Friday, May 31, 2002

TiVo Spam The BBC paid TiVo to record an unsolicited program on its UK subscriber's boxes. Consumers are incensed, and MPs are calling for legislation.

The problem here is that property rights haven't been defined. If TiVo offered customers a lower subscription rate if they agreed to let paid programming be recorded on their boxen, some would take it, and some wouldn't. But no one would complain when it happened, and everyone would be better of. After all, content producers pay TiVo to broadcast their programs, it's only fair they pass some of that onto subscribers.

Thursday, May 30, 2002

Unix fragmentation It'll be interesting to see whether anything comes of non-Red Hat GNU/Linux vendors' United Linux initiative. I'm guessing not. Each flavor of GNU/Linux is pretty substitutable, so lock-in moves to the (weaker) relationship level. United Linux is an engineering solution to what is essentially a business problem: poor distribution, services, and marketing.

(Don't) weep for local radio Here's a sappy article bemoaning both the death of old-style, mom and pop, idiosyncratic commercial radio and radio megalith ClearChannel's current financial woes.

First, it doesn't understand who lost from mom and pop radio--the public. I have nothing against small, local radio stations, but I want my bandwidth used as productively as possible and I'm guessing a three person shop playing their favorite tunes from 1972 isn't it. Secondly, it has the usual elitist attitude that people really want to listen to very unique and varied things, when really they just want to listen to what all their friends are listening to. Anyone with skin in the game understands this or goes out of business, but media's insulated from it's ignorance about, well, media.

Wednesday, May 29, 2002

DMCA and the First Ammendment My old roomie sent me this link pointing out that news stories on circumventing CD digital restrictions management are illegal under the DMCA. It uses the typical "trampling the First Ammendment" language, not realizing how flimsy constitutional protection really is. I say again, write to your Congressman.

Good Easy Email A while ago I wrote a little piece on how to handle 500+ emails a day. Mark Hurst, my old boss and the fellow who built the Good Easy system, has finally written up his email system himself. Worth reading and following.

Tuesday, May 28, 2002

It's better to be a star than a cow Let's face it, tech companies mature too. Although the Microsofts, Oracles, Ciscos, Siebels etc. of this world like to believe they're in the innovation business, they're actually in the PC OS, RDBMS, router, or CRM business, and once you've won the market you're done. But your shareholders and employees have drunk the innovation Kool-Aid, "mature" sounds suspiciously like "stagnant", and hey, you've paid out all these options promising high growth you have to be a middle manager or a Wall Street analyst to still beleive that.

After an initial product innovation, organizations specialize around processes to take that innovation to market. This makes them uniquely unsuitable to new product innovations which is why the Valley (and the world) has few companies who have successively taken genuinely innovative products to market. On paper, this looks like companies with no debt and lots of outstanding options accumulating large cash hordes. I had an excellent conversation with Ed Gilmartin about Microsoft's financials, which I've printed below. Even better, Ed calculated how much tax Microsoft could avoid by buying back shares, so I didn't have to. Great stuff.
In the 5 years to 30 June 2001 (sourced from MS's 10K filings and using cumulative figures for brevity), MS reported a profit before tax of approx. $50.1bn in total. In the same period it actually paid $5.2bn in tax - a tax rate of about 8.5%, and much lower than the prima facie tax rate in the accounts of about 34.5%. The reduction came because the company received tax rebates totalling $13.1bn due to its option program. This is a large rebate! In fact in the year to 30 June 2000 MS reported a pre-tax profit of $14.3bn and received a net tax rebate of $700m - it paid no tax, but in fact was paid $700m by the tax office!! What you say about debt reducing tax is right, but they seem to be doing OK in reducing tax & debt also has a cost in the form of interest - and, efficient market theory/Miller Modigliani etc would argue that company capital structure doesn't make a difference to value.

In the same 5 year period the company repurchased $19.5bn of stock. These guys are "super smart" in the words of the chief technology officer - they are paying a very low tax rate AND returning $$$ to shareholders.

I have not looked at the figures in detail but, in its last 10K, Cisco reported net cash and short term investments of about $7bn (no debt), Dell had $3bn or so reported in its last 10K - there is a pattern here - it is not just MS who is throwing off more cash than they can deal with. The others also have large share buy-back programs, so they are giving some back to shareholders. This looks like a wider situation than just MS "out of ideas" (although that may still be the case, see below); it seems to be a generation of large successful companies are throwing off cash and sitting on it. This could be caution (retaining option value to buy into the next big thing), the fact that (as you point out) dividends just don't make sense, the fact that they are all out of ideas, or some other reason that I can't articulate.

