Friday, November 28, 2003

Musicians go it alone

A nice piece in, of all places, American Airline's inflight magazine, outlining how some small musicians are making a reasonable living off live performance, direct album sales, and merchandise. If their music is distributed online, well, that's just valuable free advertising.

Tellingly, I have not heard of any of the artists profiled. File trading alone is not a good way to build visibility -- you need promotion (like being written-up for inflight magazines) as well.

Wednesday, November 26, 2003

Thankless work

I'm amazed that Ken has the stamina to trawl through editorial pages, score op-eds, and then keep records. But it's data like that which makes his rebuttal to Brad DeLong so solid. Brad argues that since the Economist only named one Krugman critic (Ken) and hinted at others, their claims are baseless since (presumably) Brad can name many supporters. But Ken isn't just any critic, he's systematically recorded 372 (372!) columns and this is how the numbers turn out. If you don't like them, run your own numbers, but you can't dismiss (or confuse) this sort of data collection as opinion.

(I will add that Krugman's stock had fallen very low at U Chicago amongst those who knew that supply curves sloped up, and demand curves down. By the time I graduated, he had become a joke.)

It's worth thinking about opinion. I am no expert on this subject, but I've been to enough debates to know that, well, debates don't change anyone's mind about anything. Neither side, nor spectators, come away thinking anything different from what they felt going in.

So if arguing does not work, what does? My personal hobby horse these days are bets -- after all, if the other guy is such an idiot surely you can get some money off him, and if it turned out you were wrong, well, consider it paying for knowledge, always a worthy transaction.

One way to think of the way Reagan won the Cold War was as a bet -- communists believed (as many people still do) that a centrally planned economy outperforms a market economy and Reagan believed the opposite. This was a bold belief at the time, btw, as conventional wisdom was firmly on the USSR's side. The way Reagan placed the bet was through an arms race that would bankrupt the less productive economy over time. He won the bet.

I don't know what the central bet is between Islamic fanatics and the free world, but thinking about that clearly is probably a better way to understand what's going on and avoiding futile arguments.

Happy Thanksgiving

To all US readers -- Happy Thanksgiving!

Bloggers and politics

Andrew Sullivan has a good piece outlining how Dean's campaign used the Internet to bypass the DNC and built a cheap, grassroots machine to drive the very successful campaign. Sullivan thinks this trend is good for Democracy:
That means real dynamism in the campaign next year. With the web operation in place, a burst of enthusiasm after an early primary win could mean an instant infusion of web cash that could then cover a key state with advertising and keep the momentum going. One good showing in a debate and, again, the response is instant. This insta-democracy could well have its disadvantages, of course. It could remove some of the barriers that deliberately slow democratic decision-making to avoid too much fad and not enough substance. But there is no denying its power"
Firstly, I must admit that I am less impressed by current democratic decision-making than Sullivan because I don't believe that making it more faddy and less substantive would alter it in the least. Indeed, I don't think anyone would even be able to tell the difference.

Secondly, while the Internet is good at making activist voters contribute money, I would try to be realistic about how much influence money really has on elections in the US. (U Chicago economist Steven D. Levitt has tried to quantify it here (.pdf) -- hint: it's pretty small). And while these online activists bring cash, they also bring ideological baggage that would be harmful in a genuine election. Given voter participation rates in the West, anyone thinking about or supporting politicians 18 months before the election is clearly an outlier.

Lastly, it would be nice if people just dropped the pretension that democracy can somehow be improved by "bringing it closer to the people" or "getting people more involved" or whatever the power-to-the-people sentiment of the day is. Ken Arrow (U Chicago, econ, nobel prize) mathematically demonstrated that voting systems simply cannot be perfect because it is impossible to produce a decision, through polling, that is intransitive. The fact that Nader supporters voted for Bush is a fundamental feature of the system, not some aberrant flaw. Given that the election winner is an artifact of the voting schedule, you have to wonder how the quality of the inputs matters much one way or the other. This is without going into the fact that, given how unimportant the marginal vote is, and the tremendous disutility associated with listening to politicians drone on and on, and only rational decision is to ignore the whole business. Empirically, this is also the popular choice.

I don't think the web nature of Dean's campaign is revolutionary in the way people seem to think it is, because I don't believe that the "authentic", "grassroots" support it has is anything more than a new special interest group that can now work the system. I think it will help and hinder Dean to the same extent any special interest group helps or hinders a candidate.

