Simpler Adjusted IPT

**********************************************************************************

This file calculates an alternate and simpler adjusted IPT, following

Blankespoor, deHaan, and Zhu (2017).

Assumption is that the daily return accumulation is immediately at the open.

Contact: Christina Zhu, czhu@stanford.edu

Last updated: 4/14/2017

**********************************************************************************

Input variables:

  _dsetin is your input dataset that has abnormal returns for each firm-date you

    are interested in

  _dsetout is your output dataset, which will contain all of the variables in

    _dsetin, in addition to two Output variables

  AR is the name of the prefix for the variables specifying cumulative abnormal returns

    e.g, &AR._0 to &AR._5 (or &AR._10 or &AR._75 or any number) are the cumulative

    buy-and-hold abnormal returns realized up to and including a given day,

    including day 0

  max is the end day of the calculation period, relative to day 0 of the event

    (e.g. max=10 if you want to calculate 10-day IPT)

**********************************************************************************

Output variables:

  std_ipt is the alternate standard IPT in the time period specified 

    (e.g., 10 days if max=10)

  ipt_0_max (i.e., output variable name is ipt_0_10 if max=10, ipt_0_5 if max=5)

    is adjusted IPT

**********************************************************************************

Example implementation:

  %iptcalc(returns, ipt, AR, 10);

  * Takes a dataset called returns and outputs a dataset called ipt, with the 

  IPT variables called ipt. The abnormal returns variable in the input dataset

  has prefix 'AR' and the calculated standard and adjusted IPT are for days [0, 10]

**********************************************************************************;

%macro alt_iptcalc(_dsetin, _dsetout, AR, max);

%let maxminusone=%eval(&max.-1);

* 1. create &AR._max variable that is the cumulative abnormal return of the max period (i.e., 10 day abnormal return if interested in IPT_0_10);

* This is because day 10 (or 9, or 8, ...) abnormal returns might be missing (e.g., firm is delisted day 10, day 9, etc.);

data ipt_prep;

set &_dsetin;

&AR._max=&AR._&max.;

%do day = &maxminusone. %to 1 %by -1;

  if &AR._max='' then &AR._max=&AR._&day.;

%end;

run;

* 2. Calculate IPT;

data ipt_prep;

set ipt_prep;

* prep to calculate total alternate std_ipt and total alternate adjusted IPT;

%do day = 0 %to &max.;

  stdipt_&day.=&AR._&day./&AR._max;

  adjipt_&day.=1-(abs((&AR._max - &AR._&day.)/&AR._max));

%end;

run;

data &_dsetout;

set ipt_prep;

* standard IPT is the sum of each day;

retain stdipt_: adjipt_:;

alt_std_ipt=sum(of stdipt_0-stdipt_&max.);

alt_ipt_0_&max.=sum(of adjipt_0-adjipt_&max.);

run;

%mend alt_iptcalc;