The music market

Matt Yglesias gets the key fact about recorded music. It’s a non-rival good:  when I play an mp3 file of a song, there’s no less of it for you, so the marginal cost of my consumption is zero.  He also gets the profoundly illuminating and useful principle that everything should be sold at marginal cost.  From this, and I note that it’s a short post and he doesn’t undertake to get much into the weeds, he tries to turn down the alarm volume over the 50% decline in recorded music sales (correctly distinguishing it from listening) since 2001.

However, he slips in this reassurance

Thanks to copyright, a recordings-seller does have some level of market power to allow him to seek monopoly rents. But there’s a pretty high degree of substitutability between different songs, so the competition is still pretty intense and the prices are low.

which is completely wrong in two ways. Continue reading “The music market”

Two big industries that don’t understand their business

What does cheating in college have to do with markets for digital goods?  More than you might think, and two links connect this weekend’s report from the battlefield between professors and students and a book review by the interesting and insightful jazz critic Devin Leonard.

The cheating story is profoundly depressing; the University of Central Florida has a “testing center” with snooping video cameras, keylogger software, dated scratch paper, and the like, all intended to drive cheating below its current intolerable, flagrant, pervasive rate of … of 14 suspected incidents out of 64,000 exams??!!  The article details an arms race between students widely and blithely determined to cheat and professors happily leaping into a technical battle to nail them for it.

What no-one in the story seems to get is how off-the-rails the institution must have gotten for students to believe that the point of college is to submit right answers and get grades. The easiest way to do this, a game that has precisely nothing to do with, um, learning, is to find such answers the fastest way possible (for example, from classmates with whom you divide up a problem set, or on the internet).  The pettiness of the colleges’ position is astonishing. For example, and only one:  plagiarism, operationally, is copying text, while copying ideas and rephrasing them is not pursued. Should the difference between an A and an F really be rewriting a sentence someone else wrote so  Turnitin doesn’t find it? Might students have gotten the idea that parroting the thinking of others is what college is about from profs who reward them for repeating back true propositions they ladled out in lectures? Cheating is a big deal, but responding to it at retail and confrontationally is just wilful blindness to the problem cheating really indicates – and not incidentally, allowing ourselves to play a game of mutual infantilization.

Leonard reviews a biography of Edgar Bronfman, Jr. (who bought Warner Music six years ago [yes, those last three words appear in just that order in Leonard’s piece; plagiarism!]) as a frame for reflections about the music industry. This industry, and Leonard, I think,  don’t understand the real business of music, just as the colleges have lost touch with their real value creation model. Music executives think they are in business to sell physical objects like CDs, or countable mp3 copies of performances, to listeners, and have chosen to do battle on a field of piracy and sharing suppression.  But that business model was an accident of technology history, like the temporary conditions during which newspapers could sell readers to advertisers.  The value music companies add to musicians and listeners is search efficiency, helping you find something you will be glad you heard in a universe of possibilities you don’t have time to begin to browse in, but they are leaving that whole territory, the only one in which they can conceivably survive, to peripheral players like Pandora and Rhapsody because they are too scared, or maybe just too dumb, I guess, to really think how they can monetize that service. Leonard, by the way, correctly recognizes that this mental straitjacket has not only cost the labels and artists a lot of money, but also impoverished consumers with “all Lady Gaga all the time”.

A big problem for them, of course, is the universal implicit recognition by music listeners that a digital recording is a non-rival good: because there’s no less of it for anyone else if I copy a song for you, “it just doesn’t feel like stealing” and in fact its efficient price is zero, its marginal cost.  Of course their response to this recognition has been spectacularly dysfunctional and maladaptive, suing downloaders and messing with hopeless DRM technologies. This divide between the moral stance and perceptions of provider and customer matches what’s happening in the colleges: students are comfortable finding useful stuff on the whole internet and, I think, have a sense of the non-rival property of all digital discourse, including text, close to the way they see recordings.

