This PDF is a selection from a published volume from the National Bureau of Economic Research
Volume Title: The Governance of Not-for-Profit Organizations Volume Author/Editor: Edward L. Glaeser, editor Volume Publisher: University of Chicago Press Volume ISBN: 0-226-29785-3 Volume URL: http://www.nber.org/books/glae03-1 Conference Date: January 17-19, 2002 Publication Date: January 2003
Title: Does Governance Matter? The Case of Art Museums Author: Sharon Oster, William N. Goetzmann URL: http://www.nber.org/chapters/c9966
2 Does Governance Matter? The Case of Art Museums Sharon Oster and William N. Goetzmann
2.1 Introduction Art museums provide a classic example of organizations operating with multiple objectives. On the one hand, many American museums take as their central function the education of the populace. At the same time, there is a long tradition in museum management of conservation and appeal to the narrower elite. In the past decade, the balance between these objectives seems to have tilted in favor of the broader populace. In writing of this change, one museum activist, Kenneth Hudson, has argued: “The most fundamental change that has aﬀected museums is the now almost universal conviction that they exist in order to serve the public” (Kotler and Kotler 2001, 171). Sociologists have explored this tension at some length. D’Harnoncourt et al. (1991), for example, describe the movement of art museums from secluded temples of culture to the present-day more public institutions. Grana (1971) similarly contrasts patron-oriented museums, focused on “men of leisure from the upper classes,” with publicoriented ones. This paper uses cross-sectional and time-series data on U.S. museum finances and operating characteristics to explore the eﬀect of governance structure on performance. We are particularly interested in whether the ownership structure of a museum influences the balance it strikes among competing constituents. Increasingly, economists have come to appreciate Sharon Oster is the Frederic D. Wolfe Professor of Management and Entrepreneurship and faculty director of the Partnership on Nonprofit Ventures at the Yale School of Management. William N. Goetzmann is the Edwin J. Beinecke Professor of Finance and Management Studies and director of the International Center for Finance at the Yale School of Management. We thank Cathy Shu for collecting the data. We thank the numerous museums who shared their data with us. We thank workshop participants for their suggestions.
Sharon Oster and William N. Goetzmann
the role played by governance structures on decision making in organizations, and the diﬀerentiated structure of the industry makes museums an excellent case study. 2.2 The Role of Museums We begin our discussion by considering the objective function of the typical museum. In the literature, there are three oft-cited museum goals: art preservation, education of the populace, and the providing of a social signal for the elite of a community. The first two of these goals appear frequently in the mission statements of museums. The mission statement of the Portland Art Museum in Oregon is typical: “The mission of the Portland Art Museum is to serve the public by providing access to art of enduring quality, by educating a diverse audience about art and by collecting and presenting a wide range of art for the enrichment of present and future generations.” The opening lines of the mission statement of the Boston Museum of Fine Arts strike a similar theme: “The Museum of Fine Arts houses and preserves preeminent collections and aspires to serve a wide variety of people through direct encounters with works of art.”1 The interest in both art preservation and education for the public are clear. The role of museums in reinforcing a social elite within a city is less often articulated in mission statements. Yet, until well into the twentieth century, most American museums depended on private philanthropic dollars for their support (Anheier and Toepler 1998, 235). Indeed, wealthy industrialists, to whom Dimaggio refers as “cultural capitalists,” founded many of our most well-known museums (1986). Dimaggio describes in some detail the way that these industrialists, in cities like Boston, used art institutions to build cultural boundaries separating themselves from the rest of society. As Temin suggests, displaying one’s art validates both a patron’s possessions and his or her position in society (1991). Consequently, one might expect that the more aﬄuent the society, the greater the need to signal taste through support and display of the arts. The growth of art museums was thus based not only on an aesthetic tradition in American society, but upon a philanthropic one. John Ingham (1997) and Ruth Krueger Meyer and Madeleine Fidell Beaufort (1997), in an exhibition catalogue to a major exhibition, Collection in the Gilded Age: Art and Patronage in Pittsburgh: 1890–1910, describe the art-collecting and philanthropic activities of Gilded-Age Pittsburgh through the lenses of class and society. Wealthy Pittsburgh families vied with each other to create spectacular collections of European art and also, in many cases, made gifts of these collections to the public. Andrew Carnegie, the city’s (and perhaps the nation’s) wealthiest citizen at the time, led by example in do1. See the Web mission statement at http://126.96.36.199/mission.
Does Governance Matter?
Fig. 2.1 The distribution of founding dates for museums in the 1989 and 1999 AAMD sample Sample: 1989 and 1999 AAMD with reported founding information Note: Founding dates were collected from annual AAMD directories, as available.
nating much of his wealth in order to improve the access of Pittsburgh’s citizens to higher arts and education (Ingham; Meyer and Beaufort). Subsequent gifts by leading Pittsburgh citizens enriched the artistic horizon of the nation as a whole. For example, Andrew Mellon’s collection became the core of the National Gallery of Art, and Henry Clay Frick founded the Frick Collection of New York. While the Gilded Age was an important period for museum founding and support, patronage of the arts through museum foundation has continued vigorously since. Figure 2.1 is a chart of museum-founding dates from the sample we study in this paper. It suggests that the most active periods for museum founding appear to have been the two decades preceding and the two decades following the Second World War. In fact, this probably understates the contributions of the most recent era. The figure shows a tailing-oﬀ at the end of the sample period that is most likely due to younger museums’ not reporting statistical information to the Association of Art Museum Directors (AAMD) as commonly as more established institutions. Not only was the “birth process” of museums sustained through the last century, but the social context of arts patronage has also continued to be an important factor in museum management. Museums today, as in the past, rely upon gifts for collection development and operations support, and wealthy donors and founders remain key constituents of American art museums. The continuation of the philanthropic tradition— founded on the Gilded-Age sense of civic duty and, to some extent, main-
Sharon Oster and William N. Goetzmann
tenance of social position through public giving—is an important economic foundation for art museums. Indeed, in this paper, we test the extent to which urban concentrations of wealth are related to institutional reliance upon gifts and donations. Consider now the role of governance structure in determining how museums pursue their varied objectives and balance the interests of their constituents. Approximately one-third of the art museums in the United States are public institutions. These public museums were most typically founded with service to the public in mind and are likely to emphasize public attendance as an objective. The remaining two-thirds of American museums are overwhelmingly nonprofit, but within this pool there are institutional diﬀerences, such as between university-based museums and free-standing nonprofits. University art museums, which emerged largely in the nineteenth century, were principally intended to serve the students and academic staﬀs of their own institutions (Boylan 1999). While many university museums have clearly broadened their reach to serve the general public, one might well expect some residual focus on the less popular end of the art spectrum and on curatorial and educational functions as opposed to mass appeal. Thus, we hypothesize that public museums will service the general public the most and university museums the least as they go about their respective businesses. In pursuing these three objectives, museums have a number of instruments available. To the extent that public museums emphasize public access, one would expect them to maintain low prices, focus collection eﬀorts on broadly accessible art and programs, and emphasize more popular exhibitions. University-based museums would be expected to focus on more sophisticated art and programs and be less concerned with keeping admission prices low for the general public, although free student access might well be important. Free-standing nonprofits, operating without other support, might be expected to charge higher prices and pay more attention to the interests of elite donors. Unfortunately, it is diﬃcult to gather data directly on many of these strategic variables. Locating pricing data is, for example, quite problematic. Many museums characterize admissions fees as “suggestions,” where the suggestion carries varying levels of force at diﬀerent museums. Hence, while the broad-brush data do support our hypotheses in that low or zero price levels are correlated with public ownership, it is hard to go much further simply looking at this variable. Assessing the collection eﬀorts of museums on the spectrum of popular versus more esoteric is also diﬃcult, although we have some relevant evidence in section 2.2.3, where we examine the special exhibits of the various museums. Two elements of museums operations, space utilization and financing, aﬀord some perspective on the objective functions of museums. Some measure of the emphasis that a museum places on the various elements of its
Does Governance Matter?
