The National Football League has experienced both expansion and relocation of its franchises in the past decade. It is a dynamic market; the relocation of a NFL franchise is an annual possibility. This study looked at the demographic and economic factors that determine the current locations of NFL teams. The top 50 metropolitan areas were empirically examined to explain why some cities have an NFL team and others do not. These factors included population, per capita income, the number of other sports franchises, and the number of Fortune 500 companies, geographic factors, and television ratings for “Monday Night Football.” This model can identify cities for possible expansion and those that would serve as relocation sites in the future. Special attention was paid to the Los Angeles and New Orleans markets.


The National Football League has experienced a dynamic period of expansion and relocation in the past decade; the league seeks to position itself with the optimal configuration for long-term growth of the professional football market. Although expansion is not a current short-term goal for the NFL, the relocation of weak teams remains an annual possibility.

Moving an existing sports franchise is not new. After the 1995 season, Los Angeles lost both of its football teams. The Rams moved from the old Rose Bowl in Pasadena to the brand-new TWA Dome in downtown St. Louis. The Raiders moved from the Los Angeles Coliseum to the newly renovated Oakland Coliseum. In 1996, the Cleveland Browns moved into a new stadium in Baltimore and became the Ravens. Most recently, the Houston Oilers moved to Tennessee and became the Titans in 1999.

The city of Los Angeles, which lost its chance to gain an expansion team in 2002 to Houston because it was unable to approve financing for a new stadium, remains without a team. Although this leaves the second largest television market without its own team, it also offers a credible relocation threat for existing team owners in their new stadium negotiations with local authorities.1

Expansion and relocation of franchises in professional sports leagues have been studied for baseball and basketball; however, as far as the researchers know, a location model had never been created for the National Football League. Bruggink and Zamparelli (1999) produced a location model for Major League Baseball. The top 50 metropolitan areas were chosen for the sample. The explanatory variables for the regression model are population, population growth, and per capita income, in addition to the number of other professional sports teams in the area, the number of headquarters for Fortune 500 companies in the area, and the distance to the closest city with a baseball team. As one would expect, all variables had positive coefficients in the regression model. A more recent sports model by Rascher and Rascher (2005) estimated the probability that a particular city will have a National Basketball Association team by using the same core set of variables and adding factors such as the average NBA Nielsen television ratings for each city.

One interesting application of our model was applied to the New Orleans Saints football team. Even before the Hurricane Katrina damage to the Superdome, the owners of the Saints hinted that a move to a new location was in the offing.2 Of course, this is the typical ploy to gain public subsidies for a new or improved stadium, but the closure of the Superdome for the 2005 season made this relocation potential very real. The re-scheduled 2005 games found the Saints playing in the welcoming city of San Antonio with capacity attendance at the Alamodome. Although the Saints played the 2006 season in a repaired Superdome, they are in position to pursue one of two options after 2006: 1) stay in New Orleans with a long term city commitment to help build a new or improved stadium, or 2) move to a new location in Los Angeles (Maske and Shapiro, 2005) or San Antonio (Orsborn 2006). Our location model used estimates of New Orleans depopulation to give a perspective on the potential consequences for the city’s viability as a home to a NFL franchise.

Location Model

There are a number of factors that influence the selection of cities for NFL relocation or expansion. Based on demographic and economic factors in the largest 50 metropolitan areas, the researchers constructed a logit model to determine the relationship of these factors (X1, X2, X3, X4, X5, X6, X7) to the expected conditional probability (Pi) that each city (represented as i) would have one or more NFL teams:

Pi = E (Yi = 1| X1i,X2i, X3i, X4i, X5i, X6i, X7i)


Y = 1 if metropolitan area has one or more NFL teams; 0 otherwise

X1 = POP2000 = population in metropolitan area i (millions)

X2 = POPGROWTH = % population growth of a city the decade before the current team
located in city i (if the city does not have a team, the current
decade of growth is used)

X3 = INC = income per capita in metropolitan area i ($1000)

