Recalculating the Global Virtual GDP, Yet Again

Albania? Nepal? Try Lebanon, noob.

Very soon after the New York Times Magazine published my account of life in the gold farms of China, I received an email informing me of a ghastly factual error in the second paragraph. “I’m not quite sure who you are targeting with the article,” the message began, “but if it was any of the ‘8 million’ players of World of Warcraft they all stopped reading at ‘night-elf wizards.’” I was mortified, of course (everyone knows night elves can’t be mages, for God’s sake), but there was an upside. If all those WoW players were bailing at paragraph two, that meant 8 million fewer chances I’d be busted for the even more horrific cock-up several paragraphs below: My butchering of world geekery’s most vital economic statistic, the total gross domestic product of all known virtual economies.

The offending passage follows. It is a scene of wanton, unadulterated factual imprecision, but if you think you’ve got the stomach for it, read on…

“In 2001, Edward Castronova, an economist at Indiana University and at the time an EverQuest player, published a paper in which he documented the rate at which his fellow players accumulated virtual goods, then used the current R.M.T. prices of those goods to calculate the total annual wealth generated by all that in-game activity. The figure he arrived at, $135 million, was roughly 25 times the size of EverQuest’s R.M.T. market at the time. Updated and more broadly applied, Castronova’s results suggest an aggregate gross domestic product for today’s virtual economies of anywhere from $7 billion to $12 billion, a range that puts the economic output of the online gamer population in the company of Bolivia’s, Albania’s and Nepal’s.”

Longtime Terra Nova readers will recognize that “$7 billion to $12 billion” figure from a calculation I posted here in April 2005, and if you’re satisfied with two-year-old economic statistics, I suppose it will do. But even at the time, my calculus rested on a fairly wobbly foundation: a very rough estimate of worldwide RMT revenues (somewhere between $540 million and $880 million), which I multiplied by a presumed constant of 13.5 to get my total GDP. Since then, the hardworking folks at the Virtual Economy Research Network have come up with a much more reliable assessment of total RMT, and since I cited their figures elsewhere in the article, I have no excuse—none!—for not having plugged their superior data into my GDP formula and recrunched the numbers.

Forgive me if you can, econ geeks, and while I know there’s no making up for my failure to get the correct number into the pages of America's newspaper of record, please accept the following recalculation as a token of my regret:

So: Take VERN’s final calculation of $2.09 billion for total RMT, multiply by 13.5, and voilà. The official total worldwide virtual GDP now stands at $28.215 billion—several ranks higher than Albania and Nepal, all the way up in the lofty precincts of Lithuania ($29.784 billion), Sri Lanka ($26.794 billion), and Lebanon ($22.622 billion).

And you can take that to the bank.

Comments

[Wolfe]

1.

Now I dont know the math behind this type of extrapolation of a market, but I would find the conclusion of the total value of the Global Virtual GDP to suffer more from the part where all transaction that can be measured are of luxury goods.

Now its not easy for me to describe what I mean but I can try make a comparison. I'll use cars.

Imagine a world where only cars that sell for more than $1.000.000 were known to the person doing the math, and the total number of potential car buyers were a billion. Im sure there is some type of formula which looks complicated that describes how valuable the market for cars is which lies outside the known data points. (But im sure its likely to be quite wrong.) The only thing I know for sure is that the car market is worth less than a million billions of dollars.

[Sam]

2.

Why multiply the figure in the first place with a constant?

I would argue that the value of the traded figure depends very much on the rest not getting traded. In other words - the assumed non-availability of all non traded currency is what makes the traded currency so valuable. Does the nontraded amount have any value at all?

It could be argued that the Dollar amount is fixed (e.g. what subscribers are willing to spend given the current constants) - no matter how much virtual currency gets generated.

[Dennis Connolly]

3.

I might have to suggest avoiding that constant as well, and focusing more on a point to point basis: any given MMO's gold-farmer-to-normal-player rate will depend heavily upon the community itself as much as the developers. In Ragnarok Online, it's considered something of a norm to watch a dozen farming bots teleport away the moment you walk into a map; in Anarchy Online, there are very few gold or material farmers last I checked, perhaps due to the development situation; and in A Tale in the Desert, any attempts to sell certain...services has led to a very swift and brutal ban by the players.

