3 ms·
It doesn't matter to an insurance company how likely car crashes are -- that just determines what the company charges in premiums. All that matters is that the
by a-priori 13y ago
It doesn't matter to an insurance company how likely car crashes are -- that just determines what the company charges in premiums. All that matters is that the likelihood is well understood and that the company has enough customers to stay solvent.
Insurance companies essentially work by the Law of Large Numbers: that in a large enough number of trials, the average value should approach the expected value. In real world terms, if the chance of a house burning down in a fire is 1 per year per 1000 homes at an average cost of $200k each to repair (totally bogus numbers), then an insurance company can break even by charging a homeowner $200/year to insure the home against fire.
Over a large enough number of homes, that should be reasonably close to the actual amount they have to pay out. Tack on a profit margin, and you have a business model.
Edit: Fixed numbers a bit. Thank goodness I'm not an actuary.
- gohrt 13y agoI think you overlooked that in GTA we have a > 100-year-storm of destruction In the game world, insurance companies wouldn't have predicted the level of claims.
- a-priori 13y agoI haven't played the game. If the game's premise is that there's a one-of-a-kind rampage going on that's causing all sorts of mayhem that couldn't have been predicted, then I agree that buying shares in an insurance company would be a bad idea. But if the game's premise is that the mayhem is normal within the world, then insurance would be stupid expensive. But, assuming the market can bear those prices, insurance would be a good business to be in.
- jlgreco 13y agoIt's a game where getting into a fender bender with a cop car often results in hundreds of blown up cars, numerous car-jackings, and several crashed helicopters an hour or two later, but only if your character is involved.
- drbaskin 13y agoI'm sorry to be so pedantic, but your math doesn't seem quite right. Should it instead be (1/100,000 chance of burning down each year) * ($200,000 cost if it does) = $2 per near expected damages?
- a-priori 13y agoErr, yes you're right. I was off there by a couple orders of magnitude :)
- brads84 13y agoInsurance makes money on float and usually loses money on insuring something over the long term. I can't remember the exact numbers, but GEICO might payout something like 103% (or 3% MORE than you paid them) over the term of your policy with them. How do they make money? http://seekingalpha.com/article/419731-buffett-on-insurance-and-investing-its-about-the-float http://seekingalpha.com/article/419731-buffett-on-insurance-... Basically they invest your premiums during the time you are paying them and make money off of holding your money. Then the other insurance agencies might lower their costs and end of paying out 104% (or whatever) of what you've paid in. This competition pressures the market to close the gap between what you can make on the float and what you can charge your customer, but there is going to be some gap and that is the profit margin. If the insurance company can make a high ROI on the float, they can charge less in insurance rates and still make money.
- tanzam75 13y agoInsurance companies usually lose money on underwriting. However, GEICO is not a good example, because it is the exception to the rule. GEICO usually makes an underwriting profit -- often a very large underwriting profit. In fact, all of Berkshire Hathaway's insurance operations make a long-term underwriting profit. Unlike GEICO, the reinsurers do not make a profit every year, because hurricanes and earthquakes are unpredictable. (Whereas car crash statistics change very slowly from year to year.) But averaged out over a long period of time, they make a profit. This is the secret of Warren Buffett's success as an investor. He is essentially buying stocks on margin, at a negative interest rate. Data collected from Berkshire Hathaway annual reports: 2012: $680 million profit on $11,578 million float 2011: $576 million profit on $11,169 million float 2010: $1,117 milion profit on $10,272 million float 2009: $649 million profit on $9,613 million float 2008: $916 million profit on $8,454 million float 2007: $1,113 million profit on $7,768 million float 2006: $1,314 million profit on $7,171 million float 2005: $1,221 million profit on $6,692 million float 2004: $970 million profit on $5,960 million float Not broken out separately prior to 2004 This profitability streak is truly exceptional. As Warren Buffett pointed out in his 2012 letter, State Farm has made an underwriting loss in 8 of the previous 11 years. The insurance industry as a whole has made an underwriting loss in 37 of the previous 45 years.