X4 = DISTANCE = distance to closest football franchise city (miles) from city i

X5 = F500 = the number of Fortune 500 company headquarters in metropolitan area i

X6 = OTHER/POP2000 = the number of other professional sports teams per million
population in city i (men’s basketball, hockey, and baseball)

X7 = NIELSEN = “Monday Night Football” Nielsen ratings in metropolitan area i using
the 2002-3 season

The NFL wants to expand or relocate to a metropolitan area with a large and growing population in order to maximize stadium revenues from attendance, concessions, and parking fees. The researchers expected positive coefficients for the population and population growth. In addition to market size, standard microeconomic demand theory suggests that per capita income will be a positive influence. According to a 2003 survey, football fans had the highest median salary in all sports at $55,115 (USA Today, 2003).

The distance to the closest NFL franchise city is a major aspect for the location of a franchise. The NFL does not want all of its teams in one part of the country because there will be interest only in that part of the country, not the whole United States. This is especially important for a sport driven by national television revenues. Furthermore, locating a new franchise close to an existing franchise (generally up to 75 miles) would financially hurt the owner of the latter. The league does not approve of these territorial infringements.

The number of other professional sports teams (baseball, basketball, and hockey) in the metropolitan area can have an impact on the demand for football games. The other sports could be considered substitute goods, or, on the other hand, a measure of fan interest for sports in general (a complement good). Rascher and Rascher (2005) found a negative and significant relationship between the number of other teams and the probability that a city had a professional basketball team. However, Bruggink and Zamparelli (1999) found just the opposite for baseball, supporting the fan intensity argument. The difference could be in overlapping the season with other sports. Basketball overlaps the most, competing with hockey, football, and the beginning of baseball. Baseball and football face less severe overlaps, and at times have the season to themselves.

Another location factor is the number of headquarters for Fortune 500 corporations in the metro area (F500). NFL owners are allowed to keep all the stadium revenue from luxury box receipts (i.e., no revenue sharing), and corporations are the largest patrons of these seating sections (which are also called corporate sky boxes). This is one of the reasons that owners desire new stadiums, because it affords them an opportunity to maximize the number of luxury box seats.

The last location factor is the number of households in the metro area watching football on television. In this study, the average Nielsen ratings3 for the 2002-3 “Monday Night Football” games are used for the 50 metropolitan areas. “Monday Night Football” games were selected because the same game is watched by the entire nation, whether a city has a NFL team or not. Nielsen ratings are particularly important for the National Football League because there are no local television contracts, only a national contract divided equally among the 32 teams, and it is the single largest source of income.

Empirical Results and Simulation:

The preliminary estimation of the model included the top 50 metropolitan areas (those with populations of approximately one million or more). About half had one or more teams. However, with Los Angeles in the data set the estimated coefficient of the population variable is negative instead of positive. This preliminary model showed how sensitive the population factor is to the inclusion of Los Angeles (by far, the largest metro area in the sample that has no team). As discussed earlier, league owners receive value from having a viable city without a team because it poses a credible threat for a team to relocate, allowing them to negotiate with local government for sports subsidies. In this sense, the no-team status of Los Angeles is not a market outcome but a strategic ploy. When Los Angeles was excluded, the researchers actually had a sample of cities that was more representative of market conditions. The estimated model without Los Angeles had a positive coefficient for population and a better overall fit. This sensitivity is the reason the researchers relied on this as the final model for the predictions.

In this study, the researchers: 1) examined the statistical results, 2) tested the within-sample predictions for each city, 3) determined a list of viable cities for the NFL expansion or relocation, and 4) ran a simulation on the effect of New Orleans’ recent depopulation on its viability of retaining its football team.