Ms's lack of investment opps. requires a long email of its own. Suffice it to say that they need X-Box to work, they need .Net to work, they need the Navison & Great Plains acquisitions to work; in fact they need everything to work to keep their growth rate up. What they seem to have become, as it turns out, is a cash cow - kudos to them for not behaving like other cash cows in the past & throwing their cash away (cf. big tobacco buying into food & then having to get out etc.) - but they are suffering the not unusual fate of the growth star that turns into a cash cow and then stays in denial about that fact, 'cos lets face it it's better to think of yourself as a star than as a cow.......

I think you are right that the cash pile signifies that MS is bereft of ideas (but I bet that most folks in the financial markets (if not the employees) know this too, it's just not in anyone's interest to point it out, especially while .Net etc seem to "promise" at the moment), but I'm not sure that's the whole story (see pt4 above).

Regarding their buying back shares, I'm not sure what the rules are in the US (UK & Australia are my areas), but there is usually a limit (10% or something) to how many shares a company can buy-back in any year without getting special shareholder approval - I'm not sure whether MS are near their limits - but if they need to get this approval it all gets very public & they probably have to publish a circular with reasons etc, so the lack of growth issues you allude to could get dragged into the open.

Also, bear in mind when you look at MS taking on debt/reducing cash that there is no real point in them going tax negative - they would reduce interest income or even go into debt through buying back lots of shares only to the extent that they could get the tax to zero, and since they are only paying 8-9% tax, that is all they have to play with. (Again, though, I'm not 100% sure how the option tax works: if it is actually a stand-alone cash rebate rather than a charge against taxable income, they may be able to get the option rebate & still have all their profit to service debt (although I would guess this is not likely).)

Finally, on the back of my envelope, (I'm jumping, so this may not be clear....), I would calculate as follows:

- MS had $38.7bn in cash and short term investments as at 31 March 2002; this could all be used for a share buy-back. - Assuming they have an EBIT of about $12bn (their 2001 number was $11.7bn), an EBIT/interest coverage ratio of 5-6x would be comfortable, so they might take on long term debt of about $40bn (assuming 5.5x interest cover and long term interest rate of 5.5% - they would not be raising this at the Fed. short term rates, but closer to the 10yr bond rate + a small margin).

- They could therefore buy back about $80bn of shares (as you note, much higher than what they actually bought).

- However, assuming their EBT was $14.1bn ($12bn EBIT + interest on the $38.7bn cash also at 5.5% for simplicity), and assuming they have 8.4% of tax charge to play with (ie. their 5-year average tax charge after option rebates, reduced from a prima facie tax charge of 34%), they could afford to reduce interest income/take on debt interest to the tune of $3.5bn (8.4/34*14.1); ie they could do a buy-back of "only" $63.6bn ($3.5bn of interest at a 5.5% rate) before they ran out of tax capacity.

PS. MS says in their 2001 10K that the value of options outstanding at that time was $66bn - there are a lot of those suckers still out there at attractive strike prices!!
Couldn't have said it better myself. To summarize: Microsoft's a mature company, but it's not acting like it. It's avoiding tax by exploiting sloppy option accounting rules (like everyone other tech company in the US), avoiding tax disadvantaged dividends, and has enormous hidden compensation liability (which it will renege on--watch for disgruntled employees trying to get theirs out of the cash horde, screwing current employees, who will end up taking whatever they can get from the cash horde).

Monday, May 27, 2002

When regulation is bad Clair Tristram does the usual reporter thing and casts the content cartel vs technology war in moral terms without weighing in on why the government should ever be in the position of selling injunctive relief against the future. Taxes, provisions, encumbrances, public property given (not sold) to private corporations ALL block innovation by protecting incumbents against entrants, and cause huge harm to society through the dead weight loss associated with an entire market failing to materialize. I seriously question the legitimacy of a government engaged in this type of behavior.

While the Valley wins on sanity points, I think the only way to stop Hollywood (barring mass grass roots activity) is by following rich historic precedent and simply out bribing them. Technology is more valuable to society than movies.