Saturday, November 22, 2003

Three cheers for ESR

Eric Raymond has finally completed "The Art of Unix Programming". It's a pity O'Reilly did not publish it -- I'd be interested to find out why.

Cellphone pricing

With number portability set to hit this Monday, and the fact that Sprint cannot reach my new basement apartment Boston, I've been looking into getting a new carrier, along with a new phone and plan.

It's been very complicated. The plans themselves are so difficult to understand, that I feel like I'm buying an airline ticket.

Perhaps this is because the cellular phone business is like the airline business. David Anderson, who I've known for a while, has written a good piece outlining why this is so. Essentially, David points out that a wireless network, once built, is a sunk cost and that only the variable costs associated with utilizing that network should be considered -- and they happen to be zero. This is like airlines in that the planes are sunk but empty seats cost the airline money, so airlines should be built around maximizing capacity utilization -- ie. having planes filled with passengers in the sky at all times.

This model applied to phones would focus on getting 1) maximum revenue per customer and 2) getting the customer to use the phone as much as possible. So, when designing phone services and figuring out how to price them, we move from cost/megabyte to revenue/subscriber communication.

Camera phones seem to be popular now, and actually I think they are pretty cool myself, but I would prefer a simple phone with long battery life and ubiquitous reception. It seems that building out reception is actually expensive compared to getting people to shove more data through the pipe they already have, all wrapped around complex pricing schemes, so I don't think I'll see my ideal phone any time soon. Pity.

Friday, November 21, 2003

Mutual Fund Scandals

Three cheers for Michael Lewis. Lots of people at U Chicago reckon that mutual fund managers aren't worth the money, and have run lots of numbers looking for places where they earn their keep. They don't find any. More worryingly, this seems to also be true in the private equity business, except some people there actually do seem to outperform but then take all that performance back through hefty fees, bonuses, and salaries. Good for them, not so good for the investor. Is it any wonder that most of my classmates were very excited about PE jobs and less interested in the mutual fund business?

Michael Lewis nails the real scandal in the mutual fund industry -- that people waste their money investing in them at all. The really smart money is in the "carry" ("carried interest") of very competant investors who supplement their skills with institutional cash. These guys outperform, and pocket their winnings. The rest of us should have our money in broad index funds and pay accountants to sheild us from taxes.

Thursday, November 20, 2003

WSJ note on the Post-iTumes Music Store World

The WSJ has a note echoing my recent observations on the post iTMS world (see below). Since you need a subscription, yadda yadda, I'll post an excerpt:
Crowded House: With the Web Shaking Up Music, A Free-for-All in Online Songs --- Companies Race to Stake Out Turf in Fledgling Market; A Shakeout on the Way? --- Sony vs. Wal-Mart vs. Apple

11/19/2003, 17:59 [The Wall Street Journal]

When Apple Computer Inc. started selling songs over the Internet for 99 cents each, the company had the field almost to itself. Now, just seven months later, so many competitors are jockeying for position in the music-download business that a brutal shakeout is all but certain.

A half-dozen companies have already followed Apple's lead. Roxio Inc.'s Napster, the once-renegade pioneer of Internet music, now lets you download songs at 99 cents a pop. And some of the biggest names in retail, technology and media -- including Wal-Mart Stores Inc., Microsoft Corp. and Sony Corp. -- will soon jump into the fray.

It's a rare moment in which technology has jolted an industry's business model and past practices, kicking open the door to a radical new distribution strategy. For decades, retailers have dominated the sales of physical copies of music on compact discs, vinyl and other formats. As more music fans download songs over Internet connections and organize music collections on their PCs, it suddenly isn't clear just who will sell music to consumers in the future and how.

But even as companies race to claim market share in what has rapidly become one of the hottest sectors of the Internet, online music remains unproved as a business. Profit margins appear to be so thin that many companies see song-selling sites as loss leaders to help them hawk other products.

Executives predict that the sheer number of companies rushing into the market and the likelihood of price battles will further crimp profits, leaving only a handful of significant players standing. "I think we will shrink within 12 months down to five or six or even down to three" companies in the online music market, says Sean Ryan, vice president of music services at Seattle-based RealNetworks Inc., an Internet software and services company that will soon begin selling song downloads through its Rhapsody subscription service.


But as they fight bootleggers, the record labels have finally given up on their long-held resistance to establishing a legitimate online music business. For years, the major record labels balked at licensing their song catalogs to legitimate music sites, and most of them burdened the music with unwieldy technical safeguards that prevented consumers from recording songs onto CDs or transferring them to portable music players.

