The able research assistance of Matthew Hotz is gratefully acknowledged, as is the help of Cathy Shannon, William McKinnon, Jeff Mallory and Bud Meredith, all of whom provided valuable information on the finances and prospects of PSU facilities.

0. Preliminaries

This report is concerned with the economic impact of (1) the construction phase of the proposed Beaver Stadium expansion; (2) Beaver Stadium itself; and (3) the Bryce Jordan Center on the State College Area economy.

This economic impact arises from use of the facilities, and therefore expenditure at the facilities by both residents of the State College Area and by those visiting from outside the area. We concentrate on the latter, because visitors' expenditures are new injections of resources in the local economy, whereas locals' expenditures are to a very large extent replacing other kinds of local expenditure and therefore do not constitute a net new economic impact.⁽¹⁾ The report tries to estimate the expenditure within the State College area which is generated, directly or indirectly, by the facility visitors.

The economic impact is composed of three parts. The first is visitor expenditure at the facility itself. The second is visitor expenditure within State College, but not at the facility. Much of the report is an attempt to measure these two numbers. The third type of impact is the indirect impact of local re-spending of this visitors' expenditure. State College merchants and employees who receive this money spend some part of it within the area, and so it gets respent again and yet again (in ever smaller increments) This is commonly referred to as the multiplier process, and its extent is captured in a number called the multiplier ⁽²⁾ which is multiplied by the direct expenditure discussed above to obtain a total (=direct +indirect) expenditure. Previous studies have used varying multipliers for the State College region: The recent report from the Institute of Policy Research and Evaluation⁽³⁾ assumed a value of 1.5, although this was not consistently used throughout. Anderson, et al's⁽⁴⁾ prospectus for the Bryce Jordan Center deduced from theisimulations with the IMPLAN model that the multiplier was approximately 1.6. The 1987 report⁽⁵⁾ assessing the impact of Beaver Stadium calculated that the multiplier was 1.97, using the minimum requirements estimator. Only this latter estimator permits replication for the purposes of this study.

The minimum requirements estimator calculates the multiplier by dividing the extant work force of a city into that which is needed for local production and the remainder, which is assumed to produce export goods and services. The local-serving employment is calculated for each sector as the minimum percentage of the labor force employed in that sector for cities in the same size class. The percentage of the work force devoted to local demand is then the aggregate of these percentages across sectors. Then a linear regression relating this percentage to metropolitan area population can be estimated. The 1987 report presented such a regression relationship which estimated that a metropolitan area with a population of 120,000 (the State College area's 1987 population) would have 49.7% of its workforce devoted to local production, which translates into a multiplier of 1.97 (see footnote 2). In order to update the multiplier for this study, one could either use the 1987 formula using the latest estimates of State College's population, or one could go further and attempt to recalculate the minimum requirements for the variously sized cities. The former method yields a multiplier estimate of 1.99, given the estimated rise in the State College area's population to 132,000. (Note that throughout this report, rounded estimates are used, for ease of presentation.) We also attempted a complete recalculation of the minimal requirements using the most recent Bureau of Labor Statistics data on the sectoral composition of metropolitan labor forces, but the censoring of this data is fairly severe, due to privacy constraints. Given the available information we arrive at a multiplier of approximately 2.3, but in the interests of reasonably conservative inference the multiplier to be used in the remainder of this report will be the 1.99 figure mentioned above. In any case the use of the multiplier is transparent enough that reader will be able to infer economic impacts with or without the multiplier process.

1. The Economic Impact of the Construction Expenditure for Beaver Stadium Addition

The expansion and remodeling of Beaver Stadium are estimated to cost $80 million. This section of the report is an estimate of the impact on the local economy of this $80 million expenditure.

Estimates for the location of the direct expenditure of this money have been provided by Jeff Mallory of PSU's Office of the Physical Plant. The estimates contained therein are provided and summarized in Table 1. It can be seen there that local (Centre County) expenditures by the project will total $14.4 million. Using the 1.99 multiplier it is straightforward to calculate the direct plus indirect gross expenditure in Centre County as $28.66 million

In 1981, Robert Ball⁽⁶⁾ provided estimates for the number of jobs created per thousand dollars of expenditure. We inflate the $1000 from his 1981 study to 1998 dollars (using a general version of the Consumer Price Index for the Northeastern US) and calculate the number of job years created locally by the $14.4 million expenditure.⁽⁷⁾ We find that roughly 70 local job-years are directly created out of this new expenditure. Given the multiplier factor of 1.99, we find that 139 new job-years are created out of this expenditure⁽⁸⁾.

