What are the determinants of television ratings for NASCAR Nextel

Joshua Van Cleef ECO 4451 What are the determinants of NASCAR Nextel Cup Television Ratings? April 9, 2004 What are the determinants of NASCAR Nextel Cup television ratings? Abstract Automobile racing has been one of the fastest growing sports in the United States for many years, and has been led by the National Association for Stock Car Auto Racing, or NASCAR, with the Nextel Cup series as its crown jewel. Today, primary sponsorship of one car in the Nextel Cup series costs upwards of twenty million dollars, and advertising costs for media coverage of the events continues to climb. This analysis attempts to isolate those factors that cause television ratings for one race to be higher than those of another, which is useful information to many interested parties. I have built a regression model incorporating a broad array of variables, some surrounding the races themselves, and others focused on outside influences on a viewing decision. The final results shed light on reasons that viewers are drawn to races, and some of the factors that keep them away. Automobile racing has been one of the fastest growing sports in the United States for many years, and has been led by the National Association for Stock Car Auto Racing, or NASCAR. NASCAR consists of many layers of regional and national level racing, but the pinnacle of the organization is its Nextel Cup series (formerly known as Winston Cup). As the sport has grown in popularity, the advertising dollars thrown toward it have grown even faster. Today, primary sponsorship of one car in the Nextel Cup series costs upwards of twenty million dollars, and advertising costs for media coverage of the events continues to climb. As in other sports, television advertising is a key opportunity for a firm to get a message out to a reasonably predictable demographic. However, over the course of a season that spans 10 months, race viewership moves up and down, which must impact the decisions made by advertisers. The purpose of this analysis will be to isolate those factors that cause one race to be more popular than another. A successful analysis of these factors would be useful to a wide variety of people in and around motorsports and marketing. Accurate forecasts of television ratings would allow broadcast networks to set advertising rates for their telecasts. Advertisers could select which races to place emphasis on based on their goals and budgets. Team managers of major teams could sell additional sponsorships for key races, and low-budget teams that only enter selected races could choose events that would lead to the best exposure. Determining these factors will allow a buyer to estimate which races will draw the highest ratings and make their decisions accordingly. The final objective will is to have a functional model that will take in variables for a race and forecast the television ratings for that event. Literature Review In pursuit of the determinants of television for NASCAR ratings, I found that previous work on the topic has been sparse, particularly in the realm of primary research studies. However, a broader search across the world of sports did yield some useful results. Much of the research on television revenue and ratings has been done using the National Basketball Association, but some of the findings hold value against the backdrop of NASCAR. The key difference and resulting challenge, however, is that the NBA, as well as all of the other premier sports leagues are team sports, made up of approximately 30 teams. Each week, certain teams are featured on national broadcasts, while the others are left to local networks or not aired at all. NASCAR, however, features the same competitors from week to week racing at different venues around the country. One aspect of my research will be looking at the data compiled on other leagues, and finding connections that can be made to different arrangement of NASCAR. Hausman and Leonard (1997) did a study of the impact of superstars on NBA revenues at large, and included a discussion of television ratings. The found that television ratings were increased approximately 30% for games that featured one of 3 “superstars” that they designated. This finding is obviously an important one, but on the surface NASCAR would have little use for it- all of their drivers compete each week, barring an injury or other unlikely event. However, unlike team sports, each venue that the series visits is different. Various racetracks on the circuit feature different speeds, levels of contact between cars, historical legacies, etc. Is it possible that some tracks could hold the same effect on ratings as a superstar in basketball? This is something I will attempt to analyze in the future. Caudill and Mixon (1998) produced a paper discussing the length of the basketball playoffs, and included mention of the ratings effects of longer playoff series. Their research suggested that in NBA best-of-seven playoff series (the first team to win four games wins the round), the ratings for a game 6 are consistently much higher for those of a game 5. Thus, it benefited the NBA when they modified their playoff structure slightly, resulting in more game sixes. The applications of this information to NASCAR are more direct than the superstar question. Under the system that NASCAR has used for quite some time, championship points are tabulated from the first race, each event counting equally. In