Now, the labels have gone headlong in the other direction. Increasingly, the recording industry has made attractive licensing agreements far easier to come by, as it grows more comfortable with the Internet, and the terms of the contracts it strikes with online distributors become more routine.


Selling song downloads turns out to be a low-profit-margin business. Of the 99 cents Apple and other sites charge for a song, the companies pay anywhere from 65 cents to 79 cents in wholesale costs to music companies, analysts and industry executives say. Credit-card processing fees, bandwidth charges and costs related to customer service can, in some cases, eat up whatever profit is left over.


Competitors argue that Apple's dependence on the iPod is treacherous. Apple for now commands a premium for the device, but the iPod advantage may be short-lived, as competitors gradually improve their own music players. That will, in turn, increasingly put pressure on iPod profits for Apple, which has a long history of seeing its technical innovations copied by competitors. Already, competition is getting more intense: Dell Inc. recently began selling a portable music player that costs $50 less than the least expensive iPod. Samsung Electronics Co. makes a music gadget bearing the Napster logo.

Sony could make life even harder for Apple. The Japanese electronics giant has said it will introduce a competitor to iTunes in the spring that it is for now calling Music Box. Sony will also offer a family of portable music players, including a device that sells for $60, a fraction of the iPod's price tag. The company has hired Jay Samit, a music-industry veteran who worked for years on Internet deals at EMI, to help run Music Box.

Tuesday, November 18, 2003

Not Useless

I actually kinda like MSFT's SPOT watches idea, except I would want the watch to provide a more convenient screen for my phone and palm pilot. I would prefer to get my appointment alerts and incoming phone numbers on my wrist.

Monday, November 17, 2003

The Post-iTunes Music Store World

OK -- iTunes music store (iTMS) is now available for both Macs and PCs. There is broad agreement that it combines reasonable value ($1/track, good selection) with reasonable protection (burn to your hearts content, then re-rip if you really want to, otherwise tricky to share). iTMS is tightly integrated with the iPod, but the software is free, runs on the vast majority of computers, and good. So the overall strategy is to make the fairly popular iPod even more desirable than other mp3 players.

The iPod is kind of a surprise hit for Apple, being the first true piece of consumer hardware that the traditional PC hardware company has ever produced. People are talking about Sony being on the ropes, while others opine that AOL, MSFT, Yahoo!, MTV etc. will dominate.

This is all a little nuts. There are three pieces to this technology stack -- the music copyright holders, the network distributors, and the hardware playback manufacturers. Napster played in the middle, was great for the hardware guys, but seriously threatened the copyright holders, who have since used the fact that copyright infringement is illegal to shut the system down.

Apart from lawsuits, the RIAA has also begun to license music for various online systems, and they do not seem to be very exclusive about who they offer distribution rights to. This suggests that copyright holders do, in fact, hold all the cards and they want to commodify the distribution and hardware parts of the stack down to marginal cost so they can re-licence their content at 100% margin. If a distributor (or hardware producer) enjoyed market power, they would require the copyright holders to give them exclusive rights to the song, weakening rival distributors (or hardware producers).

There are two different models of hardware producers -- integrated and high price (iPod+iTMS) and open and low price (Dell, or Gateway, if they have one). Sony briefly flirted with an poorly integrated and high price option which looked pretty but was horrible to use and thankfully is no longer for sale. We've seen the integrated and high price strategy play out for Apple in the PC market, and we'll see what happens with music. With the consumer electronics business being as cutthroat as it is, and the labels holding all the cards around distribution, neither of those are going to make any profit long term as separate segments. So the only person left to create the sort of ease-of-use and value that makes the for-pay music biz compete with the P2P networks is the RIAA, which means they are a cartel that play nice sometimes and fight other times. If the RIAA does not create standards that allow easy integration and commodification down the distribution and hardware parts of the stack, someone like Apple can potentially take some of their profit from them by building a powerful enough distribution+playback system.

So, expect consolidation within the RIAA, which I believe we see through the Bertelsmann/Sony deal. Look to see many more distribution services and hardware producers, and the labels license their music more aggressively and desperate PC manufacturers enter the mp3 player market. A standard, single price rate (ideally enforced by law) would cap the cartels power, and allow it to collude on prices with the government acting as contract enforcer.

