Capsule Review: Attendance and Participation in the Performing Arts

Title: Attendance and Public Participation in the Performing Arts: A Review of the Empirical Literature

Author(s): Bruce A. Seaman

Publisher: Georgia State University

Year: 2005

URL: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=895099

Topics: economics, econometrics, arts participation, lit review

Methods: literature review

What it says: This is a critical review of cultural economics research literature, starting with Baumol and Bowen’s work in the 1960s up through research published in the early 2000s. Broadly, Seaman examines the conventional wisdom that studies finding that the arts are a luxury good are “carefully designed confirmations of the obvious.” According to Seaman, while this perspective might make sense from a layman’s perspective, findings that confirm arts as a luxury good are hard to replicate because of challenges facing performing arts data and conceptual  confusion. The lit review is structured into four chapters:

1. Introduction

Introduces the idea that cultural economics studies tend to rely more on survey data or data derived from other studies on arts participation that have used survey data. The studies are thus more about parsing out the different predictors of arts attendance than about calculating precise price elasticities of demand.

Two major economists in the field, Louis Lévy-Garboua and Claude Montmarquette, claim that there needs to be more progress in the use of econometrics and large data sets in the field of cultural economics, and in particular, that there needs to be a rigorous consideration of how to model “cultivation of taste” before questions about price elasticity of demand can be answered definitively. Seaman’s lit review attempts to evaluate these claims based on past research and results.

      2.  Observations about arts audiences from survey data

This chapter discusses observations about arts audiences based on survey data and econometric analysis. Early reports and papers on the socioeconomic makeup of arts audiences, including Baumol and Bowen (1966) and a study from the Ford Foundation (1974) showed that higher education level is more associated with arts attendance than higher income, and this finding is not restricted to the United States or to a single arts type. While the Ford Foundation study highlighted a causal effect of education level on art participation, the effects of income and education are generally difficult to separate in econometric analysis. In summary, there are differences in the degree to which education level plays a role in arts participation, and a more recent study (DiMaggio & Useem, 2004), was reluctant to conclude that education was the dominant driver of arts participation.

This chapter also looks at audience overlap between audiences of different types of arts events. In general, there is some conflicting evidence of audience overlap, or people attending more than one type of arts event. Some studies show limited overlap, suggesting that  while audiences of different types of arts events (like theater and ballet, for example) are similar in socioeconomic factors, they are not the same audiences. In comparing rates of participation between attendees for different arts events, it is important to consider the “base rate” of participation for each group before considering crossover to different categories, and this may be a reason for conflicting findings. Further, there is some literature that looks at “snobs,” who consume just one type of arts event vs. “omnivores,” who consume a variety of arts. Behavioral changes that have made people tend to eclecticism instead of snobbery should be somewhat encouraging to arts providers.

Finally, this chapter considers how audiences have changed over time. First, the evidence suggests that arts audiences are getting older as the percentage of audiences 18-29 has fallen while audiences over the age of 50 has risen. There is conflicting evidence on the changing “elitism” of arts audiences and the role of education over time. In the 2002, there was a stronger boost in arts attendance from having attended some college instead of having a college degree, and there have been other indicators that education level has “smoothed’ in attendees. DiMaggio and Mukhtar (2004) calculated the “odds” of arts attendance for a particular group, and found that arts attendance declined for both high school and college graduates, but declined further for high school graduates.

    3. Arts demand studies: price and income elasticities

Of the 44 econometric studies that have been published since 1966, 29 have attempted to either derive or impute own price or income elasticities of demand for the performing arts. Despite the conventional wisdom that frames the arts as a luxury good, this view has not been confirmed by the data, with 12 of the 29 studies finding arts prices to be relatively inelastic and 4 finding strong evidence of high price elasticities. On the whole, studies that use more aggregated arts price data, or examine the data from many organizations, tend to find more price inelasticity.

Seaman suggests that studies with less aggregated data are analytically superior to studies that use more aggregated data because they are able to look at a particular organizational type, size, and mission. One of the problems with aggregating price data to derive price elasticity of demand is that the “average” ticket price, calculated as the total number of tickets over the total revenue, is a price that no consumer actually faces. Since arts organizations are able to price discriminate, looking at a more disaggregated measure of price data allows for more refined analysis through a consideration of seating type (main floor vs. box seats, for example) and scale of ticket prices. The most disaggregated studies show fairly high price elasticities of demand, though there are no unambiguous findings of price elasticity greater than one at the most disaggregated level.

The low price elasticities may be related to the fact that arts organizations, like sports events, systematically under charge for tickets. Another explanation is that the price of a ticket to an arts event is not the “full price” of attendance, meaning that it does not include the cost of transportation, dining, and other costs associated with attending an arts event. Additionally, Seaman notes that arts organizations may be under-charging for tickets may be because organizations are attempting to incentivize donations from some individuals while allowing for a broad range of people to attend arts events.

