15.5% to 21.4%Real lending data from an open research dataset
The number that decides pricing
of attainable profit given up, in simulation, by pricing on the wrong belief about whether price causes default
- What was observed
- A lender's prices, defaults and repayments, segment by segment, with take up and funding costs assumed.
- What a generic approach says
- Optimise prices on the book's history and take the most profitable.
- What the engine read
- The answer turns on one number history cannot settle: whether a higher price itself raises default. If it does, the best prices cut the riskiest grades' rates and raise the safest; if it does not, they raise five of seven grades and shrink the book by about a fifth of its loans. Pricing on the wrong belief gave up 15.5% to 21.4% of the profit available.
- What it meant for the decision
- Run a small randomised price test before choosing prices: it measures the one number the history cannot. This is a simulation on the lender's own history, not a test of prices in the market.