some cases, the eventual champion establishes a clear point lead early, meaning that the championship battle is not compelling near the end of the season. Other years, the point title has come down to the final laps of the season, and it could be assumed that ratings are better when this happens. Recently, NASCAR has announced a new points system for 2004, in which the top ten competitors are set virtually equal with ten races to go. The Caudill and Mixon study suggest that this will boost ratings, particularly in the later races. This remains to be fully seen and has not yet been explored in the literature, but I hope to devise a way to consider it in my research. The article from Consli (2003) offers another perspective for consideration. In its discussion of sports television ratings across the industry, it notes that NASCAR ratings were flat in the first half of 2003, but that the industry as a whole were down. However, A Fox TV executive noted that major events in Iraq occurred on the same day or weekend as races they broadcasted, suggesting that significant world events would draw away from their broadcast. This article offered little in the way of results, but certainly enters a new variable into the analysis. The article from Gold (2003) focused on tennis, but again offered insight into viewership for sports in general, and is somewhat more relevant because tennis players compete individually, so some of the challenges are the same. Professional Men’s Tennis is currently suffering due in part to a lack of charismatic stars to grab the public attention. This comes as no surprise, particularly in light of the Hausman and Leonard study pertaining to superstars. What this suggests is that these individual sports must try to do is generate star power for its competitors. NASCAR has been very successful at doing so in the last ten years or so, particularly in transitioning from one star (Richard Petty and Dale Earnhardt, Sr. generally dominated in fan support and attention) to sustaining multiple stars and generating wider appeal. Accounting for this may be difficult statistically, but it is something that needs to be done for the research to be thorough. The article by Ewing and Passariello (2003) pertains to Lance Armstrong and the sport of cycling, another individual racing sport, much like NASCAR, except that it’s U.S. following is generally limited to one event per year, the Tour de France. Armstrong has been dominant at that event, and as a result the marketing interest and European television ratings have been declining. However, interest in the both the Tour de France and cycling in general in the U.S. has increased each year that Armstrong has been successful. What we see is that as a new popular star rises, interest will increase steadily, but that there is some sort of maximum that the public will tolerate when it comes to dominance. In Europe, Armstrong is not nearly as popular, and the Europeans have grown bored of watching him win. Such a problem must be a concern for NASCAR, but as yet has not been too apparent. Depending on the scope of the final analysis, it may be insightful for my research to evaluate NASCAR data going back to seasons where one driver dominated, and see what the effects appeared to be on television ratings. The absence of any in-depth discussion of what drives NASCAR ratings was surprising, as this is the type of information that most advertisers would love to have. Using the information and ideas offered by the research that has been done before, I have formulated the framework for a regression model that opens the door to an analysis of NASCAR and it’s viewers. Discussion of Methods Building the formula for NASCAR ratings was intended to be a 2 step process. The necessary ratings information was on the web at motorsportstv.com, and NASCAR.com had all of the individual race statistics that would be a part of the analysis. However, motorsportstv.com is now defunct. A great deal of work was now required to find the Nielsen ratings information, because no publicly available ratings database exists at this time. Two sites were used in compiling the ratings numbers. One is that of a NASCAR watchdog/news source, www.jayski.com, and the other is SBR, a sports business database. Extensive searches of the material offered on these sites, only in the form of newspaper style articles, resulted in the ratings information shown. NASCAR.com was vital in providing the other statistical information. The following are the variables used in building the regression model: RATING: The Nielsen TV Rating for the broadcast of each race, and the dependent variable of the equation. Each rating point represents one percent of the potential viewers who had access to the broadcast (ie, have televisions and have the channel that the broadcast is on). ROUND: The championship round in question Months: MARCH through NOV, dummy variables for each month except February, which serves as the default Days: Dummy variables for the day and time of a weekend that a race is held: SAT, for Saturday night prime-time races (none are run on Saturday afternoons), SUNNITE, for races run in prime-time on Sundays (generally before Memorial Day), MONDAY, for races run and televised on Mondays due to rainouts from the weekend, and FRIDAY, for the one instance that a race was held the Friday following Thanksgiving (the event was originally scheduled for the week following September 11, 2002). Sunday afternoons are the default