Takeout - What's Optimal?

Everyone knows that the takeout -- the amount of total bets that the track holds onto before paying out the balance to holders of winning tickets -- is too high.The competition -- slot machines, casino table games, poker, sports betting -- typically sets takeout somewhere between 5% and 10%, while US racetracks have a blended takeout, averaging the different levels on different types of bets, of somewhere just over 20%. It doesn't take a genius to see that we're pricing our product too high.

Sure, if you have a better product, you can often charge a higher price. But is our product so superior to that of the slot parlors, casinos, poker rooms or sports books? Those of us who love Thoroughbreds and who admire the athleticism and competitive spirit of the horses and the skill and courage of the jockeys might say so, but we're probably not a very big slice of the population. And even then, we should question whether the product offered on a Saturday at Saratoga or Keeneland is no better than the Tuesday card at Finger Lakes, even though we may charge the same price (i.e., takeout) for all of them. It's no wonder that thoroughbred racing handle has stagnated over the past few decades -- actually, measured in inflation-adjusted dollars it has decreased -- while the amount bet on alternative forms of gambling has increased. The competitors offer a product that many people find to be just as good -- in the eyes of some even better, because it doesn't have the steep learning curve that betting on horses does -- and they offer it at a cheaper price.

So, with the exception of some horsemen's groups that are oblivious to the economics of betting and some track owners who don't care that much about profits from racing, because they're really in the ADW business or in slot machines or in producing games for smart phones (see Churchill Downs, Inc.), most people in horse racing who think about the takeout issue agree we should do better. The more difficult question is how to adjust our prices. There have been lots of economic studies of "optimal takeout," but, like most economists' studies, these incorporate unreasonable assumptions that ignore where bets are placed and what kinds of bets they are. Some practical experimentation is needed.

Recently, some tracks have introduced "low takeout" (i.e., 14-15%) bets, typically multiple-race wagers like the Pick 5. A few (Canterbury, for example), have tried reducing takeout across the board.  So far, these experiments haven't demonstrated that price-cutting really pays. But have we done the kind of analysis that would suggest more effective approaches? I think not.

As an example, let's take the takeout structure and the distribution of bet types at the NYRA tracks -- Aqueduct, Belmont and Saratoga. Single-horse bets (win, place and show) have a 16% takeout, two-horse bets (exacta, quinella and daily double) are 18.5%, and multi-horse bets, from a trifecta or Pick 3 through the Pick 6, are 24%, except that the Pick 5 (all the time) and the Pick 6 (on non-carryover days) are 15%. You can see the overall effect by looking at the NYRA financial reports. For example, in the second quarter of this year, NYRA's "on-track" handle was $165 million, and its gross revenue, i.e., the takeout, from that amount was $33 million, almost exactly a blended takeout rate of 20%.

In addition to that "on-track" handle (half of which, by the way, was bet through NYRA's own "NYRABets" online and phone betting app, rather than actually being handed across the teller windows at the track), NYRA's 2nd-quarter handle also included $460 million from simulcast and ADW revenue, from which it earned another $33 million. So its share of the takeout from exporting its signal was only about 7% -- better than for lower-class tracks, but a far cry from the 20% it makes on its home turf (or dirt as the case may be). And then NYRA also counted $56 million from bets placed on NYRA races at the various OTBs across New York State, on which it netted $6 million, or just under 11%. The difference between the overall 20% takeout and the 7% or 11% that NYRA actually receives from off-track and off-network bets represents money that is available to the entities receiving the NYRA signal to use for their operations or, importantly, use for rebates to big players. NYRABets' own rebate plan, according to its financials, costs less than 1% of the amount bet through its network, but for some single-purpose ADWs, especially those effectively owned by the bettors themselves, it can be as high as 10%.

So cutting takeout overall would have different effects depending on where a bet is placed. For bets on-track or on-network, takeout could be cut 5% at each level, to, say, 11%, 13.5% and 19%, without necessarily bankrupting the track, and allowing time to see if the takeout cut would pay for itself in increased handle. But if NYRA cut takeout by that much and still charged the ADWs the 7% sending fee that its financials suggest, then what would happen to the big rebates that the "whales" get through their ADW bets? At present, these "whales," mostly using sophisticated computer/robotic wagering programs to exploit small leverage opportunities in the pools, account for perhaps 20% of total US handle, concentrated in the major tracks, because that's where the pools are large enough to accommodate their big bets. If, say, NYRA cut its win takeout to 11% and still charged the ADW 7%, that would effectively limit rebate possibilities to 3-4%. Would the whales still be at that rebate level? Maybe, because they would face the same net, post-rebate, takeout as at present. But we don't know for sure.

And the effect of a takeout cut also varies by the type of bet. One of the reasons groups like HANA (the Horseplayers Association of North America) favor cutting takeout is that it would increase "churn," the number of times a player can keep betting before her money runs out. But if takeout is cut on a Pick 5, then that money, even if low-priced, is tied up for five races. Better than all day, but perhaps not so much. If churn is a factor,one would expect the effects of low takeout to be greatest at the single-horse bet. Yet few tracks have tried it there.

NYRA's most recent quarterly financials don't include a balance sheet. (By the way, it's now December; where are the 3rd quarter financials for the period that ended two months ago?) But still, it's reasonable to assume that there's a bit of cash in the bank. While I think an overall reduction of at least 5% is probably the right number, it would be great to see one of the major players in the game, like NYRA, try an across-the-board takeout reduction at, let's say 3%, reducing win place and show bets to 13%, two-horse bets to 16.5%, and everything else to 22%, for a blended rate of 17%. Even though that's still way more expensive that the non-racing competition, it might be enough to move the needle and start pulling in additional handle, perhaps even enough to make up for the price reduction. And let's pledge to keep the experiment in place for at least a year, so as to give it a fair test. That would go a long way to settling some of the so far mostly theoretical arguments about optimum takeout.

I know, I know. Changing the takeout, especially in New York, isn't easy. One has to go to the Legislature, the Gaming Commission, etc. etc. But NYRA's authorizing legislation for converting back to a private entity will be coming up in Albany in the next session anyway. Why not use the occasion to free NYRA to adjust takeout as it sees fit and maybe, just maybe, answer the question about whether a significant takeout reduction would bring gambling money back to the races.