Selling the game: estimating the economic impact of professional sports through taxable sales.

Baade, Robert A. ; Baumann, Robert ; Matheson, Victor A. 等

Sports leagues, franchises, and civic boosters tout the economic benefits of professional sports as an incentive for host cities to construct new stadiums or arenas at considerable public expense. Past league-sponsored studies have estimated that new stadiums, franchises, and mega-events such as the Super Bowl increase economic activity by potentially hundreds of millions of dollars in host cities. A detailed regression analysis of taxable sales in Florida over the period extending from 1980 to 2005 fails to support these claims. New stadiums, arenas, and franchises, as well as mega-events, appear to be as likely to reduce taxable sales as increase them. Similarly, strikes and lockouts in professional sports have not systematically lead to reductions in local taxable sales.

JEL Classification: L83

1. Introduction

Sports boosters often claim that sports teams, facilities, and events inject large sums of money into the cities lucky enough to host them. Promoters envision hoards of wealthy sports fans descending on a city's hotels, restaurants, and businesses and showering them with fistfuls of dollars. For example, the National Football League (NFL) typically claims an economic impact from the Super Bowl of around $400 million (National Football League 1999), Major League Baseball (MLB) attaches a $75 million benefit to the All-Star Game (Selig, Harrington, and Healey 1999) and up to $250 million for the World Series (Ackman 2000), and the estimated effect of the National Collegiate Athletic Association (NCAA) Men's Basketball Final Four ranges from $30 million to $110 million (Mensheha 1998; Anderson 2001). Multiday events such as the Olympics or soccer's World Cup produce even larger figures. The preOlympics estimates for the 1996 Games in Atlanta indicated that the event would generate $5.1 billion in direct and indirect economic activity as well as 77,000 new jobs in Georgia (Humphreys and Plummet 1995). A study of soccer's 2002 World Cup (by the Dentsu Institute for Human Studies) estimated a $24.8 billion impact for Japan and an $8.9 billion impact for South Korea. As a percentage of national income, these figures represent 0.6% and 2.2% of the total Japanese and South Korean economies, respectively (Finer 2002). Initial economic impact studies of the 2010 Winter Olympics in Vancouver/Whistler predict a gain to the local economy of up to $10 C billion.

Even regular season games prompt claims of huge benefits. For example, the Oregon Baseball Campaign, a group dedicated to bringing MLB to Portland, reported that "a MLB team and ballpark would generate between $170 and $300 million annually in gross expenditures to the state of Oregon" (Oregon Baseball Campaign 2002), while a similar analysis completed for the Virginia Baseball Authority stated that a "a major league baseball franchise and stadium in northern Virginia would pump more than $8.6 billion into the economy over 30 years," or $287 million annually. The St. Louis (Missouri) Regional Chamber and Growth Association estimated that the Cardinals brought $301 million in annual economic benefits to the region, with another potential $40 to $48 million in benefits from a post-season appearance (Saint Louis Regional Chamber and Growth Association 2000). Of course, baseball is not the only sport to provide rosy economic impact numbers. A study of the NFL's New Orleans Saints estimated the impact of the team on the state at $402 million in 2002 (Ryan 2003), and the Seattle Supersonics of the NBA claimed that they pump $234 million into the area's economy annually (Feit 2006). Boosters are often vague about exactly what is being measured in these claims of hundreds of millions of dollars of benefits, making direct comparisons difficult, but the overall claims are clear: Professional sports provide huge economic windfalls for host cities.

Of course, leagues, team owners, and event organizers have a strong incentive to provide economic impact numbers that are as large as possible in order to justify heavy public subsidies. When leagues consider expansion or franchise relocations, they frequently highlight the potential economic benefits of a new franchise in order to minimize the team's or league's required contribution to the funding of the stadium or arena in which the team will play. Similarly, the NFL and MLB use the Super Bowl and baseball's All-Star Game as carrots to prompt otherwise-reluctant city officials and taxpayers to provide lavish funding for new stadiums to the great financial benefit of the existing owners. For example, in baseball, of the 15 new major league stadiums built between 1970 and 1997, 13 were selected by the MLB to host an All-Star Game within five years of their construction (Baade and Matheson 2001). Similarly, during a visit to the Dallas-Fort Worth, Texas, area just before a crucial vote on public funding for a new stadium, NFL Commissioner Paul Tagliabue suggested that the construction of a new stadium would lead to the opportunity for the metro area to host the Super Bowl in the next decade. Since the NFL touts economic benefits from hosting the Super Bowl of $350 to $400 million, an amount that exceeded the proposed $325 million public subsidy for the stadium, in effect, Commissioner Tagliabue was saying that combined with a Super Bowl, Arlington, Texas, would be getting a new stadium for free.

With an event like the Olympics, the huge costs of hosting the event to the standards now required by the International Olympic Committee, as well as those associated with providing adequate security, almost necessitate an infusion of taxpayer money. For example, while on paper the 2002 Winter Olympics in Salt Lake City, Utah, made a profit, the cost figures did not include millions of dollars of additional security provided by the U.S. Department of Defense at no cost to the local organizing committee. For the 2004 Summer Games, the government in Athens, Greece, spent $1.5 billion on security alone. These figures illustrate why organizers often rely on lofty reports that promise huge monetary windfalls to host cities. Since many economic impact studies are commissioned by owners, leagues, or event organizers, which stand to benefit directly from the public subsidies such reports are designed to elicit, one must question whether such studies can be believed.

2. Ex Ante versus Ex Post Studies

A typical ex ante economic impact study used by league and event promoters estimates the number of visitors an event or team is expected to draw, the number of days each spectator is expected to stay in the city, and the amount each visitor will spend each day. Combining these figures, an estimate of the "direct economic impact" is obtained. This direct impact is then subjected to a multiplier, usually around two, to account for the initial round of spending recirculating through the economy. This additional spending is known as "indirect economic impact." Thus, the total economic impact is roughly double the size of the initial spending. While such an estimation method is relatively straightforward, academic economists have been quick to point out the failings of such ex ante studies, as they often rely on poor methodology and also suffer from several theoretical problems.

First, many booster estimates are wildly optimistic about the number of potential guests and their spending habits. In March 2005, Denver, Colorado, tourism officials predicted 100,000 visitors for the NBA All-Star Game. Considering that the Pepsi Center, the game's venue, only holds 20,000 fans and taking into account that Denver has only about 6000 hotel rooms, it is not clear exactly how such an influx of basketball fans would be possible.

