April 16, 2008

E pur si muove

I finally had time over the weekend to catch up with Robert Flanagan's report on orchestra finances, commissioned by the Mellon Foundation. It's a comic masterpiece—a hilarious parody of those diagnoses of dire orchestral health that appear like 17-year cicadas....

No, actually, it's supposed to be serious. But the report is a bit of a familiar mess. I'm beginning to think that when economists take a look at the performing arts, especially economists who don't have any experience in that sector, on some level, it fries their brain. Remember, the fundamental activity of economics isn't analysis, it's the construction of predictive models. And the default assumptions behind those models are usually based on the idea that people will always act in their own financial self-interest. Confronted with an orchestra—an organization pretty much guaranteed to lose money—economists used to for-profit assumptions often seem in danger of self-destructing, like a sentient robot confronted with a clever Captain Kirk paradox. (Robert Levine summed up the difference nicely: "Businesses make things, or provide a service, in order to make money. Orchestras make money in order to provide a service.")

Professor Flanagan retreats into his biases, which appear to be pro-management and pro-market. The first one is so transparent, it's almost ridiculous. Not only—as ICSOM's Brian Rood has pointed out—does Flanagan rely uncritically on management-supplied, non-audited financial data, he changes his methodology depending on the point he wants to make. Revenue streams, for example, are reported in real terms, in inflation-adjusted dollars. Musician salaries, on the other hand, are presented in relative terms:
Between the 1987 and 2003 seasons, the weekly salaries of symphony orchestra musicians also increased more rapidly than the average wages and salaries of other unionized workers in the United States, of nonunion workers, and of other professional service workers. (pp. 52-54)
Flanagan doesn't bother comparing that salary increase to highly-skilled workers, which is the more pertinent analogy; nor does he consider the possibility that musicians' salaries in 1987 lagged sufficiently behind those of other highly-skilled workers such that a more rapid relative increase was necessary to bring them back into line. Flanagan is also conspicuously silent on administrative salaries, even though his own graph (p. 20) shows that while artistic costs have declined as a percentage of overall expenses, general administrative costs have increased. Flanagan is quoted on the report's home page (linked above):
"Some [orchestra managers] say it doesn’t surprise them because many symphonies have a bias towards revenue growth strategies and a bias against cost-cutting strategies," Flanagan said, adding that nonprofit board members often shy away from conflict. "It’s not clear that they’re willing to be as tough minded about costs as directors in the private sector."
I was wondering when the hallowed tough-minded capitalist was going to put in an appearance. Based on his report, Flanagan's idea of a hard-nosed manager is one who cuts everybody's pay except his own.

But it's the free-market bias that's more interesting. Flanagan's main focus is the "performance income gap," the discrepancy between ticket revenues and overall expenses. Given that ticket revenues haven't covered orchestral expenses since the days when people seriously imagined flying blimps to the moon, one might think that orchestras had some awareness of this particular issue. But Flanagan pounds it home—and for a supposed "fact-finding study" (p. v), he uses some revealingly loaded language:
  • The resulting performance income gap has worsened over time and will worsen in the future. (p. v)
  • Even this group of comparatively healthy orchestras has encountered significant economic challenges, including a worsening of the performance income gap (p. vi)
  • The continuation of the historical worsening in the performance income gap (p. 22)
  • The difference in the cyclical sensitivity of revenues and expenses results in a worsening of the performance income gap (p. 27)
  • the industry confronts a worse performance income gap (p. 27)
  • In short, the majority of orchestras have continued to experience a long-term worsening of the performance income gap (p.28)
  • ...and so on. (Italics mine, except for that last one.) The factual statement is this: the so-called "performance income gap" is widening. Worsening? That's a value judgment. That "gap" has been an economic fact of life for the better part of a century, and a recognized fact of life since Baumol and Bowen analyzed it in the 1960s. Flanagan, of course, knows about the cost-disease, and, to give deserved credit, summarizes it a lot more concisely than I did:
    Variously referenced as “Baumol’s curse” or “Bowen’s disease,” the crucial facts are that labor productivity advances more rapidly in the goods-producing sector than in the performing arts (and in many other service industries), but broadly speaking, both sectors compete in the same labor markets. (p. 2)
    But the customary way for getting around the cost-disease—philanthropic support—involves what are, for all practical purposes, investments with no possibility of financial return, and that flies in the face of economic assumptions about self-interest. (There is a self-interest factor, of course, but one that's nowhere near as eminently quantifiable as money.) Flanagan can't shake the feeling that, if an organization isn't covering its costs with the direct revenue from its output, it's somehow cheating.
    Concert performance is the central objective of symphony orchestras. Addressing the relentless deficits—performance income gaps—that accompany this central activity requires attention to enhancing performance revenues and controlling performance costs. (p. 50)
    Note that if you rewrite that second sentence:
    Concert performance is the central objective of symphony orchestras. Addressing the relentless deficits—performance income gaps—that accompany this central activity requires attention to enhancing philanthropic revenues and controlling administrative costs
    ...it makes just as much economic sense as Flanagan's prescription.

