Rockonomics

by Alan B Krueger

  The underlying concert data are reported to Pollstar magazine by venues, promoters, and managers. The reported information consists of the date of the event, the headline act and opening acts, the promoters, the name and location of the venue, box office revenue, number of tickets sold, and high and low prices. In this Appendix, I compare the Pollstar data for a subset of concerts to more accurate and comprehensive data from other sources. The main conclusion of my analysis is that although there are some lapses in the coverage of the Pollstar data, the Pollstar Boxoffice Database provides reasonably comprehensive, accurate, and reliable data on concert revenue for top artists, especially since the 1990s and in the United States.

  A comparison of the Pollstar data to other sources is possible because (1) the website Setlist.fm crowdsources information on concerts for major performers and (2) a large state university that operates two sizable arenas shared its confidential box office data with me on the condition that the name of the university remain anonymous.1 These two arenas do not directly report information to Pollstar, but managers and promoters of shows performed at the venues often do report the data.

  Both of these alternative sources have their own strengths and weaknesses: Setlist.fm lacks information on prices, attendance, and revenue, but it is quite comprehensive for top artists. The large state university represents only two venues that do not report data directly to Pollstar, but it provided me with unassailable box office records on all of its shows. Taken together, the two sources help paint a picture of the coverage, accuracy, and reliability of the Pollstar data. I first considered the coverage of the Pollstar data by comparing it to Setlist.fm, and then I turned to the accuracy and reliability of the price and revenue data.

  Evidence on Coverage

  This analysis focuses on twenty-six major acts who headlined more than 23,000 shows reported to Setlist.fm since the early 1980s. Specifically, the artists are Beyoncé, Billy Joel, Bon Jovi, Britney Spears, Bruce Springsteen, Bruno Mars, Cher, Coldplay, Céline Dion, Dave Matthews Band, Jay-Z, John Mayer, Justin Bieber, Kenny Chesney, Lady Gaga, Madonna, Mariah Carey, Maroon 5, Metallica, Pink, Rihanna, Shania Twain, Taylor Swift, U2, Usher, and Whitney Houston. I scraped data on the date, location, and headline act for each of their shows from Setlist.fm. Once the data were collected, an effort was made to exclude award shows, TV appearances, MTV, and other non-concerts from the Setlist.fm data. Nevertheless, it is possible that some non-concerts remain in the Setlist.fm list, which would make the Pollstar data appear less comprehensive than is actually the case. The Pollstar sample of concerts is almost a proper subset of shows reported to Setlist.fm, which suggests that Setlist.fm covers close to the entire universe of concerts.

  Table A.1 reports the number of concerts contained in the Pollstar database as a proportion of the number of concerts in Setlist.fm for those twenty-six major headliners. The results are presented separately for events in the United States (column 3) and for all countries in which Pollstar and Setlist have data (column 6). If the Setlist.fm compilation of concerts is taken as the universe of all concerts, then the Pollstar data captured more than 80 percent of concerts performed by top acts in the United States in the 2000s. The proportion of concerts covered by Pollstar slipped for a period when Live Nation ceased reporting to Pollstar, though it has since resumed reporting. Pollstar’s coverage outside the United States has improved each decade and currently approaches the coverage rate of concerts in the United States.

  Table A.1: Number of Concerts in Setlist.fm and Pollstar for 26 Headliners

  Notes: Author’s analysis of Pollstar Boxoffice Database and Setlist.fm. See text for list of artists. An effort was made to remove non-concerts from the Setlist.fm list of events.

  Evidence on Accuracy and Reliability

  Even though the operator of the two venues at the large state university that shared data with me does not report data directly to Pollstar, artist managers and promoters of shows performed at the venues can report the data to Pollstar. The university shared data with me on live events held at its venues from mid-2010 to the end of 2017. In this period, there were 135 concerts that appeared in the Pollstar Boxoffice Database. This represents 57 percent of the total number of concerts that were performed in the two venues, which not surprisingly is less comprehensive than the Pollstar data overall, given that this venue operator does not report directly to Pollstar.

