Content uploaded by Steven Salaga
Author content
All content in this area was uploaded by Steven Salaga
Content may be subject to copyright.
http://jse.sagepub.com/
Journal of Sports Economics
http://jse.sagepub.com/content/early/2013/02/07/1527002513477662
The online version of this article can be found at:
DOI: 10.1177/1527002513477662
published online 11 February 2013Journal of Sports Economics
Steven Salaga and Jason A. Winfree
League Personal Seat Licenses and Season Ticket Rights
Determinants of Secondary Market Sales Prices for National Football
Published by:
http://www.sagepublications.com
On behalf of:
The North American Association of Sports Economists
can be found at:Journal of Sports EconomicsAdditional services and information for
http://jse.sagepub.com/cgi/alertsEmail Alerts:
http://jse.sagepub.com/subscriptionsSubscriptions:
http://www.sagepub.com/journalsReprints.navReprints:
http://www.sagepub.com/journalsPermissions.navPermissions:
What is This?
- Feb 11, 2013OnlineFirst Version of Record >>
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Determinants of
Secondary Market
Sales Prices for
National Football
League Personal Seat
Licenses and Season
Ticket Rights
Steven Salaga
1
and Jason A. Winfree
2
Abstract
A strong secondary market has emerged where National Football League (NFL)
personal seat licenses (PSL) and season ticket rights (STR) are sold electronically.
These data allow us to estimate determinants of market prices for long-run access to
NFL attendance. The analysis finds that high-quality seating locations are a strong
determinant of price. Clear differences exist between PSL and STR markets in
regard to both general interest in the live NFL product and the effect of team quality
on market price. Furthermore, we find that higher face value ticket prices are asso-
ciated with lower secondary market PSL and STR sales prices.
Keywords
National Football League, personal seat license, season ticket rights
1
College of Business, Florida Institute of Technology, Melbourne, FL, USA
2
University of Michigan, Ann Arbor, MI, USA
Corresponding Author:
Jason A. Winfree, Program in Sport Management, University of Michigan, Observatory Lodge, 1402
Washington Heights, Ann Arbor, MI 48109, USA.
Email: jwinfree@umich.edu
Journal of Sports Economics
00(0) 1-27
ª The Author(s) 2013
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1527002513477662
jse.sagepub.com
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Introduction
Despite the league’s immense popularity, there are a limited number of academic
contributions pertaining to the economics of the National Football League (NFL;
Fizel, 2006). This is especially true of studies analyzing the factors contributing
to overall consumer interest in the league, the determinants of ticket prices and
game-day attendance. This study addresses this void in the literature by identifying
the factors which influence prices for access to NFL attendance. This is achieved by
analyzing a unique data set of secondary market personal seat license (PSL) and
season ticket right (STR) sales transactions.
Possession of a PSL or STR grants the owner the right to buy NFL season tickets
at face value for a particular seat in the stadium generally for as long as that venue is
in operation. In the vast majority of markets, there is a higher quantity demanded for
attendance than what NFL teams have chosen to accommodate based on their com-
bination of ticket prices and seating capacities. Numerous studies have verified that
professional sports franchises price in the inelastic portion of demand, and this lit-
erature is summarized and expanded upon by Fort (2004). Given that nearly all NFL
games are sold-out (DeSerpa, 1994; Noll, 1974) face value ticket prices set by NFL
franchises do not allow for an accurate representation of the market price of attend-
ing a game. Likewise, utilizing attendance as the dependent variable in estimating
demand for the live NFL product is problematic as the actual quantity demanded
given face value ticket prices is truncated by fixed stadium capacities. Therefore, the
‘‘venue capacity’’ variable has dominated the majority of previous estimations
aimed at analyzing NFL game-day attendance. Subsequently, this contribution
offers a way around this issue. As an alternative to utilizing attendance as the depen-
dent variable, we employ a unique data set of secondary market PSL and STR sales
transactions and uncover the factors which drive PSL and STR prices. It should be
noted that this article analyzes market prices of PSLs and STRs and not secondary
market supply and demand.
This study contributes to the literature on professional sports ticket pricing and
attendance in a number of ways. First, a brief theoretical model of the secondary PSL
and STR market is provided. In this market, prices are not fixed, which allows us to
avoid statistical problems and revisit a line of investigation which has been relatively
dormant since Noll’s (1974) seminal contribution. The use of a unique data set pro-
duces an unobstructed view of the factors that drive market prices for access to NFL
attendance. Our modeling shows that seating location is a key driver of secondary
market sales prices for both PSLs and STRs. Also evident is a clear inverse relation-
ship between PSL and STR acquisition prices and the face value price of season tick-
ets. The results also illustrate that PSLs depreciate in value over time as the option
time to buy season tickets directly from the franchise shrinks. Finally, the findings
illuminate clear differences between PSL and STR markets in regard to both general
interest in the live NFL product and the effect of team quality on market price.
2 Journal of Sports Economics 00(0)
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
The article proceeds with some background on PSLs, a brief literature review, a
theoretical treatment of PSL and STR sales in the secondary market, outline of the
data and statistical models, estimation results, and conclusion.
Background
There are a very limited number of contributions investigating the role of PSLs in the
revenue generation landscape of professional sports franchises (e.g., Fort, 2003;
Howard & Crompton, 2004; Noll & Zimbalist, 1997; Reese, Nagel, & Southall,
2004). Furthermore, we are not aware of any empirical work examining PSLs or
STRs in either the primary or secondary market. Based on the overall lack of atten-
tion paid to the topic, a brief overview appears to be warranted.
PSL sales in the NFL are regulated by the league’s Collective Bargaining Agree-
ment (CBA). Specifically, franchises are only able to sell PSLs in conjunction with
the construction or the renovation of a venue. One significant advantage of imple-
menting a PSL program is that revenue collected from PSL sales is eligible for
exemption from the NFL revenue sharing formula. If franchises successfully apply
for a league ‘‘waiver,’’ they are permitted to retain revenues generated from PSL
sales. Under the alternative, which is simply increasing the face value price of season
tickets, franchises would be forced to share a portion of these increased revenues
under ‘‘gross receipts’’ as is specified under the league’s revenue sharing formula
(National Football League Players Association, 2006). Consequently, for the fran-
chise that is unable to secure public funding, the implementation of a PSL sales pro-
gram can be used as a tool to raise capital toward the private construction of a venue.
Since PSL sales programs are regulated by the NFL CBA, franchises are able to
generate additional capital through a PSL sales program and are then able to parlay
those PSL revenues into an upgraded venue with enhanced revenue production cap-
abilities. The ability to construct or renovate venues stocked with luxury boxes and
premium seating is a vital mechanism needed to generate revenues. Since 1995, 24
of the 32 NFL franchises have either renovated or constructed new venues. Fifteen of
those clubs have used the sale of PSLs in order to generate capital to fund these ven-
tures. Those which did not incorporate a PSL sales program were either the recipi-
ents of publically funded stadiums or used the sale of stadium naming rights to cover
the franchise’s private contribution toward venue construction. Table 1 shows the
growth of PSLs in the NFL.
Over the past two decades, PSL sales have become a significant source of revenue
for NFL franchises. Though access to team financial statements are unavailable, rev-
enues collected from the implementation of PSL programs are undoubtedly substan-
tial. In 1993, the Carolina Panthers were the first NFL franchise to utilize a modern
PSL sales program, which raised an estimated $100 million in after tax revenue (Ost-
field, 1995). More recently, Dallas Cowboys owner Jerry Jones charged a PSL fee
ranging from $16,000 to $150,000 for each of his 15,000 club seats in the new
Salaga and Winfree 3
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Cowboys Stadium (Sandomir, 2009). Fees collected by the Cowboys in association
with the construction of their new venue are estimated at $720.40 million (Vrooman,
2010). Clearly, PSL fees represent a sizable portion of ‘‘gate receipt’’ revenues col-
lected by many NFL franchises. Though the case of the Cowboys may not represent
the league norm, it does provide a frame of reference for the potential revenue a PSL
program may generate.
