Forum Discussion

gguadalupe's avatar
gguadalupe
Frequent Visitor
9 years ago
Solved

Insurance Earned Premium/Loss Ratio Calculation

Hi All! I'm new to PowerBi and I'm trying to implement an insurance dashboard. All is fine, but now I'm trying to implement the loss ratio formula controled by a slicer filled with my Dates table. ...
  • gguadalupe's avatar
    gguadalupe
    9 years ago

    v-jiascu-msft

     

    Hi Dale,

     

    I have been thinking about this, and I finally understood that my data model was wrong.

     

    Now the model has the info "flat" month by month, and I let Power Bi do just aggregations.

     

    With this model now I can see the Earned Premium month by month, and calculate the Loss Ratio change month by month.

     

    Thank you Dale for your help!!

     

    Gus.

v-jiascu-msft's avatar
v-jiascu-msft
Microsoft Employee
9 years ago

gguadalupe

 

Hi,

 

Is the PremiumID of the records (rows) in PremiumTable unique? If yes, here could be the solution.

 

It’s better to take Premium Per Day as a calculated column. The formula is: ( pay attention to the +1 in blue. )

 

Premium Per Day =
'PremiumTable'[PremiumAmt]
    / (
        DATEDIFF ( 'PremiumTable'[EffectiveDate], 'PremiumTable'[ExpirationDate], DAY )
            + 1
    )

Then we are going to create two measures.

Earned Premium =
MIN ( PremiumTable[Premium Per Day] )
    * CALCULATE (
        COUNTROWS ( DateTable ),
        FILTER (
            DateTable,
            DateTable[Date] >= MIN ( PremiumTable[EffectiveDate] )
                && DateTable[Date] <= MIN ( PremiumTable[ExpirationDate] )
        )
) 

 

Accumulated Earned Premium =
MIN ( 'PremiumTable'[Premium Per Day] )
    * CALCULATE (
        COUNTROWS ( DateTable ),
        FILTER (
            ALL ( DateTable ),
            'DateTable'[Date] <= MAX ( 'DateTable'[Date] )
                && 'DateTable'[Date] >= MIN ( 'PremiumTable'[EffectiveDate] )
                && DateTable[Date] <= MIN ( PremiumTable[ExpirationDate] )
        )
    )

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Best Regards!

Dale

 

gguadalupe's avatar
gguadalupe
Frequent Visitor
9 years ago

Hi v-jiascu-msft

 

Thank you for the script.

I tried it with these 3 records and evething is perfect, but when I load the rest of the 200k records, it start giving me some weird result, like negative amounts.

I did tried another posible solution (still working on) creating a calculated table with the following script.

EarnedPremium = FILTER(
    CROSSJOIN(PremiumTable,DateTable),
    DateTable[Date] >= PremiumTable[EffectiveDate] && DateTable[Date] <= PremiumTable[ExpirationDate]
)

This gives me a table of 52 MILLON records, so I will try to finish the implementation of your solution.

 

Thank you again!!

 

Gus.

  • v-jiascu-msft's avatar
    v-jiascu-msft
    Microsoft Employee
    9 years ago

    gguadalupe

     

    Hi,

     

    It's complicated in the production. Please take these things below into consider.

    1. The 'Premiumtable[EffectiveDate] should be less than Premiumtable[ExpirationDate];

    2. DateTable should be complete and continuous;

    3. The report should have at least one unique column;

    4. Premium Per Day is a calculated column in the table, while Earned Premium and Accumulated Earned Premium are measures.

     

     

    Best Regards!

    Dale

    • gguadalupe's avatar
      gguadalupe
      Frequent Visitor
      9 years ago

      v-jiascu-msft

       

      Hi Dale,

       

      I have been thinking about this, and I finally understood that my data model was wrong.

       

      Now the model has the info "flat" month by month, and I let Power Bi do just aggregations.

       

      With this model now I can see the Earned Premium month by month, and calculate the Loss Ratio change month by month.

       

      Thank you Dale for your help!!

       

      Gus.

      • v-jiascu-msft's avatar
        v-jiascu-msft
        Microsoft Employee
        9 years ago

        gguadalupe

         

        Hi,

         

        My pleasure. That's great! Thank you for sharing this with us!

         

        Best Regards!

        Dale

    • Actuary's avatar
      Actuary
      Frequent Visitor
      2 years ago

      I am also suffering with something similar, can you please help?