4 ms·

> But the general trigger of insurance for people is the non-linear utility function. I would respectfully disagree. Homeowners insurance is a great example.

> But the general trigger of insurance for people is the non-linear utility function. I would respectfully disagree. Homeowners insurance is a great example. Most people buy it and never make a claim, and they "lose" insofar as their premiums are gone for no monetary return. But some people who buy it "win" and get to collect enough money to roughly replace their house, belongings, etc when a hurricane, tornado or whatever destroys it all. Not all insurance is the kind where you pay the same amount whether you make 1000 small payments or one big one.

ucaetano 11y ago
"Not all insurance is the kind where you pay the same amount whether you make 1000 small payments or one big one." But I never said that! It's just like the lottery, the payout of the lottery is smaller than the price divided by the odds (that's why a lot of people call it a tax on people who don't know math), but as the utility function of money isn't linear, for most people a small chance of gaining a lot of money is more valuable than a little bit of money. The same is valid for insurance. The value of losing a lot of money is far more negative than the cost of it, even when taking the risk in account. That's because the utility function doesn't depend only on the gain, but also on how much you have. A billionaire might not care about home insurance for a 100k shack.
- msandford 11y ago
  Even if the utility function was linear, I'd still prefer to insure some things, even with a negative expected value, for the sake of smoothing variance. EDIT: I'll clarify this to mean; even if the EV of insurance was exactly neutral, I'd still buy it for smoothing. And even if the EV was negative, I'd still buy it for smoothing. There's a difference between smoothing and loss aversion.
  - ucaetano 11y ago
    Not for smoothing variance, as you don't care about positive veriations. You'd buy it to limit your downside. If people would buy it to minimize variance, they would also pay an insurance against winning the lottery.
    - msandford 11y ago
      Some people choose to take 20 years of monthly payments from lottery winning rather than one lump sum. "Rationally" anyone who does that is a fool, and yet people constantly do that. Perhaps there's value to variance minimization, aside from the non-linear utility function of money?

ucaetano 11y ago

"Not all insurance is the kind where you pay the same amount whether you make 1000 small payments or one big one." But I never said that! It's just like the lottery, the payout of the lottery is smaller than the price divided by the odds (that's why a lot of people call it a tax on people who don't know math), but as the utility function of money isn't linear, for most people a small chance of gaining a lot of money is more valuable than a little bit of money. The same is valid for insurance. The value of losing a lot of money is far more negative than the cost of it, even when taking the risk in account. That's because the utility function doesn't depend only on the gain, but also on how much you have. A billionaire might not care about home insurance for a 100k shack.

msandford 11y ago
Even if the utility function was linear, I'd still prefer to insure some things, even with a negative expected value, for the sake of smoothing variance. EDIT: I'll clarify this to mean; even if the EV of insurance was exactly neutral, I'd still buy it for smoothing. And even if the EV was negative, I'd still buy it for smoothing. There's a difference between smoothing and loss aversion.
- ucaetano 11y ago
  Not for smoothing variance, as you don't care about positive veriations. You'd buy it to limit your downside. If people would buy it to minimize variance, they would also pay an insurance against winning the lottery.
  - msandford 11y ago
    Some people choose to take 20 years of monthly payments from lottery winning rather than one lump sum. "Rationally" anyone who does that is a fool, and yet people constantly do that. Perhaps there's value to variance minimization, aside from the non-linear utility function of money?