MMO GDP, QED (WTF?)

It's Estonia, Basically. Or Maybe Cote D'Ivoire. OK, Malta.

Last August, our intrepid econometer Dr. Ted gave us a back-of-the-envelope GDP for the virtual world economy that pegged it at about $4.74 billion, or roughly level with Namibia. It’s time for an update.

I bring new numbers to the task, and a new methodology. Where Ted took his early finding of a $2000 GDP per capita for EverQuest and generalized from there, I start with a different piece of his data: the ratio between real-money trades, or RMT (an excellent bit of jargon just introduced by commenter FantasyMeiser) and total GDP. For EverQuest, Ted gave an eBay market total of $5 million, but for total RMT it’s probably best, as a rule, to double that, for $10 million. The GDP was $135 million, and the RMT:GDP ratio, therefore, was 1:13.5. Assuming that ratio holds constant across all MMOs, we can then extrapolate from any given RMT to a GDP 13.5 times as large.

But what figure to use for total worldwide RMT? Ted has given a “conservative” estimate of $100 million, but again that’s the total for eBay sales (plus ItemBay in Asia), so best to double that to $200 million. On the other hand, Steve Salyer, CEO of IGE (and yes, the score is 5 so far for those of you playing TLA Bingo) (well, 6 now) says it’s $880 million. Ted is probably more objective, Steve is probably better informed, so let’s split the difference and call it $540 million.

Worldwide MMO GDP, therefore, is $540 million times 13.5, or a booming $7.29 billion!

Now for the rankings. But here we run into a bit of a philosophical problem. The traditional method for comparing one gross domestic product to another simply translates the two GDPs into U.S. dollars at current exchange rates and ranks them accordingly. But this turns out to be a poor way to compare relative wealth, since a dollar buys a lot more in Mexico, say, than it does in France. So now your hipper economist prefers to reckon GDP values by indexing the local prices of a fixed basket of goods, thus measuring what’s known as purchasing power parity (or PPP) (do I hear “bingo”?), and this gives strikingly different results from the exchange-rate method.

Great, well, we’re nothing if not hip here at TN, so we’ll go with PPP, right? Not so fast. If we were going by a strict role-playerish or magic-circularist conceit here, in which players live exclusively in their virtual worlds while they are playing, then yes, PPP might make a certain sense, since players would value virtual goods exclusively in terms of what they could be traded for in-game. But the truth (at least to a ludological “realist” like myself) is that players only ever live liminally in their worlds, always half in and half out, and always semi-conscious, therefore, of the opportunity cost of their virtual goods – i.e., the real-world purchasing power that they forgo by not selling them off for dollars. Obviously, for a Chinese player that cost is much greater than for a French player, so in that sense, PPP for a virtual world is literally all over the map, and therefore essentially incalculable.

Obviously too, of course, that cost fluctuates according to the player’s mindset. A farmer, for instance, will only value his goods in terms of their real-world price, while a hardcore role-player might be so unlikely ever to eBay her stuff that its effective opportunity cost must be discounted almost to zero.

So let’s be agnostic about this, and go with a smorgasbord of rankings, according to player type:

Role players and other valiant defenders of the magic circle may go with the PPP rankings, which place the virtual world economy 134th among the planet’s 185 national economies, right between Benin ($7.49 billion) and Malta ($7.04 billion).

Farmers, powergamers, and other dedicated breakers of immersion should use the exchange-rate rankings, which move things up to 101st place, between Jamaica ($7.88 billion) and Estonia ($7.28 billion).

And finally, IGE fanbois, whoever you are, should feel free to plug Steve Salyer’s numbers into the formula, for a total GDP of $11.88 billion and an exchange-rate ranking of 86th place – not quite as good as Panama ($12.127 billion) but better than Côte d’Ivoire ($11.12 billion) and, for that matter, over half of the world’s sovereign nations. You go!


Comments on MMO GDP, QED (WTF?):

Gordon Calleja says:

SO my country's (Malta) GDP is lower than that of virtual worlds. I thought it was people being generally lazy that made standard of living sucky there.

Posted Apr 22, 2005 6:15:14 PM | link

A axe says:

You have a fallacy - values are simply inflated due to the grey-market nature of selling. Once it becomes legitimized [akin to what SOE did], prices will fall more than the demand will raise.

Posted Apr 23, 2005 1:42:40 AM | link

Edward Castronova says:

Excellent stuff. The methodology is well within the paradigm of contemporary economic value theory (for what it's worth).

And if you're a naysayer - the principle is, when people create things of value, it's worthy of measurement and reporting. If you don't like the possibility that synthetic worlds have the same economic impact as Panama, you're going to have to argue that the magic swords created by human manipulation of this software have no value at all. And you're going to be in trouble, especially if you make your living by selling stuff you write in C, Java, Perl, or MS-Word.

Posted Apr 23, 2005 9:26:24 AM | link

Bernardo Sanchez says:

Is there someway to estimate the growth this virtual world economy has? And if it continues to grow, what type of in-game services could show up, other than vgear commerce and services like leveling? Could this be a new frontier for on-line services, and i mean the more traditional services?

Posted Apr 23, 2005 10:25:36 AM | link

bruce boston says:

"And if you're a naysayer - the principle is, when people create things of value, it's worthy of measurement and reporting."

Ok, I'll bite.

Unless I'm mistaken there is a ton of value being created in RL that never ends up in GDP, PPP or any other TLA bingo spaces.

Examples include;

While a maid gets paid for housecleaning, the average person doing the exact same activities does not. Maid's work: included in GDP; average person's work: nope.