value. AGE: Representing the number of years the race has been run. PURSE: The money that the winning driver collected. Track Types: Dummy variables for the type of track the races are held on: PLATE- A track 2.5 miles in length where restrictor plates are used on the cars’ carburetors, causing very close and high speed racing ROAD: A racetrack consisting of left and right turns. SHORT: A track one mile or less in length. A standard speedway, between 1 and 2.5 miles, is the default value. Tragedies: Dummy variables, TRAG1, Indicating that a fatality occurred the week before, TRAG4, indicating that a fatality occurred within 4 weeks, and TRAGYR, indicating that a fatality occurred during that season. WAR: A dummy variable, marked yes if we are officially at war. For this time period, wars in Afghanistan and Iraq were counted, during the timeframes the government officially announced TERROR: Dummy variable used during the time period after September 11, 2001, but before the war in Afghanistan began. Other sports: Dummy variables while other sports are in season: NFLREG, for the National Football League’s regular season, NFLPRE, for the NFL preseason, MLBREG, for Major League Baseball’s regular season and MLBPOST for their playoffs, NBAREG and NBAPOST for the National Basketball Association, and NCAAS, representing the NCAA Men’s Basketball Tournament. AVSPEED: Finally, the average speed of the race winner. Limdep software was used to run the regression using all of the above variables. Results & Analysis Variable Constant ROUND MARCH APRIL MAY JUNE JULY AUGUST SEPT OCTOBER NOV YEAR SAT FRIDAY SUNNITE MONDAY AGE PURSE PLATE ROAD SHORT TRAG1 TRAG4 TRAGYR WAR TERROR NFLREG MLBREG NFLPRE MBLPOST NBAREG NBAPOST NCAAS Coefficient 7.29428 0.1027 -1.56069 -1.53149 -2.72872 -2.76454 -3.11282 -2.30908 -3.51564 -4.7019 -4.98167 -0.23602 -0.14694 -0.71012 0.480909 -2.08582 -0.00061 .138910D-05 1.031615 0.100026 -0.08092 -0.22169 -0.26884 -0.21781 -0.54438 0.752742 -1.47716 -1.16545 -1.01903 0.032063 0.002672 0.026028 -0.15049 T-Ratio 8.921 1.824 -4.463 -2.795 -3.722 -3.581 -3.32 -2.088 -2.794 -3.043 -3.003 -1.988 -0.67 -1.338 1.038 -5.324 -0.142 4.529 4.092 0.344 -0.382 -0.541 -0.574 -1.02 -2.941 1.873 -2.671 -3.02 -2.493 0.061 0.007 0.088 -0.571 AVSPEED 0.000912 0.28 The majority of the results seem to make sense, although some do go opposite of what might be expected: ROUND: The trend, on a seasonal basis, indicates that ratings generally rise as the season progresses. This is somewhat counterintuitive, as the Daytona 500 starts the season and was the highest rated broadcast in each season. However, the other variables seem to dwarf the trend. Months: Here is where expectations came through somewhat. Very clearly, the ratings are inclined to move downward each month as the season moves along, except for a notable uptick in August. This may be because the sporting world slows down in August- the baseball season is old, football is not yet in full swing, and the winter sports are far away. Otherwise, it makes sense that the ratings would fade as the season wears on. The championship battle spreads out, leaving fewer peoples’ favorite drivers in the mix, and the novelty wears off as the 36 race schedule gets older. YEAR: Viewership is moving downward slightly from year to year, which is the norm across the sports world. Days: While the low T-Ratios do not give much significance to the time each race is broadcast over the weekend, the results do make sense. When the races are broadcast outside of their normal times, fewer people would be expected to remember or have that alternate time available. The difference is on Memorial Day, when the race is on later. Memorial Sunday is traditionally dominated by auto racing, with the Indianapolis 500 occurring during the early afternoon and the NASCAR event being held in the evening, and into the night hours. Again, it is fitting that the ratings would be somewhat higher on that particular day. AGE: The historical value of a race has almost no bearing on television ratings. This ultimately is not startling, since the newest tracks are built to be exciting and can compete very well with the older tacks on the circuit. PURSE: Following along the theories behind reality TV, the ratings increase by approximately 0.1 for each $100,000 that the drivers could win. Type of Track: The only track that proved to have an influence on the ratings were the Restrictor Plate tracks (Daytona and Talladega). These are known for very close racing at high speeds. Additionally, these races are considered some of NASCARs most prestigious, particularly the two stops at Daytona. While much of that is hopefully explained in other variables, it may also have an impact on the Type variable. Tragedy: Surprisingly, the death of Dale Earnhardt in February 2001 did not have a significant impact on the broadcast ratings. While the ratings were higher over the course of 2001, very little of those increases can be directly linked to the fatality. It might be suggested that the influx of curious onlookers inspired by the tragedy was offset by the people who lost interest because of it. WAR and TERROR: War decreases viewership, but the period after the attacks on the United States saw increased ratings. This may be explained by the fact that sporting events proved very popular after the attacks, and featured patriotic ceremonies and events. On the opposite end, war coverage often cut in on race broadcasts, and the round-the-clock