Thursday, May 23, 2002

Memorial Day hiatus I got my driver's license this morning (no, I'm not that young, I just used to live in NY) and I'm moving this weekend, so will not have time to post. But next week I promise all sorts of econ/law/technology goodness. Enjoy Memorial Day weekend!

Tuesday, May 21, 2002

Alan Cox gets it backwards I have great respect for Alan Cox, but he is precisely wrong to say Microsoft's cash hoard means its avoiding taxes. He argues
If Microsoft paid typical US dividends they would have under 20% of their current slush fund. (under because at 80% dividends the investors not the corporation got the benefit of reinvestment of most of the interest)
If they choose to sit on that $40 billion they should be paying tax on it because I really doubt they can demonstrate its neccessary for operational overheads. In which case 39% of it belongs to the US people. Which on a quick back of the envelope calculation is a bit over $50 per US citizen
As a matter of fact, Microsoft does pay taxes on retained earnings, just like any other company. It's true the cash would be taxed again if distributed as dividends, but that helps neither investors nor Microsoft, which is why fewer and fewer companies issue dividends any more. In fact, if Microsoft wanted to avoid taxes, they would take on stacks of debt and buy back huge swathes of their own shares. Debt is tax advantaged, so reduces corporate taxes, and buying back shares returns cash to investors via (untaxed) appreciation.

Microsoft keeping around stacks of cash means it's paying more tax than it has to. Microsoft refuses to take on debt and buy back shares because that would reveal it has no more good investment ideas.

Monday, May 20, 2002

Sad New Stephen J Gould passed yesterday or today, only a few weeks after an operation for brain cancer. It's not up on any of the news sites, but this is what my friends in Cambridge tell me. Very sad :(

MSFT's X-Box Strategy Microsoft's X-Box tanked at retail, so now the price wars begin. The Borg is hoping that online subscription services will revitalize this nascent franchise, and build a new division for the increasingly conglomerate like Beast of Redmond.

Greg Costikyan, whom I met when I lived in New York, wrote an extremely insightful analysis of the computer games industry (originally published by Creative Good, but now available for free at costik.com) outlining why the shrink-wrapped market was dying, and how distributed games, as well as modes of delivery, were going to become increasingly important. So Microsoft's multiplayer, networked strategy is a good one, but they're crippling themselves by insisting it all runs on .NET servers, a strategy that scared Electronic Arts into Sony's arms a few days ago.

Microsoft makes the usual arguments about walled gardens being safer and more pleasant than the wild internet, but given the treacherous and hostile landscape that is Windows 95, 98, ME, 2000, NT, and XP, I don't think Microsoft can add much value by integrating its offering. Also, I'm not sure how many customers they need to break even on this thing, but there aren't many hardcore gamers out there, and this isn't going to change anytime soon, no matter what industry hacks claim.

And on the topic of industry hacks, ignore cries about there being no broadband. Broadband grew at a healthy 90% last year, and I'm sure many broadband subscribers are also techie-gamers, suggesting good penetration in MSFT's target segment. I don't know whether these numbers are absolutely high enough, but they're pretty good relatively.

Epstein on Copyright Richard Epstein is a libertarian legal scholar here at Chicago. He's also quite brilliant. I found this remarkably intelligent chat transcript where he discusses copyright. He clearly understands why the Mickey Mouse Copyright Extensions act was a transfer of wealth from public to private hands for no good reason, but also notes that this case will be argued on narrow, constitutional grounds regarding "limited term" and "useful progress," and so may ignore this central issue. Ultimately, reasonable copyright terms rests with Congress.

Saturday, May 18, 2002

Mashes Weblogs have made DRM for text useless. It doesn't matter how sophisticated Salon or Slate become about locking up their content, people are too busy reading weblogs to care (or notice).

I've been looking out for the same phenomenon for music. Unknown garage bands get no traction on Napster because no one knows them and there's no difference between downloading a famous hit or obscure dud. I don't think mashes are 100% of the answer, but I think they'll play an important part.

Vinyl singles sometimes contain the complete song as well as decomposed a capella tracks which isolate, say, the music from the vocals. This allows DJs to play the vocals of one song against the music of another producing a "mash-up" that's proven a big hit in clubs. Inevitably, some of these mash-ups are released as combined tracks, and become illegal derivative works under current copyright law.