mission may be revealed by the proportion of space it gives to education versus exhibitions, for example. The structure of a museum’s financing may also aﬀect the way it pursues various objectives. Finally, we use the attendance levels at museums with diﬀerent ownership structures as an index of how vigorously these museums are pursuing public education and entertainment over their alternative goals. 2.2.1 Space and Money The empirical work described in this section of the paper is based on 1999 data collected by the AAMD, the principal art museum membership organization, consisting of just over 200 museums located in the United States and Canada. The AAMD conducts annual surveys of its members, covering a wide range of information about finances, operations, and museum collections. While the survey data generally are not publicly available, we were given access to the data for 1989 and 1999. For the analysis of space utilization and financing in this section of the paper we have used the 1999 data. In a later analysis of attendance, we use both survey years. In the full sample, there are 148 U.S. museums in 1989 and 140 in 1999 with substantial institutional overlap between the two years, although many of the museums have at least some missing data. The museums surveyed are quite diverse, ranging in size, for example, from the Metropolitan Museum of Art in New York, with 1,835 full-time employees in 1999, to the California State University Art Museum, with only 4 full-time employees. There is a similarly large range in the attendance figures. The National Gallery of Art in Washington, D.C. and the Metropolitan Museum of Art in New York both attract more than 5 million annual visitors, while the Yale University Art Gallery has a more modest 50,000. The summary statistics on the sample used in this paper are given in table 2.1. Table 2.1
Summary of Variables Full Sample
Collection expenditures Attendance Type of collection Survey Modern American Governance College Public Other nonprofit Endowment Observations
72% 8% 10% 19% 26% 55% 190
76% 6% 8% 16% 24% 60% $46,400,000 166
76 Table 2.2
Sharon Oster and William N. Goetzmann Museum Space Utilization, 1999 Association of Art Museum Directors Survey Statistically Diﬀerent?
Exhibition space/all space Public museums Nonprofits, not university University nonprofit Educational space/all space Public museums Nonprofit, not university University nonprofit Museum store space/all space Public Nonprofit/not university University/nonprofit
.331 .342 .345
No No No
.043 .074 .065
Yes No No
.018 .020 .016
No No No
From Nonprofit Not University
From University Nonprofit
The first question we explore using the AAMD data is the way in which diﬀerent museums use their space. The survey itself distinguishes a number of space categories. For this analysis, we have focused on three: space for exhibitions, space designated for educational use, and museum storage space. Our particular question is whether university-based museums have more educational space and less storage space than their public or general nonprofit peers. As we see from table 2.2, just over one-third of the space for the museums in our sample is used in permanent exhibition space, while a more modest area is used for either education or the museum store. There is no diﬀerence by governance type either in exhibition space or in storage space. Simple regressions holding overall museum size and age constant confirm the results of table 2.2, revealing no influence from governance. The data do suggest that nonprofit museums are devoting significantly more space to educational uses than are the public museums. In table 2.3 we compare revenue sources for the three museum types. The four major revenue streams of museums are considered: gross earned revenues, which include admission fees (suggested and otherwise), exhibition fees, museum store sales, and rentals; private philanthropic support, including corporate, individual, and foundation; government support; and finally, endowment support. Considerable diﬀerences in the funding patterns of museums by governance types are clearly revealed in table 2.3. Nonprofit, nonuniversity museums are most dependent on earned income and private support. Public museums, not surprisingly, depend principally on public support. University museums, with access to university support, are less dependent on any
Does Governance Matter? Table 2.3
Revenue Shares by Governance Structure, 1999 Association of Art Museum Directors Survey Statistically Diﬀerent?
Gross earned revenues/all revenues Public Nonprofit, not university University Private support/all revenues Public Nonprofit, not university University Government support/all revenues Public Nonprofit, not university University Endowment support/all revenues Public Nonprofit, not university University
From Both Combined
.20 .25 .11
No Yes Yes
.21 .33 .21
Yes Yes Yes
.41 .12 .08
Yes Yes Yes
.08 .19 .12
Yes Yes No
From Nonprofit Not University
From University Nonprofit
of the three constituent-based revenue sources than are public or general nonprofit museums. We explore some of the consequences of these diﬀerent financing patterns in the next two sections of the paper as we look at museum attendance and special exhibitions. 2.2.2 Attendance Attendance levels are one of the traditional output measures used by many museums. We now consider how attendance may be influenced by governance. While governance is expected to influence the aggressiveness with which museums pursue audiences, characteristics of the collection itself likely aﬀect its inherent attractiveness to the public. Finally, since museums deliver their output on site, we expect the city characteristics to help determine demand. Here we ask: Are museums like Wal-Mart, where all that really matters for attracting customers is the organization’s location? Or will a museum attract its own audience despite location-specific features? Before we turn to the econometrics, the raw data suggest something of the governance-attendance relationship. Consider the ratio of attendance to museum exhibition space as one (admittedly crude) measure of the “productivity” of a museum. By this measure, university-based museums are heavily overrepresented in the list of the twenty least productive museums.
Sharon Oster and William N. Goetzmann
Thirty-five percent of the museums on this list are university aﬃliates, as compared to a population of 23 percent. Among the twenty most spaceproductive museums, there is only one university aﬃliate. Similarly, public museums are overrepresented in the productive class and underrepresented in the underperformers. In order to explore these diﬀerences across museums more thoroughly, we estimate a simple model of museum attendance. The attendance levels at museums are modeled as a production function, where the inputs include museum and city characteristics. In particular, we estimate a production function for museum attendance as follows: (1)
Ait Xit Zit G
where Ait is the attendance at museum i at time t, Xit is a vector of characteristics associated with the collection of museum i at time t, Zit is a vector of characteristics at time t of the city in which museum i is located, and G is an indicator for governance structure. Data on attendance levels and collection characteristics come from the 1989 and 1999 AAMD surveys. The survey data are not without problems, some of which are described by Rosett (1991) for the earlier 1989 data. From our point of view, the collection data are most problematic. Ideally, we would like a measure of the value of the museum collection to use as one element of the X vector. In the more usual industrial-production-function context, this would be equivalent to a capital stock figure. As is well known, however, museum collections are not valued in the financial statements of museums; indeed, the standard procedure is to list the value of art assets at $1. In the AAMD survey, there are some data provided on the total value of a museum’s collection based on insurance coverage.2 These data are problematic both because insurance readjustments are likely to be sticky and because many of the museums self-insure and thus drop out of the sample when we measure collection value this way. Moreover, the censored museums are not representative since it is many of the large public museums that self-insure. An alternative measure of collection value is the current expenditures on the collection. While we may presume that acquisitions are a major component of this category, expenditures on the collection may also include restoration, framing, and other expenses. Nevertheless, this measure has the advantage of being “real” data, and is also available for a broader set of museums. Clearly what we are measuring here is a flow (analogous to investment) rather than the preferable asset value, although the flow and stock values do appear to be highly correlated. Using current expenditures 2. Museum directors were asked to provide information on both the payoﬀ of the insurance and the fraction of the collection covered. These two figures were then used to generate a total value figure.