In the latter, only one person has successfully maintained a sale of virtual goods for real-world money, and even then it's a focus on the *time* he provides, rather than the *materials*. That's out of a projected playerbase of over a thousand.

I'd have to agree with what I think Sam is saying here - there's a given 'maximum value' every subscriber has for the game on average, and that can be used to determine the saturation point for a given RMT economy. For Everquest, World of Warcraft, etc., it's high owing to status symbols; in other games, it could be low, or nonexistent.

[Julian Dibbell]

4.

Guys, you’ve raised good points, but as far as I can tell, your complaints are with the fairly standard econometric practices I’ve applied to virtual economies -- not with the way I’ve applied them. Let me see if I can explain.

One of these standard practices is “shadow pricing,” which is defined by Investopedia.com et al. as “the arbitrary assignment of dollar values to non-marketed goods.” Typically you’ll see shadow pricing come into play in cost-benefit analyses, where certain intangible costs or benefits typically have to be accounted for, as for instance the scenic views that are lost if a mining operation goes forward. Putting a price on this sort of good can be a black art for sure, but we do know that these goods have real value to people, and we can be pretty confident that a market structured to capture that value would arrive at some specific price for them.

More straightforwardly, sometimes the “non-marketed goods” in question are in fact marketed plenty, but in a way that makes it hard to find reliable prices for them, and in that case shadow pricing becomes a simple matter of locating some more transparent, “parallel” market and extrapolating from the going prices there. So for instance, in a command economy like Cuba’s or North Korea’s, where everyday prices are set by government fiat, you go to black markets and use the prices there to assign value to goods in the economy at large. Or in some barter-intensive rural economy you go to the nearest market town and find the price of melons or cows or whatever else gets bartered most and work back from there.

Now, unless I’m missing something, this is precisely the case of virtual items. Is every item traded in a game traded for real money? Not hardly, no, but shadow pricing doesn’t care about that. What matters is that every item, typically, has a knowable price in in-game currency, and that this currency, typically, has a well-known price in the RMT markets. Bada bing bada boom: there’s your shadow price.

Nor does it matter much, for the purposes of figuring GDP, that many items “produced” within a game (looted from mobs, crafted from mats) are never traded at all but end up consumed, instead, by the same players who produce them. As with crops grown and eaten by the grower, this is simply a case of subsistence production, and while GDP calculations typically don’t include subsistence goods, this is thought by many to be a bug rather than a feature. In principle, GDP measures all goods and services produced in a given region, and even in practice it includes the value of inventory accumulated by producers but not yet sold. Subsistence goods are excluded from GDP mainly because they’re harder to account for, not because they don’t belong there.

So: Is my approach starting to make more sense? I know I’ve left a few key questions hanging, but I want to see if I’ve at least cleared up some of them before getting to the rest.

[Moreton]

5.

If I may say so, the problem I have with most of the measures of virtual world GDP is that they ignore imports.

Let's consider the traditional formula for GDP, which is:

GDP=C+I+G+(X-M); ie:

GDP=consumption+investment+government spending+(exports-imports).

The production of goods in the virtual economy depends critically on the import of what could describe as services -- electricity, telecommunications charges, GMs, rent and so on.

One way to represent the cost of these imports would be to take the value of the monthly fees paid by players.

I don't have a figure for this total, but I suspect that once it's taken off we will find that the virtual world sector is not an economic powerhouse, but instead has a much, much lower GDP.

[Matt Mihaly]

6.

Julian: I'm no economist, but does shadow pricing really work in situations like this?

If I went ahead and copied Office a billion times and then went ahead and subsequently destroyed that billion copies, are you saying that the theory of shadow pricing would insist that the GDP of the US had just gone up by 250 billion or whatever?

How about if I've got a thousand bots farming for gold for my own personal private collection in WoW, and that gold will never leave my hands because I'm addicted to having big numbers in Blizzard's databases. Has the GDP actually gone up? If it has, what is GDP actually measuring that is useful in that case?

I can't put my finger on it but something seems wrong with applying shadow pricing to things that are so easily replicated/instanced.

--matt

[nate_combs]

7.

Julian > shadow pricing.

The real world doesn't shadow price all activity in the RW - e.g. the contribution of homemakers (which would be enormous), when computing GDP. Why should in-game transactions be shadow priced into RMT-based GDP calculations?

