Hacker News new | ask | show | jobs
by _dps 4914 days ago
I think the notion of "effective annual interest" for such short-term and small-principal loans is a little stretched, and can easily yield crazy sounding numbers. The core problem is that any costs that are fixed-per-applicant but zero-per-dollar get amortized over a very small principal amount.

Suppose, for example, that each loan application has some fixed cost for evaluation, diligence, regulatory compliance, etc. Say it's $1 per applicant, and that one in ten evaluated applications proceeds to a loan. This means that for every loan made, one has spent $10 before any money has even gone out the door to a borrower. So now consider the effective annual interest rate of a zero-profit loan for various size/duration combinations:

  1) $200, 2 weeks => 255% [100*(210/200)^(52/2) - 100]
  2) $2000, 2 weeks => 15% [100*(2010/2000)^(52/2) - 100]
  3) $200, 1 year => 5% [100*(210/200)^(52/52) - 100]
  4) $2000, 1 year => 0.5% [100*(2010/2000)^(52/52) - 100]
As you can see, even for a lender making zero profit, small durations and small principal amounts dramatically increase the effective APY. Now if you assume your borrowing costs to be zero, and all you want for profit is a 50% gross margin after the diligence costs, we'll be asking for $20 upon repayment instead of $10. The numbers are now:

  1) $200, 2 weeks => 510% [100*(220/200)^(52/2) - 100]
  2) $2000, 2 weeks => 30% [100*(2020/2000)^(52/2) - 100]
  3) $200, 1 year => 10% [100*(220/200)^(52/52) - 100]
  4) $2000, 1 year => 1% [100*(2020/2000)^(52/52) - 100]
I don't know if the inputs I used are reflective of reality ($1 per applicant, 10% conversion rate), but if they are close to accurate I think using the effective annual rates on short-term small-principal loans to call them "predatory" is misleading.

[Edit: I'll add that there's a lot of buffer for overly-pessimistic model inputs in that I've assumed 0% delinquency among accepted loans.]

1 comments

It's also important to note the role of short terms here.

Consider a 2 week loan with a 10% chance of default. You need to charge an 11% risk premium to break even. Translating that 11% risk premium into an APR yields 1400%. A 1 year loan with the same default probability would have only an 11% APR.

So it's both fixed costs and division by short durations which make the APR seem ridiculously high.