Traditionally, single-game ticket sales have dominated the secondary market, but
over the past two decades, PSL and STR sales have become more common. In the
primary market, PSLs and STRs are clearly two distinct products. PSLs require an
up-front fee which grants the consumer the ability to purchase both season tickets
and playoff tickets from the franchise each year.
1
STRs represent the right to pur-
chase season and playoff tickets directly from franchises that do not implement for-
mal PSL programs and require no up-front fee. For both products, a failure to pay the
franchise the face value price for season tickets results in the loss of the PSL or STR,
while continuation of payment on a yearly basis maintains the consumer’s status as a
PSL or STR holder. In the case of a PSL, the consumer is purchasing an asset which
grants the ability to purchase face value season and playoff tickets for an extended
period of time. As an STR holder, the same right is conveyed to the consumer
without the up-front fee.
While there are differences between these two products in the primary market-
place, the distinction between these commodities is eliminated in the secondary
market. In the case of both products, the asset being transferred from consumer to
consumer is simply the right to purchase season and playoff tickets directly from the
franchise. The rights to NFL season tickets are a valuable piece of property (Reese
Table 1. Timeline of NFL PSL Programs.
Year Venue Opened Team Name of Program
1995 St. Louis Rams PSL
1996 Carolina Panthers PSL
1999 Tennessee Titans PSL
1998 Baltimore Ravens PSL
1999 Cleveland Browns PSL
2001 Pittsburgh Steelers Seat License
2002 Houston Texans PSL
2002 Seattle Seahawks Charter Seat License
2003 Chicago Bears PSL
2003 Green Bay Packers PSL
2003 Philadelphia Eagles Stadium Builder License
2004 Cincinnati Bengals Charter Ownership Agreement
2009 Dallas Cowboys Seat License
2010 New York Giants PSL
2010 New York Jets PSL
Note. PSL ¼ personal seat licenses.
4 Journal of Sports Economics 00(0)
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
et al., 2004) and a consumer purchasing either a PSL or an STR in the secondary
market has equivalent options.
2
First, the consumer can simply choose to purchase
season tickets from the franchise each year and maintain their PSL or STR rights. An
alternative option is to resell the PSL or STR asset on the secondary market, where a
profit or loss may be realized based on market conditions. A final and highly
unlikely option is for the consumer to decline their option of purchasing season tick-
ets from the franchise, in which case the secondary market resale value of the PSL or
STR would be forfeited along with PSL and STR ownership rights. Therefore, in the
secondary marketplace, a PSL or STR sales transaction represents a purchase of the
rights to NFL attendance.
It is important to note, however, that there are differences between pricing
mechanisms in the primary and secondary market. In the case of a primary market
PSL sale, the consumer has the opportunity to purchase a PSL from the franchise that
is typically at a predetermined fixed price based on seat quality and other compo-
nents associated with individual franchise demand. In the secondary market, pricing
represents a dynamic process in which transaction prices fluctuate based on numer-
ous determinants. Both buyers and sellers have the ability to view recent sales trans-
actions resulting in secondary market prices changing more fluidly based on existing
market conditions. This distinction is one component which makes these data
unique.
The use of secondary market PSL and STR sales data also allows for the oppor-
tunity to model a more representative purchase decision by the consumer. NFL fran-
chises typically sell the vast majority of their seating capacity in the form of advance
season tickets, which largely eliminates the game-day ‘‘walk-up’’ consumer. There-
fore, examining the purchase decision in advance, which is the case when a con-
sumer purchases a secondary market PSL or STR, may more accurately reflect
the true purchase decision regarding NFL attendance. When a consumer acquires
a PSL or STR, they are committing to the purchase of not only a single game, but
10 total (two preseason and eight regular season) home games per season as well
as the rights to purchase season tickets into the future and the option to sell the asset.
Furthermore, there has been a lack of empirical work inspecting the sale of PSLs
and STRs.
3
At this point in time, we are not aware of any published research analyz-
ing either primary or secondary market data on either item. Accordingly, this article
uses secondary market PSL and STR sales data to estimate market prices for the
rights to NFL attendance. Over 3,800 secondary market sales transactions occurring
from 2005-2009 are used in the analysis.
4
Literature Review
Despite the fact that this study estimates the determinants of secondary market PSL
and STR sales prices, the foundation of this work stems from the general attendance
demand literature in professional sports and in particular the issues associated with
Salaga and Winfree 5
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
estimating NFL demand. Empirical research on demand for professional team sports
has received a significant amount of attention in the sports economics literature. In
empirical demand studies, game-day attendance has traditionally been used as a
proxy for demand (Krautmann & Hadley, 2006). The earliest research on attendance
demand in professional team sports was completed by Demmert (1973) and Noll
(1974). Demmert’s work examined professional baseball attendance while Noll
highlighted the similarities and differences between the four major North American
team sports. Both of these early empirical pieces paved the way for future research in
the area.
On the whole, the attendance demand literature has largely focused on Major
League Baseball (MLB) and European football. There are also studies that estimate
attendance in the National Basketball Association (NBA) and the National Hockey
League. However, despite the relative importance of the NFL, few studies have
examined the demand for league attendance. In part because of the league’s rela-
tively short regular season schedule, the NFL has traditionally played to capacity
or near-capacity crowds. Noll’s seminal work (1974), which was the first to estimate
demand for pro football, noted that this capacity constraint explained almost all of
the interteam differences in game-day attendance. Outside of stadium capacity,
short-term team quality was the only other variable providing explanatory power
despite including many of the traditional control variables and demand drivers. Noll
also noted that since almost all sales were for advanced season tickets, short-term
team quality was a fundamentally important factor in determining attendance.
Following Noll’s initial work on the topic, relatively few contributions focused
on analyzing NFL attendance data. Welki and Zlatoper (1994, 1999) examined NFL
attendance data from the late 1980s and early 1990s in two separate contributions.
They uncovered negative relationships between game-day attendance and both
ticket prices and income. Predictably, the quality of the game matchup, and in
particular, the quality of the home team was associated with increased attendance.
Also of interest was the finding that divisional matchups and contests played on
non-Sundays were associated with higher attendance figures.
Putsis Jr. and Sen (2000) followed by analyzing NFL attendance demand in asso-
ciation with the league’s blackout rule. The authors expanded on the previous liter-
ature by estimating demand for both individual game tickets and season tickets.
Coinciding with previous empirical work, the authors uncovered that demand for
NFL attendance was inelastic as most franchises could easily increase ticket prices.
The effect of income on demand was ambiguous as income was positively associated
with season ticket demand but negatively tied to single-game attendance. Also in
agreement with previous research was a positive relationship between team quality
and demand as teams that reached the playoffs and had higher winning percentages
in the previous season experienced attendance increases.
Coates and Humphreys (2007) estimated demand for the NFL, MLB, and NBA,
but also incorporated the costs of ancillary attendance items specified by the fan cost
index (FCI). The authors concluded that deriving demand for the NFL was
6 Journal of Sports Economics 00(0)
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
fundamentally different as compared to the other leagues. Ticket price and the FCI
were found to be insignificant in determining demand in the NFL, but not so in MLB
and the NBA. The authors claim that this between-league variation can be attributed
to the fact that the NFL sees such a large percentage of their contests played to
capacity.
More recently, researchers have explored estimating consumer interest in the
NFL through the use of nonattendance based response variables. Nagel and Rascher
(2007) used franchise merchandise sales to uncover between-team variation in the
demand for licensed products. Alternatively, Tainsky (2010) used television ratings
as a proxy for NFL demand and uncovered strong consumer preferences for both
home and visiting short-term team quality. In agreement with the findings of both
Noll (1974) and Welki and Zlatoper (1994), an inverse relationship between
personal income and viewership was also revealed.