  Figure A.1: Revenue According to University vs. Revenue According to Pollstar (Millions of Dollars)

  It is reassuring that the data for these 135 concerts reported to Pollstar were quite similar, on average, to the data for those same shows that the large state university provided. Revenue per show was $801,086 according to Pollstar and $776,170 according to the university data, a discrepancy of just 3 percent for the 135 shows in both data sets. The average number of tickets sold was 11,324 according to Pollstar and 11,191 according to the university records, yielding a discrepancy of just 1 percent. And the average price (derived as revenue per number of tickets sold for each show) was $70.74, according to Pollstar, and $69.36, according to the university records. Surprisingly, there was a tendency for shows with more tickets sold to be underreported to Pollstar. Nonetheless, this comparison suggests that the data reported to Pollstar are quite accurate.

  The Pollstar data are also remarkably highly correlated with the actual box office data. Figure A.1 provides a scatter diagram of revenue per show as provided by the university (vertical axis) and as reported to Pollstar (horizontal axis). The correlation between the two measures is 0.99, and a fitted line is close to a 45-degree line, which implies that in most cases the revenue figures reported to Pollstar were almost identical to the revenue data recorded by the university. Interestingly, in the handful of outlying data points, the deviations from the fitted line tend to be below the line, which suggests that the revenue figures reported to Pollstar were occasionally exaggerated by managers or promoters.

  Figure A.2: Tickets Sold According to University vs Tickets Sold According to Pollstar

  Figure A.2 shows there is also a tight correspondence between the number of tickets sold per show as reported by the university and as reported in the Pollstar data. (Note that the one show with 30,000 tickets sold represented two performances by the same performer on adjacent days combined.) And the average price per show is similar in both the Pollstar and university data. The correlation between the two sources is 0.99 for both tickets sold and average price, suggesting that the data available in the Pollstar Boxoffice Database are highly reliable.

  Lastly, the university-reported data and Pollstar data also display similar trends over time. From 2011 to 2017, for example, the average ticket price rose by 27 percent for shows held at the two venues according to the Pollstar data, while the average ticket price for the complete universe of shows held at the two venues increased by 24 percent according to the university data. The correlation between the annual average ticket price calculated with the Pollstar data and the university data from 2010 to 2017 is 0.93.

  Cleaning the Pollstar Data

  The Pollstar data include some events that are not concerts, such as comedy shows, traveling Broadway shows, and sporting events. I made an effort to remove non-concerts from the database. In addition, I recoded names reported for some artists to make them consistent (e.g., consistent use of ampersands) across records. Revenue from festivals and joint performances (with more than one headliner) was distributed evenly among the various headliners.

  Although box office data on a majority of the major concerts held in the United States and around the world are reported to the Pollstar Boxoffice Database, in early years shows were more likely to be underreported, and even in recent years the database is not complete. Missing data on concerts is a concern when it comes to examining multi-decade trends in total revenue for individual artists. Consequently, to derive the figures on artists’ cumulative revenue
in Figures 4.3a and 4.3b in Chapter 4, the Pollstar data were scaled up according to the inverse of the share of each artist’s concerts that were not reported to Pollstar, using the artist-specific ratio based on the Setlist.fm and Pollstar data in the relevant decade. If ticket revenue earned in the excluded shows, on average, equals the revenue earned in the included shows, this approach would correct for incomplete reporting of concerts.

  Conclusion

  The Pollstar sample does not contain the universe of concerts, but it represents the lion’s share of concerts performed by top headliners in recent years. Incomplete coverage is not necessarily problematic for determining trends in average prices or the distribution of revenue because shows may be missing at random and because the available data appear to be accurate and reliable. But relying on Pollstar figures alone to determine the size of the concert industry as a whole or individual artists’ total income results in underestimates, because approximately 20 percent of shows over the past two decades were not reported to Pollstar.

  * These terms have a specific statistical meaning. Coverage refers to the share of concerts contained in the Pollstar data; accuracy refers to whether statistics computed from the data tend to be correct, on average; and reliability refers to the strength of correlation between the data reported to Pollstar and the actual box office records.