Table 1 shows all the statistical results for the fitted model using the logistic function shown below. The logistic function is the natural log of the odds ratio in favor of a metropolitan area having a team (P = 1 if the metro area has a team, P = 0 if it does not). There are three advantages to using the logistic curve rather than an ordinary least squares regression: (1) the predicted probabilities for a city having a team are constrained to lie between 0 and 1 for the logistic curve, whereas for a linear regression model the predicted probabilities could exceed 1 or fall short of zero, both of which are impossible values, (2) the slope coefficients in the logit model are more realistic than ordinary least square because they vary in magnitude, depending on the size of the corresponding explanatory variable, and (3) the variance is more constant in the logit model than with ordinary least squares, which makes the t-tests more valid.

log (P / (1-P)) = -17.6 + 0.98 POP2000 – 0.0063 POPGROWTH + 0.21 INC + 0.114 DISTANCE +
0.4573 F500 + 0.33 NIELSEN + 3.656 OTHER/POP2000

Table 1: Location Model

Logit Model
Included observations: 49
Convergence achieved after 8 iterations
Variable Coefficient Std. Error t-Stat Prob.
C -17.5895 8.8715 -1.983 0.0474
POP2000 0.9754 1.1482 0.849 0.3956
POPGROWTH -0.0063 0.0363 -0.174 0.8619
INC 0.2096 0.2096 1.000 0.3173
DISTANCE 0.0114 0.0087 1.315 0.1884
F500 0.4573 0.2747 1.665 0.096
NIELSEN 0.3347 0.2401 1.394 0.1634
OTHER/POP2000 3.6594 2.3024 1.589 0.112
Mean dependent 0.59184 S.D. dependent 0.4966
S.E. of regression 0.27766 McFadden R-s 0.6609

All the coefficients but one have the correct sign. Four of the seven are statistically significant at a 10% level or better in a one-tailed test. The statistically significant coefficients are for the following variables: the distance from the nearest NFL city, the number of other sports teams in the city per million population, the “Monday Night Football” television ratings, and the number of Fortune 500 headquarters in the city. The income variable missed being significant at only a 10% level. The only high correlation among the independent variables is between F500 and POP2000. This may explain why the coefficient of POP2000 does not appear statistically significant.

Table 2 shows the standardized coefficients in the logit model. Standardized coefficients scale the coefficients in the model using the standard deviations of the each independent variable and the dependent variable. By this method, the Fortune 500 and population variables have the most effect on whether a city has a team or not. Each have more the twice the size and therefore twice the impact than the other standardized coefficients. Distance and the presence of other professional teams rank next in importance followed by Nielsen ratings and income.

Table 2: Standardized Coefficients

Variable Standardized Coefficient
F500 11.33
POP2000 7.35
OTHER/POP2000 2.56
INC 1.89

Table 3 shows the results of the in-sample forecasts. The logit model correctly predicted the current NFL franchise status in 45 out of the 50 metropolitan areas. The missed predictions included Los Angeles (this outcome is made using an out-of-sample prediction), San Antonio, Salt Lake City, Buffalo, and Jacksonville. For the first three misses, the model predicted the cities would have teams but they do not (for our purposes, any probability greater than 0.50 means the city should have a team). Buffalo and Jacksonville have teams but the model predicted that they do not.

Table 3: Predictions* All predictions are within-sample, except for the out-of-sample forecast for Los Angeles.

Metropolitan Area Actual
of a Team
New York 1 1.00
Los Angeles* 0 1.00
Chicago 1 1.00
Washington- Baltimore 1 1.00
SF- Oakland-San Jose 1 1.00
Philadelphia 1 1.00
Boston 1 1.00
Detroit 1 1.00
Dallas 1 1.00
Houston 1 1.00
Atlanta 1 1.00
Miami 1 0.91
Seattle 1 1.00
Phoenix 1 0.95
Minnesota 1 1.00
Cleveland 1 0.99
San Diego 1 0.98
St. Louis 1 0.99
Denver 1 1.00
Tampa 1 0.79
Pittsburgh 1 1.00
Portland 0 0.22
Cincinnati 1 0.78
Sacramento 0 0.43
Kansas City 1 0.97
Green Bay-Milwaukee 1 1.00
Orlando 0 0.09
Indianapolis 1 0.60
San Antonio 0 0.56
Norfolk 0 0.03
Las Vegas 0 0.25
Columbus 0 0.39
Charlotte 1 0.80
New Orleans 1 0.85
Salt Lake City 0 0.51
Austin, TX 0 0.05
Nashville 1 0.85
Providence 0 0.00
Raleigh-Durham 0 0.09
Hartford 0 0.04
Buffalo 1 0.17
Memphis 0 0.35
West Palm 0 0.12
Jacksonville 1 0.03
Grand Rapids 0 0.00
Oklahoma City 0 0.04
Richmond 0 0.05
Greenville 0 0.00
Dayton 0 0.00
Birmingham 0 0.10