Each of the empirical pieces mentioned above has enhanced our understanding of
the primary demand determinants in professional football. However, there are still
some lingering issues with estimating consumer interest in attending the live NFL
product—namely the prevalence of advance season ticket sales and the venue capacity
constraint. As both Noll (1974) and Putsis Jr. and Sen (2000) previously identified, the
decision on purchasing a season ticket package, which is achieved through ownership
of a PSL or STR, is much different than the decision on purchasing a single-game
ticket. Specifically, there are differences in many of the factors that would influence
the purchase decision of a single-game ticket as opposed to a full season package
(Fizel & Bennett, 1989; Putsis Jr. & Sen, 2000). For example, game-specific factors,
such as variables capturing weather conditions, opponent team quality, and temporal
variables such as whether the game is played on a weekend or nonweekend are not
appropriate for inclusion when estimating prices for PSLs and STRs.
Second, previous empirical research estimating demand for game-day NFL atten-
dance has found that venue capacity has been the dominating variable. With this
capacity constraint dictating the results, a lack of significance in other independent
variables has routinely been discovered. The work of Noll (1974), Putsis Jr. and Sen
(2000), and Coates and Humphreys (2007) previously identified this issue and the
consensus is that there are fundamental differences in estimating NFL attendance
demand as compared to the other North American leagues.
We account for the two concerns above by estimating the factors which influence
the actual market price of long-run access to NFL attendance. Using secondary mar-
ket sales prices as the dependent variable, there is no upper limit truncation as is tra-
ditionally seen when using game-day attendance as the response variable—therefore
avoiding the venue capacity constraint. This choice of response variable also
accounts for the fact that advance season ticket sales are undoubtedly the norm in
the NFL.
5
This allows for the ability to estimate market prices for access to NFL
attendance based on market conditions at the time of the sale as opposed to the day
of the game. Given the combination of the high percentage of advanced season ticket
sales, the negligible walk-up game-day customer and the explosion of the secondary
Salaga and Winfree 7
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
market (Drayer & Martin, 2010), it is a reasonable assumption that this is more
representative of how consumers gain access to NFL season tickets.
Theoretical Model
In this market, consumers are charged both a fixed up-front fee in the form of a PSL
or STR and then must also purchase season tickets at face value from the franchise
each year. But while the PSL might technically be a two-part pricing mechanism,
given that PSL sales do not increase the quantity of tickets sold in any way, the tra-
ditional economic model of two-part pricing falls short when evaluating this product
in either the primary or secondary market. In other words, this type of two-part pric-
ing is not a way to allow teams to capture consumer surplus that could not otherwise
be captured with an increase in ticket price. This is clear given the propensity of
sellouts for NFL games regardless of PSL sales.
6
This section shows the economic
intuition behind secondary market PSL sales.
Since a PSL or STR represents the right to buy season tickets for the lifetime of
the stadium, the value of the PSL (or STR) can be written as
PSL
M
¼
X
N
t¼1
GEP
M;t
P
FV;t
1 þ iðÞ
t
; ð1Þ
where PSL
M
represents the market price for the PSL, G is the number of home
games in a season, EP
M;t
P
FV;t
is the expected difference between the market
price (P
M
) and the face value (P
FV
) of one ticket for one game in year t, i is the dis-
count rate, and N is the number of seasons that the PSL gives the owner the right to
purchase the tickets. The PSL only has a value if the expected market value of the
ticket exceeds the face value. Furthermore, in most, if not all cases, there is no expli-
cit agreement by league franchises as to what the face value of the tickets will be in
the future. Yet, it appears to be the case that there is an implicit agreement that ticket
prices will not increase to the market price.
Since it is not the focus of our article, this model does not describe why teams
issue a PSL as opposed to simply increasing ticket price. However, there is a clear
incentive for teams to issue PSLs because this revenue is not shared under the
current CBA, while ticket revenue is shared. So, it seems as though it is in the best
interest of the franchise to capture as much of the ticket revenue through the sale of
PSLs as possible.
7
Equation 1 illustrates the relationship between the value of PSLs and the price of
tickets. If we make the simplifying assumption that there is a constant difference
between the market price and the face value price of the ticket in perpetuity, then
the market price can be written as
PSL
M
¼
G
i
P
M
P
FV
½; ð2Þ
8 Journal of Sports Economics 00(0)
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Consequently, under this assumption, the price of a PSL has a direct linear rela-
tionship with the price of a ticket. Hence, estimating PSL sales prices is analogous to
estimating the market price of tickets. We note that in the secondary market, there is
no distinction between PSLs and STRs.
Empirical Models
Three fixed effects models and three random effects models are used to estimate
market prices for NFL PSLs and STRs. The fixed effects models include team indi-
cator variables for each franchise represented in the data set. The random effects
models differ from the fixed effects models in that the team indicator variables are
omitted from the estimations. Both the fixed effects models and random effects mod-
els include (1) pooled PSL and STR observations, (2) PSL observations only, and (3)
STR observations only. Following in line with the early empirical work of Demmert
(1973), Noll (1974), and Schofield (1983), these six models utilize demographic,
temporal, team-specific, stadium-specific, and economic variables in an attempt to
isolate the determinants of sales prices in this secondary marketplace.
The general form of the random effects model follows:
LN SEAT PRICEðÞ¼b
0
þ b
1
YEAR06 þ b
2
YEAR07 þ b
3
YEAR08
þ b
4
YEAR09 þ b
5
LN ROWðÞþb
6
SEATQUAL1
þ b
7
SEATQUAL2 þ b
8
SEATQUAL3 þ b
9
SEATQUAL4
þ b
10
LN STADIUM AGEðÞþb
11
LN STADIUM CAPACITYðÞ
þ b
12
TYPEPSL þ b
13
LN TICKET PRICEðÞþb
14
AISLE
þ b
15
WIN3 þ b
16
LOCAL UNEMPLOYMENT
þ b
17
LN LISTðÞþb
18
LN POPULATIONðÞ
þ b
19
LN INCOMEðÞþb
20
DOME þ e:
The general form of the fixed effects model follows:
LN SEAT PRICEðÞ¼b
0
þ b
1
YEAR06 þ b
2
YEAR07 þ b
3
YEAR08 þ b
4
YEAR 09
þ b
5
LN ROWðÞþb
6
SEATQUAL1 þ b
7
SEATQUAL2
þ b
8
SEATQUAL3 þ b
9
SEATQUAL4
þ b
10
LN STADIUM AGEðÞþb
11
LN TICKET PRICEðÞ
þ b
12
AISLE þ b
13
WIN3 þ b
14
LOCAL UNEMPLOYMENT
þ b
15
LN POPULATIONðÞþb
16
LN INCOMEðÞ
þ b
17:38
TEAM INDICATORS þ e:
Table 2 lists the variables and provides a brief explanation of each.
Salaga and Winfree 9
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Table 2. Variable Descriptions.