    We can open up a whole can of economic worms by looking at one of Flanagan's other assumptions, one he makes almost in passing. In discussing the NEA's 2002 Survey of Public Participation in the Arts, he notes that the percentage of respondents who attend classical music or opera performances has remained pretty much the same since 1982. Good news, right? Wrong.
    It is therefore puzzling that the proportion of the population attending symphony concerts has been so stable during a period in which both real income and the proportion of the population with at least a college education increased (p. 34)
    —because historically, classical audiences have been rich and college-educated. Given the emphasis on regression analysis in the rest of the report, I waited for Flanagan to decouple those two factors, income and college education. He never does—because he hasn't bothered to factor increasing income inequality into his analysis. Aggregate real income may have increased, but middle-class wages, for example, have remained largely stagnant, even as the middle class has become increasingly college-educated. Even if you accept that a college education makes one more likely to enjoy classical music (I've never been convinced of that causality, in either direction), there's the very real possibility that the spirit is willing, but the pocketbook is weak. And rich people can only go to so many concerts.

    Just for fun, let's bring in, as befits any discussion that touches on class, the one and only Karl Marx. It struck me not long ago that the mechanism of Baumol and Bowen's cost-disease is awfully reminiscent of Marx's labor theory of value. Marx was looking to explain how capitalists made money, and he famously proposed that they do so by cashing in the surplus value created by workers, that is, the amount of labor they perform above and beyond what's necessary for their subsistence. In Marx's theory, the price of a product was the sum of three factors: constant capital (factories and machines), variable capital (labor-power sufficient for the worker's subsistence), and that surplus value. That's at least how he put it in Volume I of Capital; by Volume III, there were some difficulties to deal with, mainly the fact that most products are made from raw materials that have their own set of values, and that one couldn't necessarily assume that the price-value equality would remain constant throughout the chain of production. Marx shifted the equality to the macroeconomic level, saying that, as long as profit rates remained constant across the economy, prices and values would equal out in the aggregate, even if they were unequal within a given industry. The economist Meghnad Desai explains:
    Marx's solution was to say that money rates of profit are equalized by some commodities' prices being above value, while others would be below. Commodities which had a high organic composition of capital (ratio of constant capital to total capital) will have price/value ratio larger than one. For those with a low organic composition of capital, the price/value ratio is smaller than one.

    Another way of putting Marx's solution is to say that while living labour generates surplus-value, there is a compensation for those companies (capitals) which use more machines to labour than the average because their price/value ratio is larger than one. Labour-intensive goods cede surplus-value to machine-intensive goods via this imbalance in the pricing process.
    If that imbalance is directional over time, you've got yourself a cost-disease.