[Amarilla]

8.

you have a problem : except the game's CEO and high ranked staffers , nobody has any ideea about the real numbers involved ;
what you miss is : you dont know the actual number of players , of subscribers , you dont know Company's RL expenses , you dont know individual player's expenses , you dont know wich and how many " trades " are made unofficial- just private, between players ....yet you're making appreciations about...GDP :)
The only thing you can count on , is : any and every time an item " changes ownership " in game , its " traded value " is called ...." turnover ". And even on that, you have to believe Company's word on it. Get real, you dont have even 0.0001 % of that , in terms of real cash, changing hands for real.
"Why should in-game transactions be shadow priced into RMT-based GDP calculations?" .....i tell you why : to make the VWs looks more " wealthy " than they are. PR. To make you think : " woooow, i want a chunk from that big money ". Same as it was with the " woooooow , that game already have 2 godzillion of players there ,and i can play it for free... it must be something interesting there ".

[Julian Dibbell]

9.

@Moreton: That's a great point, actually. A proper virtual GDP really should factor in imports. But I'm not sure doing so makes much of a dent in the final numbers.

Ted's survey of EverQuest players, for instance, found them producing in-game wealth at a rate of about $3.50 worth of stuff per hour, and I'm guessing 30 or 40 hours per month is pretty typical play time across MMOs, which gives an average monthly production of around $100. Stack that against the $15 or less in monthly fees (and I think you're right to identify those with imports, just as I would identify exports with the RMT market itself), and GDP is still looking pretty healthy.

[Amarilla]

10.

Woah ! I have to pay the subsciption of only $ 15 , and then i can have $ 100 / month , just from playing EverQuest ?! And all that monies , just for playing 30-40 hrs/mo ?! Where do i have to apply , please ?

[Wolfe]

11.

An interesting approach to test to find an alternative method of this type of value could come from figuring out what type of entertainment expense is an equivalent between RL and Virtual entertainment.

For example I asked a fellow WoW player what type of price range and RL purchase would be likely to sit at to give an emotional impact roughly equivalent to that of getting the Season two arena weapon. He came up with a number around $600. I can see this as some type of psychological satisfaction value that probably can be attached to virtual gains relatively reliable.

He would never pay RL money to get hold of the item for various reasons, but the item does rob some retailer from another sale of a $600 toy, like a PS3.

From a decade of mmo gaming I have noticed my own habits changing. My own shopping habits are driven by practical reasons, not to feel satisfaction or pleasure. There are rarely any practical reasons for me to buy anything on a regular basis other than food.

It probably took my friend about a hundred hours of game "labour" to get his Season 2 weapon, altho he did of course get a lot of other little things during that journey which I dont know if they should be added to the value or part of it.

[Julian Dibbell]

12.

@Nate: Hmm, good question, but let me turn it around: Why *shouldn't* real-world GDP count unpaid housework? Plenty of respectable economists and social theorists think it should. So it's hardly a stretch to argue that GDP should also count unpaid virtual production. In fact, it's rather less far-fetched. Last I checked, no one's heard of households equipped with little domestic currencies that housewives can earn for their efforts and exchange explicitly for little household amenities or even real cash. MMOs, on the other hand -- well.

[Julian Dibbell]

13.

Oh, the irony. I've been trying to post a reply to Matt's message for the last half hour and Typepad keeps giving me this:

"An error occurred...
Your comment has not been posted because we think it might be comment spam. If you believe you have received this message in error, please contact the author of this weblog."

And I swear my response was less than 400 words!

[Amarilla]

14.

Shhhht ! Don't tell that to my husband !

[Matt Mihaly]

15.

Well keep trying Julian! I'm honestly interested as to how shadow pricing deals with someone doing something like copying Office over and over.

--matt

[Thomas Malaby]

16.

I think Matt's question is a great one, and points to a fundamental fact about virtual worlds (though most directly to those without contrived scarcity), and that is the vast reduction in production and distribution costs that are possible. In a UGC vw, a disproportionate (as against conventional material production) amount of labor by the producer is initial creation -- the creative act of conceiving of and building something. Once it is done and can be replicated and sent, a lot of the effort is already done. Sure, you have marketing and other costs, of course, so it's still a complex picture. The point is only that the overhead in market exchanges changes significantly, which potentially changes the relationship (and importance?) of market exchange to other kinds of exchange, such as reciprocity (social capital) and learning/authorization (cultural capital). This is a fascinating issue. I think what it means more directly for Matt's question is that there is a difference for imposed scarcity worlds (where one can imagine there is a consistent estimate of the labor over time required to acquire item x) vs. places like SL, where the imposed scarcity is (roughly) restricted to land and money.