While a professional pizza deliver gets paid to drive from the pizza place to someone's house, if an average person goes to pick up the pizza he doesn’t. Pizza guy’s driving: included in GDP, average guy's driving: nope.

If I take my car to a gas station to get the oil changed I pay, but my Dad changes his own oil. My oil change: included in GDP, my Dad's: nope.

If someone calls a plumber to come fix the sink vs. the guy that does it himself, again, same principle.

Even, Julian writing an article for Wired vs Julian writing an article for TerraNova. GDP, not GDP.

This basic principle has a tendency to get exaggerated the more 3rd world you go, mainly because those countries tend to have a much lower degree of specialization when compared to more modern countries.

For example, while fish taken to the market and sold might end up in GDP, fish eaten on the boat, fish given to family/friends/extended family/local religious leaders/the guy that fixed your sink last week/etc/etc might not show up in the GDP calculation.

I think we may also be assuming that the 'paid' work via eBay etc does not show up in local GDP. But, the minute anyone pays taxes on their earnings, they show up in the local GDP, but this may be a different topic for another time.

-bruce

Posted Apr 23, 2005 10:46:48 AM | link

bruce boston says:

Ok, so I did have more.

Before any tax happy government agencies get too far in this article, I wanted to add one other concept to these discussions, and that is the concept of ‘value added’.

It’s my understanding that gross revenue, or gross sales don’t always end up in GDP calculations. For example, the estimated future market of $880M may be the total amount of sales of virtual items in the future, but that may not accurately map back to GDP numbers.

So let’s take two real world examples;

First, do used car sales end up in GDP calculations? Say you have a car and you sell it 12 times, does the total sales number get included in GDP? If I remember correctly, no, just the difference between the buy price and the sell price of each transaction. There is a fair amount of difference if you take that into consideration. My guess is that the average mark-up in virtual goods is about 40%, so we may be looking at a number smaller than the $880M when we see how that contributes back to GDP calculations. (note: I also agree with A axe, and I think we can expect the mark-up percentage to decrease in the future, as such I’m not sure what the average mark-up will be by the time we get to the $880M market)

The second example is more specific, and that is removing the cost of production from the actual revenues of the goods produced. My guess is that before IGE reposts to their local tax agency how much they are contributing to the local GDP, they are allowed to back out a number of expenses, including cost of rent, cost of computers, cost of software, cost of subscriptions, eBay fees, advertising fees, etc, etc. Take this concept a step further, and if you really want to see what the average MMO gamer is contributing to GDP I think you would have to take their average RMT revenue, and back out things like cost of computer, cost of software, cost of subscription, etc, etc. Frankly, I don’t think there are that many gamers that would end up in the black if we did it that way, and as such I think the amount that they are contributing to any RL consistent GDP number is trivial at best.

Now, if the Taxman would please step away from the virtual communities, these are not the droids you are looking for.

-bruce

Posted Apr 23, 2005 11:44:58 AM | link

[email protected] says:

Good point - since we don't include unpaid work in the GDP, should the total not be more near the $800 million figure?

Though, really, unpaid work *should* be included in GDP. However, we'd probably find that the relative position of the MetaVerse would not change in those circumstances.


Posted Apr 24, 2005 6:06:28 AM | link

blaze says:

If we did include unpaid labour and if the GDP per capita was 2000, I think most would end up in the black.

Posted Apr 24, 2005 6:11:56 AM | link

[email protected] says:

Hmm, interesting:

From the wikipedia:

For instance, buying a Renoir doesn't boost GDP by $20m. (If it did, buying and selling the same painting repeatedly to a gallery would imply great wealth rather than penury.)

Note that the Renoir purchase would effect the GDP figure, but not as a $20m receipt, the auctioneer's fees would appear in GDP under Consumer expenditure, because this is a final service.


So, really, the total GDP here then is orthogonal to RMT (which are really just renoirs (heh) getting passed back and forth) and the GDP is actually the brokering fee that IGE or EBAY or whoever charges.

So really, we need to measure the yearly differential in content that people are buying to get a better sense of GDP, and *not* the total purchased.

I suspect this is where Steve gets his numbers (he runs a currency exchange, right?) and why $800 million isn't that hard to doubt - unfortunately his take on that $800 million is probably pretty meagre.

Another example is the gamingopenmarket.com. The total "RMT" is $1492427.78, however their percentage of that is somewhere around 1-3% (they lowered their fees somehwere along the way) so it is ~25K or so which is their contribution to the GDP (not 1.4 million or whatever)

Posted Apr 24, 2005 6:24:33 AM | link

[email protected] says:

In case it wasn't clear, and I realise I really wasn't, by running a currency exchange I mean people are speculating on the rise on fall of currencies, ie: buying and selling the same renoir.

I know for a fact people speculate on the GOM. I've done it myself.

Posted Apr 24, 2005 6:59:43 AM | link

[email protected] says:

last one, I promise :)

What if I spend 800 USD to buy a megalevel character and then went on, over the course of the year, to generate 800 USD worth of plat and items.

What is the GDP there? I'm assuming net zero.

Posted Apr 24, 2005 7:09:20 AM | link

Edward Castronova says:

Guys, don't get hung up on the definition of GDP in the textbooks or in US government surveys. The beancounters decide what things of value get counted and what don't, and they make a lot of those decisions based on practicality. Some things are a pain in the butt to measure. Fortunately, in our system, almost everything of value does get traded, so we can use trades to approximate the aggregate of value creation. But don't then look at definitions of approximations and say that we MUST adhere to those definitions. No, you have to keep returning to the underlying concepts. This is especially so when most of the economy is subsistence - most things of value are consumed by the producers. If you used our RL approximations on that economy, you'd be missing 90 percent of the value creation.