Variable Description
LN(SEAT PRICE) logged per seat sale price of a PSL or STR (dependent variable)
YEAR05 indicator variable; 1 ¼ PSL or STR sold in 2005, 0 ¼ not sold in
2005 (baseline category)
YEAR06 indicator variable; 1 ¼ PSL or STR sold in 2006, 0 ¼ not sold in
2006
YEAR07 indicator variable; 1 ¼ PSL or STR sold in 2007, 0 ¼ not sold in
2007
YEAR08 indicator variable; 1 ¼ PSL or STR sold in 2008, 0 ¼ not sold in
2008
YEAR09 indicator variable; 1 ¼ PSL or STR sold in 2009, 0 ¼ not sold in
2009
LN(ROW) logged row of seating location associated with PSL or STR sale
SEATQUAL1 seat location: lower level between 30-yard lines
SEATQUAL2 seat location: upper level between 30-yard lines
SEATQUAL3 seat location: lower level between 30-yard lines and rear end
zone lines
SEATQUAL4 seat location: upper level between 30-yard lines and rear end
zone lines
SEATQUAL5 seat location: lower or upper level end zone seating (baseline
category)
LN(STADIUM AGE) logged age of team’s stadium during year of PSL or STR sale
LN(STADIUM
CAPACITY)
logged capacity of team’s stadium during year of PSL or STR sale
TYPEPSL indicator variable; 1 ¼ PSL sale, 0 ¼ STR sale
LN(TICKET PRICE) logged single-game ticket price for the seating location associ-
ated with the PSL or STR sale
AISLE indicator variable; 1 ¼ seating location is on aisle, 0 ¼ seating
location is not on aisle
WIN3 cumulative team win percentage over the three seasons prior to
the PSL or STR sale
LOCALUNEMPLOYMENT local MSA unemployment rate during the month of the PSL or
STR sale
LN(LIST) logged length of team’s season ticket waiting list
LN(POPULATION) logged MSA population during the year of the year of the PSL or
STR sale
LN(INCOME) logged MSA per capita income during the year of the PSL or STR
sale
DOME indicator variable; 1 ¼ PSL or STR is located in a domed or
retractable roof venue, 0 ¼ otherwise
TEAM CITY/STATE team indicator variables (included only in fixed effects models)
Note. MSA ¼ metropolitan statistical area; PSL ¼ personal seat licenses; STR ¼ season ticket rights.
10 Journal of Sports Economics 00(0)
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Data Description
The data used in this analysis are secondary market PSL and STR sales transactions
occurring between 2005 and 2009. Three thousand, eight hundred twenty-one obser-
vations were collected in total. All transactions were gleaned from either www.season-
ticketrights.com or www.ebay.com. The transactions from the former location were
exclusively fixed-price transactions, while data from the latter were either fixed-
price or timed auction style transactions. Only actual sales transactions are included
in the sample, as many PSLs and STRs were listed for sale on these secondary market
locations over the examination period, but were not sold at the listed asking price.
The dependent variable, LN(SEAT PRICE), is the logged per seat sale price of the
total PSL or STR sales transaction. In certain cases, where ancillary items, such as
yearly parking passes were included in the sale, the value of the sale price was
adjusted accordingly. Other explanatory variables that are specific to the transaction
were also collected from the aforementioned websites. These variables include
AISLE and LN(ROW). AISLE is an aisle seating indicator and LN(ROW) is the logged
row number associated with the season ticket location. A lower row number repre-
sents a seat location closer to the field of play within a section and therefore a higher
quality seat location.
SEATQUAL1, SEATQUAL2, SEATQUAL3, SEATQUAL4, and SEATQUAL5 are
indicator variables included to capture consumer preferences for specific seating
locations within a venue. Information on the exact seating locations for each variable
is available in Table 2. While LN(ROW) attempts to capture seat quality as proxied
by row location within a specific section, these variables account for the overall
quality of the seating location in terms of closeness to both midfield and the field
of play.
Stadium-related data, such as LN(STADIUM AGE), LN(STADIUM CAPACITY),
LN(TICKET PRICE), and DOME were collected from each team’s official website.
LN(STADIUM AGE) represents the logged age of the stadium. LN(STADIUM
CAPACITY) accounts for the logged number of seats in the venue and controls for
the relative quantity of PSLs and STRs available for a given team in the secondary
market. LN(TICKET PRICE) is the logged per game season ticket price (face value)
of the seat. Ticket prices associated with specific seating locations were not able to
be obtained for the Pittsburgh Steelers and the Washington Redskins. Subsequently,
average ticket prices for both standard and premium seating locations (collected
from Team Marketing Report, 2009) were used for each franchise.
8
DOME is an
indicator variable equal to one if the NFL franchise plays their home games in a
dome or retractable roof venue.
TYPEPSL is an indicator variable coded equal to one if the sale is a PSL transac-
tion and coded zero if the sale is a STR transaction. TYPEPSL may affect secondary
market sale price since this variable compares secondary market sales for franchises
with PSL programs against those with STR programs. Franchises with PSL programs
Salaga and Winfree 11
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
were able to utilize this sales strategy due to significant demand for their product.
Alternatively, clubs not implementing PSL programs fall into one of three possible
scenarios. These franchises (1) have either constructed or renovated new venues and
decided not to use a PSL sales program, (2) did not have enough season ticket demand
to justify implementing a PSL sales program, or (3) have not built a new venue or
renovated their current venue since 1993.
LN(LIST) is a variable specifying the logged length of each franchise’s season
ticket waiting list. This information was collected from numerous sources, including
official team websites, www.forbes.com, and from telephone correspondence with
NFL franchise ticket sales employees. The length of a team’s season ticket waiting
list could affect secondary market sales prices as a longer waiting list to acquire sea-
son tickets directly through the franchise may spur fans to search for purchase
options in the secondary market.
WIN3 represents the cumulative team winning percentage over the three seasons
prior to the PSL or STR sale.
9
A higher 3-year winning percentage would represent a
higher quality on-field product. If preferences for high team quality hold, differences
in 3-year team winning percentage should affect secondary market sales prices.
10
YEAR05, YEAR06, YEAR07, YEAR08, and YEAR09 are indicator variables repre-
senting the year in which the PSL or STR was sold.
11
Secondary market sales in
2005 were used as the measurement baseline. Twenty-two team indicator variables
are included, representing each of the NFL franchises in the data set.
12
Each variable
is labeled using the home city or state of each franchise.
LOCALUNEMPLOYMENT is the local metropolitan statistical area (MSA)
unemployment rate during the month of the PSL or STR sale. This variable was
gathered from the U.S. Bureau of Labor Statistics (http://www.bls.gov/) and was
included to measure the effect of the health of the local economy on secondary mar-
ket sales prices. A higher local unemployment rate during the month of the sale
would represent a weaker local economy, making it reasonable to infer that fluctua-
tions in rates could alter secondary market sales prices.
LN(POPULATION) is logged MSA population during the year of the sale and
was collected from the U.S. Census Bureau (www.census.gov). LN(INCOME) is
logged MSA per capita personal income during the year of the sale and was gleaned
from the U.S. Bureau of Economic Analysis (http://bea.doc.gov/).
13
If PSLs or STRs
are normal goods, an MSA with a larger per capita level of income would suggest an
increase in secondary market sale price.
Table 3 shows the summary statistics for all variables. One benefit of these data is
the ability to examine the total cost of attendance for the consumer separated out
between the PSL or STR fee and the face value price of season tickets. If we assume
adiscountrateof10%, and given that there are 10 home games per season, our sample
shows that on average 19.1% of the total market price of attendance is paid in the form
of a PSL or STR (11.4% if we use a 5% discount rate). This percentage represents the
average percentage of the total discounted payment a fan pays for tickets through a
PSL or STR. However, if we sum up the total value of the PSLs and STRs in the
12 Journal of Sports Economics 00(0)
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
sample and compare them with total discounted market price, PSLs/STRs represent
22.5% of the total value (12.7% at a 5% discount rate). The maximum value in our
sample is 77.6% (63.5% if we use a 5% discount rate) of the total discounted market
price of attendance paid through a PSL. However, the possibility exists that our num-
bers may be skewed as this sample does not account for the entirety of the secondary
market.
Results
Robust standard errors were specified in all six models based on significant
Breusch–Pagan/Cook–Weisberg and White (1980) test results. Table 4 outlines the
price estimation results for the fixed and random effects models using the pooled
PSL and STR observations. TYPEPSL, LN(STADIUM CAPACITY), LN(LIST), and
DOME were withheld from the model since these variables do not vary within each
team and are therefore perfectly collinear with the team indicator variables.