In this excellent BBC documentary, one of the DJs points out that if the labels really wanted to stop mash-ups, they should just stop releasing a capella tracks. The recording cartel likes mash-ups because they breathe new life into 10-30 year old songs which are still under copyright but no one buys. So a wasteful rent-seeking (are there other kinds?) legal apparatus has sprung up where mash-ups are supported through a capella tracks but legal heavies swoop in to extort money from anyone who becomes too successful.

This selective enforcement of law means property rights around a capella tracks are poorly defined, leading to terrible waste (the Coase Theorm). I'm guessing that this legal murk will eventually be cleared by registering a capella tracks under copyleft style licenses, perhaps developed by the Creative Commons, so DJs know what they can and cannot do with a capella tracks. Once labels start competing to offer better terms to DJs, barring collusion, they'll be forced to price at marginal cost and give the tracks away.

Friday, May 17, 2002

Durable goods Apple's new iBook comes with the slower G3 processor, and customers are complaining. They are also wondering whether it will ship with the old 8 MB ATI Rage 128 Mobility graphics card or something newer which will take advantage of the graphic acceleration system that will ship in Mac OS X.2.

For Apple to keep charging high-prices for its TiBooks, it has to keep prices of iBooks high (so people have less incentive to trade down) and the power of iBooks low (so people have more incentive to trade up). And it wants people to buy new hardware. The technology industry was always susceptible to the durable goods problem (i.e. stuff never wears out) but this has only become a major issue recently. As Apple has always understood, OS X's greatest competitor isn't Microsoft but OS 9. And given the healthy margins it makes on hardware, Apple has more incentive to use bloated software to drive new hardware sales than Microsoft, even though its integrated platform can supposedly run stuff more efficiently.

And besides, what are irate OS 9 on old hardware users going to switch to? If they're still using Mac's in 2002, their demand has to be pretty inelastic, so I'm betting Apple will include a powerful graphics card to encourage people to upgrade their hardware and enjoy the graphics accelerated goodness of OS X.2, but I wouldn't expect a G4 in iBooks until TiBooks ship with a G5.

Thursday, May 16, 2002

Am going to see Star Wars II I'm off to see "Attack of the Clones". I hope it sucks less than Episode I. I also just bought Moby's "18". I hope it's half as good as "Play".

Update: "Attack" sucks less than "Episode I." "18" is half the album "Play" was. *Sigh*.

Wednesday, May 15, 2002

Many Pipes Daniel seems pessimistic about many different types of (natural monopoly) internet service providers competing for consumers. I'm not sure why. Cable reaches about 95% of the population, and all of that will eventually be able to provide broadband service. I know you have to be within three miles of a switching station to get DSL, and I don't know how many folks are eligable, but let's guess 60%. Satellite, an expensive and lousy option, is available to everyone, although tellingly used only by people living in the middle of nowhere. I don't see why WiFi shouldn't ultimately be as ubiquitous as cable and DSL. And they all have to compete against the dialup access (available to everyone).

So I count 4.5 pipes available to at least 60% of the population, which doesn't sound terrible.

(BTW. Here's a Kalsey discussion thread on Apple+iChat vs Microsoft+IM that gets into tedious anti-trust economics.)

Web radio The government mandated royalty rates for Web Radio are finally being debated up on Capital Hill. But it's hard to set rates right, and its hard to change rates once they're set. In a market, so long as rate setters can discriminate between customers, they will fall (or rise) to the appropriate level. For Web Radio, this number will be zero.

It may seem counter intuitive that fixed, nondiscriminatory rates actually support collusion, but only this type of government mandate can effectively keep copyright holders from surreptitiously undercutting each other.

Similarly, it is copyright's weakness that's slowing the development of "open source" media with no copy restriction. Right now, there's no difference between downloading a copyright song and a non-copyright song. The more perfect a content holder's ability to price discriminate, the more attractive the non-encumbered substititues.

Tuesday, May 14, 2002

Supreme Court vs TELRIC The Supreme Court ruled that TELRIC pricing is OK (NY Times). Reading through their tortured opinion shows how appalling suited lawyers are to this type of decision making.

Pages and pages of rambling, turgid prose, contortions over jurisdiction, parsing "market cost" vs "book cost" vs "embedded cost", flawed accounting notions of "cost" itself, cluelessness of Ramsey Pricing, and unwitting support of government mandated collusive contracts. Sigh. Business as usual from the regulated industry front.

