An Economic Model for the Trade
in Free Goods & Services on the Internet

Rishab Aiyer Ghosh

< what is value, or: is the internet really an economy? >

Much of the economic activity on the net involves value but no money. Until a few years ago, there was almost no commercial activity on the internet. The free resources of the net still greatly outweigh all commercial resources. It is quite hard to put a price on the value of the internet's free resources, at least in part because they don't have prices attached. They exist in a market of implicit transactions.

< the economics of gossip >

Every snippet posted to a discussion group, every little webpage, every skim through a FAQ list and every snoop into an online chat session is an act of production or consumption, often both. There is no specific economic value inherent in a product. Value lies in the willingness of people to consume a good, and this potentially exists in anything that people can produce and pass on.

Even bad writing and even junk mail are parts, however reprehensible, of the internet's economy, but let's look at a more obvious case, Linux. After all, software, in particular large operating-system software occupying up to six CD- ROMs when distributed offline, is undeniably an economic good (for example, Red Hat Software). And Linux, with its loosely organized community of developer-users and its no-charge policy, undeniably has an economic logic that seems, at first, new.

< something for nothing? >

Linus Torvalds did not release Linux source code free of charge to the world as a lark, or because he was naive, but because it was a "natural decision with- in the community that [he] felt [he] wanted to be a part of " (quoted from personal correspondence with Torvalds). Any economic logic of this community -the internet-must be found somewhere in that "natural decision." It is found in whatever it was that motivated Torvalds, like so many others on the net, to act as he did and produce without direct monetary payment.

Of course, it is the motivation behind people's patterns of consumption and production that forms the marrow of economics. Figuring out what motivates, let alone measuring it, is always difficult but it is even tougher when price tags don't exist. It is simpler just to assume that motivations only exist when prices are attached, and not attempt to find economic reason in actions motivated by things other than money; simpler, therefore, just to assume as we often do that the internet has no economic logic at all.

This is wrong. The best portions of our lives usually do come without price tags on them; that they're the best parts imply that they have value to us, even if they don't cost money. The pricelessness here doesn't matter much, not unless you're trying to build an economic model for love, friendship, and fresh air. On the internet, through much of its past, the bulk of its present, and the best of its foreseeable future, prices often don't matter at all. People don't seem to want to pay-or charge-for the most popular goods and services that breed on the internet. Not only is information usually free on the net, it even wants to be free, so they say.

But free is a tricky word: like love, information-however free in terms of hard cash-is extremely valuable. So it makes sense to assume that the three million people on the internet who publish about matters of their interest on their home pages on the web, and the several million who contribute to communities in the form of newsgroups and mailing lists, and of course anyone who ever writes free software, believe they're getting something out of it for themselves. They are clearly not getting cash; their "payment" might be the contributions from others that balance their own work, or something as intangible as the satisfaction of having their words read by millions around the world.

While writing my weekly newspaper column on the information society, Electric Dreams (, I was distributing an email version free of charge on the internet. A subscription to the e-mail column was available to anyone who asked, and a number of rather well known people began to receive the column each week. My readers often responded with useful comments; I often wondered whether people would pay for a readership like this. Having many readers adds to your reputation; they make good contacts, helping you out in various ways. Simply by reading what you write, they add value to it-an endorsement, of sorts. So who should pay whom-the reader for the work written, or the writer for the work read?

The notion that attention has value is not new and has been formally analyzed in the advertising industry for decades. The "attention economy" has been described in the context of information and the internet by M. Goldhaber in "The Attention Economy" and R. A. Lanham in "The Economics of Attention." It would be facile to suggest that attention necessarily has innate value of its own. However, more often than not, attention is a proxy for further value. This may appear in the form of useful comments (or bug reports from Linux users), assistance, and contacts, or simply as an enhanced reputation that translates into better access to things of value at a later point.

