Value and Veracity

If, like me, you don’t have time to read Sach’s new book, take an hour to watch his recent talk at Google. It starts out a bit slow but is ultimately worthwhile.

One great point I hadn’t heard before was for direct carbon taxes over a cap & trade system: since carbon emissions are almost exactly proportional to the carbon in coal that comes out of mines, plus petroleum from oil wells, plus maybe a couple other sources, we can very efficiently monitor and tax the carbon coming straight out of the source (eg, the quantity of coal coming out of the mine) rather than monitoring the emissions out of every smokestack at every power plant and factory. Cap and trade does have the theoretical advantage of setting a well-defined limit on emissions, but it comes with a much larger implementation cost.

As Sachs says, financial types might find the idea of a carbon market sexy, but considering the failures of some recent “financial innovation,” I think it’s prudent to avoid a complex financial solution when a simple and effective tax-based solution is available.

Thanks again to Jamie for this tip and another one: haikus on financial markets.

As I mentioned in an earlier post, our standard methods of providing an incentive to innovate - patents or copyright - come with the significant cost of creating a temporary monopoly (people will have to pay too much for a while to enjoy the new product). But, as Nobel laureate Joseph Stiglitz pointed out in a talk given at Google, there are other possible incentives for innovation. He suggests setting up large prizes to encourage research into new medicines.

To motivate the production of knowledge, academics already have something similar to what Stiglitz suggests in the form of Nobel Prizes and tenure: the possibility of secure employment, intellectual prestige, and a good chunk of money does a lot to motivate young intellectuals to create knowledge (which the rest of the world may freely benefit from). Why not do the same thing for open source software, another public good which, like good ideas and new medicines, benefits everybody when it is produced? Turns out Sun Microsystems recently realized the wisdom in this possibility (Or at least it was recently when I first started writing this post). They are going to give out a million dollars in prizes to people who contribute to open source software development.

That’s is a good start, but I would say it’s a long shot from approaching anything near an economically efficient level of public funding for open source innovation. Here’s why:

Open source software (OSS) is a public good in the purest sense of the term. This means if I create an OSS program, everybody else in the world will be free to use it and benefit from it. It also creates a free-rider problem: once somebody pays the costs to produce an OSS program (primarily the labor involved in writing the code), everybody else will be able to use it, and “free-ride” off of the efforts of the creator, who might not get paid squat. So why would anybody want to write an OSS program in the first place?

That’s a good question. A complete answer is probably very complicated, but whatever inspires hackers to work on open-source code, they could probably be more inspired if considerable wealth and fame were on the line. But would it really be worth many millions? I say hell yes.

Think of how much money people in the world economy spend on software: $200 here for OS X, $300 there for Vista, maybe $650 for Photoshop, and $300 for MS Office. If Americans currently average around $100/year on everyday software (stuff than can feasibly be replaced by open-source alternatives), then American as a group are spending tens of billions in total. Now, if just ten percent of that was turned into public funding and used to sponsor OSS, we would have something on the order of a billion dollars to spend - quite a lot more than Sun is putting up.

Maybe the most interesting question here is how we could spend that money effectively. But putting that aside (please do comment if you have thoughts), with that kind of money, I bet we could somehow get the world’s programmers to put a lot more effort into OSS. Just for fun, let’s imagine an annual “Torvalds Prize” given out in Sweden (or somewhere) every year that includes a six-or-seven figure cash prize and a great big gold medal. If we had it five years ago, probably some people associated with Apache, Mozilla, and maybe Ubuntu could be driving around in Ferraris wearing their great big gold medals. This would be a huge motivator for young computer whizzes everywhere, and surely within a few years the quality of Linux, Gimp, and OpenOffice would fly past that of their closed-source rivals (Windows/Mac OS, Photoshop, MS Office, respectively).

That’s a bit hyperbolic, and high-profiles prizes are not the only interesting potential incentive for OSS development, but I think this sort of analysis has serious policy relevance. Software is a huge part of our economy these days, and with Apache and Firefox we have already seen widely-used open-source software that is a perfect substitute for (or even superior to) the proprietary version. Other programs have been catching up. With just a little public funding, who knows how many other open-source programs would fly past propriety rivals and into widespread use?