Table 4 shows the top five candidate cities to have teams either through expansion or relocation. The logit model predicted that Los Angeles would have a team with a probability of 1.0, which is not surprising given that it once had two teams. However, if NFL owners continue to use the city as a credible threat, and if a new stadium is not in the package for the Los Angeles team, then the other cities in this list deserve consideration. Next is San Antonio, where the Saints have already tested the waters with great success. The logit model estimated that San Antonio had a probability of 0.56 in obtaining an NFL franchise either through relocation or expansion. Salt Lake City is a more marginal candidate at 0.51, and the model suggests that both Sacramento and Columbus are not viable candidates (their predicted probabilities are less than 0.5).

Table 4: Predicted Probabilities for Candidate Cities from the Sample

Los Angeles 1.0
San Antonio 0.56
Salt Lake City 0.51
Sacramento 0.43
Columbus 0.39

League expansion and the open Los Angeles market have been discussed in the NFL by both the outgoing Commissioner Paul Tagliabue (NFL.com, 2004) and the new Commissioner Roger Goodell (Farmer, 2007), but there is no short term timetable for expanding to a 33rd team or moving a troubled franchise there. Nonetheless, the league has been working with investor groups representing sites at the Los Angeles Coliseum and the Rose Bowl in Pasadena (NFL.com 2004). Besides New Orleans, Buffalo and Jacksonville are mentioned as cities that might lose their franchises (The Sports Economist, 2006).

Do the Saints Go Marching Back In?

Under pre-Katrina conditions, New Orleans had a probability of 0.85 for its in-sample prediction for having a team (see Table 3). But this changes when the potential depopulation of New Orleans is considered. The extensive damage to the city of New Orleans was not only to the industrial and commercial structures. Whole residential sections of the city were destroyed and depopulated.

The model prediction included depopulation in two parts. First, the New Orleans metropolitan area population was reduced by 10, 20, 25, and 30%.4 Second, the Nielsen television ratings were correspondingly reduced. No adjustment was made to the Fortune 500 headquarters because New Orleans has only one such company, Entergy Corporation, and it will remain in the area.

Table 5 shows the simulation results for different assumptions about permanent depopulation for New Orleans. The most recent population estimate from the Census Bureau dates from July 1, 2006. At this time, a 400,000 loss was announced (Whoriskey, 2006). This 30% decline would put the predicted probability for New Orleans at approximately 0.43. Only Buffalo and Jacksonville have teams with lower probabilities than this. But this worse case scenario is outdated. A portion of the 400,000 have returned to the metro area since June 20065, but how many will ultimately return? At this time, there is no planned Census Bureau update for the New Orleans metropolitan population.6 Should half of the displaced 400,000 return, the model would put place a probability of 0.59 that New Orleans will have (in this case keep) a team. The most optimistic non-official estimate (as of December 20, 2006) put the metro area at 1.2 million (Savidge, 2006). This is less than 10% depopulation, and the model provided a more comfortable 0.74 probability of having (retaining) a football team.

