You’re about to launch a new product or enter a new line of business. You have some industry data, but not enough to build a reliable business plan. What can you do?
Traditionally, you would need a lot of data to build a predictive model. But what if you don’t have that data?
That’s where compartmental claims models come in, which take advantage of new probabilistic programming languages such as Stan or PyMC. These models can be used to estimate the number of claims that will be filed, the amount of money that will be paid out, and when those claims will be filed and paid.
To build a compartmental claims model, you need to know a few things:
- The speed of earnings. This is how quickly risk are underwritten and earned
- How quickly claims are reported and settled
- How rates are changing over the coming years
- Assumptions on parameter uncertainties and volatility of the data
This model can be updated as actual claims data becomes available. More importantly, the model can be used to assess whether the business plan assumptions about rate changes, loss ratios, and reporting and payment delays are still valid.
So, if you’re planning to launch a new product or enter a new line of business, don’t let the lack of data stop you. Use a compartmental claims model to build a reliable business plan and make sure your assumptions are sound.