what now

After seeing an insured go to the national carrier they are with and lower themselves from our minimum to the state minimum. 

Completely understand this younger person who needs to drive to work, pay other bills, etc.  On the other hand, 8 of our 10 options would not offer a rate so he is left with 2 of our choices and maybe 2-5 others elsewhere.  Now, we have a woefully underinsured driver on the road because this is what he can afford. So, his carrier is doing what they can to take on some of the riskier populations.  But riskier is relative to the criteria they are basing rates on.  Charging for driving infractions versus for paid claims.  But, those other 8 companies of ours, not to mention a few dozen more will then shoulder the cost of the property damage from any claims he is involved in, not to mention the potential for large medical bills.  As I learned this fall, it is very likely, in the event of a claim that exceeds the property damage, the carrier that pays it can then sue, garnish wages, etc.  None of these are positive.  None of these are healthy for the finances of the overall population.  None of these are good activities for expense ratios. But, if a contract is not enforced, if consequences for bad choices are ignored, where are we left?
We’re left where we are now.  With companies who may or may not know who they want to insure.  Then, they apply data to that uncertainty and it seems to be when in doubt, apply a higher rate.  But it gets better, see when they cannot afford that higher rate and then cancel that policy, the new company can choose to negatively impact the rate for this person. This now locks them into a typically higher rate for another six or so months, maybe longer.  Not good for anyone.
Now, in theory, they could have provided him with coverage and collected premium, maybe secured with a telematics device?  What if we took a longer play and said, you know, this person fits all our criteria and seems to have had a rough stretch 2 years ago.  But, no claims were paid out and there were no accidents reported.  Looks like he/she is in college and engaged now(we know because we are using data confidently and to the benefit of the customer, not the carrier) maybe this is a person we should work with?  

No solutions, just concepts that can be used better.