To be fair, dealing with natural monopolies, such as local loop telco bottlenecks, is tough. On the one hand, you want marginal cost pricing because its best for society. On the other hand, if companies have NO opportunity to harvest residual rents, then they won't invest in new equipment, which means above marginal cost pricing.

TELRIC pricing, which the Supremes upheld, requires incumbent telcos to offer bottleneck network elements at marginal cost. Therefore neither entrants nor incumbents have any incentive to upgrade equipment, and neither of them have since this pricing scheme was installed in 1996. Moreover, incumbents have to offer telephone service to all people, so they charge high business rates to subsidize low consumer rates. Entrants have no such obligation, so they "cream skim" by offering cheap services to businesses, which means fewer subsidies for consumers.

There's no magic bullet for natural monopolies. If you want the efficient solution, you simply grant the monopoly the ability to perfectly price discriminate. This means they capture ALL the surplus, infuriate consumers (who will pay high prices) but create no dead weight loss (so is efficient) and makes entry by close substitutes very attractive (as cross market price elasticity becomes infinite). Essentially, everyone will be charged their maximum willingness to pay, and this'll make them so mad they will gladly switch to some near substitute. Ultimately, competition between close substitutes is the best way to force down prices in natural monopolies, e.g. cable vs DSL vs WiFi vs satellite competing for residential internet service.

Monday, May 13, 2002

Hollywood and copyright Three nice articles on Hollywood's jihad against technology, and Ernest Holling's (Dem. SC) willingness to sell them injunctive relief against the future. Here's the economics.

Hal Plotkin asks why Hollywood just doesn't sell its own distribution network that controls content perfectly and leaves computers alone. It's simple: such dreck won't sell and the content cartel knows it. Each manufacturer would have incentive to cheat and sell cheaper, uncrippled hardware people actually want to buy and thus bust the cartel. Only a collusive contract can bind cartels, and since these are illegal, Valenti has paid Congress to write it for him. "Industry standards" bodies have long supported this sort of collusive activity, but John Gilmore gets his economics wrong when choosing who will lose from Intel's "standards" suggestions -- if manufacturers have a choice they'll support open uncrippled hardware that people will buy.

Lawmeme has a good article on why digital copies are as lousy as analog copies, but at one point concludes that since bootleg movies only work when legitimate movies aren't available, Hollywood can easily fix this by releasing everything at once. Hollywood delays international movie distribution because of the option value in dropping duds that probably won't do well abroad, and delays DVD releases to price discriminate between consumers with different tastes. If the internet reduces their ability to price discriminate, they get less profits, and that for them is a very real problem.

Finally, USA Today mentions how Microsoft's draconian school licenses are driving people to free alternatives. Perfect price discrimination in one market makes the demand curve for close substitutes perfectly elastic (i.e. cross market price elasticity becomes infinite) which greatly increases their ability to steal share from the incumbent. The more content owners perfectly prices their wares, the faster free alternatives will arise if they have open distribution channels. This is why even perfect DRM won't save the content cartel so long as people can still offer free content.

Mark Hurts and voice interface My good friend and old boss, Mark Hurst had the following good comments on my voice HCI posts:
Observing people "in the wild" using stupid, complex computers says nothing about how they might use (a) better tools (b) with better training up-front (think re goodeasy training on Creative Good-oriented macs - not perfect, but at least qualitatively better than the ordinary PC set up by IT).

[Good work habits are important...] and good tools! Despite anything else, we'll never get anywhere by watching people using the wrong tools. Windows slothware can't be reformed - this system needs a radical improvement.
No arguments here.

Saturday, May 11, 2002

Re: HCI will stay visual Will Fitzgerald negs my post on Maryland HCI's contention that voice interfaces are irrelevant beyond niche applications.

He cites search engines as an example of language and speech interface that lets people jump to data directly. While I agree people do not mentally organize information heirarchically and this makes good search engines useful, it says nothing about whether people will prefer to type "winterspeak" into google or say "google for winterspeak" at their computer. Here, my money is still on the former.

He also discusses "conversational interfaces" for distributed devices and contextual collaboration. The distribution question has been best answered by ubiquitous computing, and again, has nothing to say about speaking at machines vs. typing. Contextual collaboration is an AI question, and beyond narrow expert systems, it's failed to produce anything useful in thirty years.