And they have a point: grownups in any workplace do not write closed-book memos with ballpoints in bluebooks, sitting in a room with cops watching to be sure they don’t learn anything from anyone else: they vacuum up everything they can get from peers, media, and the like to find a solution to a problem they know they have.  Of course, it’s wrong – and known to be wrong – to steal that work from the guy in the next cubicle and offer it as your own.  But as Lauren Resnick immortally observed, “collaboration in a workplace is essential and rewarded; collaboration in school is cheating and punished.” Students know this and it doesn’t help the college to get up on a soapbox and preach; the students see a lot of the hoops they are expected to jump through as unrealistic artificial exercises set to maximize the comfort of a bunch of stuffed shirts from another era, not to enable them (the students) to create value in the world. Possibly excepting those very few who will create value as professors, but almost no students are baby professors, or apprentice professors.  If we don’t learn to discern intellectual merit among our students (and other kinds – Howard Gardner describes eight kinds of intelligence, of which maybe three will help your grades the way we operate now)  in a more adult and realistic fashion, we are fighting the wrong war with the wrong weapons on the wrong field against the wrong ‘enemy’.

Handwringing is actually different from knowing something

Megan McArdle is the business and economics [sic] editor  of the Atlantic. She and I have crossed swords mice before and if she recalls those debates at all probably thinks I have no manners and don’t like her.  As to the first, I do wish I were more gracious in person and in print, and as to the second, I don’t think we’ve ever met and presume she’s a perfectly nice person; she is certainly decent and just.  She has more formal economics training than I have (an actual MBA) and beats me on credentials right off the bat.

Unfortunately, her current column,  on the business prospects and options for digital media requires attention both because (i) the topic is really, really, important and (ii) it’s in a widely read publication, and because rather than shedding the least ray of light, it raises serious questions about her employer’s understanding of the words business and economics.

She records a bunch of anecdotes about how hard it is to make a living generating music, video, or text these days, and a few more about folks who have managed to scrape by in niches, or issuing recordings effectively free but charging for tickets at live events.  She rails that we have become ethically unmoored (stealing content) along with a little potted psychology. But she seems to be completely innocent of the most basic economic ideas that could actually illuminate her topic for an educated reader, and to have read absolutely nothing written by real, um, economists or other experts in this field. There is a large literature in copyright and intellectual property (if you haven’t met Lessig, Fisher, Towse, Scotchmer, and Pam Samuelson you really have a nerve writing about this).

Her theme sentence is “Maybe it’s time to admit that we may never find a way to reconcile consumers who want free entertainment with creators who want to get paid.” Only a stone [deaf] libertarian, who believes market failures are imaginary, could write that sentence and not think, like instantly, “does this sentence make any sense at all if I plug policing, parks, or clean air to breathe into it in place of entertainment?”

OK, let’s do it again: digital content is a non-rival good. If I consume it, there’s no less for you, so its marginal cost on the consumption side is zero. It’s like a spot on the grass in the park, or a (literal) breath of fresh air. The efficient price of anything to consumers is marginal cost, period, end of story, line feed, -30-. This is what people who download music understand: it doesn’t feel like stealing because it doesn’t leave any less for anyone else.

But oh dear and oh gosh, how can we possibly provide something that’s costly to make free to its users – has a civilized society ever done anything so difficult? Um, yes, lots of stuff (cf parks, defense, the Smithsonian museums, streets to walk on, foundation grants, on and on). What I think makes McArdle crash around in the dark bumping into things here is that the mechanism by which we do this is (outrage alert, not safe for work at AEI, faint of heart stop here!) government, sometimes including its distinctively American non-profit arm. If you think government is ab initio a moral outrage and a tool of the devil, you simply should not be writing for pay about digital media (or a lot of other stuff). Market failures are technical properties of certain goods, not ideological choices or figments of an overheated lefty’s imagination. If you don’t understand that, you just have to stay in chapters 1 to 3 of the textbook and stick to guns and butter and stuff that can be put in a burglar’s sack and carried away; you will get a lot of egg on your face and a terrible headache if you get into environment, education, defense, health policy, crime, recordings, and all that ill-behaved stuff.

Machinery to give away music, video and text, but still pay artists and other creators properly (that is, according to the value they create so they get the right signals about what to do more of) is not trivial to design, especially because unlike a park, we don’t want the government making quality judgments about it, nor observing what individuals are consuming.  But it’s not impossible, and Terry Fisher has laid out the basics in a scheme that could be made entirely workable (not perfect) with recent advances in hashcoding identification like 411-song or Shazaam.

The desperate desire of publishers, record companies, and their ilk to keep rents they have come to enjoy, and to make us treat non-rival goods like a ton of coal or an oil painting even though they are nothing like those things, is not only impoverishing all of us in the ways McArdle worries about, but in combination with the use controls technology provides, doing a lot more, and more serious damage. What libertarians could usefully get exercised about in this area is the assaults on individual freedom that outrage Netanel and Lessig, not a bunch of kids who intuitively understand marginal cost pricing downloading their music.

