# Help with credibility

It’s been a while…

Trying to find the credibility standard for this indication I’m working on, and I have no idea where to start.

Not a traditional line of business, so traditional auto/home credibility standards don’t really work for me.

What parameters do I need to estimate, and what’s the overall formula?

I want the credibility standard for the indication (with complement being 0% change).

I have a derivation somewhere. I think it starts with probability of claim and assumes a normal distribution. Then you have to choose your confidence interval.

Can you share? Thanks!

Yeah, I’ll just have to dig it up. Might take a bit.

But also, that only comes up with the standard for the severity, no? Since it depends solely on the # of claims, and not considering the exposure at all.

I want the standard for the overall indication.

Be sure to charge a standard rate.

you shh

1 Like

My two cents:
If it’s non-traditional, you’ll just have make your own credibility standard.
And how would you do that?
Well, you’ll have to model some large population, see if its size allows you to predict future claims. If it doesn’t, you’ll have to model a larger size.
If there is no finite size of population such that its claims are predictable (in a range you’ll accept), then too bad.

That will be two cents, please.

1 Like

you can also use least squares credibility.

you need to estimate the expected squared error of your (usually unbiased) estimate from data.

then estimate the expected squared error from your compliment.

then weight by the inverse of each expected squared error.

2 Likes

if we ever meet in person you can borrow my vape pen

If my notes are right, the formula for credibility is {Phi^(-1)[1 - (1-0.95)/2]sigma/mu}^2

where Phi is the normal/gaussian cdf, 0.95 the confidence interval for the derivation, sigma is the standard deviation of the estimate, and mu is the mean.

if you are estimating counts then you can set sigma = sqrt(mu).

remember that this aims to limit fluctuations and gives no account of how reliable is the other measurement.

however, especially for claim counts, it has basically a single parameter, so it’s easier to estimate with less data.

Just make it up. Who’s gonna check?

one of the hardest states DOI

Do they REALLY check?

they certainly do ask for your derivation of it if they want to

Tell them you used:

2 + 2 * 4x where x is the variable.

btw I did some more googling and have concluded that your Excel derivation is correct for the bernoulli/binomial frequency distribution.

It just so happens that my LOB is also one with single lifetime frequency, so it’s perfect.
I’m satisfied with my work now.

1 Like

But don’t quote me on that. I have no idea how insurance works.

why are you even on this thread again

Just tryin’ to help, my man, just tryin’ to help.