A consistent finding in this model is evidence of strong consumer preferences for
high-quality seating locations. LN(ROW), or the row location associated with the
Table 3. Summary Statistics.
Parameter N Mean SD Minimum Maximum
LN(Seat Price) 3,821 7.134 1.518 0 10.742
Year05 3,821 0.026 0.158 0 1
Year06 3,821 0.061 0.239 0 1
Year07 3,821 0.121 0.326 0 1
Year08 3,821 0.287 0.453 0 1
Year09 3,821 0.505 0.500 0 1
LN(Row) 3,821 2.562 0.875 0 4.159
SEATQUAL1 3,821 0.109 0.312 0 1
SEATQUAL2 3,821 0.080 0.271 0 1
SEATQUAL3 3,821 0.305 0.460 0 1
SEATQUAL4 3,821 0.195 0.396 0 1
SEATQUAL5 3,821 0.311 0.463 0 1
LN(Stadium Age) 3,821 2.170 0.570 0 3.912
LN(Stadium Capacity) 3,821 11.144 0.058 11.027 11.426
TYPEPSL 3,821 0.921 0.269 0 1
LN(Ticket Price) 3,821 4.426 0.523 2.890 6.310
AISLE 3,821 0.027 0.163 0 1
WIN3 3,821 0.485 0.119 0.201 0.813
Localunemployment 3,821 6.732 2.045 3.4 17
LN(List) 3,821 4.486 4.622 0 12.206
LN(Population) 3,821 14.911 0.580 13.932 16.764
LN(Income) 3,821 10.668 0.114 10.472 11.031
DOME 3,821 0.179 0.384 0 1
Salaga and Winfree 13
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Table 4. Random and Fixed Effects Models (Pooled PSL and STR Observations).
Random Effects Fixed Effects
Coefficient Robust SE t-Statistic Coefficient Robust SE t-Statistic
Year06 0.076 0.091 0.84 0.117 0.146 0.80
Year07 0.022 0.088 0.25 0.059 0.218 0.27
Year08 0.318*** 0.092 3.45 0.309 0.272 1.14
Year09 0.845*** 0.114 7.43 0.339 0.303 1.12
LN(row) 0.232*** 0.017 13.34 0.224*** 0.015 14.84
SEATQUAL1 1.027*** 0.059 17.25 1.204*** 0.051 23.55
SEATQUAL2 0.172*** 0.056 3.06 0.133*** 0.050 2.67
SEATQUAL3 0.472*** 0.040 11.95 0.578*** 0.037 15.74
SEATQUAL4 0.735*** 0.044 16.71 0.709*** 0.039 18.37
LN(Stadium Age) 0.008 0.047 0.16 1.738*** 0.229 7.58
LN(Stadium
Capacity)
4.522*** 0.438 10.32 x x x
TYPEPSL 2.236*** 0.092 24.28 x x x
LN(Ticket Price) 0.203*** 0.038 5.38 0.314*** 0.035 9.01
AISLE 0.139 0.102 1.36 0.144* 0.077 1.87
WIN3 3.380*** 0.196 17.27 1.189*** 0.219 5.42
Localunemployment 0.001 0.018 0.08 0.062*** 0.018 3.36
LN(LIST) 0.080*** 0.005 15.01 x x x
LN(Population) 0.653*** 0.042 15.69 0.169 0.127 1.33
LN(Income) 0.646*** 0.227 2.85 8.55*** 1.987 4.30
DOME 0.328*** 0.068 4.82 x x x
Constant 61.975*** 5.386 11.51 75.346*** 21.450 3.51
PSL teams
Baltimore x x x Baseline Baseline Baseline
Carolina x x x 1.426*** 0.365 3.91
(continued)
14
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Table 4. (continued)
Random Effects Fixed Effects
Coefficient Robust SE t-Statistic Coefficient Robust SE t-Statistic
Chicago x x x 0.089 0.224 0.40
Cincinnati x x x 1.537*** 0.441 3.48
Cleveland x x x 0.154 0.354 0.43
Dallas x x x 3.067*** 0.615 4.98
Houston x x x 1.724*** 0.162 10.62
Philadelphia x x x 0.809*** 0.193 4.18
Pittsburgh x x x 1.211*** 0.237 5.11
Seattle x x x 2.047*** 0.196 10.42
ST. Louis x x x 0.528* 0.290 1.82
Tennessee x x x 0.261 0.368 0.71
STR Teams
Buffalo x x x 0.692 0.611 1.13
Detroit x x x 2.537*** 0.407 6.24
Indianapolis x x x 4.365*** 0.616 7.09
Jacksonville x x x 2.171*** 0.437 4.97
Kansas City x x x 0.541 0.438 1.23
Miami x x x 0.248 0.303 0.82
New Orleans x x x 0.829** 0.366 2.26
Oakland x x x 5.72*** 0.554 10.32
San Francisco x x x 2.990*** 0.636 4.70
Tampa Bay x x x 2.189*** 0.653 3.35
Washington x x x 3.910*** 0.480 8.15
N ¼ 3,821 ***p <.01 N ¼ 3,821 ***p <.01
R
2
¼ .6911 **p <.05 R
2
¼ .7561 **p <.05
*p <.10 *p <.10
15
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
PSL or STR sale, is significant as seats closer to the field of play are associated with
higher secondary market sales prices. Additionally, SEATQUAL1, SEATQUAL2,
SEATQUAL3, and SEATQUAL4 are all significant. This suggests significant differ-
entiation in sales prices based on seating location and strong consumer preferences
for seating locations located in the lower level and on the sidelines.
Short-term team winning percentage also has a substantial impact on the depen-
dent variable. More specifically, if a team has a 0.563 winning percentage as
opposed to a 0.500 winning percentage over the previous three seasons, secondary
market sales prices are expected to increase by 28.32% all else equal. This change
in winning percentage is the difference of only three regular season wins over the
course of three seasons. In other words, this equates to the difference of a 9-7 record
each season versus an 8-8 record each season. This finding supports previous empiri-
cal work, including Noll’s (1974) seminal piece on pro football demand, which
noted that consumer preferences for short-term team quality was the largest factor
impacting attendance outside of stadium capacity.
The coefficient on LOCALUNEMPLOYMENT illustrates the effect of economic
conditions on prices. All else equal, secondary market sales prices of PSLs and STRs
are reduced when the local MSA unemployment rate increases. A 1% increase in the
local unemployment rate in a given MSA is expected to reduce secondary market
PSL or STR sales prices by 6.40%.
In examining the team indicators, 16 of the 22 were significant at traditional lev-
els against the baseline of Baltimore. Cleveland, Tennessee, and Chicago were the
PSL franchises that were shown to be the most statistically similar to the baseline.
Interestingly, two of those three franchises are relatively new to their respective
cities. Baltimore and Tennessee are two of the most recent NFL franchises to
relocate and Cleveland was awarded a franchise following the Browns move to
Baltimore in 1996.
Table 4 also displays results from the pooled data random effects model. Many of
the results are similar to those found in the fixed effects model and will not be
rehashed here. Despite this, the pooled random effects model does provide relevant
information. Specifically, TYPEPSL has a large impact on determining secondary
market sale price. If the transaction is a PSL sale as opposed to a STR sale, second-
ary market sale price is expected to increase by 835.58% [exp(2.236) ¼ 9.3558 ¼>
9.3558 1 100 ¼ 835.58%)]. This finding makes intuitive sense for a few rea-
sons. First, in the primary market, PSLs have an acquisition fee tied to them when
they are purchased directly from the franchise while STRs do not. This means that
if a PSL becomes available on the secondary market, all else equal, the seller is
likely to both set and receive a higher sale price as compared to a STR, because
of the higher perceived value of the product. Second, franchises which enacted PSL
programs in the first place, did so because of sufficient demand for NFL football in
their market. Those which allow STR transfers through STRs either did not have suf-
ficient demand to sell PSLs or likely chose not to sell PSLs in exchange for the
ability to alter ticket prices based on demand fluctuations (e.g., the New England
16 Journal of Sports Economics 00(0)
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Patriots). Regardless, this finding demonstrates that there are significant differences
in consumer interest for the rights to attendance in PSL franchises as compared to
STR franchises.