In terms of unique challenges that might face cultural economists when calculating the demand for arts events, there may be some particular considerations for thinking about product homogeneity, established consumer taste, and assessing quality variation in arts products. In terms of problems with measurement for determining price elasticity, the biggest problem is that available data sources do not incorporate a demand-determining variable.  For example, a researcher might have data about ticket sales for a particular organization, but may not have data on consumers’ incomes. Some studies have successfully decomposed the effect of rising income and used Gary Becker’s idea of “full income” and found low elasticity results. Additionally, some have suggested that econometric research needs to include a variable that accounts for consumers’ ability to learn by doing, or to acquire and refine their taste in the arts as their consumption increases, though Seaman does not think that prior attempts have been successful to account for this in formal modeling. In summary, the evidence suggests a low elasticity of demand for arts organizations, which means that ticket prices at arts organizations may be kept intentionally low, perhaps to encourage donations from consumers or allow consumers from a broad range of income levels to participate in events.

Efforts to examine cross-price elasticities and complements for arts events have drawn some interesting conclusions. The historical approach has been to examine the “recreation and reading component” of the Consumer Price Index, thus comparing the price and demand for other recreational activities to arts performances. Gapinski (1986) finds evidence of cross price elasticities in smaller geographic markets, which suggests that considering location and transportation to events might be an important consideration for researchers. Looking at television subscribers in particular has shown a strong negative effect on ticket sales in some geographic markets, though there is some conflicting evidence on the effect of television.

   4. Evaluation and conclusion

Seaman highlights major findings from the arts demand literature that may have some bearing on how we understand demand for the arts:

1. Substitution evidence shows that there substitution across performing arts may be significant, though there is less evidence of substitution within an art form.

2. Aggregate price elasticities are lower than single firm price elasticities.

3. In future studies, income elasticity needs to account for full income and opportunity cost of leisure time.

4. Quality matters in arts demand, but it’s hard to measure and model.

5. Considering how people’s attitudes and tastes evolve in the arts will be important to consider as researchers estimate demand.

6. Education level is a strong determinant of arts participation in survey data, but regression analysis suggests that formal training in the arts, family socialization, and arts training are stronger determinants of arts participation.

In summary, Seaman finds that there are surprisingly few “arts axioms” in arts demand, and that while there are certainly trends in the literature, there is also contradictory evidence highlighting the need to think critically about analytical methods and units of analysis.

What I think about it: This is a very thorough examination of the major econometric studies that have been published over the 40-year period from 1966-2005. Seaman finds many examples of inconsistencies in arts data and arts analysis and has sound reasoning for why those inconsistencies might exist, both in the econometric methods used to derive the results and from intuition about how arts organizations and arts consumers behave.

Seaman examines the idea of economic disadvantage through a largely theoretical lens, and I think two of the major topics in the article might have some bearing on arts and economic disadvantage hypothesis #1:

1. Looking at summary statistics of survey data suggests that education level is a strong driver of arts participation and attendance, even more so than income. However, regression analysis suggests that some type of formal arts training and other social lifestyle factors may be a more important driver of arts participation than education. This may be due to the fact that education level has a multicollinearity problem with other socioeconomic factors, but some evidence shows that, after considering lifestyle and attitude toward the arts, socioeconomic status is no longer a significant predictor of arts participation.

2. While a low price elasticity of demand at the aggregated level suggests that arts organizations could theoretically raise their prices to increase their revenue, Seaman points out that arts organizations may be intentionally under charging to both encourage people at a number of income levels to attend their events and to encourage some consumers to make donations to organizations.

What it all means: There are few “axioms” with regard to understanding arts demand, but there are some trends pertaining to economic disadvantage that may be important for our understanding of access and opportunity in the healthy arts ecosystem. First, the low price elasticity of demand for arts organizations suggest that cultural institutions could be charging more for arts tickets to increase their revenue, but choose not to because they are both trying to incentivize donations and make it possible for lower income people to attend arts events. However, the idea that current price elasticity estimations are omitting “full income” and opportunity cost estimates mean that arts institutions may have more work to do to ensure that a broad range of people are able to attend arts events. Second, evidence of substitution, particularly the study that found that increased television subscription decreased arts participation, suggests that there might be some evidence to our previous hypothesis that consumers may be substituting arts attendance with television. Finally, considering attitudes and formal training in the arts may be an even more important driver of arts participation than income level, which furthers the idea that choice and lifestyle may play a greater role in predicting arts participation than any measure of socioeconomic status.