Takeout is an issue that divides North American racing, yet research on the subject has not caught up to the modern simulcast market. Following TRC Publisher Charles Hayward’s commentary piece on takeout, research and analytics professional Dan Needham, a longtime racing fan, horseplayer, and founder of research and analytics firm Thorotrends, outlines one way forward.
The importance of pari-mutuel takeout percentages remains a contentious topic in North American Thoroughbred racing and beyond. The shelf life of existing economic literature on takeout, while likely expired, continues to inform us that lower takeout rates will result in increased wagering handle. The language in this literature, replete with Laffer Curves and elasticities, has not been persuasive.
But, if you believe that takeout rates play a pivotal role in driving pari-mutuel handle, the paramount question becomes: "How should racetrack operators and regulators decide upon specific percentages?"
Many who maintain that current pari-mutuel takeout rates are too high do not have a ready answer to that question. And those who do will often base their answer on little more than intuition.
A cursory survey of takeout rates across racing nations only makes the issue more opaque. The blended takeout rate in North American pari-mutuel pools, where handle has eroded dramatically in the last decade, is approximately 21 percent. In Japan, where Thoroughbred racing is vibrant and wagering handle remains robust despite declines in recent years, the takeout is approximately 25 percent. The lesson here is not that those who oversee wagering in Japan understand takeout rates better than their counterparts in the United States, rather, every racing jurisdiction is different. While the sharing of racing insights across nations is certainly worthwhile, it's essential to be mindful of each unique environment. What works in Japan, Hong Kong, or France, may not work in the United States, and vice versa.
With that in mind, we'll limit our focus to the wagering environment surrounding American racing. Nevertheless, a common call to action regarding takeout rates is applicable wherever pari-mutuel wagering occurs.
In 2011, management consulting firm McKinsey & Company presented a plan to The Jockey Club designed to drive sustainable growth in the Thoroughbred racing and breeding industry. The plan detailed a host of different initiatives to increase awareness of Thoroughbred racing and to drive handle. Many observers expected that McKinsey, a world-class consulting firm with vast analytical capabilities, would surely include pari-mutuel takeout rates among its areas of focus. Racetrack operators and state regulators would finally be furnished with expert guidance and data-driven insights based on a rigorous analysis of the modern pari-mutuel wagering landscape.
Instead, while dutifully noting that takeout rates are of high concern to heavy bettors, McKinsey observed that only a small minority of racing fans understood issues around takeout rates.
"We strongly prefer rebates as the method to address the price-sensitive bettor," we were told. And with that, and with The Jockey Club's tacit approval, the McKinsey team passed on pari-mutuel takeout rates.
We are left to imagine what McKinsey might have been able to accomplish had they deemed takeout rates worthy of study. And now, we'll begin to imagine how they may have approached that challenge.
The idea that optimal takeout rates exist and can be estimated is straightforward. There are price points that are more favorable than any other. More favorable to whom? Not the bettors. They would strongly prefer 0 percent. The optimal takeout rates are those that result in the highest possible revenue to the racetracks via increased handle.
How might we discover optimal pari-mutuel takeout rates? Let's go back to the beginning for some clues.
Pari-mutuel wagering was invented in France in 1867 by Joseph Oller. We know that Oller's commission, i.e. takeout, was 10 percent. But Oller made a very interesting adjustment to his commission on the busiest racing days when wagering volume was particularly high. He cut his take in half to 5 percent. We can't explain with certainty Oller's motivation in dropping his standard commission by 50 percent on large-handle days. But a logical explanation would be that Oller understood that churn (winnings that are repeatedly re-bet) worked to his advantage in the long-term.
Today, the wagering landscape is infinitely more complex compared with Oller's time. Simulcasting, advance deposit wagering (ADW), and the proliferation of exotic wagers have forever tangled the system. Meanwhile, the derivation of takeout rates has remained simplistic and capricious.
You could rightly argue that nearly 150 years ago pari-mutuel wagering pioneer Joseph Oller manipulated takeout percentages in a more sophisticated manner than we see today.
