing the amount of acquisitions that are penetrating of all the acquisitions in that strategy
    -Industry diversifying, percentage showing the amount of acquisitions that are diversifying of all the acquisitions in that strategy
    -Neighbouring, percentage indicating how many acquisitions are closely related to P_BVDID
    -Time_between_First_addon, measured in years
    -Time_between_first_and_last, measured in years
    -Time_between_last_and_exit, measured in years
    -Time_between_addons, the average time between all the acquisitions, measured in years
    -Distance, distance between the firms
    -Language, dummy variable indicating whether it is the same language (1) or not (0)
    -GDP per capita
    -Hofstede, score of hofstede dimension
    -Institutional quality
    -Investment Attractiveness

    Note: A value of zero, for some variables, needs to be included. Missing data is left blank, not indicated with a zero.

    the above variables have complete information, whereas the control variables below don't.

    Control variables:
    -Private_private, dummy
    -Public_private, dummy
    -Divisional, dummy
    -Privatisation, dummy
    -Receivership, dummy
    -Management participation, dummy
    -Deal syndication_total, counting variable
    -Deal_syndication, dummy
    -PE_Experience, indicating previous acquisition experience, counted as number of prior acquisitions
    -PE_domestic, percentage showing the amount of acquisitions that are domestic of PE_Experience
    -PE_cross_border, percentage showing the amount of acquisitions that are cross border of PE_Experience
    -PF_assets
    -PF_Roa
    -PF_Interest_coverage
    -PF_Experience_total
    -PF_Domestic, percentage of PF_Experience_total
    -PF_Crossborder, percentage of PF_Experience_total
    -PF_Penetrating, percentage of PF_Experience_total
    -PF_Diversifying, percentage of PF_Experience_total
    -PF_Deal_value

    Note: Domestic & Crossborder and Diversifying & Penetrating involve collinearity.

    In my research I would like to analyse the effect of the complexity measures on the duration first, and then add the additional control variables in three stages (from private to deal syndication, and from PE experience to PE cross border, and from PF_assets to PF_Deal Value)

    I really appreciate all the help, and hope that the explanation above is adequate for someone to help me.

    Kind regards,

    Michael


    Attached Files

  • #2
    You didn't get a quick answer. You'll increase your chances of a useful answer by following the FAQ on asking questions - provide Stata code in code delimiters, readable Stata output, and sample data using dataex. Do not append files - many of us won't open files from strangers. Also, try to cut down your posting to the minimum necessary - you're asking us to read a pile of stuff just to help you. We don't need to know all your control variables etc. Try to write out acronyms - what ATF refers to is by not means clear. You claim to have read the survival manual but AFT is not a procedure in that manual.

    I suggest you read some material on survival analysis. Allison has a nice intro in the Sage series.

    Comment


    • #3
      Michael:
      see also the excellent materials on survival analysis with Stata prepared by Stephen Jenkins
      https://www.iser.essex.ac.uk/resourc...sis-with-stata
      Kind regards,
      Carlo
      (Stata 18.0 SE)

      Comment


      • #4
        Dear Carlo and Phil,

        Thanks for your clarifications.
        I've been able to declare the dataset to be ST-data. However, I'm not sure whether the output below is correct.
        The first observed strategy is on 04/02/1997, and has no reported 'failure' until the last observed date.
        The most recent observed strategy is on 06/07/2019, and has the 'time' value of 0.1615, and also no reported 'failure'.
        After watching some tutorials and reading through some examples, is it correct how my list is presented? (In the third code shown below)
        With the latest observed entry being the top line, and for the 'shortest' strategy that no information is given regarding Std. Error etc.

        E_Date is the date the strategy is exited, or the latest observed date in the sample.
        P_Date is the date at which the strategy is started in the sample
        Event, 1 for strategies that are exited before the last observed date

        Code:
        stset E_Date, failure(Event==1) id(Strategy_Number) enter(time P_Date) origin(time P_Date) scale(365.25)
        sts list, survival
        Code:
                     id:  Strategy_Number
             failure event:  Event == 1
        obs. time interval:  (E_Date[_n-1], E_Date]
         enter on or after:  time P_Date
         exit on or before:  failure
            t for analysis:  (time-origin)/365.25
                    origin:  time P_Date
        
