After a little more digging, I can mostly answer my own question:

First, I had to relearn that qp_date(0,"ABC") = qp_date(0, "XYZ") is
NOT generally true. In other words, the IRL index prices were not yet
updated for the week, when I extracted my data. This accounts for the
bulk of the deviation between the IRL index return and the constituent
return (however averaged). I am still not sure, whether the IRL index
prices are supposed to be market cap adjusted, or not. I can't get the
numbers to line up, either way.

Second, the volume data for the IRL indexes is equal to total
unweighted share volume of the constituent stocks times 100.

Third, the IRLDaily database appears to contain a rolling 104 weeks of
data, with prices being indexed to the oldest datapoint. Again,
how the returns are averaged, I am not sure.

One important conclusion following from the last point is that IRL
index prices for a given DATE, will move every week, as they are
recalculated to a new baseline value.

Another (and more positive) result from the IRLDaily setup is that IRL
index prices have an apparent interpretation as two-year returns. For
example, the current closing price for !ID001 (Aerospace, Defense) of
140.81 implies a 40.81% cumulative two-year return for this industry.

So my only remaining question is, how idex returns are averaged from
constituent returns. Your help will be greatly appreciated.

Thanks,
Matthias

Reply via email to