In many cases, the variation in estimated benefits alone is enough to question the validity of the studies. A series of studies of the NBA All-Star Game produced numbers ranging from a $3 million windfall for the 1992 game in Orlando, Florida, to a $35 million bonanza for the game three years earlier in Houston, Texas (Houck 2000). Similarly, the 1997 NCAA Women's Basketball Final Four was estimated to have an economic impact of $7 million on the local economy of Cincinnati, Ohio, but the same event was predicted to produce a $32 million impact on the San Jose, California, economy just two years later (Knight Ridder News Service 1999). The 10-fold disparity in the estimated impact for the same annual event illustrates the ad hoc nature of these studies. In some cases, economic impact figures appear to be completely fabricated. While city or league officials may suggest a certain monetary figure for a particular event, when pressed on the details, the "missing study" syndrome arises (Anderson 2004).

Even when ex ante studies are done in a carefully considered manner, they suffer from three primary theoretical deficiencies: the substitution effect, crowding out, and leakages. The substitution effect occurs when consumers spend money at a sporting event rather than on other goods and services in the local economy. A local resident who goes to a baseball game is spending money at the game that likely would have been spent at local restaurants, theaters, or retail establishments in the absence of the game. Therefore, the local consumer's spending on a sporting event is not new economic activity; rather, it represents a reshuffling of local spending. For this reason, most economists advocate that spending by local residents be excluded from any economic impact estimates.

Even including only out-of-region visitors in impact studies may still result in inflated estimates if a large portion of the non-local fans at a game are "casual visitors," that is, out-oftown guests who go to a sporting event but are visiting the host city for reasons other than the sporting event itself. For example, a college professor at an academic conference may buy a ticket to a local game, and therefore the ticket would be counted as a direct economic impact of the sports contest. The professor, however, would have come to the city and spent money on hotels and restaurants in the absence of the sporting match, and again, the money spent at the game substitutes for money that would have been spent elsewhere in the local economy.

Similarly, ex ante estimates may be biased upwards if event guests engage in "time-switching," which occurs when a traveler rearranges a planned visit to a city to coincide with a mega-event. One example of time-switching is someone who has always wanted to visit Hawaii who plans a trip during the NFL's Pro-Bowl. While the Pro-Bowl did influence the tourist's decision about when to come, it did not affect the decision whether to come. Therefore, total tourism spending in Hawaii is unchanged; the Pro-Bowl simply affects the timing of such spending.

Accounting for the substitution effect is likely to result in large reductions in the estimated economic impact of regular season games. In the case of mega-events, however, the substitution effect may be much smaller. Since these premier events are thought to attract large audiences from outside the local economy, many of whom come specifically for the event, the amount of spending that is new to the economy is thought to be quite a large proportion of the total amount of spending, whereas 5-20% of fans at a typical MLB game are visitors from outside the local metropolitan area, the percentage of visitors at an event like an All-Star Game or the Super Bowl is thought to be much higher (Siegfried and Zimbalist 2000).

A second source of bias is "crowding out," which results from the congestion caused by a game that dissuades local citizens from venturing near the playing venue during the game and thereby reduces economic activity. Attractions such as Chicago's Field Museum or Cleveland's Rock and Roll Hall of Fame are located next door to NFL stadiums, and their attendance suffers during the Chicago Bears or Cleveland Browns home games. Similarly, mega-events may dissuade regular recreational and business visitors from coming to a city during that time. While a city's hotels may be full of sports fans during the Super Bowl, if the city's hotels are generally full of vacationers or conventioneers anyway, the Super Bowl simply displaces other economic activity that would have occurred. High prices charged by hotels and other businesses in the hospitality industry also tend to dissuade casual visitors during mega-events. In other words, the economic impact of a mega-event may be large in a gross sense, but the net impact may be small. Scores of examples of this phenomenon exist. As a case in point, during the 2002 World Cup in South Korea, the number of European visitors to the country was higher than normal, but this increase was offset by a similar-sized decrease in the number of regular tourists and business travelers from Japan who avoided South Korea as a result of World Cup hassles. The total number of foreign visitors to South Korea during the World Cup in 2002 was estimated at 460,000, a figure identical to the number of foreign visitors during the same period in the previous year (Golovnina 2002).

A third source of bias comes from leakages. While money may be spent in local economies during sporting events, this spending may not wind up in the pockets of local residents. The taxes used to subsidize these events, however, are paid for by local taxpayers. The income multiplier for sporting events is likely to be much lower than for general expenditures as a result of the specialized nature of the service provided. In the NBA, for example, only 29% of players live in the metropolitan area in which their team plays (Siegfried and Zimbalist 2002).

Leakages during mega-events may also be quite high. The economic multipliers used in ex ante analyses are calculated using complex input-output tables for specific industries grounded in inter-industry relationships within regions based on an economic area's normal production patterns. During mega-events, however, the economy within a region may be anything but normal, and, therefore, these same inter-industry relationships may not hold. Since there is no reason to believe the usual economic multipliers apply during mega-events, any economic analyses based on these multipliers may, therefore, be highly inaccurate.

In fact, there is substantial reason to believe that during mega-events, these multipliers are highly overstated, which overestimates the true impact of these events on the local economy. Hotels, for example, routinely raise their prices during mega-events to three or four times their normal rates. The wages paid to a hotel's workers, however, remain unchanged, and, indeed, workers may be simply expected to work harder during times of high demand without any additional monetary compensation. As a hotel's revenue increases without a corresponding increase in costs, the return to capital (as a percentage of revenues) rises, while the return to labor falls. Capital income is far less likely than labor income to stay within the area in which it is earned, and, therefore, one might expect a fall in the multiplier effect during mega-events as a result of these increased leakages (Matheson 2004).

While ex ante estimates often do a credible job of determining the economic activity that occurs as a result of a sports team or mega-event, and although they may also address the issue of the substitution effect by excluding spending by local residents, they generally do a poor job of accounting for crowding-out, and they almost never acknowledge the problems associated with the application of incorrect multipliers. For these reasons, numerous studies have looked back at the actual performance of economies that have had professional franchises, built new playing facilities, and hosted mega-events and have compared the observed economic performance of host cities to that predicted in ex ante studies. These ex post analyses of stadiums and franchises, including those of Rosentraub (1994), Baade (1996), Coates and Humphreys (1999, 2003), and Siegfried and Zimbalist (2000), to name just a few, generally find little or no economic benefits from professional sports teams or new playing facilities.