Even those who have never studied economics have an idea of its basic principles: that prices rise with scarcity and fall in a glut, that they are settled when what consumers will pay matches what producers can charge. These principles obviously work, as can be seen in day-to-day life. But that's the "real world" of things you can drop on your toe. Will they work in a knowledge economy? After all, this is where you frequently don't really know what the "thing" is that you're buying or selling, or clearly when it is that you're doing it, or, as in the case of my column, even whether you're buying-or selling. Contrary to what many doom-sayers and hype-mongers suggest, it always seemed to me that the basic principles of economics would work in an economy of knowledge, information, and expertise. They are, after all, not only logical on the surface but also practically proven over centuries-a powerful combination. Even if the internet appears to behave strangely in how it handles value, there is no reason to believe that if it had an economic model of its own, this would contradict the economic principles that have generally worked. However, if a textbook definition of economics as the "study of how societies use scarce resources to produce valuable commodities and distribute them among different people" remains as valid now as ever, almost all the terms in there need reexamination (P. A. Samuelson and W. D. Nordhaus, Economics, 15th ed., NY: McGraw-Hill, 1995). This is because the same peculiar economic behavior of the net suggests that it has developed its own model, the economic model of the information age.

The Times of India sells some three million copies every day across India. The whole operation, particularly the coordination of advertising and editorial, depends on RespNet. This internal network won the Times a listing in ComputerWorld magazine's selection of the world's best corporate users of information technology. RespNet runs on Linux and other similar free software off the net.

Raj Mathur, who set up Linux on RespNet, agrees with Torvalds when the latter says, "people who are entirely willing to pay for the product and support find that the Linux way of doing things is often superior to 'real' commercial support." This is thanks to the large community of other developers and users who share problems and solutions and provide constant (sometimes daily) improvements to the system. The developer-users naturally include operators of networks similar to RespNet. So many of them can separately provide assistance that might not be available if they were all working together in a software company-as Linux Inc.-where they would be producers of the software but not consumers. This shifting base of tens of thousands of developers-users worldwide working on Linux means that the Times of India would have a tough time figuring out whom to pay, if it wanted to.

The fact that people go looking for other people on the internet, and that Linux developers look for others like them, is just one instance of the immediacy of much of the trade that takes place on the net. When you post your message to rec.pets.cats, or create a home page-whether personal or full of your hobbies and work-you are continuously involved in trade. Other cat-lovers trade your message with theirs, visitors to your homepage trade your content with their responses, or perhaps you get the satisfaction of knowing that you're popular enough to get a few thousand people discovering you each week. Even when you don't charge for what you create, you're trading it, because you're using your work to get the work of others (or the satisfaction of popularity) in a discussion group through your website. What is most important about this immediacy of the implicit trades that go on all the time on the net is its impact on notions of value. Unlike in the "real world," where things tend to have a value, as expressed in a pricetag, that is sluggish in response to change and relatively static across its individual consumers, on the net everything is undergoing constant revaluation. Without the intermediary of money, there are always two sides to every transaction, and every transaction is potentially unique, rather than being based on a value derived through numerous similar trades between others-that is, the pricetag.

As we continue to alternate between examples from the worlds of free software and usenet-to reiterate their equivalence in economic terms-we can see the two- sided nature of trade in this hypothetical example about cats. You may value the participants in rec.pets.cats enough to post a long note on the nomadic habits of your tom. In a different context-such as when the same participants are quarreling over the relative abilities of breeds to catch mice-you may not find it worthwhile contributing, because the topic bores you. And you may be far less generous in your contributions to rec.pets.dogs. You value the discussion on dogs, and catching mice, much less than a discussion on tomcats, so you're not willing to make a contribution. This would be "selling" your writing cheap; but when you get feedback on tomcats in exchange for your post, it's the right price.

Unlike noodles and bread, readers on internet newsgroups don't come with pricetags pinned on, so commonplace decisions involving your online acts of production require that you figure out the relative values of what you get and what you give, all the time. Others are figuring out the worth of your contribution all the time, too. Life on the internet is like a perpetual auction with ideas instead of money.

That note on your tomcat probably does not deserve the glorious title of idea; certainly the warm feeling that you got in exchange for posting it- when people responded positively and flocked to your homepage to see pictures of your cat-couldn't possibly be classed with "real ideas." Still, for the sake of convenience the subjects of trade on the net can be categorized as idea (goods and services) and reputation (which when enhanced brings all those warm, satisfied feelings, and more tangible benefits too).

Ideas are sold for other ideas or an enhanced reputation; reputations are enhanced among buyers of ideas, and reputations are themselves bought and sold all the time for other reputations, as we shall see later. The basic difference is that reputation (or attention) is, like money, a proxy. It is not produced or consumed in itself, but is a byproduct of the underlying production of actual goods ("ideas" in our binary terminology).