This breakdown of construction expenditures included not only local but also regional expenditure. This is expenditure in metropolitan areas within the central part of the Commonwealth and mainly includes expenditure in the Altoona and Harrisburg metropolitan areas. We can therefore attempt to analyze the spillover impact of such expenditure. To the extent that State College is a market destination for Altoona and (probably to a lesser extent) Harrisburg consumers, autonomous expenditure in those regions will have some impact on the State College economy. A simple econometric model (which also accounts for the multiplier effect within State College) was developed to assess the size of such spillovers over the past two decades⁽⁹⁾, and finds that for every increase in 100 jobs in those two regions (taken as a whole) creates approximately 9 jobs in Centre County. The regional expenditure is given as $15.0 million. Following the above calculations, this indicates that in the region described there will be approximately 73 jobs created directly and (assuming a 1.99 multiplier for this region) 145 jobs including the induced effects. This will account for about 13 extra new jobs in the Centre County area. The total employment impact is therefore estimated to be about 152 new jobs.

2. The Impact of Beaver Stadium Expenditure on the Local Economy

This section attempts to update the September 1987 report The Economic Impact of Penn State Football in the State College Region. The update is simply based on the recognition of changing prices and increased seating capacity of the Stadium since that report was issued.

How Many Visitors in 1998?

Seating capacity is currently around 93,000 seats since the completion of the North End addition. Of those, about 20,000 of those are allocated to students. Of the remaining seats, about 64,000 are distributed to season ticket holders both within and outside the State College region, and to other purposes. The remaining 9,000 seats are distributed to other Penn State students (e.g. from other campuses in the PSU system), to visiting universities and to others. Of the 64,000 season ticket holders, zip code data on the location of season ticket holders provided by the Penn State Athletic Ticket Office ⁽¹⁰⁾ approximately 10,500 come from zip codes beginning with 168 so that (by that measure) approximately 83% or 53500 come from outside the area.⁽¹¹⁾ Applying this same percentage to the unaccounted 9000 seats yields about 7500 more

visitors for a total per game influx of about 61,000(=53500+7500) visitors.⁽¹²⁾

The 1987 report noted the almost ritualistic behavior to which these fans adhere. For this reason we feel reasonably confident that the spending patterns uncovered by the survey team in Erickson et al will, with appropriate adjustments, provide a reasonable estimate of the fans' current economic impact on the community. On that account, we provide as an initial estimate of that expenditure, Table 2.

How Much Spending?

The first column of Table 2 is the category of expenditure in State College for visitors from greater than 25 miles as developed in the 1987 report. The second column reports the percentage of respondents who had expenditure in this category in that report. The third column represents our attempt to categorize these expenditures according to categories of expenditures in the Consumer Price Index. For all such categorizations these indexes are for Northeastern Urban Consumers and represent the CPI index for 1998 divided by the same categorical index for 1986. In those cases where specific indexes are not available we used the "all items index." For all categories of expenditure the commodity category and inflation factor are given in Column (B), and the 1986 per (out-of-town) attendee expenditure as calculated in the 1987 report is given in Column (C). Multiplying this latter expenditure by the inflation factor gives 1998 per-visitor expenditure. The out-of-town attendance has been given as 61000, so column (E) provides the approximate aggregate expenditure on game weekends by visiting attendees by multiplying column (D) by 61000. Finally, the season expenditure is given by multiplying the per game expenditure by 6, the typical number of home games in a (Big 10) football season⁽¹³⁾.

The row marked 'total' gives the corresponding total expenditures by adding up expenditures across categories. It can be seen that the typical football game creates almost $5 million in total expenditure by nonlocal fans, and that over the season nearly $30 million is injected into the local economy. Using the multiplier of 1.99 the direct plus indirect expenditures bring over $59 million into the local economy.

Additional Impact of the Beaver Stadium Expansion

When the south end and luxury box expansion is completed, it will add 10,000 seats to the capacity of Beaver Stadium. A thousand seats will be reallocated to students, so that a net increase of 9,000 new season tickets will be sold. If 83% of the seats are sold to visitors from outside the State College area, then approximately 7500 people will add their expenditures into the injections already calculated. Note from Table 2 that the total per visitor impact can be approximated as $81.16. Multiplying this by 7500 yields an increase in game-weekend expenditure of approximately $608,700. Over the course of a six game season direct expenditures will increase by over $3.65 million. With the multiplier the direct plus indirect impact can be estimated to be about $7.26 million in additional expenditure⁽¹⁴⁾.