In the area of mega-events, Baade and Matheson (2001) examine the MLB's All-Star Game and find that employment growth in host cities between 1973 and 1997 was 0.38% lower than expected, compared to other cities. A similar examination of the 1996 Summer Olympics in Atlanta, Georgia, found employment growth of between 3500 and 42,000 jobs, a fraction of the 77,000 new jobs claimed in ex ante studies (Baade and Matheson 2002). An examination of metropolitan area-wide personal income during 30 NCAA Men's Final Four basketball tournaments found that, on average, personal incomes were lower in host cities during tournament years (Baade and Matheson 2004a). A similar study of the 1994 World Cup in the United States found that personal income in host cities was $4 billion lower than predicted, a direct contradiction to ex ante estimates of a $4 billion windfall (Baade and Matheson 2004b). Coates and Humphreys (2002) examined the effect of post-season play in all four major U.S. sports on per capita personal incomes and found in all cases that hosting playoff games had a statistically insignificant impact on per capita incomes.

The remainder of this paper adds to the already-substantial body of work regarding ex post analyses of franchises, stadiums, and sporting events by using taxable sales data to estimate the effect of professional sports on local economies. In addition, this paper examines labor disputes, which serve as natural experiments for determining the economic impact of professional sports on host communities. If franchises do indeed provide large positive impacts on local economies, then their sudden absence as a result of work stoppages should result in observable negative effects on the city. Several previous studies examine the impact of sports on local metropolitan areas using strikes and lockouts as test cases. Zipp (1996) examines the effect of the 1994 MLB strike on 17 metropolitan statistical areas (MSAs), and he later extends his work to cover the effect of this strike on spring training venues in Florida (Zipp 1997). Baade and Matheson (2005) examine the 1981 and 1994/95 MLB baseball strikes, using personal income data, to arrive at an average net annual economic impact of a MLB team on a host city of between $16.2 million and $132.3 million, or between 5% and 50% of the figure generally suggested by baseball's boosters. Coates and Humphreys (2001) present the most comprehensive analysis of the economic consequences of sports strikes and lockouts. Their analysis of real per capita personal income finds no statistically significant effects from the strikes in MLB in 1972, 1981, and 1994/95 and strikes in the NFL in 1982 and 1987.

The existing studies of sports labor interruptions have two major weaknesses. First, all four previous studies examined the 1994/95 baseball strike, but only Coates and Humphreys (2001) examine a sport other than baseball. To the best of our knowledge, no other study has examined the effect of labor stoppages in the NBA and the National Hockey League (NHL).

Second, a major difficulty of measuring the economic impact of sports teams, events, and strikes is that even the impact of large businesses may be hard to isolate within the large, diverse metropolitan economies in which they reside. For example, even if a MLB franchise or a Super Bowl does result in a $300 million boost to the host city, this is less than 0.1% of the annual personal income of a large metropolitan area like Los Angeles. Any income gains as a result of a franchise or the "big game" would likely be obscured by normal fluctuations in the region's economy. This problem is further compounded if the labor interruption or the mega-event lasts for only a few months or even just a few days. Even if the effects of a labor dispute or megaevent are large in the time period immediately surrounding the event, this impact is likely to be obscured in annual data. The studies of Coates and Humphreys (2001) and Baade and Matheson (2005), which examine the ex post economic impact of labor interruptions in sports, both suffer from this limitation. For example, Coates and Humphreys (2001) faced a daunting task when they examined the impact of the 1972 MLB strike, which lasted 13 days, using annual data. Similarly, Coates and Humphreys' (2001) and Baade and Matheson's (2000, 2001, 2002, 2004a, b) studies of mega-events all examine events lasting at most one month, often as little as a single weekend, using annual data.

3. Use of Taxable Sales

Taxable sales are ideally suited to measuring the economic impact of stadiums, teams, and large sporting events for several reasons. First, there is a direct connection between sales tax collections and sporting events or facilities. Boosters often include large sums for visitor spending in their ex ante estimates of the economic impact of an event. In one of the few examples of a league-sponsored ex post study, the NFL reported that Super Bowl XXXIII in 1999 was responsible for a $670 million increase in taxable sales in South Florida, compared to the equivalent January-February period in 1998 (National Football League 1999). Numerous publicly funded sports facilities have also been financed specifically from sales tax collections or through specific increases in the sales tax rate, making an examination of taxable sales especially relevant. For example, of the 22 new stadiums constructed for NFL franchises between 1992 and 2005, six were funded, at least in part, through increases in the local general sales tax rate, while another eight were funded through increased excise taxes (i.e., sales taxes on specific goods and services, such as rental cars or hotel rooms) (Baade and Matheson 2006). In addition, consumer spending, much of which is captured by taxable sales, is the single largest component of gross domestic product and, therefore, is a good proxy for economic activity.

A second major reason that taxable sales are a useful tool in measuring the economic impact of professional sports is that, as noted previously, even significant economic events may be hard to isolate within the large, diverse metropolitan economies in which they take place. Any income gains as a result of the game or team would likely be obscured by normal fluctuations in the region's economy. If the event or franchise can be isolated within space and time, however, any potential impact is more likely to be identified. For example, while the presence of a World Series might have a large effect on neighborhood businesses, the overall effect on a state or country's economy will be minuscule and hard to identify. Furthermore, these same economic effects may be large for the time period immediately surrounding the event, but over the course of an entire year, the impact during a week-long period is not likely to show up as an important change.

Most previous studies of professional sports have used personal income (Baade and Matheson 2004a, b), per capita income (Coates and Humphreys 1999, 2002, 2003), or employment data (Baade 1996; Baade and Matheson 2000, 2001, 2002; Coates and Humphreys 2003) to estimate the ex post economic impact of sports. Generally, these data are available only annually and at the county or metropolitan area level, and, therefore, these studies suffer from the limitations mentioned previously. Taxable sales data, on the other hand, are often published either monthly or quarterly and can cover areas down to the city level or smaller. Therefore, these data can be analyzed to identify activities that are much smaller in scale and duration.

Several previous attempts to measure the effect of teams and mega-events through taxable sales data have been made. Baade and Matheson (2000) challenge the NFL's claim of a $670 million boost in South Florida's taxable sales from the Super Bowl and arrive at a figure of a mere $37 million increase. Their analysis is quite simplistic, however; their estimates account for only GDP growth, inflation, and population growth. Baade and Matheson (2001) examined taxable sales in California to determine the effect of MLB's All-Star Game on local economies. They found that the three California cities that hosted All-Star Games between 1985 and 1997 suffered an average drop in taxable sales of roughly $30 million in the quarter in which the game took place. Their study, however, is limited only to baseball's All-Star Game.

