How insurance works is described most easily for life insurance. The ideas are very similar for health insurance, but much more complicated to calculate or explain.
So, pretend that we’re an insurance company, and we have 1000 policy-holders, all age 35, each with a $100,000 policy. We look up on a table and find that the probability of a person age 35 dying before they reach age 36 is 1/1000.
So, we expect that our group of 1000 people age 35 will experience one death this year (one “claim”), and we’ll have to pay out $100,000 to the beneficiary. If we collect $100 per person as premium, we’ll have collected $100,000 and we’ll have just enough to pay the $100,000 claim.
So, we’re going to charge each person $100 in premium for the year. OK, we have to charge a bit more to cover our expenses, profit, administration, etc., but let’s ignore that for our example.
The problem is that we’re working on probabilities. We could have bad luck, and have two or three deaths from our group of 1000 people age 35. In that case, we would have to pay off $200,000 or $300,000 and we have NOT have collected enough premium, and we’d be in big trouble.
There are things we can do to protect ourselves, such as charging more  – but if we charge $200 per person, or $300, we protect ourselves against adverse claims but may no longer be competitive, and we’ll lose business to the competition. We can also, as Hemlock says, insure ourselves – find another insurer who will take on the risk of us having more than one claim. That will cost us, and so will raise our premium, but would be a good protective device.
Life insurance is fairly predictable for very large groups (it’s just a probability experiment.) Thus, the arrangement I’ve described is usually fairly successful.
When it comes to health insurance, however, the situation gets much more complicated. Life insurance statistics are pretty simple – people either live or die, there’s not a lot of quibbling. A claim means someone died and gets paid the fixed amount.
With health insurance, we have a wide range of claims, from small amounts like for a doctor’s visit, to very large amounts, like for heart transplants. For some of these, we can look at past statistics of the number of occurrences, but there can be a lot of variation. For instance, there’s new technology, which is available faster than the statistics about its use. There are subjective factors – an article about X treatment in a local newspaper usually means a lot of people suddenly ask for X treatment. There are other subjective factors – state of health or how someone feels is not objectively measured.
So, if there have been years and years of claims exceeding premiums, the insurance company is in major financial trouble… even if they’ve been charging a lot in premiums.
That explanation help?