  ACKNOWLEDGMENTS

  I have a confession to make: I am only an average music fan. That means that I listen to music about a couple of hours a day. True, I have gone to more than my share of concerts—starting with Billy Joel, Judy Collins, Steve Miller, George Benson, and Henry Mancini in the late 1970s—and I played the drums (very poorly)in the orchestra until eighth grade. But I am not a fanatic. I like to sing along and tap my feet to the rhythm, but I am hardly a music encyclopedia.

  Throughout my career, I have followed U3 (the official unemployment rate) much more closely than U2 (the Irish rock band). But having some remove from a subject can help to understand it better, or at least to see the big picture, which I hope to have conveyed in this book. And the fact that music has played such an integral role in my life, the life of an average fan, suggests that it is at least as important in the lives of millions of readers.

  As I reflect on it, music has had a profound impact on my life. I listened to music in my bedroom while trying to figure out what I wanted to be when I grew up, and how to fit in. I listened to music when I exercised, and when I took the bus home from sporting events. I danced to music at my prom and at my wedding. Like many others in my generation, I quoted song lyrics in my high school yearbook that were particularly meaningful to me. I chose the prologue to one of ABBA’s least popular songs, “Move On”:

  If I explore the heavens, or if I search inside

  Well, it really doesn’t matter as long as I can tell myself I’ve always tried

  That sentiment has been something of an inspiration and beacon for me. This book is a product of that journey.

  I get by with more than a little help from my family, friends, research assistants, and students. I could not have written this book without the assistance, encouragement, and advice of many others. My wife, Lisa, read every chapter and encouraged the project from the start. Rafe Sagalyn, my book agent and friend, enthusiastically encouraged me to write about rockonomics. My editor, Roger Scholl, instantly saw what I was trying to achieve and guided me through the process of writing a trade book. Erin Little at Penguin Random House skillfully assisted with the production of this book.

  Several research assistants helped with data analysis and other research tasks. While training for the U.S. Olympic rowing team, Max Meyer-Bosse assisted with the analysis of the Pollstar Boxoffice Database. Amy Wickett and James Reeves expertly provided assistance analyzing streaming data, tabulated ACS and Billboard data, and performed countless other tasks, from making graphs to running down footnotes. Grace Kwon, Brendan Wang, and Jun Ho Choi spent a summer tracking down the backgrounds of top musicians.

  Princeton University has been a supportive employer and environment for my research for three decades. Ed Freeland of the Princeton University Research Center helped design and carry out the MIRA Musician Survey.

  My brother, Richard Krueger, and childhood friend Jeff Kolton provided feedback on several chapters and tutored me on the intricacies of music, musicians, and bands from their extensive experience.

  My children, Ben and Sydney, read several chapters of the manuscript and provided valuable advice. They are also responsible for reinvigorating my interest in live music, when they lobbied to see *NSYNC perform many years ago. Sydney’s Columbia classmate Paul Bloom provided an informative summary of research on the neuroscience of music.

  Several people in the music industry provided me with data. Gary Smith and Gary Bongiovanni, the founders of Pollstar, were particularly generous in giving me access to the Pollstar Boxoffice Database, and the new owners of Pollstar honored their arrangement. Tinko Georgiev of Kanjian provided streaming data and tutored me on the Chinese music market. David Price of IFPI provided me with the latest international data on recorded music revenue. Marco Rinaldo, Karolina Bressan, and Andrea Barattini generously provided streaming data from Rehegoo.

  Too many people in the music business to mention individually tutored me on the inner workings of the business, but I want to especially acknowledge Adam Fell, Quincy Jones, John Eastman, Eric Berman, Don Passman, Marty Gottesman, and Rob Levine for their advice and help. John Cappo, Eric de Fontenay, Qingman Liu, and J. P. Mei helped me to understand the Chinese music market. And I am particularly grateful to Cliff Burnstein and Peter Mensch for sharing their stories and insights, arranging for me to see Metallica from the mixing station, and encouraging my research. I hope to join them at spring training next year.