Does Governance Matter?
on the collection may also create an endogeneity problem. Increased attendance at a museum typically contributes to the earned income of a museum, through either admissions fees or concession revenue, and thus may increase funds available for collections. To deal with this issue, we provide an alternative estimate of the attendance regression, instrumenting for collection expenditures using the market value of the endowment at the end of the prior period. Endowment value should be both independent of attendance and correlated with collection expenditures. Since a number of the museums in the sample do not report endowment values, instrumenting in this way reduces the sample size somewhat. In addition to the variable measuring collection value, we also identify each collection by type. Narrative summaries of each museum provided by the AAMD were used to categorize each museum as either survey, modern, American, or other. We are interested here in whether there is any evidence of a type bias in American museum goers. The Z vector contains a set of variables describing the characteristics of the site of the museum. The typical museum attracts both residents and tourists. To capture local demand, we used the size of the local population and the percentage of the population with a college degree. Prior work (Dimaggio 1987) suggests that educational level is a better predictor of local demand than income. We used two measures of tourist demand: hotel expenditures per capita and mean January temperature. High January temperatures are intended to capture substitution possibilities for tourists and local residents alike. We expect that, holding tourism levels constant, museums do better in climates with cold winters. Finally, we use dummies to capture governance type, distinguishing public, university-based, and other nonprofit museums. The public museums include those run by city, state, and federal governments. The set of independent variables used and the means of the data are given in table 2.1. We note that the problem of missing observations reduces the overall sample considerably, essentially halving the population of 300 museums we started with. Table 2.4 reports the results of the estimation. In the estimation, all variables were transformed to logs, given the expected nonlinear relationship between attendance and museum and city characteristics. Thus, in this specification, we can think of the coeﬃcient estimates as elasticities. The results in table 2.4 suggest that both museum and city characteristics matter for a museum’s ability to draw an audience. Collection expenditures exert a large, positive, and highly significant eﬀect on attendance. A 10 percent increase in the expenditures on collections increases current attendance by 2.5 percent to 4.0 percent, which seems to be a relatively large eﬀect given the durable nature of collection expenditures. There is some evidence that survey collections have more drawing power than other collection types. In fundamental terms, these results suggest that art matters. Our results
Sharon Oster and William N. Goetzmann
Independent Variable Log collection expenditures Type Survey American Modern MSA population (log) Percent of population with bachelor’s degree (log) Hotel expenditures per capita (log) January mean temperature (log) Governance College Public Other nonprofit Constant Observations R2
.501 (2.78)** .145 (.226) .296 (1.4) .205 (4.53)** .183 (1.22) .240 (3.46)** –.442 (–2.47)**
.454 (2.00)* .080 (.27) .408 (1.65) .124 (2.00)* .179 (1.04) .156 (1.88) –.32 (1.53)
Omitted .804 (4.61)** .539 (5.52)* 6.34 (5.56)**
Omitted .863 (4.05)** .552 (3.03)** 4.98 (3.59)**
Note: MSA = metropolitan statistical area. OLS = ordinary least squares. IV = instrumental variables. **Significant at the .01 level. *Significant at the .05 level.
are consistent with the hypothesis that collections function as economic assets, with larger collections drawing more customers. In fact, we can go further and use the coeﬃcient estimates to answer the question of what the economic impact of an increase in collection expenditures would have on the museum. The data in table 2.1 suggest that in our sample the mean annual collection expenditure is about $1.5 million, while average attendance in the sample is about 379,000. If we apply the lower elasticity figure of 0.25 generated in table 2.4, we see that an increased expenditure on the collection of $150,000 (10 percent) would yield approximately 9,500 more museum attendees each year. For this to pay oﬀ in strictly a one-year economic impact, each new attendee would have to spend $16 in a visit, which is likely high. Of course, one would not expect art investment to pay oﬀ this quickly for a museum or else they would be doing more of it!
Does Governance Matter?
In terms of location, all of the variables are of the right signs in both regressions, although only the population variable passes the usual significance tests in both specifications. We note again the truncated sample in the instrumental variable (IV) regressions. The tourist-related variables suggest that the ideal museum location from an attendance perspective is a tourist location in a cold area. For Tom Krens’ new Guggenheim museum branch in Las Vegas, the regression gives a mixed prediction: Based on tourist beds, Las Vegas looks like a good site; based on January temperature, Krens may have a failure on his hands. The results further suggest that governance type matters a good deal in terms of audience attraction. Public museums strongly outdraw nonprofit museums of either type, and university-based museums clearly deliver the smallest audiences. These results are consistent with the view that public museums stress public education, while college museums in particular may focus more on higher education, connoisseurship, and other aspects of the museum mission. These results further support Hansmann’s (1981) observations on the diﬀerences in the focus on attendance by performance arts organizations. We turn now to look directly at the role of special exhibits in museums of varying ownership types. 2.2.3 The Role of Traveling Exhibitions Special exhibitions play two important roles for museums. In some cases, these exhibitions are mounted by a museum’s own curators and represent the historical vision of that curator, expressing a particular point of view about a body of work. Thus, at one level, special exhibitions represent a curatorial research product. On the other hand, some special exhibitions—the blockbusters—serve in large measure as a way to attract large, new audiences to a museum. Attracting large audiences has financial benefits as well. Even those museums that charge no admission fees benefit through their concession and museum shops from increases in visitorship. Indeed, for the average museum, revenues from audience-related concessions exceed admissions fees (AAMD Survey 1999). The traveling special exhibition is particularly interesting in terms of function. In many cases, exhibitions travel from one museum to another and provide a way to expose a local audience to new work. For moderatesized art museums, some reliance on traveling exhibitions is common. The St. Louis Art Museum, for example, had thirty-five special exhibits in the 1990s, 35 percent of which were organized outside of the museum itself, including most of the very high-attendance shows. As such, traveling exhibitions are a way of temporarily augmenting a museum collection through, in eﬀect, leasing more-valuable works from major museums. Much of the discussion by critics on the changed role of the museum has focused particularly on the use of the special exhibition as a crowd pleaser. By mounting a recent exhibit of guitars, the Museum of Fine Arts in Boston was described
Sharon Oster and William N. Goetzmann
as “turning itself into a gigantic Hard Rock Café” (Leo 2001). Of New York’s Guggenheim, which is well known for its unusual exhibits, Heather Macdonald opined that “the Giorgio Armani show at the Guggenheim reminds us that ‘art’ in an art museum these days is optional” (Leo). There is a tension, then, between the smaller-scale special exhibit, which principally serves a research or educational function, and the audiencegenerating, revenue-producing blockbuster. In line with our earlier discussion, we expect to see diﬀerent museum types specializing in each of these forms. In particular, university-based museums are likely to be overrepresented among museums mounting specialized exhibits, while public and nonprofit museums, lured by both revenues and audience, will focus on the blockbuster segment. Before we can consider the diﬀerent production of special exhibits by diﬀerent museums, it is useful to touch briefly on the economics of exhibition production more generally. From the point of view of an industrialorganization economist and a finance professor, it is a curious process indeed. Producing special exhibits requires essentially two inputs: curatorial time and art objects. While museums can and do use visiting curators, the ability to regularly mount a diverse group of special exhibits requires a substantial curatorial staﬀ. In the modern blockbuster age, a staﬀ of exhibit designers has become increasingly important (Silver 1982), further increasing the fixed costs burden for the smaller museum. A more important barrier to mounting major exhibits by the small museums is created by the economics of art-object lending. The typical special exhibit relies on both a museum’s own objects and borrowed objects. It is the custom in the museum business that these loans are made without a fee, although it is usual for the borrowing museum to pay for travel and insurance costs. Even objects from private collections are borrowed rather than rented, although there is, at times, some restoration work serving as a quid pro quo. Initially, one might think that the borrowing tradition would make it easier for smaller museums to mount exhibits, by lowering costs. We would argue, however, that this system may discriminate against the smaller museums. In the barter system used, the smaller museum may find itself with few objects of any appreciable “trade” value and thus more often find its requests for loans refused. Similarly, private exhibitors likely prefer lending to big-name museums. As with many barter systems, this one may create an ineﬃciency by reducing the ability of the creative curator in the smaller museum to exploit his or her skill. As we will shortly argue, however, the university museum—even the relatively small one—is in a somewhat advantaged position in the borrowing business. The evidence suggests that production of traveling exhibitions among art museums is indeed a highly concentrated business. One way to measure concentration is to look at participation fees earned by museums. In 1999,
Does Governance Matter? Table 2.5
Exhibit Census Exhibit
Blockbusters in 1998, 1999 (attendance 400,000 at one museum) Monet in the Twentieth Century Boston MFA The Private Collection of Degas Metropolitan Museum Van Gogh’s Van Gogh National Gallery Mary Cassatt: Modern Woman Art Institute, Chicago Pierre Bonnard MOMA Cézanne to Van Gogh: Dr. Gachet Metropolitan Museum John Singer Sargent National Gallery Renoir’s Portraits Art Institute, Chicago Mini-blockbuster (attendance 200,000 and 400,000) Monet: Portrait of Giverny Walters Art Gallery Alexander Calder National Gallery A Collector’s Cabinet National Gallery Manet, Monet, and Gare St. Lazere National Gallery Degas at the Races National Gallery Collecting Impressionism High and Seattle Picasso and the War Years Guggenheim From Van Eyck to Brueghel Metropolitan Museum Picasso: Painter and Sculptor in Clay Metropolitan Museum Hans Hoﬀman in the Metropolitan Metropolitan Museum Jackson Pollock MOMA Delacroix: The Late Work Philadelphia Museum Portraits by Ingres National Gallery Notes: MFA = Museum of Fine Arts. MOMA = Museum of Modern Art.