This model also highlights the relationship between PSLs, STRs, and the face
value price of season tickets. The negative and significant coefficient on
LN(TICKET PRICE) illustrates that as the face value price of a season ticket
increases, the price a consumer is willing to pay for the corresponding PSL or STR
decreases. All else equal, if a franchise increases the face value ticket price of a
given seat by 10%, the secondary market sales price for the corresponding PSL or
STR is expected to decrease by 1.95%.
The pooled random effects model also illustrates a strong positive relationship
between LN(POPULATION) and secondary market sales prices. As MSA population
increases, secondary market sales prices also increase, as a 10% increase in logged
MSA population results in a 6.42% increase in secondary market PSL and STR sales
prices [1.1
.653
¼ 1.0642].
Table 5 shows the estimation results using only the PSL data. In the fixed effects
model, the Baltimore team indicator was again treated as the baseline and TYPEPSL,
LN(STADIUM CAPACITY), LN(LIST), and DOME were excluded from the model
due to perfect collinearity with the team indicator variables. This model produces
results that are similar to the previous pooled fixed effects model, which is due to
92.1% of the data being comprised of PSL observations.
The positive and significant coefficient on LN(INCOME) impl ies that second-
ary market PSL sales prices increase along with per capita levels of income in a
franchise’s MSA. Additionally, LN(STADIUM AGE) shows a strong negative
effect, suggesting that once franchise specific effects are accounted for, fan pre-
ferences for newer venues emerge. This could also suggest that consumers are
willing to pay higher PSL prices in venues where they hold a n option to pur-
chase season tickets directly from the fra nchise for a longer pe riod of time.
Once again, consume r prefer ences f or pr ime sea ting loc ations a re evident as
LN(ROW), SEATQUAL1, and SEATQUAL3 drive sales prices in the secondary
market.
Also of specific interest are the coefficients of increasing magnitude on the
YEAR08 and YEAR09 variables. All else equal, PSL sales prices have declined over
time as compared to the 2005 baseline. Since other influential variables are con-
trolled for, this could suggest that a PSL decreases in value over time as a PSL own-
er’s window to purchase season tickets from the franchise shrinks. These findings
could represent the possibility of asset depreciation or simply consumer awareness
of the aforementioned scenario.
The random effects model for the PSL observations is also displayed in Table 5
and because no STR observations are included, TYPEPSL is omitted from the esti-
mation. This model produces effects that are similar to the fixed effects PSL model.
Specifically, consumer preferences for both prime seating locations and high team
quality drive prices on the secondary market. Likewise, markets with larger
Salaga and Winfree 17
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Table 5. Random and Fixed Effects Models (PSL Observations Only).
Random Effects Fixed Effects
Coefficient Robust SE t-Statistic Coefficient Robust SE t-Statistic
Year06 0.114 0.095 1.21 0.145 0.157 0.92
Year07 0.055 0.093 0.59 0.152 0.258 0.59
Year08 0.299*** 0.097 3.07 0.427 0.327 1.31
Year09 0.534*** 0.122 4.36 0.403 0.357 1.13
LN(Row) 0.212*** 0.017 12.65 0.208*** 0.015 14.06
SEATQUAL1 1.107*** 0.058 19.18 1.231*** 0.049 24.88
SEATQUAL2 0.143*** 0.054 2.64 0.149*** 0.048 3.10
SEATQUAL3 0.499*** 0.036 13.81 0.592*** 0.037 16.2
SEATQUAL4 0.706*** 0.043 16.42 0.706*** 0.037 19.27
LN(Stadium Age) 0.274*** 0.054 5.10 2.027*** 0.234 8.68
LN(Stadium Capacity) 5.559*** 0.387 14.35 x x x
TYPEPSL x x x x x x
LN(Ticket Price) 0.233*** 0.038 6.20 0.315*** 0.034 9.30
AISLE 0.122 0.103 1.19 0.171** 0.082 2.08
WIN3 3.408*** 0.188 18.15 1.102*** 0.208 5.29
Localunemployment
0.095*** 0.019 5.00 0.070*** 0.018 3.92
LN(List) 0.072*** 0.005 15.23 x x x
LN(Population) 1.073*** 0.058 18.52 1.887 1.467 1.29
LN(Income) 0.456 0.359 1.27 10.591*** 2.424 4.37
DOME 0.624*** 0.069 9.00 x x x
Constant 65.751*** 4.398 14.95 126.887*** 43.122 2.94
PSL teams
Baltimore x x x baseline baseline baseline
Carolina x x x 2.833*** 1.027 2.76
Chicago x x x 2.592 1.806 1.43
(continued)
18
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Table 5. (continued)
Random Effects Fixed Effects
Coefficient Robust SE t-Statistic Coefficient Robust SE t-Statistic
Cincinnati x x x 0.919 0.749 1.23
Cleveland x x x 0.969 0.702 1.38
Dallas x x x 5.343*** 1.266 4.22
Houston x x x 3.510*** 1.179 2.98
Philadelphia x x x 2.585** 1.151 2.25
Pittsburgh x x x 1.605*** 0.417 3.85
Seattle x x x 2.757*** 0.460 6.00
ST. Louis x x x 0.317 0.304 1.05
Tennessee x x x 1.721 1.131 1.52
STR teams
Buffalo x x x x x x
Detroit x x x x x x
Indianapolis x x x x x x
Jacksonville x x x x x x
Kansas City x x x x x x
Miami x x x x x x
New Orleans x x x x x x
Oakland x x x x x x
San Francisco x x x x x x
Tampa Bay x x x x x x
Washington x x x x x x
N ¼ 3,520 ***p <.01 N ¼ 3,520 ***p <.01
R
2
¼ .6706 **p <.05 R
2
¼ .7356 **p <.05
*p <.10 *p <.10
Note. * denotes p<.1, ** denotes p<.05, *** denotes p<.01.
19
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Table 6. Random and Fixed Effects Models (STR Observations Only).