    Well, you might say, that's all very fascinating, but we were talking about the Flanagan report, remember? Bear with me—we're getting there. Desai's explanation of Marx's theory is one of the more intelligible I've ever read, unlike Marx's own explanation (part of the sheaf of notes and incomplete drafts that Engels fashioned into the second two volumes of Capital), which is impenetrable even by Marx's standards. The murkiness of Marx's writing led to all sorts of refutations and counter-refutations (though recent research has at least demonstrated the empirical plausibility of Marx's theory). But take a look at an early mathematical approach, by Ladislaus von Bortkiewicz. In 1907, Bortkiewicz realized that the fuzzy heart of Marx's theory—the "transformation problem," how to correlate values and prices—had an elegant solution under a set of certain simplifying conditions. Bortkiewicz divided all commodities into three categories: (1) means of production, (2) workers' consumption goods, and (3) capitalists' consumption goods—luxury goods, in other words. And then he related their constant capital (c), variable capital (v), and surplus value (s) like this:
    Three equations, three variables—that means there's a solution. Caveats: Bortkiewicz's approach is, again, wildly simplified (later on, the transformation problem would be tackled in the general case with linear and matrix algebra), and by no means uncontroversial (even some Marxists consider it far too distorted a caricature)—and the whole idea of trying to equate value and price leaves some profoundly unconvinced (both the Marxist Joan Robinson and the anti-Marxist Paul Samuelson found the transformation problem ultimately pointless). But remember what Marx was doing: trying to see capitalism in the big picture. And take another look at that last equation: the total value of luxury goods is equal to (or at least proportional to) the total surplus value created in the entire economy. And then consider that music is, after all, a luxury good—and that classical music is hardly the only genre feeling the pinch these days.

    If you buy Marx's analysis (and I think he was on to something), then Bortkiewicz's simple model hints at how the available capital for luxury goods would be squeezed as income inequality increases—if wages remain stagnant, i.e., don't keep pace with inflation, then more labor is required to reach subsistence levels, leaving less available for surplus value. And here's where Flanagan (see, I told you we'd make it back) doesn't take a wide enough view. The point of the report is to separate cyclical factors (economic factors that orchestras have no control over) from structural trends (economic factors brought upon orchestras themselves by their own policies). Flanagan clearly regards the flat attendance among affluent-and-educated potential concertgoers as a structural trend—a failure of programming and marketing. But by failing to separate out education from income (come on—any graduate music student can anecdotally confirm a very weak correlation) he misses the possibility that such data is cyclical, the result of external economic conditions, in which case the proper prescription would quite possibly be an increased focus on fundraising and development—go after the only sector with sufficient surplus—the opposite of what the report recommends.

    This may seem like a minor, if long-winded, point. I don't think so—I think it gets at the heart of what's wrong with the report (and many like it). Marx's guest appearance may seem either oblique or jarring—especially if you're rabidly laissez-faire or prone to confuse Marx with "Marxism"—but I like having him here because his holistic intellectual cast, trying to analyze the entirety of society, is exactly what the report needs more of. Flanagan expends much effort in trying to isolate possible factors—but never considers how those factors might interact. (And when he can't isolate factors—as with unemployment rates and stock prices—he sort of just throws up his hands.) Most importantly—and this is getting downright endemic among analyses of The State Of Classical Music—there's paltry indication of anything beyond the largest organizations. String quartets, entrepreneurial chamber ensembles, new-music groups, community choirs, we'll-try-anything festivals: perhaps the presence and efforts of all those musicians might have some salutary effect on the viability of orchestras? (Sure, that might not be reflected in the orchestras' own data, but that didn't stop Flanagan from, for example, pulling in the NEA study.)

    In other words, the report doesn't put orchestras into a sufficiently community- or society-wide context such that an orchestra does make economic sense. Orchestras exist, after all, and persist even as report after report predicts their downfall. It's like plugging values into the Drake equation that make even intelligent life on Earth improbable. The normal plea at this point would be for an economic analysis that focuses not on why orchestras will disappear, but why they continue to survive. But, honestly, I enjoy reading demonstrations of the impossibility of heavier-than-air flight or a sub-four-minute-mile. How long until the next one?