for example, the AAMD data indicate that the top four museums providing data on participation fees earned 55 percent of the total fees earned.3 A decade earlier, in 1989, this figure was slightly lower. There are no university museums among this top list. Another way to estimate concentration is to look at the originating museum for recent large exhibits. This allows us to look at some museums that do not provide AAMD survey data. This information is provided in table 2.5. Of the twenty-one exhibits we identified in the 1998–99 period with attendance levels over 200,000 in a single museum, the National Gallery had one-third and the Metropolitan one-fourth of the exhibits. Again, high concentration is clearly in evidence, public and nonprofit museums are represented in proportion to their place in the pool, and no university museums are present. One might also notice that almost all of the blockbuster shows are of Impressionist painters. The 1999 AAMD list of museums with the highest earned income from 3. This figure is based on the approximately two-thirds of the museums responding to this question.
Sharon Oster and William N. Goetzmann
participation fees is principally dominated by the very largest museums. Interestingly, the smaller museums earning participation fees are disproportionately university-based museums. Here we see the importance of the more specialized traveling exhibition to the research life of the university museum. In 1999, the Harvard University Art Museums were among the top ten in participation fees among reporting institutions. These fees appear to be the result of a show mounted in 1998, Inside Out: the New Chinese Art, which traveled throughout the country in 1999 and 2000 and was mounted in cooperation with the San Francisco Museum of Modern Art. Williams College, Smith College, and Yale University all earn more from participation fees than you might expect from their operating budgets. The Harvard and Yale art galleries routinely mount special exhibitions that travel to other museums. The university museum may well have cost advantages in mounting these exhibits, as well as enhanced mission-driven reasons to support such activity. Here we see some of the advantages of the university museum in terms of ability to use curatorial talents outside the museum budget, in the quality of their history of art departments, and in terms of their ability to borrow, particularly from aﬃliated collectors. Colleges with well-endowed alumni may be able to call on these alumni to lend art to their museum exhibitions and in this way are less hampered by the borrowing culture of the art world than their similarly sized cohorts. 2.3 Museums as Social Institutions We have thus far explored the way in which museum ownership and governance structure may influence the emphasis it places on audience attraction. We turn now to look more directly at the role of a museum vis-à-vis the social elite in a city. Founding a museum, sitting on the board of a local arts institution, and contributing conspicuously to a public museum have long been an avenue into society. The role of the single philanthropist in founding museums like the Guggenheim and the Whitney in New York is well known, but the pattern is common in the rest of the country as well. In Minneapolis, T. B. Walker, who made his fortune in lumber, started the Walker Art Center in the mid-nineteenth century. The Center for British Art at Yale University is the gift of philanthropist and collector Paul Mellon. In Chicago, the Terra Museum of American Art was founded, funded, and named by its principal donor, Daniel Terra. What has happened to the museum’s role as a validator of social position? As we suggested earlier, the typical museum in the last several decades has attempted to broaden its public appeal in part to attract new audiences for revenue reasons. As museums have become democratized in their exhibitions, there is some question about whether they have lost their role as promoters of the social elite.
Does Governance Matter?
As part of their required Form 990 filings with the Internal Revenue Service, museums are asked a series of questions pertaining to their “public support” basis for tax exemption. As part of this set of questions, museums are required to indicate funds raised from individuals who have contributed over the past four years an amount in excess of 2 percent of the museum’s total funds. We use this information as one measure of the “elite focus” of the museum’s funds. As table 2.6 suggests, there is considerable variation in the reliance of museums on very large contributions. Some museums report having no patron who, in the period 1994–97, contributed more than 2 percent of museum support, while several museums receive almost half of their private support from this source. Among the museums with substantial reliance on the large gift are included several very large, high-profile museums (e.g., the Whitney Museum and the San Francisco Museum of Modern Art [MOMA]), as well as a number of smaller, less well-known museums, including the Arkansas Art Center and the Akron Art Museum. In table 2.7, we report the results of a simple regression intended to tease out some of the determinants of museum dependence on concentrated donors. The dependent variable is the ratio of donations raised from donors contributing each in excess of 2 percent of the pool to the total support pool. As independent variables, we consider two city characteristics: percentage of the city population in the top income group ($150,000 in 1990), and population stability (percentage of the population living in the same county between 1985 and 1990). Our expectation is that a museum’s reliance on high-end donors will be positively related to both measures, the intuition being that the social elites supporting museums have historically been high-income and stable in residence. In addition, we look at the museum’s age, recognizing that in early stages museums are often the product of a few wealthy benefactors, and that through a museum’s life cycle, the donor pool will tend to spread. While all variables are of the expected sign, only the income variable is statistically significant. The significance of the high-income variable is consistent with the conspicuous consumption function of museums. The greater the density of aﬄuent citizens, the greater the need to signal social status through support of the arts. It is also interesting to consider the way in which the importance of the big donor to museums may have changed over time. In panels A and B of table 2.8, we have briefly summarized the history of the museums listed in the AAMD survey founded in two historical periods: before 1920, a period in which many of the premier U.S. museums were founded, and since 1960. We note first that the ownership structure in these newer museums parallels those of the earlier museums: Two-thirds of the new museums are nonprofits, and one-third, public. There is no indication of an evolutionary trend toward one “ideal” museum form, the way we have seen in other areas of nonprofit management. A somewhat higher than expected fraction
Museum Reliance on Large Donors
Museum 1. Akron Art Museum 2. Albright-Knox Art Gallery 3. Allentown Art Museum 4. Arkansas Arts Center 5. Asia Society and Museum 6. Butler Institute of American Art 7. Boston Museum of Fine Art 8. Chrysler Museum 9. Columbus Museum of Art 10. Columbus Museum 11. Contemporary Arts Center 12. Cummer Museum of Art 13. Currier Gallery of Art 14. Dallas Museum of Art 15. Dayton Art Institute 16. Detroit Institute of Arts 17. Dia Center for the Arts 18. Flint Institute of Arts 19. Honolulu Academy of Arts 20. Huntington Library and Art Gallery 21. Huntington Museum of Art 22. Huntsville Museum of Art 23. Indianapolis Museum of Art 24. International Center of Photography 25. Isabella Stewart Gardner Museum 26. JB Speed Art Museum 27. Jewish Museum 28. Joslyn Art Museum 29. Long Beach Museum of Art 30. Marion Koogler McNay Art Museum 31. Metropolitan Museum of Art 32. Milwaukee Art Museum 33. Mint Museum of Art 34. Museum of Contemporary Art 35. Neuberger Museum of Art 36. New Museum of Contemporary Art 37. New Orleans Museum of Art 38. Newark Museum 39. North Carolina Museum of Art 40. Palm Springs Desert Museum 41. Parrish Art Museum 42. Philadelphia Museum of Art 43. Philbrook Museum of Art 44. Phoenix Art Museum 45. Pierpont Morgan Library 46. Portland Art Museum 47. San Antonio Museum of Art 48. San Diego Museum of Art
Proportion of Funds from Large Donors .1899962 0 .0012145 .4329223 .1664267 .056930 0 0 .0743780 .1177242 0 .0296322 .0756367 .0680886 .0931211 .0431507 .3419761 .0997698 .1077176 .1057051 .0330222 0 .1308966 .0312683 .0789347 0 .0891177 0 .0056818 .1067609 .0564407 .0731025 .0057806 0 .2936345 0 .0376919 .0036008 .1515550 .1635293 .0595174 .0298572 .1342124 .2005516 .2166278 .1950636 .0700298 .0187342
Does Governance Matter?