And value creation is where money-making, usiness strategy, policy deicsions, individual decision-making, etc., should focus. NOT on the approximations.

What we care about is economic rationality, equity, and well-being. When Jiffy Lube changes your oil, it counts in GDP. When your dad does it, it does not. Are you telling me that if we outlawed at-home oil changes that there'd be no loss of value? Are you going to argue that all of the at-home parenting that goes on has no economic value?

Think about value and recognize that, in the MMOG context, using the standard RL approximations results in a huge underestimate of what's going on.

Posted Apr 24, 2005 8:51:51 AM | link

blaze says:

I agree.. unpaid labor should be counted.

However, the question does remain .. if I spend $800 USD to buy an avatar and only produce $800 USD worth of plat in a year.. where is the economic value in that?

Or, what if I buy something for $100, play a month, sell something for $100 and then buy something else for $100, and then repeat this 12 times over.

Aren't we double (triple, quadruple, etc) counting the same thing over and over again?

Finally, if all I'm doing is simply buying and selling currency in order to perform market arbitrage, well that's a lot different than someone who sold $400 USD of plat and then went and bought himself a very nice ipod.

On $50 alone I could easily buy and sell $10,000 worth of currency over the space of a few months.

Posted Apr 24, 2005 9:20:21 AM | link

Josh de L-H says:

So then, at the end of the day, just how big is the secondary market for virtual property?

Posted Apr 24, 2005 8:35:12 PM | link

bruce boston says:

"If you used our RL approximations on that economy, you'd be missing 90 percent of the value creation."

As you do in most 3rd world economies.

I think the point is that if we are going to try and estimate where "virtual economies" would rank in the following list:
http://www.worldbank.org/data/databytopic/GDP.pdf, then, I think it's prudent, if not requisite, that we use the very same calculations that were used to determine the ranking of every other entry on that list.

And, yes, I think this means that we should not be including in our calculations of the aggregate size of virtual economies anything that would not be included in the calculations that we are comparing against.

I think Blaze also makes a good point, and that is when accounts/characters/items are resold, we should not be including this in our calculations. Else, maybe we should be including things like daily commodity trading volumes, daily currency exchanges, and daily stock trading volumes in the RL economy numbers. Don't think I'd recommend this, but at some point, we should be using consistent calculations before we start ranking things.

Also, I'm still not sure how we are ensuring that we are only measuring 'added value'. Again, why are we not backing out the directly related production costs, like computer hardware, game software, subscription costs, connection costs, etc.

In fact, I think if we were to do this right, I think we would need to be removing in-world production costs as well.

If it costs 2 bars of truesilver, a star ruby and a mithril tube to produce a sniper scope, then we shouldn't be counting the value of the materials and the scope in our calculations, I think we should only be including the value of the scope. Unless of course that scope was attached to a Thorium Rifle, then I would say we should only include the value of the rifle w/scope. That was unless….the rifle was used to go hunt Colbalt Mageweavers and try and find a Arcanite Dragonling Schematic. Then we should only be including the value of the schematic and any loot gained there. That is unless, that loot and schematic were used to build an actual Arcanite Drangling. Then we really should only be counting the value of the Arcanite Dragonling, and not the value of any of the loot used to build it, or the value of the rifle used to hunt for the schematic, or the value of the scope or the value of any of the components used to create the rifle.

I think you get where this goes.

Because the sole purpose of 99% of everything in MMORPG Virtual Economies is basically the production costs for what we hope to build the next time we log on, it all gets subtracted, except for the current value of the current items in-world.

Basically, I think that most likely the best way to determine the size/value of virtual economies is to add the relative value of the accounts/characters/items currently in that economy, and then of course subtract the RL costs incurred in the production process as listed above (computers, software, subscriptions, connections, etc)

To get an annual number, I think we would then need subtract the total economic value that was present 1 year previous, and subtract it from the current calculation. Surely if an Arcanite Drangling was created last year, just because it was traded this year shouldn't mean that we include it in this year's GDP ranking number.(granted all Drangonlings are less than a year old at this time, but thinking forward).

My guess, if you did the above, we would be in the red. Basically, a house of cards. This is why the government, thankfully, currently calls someone that buys a $1500 computer, $50 in game software and then pays a gaming company $15 a month, a consumer, not a micro business.

Listen, even at a gut level, I’m not sure I understand the Namibia claim. Let’s take everyone that we know on a personal level that has made a buck off of MMORPG economies in the last five years, and then compare them to everyone that we know, on a personal level, that spent a bunch of money buying game quality computers, buying MMORPG games, paying monthly subscriptions, and paying for gaming quality internet connections. Maybe it’s just me, but the spending far outweighs any virtual income being made. Maybe I’m in the wrong crowd. But, if that is the activity at the ground level, how are we not surprised that adding a bunch of negative numbers, results in one big negative number?

-bruce

Posted Apr 24, 2005 10:23:32 PM | link

Tobold says:

What you are saying is that if 13.5 swords of uberness drop per month, and one of them is sold on the black market for $100, while the other 12.5 are kept by their owners or given to friends for free, the GDP rises by $1,350 dollars. I can't believe that is orthodox economics.

Either you only count the RMT as GDP, in which case the GDP goes up by only $100. Of you consider what the price would be if all 13.5 swords of uberness would be traded on a legal market, come to a value of maybe $10, and the GDP goes up by $135.

The sword of uberness has a value, but it is *not* $100. Because the $100 are divided into the actual value of the sword, plus the "transaction cost". The transaction cost is high on black markets, often much higher than the actual value. This completely distorts the calculation.