Takeout rates in American racing were set to an "Oller-ian" level or below during the sport's early years of expansion. The notion that takeout must be as low as possible for racing to succeed was widely supported by racetrack operators in the 1930s. Nevertheless, takeout rates would creep higher through subsequent decades even as less expensive gambling options began to proliferate.
Today's higher takeout rates are a byzantine product of statutory taxes and a myriad of fees. They have increased artificially rather than through any kind of data-driven knowledge.
Discussions of optimal pari-mutuel takeout rates inevitably drift to a comparison with slot machine "hold." Hold, typically less than 10 percent, is simply the percentage of coins-in that the casino keeps over the long-term. While hold and takeout are essentially identical terms, slot gaming and pari-mutuel wagering are worlds apart. Comparisons between the two are therefore usually wrongheaded.
However, we can certainly ask why slot machine hold does not vary significantly across the gaming industry. The answer may potentially help us on the pari-mutuel side. Is there foundational research and scholarship that informs decisions on hold percentages? Is there lively debate about optimal hold percentages? The surprising answer to both of these questions is "No."
Instead, slot machine hold percentages are generally market-driven. A casino manager is simply unable to significantly increase hold without customers "noticing" and taking their business elsewhere.
Similar to slots players, sports bettors are comfortable paying a 10 percent "vig" on their wagers. What price would entice these types of gamblers (as well as poker players) to test the waters in pari-mutuel pools? This is an important question that can be answered through research and experimentation. Research would help us to understand how important the cost of wagering is to those already predisposed to gambling in general. Demand forecasting could help us predict the impact on handle from "new" money attracted by lower takeout rates.
Large-scale experimentation with lower takeout rates is one way to approach the problem. The New York Racing Association, Inc., Churchill Downs, Inc., and the Stronach Group are each viable candidates but unlikely volunteers.
Of great importance to such an effort is unfettered access to player profile data through an ADW provider. Only the ADW companies can bring player-level insights to bear on key concepts such as churn. Unfortunately, a "global" picture of customer behavior is not guaranteed since bettors tend to maintain multiple wagering accounts.
Experimentation would require careful planning by a team of analysts, econometricians, and industry experts. Experimental methods would attempt to measure the effect of lower takeout rates on handle while controlling for other factors. Large-scale experimentation will require time, resources, and cooperation – especially from regulators. And the challenge of measuring incremental demand from other forms of gambling will have to be an element of the experimental design.
If large-scale experimentation into takeout rates proves too unwieldy or politically unpopular, simulation may be an excellent alternative. Simulation parameters can be informed by market research and small-scale experimentation. Compared with large-scale experimentation, simulation may require less time and would be less invasive. However, the results of simulation may be considered less authoritative than experimentation.
In addition to lower rates, one can easily imagine various unconventional takeout rate schemes that might be modeled or used in experiments. Perhaps dynamic takeout rates tied to field size or the size of the wagering pool. Or both.
Finally, another – albeit fanciful – scenario involves no research whatsoever. One of the "Big 3" racing organizations, with cooperation of state regulators, might announce a takeout rate cut to 10 percent across its wagering menus. Bettors would swarm their pools, and as with slots, market forces would rule the day. All other racing jurisdictions would have no choice but to follow suit. What may happen next is up to our informed imaginations.
The larger point remains: Pari-mutuel takeout rates in Thoroughbred racing, the proverbial low-hanging fruit of opportunity, present a topic in dire need of creative quantitive scrutiny. It is an issue that affects every bettor. Every bettor is price-sensitive since he or she has only so much money to wager.
Joseph Oller's grave is inscribed with the phrase "Son intelligence égalait son coeur," meaning "His intelligence equaled his heart". Those involved with the stewardship of the Thoroughbred racing industry have plenty of heart. That is not why the industry has declined. It is, perhaps, the unwillingness to approach problems with a maximum dose of intelligence and sophistication that has contributed to the decline.
---
Dan Needham is the principal and founder of Thorotrends, LLC, a research and analytics firm founded upon the premise that the Thoroughbred racing industry must become both customer-centric and research-driven in order to recapture lost market share, interest, and awareness.