< two sides to a trade >

Unlike the markets of the "real world," where trade is denominated in some form of money, on the net every trade of ideas and reputations is a direct, equal exchange, in forms derivative of barter. This means that not only are there two sides to every trade, as far as the transaction of exchanging one thing for another goes (which also applies to trades involving money), there are also two points of view in any exchange, two conceptions of where the value lies. (In a monetary transaction, by definition, both parties see the value as fixed by the price.)

As the poster of notes on tomcats, the value of your posting something is in throwing your note into the cooking pot of participatory discussion that is rec.pets.cats and seeing what comes out. As the author of a page on cats, what you value in exchange for your words and photographs is the visits and comments of others. On the other hand, as a participant on rec.pets.cats I value your post for its humor and what it tells me to expect when my kitten grows up; as a visitor to your webpage I learn about cats and enjoy pretty pictures.

When I buy your book about cats, it's clear that I am the consumer, you the producer. On the net, this clear black-and-white distinction disappears; any exchange can be seen as two simultaneous transactions, with interchanging roles for producer and consumer. In one transaction, you are buying feedback to your ideas about cats; in the other, I am buying those ideas. In the "real world" this would happen in a very roundabout manner, through at least two exchanges: in one, I pay for your book in cash; in the next, you send me a check for my response. This does not happen very often! (The exception is in the academic world, where neither of us would get money from the Journal of Cat Studies for our contributions; instead our employers would pay us to think about cats.)

As soon as you see that every message posted and every website visited is an act of trade-as is the reading or publishing of a paper in an academic journal-any pretense is lost that these acts have inherent value as economic goods with a pricetag.

In a barter exchange the value of nothing is absolute. Both parties to a barter have to provide something of value to the other; this something is not a universally or even widely accepted intermediary such as money. There can be no formal pricetags, as an evaluation must take place on the spot at the time of exchange. When you barter you are in general not likely to exchange your produce for another's in order to make a further exchange with that.

When the contribution of each side to a barter is used directly by the other, it further blurs the distinction between buyer and seller. In the "real world" barter did not, of course, take place between buyer and seller but between two producer-consumers in one transaction. When I trade my grain for your chicken, there's no buyer or seller, although one of us may be hungrier than or have different tastes from the other. On the internet, say in the Linux world, where it may seem at first that there's a clear buyer (the Times of India) and an equally clear, if aggregate seller (the Linux developer community), there is, in fact, little such distinction.

Just as the existence of the thousands of independent Linux developers are valuable to the newspaper because they are also users of the product-and may face similar problems-other Linux developers welcome the Times of India because the way it faces its problems could help them as Linux users.

< can you eat goodwill? >

Perhaps you will agree that when you next post a note on cats, you're not giving away something for nothing. But what you get in return is often pretty intangible stuff-satisfaction, participation in discussion, and even answers to cat-related questions are all very well, and may be fair exchange for your own little notes, but don't seem substantial enough to make much of an economy. As for Linux-it's fine to talk about a large base of user-developers all helping one another, but what has all this brought Linus Torvalds? Although Linux did get vastly improved by the continuing efforts of others, none of this would have happened without Torvalds's original version, released free. Assuming that he's not interested in Linux as a hobby, he's got to make a living somehow. Doesn't he seem to have just thrown away a great product for nothing?

First, let's see what intangible "payment" Linux brought Torvalds. In the circles that might matter to Torvalds's career, he's a sort of god. As government and academic participation has declined as a proportion of the total internet developer community, most recent "free" technology has not been subsidized, either. The main thing people like Torvalds get in exchange for their work is an enhanced reputation. So there are, in fact, lots of net gods.

Net gods get hungry, though, and reputation doesn't buy pizzas. So what does Torvalds do? He was in the University of Helsinki in October 1996, when I first interviewed him (he's now with a U.S. company where "it's actually in [his] contract [to do] Linux part-time"). "Doing Linux hasn't officially been part of my job description, but that's what I've been doing," he says. His reputation helped: as Torvalds says, "in a sense I do get my pizzas paid for by Linux indirectly." Was this in an academic sense, perhaps? Is Linux, then, just another of those apparently free things that has actually been paid for by an academic institution, or by a government? Not quite. Torvalds remained in the university out of choice, not necessity. Linux has paid back, because the reputation it's earned him is a convertible commodity. "Yes, you can trade in your reputation for money," says Torvalds, "I don't exactly expect to go hungry if I decide to leave the university. 'Resume: Linux' looks pretty good in many places."