Porter (1999) provides detailed analysis of taxable sales with respect to mega-events, using regression analysis to determine that the economic impact of the Super Bowl was statistically insignificant, that is, not measurably different from zero. After reviewing short-term data on sales receipts for several Super Bowls, Porter concluded:

 Investigator bias, data measurement error, changing production
 relationships, diminishing returns to both scale and variable
 inputs, and capacity constraints anywhere along the chain of sales
 relations lead to lower multipliers. Crowding out and price
 increases by input suppliers in response to higher levels of demand
 and the tendency of suppliers to lower prices to stimulate sales
 when demand is weak lead to overestimates of net new sales due to
 the event. These characteristics alone would suggest that the
 estimated impact of the mega-sporting event will be lower than the
 impact analysis predicts. (p. 65)

Finally, the two most recent studies examining monthly sales tax collections in Texas present the most detailed analysis of tax data. Coates (2006) examines tax receipts in Houston during both the MLB All-Star Game and the Super Bowl, finding "little evidence of either event having a beneficial impact on any type of sales, total, retail, or services, or of the parts of those sales actually subject to the sales tax" (p. 240). Coates and Depken (2006) extend this study by examining 126 jurisdictions in Texas from 1990 through early 2006. They find that regular season games in the NBA, NFL, NHL, and MLB have divergent effects, with NHL and MLB games increasing tax revenues, while NBA and NFL regular season games decrease revenue. Collegiate regular season football games are revenue generators for small cities and towns home to Division I and Division IAA football. The Super Bowl had the largest effect of any event in the sample, with Houston garnering an additional $2 million in tax revenues during the big game.

This paper expands the scope of previous work in the area of taxable sales by including a much broader array of stadiums, franchises, and mega-events; a larger number of host cities; a longer time series; and a more detailed regression analysis in its examination than were included in previous studies.

4. The Data

We use taxable sales data, which are available monthly, to estimate the economic impact of professional sports on local economies. These data include just over 25 years' worth of monthly sales tax data (January 1980 through June 2005) for every county in Florida. Florida is an ideal candidate for this analysis since the state is home to at least two teams from each of the "Big Four" American professional sports--football, baseball, basketball, and hockey. The state has welcomed multiple new sports franchises to the area over the past 25 years, including the Orlando Magic and Miami Heat in the NBA, the Florida Marlins and Tampa Bay Devil Rays in the MLB, the Tampa Bay Lightning and the Florida Panthers in the NHL, and the Jacksonville Jaguars in the NFL. In addition, at least nine major new stadiums or arenas have been constructed in the state since 1980. Furthermore, each of the major labor interruptions since 1982 has impacted at least one franchise in Florida: the 1982 and 1987 NFL strikes, the NHL's 1994/95 and 2004/05 lockouts, the 1998/99 NBA lockout, and the 1994/95 MLB strike. Only two other labor interruptions in the Big Four have resulted in the loss of games, the 1972 and 1981 MLB strikes, which we do not analyze, since no city in Florida hosted a MLB franchise until the arrival of the Florida Marlins in 1993. Finally, Florida cities have hosted championship events for each of the "Big Four" American professional sports as well as soccer's World Cup and the NCAA Men's Final Four basketball tournament. In addition, Florida cities have also hosted all-star games in professional basketball, hockey, and soccer.

In order to maximize the chance that the economic effects of the events can be isolated (i.e., to minimize statistical "noise"), it is crucial to find data as specific as possible to the area in which the franchise is located or in which the mega-event occurred, and with the highest frequency possible. Florida provides monthly data on taxable sales for individual counties, and these data meet our criteria. In the analysis, taxable sales from several counties are added together, corresponding to the four specific Florida MSAs that will be examined: Miami-Fort Lauderdale-West Palm Beach, Tampa-St. Petersburg, Orlando, and Jacksonville. We use the monthly consumer price index compiled by the Bureau of Labor Statistics to convert taxable sales data from nominal to real.

MSA taxable sales are used in lieu of county taxable sales for two reasons. First, in several cases, stadiums are very near to county lines. For example, Dolphins Stadium, home of the Miami Dolphins and the Florida Marlins, resides in Miami-Dade County but is less than 1 mile from the Broward County border. In a case like this it seems unreasonable to exclude Broward County from the economic analysis because of the stadium's close proximity. Furthermore, as noted previously, it is important to differentiate between the gross and net impact of a franchise. If a lockout simply causes residents to spend money elsewhere in the local economy rather than at the ballpark, analysis of only a single county may not capture this substitution effect. Examining an entire MSA will account for the substitution effect if money is redirected between counties within a single MSA as a result of a strike, expansion, or a new stadium.

Since the current annual GDPs of large MSAs in Florida, such as Miami or Tampa, exceed $50 billion in nominal terms, even the effects of a potential major economic event, such as a new franchise or stadium, a prolonged strike, or a mega-event, can be obscured by the normal economic fluctuations of these economies. Many factors, including the local, regional, and national business cycle; state and federal government policies; monetary policy and inflation; international factors; consumer and business confidence; wealth effects; and a host of other ingredients, tend to influence taxable sales. In order to prevent these other factors from clouding the true effects of the labor interruption, it is essential to find a method to account for them.

5. The Model

In order to examine the impact of the individual sporting events on taxable sales in the relevant MSAs of Florida, we estimate the following reduced-form model for each of the four MSAs in our analysis:

[y.sup.*.sub.t] = [[beta].sub.0] + [[beta].sub.1][x.sub.t] + [[beta].sub.2][z.sub.t] + [3.summation over (p=1)] [[DELTA].sub.p][y.sup.*.sub.t-p] + [11.summation over (m=1)][[alpha].sub.m][S.sub.m] + [[epsilon].sub.t],

where [y.sup.*.sub.t] is the natural log of taxable sales; [x.sub.t] is a vector of macroeconomic control variables; [z.sub.t] is a vector of variables that capture the impact of strikes, lockouts, expansions, stadium construction, and mega-events; and [S.sub.m] is a vector of monthly dummies that account for seasonal variation in taxable sales. We include lagged values of the dependent variable to correct for autocorrelation in taxable sales and to purge out carryover effects of taxable sales from one month to the next. We use the Breusch-Godfrey Lagrange multiplier test to determine the appropriate number of lagged dependent variables for each MSA model. As Coates and Humphreys (2001) point out, lagged dependent variables bias their own coefficients, but not the coefficients of other variables in a linear model. Thus, we attach no interpretation of these coefficients.