Example: Why is a joint more expensive than a cigarette? The marihuana plant is as easy, if not easier, to grow than the tobacco plant. It is simply the fact that the trade of it is illegal, which holds down the volume and up the price. And if you plant a field of marihuana, but without selling the product, you don't add (biomass of the field times cost of a joint) to the GDP.

Posted Apr 25, 2005 8:27:59 AM | link

Staarkhand says:

I'll chime in with a 'me too' post - I'd like to hear Ted or another (former?) economist respond to the notion that the blackmarket nature of these sales is skewing the calculations.

With SonyExchange on the horizon this seems especially significant.

Posted Apr 25, 2005 11:01:34 AM | link

magicback says:

Not an economist, but I think the amount of production not counted in GDP is called the informal sector or informal economy. UN, World Bank, IMF, etc. has done a lot of research on this area. But I don't think they have found a good way of measuring it.

But based on conventional GDP measurements, the GDP would be much lower than $11 billion. $1 billion at most.

Frank

Posted Apr 25, 2005 12:13:15 PM | link

blaze says:

Also, I'd like to take at the "creating value" comment.

I think only in a very meagre way. I think what they do in SecondLife is creating value (scripting, texturing, prim building), but what they do in EQ? WoW? I don't see it.

Posted Apr 25, 2005 10:48:37 PM | link

Edward Castronova says:

Too many things in here to respond to. Not sure where to start. One thing I will say: early in this blog's life, and then following the Namibia post, this debate about GDP has been done to death. I stand by the argument I made there.

I hate having to enter lecturing mode, but many of you are making more less classical missteps that I have encountered for years in econ principles classes. There's no point in naming names, but -

You have to distinguish between sales and value creation

You have to let the markets decide what's value and what isn't

You have to focus on core principles and try to find measures that work in the economy you're studying

Now, I don't have time to teach a macro principles course here. I'd love to have a debate with another economist at some point, but none has emerged - it surprises me, but so far, that's how it is. I am sure I've gotten a bunch of stuff wrong. Macroeconomics is not my gig.

But i stand firmly behind the bottom line: I don't care how you cut it up, if you have 10-20 million people with an incomes base drawn from post-industrial economies, devoting something like 20 hours a week to the crafting of playthings for themselves and their friends, that's a lot of value creation. And there's no doubt in my mind it rivals value creation in some jurisdiction of this Earth, whether it's Namibia or Panama or Boise, Idaho.

You can try to push back on the latter conclusion but you're not going to move me. In order to zero out or trivialize that account, folks, you have to fall back on an argument that virtual goods are just not real. It's Monopoly money that does not translate into any persistent real wealth at all. On the day that RMT vanishes, I will agree with you. Until that day, it's all real money to me.

Posted Apr 26, 2005 8:03:32 AM | link

Brask Mumei says:

About that value creation, blaze.

Let's say I was psychokinetic with the unique power to build fences. I could build hundreds of meters of fine quality of fence with a mere thought. Three seconds of my work could fully fence your yard.

As it happens, you want a new fence. You ask me how much it would cost for me to create the fence. I refuse. You then ask Joe to build you the fence. He's got to use his own two hands, so charges you $x000 for the job.

How can we say Joe has done anything of value? It could have been done with a few idle thoughts by Brask! Indeed, Brask could equally as easily dissassemble the fence, undoing all of Joe's hard work!

This is the case of Swords of Leetness in EverQuest. For a small group of people (the developers, system administrators, etc), Swords of Leetness are trivial to create. They, however, refuse to give you a Sword of Leetness. They create the artifical scarcity for Swords of Leetness. And, since scarcity is scarcity (artifical or not), creating Swords of Leetness by killing the Boss Dragon is creating value.

Posted Apr 26, 2005 8:27:52 AM | link

bruce boston says:

Dr. Castronova,

"this debate about GDP has been done to death. I stand by the argument I made there."

Yes it has. That said, I think that with Sony's recent announcement the stakes of this debate have increased significantly. And, to clarify, I think my concern is really with the terminology than with the actual conclusions being drawn there from.

In my mind, 'GDP = Taxable income' and that has little to do with if value is being created or not.

At the industry level, not that I have any say at that level, but at an industry level I don't think there is any benefit, either to the gamers or the developers, that could come from telling RL governments that there is a huge 'taxable' revenue stream that they have yet to tax. Not only does StationExchange open the industry to a very accurate way to measure trading activity, as does eBaying, but SE now has the possibility of moving the taxation that is very real on eBay, to within the games themselves.

At the end of the day, what do gamers have to lose if we drop the use of the term GDP, and instead replace it with terms like ‘markets’, ‘industry’ or ‘value’ even.

In the business world, you sign one piece of paper and use one set of language and the government charges you one tax rate, you sign another piece of paper and use other language and they charge you another rate. I would hate for us to use the wrong language, and sign the wrong pieces of paper and have gamers taxed as a result.

"But I stand firmly behind the bottom line: I don't care how you cut it up, if you have 10-20 million people with an incomes base drawn from post-industrial economies, devoting something like 20 hours a week to the crafting of playthings for themselves and their friends, that's a lot of value creation."

Totally agree. Outside of taxable income calculations, it doesn't matter what people are doing with their time, be it creating open source software or farming virtual light leather, at the end of the day it’s all value.

"Now, I don't have time to teach a macro principles course here. I'd love to have a debate with another economist at some point, but none has emerged"

That's unfortunate. I can understand the thrill of debating with RL economists; I certainly enjoy it. On the other hand, if they are not so interested in what you have to say, I would hope you understand that you have the eyes and ears of a ton of gamers that spend 20-40hrs a week in these worlds, and the eyes and ears of a ton of developers that spend 60-80hrs a week building them. While RL economists may someday find what you have to say groundbreaking, I think the players and developers have much more riding on these worlds than even you may realize, and would greatly appreciate the lecturing.

