Hello all, My data refers to CEO tenures (average tenure length is 8 years). This means that my sample includes firm-year observations regarding several measures of net profit (accounting return), and those observations are sorted by CEO tenures. I wrote a code to eliminate outliers of the variables "net profit" in the 1% and 99% (program 1). Then I wrote a second code that said that if there is an outlier during CEO tenure, all the other observations in the same tenure will be eliminated as well (program 2). According to my understanding, those programs combined should reduce my sample by an average of 16%: 2% of the sample should be reduced in program 1, and as the average tenure length is 8 years, after program 2, the reduction of observations should be 16%. However, the actual reduction is a third of the sample (twice what I expected). Enclosed are the two codes (two programs). Could you please advise if there are problems with the codes? Thanks in advance, Lior
... View more