        ------------------------------------------------------------------------------
              1,197  total observations
                  0  exclusions
        ------------------------------------------------------------------------------
              1,197  observations remaining, representing
              1,197  subjects
                483  failures in single-failure-per-subject data
          8,332.068  total analysis time at risk and under observation
                                                        at risk from t =         0
                                             earliest observed entry t =         0
                                                  last observed exit t =  22.49418
        Code:
                   Beg.          Net            Survivor      Std.
          Time    Total   Fail   Lost           Function     Error     [95% Conf. Int.]
        -------------------------------------------------------------------------------
         .1615     1197      0      1             1.0000         .          .         .
         .4162     1196      0      1             1.0000         .          .         .
         .4572     1195      0      1             1.0000         .          .         .
         .4736     1194      0      1             1.0000         .          .         .
         .5613     1193      0      1             1.0000         .          .         .
         .6188     1192      0      1             1.0000         .          .         .
          .627     1191      0      1             1.0000         .          .         .
          .668     1190      0      1             1.0000         .          .         .
         .6845     1189      0      1             1.0000         .          .         .
          .731     1188      1      0             0.9992    0.0008     0.9940    0.9999
         .7337     1187      1      0             0.9983    0.0012     0.9933    0.9996
         .8186     1186      0      1             0.9983    0.0012     0.9933    0.9996
         .8569     1185      1      0             0.9975    0.0015     0.9922    0.9992
         .8925     1184      0      1             0.9975    0.0015     0.9922    0.9992
         .9309     1183      0      1             0.9975    0.0015     0.9922    0.9992
         1.008     1182      0      1             0.9975    0.0015     0.9922    0.9992
          1.01     1181      0      2             0.9975    0.0015     0.9922    0.9992
        ***omitted***
         19.59       17      0      1             0.3261    0.0277     0.2725    0.3806
         19.74       16      1      0             0.3057    0.0326     0.2432    0.3703
         20.01       15      0      2             0.3057    0.0326     0.2432    0.3703
         20.16       13      0      1             0.3057    0.0326     0.2432    0.3703
         20.21       12      0      1             0.3057    0.0326     0.2432    0.3703
         20.39       11      0      1             0.3057    0.0326     0.2432    0.3703
         20.48       10      0      1             0.3057    0.0326     0.2432    0.3703
         20.55        9      0      1             0.3057    0.0326     0.2432    0.3703
         20.59        8      0      1             0.3057    0.0326     0.2432    0.3703
         20.61        7      0      1             0.3057    0.0326     0.2432    0.3703
         20.79        6      0      1             0.3057    0.0326     0.2432    0.3703
         20.84        5      0      1             0.3057    0.0326     0.2432    0.3703
         21.04        4      0      1             0.3057    0.0326     0.2432    0.3703
         21.38        3      0      1             0.3057    0.0326     0.2432    0.3703
         21.57        2      0      1             0.3057    0.0326     0.2432    0.3703
         22.49        1      0      1             0.3057    0.0326     0.2432    0.3703

        Code:
         list _t0 _t _origin _d _st
        Code:
               | _t0          _t     _origin   _d   _st |
              |----------------------------------------|
           1. |   0   22.494182   04feb1997    0     1 |
           2. |   0   8.3312799   04jun1997    1     1 |
           3. |   0   7.7289528   07oct1997    1     1 |
           4. |   0   5.1498973   15dec1997    1     1 |
           5. |   0   21.574264   06jan1998    0     1 |
              |----------------------------------------|
           6. |   0   4.0711841   20feb1998    1     1 |
           7. |   0    3.633128   28feb1998    1     1 |
           8. |   0   21.382615   17mar1998    0     1 |
           9. |   0   4.1259411   30mar1998    1     1 |
          10. |   0    5.467488   31mar1998    1     1 |
              |----------------------------------------
        ***omitted***
              |----------------------------------------|
        1191. |   0   .62696783   18dec2018    0     1 |
        1192. |   0   .61875428   21dec2018    0     1 |
        1193. |   0   .56125941   11jan2019    0     1 |
        1194. |   0   .47364819   12feb2019    0     1 |
        1195. |   0   .45722108   18feb2019    0     1 |
              |----------------------------------------|
        1196. |   0   .41615332   05mar2019    0     1 |
        1197. |   0    .1615332   06jun2019    0     1 |

        I hope this provides adequate information for someone to help me.

        Kind regards,

        Michael
        Last edited by MIchael Jefferson; 23 Sep 2019, 08:15.

        Comment

        Working...
        X