The macroeconomic control variables in [x.sub.t] are the state population in Florida and the two-month change in the national unemployment rate. In addition, from July 1987 to April 1988, the State of Florida temporarily imposed the state sales tax on services, thereby increasing the number of items included within taxable sales. Population data come from the Census, and since these data are annual, we take the annual difference and divide it by 12 to construct monthly data. We employed a trial-and-error approach to find the best specification for each MSA, and we omitted the controls that did not have an impact distinguishable from zero on taxable sales and that failed to produce an increase in explanatory power, as measured by the adjusted [r.sup.2].

We identify two non-sports controls that affect the taxable sales ratio for the Miami MSA to improve the fit of the model. Hurricane Andrew, which devastated the South Florida economy in 1992, had a dramatic effect on taxable sales in the Miami MSA. Taxable sales initially fell in the area in the wake of the storm, then surged as residents rebuilt homes and replaced damaged property. This pattern is modeled using two intervention variables: an initial penalty during the month of the storm (August 1992) and a convex "ramp" (above pre-storm levels) that lasted for three months after the storm. See Baade, Baumann, and Matheson (2007) for details and a sensitivity analysis of this specification. Hurricane Andrew controls for the Tampa, Orlando, and Jacksonville models and did not produce significant results. In addition, controls for other costly hurricanes during our sample frame (e.g., the many hurricanes of August-October 2004) did not produce significant results for any of the MSAs in our analysis. We also include a dummy variable for the month of September 2001 to account for the terrorist attacks of 9/11, which had a negative and statistically significant impact on Jacksonville and Miami, but not on Tampa or Orlando.

The labor interruption variables in [z.sub.t] include the NBA lockout from November 1998 through January 1999, the MLB strikes in August and September 1994 and April 1995, the NFL strikes in October and November 1982 and October 1987, and the NHL lockouts from October 1994 through January 1995 and October 2004 through April 2005. When strikes carried over into the off-seasons of their sports, only the months during which regular season games were lost were counted as strike periods. Of course, not every labor interruption occurred at the beginning or end of a month. The results in Table 1 designate labor interruption months as those in which at least half of the games were lost. Alternative specifications were attempted and made little difference in terms of the results.

We also include controls for franchise expansions, which represent the entry of sports franchises in a manner similar to that by which labor interruptions represent a temporary exit. The franchise variables in [z.sub.t] include the expansion of the NBA into Miami (Miami Heat, November 1988) and Orlando (Orlando Magic, November 1989), the expansion of the NHL into Miami (Florida Panthers, October 1993) and Tampa (Tampa Lightning, October 1993), the expansion of the MLB into Miami (Florida Marlins, April 1993) and Tampa (Devil Rays, April 1998), and the expansion of the NFL into Jacksonville (Jaguars, September 1995). In addition, we also include controls for stadium and arena construction. The NFL stadiums constructed in Florida during our sample frame are the Raymond James Stadium in Tampa (September 1998) and the Dolphins Stadium (originally Joe Robbie Stadium) near Miami (September 1987). In the NHL, we include controls for the BankAtlantic Center in the Miami metropolitan area (October 1998), the St. Petersburg Times Arena in Tampa (October 1996), and the Tampa Thunderdome (April 1990), which was used by the Tampa Bay Lightning of the NHL until the 1996/97 season and was then subsequently renamed the Tropicana Field and was used for the expansion Tampa Bay Devil Rays of the MLB. In the NBA, we include a control for American Airlines Arena in Miami (January 2000). There are not separate controls for two other NBA arena constructions during our sample frame: the Miami Arena and the TD Waterhouse Centre (originally Orlando Arena). Both of these arenas were completed just months prior to the first NBA season of their respective expansion teams, which prevents us from separately estimating the effect of the franchise and the arena. Similarly, the renovation of Alltel Stadium in Jacksonville cannot be disentangled from the arrival of the new NFL team in the city.

Finally, 19 separate mega-event dummy variables are added to [z.sub.t], including the Super Bowls in Tampa in 1984, 1991, and 2001; those in Miami in 1989, 1995, and 1999; and those in Jacksonville in 2005; the NBA Finals in Orlando in 1995 and the NBA All-Star Game in Orlando in 1992 and in Miami in 1990; the NHL Stanley Cup in Miami in 1996 and in Tampa in 2004 and the NHL All-Star Game in Tampa in 1999 and in Fort Lauderdale (Miami MSA) in 2003; the MLB World Series in Miami in 1997 and 2003; the NCAA Men's Basketball Final Four in Tampa in 1999; and FIFA's World Cup in Orlando in 1994 and the Major League Soccer All-Star Game in 1998.

As the Super Bowl generally occurs in either the last weekend of January or the first weekend of February, the dummy variables for all Super Bowl years include both January and February. This captures spending in preparation for the event, economic activity during Super Bowl week, and spending occurring several weeks after the big game, which should capture some portion of the multiplier effect as local businesses and residents spend part of their Super Bowl windfall. Similarly, dummy variables for both the NBA and NHL finals cover both May and June, since the playoffs and finals can cover portions of both months. All other sports variables cover only the specific month in which the game(s) is played.

The results of ordinary least squares analysis on the model for the four cities are shown in Tables 1-4. The tables show the coefficients, standard errors, and t-statistics for each variable in the model. The seasonal dummy variables are omitted from the tables for brevity. It should be noted that several of these coefficients imply a reasonably large economic impact from professional sports (although they are much smaller than the average residuals reported in studies using annual data, such as those of Baade and Matheson 2004a, b). Again, given the size of these economies, the effect of even a large event with hundreds of millions of dollars of potential impact is likely to be obscured by natural, unexplained variations in the economy that even a sophisticated econometric model may fail to detect. It is, therefore, unlikely that the models for an individual city will accurately capture the effects of any one specific event, such as a new stadium or a Super Bowl.

If professional sports do have a positive impact on a region's economy, however, across a large number of cities and events one should expect a recognizable pattern of increasing taxable sales during periods with new stadiums, franchises, or large events and a pattern of decreasing taxable sales during labor interruptions. In fact, the coefficients of the sports variables appear to be almost entirely random. Despite boosters' claims that professional sports result in huge economic windfalls for host cities, over half of the sports variables examined in this study resulted in reductions in taxable sales in host cities. Of the 43 sports variables included in the four models, only 21 (49%) have the "correct" sign indicating that the presence of teams and events contributes positively to taxable sales in a metropolitan area. Just six of the 11 labor interruptions have negative coefficients, only four of the 13 new stadium/franchise coefficients are positive, and only 11 of the 19 mega-event coefficients are positive. This is hardly the economic stimulus that local taxpayers are routinely promised when teams and leagues claim, "If you build it, they will come."

