If deductibles are below $1000, the from ground up lambda parameters are predicted accurately. However, if the deductibles go up to 10k-100k, the predicted lambda is over 1000. Is this an error in our Stan code or a fundamental issue with the approach?
If deductibles are below $1000, the from ground up lambda parameters are predicted accurately. However, if the deductibles go up to 10k-100k, the predicted lambda is over 1000. Is this an error in our Stan code or a fundamental issue with the approach?