Proportion of Funds from Large Donors
Museum 49. San Francisco Museum of Modern Art 50. San Jose Museum of Art 51. Santa Barbara Museum of Art 52. Seattle Art Museum 53. Southeastern Center 54. Studio Museum in Harlem 55. Tampa Museum of Art 56. Telfair Museum of Art 57. Textile Museum 58. Toledo Museum of Art 59. Wadsworth Atheneum 60. Walker Art Center 61. Whitney Museum of American Art 62. Winterthur Museum 63. Worchester Art Museum
.3304738 .0114623 .1068513 .0058311 .3356010 0 0 0 .1806287 .3462301 .0241641 .0260312 .2245290 .0056559 .0281015
Determinants of High Donor Funding Independent Variable
–.045 .961 .002 –.0002
(–.30) (2.31)* (.82) (–.41)
Constant High-income Population stability Museum age R2 N
*Significant at the .05 level.
of the new museums do, however, appear to be university based. Most significantly, nearly every one of the new museums—including those associated with universities—was founded by a large gift of money or a gift of art by a major donor. Indeed, the role of the single major donor appears, if anything, to have increased over time. Interestingly, many of the new donors come from the same industry bases as those in the earlier period— manufacturing, oil, and transportation. Our evidence suggests remarkable stability in the prevalence of founding donors and the profile of those donors in the museum world. 2.4 Museums as Aesthetic Institutions In the analyses thus far, we have emphasized the ways in which serving popular audiences and serving a narrower elite group compete for museum attention. While recent scholarship has underscored the contrasts in these
U.S. Museum, by Year Founded and Donor
Amon Carter Museum Asian Art Museum of San Francisco Brandywine River Museum Contemporary Arts Center David and Alfred Smart Museum of Art (University of Chicago) Dia Center for the Arts Elvehjem Museum of Art (University of Wisconsin) Georgia O’Keefe Museum
A. Founded Since 1960 1961 Amon Carter (publishing) 1966 Avery Brundage (construction) 1971 DuPont (chemicals) 1976 State 1974 Smarts (publishing)
Hirshhorn Museum and Sculpture Garden Huntsville Museum of Art Herbert F. Johnson Museum of Art (Cornell University) Jack S. Blanton Museum of Art (University of Texas) Jane Voorhees Zimmerli Art Museum (Rutgers) Krannert Art Museum (University of Illinois) Museum of Contemporary Art National Museum of African Art National Portrait Gallery Neuberger Museum of Art (SUNY Purchase) New Museum of Contemporary Art Salvador Dali Museum Samuel Harn Museum San Antonio Museum of Art San Jose Museum of Art St. Petersburg Museum of Fine Arts Studio Museum in Harlem Tampa Museum of Art UCLA Hammer Museum University of California, Berkeley, Art Museum University of Iowa Museum of Art Wexner Center for the Arts Yale Center for British Art Albright-Knox Art Gallery Art Institute of Chicago Baltimore Museum of Art Brooklyn Museum of Art Butler Institute of American Art Carnegie Museum of Art Cincinnati Art Museum
DeMenil (oil and banking) Faculty idea: no money
1997 1966 1970 1973
Anne and John Marion (former Sotheby’s head) Hirshhorn (finance, mining) City Johnson (manufacturing)
Herman Krannert (box manufacturing)
1967 1964 1962 1974
Daniel Brenner Government Government Roy Neuberger (finance)
1977 1971 1981 1981 1969 1961 1967 1967 1994 1970
City A. R. Morse (industry) Samuel Harn (manufacturing) City City M. Acheson Stuart (publishing) Volunteer founders DeMenils (oil and banking) Hammer (chemicals) Hans Hoﬀmann (artist)
1967 1989 1977
Owen and Leone Elliot Wexner (retail) Andrew Mellon (transport and aluminum)
B. Founded before 1920 1826 John Albright (steel) 1879 Group of businessmen 1914 M. Carey Thomas (president of Bryn Mawr; railroad money inherited) 1823 Community group 1919 Joseph Butler (manufacturing) 1896 Andrew Carnegie (steel) 1896 Citizen group
Cleveland Museum of Art
Cooper-Hewitt National Design Museum Corcoran Gallery of Art Crocker Art Museum Currier Gallery of Art Dallas Museum of Art Davis Museum Dayton Art Institute
1887 1869 1885 1919 1903 1889 1919
Delaware Art Museum Denver Art Museum Detroit Institute of Arts Fine Arts Museums of San Francisco Freer Gallery of Art Frick Collection Harvard University Art Museums (Fogg) Henry Art Gallery Huntington Library and Art Gallery Indianapolis Museum of Art Isabella Stewart Gardner Museum Los Angeles County Museum of Art Memory Art Gallery of Rochester Metropolitan Museum of Art Michael C. Carlos Museum Milwaukee Art Museum Minneapolis Institute of Arts Mississippi Museum of Art Munson-Williams-Proctor Arts Institute
1912 1883 1885 1894 1916 1920 1895 1917 1919 1883 1903 1910 1913 1870 1876 1888 1915 1911 1919
Huntington (oil); Kelley (development); Hurlburt (banks) Cooper grandchildren (railroads) William Corcoran (banking) Edwin Crocker (railroads) Moody Currier (banking) Citizen group Wellesley College Julia Paterson Carnell (National Cash Register) Citizen group Municipal Brearly (journalism) DeYoung (publishing) Charles Freer (railroads) Henry Frick (steel) William Hayes Fogg (China trade) Horace Henry (railroads) Henry Huntington (railroads) John Herron Isabella Gardner (commerce) City Mrs. J. S. Watson (telegraph) Group of businessmen Emory; Carlos (alcohol distributor)
Museum of Fine Arts, Boston
New Orleans Museum of Art Newark Museum Parrish Art Museum Philadelphia Museum of Art Phillips Collection Portland (Maine) Museum of Art Portland (Oregon) Art Museum Saint Louis Art Museum Seattle Art Museum Telfair Museum of Art Toledo Museum of Art Wadsworth Atheneum Walker Art Center Walters Art Museum Worcester Art Museum Yale University Art Gallery
1911 1909 1898 1876 1897 1883 1892 1892 1917 1875 1901 1842 1879 1908 1896 1832
Citizen association Munson (banking); Williams (politics); Proctor (manufacture) Group of citizens (Henry Kidder, finance; W. Endicott, dry goods; Charles Eliot, Harvard president) Isaac Delgado (sugar) Louis Bamberger (retail) Samuel Parrish Group: Centennial related Duncan Phillips (steel) Margaret deMedici Sweat (retail) Henry Corbett (bands) Group: St. Louis Fair Russell Fuller (medicine) Alexander Telfair (trade; agriculture) Edward Libbey (glass) D. Wadsworth (insurance) T. Walker (lumber) William Walters (railroads) Stephen Salisbury (trade) John Trumbull (artist)
Note: Includes all museums listed in the AAMD directory.