All Apple can do about this is reduce the copyright owner's power by giving distribution ability to more desperate copyright holders. So look to them cutting deals with small labels kept out of the big stores. Apple is limited in how much it can do in this area because it costs them real money to host and deliver each song -- they can't drop the iTMS price to zero even if the copyright holder wants them to because it costs them more than that to offer the song. Apple could integrate P2P functionality within iTMS that would let people share free (non-RIAA) mp3s between each other, and this would require the very strict sort of DRM that a combined distributor-playback company can offer. It is strong DRM that makes it tough for the RIAA to open the floodgates to all distributors and hardware manufacturers, and it is strong DRM that is best made convenient by integrated distribution/hardware companies.

Tuesday, November 11, 2003

Surowiecki asks "what is to be done"

This post on Brad DeLong's cite is remarkable not because of anything Brad wrote, but because James Surowiecki appears (repeatedly) in the comments thread. Once upon a time, many years ago, Slate had an excellent economics and business section, with Surowiecki and Krugman writing fantastic articles on money and business. Krugman lost his mind, and Surowiecki moved his continually excellent column to the New Yorker, where it either stands out or languishes (depending on your point of view) between deeply personal and long articles about other peoples' house pets.

I have no idea if this James Surowiecki is the same one as the Slate/New Yorker one, but the clarity of writing and general thoughtfulness suggests that it is. I recommend skipping the post and getting straight to the comments.

Monday, November 10, 2003

Matrix III

Don't worry -- no spoilers -- but I did want to say that I quite liked the third installment. I found the ending to be very satisfying. After reading all of the critical reviews I had set my expectations very low, which helped I'm sure. Also, the world is ready for Miracleman, the movie.

Saturday, November 08, 2003

Behavioral economics examples

This piece (once you scroll past the stuff on Dostoyevsky) has some of the typical questions behavioral economists post to subjects to try and figure out where their biases are. It's worth taking them yourself, because there is nothing like being proven wrong to open your mind to new ideas.

Thursday, November 06, 2003

Myth debunking

This is an interesting interview with Ray Hyman, who talks about how easy it is to fool people, for people to fool themselves, and why pseudoscience is so persistent. It's also nice to read Physicist Richard Feynman's excellent and justly famous piece on Cargo Cults. People do not look for disconfirming evidence. We are not cognitively well equipped to understand the world around us (the fact that we can at all is remarkable). Just for those of you out there who want to understand why the Web is suddenly filled with crazy people.

Tuesday, November 04, 2003

NukeUS for 23

Public outcry may have scuppered the Pentagon's plan for terrorist futures, but predictably the type of bets it would have carried have appeared in other decision markets. My old roomie Stumbling Tongue points out that the markets predict there is a 20% chance of nuke going off in the US by 2010.

Let's try this again: people think there is a 20% chance of a nuke going off in the US in the next 6 years!! It's hard to know what to say. I have no idea if that 20% figure is too high or too low, but I hope to God it's waaay too high. Three years ago I would have put the odds at zero, but now the idea of a young man wandering into Grand Central, saying "Allah Ho-Akbar," and pushing a small red button seem terrifying plausible. If such a terrible catastrophe does happen, it will make this war against terrorism (depose a brutal dictator as bloodlessly as possible and then spend $87B in cash, and who knows how much in blood, to install a democratic and free muslim state in Arabia) look like, well, a wonderful, but naive, humanitarian basket-weaving session. Worth keeping in mind.

Monday, November 03, 2003

When is a raven like a writing desk?

(Or when is a preference not a preference). Dennis Dutton writes a piece for the excellent arts & literature daily outlining how Darwinian evolution explains people's economic preferences. Given the amount of gibberish spewed on this subject in humanities departments, it's nice to see genetics and evolutionary biology start to shed real light on human cognitive function.

Dutton's point is that human nature was selected to maximize reproductive chances in a caveman world that is very different from today's world, but explains why we do the wacky things we do. He talks about Pinker's Blank Slate (which is very good) but it seems the topic is given a more complete treatment in Rubin's "Darwinian Politics: The Evolutionary Origin of Freedom" (which I have not read.) Rubin, by the way, is a Law and Economics professor, a type whom I tend not to trust because (in my experience) they like theory too much and never do the math, so I'm not sure how good his book will be.

Nonetheless, the thoughts here echo a short paper I wrote in Thaler's (U Chicago Behavioral Economist) class at school. We are asked to write about when loss aversion and mental accounting were useful heuristics, and I think were expected to jot a note about self-discipline, but I wrote about cavemen instead. So, first some background, then my note, and finally some thoughts.