[markov]

17.

No one has mentioned the Night Elf Wizard comment which deserves some discussion. If you are being contacted by the mainstream press to write articles, the community deserves a knowledgeable proponent out there. I am sorry to see that you could not get your facts straight when asked to submit a piece for the NYT. We are poorly served when one of our own can’t even get such a basic fact correct regarding World of Warcraft. Do you have an editor that has any understanding of games? I just wonder how this obvious error got through.

[Lavant]

18.

RMT measurements, Shadow prices, and the fear of inflation?

The comments here are interesting as they highlight the way we are trying to come to grips with the phenomenon of gold farming and RMT. I can’t help but consider the constraints of time and limited demand as perhaps more informative than the traditional assumption of infinite supply and the fear of inflation. It seems it is flawed to presume that were game economies like WoW to go unregulated, the economies would become more hyper inflated than post war Germany. For this is to assume away all but artificial scarcity, forgetting the constraints of time and the difficulty of the duplication. Not to mention limited (albeit growing demand). In response to Matt, I do think shadow prices apply to the duplication of software solely because time is inherently limited and the greatest constraint faced, and more importantly the opportunity cost of that time is large. You could copy Office all day, but the opportunity cost (forgone wages, time spent with family, eating, sleeping etc...) of doing so would become increasingly high. The same I believe is true of gold farming, at least from an average American perspective (versus low wage Chinese worker). You could farm gold all day and start a business with your friend, but most likely you could make a lot more money working at McDonald’s. And Chinese workers although they don't face this same opportunity cost do face the demand constraints of American and European buyers. Yet Julian and Jingle have revealed that gold farmers are underpaid when their .30hour wage is compared to the .69 nationwide avg.—opportunity cost in monetary terms of gold farming is quite high). Accordingly, I would guess (based on a bit of data) that in a completely open WoW we would probably see a lot of temporary inflation, until the economy has adjusted. Consider the inflation effect of the Burning Crusade. Gold prices plummeted following the January release (meaning inflation– more gold per dollar) but had stabilized shortly thereafter. As far as I know there were minimal bannings during this time (as opposed to last November). That 400 word limit might have just gone out the window so I will stop and ask to what extent Julian, Matt, Thomas and co think that scarcity would exist regardless of “divine” intervention on the part of developers?

399 words– I'm no prok ;)

[Matt Mihaly]

19.

Lavant wrote:

In response to Matt, I do think shadow prices apply to the duplication of software solely because time is inherently limited and the greatest constraint faced, and more importantly the opportunity cost of that time is large. You could copy Office all day, but the opportunity cost (forgone wages, time spent with family, eating, sleeping etc...) of doing so would become increasingly high.

What? All I need is a large array of harddrives and from there, I don't have to do a thing. There's literally essentially no time sink for me if I choose to make a billion copies of Office. The only problem would be storage medium and the fact that I don't have an array of harddrives built to store a billion copies is purely incidental to the discussion. Someone could, and I maintain that duplicating a billion copies of Windows either a) does not add anything significant to the GDP or b) does, in which case we need to find a new concept to use besides GDP since it's then kind of meaningless.

--matt

[Thomas Malaby]

20.

The scarcity is always there, Lavant, so it's true that we shouldn't make the mistake of thinking that the only scarcity that matters at all in virtual worlds is the imposed kind. But Matt is correct that, under certain conditions, the other kinds of costs in (re)production and distribution become incredibly small as compared to most other familiar contexts. What does this mean for GDP? Does it mean that the relationships between the kinds of resources we all use are in the process of being reconfigured to such a degree that we should start thinking about how to measure things like people's investment of effort over time in establishing social relationships? In acquiring the competencies and credentials that then enable them to produce material goods? These are not easy questions, but they lie at the heart of where all this is going, imho.

[nate_combs]

21.

Julian> Why *shouldn't* real-world GDP count unpaid housework? Plenty of respectable economists and social theorists think it should. So it's hardly a stretch to argue that GDP should also count unpaid virtual production.

----------------------------

At the very least I think if the real world GDP count included housework those GDP numbers will greatly increase, diminishing the relative position of the virtual worlds tally.