-bruce

Posted Apr 26, 2005 5:18:01 PM | link

Tobold says:

On the day that RMT vanishes, I will agree with you. Until that day, it's all real money to me.

I never said that there is zero value. But if RMT is your only pillar on which to base GDP, then GDP = RMT, and not GDP = 13.5 x RMT.

How people with an incomes base drawn from post-industrial economies spend their time has nothing to do with value creation. Some of them are hitting little white balls with an iron club over an artificial landscape. But the only contribution of that to GDP is the cost of equipment and golf fees they pay, you can't add the time they spend on the golf course multiplied with their hourly salary to GDP as well.

Posted Apr 27, 2005 7:24:17 AM | link

Julian Dibbell says:

Damn, you guys are hard for an old geezer like me to keep up with. So much to respond to!

But I’ll start with Bruce’s excellent point about the essential incommensurability of RL and virtual GDPs, because I think it gets to the heart of what’s going on here, and of why this GDP game of Ted’s is so productive and illuminating. One reason Ted and I like crunching these numbers is that they help convey to non-gamers the reality of people’s existential investment in the games they play. But speaking for myself, I also like how the numbers help convey to non-economists (like me) the un- and/or surreality of the data economists play with.

So Bruce asks, roughly, why it’s fair to compare Norrath’s GDP to Bangladesh’s, say, when Bangladesh’s official GDP leaves out vast swathes of subsistence-level and informal economic activity. The correct answer being: Who said it was fair? It’s kind of absurd, in fact. Moreover, I’m surprised no one picked up on the insanity of my even thinking about applying a purchasing power parity (PPP) standard to my rankings. To even begin working with PPP, after all, you have to be able to pick a comparable basket of goods and services available in all the economies you’re measuring. Um, OK, well, let’s see, they have gold ingots in Britannia and gold ingots in Britain – how about those? (Rimshot, please!)

But wait, here’s the punchline: As queer as these comparisons may be, the difference between them and the ones that determine the World Bank’s official rankings is ultimately just a difference of degree. No, it isn’t fair to compare the official Bangladeshi GDP to Ted’s Norrathian GDP, but by the same token, and for pretty much the same reasons, it’s not much fairer to compare it to the French GDP.

As for relative purchasing power, no, you can never find a good that compares perfectly between a virtual world and the real one, but the same holds true for any two real countries as well. For instance, there is probably not a more globally uniform but locally made product than the Big Mac (which is why The Economist magazine uses it for its yearly Big Mac Index of world purchasing power), yet even Big Macs inevitably vary from country to country, not only in quality and taste but in the ambient dining experience that is a big part of what you pay for at McDonalds. The differences might not matter much to the average consumer, perhaps, but surely a true Mickey D’s connoisseur might crave the American Big Mac yet feel indifferent toward the Romanian.

With all these numbers, in other words, economists do a lot of ontological fudging. Yes, they may concede, in reality the French and Bangladeshi economies are different beasts, in reality no two hamburgers are exactly alike. But for our purposes, they may also reasonably insist, a GDP is a GDP is a GDP and a Big Mac is a Big Mac is a Big Mac. The objects of our study aren’t the real objects themselves, in all their infinite detail, but abstract models of them that incorporate enough of their defining data to be as good – for our purposes – as the real objects. It’s not the world itself we deal with, but for our purposes it might as well be. It’s as good as the world, it’s effectively the world, it is virtually the world.

It’s a, you know, a whaddayacallit, a virtual world.

Posted Apr 28, 2005 1:01:26 PM | link

Julian Dibbell says:

Bruce also said:

Basically, I think that most likely the best way to determine the size/value of virtual economies is to add the relative value of the accounts/characters/items currently in that economy, and then of course subtract the RL costs incurred in the production process as listed above (computers, software, subscriptions, connections, etc)

And while I think the suggestion is generally sound, I don't agree that it makes sense to deduct those RL costs. Sure, it might if we were measuring gamers’ contributions to their RL national GDPs. But we’re not. We’re measuring their contributions to the GDPs of the games they play. And from that perspective, all those purchases of computers, software, monthly access, etc., aren’t so much about cost of production as about balance of trade. Do all the balancing, and you discover that they have almost no net effect on the GDP we’re measuring.

Here’s what I’m saying. Just like real economies, the virtual economy has imports and exports. Through RMT, it exports epic mounts, swords of uberness, and other domestic products. On the flipside, it imports – what? What products does the virtual world require that cannot be produced within that world? Well: computers, software, monthly access, etc. Without these, it doesn’t exist. They are the guarantors of its territorial integrity, and like some third-world country buying weaponry from Lockheed, Grumman, and other U.S. companies, it must buy them from Sony, Electronic Arts, Dell, and other RL companies.

And that, in principle, means wealth must leave the virtual economy to pay for those imports. But does it in practice? Generally speaking, no. What happens, instead, is that players pay for those outside necessities with money they make outside the game. In effect, the players’ go to their day jobs like Turks going to Germany, Filipinos to Dubai, Mexicans to the US, sending back all the dollars necessary to pay for the technological infrastructure that sustains their virtual homeland. So: the RL economy demands payment from the virtual world, the RL economy provides the money to make the payment. Net drag on the virtual economy: zero.

Posted Apr 28, 2005 4:22:28 PM | link

bruce boston says:

Hi Julian,

Thanks for the reply. I think you make some great points here.

"One reason Ted and I like crunching these numbers is that they help convey to non-gamers the reality of people’s existential investment in the games they play."