Further Projections

The proposed renovation of Beaver Stadium not only includes an expansion of its seating capacity, but also various other improvements to scoreboards, restroom facilities, walkways, etc. which promise to raise the quality of the game experience for attendees. On that account it seems certain that ticket prices will be raised over the next several years. Current projections call for a $6 increase staggered over the next few years, over and above any increase due to inflation. It is straightforward to calculate the revenue potential from this increase. By the 2001 season, there will be approximately 68,500 (=61000+7500) fans from outside the State College area spending $87.16 (81.16+6.00) at each football game for a total of $5,970,000 in direct expenditure in State College each football weekend. Over the course of a six-game season, this yields $35,822,000 in direct revenue, and multiplying this by the 1.99 multiplier yields a total economic impact of $71,287,000.

One further factor contributes to the upside potential of the stadium, which is the recent liberalization of NCAA rules regarding the number of available games in a season. It will be easier in years to come for teams schedule a 12th game in the regular season, making it more likely that PSU can schedule a seventh home game. (This was a regular feature of the PSU schedule before joining the Big 10.) During those seasons when such a seventh home game is indeed scheduled this can add (assuming expanded seating and the above ticket price increase) another $11,881,000 (=5,970,000*1.99) to the local impact for a total potential impact of $83,168,000.

3. The Economic Impact of the Bryce Jordan Center

Table 3 gives the sources and amounts of local economic impact that are derived from the Bryce Jordan Center. In what follows the derivation of these figures is discussed. Roughly speaking, the BJC holds events of two types: PSU-sponsored events, and externally sponsored events. In the former category are athletic contests (mainly Men's and Women's Basketball games), graduation ceremonies, "tailgreat" gatherings, and a few other events. In the latter category are musical concerts and other entertainment events which are brought to the BJC by a professional promotion company.

The economic impacts are of two types: first, the university (as proprietor of the BJC) receives revenue directly from the use of the facility and its services. In the case of PSU- sponsored events the revenues are direct. The university charges admission to its basketball games receives the flow of this revenue along with the concessions and other expenditures at the games. In the case of entertainment events, Penn State rents the facility and a number of associated services (such as labor, equipment rental, etc.) to the promoter of the event and also receives a share of the ticket and concession revenue.

The second type of economic impact is the incremental expenditure within the community associated with the events held at the BJC, which arises from patterns of behavior not fundamentally different from that exhibited in association with Beaver Stadium football games. The assumption in the Beaver Stadium analysis is followed here as well: that local residents' spending associated with BJC events merely replaces expenditures that would otherwise take place some other time, and hence should not be counted among the economic impacts.

Direct Impacts of Athletic Events

We discuss first the revenue generated by PSU as a result of holding athletic events at the BJC. The vast majority of such revenue is generated by Men's and Women's Basketball games and the analysis here is concentrated on that.⁽¹⁵⁾ In some respects the estimation of the direct impacts is fairly straightforward. The university collects revenue from ticket sales, concessions and parking at the games and spends it either non-locally or locally (on wages or on locally produced goods and services) to create the induced effect (about which more later). We take the same line of reasoning as in part 2 of this report, that only those tickets that are sold outside the State College area (i.e. the 168 zip codes) constitute injections into the local economy.

During the 1997-98 basketball season, there were approximately 143,000 tickets sold for Men's basketball games held at the BJC, which generated approximately $1.46 million in revenue. Of this revenue, the substantial majority, $1.05 million came from season ticket sales to non-students. Ticket office zip code records indicate that approximately 50% of these tickets were purchased by fans from outside the State College area, and so it is sensible to correspondingly assume that half of the season ticket revenue similarly comes from outside the area⁽¹⁶⁾. The remaining $400,000 in revenue arises from a variety of sources: student season tickets, individual ticket sales (to both students and non-students), group ticket sales for individual games, and the like. Excepting student sales and some trivial sources of revenues, there is approximately $320,000 in individual game ticket revenue sold to the general public. If it similarly assumed that half of this revenue arises from sales outside the region, then approximately $685,000(=0.5*($1.05million +$320,000)) in revenue from Men's Basketball is collected by Penn State from outside the State College area and thereby injected into the State College economy from outside the area.