My real point, however, is this. Why should the agreements about what to count and what is what in real world GDP calculations directly map into in-game simulated actions (via shadow pricing)? In other words, yes there can be a virtual worlds GDP, but why is that GDP and not GDP' (prime) and why should GDP and GDP' be directly comparable?

I thought (disclaimer: speaking from a lay person's recollection) there were these gyrations economists did even with RW GDP calculations to compensate for economies that were less market-oriented (e.g. developing world - where folks may use currency less) to get them comparable.

If true, what does that say about those game worlds where the rules, motivations, and mechanisms of "market" may even be different. Why are these directly comparable?

For example, does a game world which features an internal market and uses currency more extensively have greater GDP per capita than one that doesn't?

[Julian Dibbell]

22.

Holy mother of God.

Matt, I just rewrote my reply, pasted it in to the comments box, got the error message again, retyped the thing in its entirety directly into the box, and got the frickin' error message yet again.

I will mail it to you directly. Feel free to repost it here, if you can.

[Julian Dibbell]

23.

@Markov: Dude. Either you are yanking my chain or you are an emissary from some distant WoW Forums thread where my "nelf mage" faux pas has blown up into the greatest media debacle since the Stephen Glass and Jayson Blair scandals. If it's the latter, I beg you: Take me to your leader. I don't want to have to explain more than once how a 5000-word article about a game as vastly complex as World of Warcraft could have possibly, in the fury of last-minute revisions, been infected by that microscopic, three-word glitch.

Besides which, have some pity: I am Horde 4 Life. It has been an almost physically painful ordeal for me to have to give a damn what kind of spells Night Elves can and can't cast. You Alliance weendogs should be grateful I mentioned them at all.

[Lavant]

24.

@ terranova

The gliches and error messages are worse than spamming. Always an improvement when the founders can't post.

@Julian
Lol?

[Wolfe]

25.

If the average price of 1 unit of crack is $100, and I create a crack factory where I produce a million units of crack. Will I add a hundred million dollars to my nations GDP when I give this crack away for free to anyone who visit my home?

Im sure the answer is no. The RMT market is also primarily based on the trade of contraband.

[Lavant]

26.

Um....Black market activities are not counted in the GDP, so we knew that to start. Yet black market goods have notable effects on nations economies whether or not they are counted in the officially sanctioned "GDP." If they could easily be accounted for crack factories unfortunately would be counted in the GDP...Hollywood might be doing that much better if it were...

[Ren Reynolds]

27.

wrt Julian's career limiting factual error about shadow-elves or dusky-mites or what ever they are called. I read in the NYT on Sunday that DDOS stands for: Digital Denial of Service (attack). Looks like they need a few geeks in the fact check dept.

[Wolfe]

28.

There should be several problems with the crack example.

1 - Im giving my crack away for free, practically reducing the average value of crack by a significant amount. Lets say my free crack supplies 98% of the nations crack demand and the cost is only a visit to my house. It will be hard to place any real weight behind the price you have to pay when buying a unit of crack on eBay for $100.

2 - Almost no one is interested in buying crack in the first place. A few people might be interesting in testing a bit of free crack but when they have to cash up for the crack they decide crack is not really anything they want to pay for.

3 - The maximum obtainable pleasure from crack is significantly less than warrants a supply of a million units within my market. By building a madly oversized crack stockpile "for free" I can hardly do anything that has a significant value. (This is as I can figure the same type of problem as Matt copying Windows a billion times.)

[Julian Dibbell]

29.

OK, folks, the case of the Post That Dared Not Speak Its Name has been solved, and I predict that students of dysfunctional online community dynamics (paging Clay Shirky) will be writing about this one for decades. Or laughing about it anyway:

Turns out Greg L got fed up with all the spam from WoW gold sellers a while back and set Typepad to block all posts with the phrase "WoW gold" in them.

The setting has been switched off. Welcome back, IGE.

[markov]

30.

@Julian,

Your reply made me laugh, thanks. Funny thing is, I am just a regular guy who plays some WoW. When I first read your article, I didn't realize it was someone like you, a blogger and gamer. You know, I don't want to give you too much trouble over a small error (especially since you are a Hordie), but *it matters*. That is all I am saying. You (we) lose some legitimacy with errors like that.

Anyways, thanks for responding with such a funny comment.

[Julian Dibbell]

31.