Sharon Oster and William N. Goetzmann
two objectives, it is worth considering the commonalities as well. An art museum is, for the most part, a spatial technology for facilitating the personal experience of art. While connoisseurship might be the elite extreme of the aesthetic experience, and art education the populist extreme, they can be expected to share some common kernel or at least to be connected by a continuum of personal experience. Are there cultural commonalities in the “high” and “low” experience of art? Can a single institution serve both extremes? To explore the question of whether common and elite artistic tastes are connected, we used time-series analysis of art prices and attendance at museums. Clearly, art serves in some measure as an investment good, and thus its price will reflect other forces in investment markets. This has been the direction of most of the prior literature. For example, Goetzmann and Spiegel (1995) take art as a fixed percentage of wealth and show how this may explain the covariation of art with equity markets. More recently, AitSahalia, Parker, and Yogo (2001) show how this covariation between luxury goods like art might account for the magnitude of the equity premium. To date, however, there has been little theoretical work that links a socialpecking-order framework to the prices of the luxury goods and the aesthetic experience directly. On the other hand, such frameworks are common in other parts of the finance literature. For example, “keeping up with the Joneses” models in the asset pricing literature, such as Bakshi and Chen (1996) and Campbell and Cochrane (2001), show how competitive, socially determined preferences may aﬀect security prices. A natural question to ask is whether local social competition determines the demand for conspicuous consumption as well and what role museums might play in this competition. Economists have long debated the issue of whether art provides a fair rate of return to investors. The natural presumption is that some component of the return to art investment is the aesthetic dividend that accrues to the owner—the private benefits enjoyed by viewing the work. Neglecting expectations about future resale, the entire value of owning a painting would be the capitalized stream of the aesthetic dividends. Given the evidence on the social role of art institutions presented above, one could conceivably substitute “social” for “aesthetic,” however. Museums deliver a flow of these nonmonetary dividends to participants: The aesthetic dividends are delivered through viewership, the social dividends are delivered through board association, membership, and attendance. To the extent that there are common tastes and desires for social signaling, we might expect that measures of the dividend flow and its capitalized value to covary. Indeed, our cross-sectional regressions found a relationship between attendance—i.e., the demand for the flow—and the value of the stock. We also might expect art prices to covary with attendance. By the same token, the existence of common aesthetic tastes and demand for social signaling
Does Governance Matter?
should be associated with correlations in museum attendance. In this section, we test these two propositions with time-series data on museum attendance and the returns to art investment. 2.4.1 Data It is surprisingly diﬃcult to obtain time-series data on museum attendance. The AAMD was unwilling to provide us access to their annual survey for multiple years. As an alternative, we contacted the top fifty art museums in the country and asked for their annual attendance numbers. Many had to reconstruct this information specifically for us. In total, we were able to obtain annual attendance figures for twenty-six museums for diﬀerent intervals of time. Table 2.9 reports this time-series data. In order to test hypotheses about the covariation in art prices and museum attendance, we construct an equal-weighted index of annual percentage changes in museum attendance from this data. As table 2.9 suggests, the composition of this changes as museums enter and exit the sample, but it provides the best measure we can get of the annual fluctuations in national art museum attendance. Table 2.10 reports the statistical characteristics of the index for diﬀerent subperiods of the data. For our measure of returns to investment in art, we use the Mei and Moses (2002) art price indexes. These are estimated from repeated sales of art works auctioned at major houses from 1875 to the present. The technology is similar to Goetzmann (1993)—it calculates pretax and precommission investment returns based upon the auction-to-auction price relative, conditional upon resale. Hence, those works that did not sell after once appearing at auction have no influence on the estimation of the time series of returns. For our purposes, we are chiefly interested in the intertemporal variation in art prices. In small sample, repeat-sales estimators may induce negative serial correlation in the series estimates. However the Mei and Moses data set is large, and thus we may take their index estimation as a fairly accurate representation of the trends in art prices over the past forty years. 2.4.2 Do Art Returns Explain Museum Attendance? If art prices and museum attendance both reflect fluctuations in the common component of demand for the aesthetic or social dividend, we should expect to find some correlation between attendance and the art index. Figure 2.2 plots the cumulated growth in art prices and in museum attendance for the equal-weighed index and for a few representative cities. From 1961 to 2000, art prices appreciated at a considerably higher rate than the growth rate in attendance at art museums. The plot suggests little relationship between attendance and art prices, however. Art prices spiked in the late 1980s and 1990, while the attendance graph shows no such trend. To more formally examine the relationship between art prices and attendance trends, we regress the equal-weighted index of annual percentage
Sharon Oster and William N. Goetzmann
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 29,610 28,943 25,469 34,925 45,526 48,689 32,331 25,545
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 545,152 596,223 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 422,464 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 380,425
nty ou .C
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 610,102 487,753 450,817 450,000 486,847 552,299 596,419 590,075 541,557 444,094 379,096 396,695 489,917 502,635 470,692 509,292 456,824 515,058 483,964 442,238 497,482 542,813 534,676 492,624 553,503 484,849 463,938 487,861 467,064 509,377 534,162
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 63,591 73,993 71,701 83,762 73,665 77,656 67,097 84,212 66,535 72,423 67,656 74,698 71,393 71,875 66,284 68,081 72,134
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 78,966 65,003 86,802 109,000 120,000
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 54,991 84,724 92,954 90,432 100,156
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 291,100 442,200 419,600 427,000 410,700 422,300 380,000 458,100 415,200 431,500 501,661 n.a.
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 315,047 302,196 483,347 328,714 322,073 311,577 347,996 317,090 340,677 277,589 290,299
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 61,868 62,666 85,117 91,369 73,880
Ind ian ap oli s
ton ng Hu
s lla Da
1. 1960 2. 1961 3. 1962 4. 1963 5. 1964 6. 1965 7. 1966 8. 1967 9. 1968 10. 1969 11. 1970 12. 1971 13. 1972 14. 1973 15. 1974 16. 1975 17. 1976 18. 1977 19. 1978 20. 1979 21. 1980 22. 1981 23. 1982 24. 1983 25. 1984 26. 1985 27. 1986 28. 1987 29. 1988 30. 1989 31. 1990 32. 1991 33. 1992 34. 1993 35. 1994 36. 1995 37. 1996 38. 1997 39. 1998 40. 1999 41. 2000
Museum Attendance Data ore
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 1,750,000 1,500,000 1,400,000
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 481,049 138,016
n.a. n.a. n.a. n.a. n.a. n.a. 2,665,388 1,887,135 1,174,674 1,133,870 1,384,448 1,185,741 1,203,999 1,124,870 1,204,857 1,026,918 1,425,704 1,350,302 2,750,039 357,577 506,956 586,587 372,182 415,000 579,569 914,978 421,296 1,099,440 860,689 950,833 663,869 1,003,059 848,099 612,005 551,935 541,308 663,429 602,141 554,024 1,328,765 597,409
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 85,333 80,349 73,978 98,458 79,499 84,952 88,294 83,733 87,273 102,682 85,678 75,398 110,910
Note: n.a. = not available.
changes in attendance on annual percentage changes in the Mei and Moses (2002) art index. We also perform each regression separately by city, and finally we stack all cities together and estimate the coeﬃcient on art under the assumption of equality of coeﬃcients. Table 2.11 reports the regression results, showing no evidence of a relationship between attendance and art returns. Assuming our tests have power, we can interpret this negative evidence as favoring the hypothesis that the demand at the high end and the demand at the low end for the nonmonetary dividends supplied by art are essentially disjoint. Figure 2.2 also suggests little relationship among the museums in the
Does Governance Matter?
ney Ya le
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 417,380 436,040 423,362 645,799 360,793 415,340 401,305 396,554 352,099 473,259 473,074 334,033 350,044 335,996 371,672 356,801 406,910 456,825 499,693 509,123 516,568 518,398 430,252 581,590
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 671,303 919,191 745,526 788,717 789,182 875,118 1,048,302 1,029,638 1,129,366
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 209,312 225,685 216,340 212,057
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 542,656 447,436 479,738 645,738 553,853 653,016 494,848 499,944
Wa lke r
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 41,811 43,641 43,850 47,575 59,770 99,706 123,722 84,338 89,519 86,779 77,228 76,031 59,551 n.a. n.a. 61,817 61,145 68,281 n.a. n.a. n.a. 48,118 75,713 81,345 103,589 119,211 78,836 72,188 85,385 84,797 68,144 76,722 69,980
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 873,515 841,683 1,148,816 734,149 748,966 645,999
uis Lo St.