war coverage offered now draws people away from their regular viewing habits. Sport Seasons: Coming as no surprise is the fact that the NFL beats up on NASCAR ratings. By far America’s most favorite spectator sport, NASCAR is very aware that it pulls away from their ratings. On the NFL’s opening weekend, NASCAR holds its race on Saturday night to avoid going head to head with football’s debut. More surprising is that even the NFL preseason draws away from NASCAR ratings, as does Major League Baseball’s regular season. Basketball, however, has very little effect on NASCAR. Also, Baseball playoffs do not impact ratings significantly, although the races tend to overlap very little with baseball playoff games. AVSPEED: Average speed has virtually no effect on on whether or not a person is likely to tune into a race. This suggests that much more is involved in whether a race is watchable than just the speed, hence similar ratings above on short tracks and full size speedways. Conclusions While the original goal was to build a model that could be used to forecast ratings, as might be used by an advertiser or network executive, the results of this analysis have even broader implications. While they can certainly be used as a predictor of ratings, they also expose areas that NASCAR may want to explore improving upon. From the advertising perspective, these results highlight the most popular races of the season. Restrictor plate tracks consistently draw high ratings, as do the earliest races of the season. The month of August offers high ratings that buck the late season doldrums, and those events featuring large payouts draw large audiences. All of these areas are places that an investor or advertiser would do well to focus on. Also, they need not be particularly concerned about basketball events, including the NCAA or NBA playoffs. On the flipside, people depending on higher ratings would do well to think hard about off-night races. While a novelty, races on Saturdays and other days of the week do not draw as well as the traditional Sunday afternoons. Also, late season races should be evaluated closely, as the audience tapers off drastically once football season starts and the racing winds down. NASCAR can use these ratings to get a feel for what draws viewers in. Restrictor plate tracks are very popular, suggesting that fans thrive on close, fast pace racing with a high perceived risk. NASCAR can also be confident that their viewers are content with the variety of racetracks presented, because we see no great revolution one way or the other when the circuit visits short tracks and road courses. On the other hand, problems can also be found. One striking concern, as mentioned above, is that the Saturday night races are not terribly successful. Also, most glaring, NASCAR must consider a way to improve ratings as the season wears on. Fortunately for them, they have. Instead of a basic, season long points accumulation, the system has been changed. After 26 races, the top ten drivers in points will have their points arbitrarily readjusted, set almost equal. The goal is to make both the mid season races more exciting, as drivers battle to reach that top ten group, and late in the season as the race to the finish will be closer than ever. The results of this change will be fascinating in light of this analysis. This analysis only scratches the surface of the kinds of long studies that could be done on sports television ratings, attendance, and popularity as a whole. Further study could be done by adding more in-depth variables to the equation, such as popularity of drivers in the championship hunt, or somehow quantifying the legacy of certain tracks (such as Indianapolis, which has only been used by NASCAR for 10 years). In depth analysis could also be done on year-to-year changes in each race. The biggest challenge faced by researchers would continually be discovering the ratings information. Current data is not hard to find, but anything past the newest network tv contracts is seemingly impossible (the majority of races were broadcast on cable until very recently, and ratings information is not as easily found as it is for broadcast networks). Also, statistics on NASCAR events become less thorough going backwards past 2001, when NASCAR re-organized their web site and archives. The research here provides a unique window on how NASCAR races are watched and followed throughout the year. Using this data or these techniques, fans, advertisers, and even the sanctioning body can get a better feel for what draws people to broadcasts, and what keeps them away. Works Cited Caudill, Steven B., and Franklin G. Mixon Jr. "Television Revenue and the Structure of Athletic Contests: The Case of the National Basketball Association." Eastern Economic Journal 24.1 (1998): 43-50. Consoli, John. "Making Fewer Hits, But Still Scoring." MediaWeek 13.26 (2003): 4. Ewing, Jack, and Christina Passariello. "Is Lance just too Good?" Business week(2003): 44. Gold, Jacqueline S. "Taking center court." Crain's New York Business 19.32 (2003): 3. Hausman, Jerry A., and Gregory K. Leonard. "Superstars in the National Basketball Association: Economic Value and Policy." Journal of Labor Economics 15.4 (1997): 586-624. Databases Used www.sbrnet.com, Sports Business Research Network http://jayski.thatsracin.com/pages/tvratings.htm, “Jayski” NASCAR News, Rumor, and Statisitics www.NASCAR.com, Official site of the National Association for Stock Car Auto Racing