    Daniel Wolf said...

    Well done, Matthew. Let me add one caveat to the assumptions about the "educated audience". I believe that the changes in the US audience for classical music are closely connected to immigration patterns -- the audience for the "big five" and other urban orchestras of the middle of the last century had a large component of first generation European immigrants; these audiences were cultivated consumers of the European tradition, but not necessarily formally educated in that tradition. The most dynamic developments -- both in growth and restructuring (including, in some cases, closures) -- of large musical institutions have been happening in communities, particularly in the West, in which European immigration is several generations past. While immigrants from other regions play some role in audience development, the bulk of the local audience is upwardly mobile and formally educated, but with an education that is not necessarily invested in a particular European canon, for whom Haydn and Bartok are equally new, and whose concert-going loyalties may well include as much pop music as classical.

    Lisa Hirsch said...

    Wow. How many econ classes did you take in college??

    Have you read Sandow's defense of Flanagan? I am about to.

    Anonymous said...

    Bravo. That's not only a very cogent argument about the economics of the concert music business, it's also the clearest summary of the Volume Three transformation problem I've ever read.

    Matthew said...

    Daniel: that's a really interesting idea. My own experience (third-generation on both sides) certainly bears it out—I came to classical music because I heard it and liked it, not because I had a cultural affinity or obligation towards it. It strikes me that internal migration might have an equal effect on the way classical music organizations have evolved in different regions of the US—start with the overgeneralized caricature of "farm kids move to the city, retirees move away from it" and you can imagine how that may have affected the expectations of particular audiences.

    Lisa: How many econ classes did you take in college?? A couple. Econ is, happily, one of those subjects where the math only rarely gets beyond differential equations, and even then, you can still usually at least follow the spirit of the argument.

    I did read Greg's defense, which was reasonable, but I think he missed the point of a lot of the criticism—most people weren't dismissing Baumol and Bowen, they were noting that Flanagan's recommendations don't do a thing to ameliorate it—in fact, they only kick down the road any reckoning with the cost-disease. (I did laugh when he got to his "non-economists shouldn't critique economics" part—I took pride in my perpetual foolhardiness.)

    Cosmo: To be fair, I cribbed a lot from both Desai and another seminal 1950s paper by Ronald Meek (who also breaks down Bortkiewicz's solution in more-or-less plain English). Another (far more critical) paper worth seeking out is Samuelson's 1971 review of the problem (summarizing his 1970 mathematical analysis), even if you don't have the math—his elegant bitchiness is so entertaining, I read the thing twice so as not to miss a quip.

    Guilherme da Fonseca-Statter said...

    Small world... Or the wonders of the www...
    I stumbled upon this blog as I am putting a few final touches on a book on the "falling tendency of the rate of profit", as explained by Marx.
    How come? One might ask...
    Well, as the mainstream economists contend that there is no such thing as a "falling tendency" in the capitalist market economy, I was planning the title to be "E PUR SI MUOVE"... eh eh eh
    The publisher is of the opinion - conveyed to me this very afternoon - that, if we want to sell a bit more that a few hundred copies - I should look for a more #catchy" title.
    Incidentally although I am a musical illiterate /cannot read music..) I love MUSIC... And the editor/publisher has a degree in Musicology...
    Also incidentally, the "transformation problem" has been misunderstood by many, many critics. In a nutshell, Marx was not trying to prove anything, He was merely trying to show how David Ricardo had missed the role of prices in the re-distribution of surplus valur amongst the varius groups of capitalists. In other words, the "transformation problem" was a ricardian problem, not a marxian problem.
    And my warmest compliments for the clarity and breath of your argument...
    Guilherme (Lisbon, Portugal)