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 52,494 54,174 55,092 40,268 87,689 98,309 123,212 150,436 69,487
ia ph del Ph
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 5,597,973 4,042,044 4,684,095 4,731,418 5,637,841 6,198,523 5,969,528 5,126,954
l na Na tio
Me tro po lita n
Mu Fin seum eA o rts f
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 490,888 390,604 327,431 341,901 335,142 437,685 491,603 507,507 511,838 665,887 560,187 510,992 760,868 544,804 579,466 1,247,768 1,259,642 n.a. 1,801,924 1,323,380 1,251,094 1,784,332
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 2,225,530 2,272,212 2,590,851 3,326,012 2,871,417 3,337,040 3,235,684 4,687,277 3,369,934 3,574,138 3,232,876 4,333,918 3,945,708 3,889,471 3,290,133 4,871,698 3,767,018 4,585,554 4,329,474 4,479,344 4,453,441 4,399,542 4,308,881 4,657,430 4,566,579 5,309,076 4,950,136 4,850,913 5,152,884
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 220,000 200,000 247,000 275,000 306,000 255,000 267,000 200,000 275,000 344,000 143,676 110,952
n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 261,342 231,829 278,981 401,489 458,547 369,791 441,405 637,578 420,150 426,547 387,743 310,595 340,781 457,471 399,564 313,143 338,090 260,800 273,986 273,426 231,100 293,040 421,867 291,800 385,836 464,244 570,255
99,196 92,989 94,372 83,440 79,302 92,019 101,424 114,211 131,811 126,253 119,004 101,482 120,946 118,366 87,496 96,293 144,290 75,392 98,546 96,423 106,677 110,223 99,346 110,914 97,130 117,746 185,951 118,467 137,867 135,981 155,085 119,834 120,630 121,436 103,786 96,873 100,968 98,848 111,547 n.a. 116,400
Note: n.a. = not available.
sample. This is even more surprising. While the low correlation between attendance and art prices may not be surprising given that auctions reflect demands by a relatively aﬄuent clientele, (indeed, a group whose wealth may depend upon a diﬀerent set of factors than does the wealth of those who regularly attend art galleries) it is surprising to us to see low intercity relationships in museum attendance trends. In fact, the average correlation among the cities, reported in table 2.12, is close to zero. One way to interpret this is that all art appreciation, like all politics, is local. In some ways, this result reinforces our earlier finding on the importance of both city- and museumspecific factors in determining attendance patterns. An alternative explana-
Sharon Oster and William N. Goetzmann
Summary Statistics, Annual Percentage Changes in Attendance Index of American Art Museums, 1961–2000 Geographical Growth
–0.0105 0.0422 0.0307 0.0205
–0.0100 0.0461 0.0414 0.0414
0.0930 0.1064 0.1046 0.0599
1961–1970 1971–1980 1981–1990 1991–2000
Fig. 2.2 Comparison of the performance of art at auction to measures of growth in attendance Notes: For an equal-weighted index of museum attendance, and for three museums: Los Angeles County Museum, New York Metropolitan Museum of Art, and the Whitney Museum of American Art. Equal-weighted average of available museums, and three large institutions
tion is that traveling shows are important determinants of attendance with the biggest drawing shows are in diﬀerent cities in diﬀerent years. 2.5 Conclusions Art museums in the United States come in a range of ownership forms. In this paper, we have found striking diﬀerences in the performance of
Does Governance Matter? Table 2.11
Regressions of Equal-Weighted Percent Changes in Attendance on Art Returns Coeﬃcient
Asia Society and Museum Baltimore Museum of Art Dallas Museum of Art DeCordova Museum and Sculpture Park Georgia Museum of Art Johnson Huntington Library and Art Gallery Illinois Art Gallery Indianapolis Museum of Art J. Paul Getty Museum Kimbell Art Museum L.A. County Museum of Art Memorial Art Gallery of Rochester Metropolitan Museum of Art Museum of Fine Arts National Gallery of Art Norton Museum of Art Philadelphia Museum of Art Princeton University Art Museum Saint Louis Art Museum Dali Guggenheim Museum Walker Art Center Walters Art Museum Whitney Museum of American Art Yale University Art Gallery
–0.033 0.381 0.210 –1.179 0.962 –0.064 –0.175 –0.747 0.000 0.528 0.000 0.286 –0.394 –0.131 –0.451 0.123 –0.749 –1.502 0.178 0.506 0.127 0.001 –0.143 –0.490 0.320 0.152
–0.037 1.006 –0.651 –2.203 1.629 –0.480 –2.181 –1.298 n.a. n.a. n.a. 0.742 –1.362 –0.851 –1.315 0.302 –0.665 –3.109 0.863 1.035 0.145 0.002 –0.709 –1.43 1.517 1.087
4 10 10 4 4 16 30 7 1 2 1 34 12 28 19 7 8 5 25 7 3 8 23 11 26 38
0.001 0.112 0.050 0.708 0.570 0.016 0.145 0.252 n.a. 1.000 n.a. 0.017 0.156 0.027 0.092 0.018 0.069 0.763 0.031 0.l77 0.021 0.000 0.023 0.127 0.088 0.032
Equal-weighted index Stacked regression
Note: City-by-city regression, index regression, and stacked regression.
these museums that are consistent with our expectations about diﬀerences in institutional economic incentives. Based on our work comparing art prices and museum attendance, we further find that the levels of demand for art by the various sectors of the market are disjoint. In this light, it is interesting to consider the recent Italian proposal to begin moving some of the major museums into the nongovernmental sector. Our own work suggests that changing governance in this way may well change the operating behavior of those museums, perhaps in ways unanticipated by the government. Our work also suggests that art collections housed in museums, although often treated as noncommercial assets, have considerable ability to generate revenues. Moreover, the productivity of a collection varies significantly by the characteristics of the city in which it is located. In our his-
V1 BaltimoreMOA DallasMOA Johnson Huntington Lacounty MemorialAG Met MFA Princeton Walker Walters Whitney Yale
1.00 0.05 0.06 0.06 –0.11 –0.10 –0.07 0.09 –0.11 0.12 –0.03 0.09 –0.03 0.01
0.05 1.00 0.09 –0.67 –0.06 –0.26 –0.32 –0.05 –0.43 0.20 –0.09 0.13 0.19 0.09
0.06 0.09 1.00 –0.51 0.37 0.03 –0.28 –0.22 –0.31 0.73 –0.24 –0.52 0.35 0.38
0.06 –0.67 –0.51 1.00 0.15 0.59 0.60 0.45 0.21 –0.23 0.19 0.11 –0.12 –0.48
n gto nti n Hu
–0.11 –0.06 0.37 0.15 1.00 0.24 0.24 0.10 0.30 0.08 –0.20 –0.23 –0.18 –0.20
nty ou L.