  I learned a tremendous amount that is reflected here from my partners in the Music Industry Research Association, including Joel Waldfogel, Alan Sorensen, Marie Connolly, Dan Ryan, Julie Holland Mortimer, and Jeannie Wilkinson. And Per Krusell tutored me on the Swedish music scene. Dan Pink helped me organize my thoughts. My colleague Orley Ashenfelter provided encouragement and advice, and paved the way for my work in this field with his research on wine economics. Amy Brundage and Sam Michel provided valuable feedback at various steps along the way of this project.

  The incredibly talented and dedicated musicians Max Weinberg, Steve Ferrone, Dan Wilson, Jacob Collier, and Gloria Estefan generously and patiently gave me a window into their lives and craft. They inspire me and their fans. I have tried my best to represent their stories and the economics of their industry as faithfully as possible.

  NOTES

  Chapter 1: PRELUDE

  1. See Mikal Gilmore, “Why the Beatles Broke Up,” Rolling Stone, Sep. 3, 2009.

  2. Jon Pareles, “David Bowie: 21st Century Entrepreneur,” New York Times, Jun. 9, 2002.

  3. Mark Mulligan, “Why Profit Doesn’t Come into It for Apple Music,” Music Industry Blog, Aug. 28, 2015.

  4. Barbara Charone, “Ladies and Gentlemen…Emerson, Lake & Palmer,” Gig Magazine, Sep. 1977, http://ladiesofthelake.com/​cabinet/​77Tour.html.

  5. This quote comes from an interview with John Eastman on Mar. 2, 2018, in New York City.

  6. Neil Young, “My My, Hey Hey (Into the Black),” Rust Never Sleeps, Reprise Records, 1979.

  7. Chris Moukarbel, dir., Gaga: Five Foot Two, Netflix, 2017.

  8. David J. Hargreaves, “The Effects of Repetition on Liking for Music,” Journal of Research in Music Education 32, no. 1 (1984): 35–47; Robert B. Zajonc, “Attitudinal Effects of Mere Exposure,” Journal of Personality and Social Psychology 9, no. 2, pt. 2 (1968): 1–27.

  9. Tasneem Chipty, “FCC Media Ownership Study #5: Station Ownership and Programming in Radio,” Federal Communications Commission, Jun. 24, 2007, https://docs.fcc.gov/​public/​attachments/​DA-07-3470A6.pdf.

  10.
Gary Trust, “Ask Billboard: Max Martin Notches Another No. 1,” Billboard, Nov. 25, 2014; Ed Christman, “Billboard’s 2018 Money Makers: 50 Highest-Paid Musicians,” Billboard, Jul. 20, 2018.

  11. Marc Myers, “Steely Dan’s Donald Fagen Tours with Young ‘Nightflyers,’ ” Wall Street Journal, Jul. 26, 2017.

  12. Alfred Marshall, Principles of Economics (New York: Macmillan, 1947).

  13. Chris Anderson, The Long Tail: Why the Future of Business Is Selling Less of More (New York: Hyperion, 2006).

  14. Author’s calculations using information from Pollstar Boxoffice Database.

  15. Facundo Alvaredo, Lucas Chancel, Thomas Piketty, Emmanuel Saez, and Gabriel Zucman, “World Inequality Report,” World Inequality Lab, 2018.

  16. Gerry Mullany, “World’s 8 Richest Have as Much Wealth as Bottom Half, Oxfam Says,” New York Times, Jan. 16, 2017.

  17. Michael Wheeler, “The Luck Factor in Great Decisions,” Harvard Business Review, Nov. 18, 2013.

  18. See Chapter 6.

  19. “Concert Tickets’ Expensive Rocknomics: Our View,” USA Today, Aug. 8, 2013.

  20. Richard E. Caves, Creative Industries: Contracts Between Art and Commerce (Cambridge, MA: Harvard University Press, 2002). See also Peter Tschmuck, The Economics of Music (New Castle upon Tyne: Agenda Publishing, 2017).