ASROL Version update: Improvements in the calculation of geometric mean and products in a rolling window and over groups in Stata

[Attaullah Shah]

Version 4.5.1 dated December 11, 2018, of asrol is now available on SSC (thanks to Kit Baum as usual). This version improves the calculation of geometric mean and products. A detailed blog post with relevant examples can be found on my website here. Below is an excerpt from the asrol help file concerning the new changes.


This version of asrol improves the calculation of the product of values and the geometric mean. Since both the statistics involve the multiplication of values in a given window, the presence of missing values and zeros present a challenge to getting the desired results. Following are the defaults in asrol to deal with missing values and zeros:

7.1 : Missing values are ignored when calculating the product or the geometric mean of values.

7.2 : To be consistent with Stata's default for geometric mean calculations, (see ameans), the default in asrol is to ignore zeros and negative number. So

the geometric mean of 0,2,4,6 is 3.6342412, that is [2 * 4 * 6]^(1/3).

And

the geometric mean of 0,-2,4,6 is 4.8989795,
that is [4 * 6]^(1/2).

7.3 : Zeros are considered when calculating the product of values. So

the product of 0,2,4,6 is 0

Two variations are possible when we want to treat zeros differently. These are discussed below:

7.4 Option ignorezero: This option can be used to ignore zeros when calculating the product of values. Therefore, when the zero is ignored,

the product of 0,2,4,6 is 48

7.5 Option add(#): This option adds a constant # to each value in the range before calculating the product or the geometric mean. Once the required statistic is calculated, then the constant is subtracted back. So using option add(1), t

the product of 0,.2,.4,.6 is 1.6880001
=  [1+0 * 1+.2 * 1+.4 * 1+.6] - 1

and the

geometric mean is .280434
 =  [(1+0 * 1+.2 * 1+.4 * 1+.6)^(1/4)] - 1.

The Stata's ameans command calculates three types of means, including the geometric mean. The difference between asrol' gmean function and the Stata ameans command lies in the treatment of option add(#). ameans does not subtract the constant # from the results, whereas asrol does.

[alyubimov]

Hi, I have just started using asrol, which is a very useful tool. I had the following question. I am using asrol to calculated stock returns compounded daily, and asrol allows me to pick a window starting before a specific observation. E.g. if the date is February 20, 2019, I can set the window to 3 and use the rolling product (which is what I need in my case) for February 18-February20. Is there any way to increase the window to after the date, so e.g. make the window February 18-February 22.
Below is the code I am using (cusip is the company identifier, date is the trading day, ret is the daily stock return; I calculate oneplus to later use for compounding)
sort cusip date
by cusip: gen number=_n
gen oneplus=1+ret
by cusip: asrol oneplus, window(number 3) stat(product)

[Attaullah Shah]

The current version of asrol does not support a forward-looking window. I have a plan to include it in its future update. Since your data has stock returns, I assume that your data can be declared as panel data or time series data. IF so, then you can utilize the forward operator of time series/panel data. First, we shall find the returns for a window of 5 observations and then forward the values by two periods.

Since you have not posted an example data set, I am posting first a dataset that will be used in my example code

clear
input float(cusip date ret)
1005 20211  .01271792
1005 20212  .04037379
1005 20213    .020725
1005 20214  .03062747
1005 20215   .0201812
1005 20216  .03498518
1005 20217  .03957188
1005 20218 .020718904
1005 20219 .011318724
1005 20220  .04568697
1005 20221 .012142092
1005 20222  .03731414
1005 20223 .017001197
1005 20224  .04407655
1005 20225 .035473723
1005 20226 .014186862
1005 20227 .025026076
1005 20228 .014784273
1005 20229  .03554296
1005 20230 .016812455
1005 20231  .03768982
1005 20232  .02161423
1005 20233 .017827658
1005 20234  .04664527
1005 20235  .03440519
1005 20236  .02711067
1005 20237 .034615275
1005 20238  .04154405
1005 20239 .008618532
1005 20240 .014368876
1005 20241 .023795227
1005 20242 .032901995
1005 20243  .04341258
1005 20244 .034236886
1005 20245  .04127531
1005 20246 .009568072
1005 20247  .04056266
1005 20248  .03183281
1005 20249 .023167025
1005 20250  .04614547
1005 20251  .02151072
1005 20252  .01955298
1005 20253  .03252285
1005 20254  .01447975
1005 20255 .027213907
1005 20256  .02810522
1005 20257  .03841366
1005 20258  .02586935
1005 20259 .016199369
1005 20260  .03890731
end
format %td date

* The new variable
* Find 5 periods rolling window product of returns, adding the value of 1 before finding the product, the same is deducted back once the product is found.
bys cusip : asrol ret, stat(product) add(1) wind(date 5) gen(cret)

* Declare the data as panel data
tsset cusip date

* Forward the values by two periods
gen cretm_lag = F2.cret


     +-----------------------------------------------------+
     | cusip        date        ret        cret   cretm_~g |
     |-----------------------------------------------------|
  1. |  1005   03may2015   .0127179   .01271792   .0754412 |
  2. |  1005   04may2015   .0403738   .05360518   .1083792 |
  3. |  1005   05may2015    .020725   .07544115   .1307476 |
  4. |  1005   06may2015   .0306275   .10837919   .1556101 |
  5. |  1005   07may2015   .0201812   .13074761   .1547193 |
     |-----------------------------------------------------|
  6. |  1005   08may2015   .0349852   .15561007   .1547124 |
  7. |  1005   09may2015   .0395719   .15471933    .133079 |
  8. |  1005   10may2015   .0207189   .15471244   .1614073 |
  9. |  1005   11may2015   .0113187   .13307897    .135774 |
 10. |  1005   12may2015    .045687   .16140732   .1333073 |
     +-----------------------------------------------------+

[Attaullah Shah]

There is a version update of asrol on SSC, thanks to Kit Baum for maintaining the SSC archives. This version is 4.6, May 2019. This version improves the calculation of gmean and products. An issue was reported of missing values when using the gmean() option of asrol in a rolling window. The issue occurred when there were large numbers to be multiplied before find the geometric mean.
To update

ssc install asrol, replace
help asrol