Random Effects Fixed Effects
Coefficient Robust SE t-Statistic Coefficient Robust SE t-Statistic
Year06 x x x x x x
Year07 x x x x x x
Year08 x x x x x x
Year09 0.232 0.349 0.07 0.485 0.389 1.25
LN(Row) 0.416*** 0.084 4.93 0.373*** 0.080 4.67
SEATQUAL1 1.006*** 0.309 3.25 0.898*** 0.338 2.66
SEATQUAL2 0.131 0.313 0.42 0.168 0.339 0.5
SEATQUAL3 0.475*** 0.179 2.65 0.478*** 0.171 2.8
SEATQUAL4 0.517 0.318 1.62 0.643 0.437 1.47
LN(Stadium Age) 0.127 0.159 0.79 0.375 1.014 0.37
LN(Stadium Capacity) 5.539** 2.284 2.43 x x x
TYPEPSL x x x x x x
LN(Ticket Price) 0.205 0.181 1.14 0.243 0.224 1.09
AISLE 0.003 0.288 0.01 0.079 0.251 0.31
WIN3 1.096 0.888 1.23 0.276 2.803 0.1
Localunemployment 0.275** 0.119 2.30 0.042 0.144 0.29
LN(List) 0.062 0.079 0.79 x x x
LN(Population) 0.057 0.169 0.34 0.004 0.253 0.02
LN(Income) 0.536 0.661 0.81 x x x
DOME 1.520** 0.646 2.35 x x x
Constant 58.240** 29.640 1.96 8.339 5.475 1.52
PSL teams
Baltimore x x x x x x
Carolina x x x x x x
Chicago x x x x x x
(continued)
20
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Table 6. (continued)
Random Effects Fixed Effects
Coefficient Robust SE t-Statistic Coefficient Robust SE t-Statistic
Cincinnati x x x x x x
Cleveland x x x x x x
Dallas x x x x x x
Houston x x x x x x
Philadelphia x x x x x x
Pittsburgh x x x x x x
Seattle x x x x x x
St. Louis x x x x x x
Tennessee x x x x x x
STR teams
Buffalo x x x Baseline Baseline Baseline
Detroit x x x 0.843 1.981 0.43
Indianapolis x x x 0.060 3.035 0.02
Jacksonville x x x 1.002 0.902 1.11
Kansas City x x x 0.469 0.565 0.83
Miami x x x 0.759 0.821 0.92
New Orleans x x x 1.667** 0.697 2.39
Oakland x x x 0.939 1.401 0.67
San Francisco x x x 0.077 0.582 0.13
Tampa Bay x x x 1.389 1.288 1.08
Washington x x x 0.419 1.253 0.33
N ¼ 301 ***p <.01 N ¼ 301 ***p <.01
R
2
¼ .4232 **p <.05 R
2
¼ .4399 **p <.05
*p <.10 *p <.10
Note. * denotes p<.1, ** denotes p<.05, *** denotes p<.01.
21
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
populations and lower unemployment rates have PSLs that sell for higher secondary
market prices.
The impact of LN(LIST) also illustrates the effect of primary market availability
on secondary market PSL sales prices. A 10% increase in the length of a franchise’s
season ticket waiting list produces secondary market PSL sales prices that are
0.6885% higher. In other words, if a given team’s season ticket waiting list increases
from 8,000 to 16,000, secondary market PSL sales prices are expected to rise by
6.89% all else equal.
Table 6 illustrates estimation results for STRs only. Because there are fewer STR
observations, these estimations are not as robust. No 2005, 2006, or 2007 observa-
tions exist in this portion of the data, so 2008 secondary market STR sales are used as
the baseline and only the 2009-year indicator variable is included in the estimation.
The Buffalo team indicator is excluded from the estimation and is used as the base-
line. Additionally, TYPEPSL, LN(STADIUM CAPACITY), LN(LIST), LN(INCOME),
and DOME are excluded from the estimation due to perfect collinearity with the
team indicator variables.
Similar to what is uncovered in the PSL models, consumer preferences for seating
closer to the field of play and to midfield are key drivers of access to STRs on the
secondary market. This is illustrated by the strong positive effects of LN(ROW),
SEATQUAL1, and SEATQUAL3. However, it is important to note that no significant
differences in STR sales prices were found for seating locations outside of lower
level sideline seating.
Despite the consistency of preferences for high-quality seating locations, there
are a couple of stark differences between consumer interest in PSLs versus STRs.
Specifically, LN(POPULATION) and WIN3 fail to produce any significant effect
on secondary market STR sales prices. The nonsignificant relationship between
WIN3 and the dependent variable is particularly unique based on the strong and con-
sistent positive relationship between team quality and demand in the sports econom-
ics literature. This result suggests that market prices for access to NFL season tickets
in STR markets are not responsive to changes in short-term team quality.
Also of note is that only 1 of the 10 team indicators in this model is significant at
any level (New Orleans at the .05 level). This is a notable difference from the first
two fixed effects models, where the bulk of team indicators were significant. This
suggests that the vast majority of STR franchises included in the data set do not dif-
fer significantly from Buffalo, the baseline franchise. This can be interpreted further
by stating that this collection of STR franchises does not exhibit characteristics
beyond what is controlled for by other variables in this model that would signifi-
cantly differentiate them from each other.
Finally, the STR random effects model is also shown in Table 6. LN(ROW),
SEATQUAL1, and SEATQUAL3 again produce significant effects, supporting the
notion that fan preferences for high-quality seating locations drive prices on the sec-
ondary market. WIN3 remains nonsignificant in the random effects model support-
ing the result seen in the fixed effects STR only model.
22 Journal of Sports Economics 00(0)
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Summary and Conclusions
The sale of PSLs as a franchise revenue generation mechanism has become a com-
mon practice in the NFL over the past two decades. The emergence of a strong sec-
ondary market where NFL season ticket holders sell their PSLs and STRs over the
Internet has led to an opportunity to estimate the determinants of secondary market
PSL and STR sales prices. This provides an opportunity to avoid the venue capacity
constraint which has hampered previous NFL attendance estimations. The choice of
dependent variable produces estimations which allow for a better understanding of
the factors influencing consumer interest in long-run NFL attendance. This approach
reopens a line of investigation which has been relatively inactive since Noll (1974).
The relationship between the secondary market PSL or STR acquisition fee and
the yearly purchase of season tickets is modeled mathematically. The corresponding
empirical modeling illustrates an inverse relationship between the face value price of
season tickets and the PSL or STR acquisition fee. The results also show that con-
sumer preferences for high-quality seating locations are prime drivers of both PSL
and STR secondary market sales prices.
14
Willingness to pay is higher for seating
locations closer to the field of play and closer to midfield.
The modeling also uncovers the possibility of PSL asset depreciation as PSL sales
prices have decreased significantly from the beginning to the end of the data set.
Intuitively, this makes sense as the window to purchase season tickets from the fran-
chise shrinks as the venue ages and the franchise moves closer to constructing a new
venue which would require the consumer to purchase a new PSL.
Important distinctions between PSL and STR markets are also evident through
the modeling. Most notably, prices in PSL markets are highly sensitive to team qual-
ity with market prices increasing along with improvements in on-field performance.
Alternatively, STR market prices do not respond at statistically significant levels to
changes in team quality. Given that the positive relationship between consumer
interest and team quality is uniform in the sports economics literature, this is a find-
ing of note. Additionally, market prices for PSLs are substantially higher than STRs
despite the fact that these are equivalent items when purchased on the secondary
market. It is clear that consumer interest in long-run access to NFL attendance is sub-
stantially stronger in PSL markets. Along with the behavior of market prices in
response to team quality, it raises the possibility that there are two distinct classes
of NFL franchises. Future research may look to further examine the differences in
consumer interest between PSL and STR markets.
Another intriguing question stemming from this research asks why NFL franchises
have almost uniformly decided not to restrict the sale of PSLs and STRs on the sec-
ondary market. Similar to what the New England Patriots and Green Bay Packers do,
clubs could simply require season ticket holders to return PSLs and STRs to the team if
the consumer no longer wishes to purchase their season ticket package. This would
allow the franchise to recapture unclaimed consumer surplus that is lost when the club
Salaga and Winfree 23
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
allows season ticket holders to sell their PSLs and STRs on the secondary market. On
the other hand, the current scenario where the majority of franchises sell consumers
the unrestricted rights to seats likely increases the original PSL value. An interesting
extension to this work would examine that question further.
Acknowledgments
We would like to thank Charles Brown, Rodney Fort, Thomas Peeters, Stefan Szymanski, and
two anonymous referees for helpful comments and insights.
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship,
and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of
this article.
Notes
1. Franchises utilizing PSL programs typically set the vast majority of their venue capacity
as ‘‘PSL seats,’’ which require an up-front PSL fee and are then sold as season tickets.
Traditionally, a very limited number of low-quality seats are sold as ‘‘non-PSL seats’’
available for purchase without a PSL either prior to the season or the date of the game.
2. In order for STRs to be sold in the secondary market, the issuing NFL franchise must
allow their current season ticket holders the opportunity to sell or ‘‘transfer’’ their STRs.
