Announcement

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

  • Listing cohorts: Number of firm-year observations

    Hello Guys,

    My sample period is 1976 to 2016 (fyear in the attached dta file and gvkey is the firm id)
    Based on the paper by Srivastava (2014, Why have measures of earnings quality changed over time?) I am trying to prepare a table similar to Table 1: The number of firm-year observations from the successive listing cohorts in each year. For my case firms prior to 1980 will be seasoned firms with new firms for 1980s wave, 1990s wave and 2000s wave

    The table will look like as following (based on full sample)
    Fiscal year Total number of firms Seasoned firms (firms from 1970s) 1980’s wave 1990s wave 2000s wave
    1980 4013 3832 181
    1981 4010 3543 467
    1982 3946 3252 694
    1983 4165 2964 1201
    1984 4204 2589 1615
    1985 4161 2199 1962
    1986 4317 1839 2478
    1987 4440 1469 2971
    1988 4392 1036 3356
    1989 4314 638 3676
    1990 4339 352
    1991 4366 680
    1992 4471 1061
    1993 4757 1601
    1994 5034 2172
    1995 5324 2816
    1996 5863 3760
    1997 5860 4265
    1998 5536 4547
    1999 5425 5070
    2000 5099 403
    2001 4816 738
    2002 4710 1028
    2003 4706 1391
    Where 1980s wave column for the year 1980 is the new firms of 1980. For year 1981 is the summation of new firms in 1980 and 1981 and so forth. What I am unable to do when I move to year 1990 and onward I can’t identify how many of 1970s firms still exists and how many 1980s firms still exists in 1990, 1991 and so on.

    I have attached partial sample Stata dta file. Your help will be highly appreciated.

    Regards
    Attached Files

  • #2
    Please read the FAQ on asking questions. You are advised to provide Stata code in code delimiters, readable Stata output, and sample data using dataex. Many of us will not open files from people we don't know due to concerns about viruses.

    If you have information on the existence of the firm in each year, you can easily make an indicator showing the firm existed in 1970 or whatever. You can use bysort gvkey: egen with either means or totals to paste that value across all observations for that firm. Then you can simply compare with the firm existed in 1970. Once you have identified the group, you can easily use table or summary to count observations.

    Comment

    Working...
    X