Understood.

"Um, OK, well, let’s see, they have gold ingots in Britannia and gold ingots in Britain – how about those? (Rimshot, please!)"

lol. At the same time, I don't disagree with the basic methodology that Dr. Castronova has been using. For example if players are deciding between spending US$8 on a meal at Mickey D's or spending the same US$8 on a new piece of armor in Norrath, then I think it's fair to say, at least in economics terms, that in the minds of those visiting Norrath, that these two items have equal value.

"No, it isn’t fair to compare the official Bangladeshi GDP to Ted’s Norrathian GDP, but by the same token, and for pretty much the same reasons, it’s not much fairer to compare it to the French GDP."

True, but at least we use a similar calculation when we make any attempt to compare the two.
GDP = consumption + investments + exports - imports

"As for relative purchasing power, no, you can never find a good that compares perfectly between a virtual world and the real one, but the same holds true for any two real countries as well."

Agreed.

"Just like real economies, the virtual economy has imports and exports. Through RMT, it exports epic mounts, swords of uberness, and other domestic products."

'hmmm. Don't exports need to be exported before they are counted as exports? If a guy in Florida owns a car in Cuba, and sells it to his neighbor in Florida, it doesn't count as exports unless the car actually gets shipped out of Cuba.

(A)"And that, in principle, means wealth must leave the virtual economy to pay for those imports. (B) But does it in practice? Generally speaking, no". What happens, instead, is that players pay for those outside necessities with money they make outside the game. In effect, the players’ go to their day jobs like Turks going to Germany, Filipinos to Dubai, Mexicans to the US, sending back all the dollars necessary to pay for the technological infrastructure that sustains their virtual homeland. (C) So: the RL economy demands payment from the virtual world, the RL economy provides the money to make the payment. (D) Net drag on the virtual economy: zero."

So there is a ton here. First, let me add two quick reasons why I have always had a hard time with these types of calculations.

1) One thing that has always perplexed me is how economics are measured. Outside of when new money is printed, in reality or even in virtual reality, the balance of payments is always zero. The amount of currency that is paid by someone and the amount of currency that is received by someone in any transaction is always equal. And, while that may seem obvious, I think that fact makes the decision, of deciding what transactions should actually count in what economic bucket, very hard.

2) Trade is particularly hard to decide. I think there is over $2 trillion in currency trades everyday, but worldwide GDP is only like $45 trillion. This doesn't even include commodity trading, or stock trading or Pokemon card trading. If you really want to have fun, next Christmas offer to give your Wife $5 million if she promises to give you the same. The balance of the payments will be equal, but the two of you can now go around telling everyone that you got $5M from the other for Christmas, and it will actually cost less than an actual gift. (just as a note, I wouldn't tell too many people, this trade may be subject to a RL gift tax)

Point A, seems to suggest that most virtual worlds are running a trade deficit with the real world.

Point B is a bit fuzzy to me. While I understand the concept I’m not sure it works well. First, when Turks go to work in Germany their productivity is counted in Germany’s economy. Second, how do we tell the difference between citizens of VWs going to RL jobs to work, and citizens of RL going to VWs to play? I think it could be argued that VWs act similar to Island Resort economies, where players pay airlines to take them, and the tourists drop a bunch of cash to have a good time. This cash is later used to buy the goods/resources that the island needs to continue to support new would be tourists. Certainly this is in line with some of the work Betsy Book has been doing with regards to virtual tourism.

Point C: agreed

Point D: Not so sure. In fact, if we hypothesize that virtual economies ultimately run an increasing trade deficit with the real world (my response to point A), I think this would help explain two to three other phenomena that we see in the space today. First, why virtual worlds ultimately shut down, (basically they become unable to pay back their RL trade deficits) and second, why virtual currencies tend to deflate vs real world currencies, and maybe even three, why virtual world economies tend to over produce currency in the first place.

These last two events also seem to pop-up in the real world when a country runs a high trade deficit. First in an effort to pay increasing debts both to its citizens and ultimately to outside creditors it’s fairly common for countries with high trade deficits to increase their money supply. While this works in the short-term, ultimately the market catches up with the currency and it deflates dramatically vs outside currencies. Much like what we see too often in virtual worlds, today. And, like in RL, this crash of the currency often leads to migration to places with more stable currencies, and the eventual implosion of that economy.

So again, I’m not sure I would say the net drag is zero, that would seem to suggest that virtual worlds are self-perpetuating, something that goes against point C.

Also, as far as exports are concerned, I think that most players would say that they are playing/paying for the fun factor, or at least one of the archetypical cousins of the fun factor. This seems congruent with thinking that the players are actually citizens of RL, going to VWs to play.

Again, I think that maybe the best way to measure the economic growth of virtual worlds is to measure the gross amount of assets that are there today that were not there last year, and subtract the on-going trade deficit that doesn't seem to be going anywhere.

One interesting conclusion that we might find, is that we may be able to predict the health of a virtual world by measuring the assets being created, assessing their value via RMT markets, and checking it against the estimated trade deficit (cost of computers, software, subscription, connection). My guess, we may find that when the virtual worlds are in the 'red' that the days of that world and its citizens are numbered.

Another interesting conclusion that may come from this is the current design of storage in most MMORPGs. If it is the raw value of assets that matters, certainly limited storage (like bank slots) has a huge influence on the amount of value that a virtual world can hold.

-bruce

Posted Apr 28, 2005 7:37:47 PM | link

bruce boston says:

Back again,

After thinking about it a few minutes;

I think that a RL Trade Deficit Theory (VW Assets - RL Trade Deficit = Value of Economy) may also help to shed light on a few other recent events.


