One of the conclusions:
- The “house money effect” [shows that] bettors ..... mostly spend the money they have won.
This is, in fact, a big reason why slot machines were taxed, at one time, at 25% juice, but this rate went lower and lower. The people playing to break even were not enjoying themselves at 25% and the casino was not making as much as it should've been. At 5% you had a whole lot of people on the floor chasing break even. Chasing is maximizing utility.
This theory is relatively simple, and it's not solely used for racing, or slots.
In finance, Tobin's Q is (loosely) the market value of a stock divided by the firm's replacement value. Companies with a low Tobin's Q use the house money effect to chase gains, and even when they have minor losses in a given year, will increase their risk, chasing break even the following year. Companies with a high Tobin's Q do not exhibit this behavior.
In racing, low Tobin's Q users are your casual user. People like to say "they don't care about takeout" but they do, because they always know where they are when it comes to break even, and when they have (or don't have) house money. Higher takeout moves them away from break even, and in response, over time they leave the market.
This is also why, in my view, jackpot bets are so terrible for the casual bettor. When you take a bunch of a user's money and tie it up, they can't exhibit the behavior they are predisposed to exhibit. Then, adding to the deleterious effect, jackpot money is not spread out for masses to chase break even, it's sent to one winner and he or she puts much of that money away.
When I type, as I often do here, "people don't take their incremental winnings and stick them in a sock" this is what, more or less, I am referring to.
When you give people more back for them to bet, they bet. This is why, 100 out of 100 times, takeout rate decreases increase handle.