My spreadsheet currently has these three columns: CompanyID, FactorScore, StockReturn.
I need to create a second sheet that will calculate the median StockReturn for each decile of FactorScores. In other words, sort the factor scores and form into 10 equally-sized groups. Then calculate the median stock return within each decile group.
Sounds like you could solve this using a PivotTable.
You would use the grouping function to group the ranks into 10s, and then calculate the median for each group.
0
With monday.comâ€™s project management tool, you can see what everyone on your team is working in a single glance. Its intuitive dashboards are customizable, so you can create systems that work for you.
I created a test worksheet with a Database Named Range R6C2:R36C4 with 3 columns and 30 rows of test data. The columns are
Factor Score Stock Return Decile In Decile
I ignore the Company ID, since it is not being used, but I left room for it in Column 1. The Decile column contains a number from 0 to 9 indicating the decile that Stock Return belongs to. The "In Decile" to the right of the table in column 5 is not part of the Database, but I show it since I will use it to extract a portion of the Factor Scores one decile at a time.
I name the range R7C5:R36C5 InDecile. This is the InDecile column excluding the row title.
I created a named range in R2C4 called DecileCritera. It will contain a number from 0 to 9, representing the decile I want to extract into "In Decile". I put a formula in this cell
=ColumnInputCell
which refers to the Column Input Cell in a Data Table that we will create later.
In R7C4, which is in the Decile column to the right of the first row of data, I put the formula
This formula first isolates the intersection between Database and R7C3:R65536C3, which is essentially the data in the Stock Return column. The RANK function returns a number between 1 and the number of data rows in the database, indicating the rank of the Stock Return in the list. We divide by the number of data rows (we have to subtract 1 since the Database range includes the column titles), multiply by 10 and drop the fractional part. What remains is an integer from 0 to 9 indicating the decile that the Stock Return belongs to.
In R7C5, which is in the In Decile column to the right of the first row of data, I put the formula
=IF(RC[-1]=DecileCriteria,RC[-3],"")
This formula compares the number in the Decile column to DecileCriteria. If it is a match, then we extract the Factor Score. If it isn't a match, we put in a blank so it will be ignored by the Median formula.
We will create a Data Table in the range R40C7:R50C8. Above this range, in R39C7:R39C8 I put two column titles
Decile Decile Median
Then in R41C7:R50C7 I put the numbers 0 to 9. These label the 10 results that will be in R41C8:R50C8, which is the deciles you are looking for.
I name the range R40C7 ColumnInputCell. If you recall, this was the name referred to in the formula in the DecileCriteria.
In cell R40C8 I put the formula
=MEDIAN(InDecile)
This returns the Median of the values in the In Decile column. That column, though, only has values for one decile at a time, depending on what is entered in the ColumnInputCell. Since this cell is blank, the 0 decile is calculated.
Highlight the range R40C7:R50C8. Use Data, Table, and enter ColumnInputCell for the Column Input Cell. Voila, the 10 Medians are calculated in the Decile Median column.
And I added two columns to the 2nd sheet: StartRank and EndRank
StartRank EndRank MedianStockReturn
Decile1 0% 10% xxx
Decile2 10% 20% xxx
Decile3 20% 30% xxx
...
Decile10 90% 100% xxx
So now I'm trying to calculate the median stock returns like this:
if PercentRank between StartRank and EndRank then calculate the median stock return