The dream behind voice-interfaces is that "computers are too hard to use, and if I could just speak to them everything would be so much easier." But computers are hard to use because they're complex and stupid, so can't infer our intentions the way people can. Communication with complex, stupid devices will be frustrating, whether mediated by typing, speech, or a brick.

Friday, May 10, 2002

Microsoft's cash woes Microsoft has $40B in cash, and this is a problem. Companies exist to generate profit for their investors via 1) dividends, 2) projects that make even more money, and 3) share repurchases. Dividends are heavily taxed, so this is a stupid way of compensating investors (although some companies still do this). New projects are good, but only if they make money and it's unlikely that Microsoft can ever match its historic profit growth. That leaves repurchases, which I'm guessing Microsoft's avoiding because it's the same as admitting it has no good business ideas.

Employees will give up salary for growth, which is how Microsoft (and others) can get away with paying below market wages. But eventually growth stops and disillusioned employees bid their salaries up, and the company makes all its residual profits in the time lag in between these two events. (Note: If there was no time lag the company would make no excess profits, for rational employees the NPV of salary is the same between average wages throughout and low wages first, high wages later).

Microsoft pays out a lot of options. If the company ceases to grow, this form of deferred compensation becomes unattractive and employees will start demanding cash up front. Profits will shrink, but more importantly, the stock valuation will fall (as there are no more growth opportunities) leading to an even higher demand for wages now.

So why is Microsoft sitting on all this cash? Because it doesn't want its employees (and the market) to know it's out of growth opportunities.

Thursday, May 09, 2002

HCI will stay visual People find it difficult to think and speak at the same time. Therefore, controlling computers through vocal commands is too mentally taxing to be useful--better to stick with hand-eye coordination.

I think Don Norman put it most clearly when he said speech interfaces were no panacea to complex computing. The biggest problem with speech programs currently is that they poorly transcribe what you say. But if you use a keyboard and type the information in, transcription is perfect AND this doesn't help make the machine any easier to use. The only thing that will do that is watching how people actually use machines "in the wild" and then design directly off human behavior.

Wednesday, May 08, 2002

EU Internet taxes In the US, (sales) tax on purchases need to be collected by the vendor if it has nexus in that state. Nexus can mean as little as one part-time employee. If the vendor has no nexus in the state, it's up to the buyer to report the purchase on his tax forms and pay the tax then. In practise no one does this (you can try this at home -- ask someone what the "use" tax is)

Europe has decided to apply VAT to software downloads. This would let European governments soak the foreign (read American) companies that make most digital goods. Within Europe, customers pay the download VAT rate in the vendor's country, not theirs. This is why all the servers are in Luxembourg.

Tuesday, May 07, 2002

iChat revisited Why is Microsoft integrating IM into the OS bad, while Apple integrating it into their OS good?

From a consumer welfare standpoint, a better IM experience is good and integration with the OS might be a prerequisite. This is certainly what Microsoft cries whenever the puny DoJ thinks about enforcing anti-trust law. But monopolies have different economic incentives than entrants, and so different regulatory regimes should apply. Apple has a puny market share so is very focused on improving its product, Microsoft just wants to keep everyone else out. Microsoft honestly beleives it is not a monopoly because it does not yet control all digital information and therefore does not feel its actions are wrong.

iChat Apple is integrating a middleware IM platfrom called iChat into OS X.2. It "seamlessly" interconnects with Mac.com members, AOL members and AOL Instant Messenger (AIM). Why is Apple investing in a closed, proprietary platfrom when perfectly good open source alternatives exist? Because of AOL. AOL has the largest chat community, and that has to be attractive to Apple. Also, Apple is something AOL can beat on Microsoft with, and that has to be attractive to AOL.

Friday, May 03, 2002

Not related to tech Two not related to technology posts today. First of all, the NYTimes has revisited the discriminatory mortgage lending story in Boston. Unfortunately, the new study is as worthless as the old one, because the people in charge of it don't know what they're doing.

Prejudice is a consumption good just like orange juice or Nintendo games. The only real evidence of prejudice would be if loans to Hispanic or black borrowers were more profitable than loans to white borrowers. Prejudiced loan officers would want to "buy" there prejudice where it was cheapest, which means the least profitable black and Hispanic customers, resulting in higher than average returns on their loan portfolios. But no one understands any of this and so the cost of borrowing will just needlessly go up for everyone, harming all.