< is reputation a convertible currency? >

Suppose you live in a world where people trade chicken and grain and cloth-a very basic economy indeed! Suddenly one day some strangers appear and offer to sell you a car; you want it, but "Sorry," says one of the strangers, "we don't take payment in chicken; gold, greenbacks, or plastic only." What do you do? It's not hard to figure out that you have to find some way to convert your chicken into the sort of commodities acceptable to car dealers. You have to find someone willing to give you gold for your chicken, or someone who'll give you something you can trade in yet again for gold, and so on. As long as your chicken is, directly or indirectly, convertible into gold, you can buy that car.

What holds for chicken in a primitive barter economy holds also for intangibles such as ideas and reputation in the part of the economy that operates on the internet ("Implicit Transactions Need Money You Can Give away," ED 70). And some of these intangibles, in the right circumstances, can certainly be converted into the sort of money that buys cars, let alone pizzas to keep hunger away. This may not apply to your reputation as a cat enthusiast, though; it may not apply to all software developers all the time, either.

On the internet-indeed in any knowledge economy-it is not necessary for everything to be immediately traded into "real world" money. If a significant part of your needs are for information products themselves, you do not need to trade in your intangible earnings from the products you create for hard cash, because you can use those intangibles to "buy" the information you want. So you don't have to worry about converting the warm feelings you get from visits to your cat webpage into dollars, because for your information needs, and your activities on the net, the "reputation capital" you make will probably do. "The cyberspace 'earnings' I get from Linux," says Torvalds, "come in the format of having a network of people that know me and trust me, and that I can depend on in return. And that kind of network of trust comes in very handy not only in cyberspace." As for converting intangible earnings from the net, he notes that "the good thing about that you still have them even though you traded them in. Have your cake and eat it too!"

There is, here, the first glimpse of a process of give and take by which people do lots of work on their creations-which are distributed not for nothing, but in exchange for things of value. People "put it" on the internet because they realize that they "take out" from it. Although the connection between giving and taking seems tenuous at best, it is in fact crucial. Because whatever resources are on the net for you to take out, without payment, were all put in by others without payment; the net's resources that you consume were produced by others for similar reasons-in exchange for what they consumed, and so on. So the economy of the net begins to look like a vast tribal cooking pot, surging with production to match consumption, simply because everyone understands (instinctively, perhaps) that trade need not occur in single transactions of barter, and that one product can be exchanged for millions at a time. The cooking pot keeps boiling because people keep putting in things as they themselves-and others-take things out. Torvalds points out, "I get the other informational products for free regardless of whether I do Linux or not." True. But although nobody knows all the time whether your contribution is exceeded by your consumption, everyone knows that if all the contributions stopped together there'd be nothing for anyone: the fire would go out. And that wouldn't be fun at all.

< cooking-pot markets >

If it occurred in brickspace, my cooking-pot model would require fairly altruistic participants. A real tribal communal cooking pot works on a pretty different model, of barter and division of labor (I provide the chicken, you the goat, she the berries, together we share the spiced stew). In our hypothetical tribe, however, people put what they have in the pot with no guarantee that they're getting a fair exchange, which smacks of altruism.

But on the net, a cooking-pot market is far from altruistic, or it wouldn't work. This happens thanks to the major cause for the erosion of value on the internet-the problem of infinity. Because it takes as much effort to distribute one copy of an original creation as a million, and because the costs are distributed across millions of people, you never lose from putting your product in the cooking pot for free, as long as you are compensated for its creation. You are not giving away something for nothing. You are giving away a million copies of something, for at least one copy of at least one other thing. Since those millions cost you nothing, you lose nothing. Nor need there be a notional loss of potential earnings, because those million copies are not inherently valuable-the very fact of there being a million of them, and theoretically a billion or more-makes them worthless. Your effort is limited to creating one-the original-copy of your product. You are happy to receive something of value in exchange for that one creation.

What a miracle, then, that you receive not one thing of value in exchange- indeed there is no explicit act of exchange at all-but millions of unique goods made by others! Of course, you only receive "worthless" copies; but since you only need to have one copy of each original product, every one of them can have value for you. It is this asymmetry unique to the infinitely reproducing internet that makes the cooking pot a viable economic model, which it would not be in the long run in any brickspace tribal commune.