For Women's Basketball 68,000 tickets were sold, and $275,000 in revenue collected. Of this, approximately $208,000 in revenue was collected from season ticket holders, and according to the zip code counts approximately 29% of these were from outside the State College area. Applying the same procedure used in the Men's Basketball analysis it is estimated that $53,000 in individual game ticket revenue was collected from the general public. Applying the 29% to both season and individual game sales yields revenue to Penn State of $75,700 (=(.29)*(208,000+53000), rounded)) from sales to those outside the area.

To these figures must be added the concession and parking revenue. Concession revenue is obtained from the BJC revenue summary and total approximately $160,000 arises from men's and $73,000 for women's games. Multiplying these figures by (.50) and (.29) respectively, yields the injected concession revenue provided in the Table. Similarly the revenue obtained from event parking at Men's Basketball totals approximately $58,000 and from women's approximately $23,000. Applying the same percentages as before yields the table entries.

B. Direct Impacts from Entertainment Events

During the 1997-98 fiscal year there were 28 entertainment events at the BJC which had yielded a total attendance of 242,000. It is necessary first to determine the number of tickets sold to those outside the State College Area. For these events, the phone and mail orders that are received by the BJC box office have a known location, and these are reported here in Table 4. As can be seen, a bit more than 42% of phone ticket orders come from the State College Area. However this does not fully account for all of the tickets sold, since a number of walkup tickets are sold at each event. It is assumed that the same percentage of walkup ticket-buyers are State College area (168 zip codes) residents.

The revenue generated by these events given in the BJC Revenue Summary FY 1997-98. Revenue is generated via the promoter rent, box office and merchandise shares, concession sales⁽¹⁷⁾, and charges to the promoter for catering, parking, advertising, equipment and labor⁽¹⁸⁾. This totaled $1.93 million from promoter payments and $400,000 in concessions over the fiscal year. As it is assumed that 57% of the ticket revenue comes from outside the area, the new injection is calculated to be 57% of these two revenue streams-- that is, $1.10 million and $228,000.

Revenue is also received from patron parking fees. The parking income statement displays $196,000 in revenue from entertainment events. Applying the 57% figure yields an external injection of parking income of $112,000

Associated Visitor Spending

The indirect revenue is the spending that entertainment and athletic event patrons spend on goods and services in the State College area but not at the BJC itself. The 1987 football report gave a flavor of the kind of spending that is meant here: hotels, restaurants, souvenirs, etc. But it would be foolish to think that the average attendee of a basketball game or rock concert spends in the same way that a football fan does. Football games are on Saturday, and usually in the afternoon; concerts and basketball games are usually (though certainly not always) at night, and often on a weeknight. This then does not lend itself to a leisurely couple of days vacation in the area, and the resulting spending patterns that are seen on football weekends.

Nevertheless, some spending of this nature probably does occur, especially to the extent that fans travel to State College from substantial distances. While the 1987 football report does not attempt to correlate fan spending with distance traveled, it is likely that the wide area from which fans travel to see Penn State football contributes to the high level of spending they engage in.

It can therefore move the analysis forward if we make an assumption that BJC attendees who come from sufficiently distant locations have spending patterns equivalent to that of the average football fan.

What is a sufficiently distant location? We assume that any visitor traveling to the BJC from outside the following metropolitan areas is so considered : State College, Harrisburg, Williamsport, Altoona, Johnstown and Dubois. This would seem to lead to a conservative calculation of the indirect impacts since as table 3 displays (for entertainment events), these areas constitute the bulk of the ticket sales.

For the BJC entertainment events note that approximately 5,300 ticket orders (out of about 55,000) came from areas outside those noted above (these "outside areas" are noted with asterisks in Table 4). Assuming 2.5 tickets per order this means that 13,250 people came to State College from these outlying areas over the course of the year.⁽¹⁹⁾ If each of these had spending patterns like those of football fans in table 1 (i.e., 28% spent the night in a State College hotel or motel, etc.) then (netting out the Table 1 expenditure on Stadium expenditure) each would spend approximately $49 in the area, for a total expenditure of about $649,000.

To perform a similar calculation for Men's Basketball attendance, the masterfile listing of zip codes for season ticket requests is used. There, we find that approximately 20% of season ticket orders come from outside the multi-county area of central PA described above. Applying this percentage to the total of tickets sold to non-students (and a few other small groups of ticket buyers) it is estimated that 21,000 people come to Penn State Men's Basketball contests from outside the central PA area. Assuming the spending patterns previously discussed, we find that $1.04 million is spent by these fans in the State College community⁽²⁰⁾.