The excitement of P2P virtual worlds; because they have the possibility of dramatically reducing the to-date inherent RL trade deficit incurred because of RL costs in non-P2P worlds. Decreased RL Trade Deficit = greater economic stability = greater chance for longevity, decreased cost/risk of customization/production >>> increased production.

The fear of RMT markets: This goes back to Staarkhands' comment "the notion that the blackmarket nature of these sales is skewing the calculations." As markets get more efficient, the value of VW items could quite possibly decrease. If in fact we are able to predict the health of an economy by using RMT markets to assess virtual asset values, the more efficient the trading markets are, the less the total value of those assets may be. Japan had a similar mixed response to Toys-R-Us putting huge efficient toy stores in Japan. While the prices dropped due to increased efficiency, the added loss of jobs/small toyshops etc, didn't happen without some deal of pain.

-bruce

Posted Apr 28, 2005 8:40:19 PM | link

Julian Dibbell says:

Interesting stuff there, Bruce, and much to respond to. But quickly, since you're going in a lot of stimulating directions with the whole balance of trade thing, I want to say that my understanding of it seems to be quite different from yours.

First, I draw the opposite conclusion about MMO trade deficits. The way I see it, not only is it the real world that's running the deficit, but it's very unlikely that the balance will ever shift in the opposite direction.

Why? Because of all that "overseas" income. Forget for a moment the question of whether day-job wages get credited to the virtual GDP or not. Let's say they don't. Even so, it's those wages that pay for the vast majority of the virtual world's "imports."

So whether you look at that money, on the one hand, as foreign receipts credited to the virtual economy then immediately sent back to the RL side of the ledger to cover the imports or, on the other hand, as generous foreign aid picking up the tab, the net impact on the virtual economy is the same: the game world gets all its imports effectively for free. And therefore the virtual economy runs a surplus approximately equivalent to RMT.

Approximately -- because don't forget that some players do pay the odd subscription bill by selling off some plat or credits or whatever. So that shaves the surplus a bit, insofar as payment for that imported access is coming straight from the virtual economy. And don't forget the farmer types, whose payments for just about everything comes from virtual economic activity.

Now, if every player were like them, paying all their RL gaming expenses with the wages of their in-game activity, then we could start talking about deficits on the virtual side of the ledger. But I think that's a long way off.

Posted Apr 28, 2005 10:42:58 PM | link

magicback says:

Of this discussion, you have to considered the fact that someone in RL is getting the RL$ in a RMT.

RMT of production in virtual worlds have at least two RL participants (excluding self-dealing). VW are just PPE or Fixed Assets in conventional measurements.

We can value VW production via RMT-implied valuation and come to a theoretical GDP, but the VW-RL balance of trade discussion needs to include the counterparty to the RMT.

One prospective in measurement is to look at VW as a virtual factory and the collective outputs are virtual objects.

Posted Apr 29, 2005 12:59:31 AM | link

bruce boston says:

Hi Julian,

So let's start with currency trades, as they are a major part of RMT.

Pure currency for currency trades, as far as I know, never get included in either 'GDP like calculations' or Balance of Trade calculations. Why? Because before the trade and after the trade, the gross balances in both currencies are equal. The net gain/loss of currency on both sides is zero. For example, if I have 500G and I sell it for $50, while I gain $50 and lose 500G, the person I am trading with loses $50 and gains 500G, and the net gain/lose of the two of us aggregated together is zero.

I think that, typically, currency trades are only used to compare relative values, not as value trades themselves. So at minimum I think we will need to exclude all currency trades right off the top from any RMT numbers used to determine BOT value exchanges.

As far as Foreign Aid is considered, typically, it is counted as a Debit, not a credit in terms of BOT calculations (see http://www.investopedia.com/terms/b/bot.asp). Thus, it’s not free money, the higher your foreign aid, the higher your deficit.

I also think that for BOT calculations it may be useful to define who is who. For example, are we domestics or foreigners in terms of a VW economy? Is my avatar separate from me, or an agent of me? How about someone from IGE? How about Sony? How about NPC vendors? How about other players?

One rule that I would hope we followed is that if we define someone as either a foreigner or a domestic that we stick to that definition for the whole BOT calculation.

Also, I think that any classification that we give to ourselves for playing the game, we should also give to someone at IGE playing the game. For example, if we call ourselves domestics, then I think that someone from IGE playing the game would then be classified as the same. As such, sales between two domestics for local property, even if the sale is denominated in a foreign currency, shouldn't be showing up in the ‘exports’ pile of a BOT calculation, as has been suggested above.

And, one last question, have there been any real world economies that had a high trading surplus and imploded? Or even, whose currency had a tendency to deflate over time?

And, yes, I agree there are way too many questions than we have time to address here.

-bruce

Posted Apr 29, 2005 3:27:02 PM | link

bruce boston says:

"Now, if every player were like them, paying all their RL gaming expenses with the wages of their in-game activity, then we could start talking about deficits on the virtual side of the ledger. But I think that's a long way off."

Actually, I think this is reversed. The inverse of Foreign Aid, ie money from VWs going to support RL costs, would be the inverse of the Foreign Aid Debit, ie a Credit, or a trade surplus.

I may have this wrong, but if this is correct, than what I think is a long way off is a BOT trade surplus.

-bruce

Posted Apr 29, 2005 3:34:54 PM | link

Martin Im says:

I read the old thread and posted there before realizing that I was at the wrong place. Not sure the posting will give any input here, but just to have something to say I'll repost it here.

