[Hemanshu Kumar]

They will result in missing values if the rt forall stocks on a given day, or the mcap for all stocks on the previous day, are missing. Otherwise, they will produce values based on the non-missing values of those variables. You can see the output in #9.

[Sartaj Hussain]

Ok. Can we add no of non-missing rt observations for each time period.

[Hemanshu Kumar]

Use the count function of -egen- for that.

[Clyde Schechter]

Re #13

Please go through #5 which i posted today. The code has to do calculations for each day. In case for any stock, if the value of rt or mcap or both is missing, then it can ignore that stock for that day. But as a whole for each time period (day), other daily rt's using the previous period mcap as weight (provided either rt, mcap or both is not missing) have to aggregate into a single daily weighted average measure.

Exactly. When mcap from the previous period is missing, as is the case with 31 May 2002 for stock id #2, it is impossible to include anything from stock id #2 for June 2002-May 2003.