My last post on BNPL’s impact on FICO drew strong interest from the community.
Why do the two recent studies show conflicting results?
I did a comparison and found the following:
📜 About the 2 studies:
Earlier Equifax/FICO study: average FICO increase of 9 points.
Latest Experian/FICO study: average FICO decrease of 3 points.
FiCO was involved in both studies.
The methodology would be the same.
The scorecard is the same FICO 8 score.
Could it be the underlying sample data?
I pulled out the key info in the summary below.
A few things worth elaboration on beyond the table.
👉 The role of BNPL trades maturity?
The Equifax data has an average repayment history of 4.5 months.
LOCs allow multiple pay-in-4 loans thus the longer repayment period.
The Experian data was based on BNPL trades opened in the previous 3 months.
Scoring: September 2020. BNPL opened: July-September 2020.
So the average prepayment history is 1.5 months only.
It was not clear when the LOCs trades were opened in the Equifax study.
Could the length of the repayment period play a role in the difference?
A longer repayment period allows good payments to show.
In that case, the utilization of LOC would be low at scoring.
👉 The effect of New Credit or the effect of BNPL?
Both studies are based on scores without BNPL vs. scores with BNPL.
The score difference is due to the addition of new BNPL trades.
FICO mentioned another comparison in the Experian study,
Which I totally missed in the first read.
FICO looked at the addition of non-BNPL new credit during the same period.
Then compared scores without new trade vs. with new trade.
The result is a negative 2 points.
That means the -3 points shift from adding BNPL trades,
Is mainly from the effect of new credit.
Even -1 point still contrasts with the strong boost in the other study.
I’m interested to see a study with longer BNPL installment loan history.
👉 Revolving vs. Installment Loan?
Regardless of whether BNPL is reported as a revolving or installment loan,
In both studies, good payments boost the scores; and vice versa.
However, LOC reporting sees score changes +/- 16 on average.
Installment loan reporting sees score changes +/- 10 on average.
Does the classification of trades matter that much?
Well, current scorecards were developed without the BNPL trades.
Now it has to fit in one of the two and uses existing scorecard specifications.
Once more BNPL trades get reported more consistently by lenders,
The next generation of scorecards can be developed to better measure the risk.
My previous post: BNPL impact on FICO (1)
Sources:
Comentarios