I guess all these questions arise because we are searching for VW analogies with nation states and other social constructs, the worlds as we know it today. There are a number of implications when calculating GDP and counted one single time for a single country (or whatever) doesn't give us any kind of answer to anything. Calculated for Sweden in 2005, and only Sweden and only in 2005, says nothing really. How you choose to do it and what you decide to include impacts the result and the accuracy of the calculation. But more important is that it's done the same way between countries and between years. Otherwise it's a completely useless figure. I have my doubts whether it's possible to actually compare a VW with a nation state (for one they are much more volatile) but I have no doubts about the potential economical impact of VW, and there will certainly be a way how to calculate the value of the players actions in real currency. So more important is to establish a generally accepted way of calculating GDP (or what it might be called) for VW and then collect this data over a number of years while understanding how this information relates and can be compared to GDP of countries. If a straight comparison isn't possible there's always the option to adjust for this or that (like in all statistics). But then again GDP is such a rough figure and there a so many other fun things to look at in VW too.

And about scarcity, you'll find a lot of artficially sustained prices in RL too. We tend to forget that there is no such thing as a perfect market.

Posted May 3, 2005 8:23:31 AM | link

Zeke says:

GDP calculation for MMORPGS is interesting, but not valid given any of the above methods. Taking the PPP, especially, is ludicrous, given that you don't live inside the game world. Even by more pedestrian methods, what you're doing is calculating the total economic value of an online game or community by extrapolating top dollar value paid by the subscribers who are willing to engage in transactions with real life dollars across the whole inventory of items/game money. This assumes away a great deal of market elasticity which I suspect exists (i.e., if you open up supply by allowing freer trading of goods on real dollar exchanges, the price of items won't fall).

A good analogy can be made with those trading card games like Magic. Some players were willing to pay hundreds of real dollars for rare cards which yielded in-game advantage and prestige, whereas others would only trade other cards, not willing to put down real bucks for the game. It would be obviously false to say that all the magic cards in existence were worth the sum of the "market value" for each card at the swap-determined rate. If that were true, the publishing co. could've sold the cards separately, outside of the opaque cellophane packages, and raked in the dough. But actually, if the cards had been sold separately, it is more reasonable to estimate that the value of most cards would've fallen. That is to say, the trading, which in real life is profitable b/c of the law of comparative advantage, takes place in games b/c of artificially created comparative advantage.

To avoid fallacy, it is best to think of standard MMORPGs as black box economic units. Their value is determined by the real cash flows associated with them, both from subscriptions and to some degree online trading of in-game items for real bucks (conditioned on the desire of players to "win" or to achieve prestige inside the game community). These cash flows could certainly increase if more companies embraced open real trading, but not by the ridiculous multiples suggested by the GDP analogy.

An important exception occurs where gamers actually create items themselves, and don't "earn" them through a treadmill process (where the designers have already created an item but withold it for game enjoyment purposes). If gamers actually script items in-game, like in SL, it could be much more valid to talk about in-game economy in a real sense, as opposed to a zero-sum player management sense.

Posted May 8, 2005 3:53:14 PM | link

blaze says:

Sorry, I know this is dead and buried but nonetheless:

"economies, devoting something like 20 hours a week to the crafting of playthings for themselves and their friends"

"Totally agree. Outside of taxable income calculations, it doesn't matter what people are doing with their time, be it creating open source software or farming virtual light leather, at the end of the day it’s all value."

People devote 20+ sitting on Yahoo Chat lines or hanging out in Match.com.

Are they then creating value as well?

Simply because you spend (waste) time at something does not necessarily mean you are creating value.


The question is not if any value is being created, though, but rather *how much* value is being creating.

I think the rather low skills and education requirements for playing games like EQ2 or WoW provide us with a useful heuristic which should encourage us to look at the low end rather then the high end for the true total # (GDP or whatever) of how much value is being created.

Posted May 8, 2005 5:06:37 PM | link

Zeke says:

Blaze: Hey, I just got here! This thread can't be "dead and buried yet!" :))

Anyway, in spite of my previous post, I would beg to differ about the skills and education of gamers on EQ2 and WoW. I'm guessing that, compared to the rest of the populace, we're talking about a relatively high-productivity group of folks. I'm guessing they contribute more than average to their share of total income in RL, and would do so in-game if given the tools.

Furthermore, economically, you are in fact creating at least some GDP simply by subscribing to WoW. OK, the game maker "created it" but it is your $15/month or whatever that is being counted, as GDP is simply the sum total cost of all finished goods and services (most sophisticated estimates also factor in for black market transactions, etc.). There is no need for these sophisticated methods here, though. Just look at what people are willing to spend on the game in real dollars and you have a measure of its real GDP contribution. And no one would pay a sub fee if not for the quality online community, which is clearly something of value.

But basically, I agree with the sentiment that the gaming enthusiasts over-estimate the current value of online virtual income in real terms. The reason is the focus on the trading of game money for artificially scarce items. Such measurements would be valid only for items that are genuinely scarce and which require player effort to invent (the boutique scripted pets of SL, or player-designed areas on the old text MUDs).

Posted May 8, 2005 9:01:26 PM | link

cartes d amour says:

amour
amoureu
amoureus
amusante
amusantes
aniversaire
anniversaires
anniversaires de mariage
annuaire cartes virtuelles
belles lettres
besoin d amour
carte postal
carte postale d amour
carte postales virtuelles
carte romantique
carte virtuelle
carte virtuelles
cartepostalevirtuelle
cartes anniversaire virtuelles
cartes d amour
cartes d amour virtuelles
cartes d anniversaire virtuelles
cartes postale virtuelles
cartes postales amour
cartes postales de paris

Posted Jan 27, 2006 10:27:53 AM | link