Financing drug and medical device R&D

We should run parallel systems: the current system driven by the promise of monopoly rents via patent, and a new system of direct public expenditure via public-benefit corporations, grants, and prizes.

Megan McArdle worries that, since some of the absurdly exorbitant cost of health care in the United States goes to finance research and development into new drugs and devices, cutting back on that expenditure is likely to cut into the R&D activity. It’s hard to argue that her worry is unreasonable, though the quantitative relationship is hard to guess.

I’d put the case more strongly: in effect, the United States is helping to finance the development of knowledge that is a global public good. (It’s odd that the people who think it horribly unjust that the U.S. pay more than a fair share of the costs of protecting the planet from global warming are fine with having us pay more than a fair share of the costs of medical advancement, and vice versa.)

John Holbo responds: Yes, but pumping money into a wasteful health care system in hopes that some of it will trickle through to R&D is a classic case of feeding the horse in order to feed the sparrows. If we need more R&D for health technology, why not pay for it publicly? After all, we already spend $30B on the National Institutes of Health.

Megan replies that she’s willing to try it – she mentions the use of prizes, in lieu of patents, which seems like a great idea – try it, but not willing to give up the profit incentive for health R&D unless and until we know that the public version works with comparable efficacy.

Fair enough. Presumably you want both systems at work: no public agency is going to spend money developing a cure for male pattern baldness. But what I don’t see is the argument that we have to wait. It’s not as if long-term changes in health care revenue streams mean an immediate shut-down in the new-drug pipeline. Why not try to squeeze some of the quasi-rents out of current pharmaceutical pricing while at the same time putting money into drug development: via one or (preferably) more public-benefit corporations devoted to developing new drugs and putting them on the market at roughly the marginal cost of production and distribution, via grants to existing or newly-created medical-research outfits, and via prizes for specific accomplishments? (A billion dollars for a malaria vaccine with 90% efficacy and no greater side effects than the tetanus vaccine? A bargain!)

Then we can look up in five or ten years and see if the changes, net, have so reduced the flow of new technology development that we need either more public R&D or greater market incentive through patent-generated monopoly rents. The basic economic point is that monopoly pricing is grossly inefficient, since the gain to the manufacturer is smaller than the losses to consumers. So the government-based mechanisms could operate far from optimally and still outperform the current system.

And while we’re at it, we might ask the EU and the rest of the G-20 to start spending as large a share of GDP on medical research as we do.

Newspapers, music, video, transit, and parks

The news news is bad news: my hometown paper is breaking up on the rocks, the LA Times is a shadow of its former self, the Rocky Mountain News is a memory, and the New York Times is on very soft financial ground. Almost every American city is a one-newspaper town, and those surviving papers are firing reporters left and right, closing capital and foreign bureaus, cutting features, and generally providing less content of lower quality. This is not, in my view, a Schumpeterian story about laudable industrial change with unavoidable Darwinian debris. News and comment provided by informed experts, at enough length to capture subtlety and complexity, are not just another product with substitutes: the Daily Show is not news, and even the Sunday morning talk shows and the Keith and Rachel hours are nothing like a real newspaper.

The social and political risks of journalistic meltdown are a lot more like social climate change: very big, and very scary, especially as the next few years are going to face us with choices from the local to the national level that are complicated, high-stakes, and not adequately guided by easy ideological bromides or slogans. Howard Kurtz provides a current roundup of misery in the WaPo today; Paul Starr has a long, reflective piece in TNR this week.

Just to give the problem some human scale, consider: the San Francisco Chronicle now has a circulation of about 340,000; assume generously another 300,000 on-line (not counting people who read both versions) and is losing $50m a year. So its readers could have at least the current, slimmed-down Chron for another $80/year each (paper subscribers pay $200 +/yr). All very well, but very few people want to read only the Chronicle; one of the wonders of the internet is that we can read newspapers all over the country and the world, and should. I dip into at least ten newspapers in the course of a month, and $800/year is not realistic.

Several stories compete to be ‘the’ diagnosis:

(1) Craig’s list and EBay are much better than classifieds, so they won and the most profitable advertising type died. Google is better for finding when the movie starts and where it’s playing, or whether Macy’s has shirts on sale this week, so other advertising left the building by the thousands of square feet.