1) Background

All else equal, a gain of $10 should be worth as much to you as not losing $10. In practice, losses loom larger than gains, and people prefer to forgo $10 new dollars over paying $10 old dollars out of pocket. People also use mental accounts, so they will create categories for expenditures and ignore the fact that money is fungible across categories. This leads to amusing anecdotes about people risking blizzards to see a play when they bought the ticket, but deciding the drive is not worth it if they were given the tickets as a gift. In standard Chicago school neo-classical economics, none of this should happen.

2) The paper (edited for length)
Question: Mental accounting, to the extent that it violates fungibility, is (according to economic theory) irrational. Do you think there are any circumstances where mental accounting makes people better off nonetheless? How?

Mental accounting (concave value gains, steeper and convex value losses, the endowment effect, transaction utility, and multiple accounts for purchases) might make people better off in a primitive "caveman" society characterized by subsistence living, zero capital formation, a barter economy, and no law except for trust and coercion within your small community.

In a subsistence society, individuals are essentially just one big mistake away from death. So while they will seek gains they will be unwilling to take big risks to do so and when faced with a potentially catastrophic loss, they may decide to gamble. This survival instinct may result in a value function where losses loom larger than gains, and individuals are risk seeking to avoid loss while risk averse in seeking gains.

An additional effect of subsistence living is that there are essentially no savings. An evaluation mechanism that evolved under these conditions would equate gains or losses from a transaction as being equivalent as gains or losses in wealth because they are, indeed, equivalent.

Similarly, in a barter economy with weak property rights it may be reasonable to underweight opportunity cost. Firstly, property that is not proximate may be easily stolen, and secondly, the risks and inefficiencies inherent in bartering make goods less fungible. The net effect would be to only count on property that you have for sure, and not anticipate changing that property into something you value more.

Subsistence living is also zero-sum, focused on redistributing wealth not creating it. In this environment, wellbeing is tied to tracking how much each group member is getting relative to each other. In this environment, it is important to track whether or not the transaction was fair, and assume that any individual gain came from another individual’s loss. The value of goods in this environment is essentially the cost to the seller (p*) since there no opportunity to add value through capital. Since the only competition is in distribution, not production, tracking transaction utility is most important.

Finally, it's worth distinguishing between two different kinds of self control problems. The first set essentially come from a mismatch between what was biologically optimal in a primitive society but is harmful today, with sugar, salt and fat being obvious examples. There is a mismatch between our biological desires and our "rational" cognitive desires. The second set of problems is where the optimal strategy is to cheat and avoid getting caught. For example, an unwillingness to indulge in luxuries, may reflect how in a primitive society individual survival is dependent on group cooperation, even though the group is made up of competitors. Therefore the individual needs to maintain the support and trust of the group, and desires to be seen as frugal, and conscientious. In this situation, people would be more comfortable indulging their luxuries in private, and only being public about luxuries that were given to them as gifts. A luxurious gift demonstrates how important others think you are and therefore furthers your reputation as a valued member of the tribe. The same mechanism of mental accounting that separates and controls private indulgence versus public thrift may have been recruited to handle the broader variety of tasks we need to juggle in our more complex world today.

If the systematic biases uncovered in behavioral economics are universal and have a genetic basis, then it seems reasonable to suppose they were selected for in an evolutionary context. One would therefore expect them to be optimal for the primitive existence that characterizes the bulk of human existence.
My grader (and good buddy, compadre, and general partner in crime) thought I was loony and said so in his comments.

3) Thoughts

In standard economic models, anything someone gives up money for is treated as a preference. If you give up money for ice cream, it means you have a preference for ice cream. If you give up money for a male colleague (by hiring him over a more talented female colleague) you have a preference for sexual discrimination. Preferences can be legitimate or illegitimate. For behavioral economics to show that people giving up money for stuff is irrational, they need to show that it cannot be simply thought of as a preference, and they do this by showing how framing effects, or circumstances, create choices that are, well, stupid.

Is it unfair to call them stupid though, as we all struggle through his modern world with our caveman minds? I don't think so, but that does not change the fact that we, as individuals or society, as traders or committee members, as elected officials or private citizens, put ourselves and our clan before others and act dumb.

Thaler used to say that once you spot a bias you should de-bias, and if you can't de-bias, re-bias. Dutton, Pinker, Rubin, Thaler, and yours truly (among others) are pointing out real biases that are probably biological, and I don't think that any degree of training will really rid them of us. Therefore, the de-biasing strategy is a non-starter, leaving only re-biasing to help correct these errors.