With a cooking pot made of iron, what comes out is little more than what went in-albeit processed by fire-so a limited quantity can be shared by the entire community. This usually leads either to systems of private property and explicit barter exchanges, or to the much analyzed "Tragedy of the Commons".

The internet cooking pots are quite different, naturally. They take in whatever is produced, and give out their entire contents to whoever wants to consume. The digital cooking pot is obviously a vast cloning machine, dishing out not single morsels but clones of the entire pot. But seen one at a time, every potful of clones is as valuable to the consumer as were the original products that went in.

The key here is the value placed on diversity, so that multiple copies of a single product add little value-marginal utility is near zero-but single copies of multiple products are, to a single user, of immense value. If a sufficient number of people put in free goods, the cooking pot clones them for everyone, so that everyone gets far more value than was put in.

An explicit monetary transaction-a sale of a software product-is based on what is increasingly an economic fallacy: that each single copy of a product has marginal value. In contrast, for each distinct product, the cooking-pot market rightly allocates resources on the basis of where consumers see value to be.

< a calculus of reputation >

A crucial component of the cooking-pot market model is reputation, the counterpoint to ideas. Just as money does not make an economy without concrete goods and services, reputation or attention cannot make an economy without valuable goods and services, which I have called "ideas," being produced, consumed, and traded).

Like money, reputation is a currency-a proxy-that greases the wheels of the economy. Monetary currency allows producers to sell to any consumer, without waiting for the right one to offer a needed product in barter exchange. Reputation encourages producers to seed the cooking pot by providing immediate gratification to those who aren't prepared to pull things out of the pot just yet, or find nothing of great interest there, and thus keeps the fire lit.

Money also provides an index of value that aids in understanding not just individual goods (or their producers), but the entire economy. Reputation, similarly, is a measure of the value placed upon certain producer-consumers-and their products-by others. The flow and interaction of reputation is a measure of the health of the entire cooking-pot economy.

Unlike money, reputation is not fixed, nor does it come in the form of single numerical values. It may not even be cardinal. Moreover, while a monetary value in the form of price is the result of matching demand and supply over time, reputation is more hazy. In the common English sense, it is equivalent to price, having come about through the combination of multiple personal attestations (the equivalent of single money transactions).

Money wouldn't be the same without technology to determine prices. Insufficient flow of the information required for evaluation, and insufficient technology to cope with the information, have always been responsible for the fact that the same things often have the same price across all markets.

The management of reputation is far too inefficient today to be a useful aspect of a working economy. Its semantics are poorly understood; moreover, it has nothing remotely akin to the technology that determines prices based on individual transactions in the monetary economy.

< conclusion >

The common assumption that the net feels at home with free goods and vague trade because its population is averse to money, altruistic, or slightly demented is wrong. It is becoming more obviously so as floods of "normal" people arrive from the world outside, and initiate themselves into the ways of the net.

An economic model based on rational self-interest and the maximization of utility requires the identification of what is useful-sources of value-as well as a method of expressing economic interaction. In the cooking-pot market model, while scarcity creates value, value is subjective, and may therefore be found in any information at all that is distributed on the net. The cooking-pot model provides a rational explanation (where a monetary incentive is lacking) for people's motivations to produce and trade in goods and services. It suggests that people do not only-or even largely-produce in order to improve their reputation, but as a more-than-fair payment for other goods-"ideas"-that they receive from the cooking pot. The cooking-pot market is not a barter system, as it does not require individual transactions. It is based on the assumption that on the net, you don't lose when you duplicate, so every contributor gets much more than a fair return in the form of combined contributions from others.

Reputations, unlike ideas, have no inherent value; like money, they represent things of value, as proxies. Reputations are crucial to seed the cooking pot and keep the fire lit, just as money is required to reduce the inefficiencies of pure barter markets. However, reputations require a calculus and technology for efficient working, just as money has its price-setting mechanisms today.

The cooking-pot model shows the possibility of generating immense value through the continuous interaction of people at numbing speed, with an unprecedented flexibility and aptitude toward intangible, ambiguously defined goods and services. The cooking-pot market already exists; it is an image of what the internet has already evolved into, calmly and almost surreptitiously, over the past couple of decades.

The cooking-pot model is perhaps one way to find a rationale for the workings of the internet-and on the net, it finds expression everywhere.

A longer version of this text was first published in First Monday, the peer-reviewed journal of the Internet, in 1998 (


about Rishab Aiyer Ghosh >>