(2) The important change is merely from paper to the screen you’re reading now, and journalism is just having teething problems learning to adapt.

(3) (a) People are no damn good (any more); lazy and ignorant, young people especially don’t care and won’t really read (b) …and they won’t pay what they should be willing to for news.

(4) The conglomerates and chains that bought all the papers cut their own throats by short-sighted cost-cutting on the editorial side, and flaccid sales effort on advertising, so the product is ruined and people rightly reject it.

The answer, of course, is (5), all of the above (in some form, I guess less snarky for (3)), plus something more fundamental: the efficient price (marginal cost) for anything that is delivered in digital form, like writing, is zero. The great challenge is to find socio/politico/economic machinery that meets three essential criteria:

(1) It will bridge the gap between that price and the real economic costs (like salaries and travel expenses for writers) of the product. Like the other declining marginal cost goods in the header of this post, writing can only be provided efficiently (enough and the right quality) by government, which raises the next criterion:

(2) It has to be accomplished without a Ministry of Information, censorship, or snooping on readers to see what they’re into.

And we don’t just want a bunch of words as cheap as possible, we want the best words, which means

(3) It has to reward creators differentially according to the value their work creates.

(One more complication, which I leave for another post, is the need for efficient search and selection. In a completely free internet world, allocating finite attention time (the only, but critical, cost of consuming news) is crippled by our inability to know in advance what’s worth attending to. Newspapers (all kinds of publishers) used to do this sorting and evaluation for us, and it’s still true that a byline story in the New York Times has, and should have, a lot more weight than a post on this blog. If something like the Times can’t make money providing this selection and quality assurance, nothing will.)

Starr begins to explore the idea of nonprofit charitable supply, or a sort of hands-off government-funded endowment of non-profit news disseminators. I can’t begin to list the obvious defects and risks of these schemes (not that they are fatal defects), but consider, as to the first, who has enough money to endow such charities, and do those people mirror society’s interests in emphasis, focus, issues, and policy? …and as to the second, what mechanism would direct the performance of these endowed bodies to behave in even a vaguely efficient manner?

There is a model, much underattended, originally designed for recorded music by Terry Fisher and described in his book Promises to Keep, with a lot of useful commentary in Lawrence Lessig’s essential Free Culture . The key idea is an appropriation to the copyright office of the whole cost of newspapers, adjusted annually (for music, taking the entire sales of the music industry, it’s about $30 per capita). When a page or article is hit, a counter clicks in the copyright office (without recording who caused the click) and at the end of the year, the appropriation is divided in proportion to clicks. This takes page hits as a proxy for value created; not perfect, but not too bad. It’s government funding, in a sense, but with any bureaucratic or political judgment excluded; if you post something on the web and you get hits, you get paid; more hits, more pay. I could not make a living from this blogging, but Paul Krugman (yes, and George Will) will do nicely…and a lot of ink-stained wretches in between will put food on the table and see very good incentives as to what to write about, and how to write better.

This is a revolution, of course, and will definitely leave a lot of Darwinian debris. But I haven’t come across another way. Weird and scary as it sounds, it’s time for (mechanistic, bureaucratic, judgment-free) socialization of the news.

Digital media again/still

The crisis in print journalism seems to be worsening, or at least ripening. It’s not just the incredible shrinking daily paper, but monthly magazines (including Wired, so it’s not outdated content) and it’s not just periodicals but books, and not just print media but music and video.

Us shadetree mechanics always look for one specific failure when a machine acts up: because the odds are so small that you will burn a valve at exactly the same time your timing belt slips a tooth, it’s a reasonable heuristic. Same in medicine: docs look for a particular morbidity that could cause the set of symptoms presented, not simultaneous attacks of flu, shingles, measles and hangnails. A lot of discussion of the media meltdown implicitly adopts this heuristic, asserting this or that single core problem, but here I think we have to put the single-cause assumption aside, because several things are actually going on at once, simultaneous failures in different parts of a large complicated machine.

Continue reading “Digital media again/still”

Do I hear one million for the drop cloth?

Rothkos aren’t the bargain they were a year ago.