Five of the sports variable coefficients are statistically significant at the 10% level, but again, roughly half of the signs are in the "wrong" direction. While the NBA expansion in Miami, the 2003 World Series in Miami, and the 2001 Super Bowl in Tampa display a statistically significant positive impact on their local economies, the 2003 NHL All-Star Game and the opening of Joe Robbie Stadium in 1987, both in Miami, produce statistically significant negative results. Of course, with 43 sports variables in the model, one would normally expect four or five coefficients to be statistically significant at the 10% level even if no true correlation between sports and taxable sales existed, so even these statistically significant results must be taken with a grain of salt.

In the most extreme cases of exaggerated economic benefits, not only do the results in this paper fail to support the boosters' claims, but these results can actually show that the boosters' claims are demonstrably wrong. For example, in 1999 the NFL reported, "Thanks to Super Bowl XXXIII, there was a $670 million increase in taxable sales in South Florida compared to the equivalent January-February period in 1998" (National Football League 1999, p. 1).

Forgetting for the moment the questionable statistical practice of drawing a conclusion based on a comparison of two years' worth of data, the data do indeed show that the Florida Department of Revenue reported that taxable sales increased by $640 million in the three-county region, including Broward, Dade, and Palm Beach Counties, in January-February 1999, compared to the same period in 1998. (The $30 million discrepancy between the official figures and the numbers reported by the NFL is of little significance, except to indicate possible sloppiness on the part of the League.)

The important issue here, however, is that taxable sales in the region could be expected to grow for many reasons other than the presence of a mega-event, reasons such as inflation, population growth, and increases in real income associated with economic conditions besides the presence of the Super Bowl. The model presented in this paper attempts to control for these factors (among others) and derives an increase in taxable sales from the 1999 Super Bowl in Miami of only $99.6 million, with a 95% confidence interval of $--434.8 to $634 million. Thus, a $670 million increase in taxable sales due to the Super Bowl can be rejected solely on the basis of this single observation. In concert with the host of other coefficients in this paper indicating little or no positive economic impact from professional sports, the typical boosters' claims can be even more soundly rejected.

One final note: Taxable sales in the area in January-February 2000, the year after the game, were $1.26 billion higher than in the same months during the preceding year, yet the NFL never publicized a story proclaiming, "Thanks to the lack of a Super Bowl, there was a $1.26 billion increase in taxable sales in South Florida compared to the equivalent January February period in 1999."

6. Conclusions

Professional sports leagues, franchises, and civic boosters have used the promise of sports franchises, new stadiums and arenas, and all-star games or league championships as an incentive for host cities to construct new stadiums or arenas at considerable public expense. In the past, league- and industry-sponsored studies have estimated that mega-events such as the Super Bowl and all-star games increase economic activity by hundreds of millions of dollars in host cities. Similar studies claim that new stadiums or franchises also can have hundreds of millions of dollars of annual local economic impact. Our detailed regression analysis of taxable sales in Florida over the period from 1980 to mid-2005 fails to support these claims. New stadiums, arenas, and franchises, as well as mega-events, appear to be as likely to reduce taxable sales as to increase them. Similarly, strikes and lockouts in professional sports have not systematically reduced local taxable sales. While these results, like any econometric estimates, are subject to some degree of uncertainty, they clearly place doubt on boosters' claims of huge economic windfalls. Cities would be wise to view with caution economic impact estimates provided by sports boosters, who have a clear incentive to inflate these estimates. It would appear that "padding" is an essential element of many games both on and off the field.

Received April 2006; accepted January 2007.

References

Ackman, Dan. 2000. "In Money Terms, the Subway Series Strikes Out." Forbes Magazine, Oct 21, 2000. Available http://www.forbes.com/2000/10/21/1021series/html. Accessed 15 August 2007.

Anderson, Patrick L. 2004. Business economics and finance with MATLAB, GIS and simulation models. New York: Chapman & Hall/CRC, pp. 131.

Anderson, Tom. 2001. "St. Louis Ready to Raise NCAA Flag if Atlanta Can't." St. Louis Business Journal, January 19, 2001. Available http://stlouis.bizjournals.com/stlouis/stories/2001/0l/22/story8.html. Accessed 15 August 2007.

Baade, Robert, and Victor Matheson. 2001. Home run or wild pitch? Assessing the economic impact of major league baseball's all-star game. Journal of Sports Economics 2:307-27.

Baade, Robert A. 1996. Professional sports as a catalyst for metropolitan economic development. Journal of Urban Affairs 18:1-17.

Baade, Robert A., Robert Baumann, and Victor A. Matheson. 2007. Estimating the economic impact of natural and social disasters with an application to Hurricane Katrina. Urban Studies 44.

Baade, Robert A., and Victor A. Matheson. 2000. An assessment of the economic impact of the American football championship, the Super Bowl, on host communities. Reflets et Perspectives 34:35-46.

Baade, Robert A., and Victor A. Matheson. 2002. Bidding for the Olympics: Fool's gold? In Transatlantic sport: The comparative economics of North American and European sports, edited by C. P. Barros, M. Ibrahimo, and S. Szymanski. London: Edward Elgar Publishing, pp. 127-51.

Baade, Robert A., and Victor A. Matheson. 2004a. The quest for the cup: Assessing the economic impact of the World Cup. Regional Studies 38:341-52.

Baade, Robert A., and Victor A. Matheson. 2004b. An economic slam dunk or march madness? Assessing the economic impact of the NCAA basketball tournament. In Economics of college sports, edited by J. Fizel and R. Fort. Westport, CT: Praeger Publishers, pp. 111-33.

Baade, Robert A., and Victor A. Matheson. 2005. Striking out? The economic impact of major league baseball work stoppages on host communities. College of the Holy Cross, Department of Economics Working Paper No. 05-07.

Baade, Robert A., and Victor A. Matheson. 2006. Have public finance principles been shut out in financing new stadiums for the NFL? Public Finance and Management 6:284-320.

Coates, Dennis. 2006. The tax benefits of hosting the Super Bowl and the MLB all-star game: The Houston experience. International Journal of Sports Finance 1:239-52.

Coates, Dennis, and Craig A. Depken, II. 2006. Mega-events: Is the Texas-Baylor game to Waco what the Super Bowl is to Houston? International Association of Sports Economists Working Paper No. 06-06.

Coates, Dennis, and Brad R. Humphreys. 1999. The growth effects of sports franchises, stadia and arenas. Journal of Policy Analysis and Management 14:601-24.

Coates, Dennis, and Brad R. Humphreys. 2001. The economic consequences of professional sports strikes and lockouts. Southern Economic Journal 67:737-47.

Coates, Dennis, and Brad R. Humphreys. 2002. The economic impact of post-season play in professional sports. Journal of Sports Economics 3:291-9.

