3 ms·

Punishments should always be greater than the financial damages, because otherwise the expected value of fraud will always be positive.

lazaroclapp 10y ago
It depends. Do you consider loss of reputation part of the punishment, or just an additional natural consequence of the fraud? If you want to think like an economist about fraud, in terms of pure monetary cost/benefit (which is not how people, even CEOs, think), then the game is something like this: Every year you can get $X by being honest, or $(X+Y) by committing fraud, you have probability p of being caught every year (assume: independent) and will have to pay $Z if caught, per year you stole $Y. Additionally, because of loss of reputation, you will only be able to earn at most $W honestly afterwards, instead of $X (assume: it won't be possible/reasonable for you to get away with fraud a second time). Thus, so long as i(X+Y-Z) + W(years_of_career - i) < Xyears_of_career any time (1-p)^i > 0.5, fraud is disincentivized. Sure, you can make the model more complex by adding the extra interest of earning an additional $Y early in your career whether or not you can leverage fraud to be promoted, but you still end up with an equation in which being finned and* removed from your position is rarely worth it if: a) the probability of being caught is high, b) the fine is in the ballpark of the damages (Y+\epsilon due to moral damages and legal fees) and c) being found out comes with an extra-legal reputation cost. Also, at the levels we are talking about here, is money anything more than a proxy for ego/reputation? Would you rather have $10 million and be well regarded or $40 million and infamous?
- crdoconnor 10y ago
  >If you want to think like an economist about fraud You don't. Thinking like an economist about control fraud is the exact mistake Alan Greenspan made when he denied that fraud could exist because an efficient market wouldn't allow it: http://www.dailykos.com/story/2010/11/7/918589/- http://www.dailykos.com/story/2010/11/7/918589/- His fatal assumption was also that reputational risk was enough to prevent it from ever happening.

lazaroclapp 10y ago

It depends. Do you consider loss of reputation part of the punishment, or just an additional natural consequence of the fraud? If you want to think like an economist about fraud, in terms of pure monetary cost/benefit (which is not how people, even CEOs, think), then the game is something like this: Every year you can get $X by being honest, or $(X+Y) by committing fraud, you have probability p of being caught every year (assume: independent) and will have to pay $Z if caught, per year you stole $Y. Additionally, because of loss of reputation, you will only be able to earn at most $W honestly afterwards, instead of $X (assume: it won't be possible/reasonable for you to get away with fraud a second time). Thus, so long as i(X+Y-Z) + W(years_of_career - i) < Xyears_of_career any time (1-p)^i > 0.5, fraud is disincentivized. Sure, you can make the model more complex by adding the extra interest of earning an additional $Y early in your career whether or not you can leverage fraud to be promoted, but you still end up with an equation in which being finned and* removed from your position is rarely worth it if: a) the probability of being caught is high, b) the fine is in the ballpark of the damages (Y+\epsilon due to moral damages and legal fees) and c) being found out comes with an extra-legal reputation cost. Also, at the levels we are talking about here, is money anything more than a proxy for ego/reputation? Would you rather have $10 million and be well regarded or $40 million and infamous?

crdoconnor 10y ago
>If you want to think like an economist about fraud You don't. Thinking like an economist about control fraud is the exact mistake Alan Greenspan made when he denied that fraud could exist because an efficient market wouldn't allow it: http://www.dailykos.com/story/2010/11/7/918589/- http://www.dailykos.com/story/2010/11/7/918589/- His fatal assumption was also that reputational risk was enough to prevent it from ever happening.