For the Women's team, the zip code file allows us to estimate that 8% of fans travel to games from outside this broadly defined area. The similar calculation implies that about 3800 fans travel to games and spend about $188,000.

Totals

As can be seen from Table 3, Men's Basketball has a direct economic impact on the State College economy is over $1.8 million, while Women's Basketball adds almost a quarter of a million dollars. Entertainment events generate more than $2 million. The total direct impact is therefore a bit more than $4 million. Using the 1.99 multiplier as before yields a total impact of over $8 million from events at the BJC.⁽²¹⁾

5. Conclusions

States and metropolitan areas have increasingly struggled with the question of whether they should participate in the financing of sports facility construction, in particular for professional sports. The overall evidence indicates that they should not, and the reason they should not is that the observable economic benefits of sports facilities are minuscule, especially when considered relative to the aggregate economic activity within a city.⁽²²⁾ The lack of impact in part derives from the fact that demand for attendance at such facilities largely arises from within the city, merely displacing other intracity expenditure. However this debate also makes clear that facility construction can have a substantial impact at the micro-level. In the facility's own neighborhood, there can be business formation and increased economic activity from the increased expenditure precisely because the demanders arrive from outside the neighborhood.⁽²³⁾

And so it is with the sports facilities on the Penn State campus. The estimable economic impacts of the Stadium and BJC are the product of three factors: the number of events, the size of the attendance, and the extent to which that attendance is drawn from outside the State College area. We have attempted to assess these factors using the data which are available, and using conservative assumptions to analyze them. While the assumptions are of course subject to discussion, the overall conclusion is not at risk: the results indicate a substantial economic impact. Football games are few in number, but they draw huge crowds, a good percentage of which come from outside the area. Hence the economic impact of a football season is much larger than that for the Bryce Jordan center events. But together their economic impacts are estimated to total more than $60 million per year, and are therefore a substantial contributor to the local economy.

Part I

"Proposed Beaver Stadium Expansion" Office of Physical Plant

Part II

"Beaver Stadium Expansion and Renovation Project Overview" (web document)

"The Economic Impact of Penn State Football in the State College Region" (Center for Regional Business Analysis) September 1987

Part III

"Bryce Jordan Center Revenue Summary, FY 97-98" (BJC Ticket Office)

"Intercollegiate Athletics Men's Basketball Ticket Accountability 1997-1998 Season"

"Intercollegiate Athletics Women's Basketball Ticket Accountability 1997-1998 Season"

"Bryce Jordan Center Masterfile Listing-- Men's Basketball" (Zip code files)

"Bryce Jordan Center Masterfile Listing-- Women's Basketball" (Zip code files)

"Bryce Jordan Center 1997/98 Event Parking Income/Expense Statement"

"Ticket Transactions for Fiscal Year 1997/1998" (Locations of ticket orders)

Footnotes:

1. Of course the existence of these facilities may cause locals to spend more of their income locally, rather than seek entertainment out of town, but the extent of this is hard to measure and so in order to provide a conservative estimate of the impacts, it is ignored. But these benefits should be kept in mind, as they are often lost in the debate over the benefit-cost analysis of sports facilities. See for example Chapter 4 of Quirk and Fort, Pay Dirt, The Business of Professional Team Sports (Princeton University Press, 1992)

2. Technically the multiplier is 1/ (1-proportion of income spent locally). In a small metropolitan area like State College, the multiplier is also going to be small as a smaller proportion of income can be spent on local goods and services. For an analytical treatment of these issues, see Brown, Coulson and Engle "On the Determination of Regional Base and Regional Base Multipliers" Regional Science and Urban Economics, November 1995.

3. "The Expenditure Impacts of Penn State University Park on the Centre Region" Institute for Policy Research and Evaluation Working Paper 99-01.

4. Anderson et al " The Economic Impact of the Pennsylvania State University Academic/Athletic Convocation and Events Center" March 1991.

5. Erickson et al, "The Economic Impact of Penn State Football in the State College Region" September 1987. Referred to hereafter as "the 1987 report".

6. Robert Ball "Employment Created by Construction Expenditures" Monthly Labor Review, December 1981

7. This brings forth the question of whether productivity changes have occurred in the construction industry and whether Ball's conversion factors ought to be altered thereby. Both Paul Allen ("Why Construction Industry Productivity is Declining" Review of Economics and Statistics, November 1985) and more recently Andrew Bernard and Charles Jones ("Productivity across Industries and Countries" Review of Economics and Statistics, February 1996) have found that productivity is declining in the construction industry. Paradoxically this indicates that a given amount of expenditures might entail a larger number of jobs created. We opt to not try and adjust the conversion factors in the interest of being conservative in our estimates.