Coates, Dennis, and Brad R. Humphreys. 2003. The effect of professional sports on earnings and employment in the services and retail sectors in US cities. Regional Science and Urban Economics 33:175-98.

Feit, Josh. 2006. Key amendments: Council member Nick Licata challenges Sonics subsidy. The Stranger, Seattle's Only News Section. 23 February-1 March 2006.

Finer, Jonathan. 2002. The grand illusion. Far Eastern Economic Review 7 March: 32-6.

Golovnina, Maria. 2002. "S. Korean Tourism Sector in Blues Despite World Cup." Forbes.com. Available http://www.forbes.com/newswire/2002/06/19/rtr636036.html. Posted 19 June 2002. Accessed 20 June 2002.

Houck, J. 2000. "High-Stake Courtship." FoxSportsBiz. Available http://www.foxsports.com/business/trends/ z000120allstar1.sml. Accessed 21 January 2000.

Humphreys, Jeffrey M., and Michael K. Plummer. 1995. The economic impact on the state of Georgia of hosting the 1996 Summer Olympic games. Athens, GA: Selig Center for Economic Growth, The University of Georgia Working Paper.

Knight Ridder News Service. 1999. "Final Four's Financial Impact Hard to Gauge." Enquirer Sports Coverage, March 25, 1999. Available http://enquirer.com/editions/1999/02/25/spt_final_fours.html. Accessed 30 August 2001.

Matheson, Victor A. 2004. Economic multipliers and mega-event analysis. College of the Holy Cross Working Paper Series No. 04-02.

Mensheha, M. 1998. "Home-Court Edge: Final Four Promises to be Economic Slam Dunk." San Antonio Business Journal, March 27, 1998. Available http://sanantonio.bizjournals.com/sanantonio/stories/1998/03/30/story3.html. Accessed 15 August 2007.

National Football League. 1999. Super Bowl XXXIII generates $396 million for South Florida. NFL Report 58:7.

Oregon Baseball Campaign. 2002. "Why MLB, Why Oregon, Why Now?" Available http://www.oregonbaseballcompaign.com/fact.htm. Accessed 5 September 2002.

Porter, Philip. 1999. Mega-sports events as municipal investments: A critique of impact analysis. In Sports economics: Current research, edited by J. Fizel, E. Gustafson, and L. Hadley. Westport, CT: Praeger Press, pp. 61-74.

Rosentraub, Mark. 1994. Sport and downtown development strategy. Journal of Urban Affairs 16:228-39.

Ryan, Timothy P. 2003. The economic impact of the New Orleans Saints, 2002. University of New Orleans Working Paper.

Saint Louis Regional Chamber and Growth Association. 2000. "RCGA Estimates Economic Impact of Cardinals Playoff Run." Available http://www.stlrcga.org/00_1002.html. Accessed 5 September 2002.

Selig, Bud, John Harrington, and Jim Healey. "New Ballpark Press Briefing: July 12, 1999." Available http://www.asapsports.com/baseball/1999allstar/071299BS.html. Accessed 29 August 2000.

Siegfried, John, and Andrew Zimbalist. 2000. The economics of sports facilities and their communities. Journal of Economic Perspectives 14:95-114.

Siegfried, John, and Andrew Zimbalist. 2002. A note on the local economic impact of sports expenditures. Journal of Sports Economics 3:361-6.

Zipp, John F. 1996. The economic impact of the baseball strike of 1994. Urban Affairs Review 32:157-85.

Zipp, John F. 1997. Spring training. In Sports, jobs, and taxes: The economic impact of sports teams and stadiums, edited by Roger G. Noll and Andrew Zimbalist. Washington, DC: The Brookings Institution Press, pp. 427-51.

Robert A. Baade, * Robert Baumann, ([dagger]) and Victor A. Matheson ([double dagger])

* Department of Economics and Business, Lake Forest College, Lake Forest, IL 60045, USA; E-mail baade@lfc.edu.

([dagger]) Department of Economics, Box 192A, College of the Holy Cross, Worcester, MA 01610-2395, USA; E-mail rbaumann@holycross.edu.

([double dagger]) Department of Economics, Box 157A, College of the Holy Cross, Worcester, MA 01610-2395, USA; E-mail vmatheso@holycross.edu; corresponding author.

Table 1. Sample 1980:1--2005:6: Miami. Dependent Variable:
[y.sup.*.sub.t] = ln (Taxable Sales)

Variable Coefficient SE t-Statistic

Constant 4.32050 ** 0.87343 4.95
Population 5.48E-08 ** 1.32E-08 4.16
Two-month change in
 unemployment -0.01137 ** 0.00297 -3.83
9/11 impact -0.07116 * 0.0326 -2.18
Services taxed 0.05992 ** 0.01595 3.76
Hurricane Andrew--initial
 penalty -0.07228 * 0.03262 -2.22
Hurricane Andrew--convex
 ramp 0.12611 ** 0.02349 5.37
NFL Strike, 1982 0.03208 0.02392 -1.34
NFL Strike, 1987 -0.01860 0.03501 0.53
NBA Strike, 1998/99 0.01321 0.02145 -0.62
NHL Strike, 1994/95 -0.01037 0.01747 0.59
NHL Strike, 2004/05 -0.01965 0.01376 1.43
MLB Strike, 1994/95 0.01139 0.01898 -0.6
Joe Robbie Stadium, NFL,
 9/1987 -0.02568 * 0.01467 -1.75
Bank Atlantic Center, NHL,
 10/1998 0.00249 0.01124 -0.22
American Airlines Arena,
 NBA, 1/2000 0.00825 0.01106 -0.75
MLB Expansion, 4/1993 -0.00871 0.01414 0.62
NHL Expansion, 10/1993 0.01736 0.01528 -1.14
NBA Expansion, 11/1988 0.03186 * 0.01419 2.25
Super Bowl, 1989 -0.03859 0.02388 1.62
Super Bowl, 1995 0.00320 0.02428 -0.13
Super Bowl, 1999 -0.00930 0.02545 0.37
NBA All-Star Game, 1990 0.00268 0.03236 -0.08
World Series, 1997 -0.01261 0.0324 0.39
World Series, 2003 0.10201 ** 0.03285 3.11
Stanley Cup Finals, 1997 0.01296 0.02291 -0.57
NHL All-Star Game, 2003 -0.07173 * 0.03257 -2.2
ln (taxable [sales.sub.t-1]) 0.33638 ** 0.06055 5.56
ln (taxable [sales.sub.t-2]) 0.23928 ** 0.06204 3.86
ln (taxable [sales.sub.t-3]) 0.15032 * 0.06289 2.39
ln (taxable [sales.sub.t-4]) -0.0615 0.05418 1.14
Adjusted [r.sup.2] -0.9977

Breusch-Godfrey Lagrange Multiplier Tests

Lags Chi-Squared Statistic p-Value

1 0.55 0.458
2 0.688 0.709
3 1.355 0.716
4 4.313 0.365
5 5.841 0.322

All dollar impact values are in 1982-1984 dollars using the CPI. The
coefficients are reported with their associated t-statistic for the
null hypothesis that the estimated value is equal to zero.
SE = standard error.