NFL franchises place varying restrictions on the resale of both STRs and PSLs (Reese
et al., 2004) and these team policies have been altered over the time period examined.
3. Previous work including Noll and Zimbalist (1997), Fort (2003), and Howard and Cromp-
ton (2004) has identified the sale of PSLs as a growing trend in professional sports, but
there has been a lack of empirical investigation regarding the topic.
4. Every effort is made to develop a comprehensive sample of NFL PSL and STR sales
transactions over the examination period. However, we must note that the sources of data
used do not necessarily reflect the entire secondary market.
5. Documentation of season ticket sales dominating over single-game ticket sales in the
NFL is referenced as far back as Noll (1974).
6. A distinguishing characteristic of a two-part pricing model is that quantity increases when
two-part pricing is implemented. However, nearly all NFL games are sold out regardless
of a presence of a PSL or STR. In the case where the quantity sold does not change, teams
could simply increase their ticket price to generate more revenue. So, at least at the mar-
gin, PSLs are not necessary to capture consumer surplus. However, in the NFL’s case,
teams do not share revenue generated from PSLs. Presumably, PSL revenue is not shared
to provide an incentive to build stadiums since teams can only issue a PSL when a new
24 Journal of Sports Economics 00(0)
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
stadium is built. Therefore, PSLs are a way to avoid revenue sharing as opposed to a
mechanism to capture consumer surplus.
7. Presumably, PSL revenue is not shared to encourage the construction and renovation of
stadiums, which is the only time franchises are allowed to issue a PSL. However, it is
unclear to us why teams with PSLs still charge a ticket price. There may be some formal
or informal agreement to generate some amount of stadium revenue that is shared.
8. The empirical models were estimated both with and without the Redskins and Steelers
data in order to determine whether using average ticket prices for these two franchises
would significantly impact the estimation results. The differences in the results were
negligible.
9. Other short-term and long-term measures of team quality were tested, and 3-year winning
percentage was selected because it showed the most explanatory power.
10. Level of game uncertainty variables and visiting team quality variables are not included
in these models because the consumer is purchasing access to an entire season ticket
package, representing a different purchase decision as compared to a single-game ticket,
in which the quality of the visiting team would be a significant determinant on the atten-
dance decision.
11. Various linear and nonlinear trend variables were tested and indicator variables based on
the year of sale were selected because they controlled for more variation in the dependent
variable.
12. A small number of observations for the following franchises were eliminated from the
data set: Arizona, Kansas City, New York Giants, and New York Jets. Observations from
both New York franchises were withheld because these sales represented ‘‘an option to
purchase’’ a PSL from the franchise prior to the opening of their new venue in 2010.
These sales did not represent actual secondary market PSL transactions. The Kansas City
and Arizona observations were removed because of questionable sales terms.
13. Because 2009 MSA per capita income was not available at the time of analysis, all 2009
observations use 2008 income data. This causes LN(INCOME) to be collinear with the
team indicator variables in the STR fixed effects model.
14. Our results illuminate consumer preferences for high-quality seating locations. However,
stadium configurations are such that there are fewer ‘‘good’’ seating locations as opposed
to ‘‘bad’’ seating locations. Therefore, these estimations illustrate consumer preferences
given the relative scarcity of the given seat location.
References
Coates, D. C., & Humphreys, B. R. (2007). Ticket prices, concessions and attendance at
professional sporting events. International Journal of Sport Finance, 2, 161-170.
Demmert, H. G. (1973). The economics of professional team sports. Lexington, MA: Lexing-
ton Books.
DeSerpa, A. C. (1994). To err is rational: A theory of excess demand for tickets. Managerial
and Decision Economics, 15, 511-518.
Salaga and Winfree 25
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
Drayer, J., & Martin, N. T. (2010). Establishing legitimacy in the secondary ticket market: A
case study of an NFL market. Sport Management Review, 13, 39-49.
Fizel, J. (2006). The National Football League conundrum. In J. Fizel (Ed.), Handbook of
sports economics research (pp. 170-171). Armonk, NY: M.E. Sharpe.
Fizel, J. L., & Bennett, R. W. (1989). The impact of college football telecasts in college
football attendance. Social Science Quarterly, 70, 980-988.
Fort, R. D. (2003). Sports economics. Upper Saddle River, NJ: Prentice Hall.
Fort, R. D. (2004). Inelastic sports pricing. Managerial and Decision Economics, 25, 87-94.
Howard, D. R., & Crompton, J. L. (2004). Financing sport (2nd ed.). Morgantown, WV:
Fitness Information Technology.
Krautmann, A., & Hadley, L. (2006). The product market for professional sports. In J. Fizel
(Ed.), Handbook of sports economics research (pp. 175-189). Armonk, NY: M.E. Sharpe.
Nagel, M. S., & Rascher, D. A. (2007). Washington ‘‘Redskins’’ – Disparaging term or valu-
able tradition? Legal and economic issues concerning Harjo v. Pro-football, Inc. Fordham
Intellectual Property, Media & Entertainment Law Journal, 17, 789-804.
National Football League Players Association. (2006). Collective bargaining agreement
between the NFL management council and the NFL players association. Retrieved
October 12, 2009, from www.nflpa.org/CBA/CBA_Complete.aspx
Noll, R. G. (1974). Attendance and price setting. In R. G. Noll (Ed.), Government and the
sport business (pp. 115-157). Washington, DC: Brookings Institution Press.
Noll, R. G., & Zimbalist, A. S. (1997). Build the stadium – Create the jobs! In R. G. Noll & A.
S. Zimbalist (Eds.), Sports, jobs & taxes (pp. 1-53). Washington, DC: Brookings Institu-
tion Press.
Ostfield, A. (1995). Seat license revenue in the National Football League: Shareable or not?
Seton Hall Journal of Sport Law, 5, 599-610.
Putsis, W. P., Jr., & Sen, S. K. (2000). Should NFL blackouts be banned? Applied Economics ,
32, 1495-1507.
Reese, J. T., Nagel, M. S., & Southall, R. M. (2004). National Football League ticket transfer
policies: Legal and policy issues. Journal of Legal Aspects of Sport, 14, 163-190.
Sandomir, R. (2009, July 17). A Texas-size stadium. New York Times. Retrieved October 12,
2010, from http://www.nytimes.com/2009/07/17/sports/football/17cowboys.html
Schofield, J. A. (1983). Performance and attendance at professional team sports. Journal of
Sport Behavior, 6, 196-206.
Tainsky, S. (2010). Television broadcast demand for National Football League contests. Jour-
nal of Sports Economics, 11, 629-640.
Team Marketing Report. (2009). NFL fan cost index. Team Marketing Report. Retrieved Feb-
ruary 3, 2010, from http://www.teammarketing.com/blog/index.html?article_id¼96
Vrooman, J. (2010). Dallas cowboys stadium analysis. Nashville, TN: Vanderbilt University.
Welki, A. M., & Zlatoper, T. J. (1994). US professional football: The demand for game-day
attendance. Managerial and Decision Economics, 15
, 489-495.
Welki, A. M., & Zlatoper, T. J. (1999). US professional football and game day attendance.
Atlantic Economic Journal, 27, 285-298.
26 Journal of Sports Economics 00(0)
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
White, H. (1980). A heteroskedasticity-consistent covariance matrix estimator and a direct
test for heteroskedasticity. Econometrica, 48, 817-838.
Author Biographies
Steven Salaga is an Assistant Professor of Sport Management in the College of Business at
the Florida Institute of Technology. He received his PhD from the University of Michigan.
Jason A.Winfree is an Associate Professor of Sport Management at the University of
Michigan. His work focuses on sports economics and other topics in industrial organization.
Salaga and Winfree 27
at FLORIDA INSTITUTE OF TECHNOLO on February 12, 2013jse.sagepub.comDownloaded from