And here’s the really relevant thought experiment: if Americans collectively pooled as much money as they currently pay for software, could it be used to produce higher quality open-source software than Microsoft, Apple, Adobe, etc are currently putting out? For the economists out there: are we in the midst of a huge market failure - are we seriously underestimating the positive externalities of open-source software (or overestimating the costs of producing it)?

It’s hard to imagine what we could do with tens of billions going into OSS, especially when we don’t have a clear idea of how we would be using that money, but this is an important question to ask.

A recent post on Apple Matters argued that the recent resurgence of Apple is bad news for Linux:

In the last two years, OS X has seen continual growth, from 4.21% in Jan 2006 (the first month of figures), to 5.67% in December 2006, to 7.31% in December 2007.

In the same time, Linux’s percentage has risen from only 0.29% to 0.63%. Although depending on how you apply the maths—you can put a positive slant on that by saying it’s more than doubled—the cold truth is Linux on the desktop is still barely worth mentioning. To paraphrase: reports of [Linux's] life have been greatly exaggerated.

I strongly disagree, so I’ll try to explain why this is a very exciting time for Linux.

If NetApplications’s numbers are right, Linux has grown 117% in two years. That’s pretty good, and considerably better than Apple’s 74%. In fact, at this rate it will take less than 13 years to have 100% of the market share! (at which point Apple will have 270%… weird).

Silly extrapolations aside, I don’t think a growth rate of almost 50% per year needs a “positive slant” to be impressive, but what really bothers me is the perception that there is some sort of conflict between Apple and Linux, and that Apple’s victories in this conflict are somehow bad for Linux. Maybe that’s not what Howard is saying, but he certainly suggests that these recent numbers reflect poorly on Linux. That suggestion is very false.

Although Linux’s market share remains small, network effects make these recent statistics very exciting. Wikipedia defines a “network effect” as “a characteristic that causes a good or service to have a value to a potential customer which depends on the number of other customers who own the good or are users of the service.” Linux has many such characteristics, the most obvious being that it is open source software, which means any developers that use Linux can fix bugs they find, which improves the product for everybody else. This improved quality attracts more users, and a positive feedback loop results just like the one that fueled Wikipedia’s rapid growth.

Open source is Linux’s long-term advantage against Mac and Windows, but there are also very powerful network effects associated with any operating system, and these may be more relevant in the short term. If everybody at your office or school is using a particular operating system, it makes a lot of sense for you to use it, too; the more people use an operating system, the more value that operating system has to a given user (holding other things fixed).

You’ve probably heard about how Microsoft in its early days, realized the power of network effects and actually paid computer manufactures to install windows on the computers they sold. By acquiring a large portion of the market, Microsoft added a lot of value to Windows, which led more people to buy Windows, which made added more value to it, which eventually meant Microsoft could charge a high price for Windows even when viable alternatives (like Linux) are free to download.

In any case, Linux has become a little bit more valuable thanks to that extra .34% market share. You might say “sure, but an extra .34% of the market can’t be that much.” I disagree. When an OS has a very small share of the market, each additional .1% adds a lot more value than when an OS has a large market share like Windows. Imagine an extra 1% of the population started using Windows, raising its market share from ninety-something to ninety-something-plus-one percent. Most Windows users probably wouldn’t know the difference. If I’m a college student, almost everybody living in my hall will be using Windows either way. But a 1% increase in market share is huge for an OS with a small market share. An increase from .5% to 1.5% might mean that there will be enough students on any college campus to have a healthy Linux-users’ mailing list and perhaps a Linux-users’ Wiki, making it much easier for students and effectively use Linux on campus. And if an extra 1% of the population switches to Linux, that translates to thousands more people every day reporting problems and finding solutions on the Ubuntu forums and other support milieus.

This means, as Windows loses momentum, Linux may be picking up speed. Furthermore, as users switch from Windows to Apple, Windows loses some value from network effects, and Linux looks relatively more attractive. So Linux and Apple are friends. For now, at least.

Apple has done a great job in recent years, and Microsoft’s blunders have bolstered its success. Probably most people that have ditched Microsoft because of Vista have wisely switched to Apple. But this is good news for Linux, too.