* Significant at the 10% level.

** Significant at the 1% level.

Table 2. Sample 1980:1-2005:6: Tampa. Dependent Variable:
[y.sup.*.sub.t] = ln (Taxable Sales)

Variable Coefficient SE t-Statistic

Constant 2.06014 ** 0.45696 4.51
Population 2.93E-08 ** 9.87E-09 2.97
Two-month change in
 unemployment 0.03951 ** 0.05393 5.08
9/11 impact 0.01107 0.03522 0.31
Services taxed 0.02206 0.01417 1.56
NFL Strike, 1982 -0.02843 0.0253 -1.12
NFL Strike, 1987 0.0277 0.03738 0.74
NHL Strike, 1994/95 -0.00783 0.01836 -0.43
NHL Strike, 2004/05 0.01078 0.0148 0.73
Raymond James Stadium,
 NFL, 9/1998 -0.01724 0.01659 -1.04
Tropicana Field, MLB, 3/1990 -0.00407 0.00994 -0.41
St. Pete Times Forum, NHL,
 10/1996 0.00813 0.01032 0.79
MLB Expansion, 4/1998 0.00413 0.01752 0.24
NHL Expansion, 10/1993 -0.00203 0.00898 -0.23
Super Bowl, 1984 -0.02158 0.02506 -0.86
Super Bowl, 1991 0.01166 0.02538 0.46
Super Bowl, 2001 0.07534 ** 0.0251 3
Super Bowl, 2004 0.02344 0.02576 0.91
NHL All-Star Game, 1999 0.03361 0.03525 0.95
NCAA Men's Final Four, 1999 0.02504 0.03517 0.71
ln (taxable [sales.sub.t-1]) 0.39080 ** 0.05754 6.79
ln (taxable [sales.sub.t-2]) 0.23144 ** 0.05883 6.79
ln (taxable [sales.sub.t-3]) 0.27388 ** 0.05393 5.08
Adjusted [r.sup.2] 0.9974

Breusch-Godfrey Lagrange Multiplier Tests

Lags Chi-Squared Statistic p-Value

1 0.38 0.538
2 0.891 0.641
3 3.562 0.313
4 3.734 0.443
5 3.863 0.569

All dollar impact values are in 1982-1984 dollars using the CPI.
The coefficients are reported with their associated t-statistic
for the null hypothesis that the estimated value is equal to zero.
SE = standard error.

* Significant at the 10% level.

** Significant at the 1% level.

Table 3. Sample 1980:1-2005:6: Orlando. Dependent Variable:
[y.sup.*.sub.t] = ln (Taxable Sales)

Variable Coefficient SE t-Statistic

Constant 1.57543 ** 0.50322 3.13
Population 3.20E-08 ** 1.14E-08 2.8
Two-month change in
 unemployment 0.02240 * 0.01277 1.75
9/11 impact 0.04147 0.05203 0.8
Services taxed 0.03387 * 0.0195 1.74
NBA Strike, 1998/99 -0.00961 0.02955 -0.33
NBA Expansion, 11/1989 -0.00239 0.01212 -0.2
NBA All-Star Game, 1992 -0.04135 0.05122 -0.81
World Cup, 1994 -0.00967 0.05127 -0.19
NBA Finals, 1995 0.00008 0.03693 0
MLS All-Star Game, 1998 -0.025 0.05114 -0.49
ln (taxable [sales.sub.t-1]) 0.27905 ** 0.06148 4.54
ln (taxable [sales.sub.t-2]) 0.21591 ** 0.06211 3.48
ln (taxable [sales.sub.t-3]) 0.17002 ** 0.06315 2.69
ln (taxable [sales.sub.t-4]) 0.17089 ** 0.06156 2.78
ln (taxable [sales.sub.t-5]) 0.07609 ** 0.05992 1.27
Adjusted [r.sup.2] 0.9974

Breusch-Godfrey Lagrange Multiplier Tests

Lags Chi-Squared Statistic p-Value

1 0.675 0.411
2 1.014 0.602
3 1.591 0.661
4 6.96 0.138
5 9.788 0.082

All dollar impact values are in 1982-1984 dollars using the CPI. The
coefficients are reported with their associated t-statistic for the
null hypothesis that the estimated value is equal to zero.
SE = standard error.

* Significant at the 10% level.

** Significant at the 1% level.

Table 4. Sample 1980:1-2005:6: Jacksonville. Dependent Variable:
[y.sup.*.sub.t] = In (Taxable Sales)

Variable Coefficient SE t-Statistic

Constant 3.1061 ** 0.62356 4.98
Population 5.06E-08 ** 1.15E-08 4.42
Two-month change in
 unemployment 0.02576 ** 0.01084 2.38
9/11 initial impact -0.47586 ** 0.04408 -10.8
9/11 permanent impact -0.06291 ** 0.01632 -3.86
Services taxed 0.04224 ** 0.01687 2.5
NFL expansion, 9/1995 -0.04423 0.04333 -1.02
Super Bowl, 2005 0.03358 0.03121 1.08
ln (taxable [sales.sub.t-1]) 0.38395 ** 0.05355 7.17
ln (taxable [sales.sub.t-2]) 0.15316 ** 0.05618 2.73
ln (taxable [sales.sub.t-3]) 0.16287 ** 0.05407 3.01
ln (taxable [sales.sub.t-4]) 0.06869 0.05464 1.26
ln (taxable [sales.sub.t-5]) 0.06063 0.04864 1.25
Adjusted [r.sup.2] 0.9955

Breusch-Godfrey Lagrange Multiplier Tests

Lags Chi-Squared Statistic p-Value

1 0.233 0.629
2 4.066 0.131
3 4.412 0.22
4 5.135 0.274
5 5.246 0.387

All dollar impact values are in 1982-1984 dollars using the CPI.
The coefficients are reported with their associated t-statistic
for the null hypothesis that the estimated value is equal to zero.
SE = standard error.

* Significant at the 10% level.

** Significant at the 1% level.