–0.10 –0.26 0.03 0.59 0.24 1.00 –0.22 0.06 –0.12 –0.04 –0.23 –0.47 0.31 –0.16
–0.07 –0.32 –0.28 0.60 0.24 –0.22 1.00 0.14 0.81 –0.25 0.48 0.08 –0.25 –0.27
Correlations in Attendance (museums with at least ten years of data)
–0.11 –0.43 –0.31 0.21 0.30 –0.12 0.81 –0.06 1.00 –0.54 0.14 –0.25 –0.39 –0.31
Me t r op oli tan M Fin useu eA mo rts f
0.09 –0.05 –0.22 0.45 0.10 0.06 0.14 1.00 –0.06 –0.07 –0.22 0.13 –0.04 –0.55
0.12 0.20 0.73 –0.23 0.08 –0.04 –0.25 –0.07 –0.54 1.00 –0.39 –0.11 0.04 0.03
Wa lke r
–0.03 –0.09 –0.24 0.19 –0.20 –0.23 0.48 –0.22 0.14 –0.39 1.00 0.10 –0.05 0.11
0.09 0.13 –0.52 0.11 –0.23 –0.47 0.08 0.13 –0.25 –0.11 0.10 1.00 –0.47 –0.11
–0.03 0.19 0.35 –0.12 –0.18 0.31 –0.25 –0.04 –0.39 0.04 –0.05 –0.47 1.00 –0.02
0.01 0.09 0.38 –0.48 –0.20 –0.16 –0.27 –0.05 –0.31 0.03 0.11 –0.11 –0.02 1.0
Does Governance Matter?
torical work on the relationship between social elite and museums, we find remarkable stability: big donors continue to found new museums and support those museums with largesse earned in traditional, old-economy ways. In this paper, we have focused on the role of governance structure in museum decision making. An interesting example of the dynamics of museum governance can be witnessed in the extraordinary set of western American art collections accessible to the public in Denver, Colorado. In the 1980s and early 1990s, the Denver area had not one, but three, superb collections of art of the American West. The Museum of Western Art (MWA) was founded in the early 1980s as a private, not-for-profit institution by cattleman William Foxley to display his personal collection of paintings and sculpture, which were on loan to the organization for which he served as the chairman of the board. The MWA collection focused on nineteenthand early-twentieth-century “masterpieces” of western art—from the action paintings of Remington and Russell to the later, much-admired modernist paintings by Taos and Santa Fe artists. The Philip Anschutz collection, similarly, is composed of major works of western American art, and it was somewhat more widely known than the MWA collection. Anschutz amassed a fortune on oil, railroads, and telecommunications, and, like William Foxley, began to collect prize western American paintings and sculpture as a private collector. Over the past two decades, he has exhibited it widely to the public by publishing a catalogue of the collection and underwriting traveling exhibitions of the works to major art museums around the country. The Denver Art Museum (DAM) recently organized a show of the Anschutz collection that traveled to the Jocelyn Museum in Omaha and the Corcoran Gallery in Washington, D.C. The third major collection in Denver was in the Denver Art Museum itself. Dorothy and William Harmsen, founders of the Jolly Rancher Candy Company, assembled a collection of noteworthy western paintings and American Indian art over several decades, which they donated in 2001 to the DAM. The artists whose works are represented in the Harmsen collection are essentially the same as those in the Foxley and Anschutz collections, but they are a part of a public museum, not a private collection or a private, not-for-profit museum. The constellation of collections is instructive, first because of the apparent rivalry within Denver among some of the leading businessmen at the time to form top western art collections—perhaps as a way of “keeping score” and perhaps as a way of demonstrating refinement, taste, and “western” values. In this respect, it is tempting to draw a parallel to the rivalries among turn-of-the-century Pittsburgh’s captains of industry as they vied to buy European masterpieces. Perhaps more interesting for our purposes is that these founders chose diﬀerent governance forms for the context of their philanthropy. The collection of the Museum of Western Art, until it was ultimately moved and
Sharon Oster and William N. Goetzmann
partly dispersed, was largely in the control of the founder, who was able to sell and to augment the exhibit. The museum relied, to a large extent, upon his financial support. Nevertheless, it was a not-for-profit organization with a mission to serve the public through its exhibitions. The Anschutz collection, on the other hand, was not necessarily formed with the public good in mind: The founder has complete control and no special mandate to use it for philanthropic goals, although lending to a traveling show is certainly a benefit to the public. Although Anschutz and Foxley undoubtedly had the option to give their collections to the Denver Art Museum, they both chose to maintain control of their collections to a greater or lesser degree. In contrast, the Harmsen collection is no longer under the control of the founder, nor does it receive top billing at the museum. The DAM prides itself on displaying an extensive survey collection of world art, as opposed to a regionally focused collection. While Harmsen can probably exert influence on the mission of the museum through his philanthropic activities, the director of the DAM has a larger range of choices about the strategic deployment of the institution’s assets. In addition, the DAM serves a broader constituency—a community with an interest in world art, not solely focused on western Americana. Thus, institutional forms facilitate diﬀerent donor and community goals, even when the art itself is similar.
References Ait-Sahalia, Yacine, Jonathan A. Parker, and Motohiro Yogo. 2001. Luxury goods and the equity premium. NBER Working Paper no. W8417. Cambridge, Mass.: National Bureau of Economic Research. Association of Museum Directors. 1989. Statistical survey of members. Washington, D.C.: Association of Museum Directors. ———. 1999. Statistical survey of members. Washington, D.C.: Association of Museum Directors. Anheier, Helmut, and Stefan Toepler. 1998. Commerce and the muse: Are art museums becoming commercial? In To profit or not to profit, ed. Burton Weisbrod, 233–48. New York: Cambridge University Press. Bakshi, Gurdip, and Zhiwu Chen. 1996. The spirit of capitalism and stock market prices. American Economic Review 86 (March): 133–57. Boylan, Patrick. 1999. Universities and museums: Past, present and future. Museum Management and Curatorship 18 (1): 43–56. Campbell, John, and John Cochrane. 2001. By force of habit: A consumptionbased explanation of aggregate stock market behavior. Journal of Political Economy 107 (April): 205–51. d’Harnoncourt, Anne, Paul DiMaggio, Marilyn Perry, and James Wood. 1991. The museum and the public. In The economics of art museums, ed. Martin Feldstein, 35–60. Chicago: University of Chicago Press.
Does Governance Matter?
DiMaggio, Paul. 1986. Nonprofit enterprise in the arts. New York: Oxford University Press, ———. 1987. Nonprofit organizations in the production and distribution of culture. In The nonprofit sector, ed. Walter W. Powell, 195–220. New Haven, Conn.: Yale University Press. Goetzmann, William N. 1993. Accounting for taste: An analysis of art returns over three centuries. American Economic Review 83 (5): 1370–76. Goetzmann, William N., and Matthew Spiegel. 1995. Private value components and the winner’s curse in an art index. European Economic Review 39 (May): 549– 55. Grana, Cesar. 1971. The private lives of public museums: Can art be democratic? In Fact and symbol: Essays in the sociology of art and literature, 95–111. New York: Oxford University Press. Hansmann, Henry. 1981. The role of nonprofit enterprise. Yale Law Journal 89:835–901. Ingham, John N. 1997. Reaching for respectability: The Pittsburgh industrial elite at the turn of the century. In Collection in the Gilded Age: Art and patronage in Pittsburgh, 1890–1910, ed. Gabriel P. Weisberg, DeCourcy E. McIntosh, and Alison McQueen, Pittsburgh: Frick Art and Historical Center. Kotler, Neil, and Philip Kotler. 2001. Can museums be all things to all people? In Museum Management and Curatorship 18 (3): 271–87. Leo, John. 2001. But where’s the art? US News and World Report, 14 May, p. 14. MacDonald, Heather. 2001. The Met’s triumphant democratic elitism. City Journal 11 (Winter): 76–91. Mei, J. P., and Michael Moses. 2002. Art as an investment and the underperformance of masterpieces: Evidence from 1875–2000. American Economics Review, forthcoming. Meyer, Ruth Krueger, and Madeleine Fidfell Beaufort. 1997. The rage for collecting: Beyond Pittsburgh in the Gilded Age. In Collection in the Gilded Age: Art and patronage in Pittsburgh, 1890–1910, ed. Gabriel P. Weisberg, DeCourcy E. McIntosh, and Alison McQueen, Pittsburgh: Frick Art and Historical Center. Rosett, Richard. 1991. Art museums in the U.S.: A financial portrait. In The economics of art museums, ed. Martin Feldstein, 129–76. Chicago: University of Chicago Press. Silver, Stuart. 1982. Almost everyone loves a winner. Museum News (November– December): 25–35. Temin, Peter. 1991. An economic history of American art museums. In The economics of art museums, ed. Martin Feldstein, 179–93. Chicago: University of Chicago Press.