So here's what you all missed while we were working out our spam-blocking issues. First, I emailed the accursed post directly to Matt, as promised:

------------

So, Matt, fair warning, I'm not an economist either, so I can't tell you for sure how the pros would shadow-price your 1 billion copies of MS Office, but I strongly suspect they would look for guidance to conventional software-market practice and treat them as exactly what they appear to be, which is backup copies. For the purposes of calculating GDP, in other words, it's a straight shot: A billion copies of Office residing on one person's hard drive (or RAID, as the case might be) adds no more or less to the wealth of nations than a single copy does. The shadow price is effectively zero.

Where things start to get interesting is when one of those copies gets copied to another person's drive, and here we do see impressive shadow prices tossed around pretty furiously. Think of all those RIAA/MPAA/BSA press releases claiming "$3.45 trillion lost annually to digital piracy" and so forth. That may be a technically valid use of shadow pricing, but as us free-culture types like to point out, it's pretty dubious as a statement of fact, given how few of those "pirated" copies would have been purchased at full market price if freebies were harder to come by. (Not that we're nearly as hard on free-software advocates who try to work up equally questionable figures for the economic value of all those Linux and Apache suites installed out there.) So I think you're right to feel a little queasy about the applicability of shadow pricing to things like software.

That said, I don't really see how you make the leap from software apps to virtual items. They have a lot of features in common, definitely, but I really don't think "easily replicable" is one of them, and it's the crux. Sure, from where you sit as an MMO developer, virtual items may look arbitrarily replicable, because to you they are. But over on the client side, it's a very different picture, and it's over there that economists are going to look for the market value of virtual items. Yes, we can fantasize all we want about WoW players keeping a thousand gold-farming bots up and running just to scratch a gold-hoarding itch, but if Blizzard's attitude toward bots and farmers were such that that were anywhere near as simple a proposition as making a billion Office backups, I guarantee you the exchange rate for WoW gold would be measured in nanocents. In the meantime, though, and until Blizzard says otherwise, WoW gold *is* a scarce resource.

Artificially scarce, to be sure, and Lavant is right to point out that there are natural scarcities to be reckoned with in these economies as well. But let's not make the mistake -- a pretty common one in discussions like these -- of confusing the artificial with the imaginary. Just because the scarcity of virtual resources is contrived, that doesn't mean it's not real.

[Julian Dibbell]

32.

And then Matt goes:

------------

> That said, I don't really see how you make the leap from software apps
> to virtual items. They have a lot of features in common, definitely,
> but I really don't think "easily replicable" is one of them, and it's
> the crux. Sure, from where you sit as an MMO developer, virtual items
> may look arbitrarily replicable, because to you they are.

What if I, as a developer, create a trillion gold and distribute it among my my secret guildmates to use, with instructions not to reveal the existence of the gold to people outside the guild?

Gold is only a scarce resource from one point of view (the player's) though. The developers can create it at will (assuming there was any will to do so among those with the power to do so).

If gold is to be counted towards the GDP, does it matter how it was created? What about when a player finds a dupe bug and creates a billion gold in the blink of an eye? Has he increased the GDP? Is the GDP then decreased somehow (is that even possible?) when Blizzard deletes the gold? What about if Blizzard shuts the game down tomorrow?

[Julian Dibbell]

33.

So then I'm like:

------------

Anyway, look, now you're talking money supply, and that shrinks the distance between real and virtual economies even further (money being the original virtual item (but I digress)):

> What if I, as a developer, create a trillion gold and distribute it among my
> my secret guildmates to use, with instructions not to reveal the existence
> of the gold to people outside the guild?

Keep its existence secret, sure, but the only way to do that is to keep it from circulating outside your guild, in which case it remains fairly valueless and goes into the GDP as a zero (think backup copies again). Once you start spending it outside the guild, the secret's out: Even if nobody knows where the gold came from, the markets know it's there, and they start adjusting the price of gold downward to compensate. By the time you're done disbursing your stash, the price of gold has dropped to the point where, all else being equal, the total value of the newly bloated money supply is exactly what it was before you started bloating it, and again, you've added net zero to the economy's total wealth. What you *have* done is exploit your first-spender advantage to move a bunch of that wealth from other people's hands into yours. You're like the first "investors" in a Ponzi scheme, the ones who get rich and leave everyone else holding the bag.