Mark’s last $1 million won’t buy much of a Rothko these days:

The witty 1944 work, Untitled, is a Surrealist rendering fashioned with a “Miro meets Dali in Madrid” influence, his style loosened and spatial with forms flattened and colors softened. The watercolor on cardboard sold at Sotheby’s in May 2005 for $350,000 (including buyer’s premium).

Or maybe a tiny scrap of a colored-block oil on canvas:

David Rockefeller’s success last spring in selling his 1950 Rothko for $72.8 million, a record price for the artist, has drawn many more Rothkos to auction. “Untitled (Red, Blue, Orange),” a classic abstract canvas from 1955, became the most expensive work of the evening.

Five bidders tried for the painting, which sold to a telephone bidder for $34.2 million, against a high estimate of $30 million.

Very bad news, stop cheering.

The New York Times appears to be about to stop charging for its opinion articles and archive searches (I only have unsourced assertions and suspicions to support this, but the noise is consistent). Several bloggers are cheering. I am not, and they are wrong.

I’m happy to get anything I can for free, but this development, however nice in the short run, is evidence of the inability of any newspaper to develop a workable business model for on-line publishing, and that failure is very bad news all around. It’s part and parcel of the slow collapse of the music business, whose sales are down another 20% or so this year, and the enduring piracy troubles of movie producers. Newspapers used to charge a nuisance fee for a physical paper whose circulation to countable readers allowed them to sell the readers’ attention to advertisers, and of course music used to be delivered on some sort of physical disk that was tedious to copy. Paper circulation is down and ad revenues with it, for every paper; yesterday the Times arrived an inch and a half narrower than it used to be. Consequently, the lights are going out in city rooms all over the country. I know there’s some Schumpeterian model of markets red in tooth and claw that should reassure me that everything is for the best, but newspapers are not just another medium, even if they have had a good century operating as profit-making private businesses.

Why can’t they just move the whole operation to the net? I never click on an internet ad and I don’t get much useful information seeing them, nor remember them, because they’re just too small to have either impact or much content. Whether because we’re so strongly acclimatized to stuff on the screen being free, or because it’s impractical to keep it from leaking out (despite a whirlwind of suing websites), or because people intrinsically understand that the economically efficient price for digital content is zero, the market is broken and the businesses are going broke. Cable TV, which people seem willing to pay for access to, seems to be hanging on, but I don’t watch TV commercials either and I’m not alone. There are a lots of Tivos out there, and advertisers are going to stop paying for eyeballs associated with skip buttons and timeshifting. When the dust clears, I see no workable mechanism on the horizon by which writers and editors can be paid to generate a newspaper (musicians can give live concerts publicized by downloaded and shared mp3s, but that’s not going to keep their business alive either).

Enjoy your free Times Select while it lasts, folks; that sound you just heard was not the triumph of free information over selfish profiteers, it was the sound of the iceberg opening up four or five compartments below the waterline.


Why is there no civilian counterpart to the Defense Advanced Research Projects Agency?

The July 21-27 issue of the New Scientist carries a story about the creation of light-senstive neurons, under the somewhat overwrought headline “Remote Control Brains.”

I have no idea whether the new technology will turn out to be useful, either in developing animals subject to remote control or in treating disease. I hope it does. But the most intriguing sentence in the article is a throw-away. Discussing a paper published in Cell by Yale’s Gero Miesenböck, the story reports:

Within an hour of the paper appearing online, Miesenböck’s phone rang: it was the US Defense Advanced Research Projects Agency wanting to know if his work had possible military applications (he now works with them).

Well, good for DARPA, I say. Maybe there are military applications.

But where was the call from DARPA’s civilian counterpart? Oh right, I forgot: there is no civilian counterpart to DARPA.

Why the hell not? If DoD needs a research-funding unit that combines speed, flexibility, willingness to take risks on low-probability but high-potential-payoff ventures, and freedom from the tryanny of the study sections, why doesn’t the civilian sector need the same thing?

Copyright and Calvin

The “generous soul” Mark thanks is, of course, generous not only with his nights staying up over a hot scanner, but also with Watterson’s and Andrews-McMeel’s property. The page he links to is a cultural treasure, a crime, and also a pithy lesson in what’s wrong both practically and morally with our current intellectual property regime. It’s obviously good for everyone that stuff like this should be available free at the click of a mouse, and obviously bad for everyone that we’re not developing a workable mechanism to send creators the right signals about what they should make more of (not to mention putting food on their tables and braces on their kids’ teeth).