In other news, my money is on The Essence of Dubai to win the Kentucky Derby. And no, this is not because I lived there. The horse is proven at 1 1/4 miles, unlike Buddha (now dropped out), Johannesburg, Medaglia d'Oro, and Castle Gandolfo. And he's a better horse than Saarland, the wiseguy pick.

Thursday, May 02, 2002

"Going to the bathroom is theft" Time-Warner's Jamie Kellner is so outrageous you have to wonder if he's for real. His rant against TiVo, where he asserted TV viewers had to watch commercials (although a limited number of bathroom exceptions were OK), was priceless. Lawmeme has a good summary.

Now let's get to the economics. Broadcast TV is where programmers assemble audiences and sell them to advertisers. TV viewers may think content is made for them, but it's not, it's just bait to "aggregate monetizable sticky eyeballs" (and people thought the Internet came up with that idea!) TiVo messes this all up by fragmenting audiences and skipping ads.

So in the TiVo future, viewers will choose between paid-programming or programming with unskippable ads. Ad supported broadcast TV, a long-time political sacred cow, no longer has any economic rationale (unless viewers keep watching ads).

And now for something controversial. The Lawmeme article starts with "Top Ten New Copyright Crimes" a Letterman-esque list of price discrimination that transfers surplus from consumers to producers. But if a producer manages perfect price discrimination, the elasticity of demand for substitutes becomes infinite. This means that if content producers put in the perfect digital restrictions management systems they crave, demand will shift to content with no such restrictions.

When people ask why the web isn't filled with music from garage bands and amateur DJs, it's because the CBDTPA hasn't been passed. An unencumbered mp3 by U2 is better than an unencumbered mp3 by Nobody, but an unencumbered mp3 by Nobody starts looking pretty good when listening to U2 costs $15 a minute and percludes bathroom breaks.

Is spectrum a rival good? I recently stated that since spectrum was a rival good, strong property rights would ensure efficient allocation. Staton pointed out people who disagreed, and then wondered how it should be allocated if it was non-rival. If spectrum is non-rival, it should be completely unregulated, but that doesn't matter since, contrary to dissenters, it is rival after all.

David Reed claims that there is actually no such thing as interference, since "radio signals just add to each other, non-destructively." This is actually false since if a signal came across its perfect reflection, they would (additively) cancel each other out, and the information would be (destructively) lost. This is how those noise-canceller airplane thingies work.

What Reed really means is that processing power in the receiver can deal with dirtier signal, and so use spectrum more efficiently. This statement is not the same as "there is no such thing as interference" and is actually true. There is a trade off between processing expense and spectrum consumption, and only a market can correctly allocate between the two.

This is why strong spectrum property rights are still the answer. If you bought 50Mhz of spectrum, and receiver technology improves so now you only need half that much, you can simply sell off the unneeded remainder. But for this to work you must own the spectrum in perpetuity and be able to use it for anything.

All the considerable waste in spectrum exists because the spectrum users don't own it in perpetuity or it has encumbrances on its use, which means it can't be freely sold. Strong property rights would solve both problems. Also note that if the signal attenuates quickly, as is the case in 802.11b, and there's no interference problem, then spectrum actually is a non-rival good and it should remain completely unregulated.

Wednesday, May 01, 2002

Bruce Sterling loses his touch I tried to like Bruce Sterling back in the late 80s, but his books (The Difference Engine, Artificial Kid, Heavy Weather) were all so terrible I couldn't. It seems he's tried to refashion himself as a cyber-visionary-writer, a la Neal Stephenson, but unfortunately, he does not understand technology, has no vision, and is a poorer writer in every sense of the term.

I came upon a transcript of his closing speech at CFP conference which sorts of sums things up. He has some interesting enough things to say about the Bush administration, and some stupid and boring arguments about how unauthorized sharing hurts artists, bringing up Bollywood actress, Kajol Devgan's tapes in the US.

Most arguments about "artists deserving to get paid for their work" are about redistribution -- they think artists should keep more of the money than record companies. But copyright law says *nothing* about redistribution, and idiotic laws like "moral rights" in Europe only impoverish artists more since distributors are less willing to pay for material with encumbrances.

Artists are poor because 1) they're common as muck and 2) no one pays much for content. Labels are rich because 1) their distribution channels are scarce and 2) promotion is costly. While web radio and the Internet reduce label's ability to make money, artists are as fungible as they ever were and no law can overcome these basic economics.