Table 5: Predicted Probability of a Team in New Orleans with Depopulation

10 % 0.74
20 0.59
25 0.51
30 0.43


Expansion and relocation of franchises in the National Football League remains an active topic when one considers the fates of both the Los Angeles and New Orleans markets. Although expansion is not a current short term goal for the NFL, relocation of teams in weak markets remains an annual possibility. The researchers have estimated a model that identifies those weak teams based on economic and demographic factors, and, more importantly, identifies candidate cities for new or relocated teams. Buffalo, Jacksonville, and a depopulated New Orleans are vulnerable to losing their teams, while Los Angeles and San Antonio are viable candidates to offer new homes to teams. What happens next depends on the interests of the current owners and the investor groups in the candidate cities, as well as the state and local government support for new stadiums in the old or new locations.


1 “The most recent NFL expansion, when the league was deciding between Houston and Los Angeles, is instructive on this point. In general terms, the decision between the two locations hinged on two considerations regarding the Houston and Los Angeles markets. First, the league considered the financial contribution that either location would make to the league. Second, it considered the value of an open location and the negotiating advantages it provided to current league membership. Keeping the best believable threat location helps owners in negotiations with their current host cities (Fort, 2006, p. 393).

2 Gary Roberts, a professor at Tulane University Law School and an expert in sports business issues, states “Everyone knows New Orleans was a marginal major league market. More and more, the NFL has come to rely on corporate dollars and New Orleans doesn’t have a very large corporate base” (Isidore, 2005).

Then Football Commissioner Tagliabue comments on whether New Orleans can support an NFL team long term: “[team owner] Benson has strong personal and professional ties to San Antonio, the suspicion remains that he would prefer to permanently locate the franchise there. Benson fears that the rebuilding of New Orleans, a process expected to take years, will threaten the team’s financial viability” (Pasquarelli, 2005).

3 A Nielsen TV rating is the percentage of households watching that particular television program out of all households with televisions. A TV share is the percentage of televisions in use that are watching that particular program.

4 The population growth variable was not altered because its coefficient 1) had the wrong sign, 2) a very small magnitude, and 3) was highly insignificant statistically.

5 “As a city in flux New Orleans remains statistically murky, but demographers generally that the population replenishment after the storm, as measured by things like the amount of mail sent and employment in main economic sectors, has leveled off” (Dewan. 2007).

6 The Census Bureau is “just not equipped to provide real time population estimates in a situation that is changing as rapidly as New Orleans” (Plyer, 2007).


Brooks, Rick A celebration at the Superdome, Wall Street Journal Online September
23, 2006

Bruggink, Thomas H. and Justin Zamparelli Emerging markets in baseball: An
econometric model for predicting the expansion teams’ new cities, Economics:
Current Research
Praeger Press (1999): 49-59.

Dewan, Shaila Fed-Up New Orleans residents are giving up, New York Times
February 16, 2007. www.nytimes.com/2007/02/16/us/nationalspecial/16orleans

Fort, Rodney, Sports Economics, Pearson Prentice-Hall (2006)

Hage, Jim A city is up and running, washingtonpost.com January 1, 2006

Isidore, Chris New Orleans’ muddy sports future, Money.cnn.com September 22,

Maske, Mark, and Leonard Shapiro, Saints could end up in L.A.
Washingtonpost.com October 27, 2005

NFL.com NFL eyes Los Angeles for 2008, NFL.com wire reports May 25, 2004.

Orsborn, Tom. Football: The San Antonio…Chargers? Mysantonio.com January 13, 2006

Pasquarelli, Len. New Orleans Saints, ESPN.com December 30, 2005.

Plyer, Allison. Correspondence with Allison Plyer of Greater New Orleans Nonprofit
Knowledge Works (gnocdc.org) January 19, 2007.

Rascher, Daniel, and Heather Rascher Optimal markets for NBA expansion and
relocation, Sports Economics Perspectives Issue 2 November 2005

Savidge, Martin New Orleans returning slowly, MSNBC.com December 20, 2006.

Wetzel, Dan. Calling a T.O. for N.O., Sports.yahoo.com September 4, 2005

Print Friendly, PDF & Email