8. This surely understates the amount of job creation arising from the project, for it assumes that the only multiplier impact is via the local spending of construction wages. In point of fact much of the local impact takes place through spending on material and equipment, which in turn has indirect impacts. Thus the expenditure figures are more to the point than the employment figures.

9. The model regresses State College employment on its own history and the history of Altoona and Harrisburg employment (summed together). The coefficient on the latter indicates the shortrun impact of Altoona/Harrisburg growth on State College.

10. See the appendix for a list of Penn State documents consulted in this report.

11. This is roughly the same percentage of "out-of-towners" as found in the 1987 report, and so provides some confidence that the 25 mile radius used in that report can be roughly equated with the 168 zip area used here.

12. This assumption was also used in the 1987 report.

13. We assume throughout this section that the State College economy has the ability to accommodate the larger number of visitors engendered by the North end expansion. Certainly the number of hotel rooms has increased substantially in the intervening years, both in quality and in quantity. Indeed one might question whether the inflation factor is too low for hotel rooms in State College given this increased quality, but inasmuch as that inflation factor was already the highest in the Table, no adjustment of this kind was attempted.

14. The revenue generated from luxury boxes, while bringing substantial revenue to the University, will not, for the first several years of its existence, bring any new revenue impacts to State College. This is because this revenue will be used to pay for the construction of the expansion itself, so that the impact has already been taken into account in the previous section. Therefore for the first several years after the expansion is in place, the revenue impacts will be limited to the ordinary expenditure described above.

15. The economic impact of graduation ceremonies are not counted here. Nor are "tailgreats" since they are associated exclusively with home football games and hence their economic impact is accounted for above. Occasionally the mis-named "minor sports" hold events at the BJC. The number of visitors to these contests is small and probably heavily weighted toward locals; ticket counts by area are not available, and they are not considered in the analysis to come. Externally promoted athletic events, such as NBA exhibitions, or NCAA or Big 10 championships, are classified with the entertainment events.

16. There are several different ticket price classifications for athletic events at the BJC so that if the mix of tickets differs from area to area there is not necessarily a direct correspondence between number of tickets and amount of revenue. However it is necessary to assume so here.

17. The concession sales accrue to Penn State Food Services, which rebates a share to the BJC. The total sales are included in the economic impact. This is different from the box office and merchandise sales, which accrue to the promoter, who then provides a share to the BJC.

18. There was a further revenue category for deferred maintenance but as the direct economic impact seems rather minimal, it is omitted from these calculations.

19. The assumption of 2.5 tickets per phone order is an approximation and the only evidence available suggests that it is a conservative approximation. It is possible to estimate the number of tickets per season ticket order for football and basketball and this is greater than 2.5 for all three major sports. (Women's Basketball is the lowest at 2.6 tickets per order.)

20. At this point it is necessary to confront the issue of "no-shows". If tickets are purchased but not used, then the associated visitor spending does not occur either. (This is not an issue for concession and parking calculations since actual revenues are available for those categories, nor does it seem likely to be an issue for concert events either). However, the recorded attendance figures for basketball games are very close in size to the number of tickets sold, with the exception of one or two exhibition games at the start of the season. Since they are different for that limited number of games, one must infer that the attendance figures are accurate for the remaining bulk of games and that we can use the ticket sales data accordingly.

21. The entire analysis has assumed that local expenditure does not contribute anything to the BJC's economic impact. This conservative assumption is true only to the extent that local ticket, concession and parking expenditure merely replaces other local expenditure. However the BJC offers a type of entertainment which is not available close by and therefore its revenue is not merely replacing other expenditure. Therefore one might legitimately count part of local resident's expenditure as part of the impact. One should not, however, count any multiplier impact ( since the actual effect here is an increase in the multiplier itself, not an increase in the revenue being multiplied). As an upper bound, one might suppose that all of the entertainment dollars spent by locals was newly generated. In this extreme case the estimated impact would rise by about $2.2 million, to well over $10 million.

22. See especially the papers in Noll and Zimbalist, Sports, Jobs and Taxes, Brookings Institution Press, 1997

23. A good example of this is Austrian and Rosentraub "Cleveland's Gateway to the Future" in Sports, Jobs, and Taxes cited above.