The even better analogy, though, is to governments that, in various ways, overproduce money in order to enrich themselves and their favored constituencies at the ultimate expense of the private sector. If the government is as corrupt as you and your guildies, this is just scoundrelry. But you can also call it deficit spending, which in the right hands can be a perfectly legitimate tool for redistributing wealth and can even be used, so the Keynesians argue, to stimulate general economic growth. Either way, though, it's basically a stealth tax, which itself adds nothing to GDP.

> Gold is only a scarce resource from one point of view (the player's) though.
> The developers can create it at will (assuming there was any will to do so
> among those with the power to do so).

Again, this is old news:

[A nation's currency] is only a scarce resource from one point of view ([the private sector's]) though. [The government] can create it at will (assuming there was any will to do so among those with the power to do so).

> If gold is to be counted towards the GDP, does it matter how it was created?
> What about when a player finds a dupe bug and creates a billion gold in the
> blink of an eye? Has he increased the GDP?

Dupers are counterfeiters. Same exact thing. It's the private-sector's attempt to get a piece of that deficit-spending action, and it likewise has zero effect on GDP. Remember: GDP measures production of wealth, not money. Increase the money supply without a corresponding increase in wealth, and you've done nothing that affects GDP.

> Is the GDP then decreased somehow
> (is that even possible?) when Blizzard deletes the gold?

Second verse, same as the first: Decrease the money supply without a corresponding decrease in wealth, and the GDP doesn't change at all.

> What about if
> Blizzard shuts the game down tomorrow?

Then all in-game production goes poof, and GDP zeroes out. It's an economic catastrophe, absolutely. But here the conversation often derails into another one of those common confusions, in this case the one where people somehow get the idea that the relative contingency of wealth in virtual economies -- the fact that it depends on the infrastructure of a single company rather than of an entire region -- means it can't be real wealth.

Nonsense. Chernobyl had a GDP once. Now it doesn't. All that means is that Chernobyl in 1985 was a very bad investment -- not that the wealth produced there previously had been a figment all along.

[Julian Dibbell]

34.

@Markov: For the Warchief!

[Lavant]

35.

Well said Julian.

Another silenced post:

@Matt I couldn't agree more that vws have tremendous cost reducing effects and possibilities. Where I disagree is the semantics of your office example and the analogy drawn between Office and WOW gold. [as Julian has now pointed out] Its comparing apples and oranges. First of all, multiple copies of office are useless and worthless on a single computer. Think Network Effects. The more computers have copies the better. And the constraints here are the number of computers with Office, the time and effort it takes to get them there and the facilitation of their interaction..Costs which are not zero...Major transaction costs here despite technological ease of duplication. People seldomly want to use the same network.

Your analogy: there are significant monetary differences between office and the new currencies of the realm. Wow Gold is money: a store of value, unit of account and a medium of exchange regularly traded against the dollar: It is volatile, but reliably valuable nonetheless. While Office is potentially valuable it is not money or actively converted to money with the ease and liquidity of gold (at least Office stored as ISOs or what have you on a computer somewhere—It means nothing to have more than one image of office on a computer whereas it means a lot to have more copies of gold on Blizzard’s servers. Where office becomes valuable is by spreading these copies to as many people as possible and these people then using the program to write Terra Nova posts. Back to WOW: a billion gold should be considered part of GDP and Office cannot be in the same way. Perhaps we need a standardized way of measuring the value of network effects (subscribers is informative, but inherently weak measure of the actual value created). So much wow gold value derives from Wow’s network effects...the ability to bring large numbers of people together....how do you measure a game’s ability to bring people together? And perhaps the we should be looking at different forms of capital besides the monetary kind...perhaps we should brainstorm measurement techniques?

[Simeon (UK)]

36.

Dear Julian

I thank you for upsetting all those over sensitive WOW fans! However, it has been brought to my attention that the biggest journalistic sin you made was to make facts up re info from the University of Indiana. Is it true that there is no University of Indiana? None the less a good article!

[Julian Dibbell]

37.

To my undying shame, Simeon, it's true: There is no University of Indiana. There is an Indiana University, which in all respects but its name corresponds to the institution I described in the article. But I don't expect a sharp tack like you to believe me if I say the similarities are anything but sheer coincidence. So there you have it: I made this "University of Indiana" up out of whole cloth. Now what'll it take to keep you quiet about this? Hm? Name your price.