The Economist has an interesting piece on the interaction between the economy in massively multiplayer games and that of the real world. The classic study of this question is Castronova’s analysis of the economy of Norrath, the setting for Everquest. Among various features of Norrath’s economy, one of the most interesting is trade with Earth through the sale of game items (weapons and so forth) via private treaty or on eBay[1]. This enables Castronova to estimate that the wage in Norrath is $US3.42 an hour, a figure that has some interesting implications.
At the Creative Commons conference last week, I heard a story to the effect that when the owners of one of these games tried to prohibit item trading they were sued and, in the course of litigation discovered that the plaintiff ran a sweatshop in Mexico where workers participated in the game solely to collect salable items. Clearly as long as the wage is below $3.42 there’s an arbitrage opportunity here. More technically sophisticated arbitrageurs have replaced human workers by scripted agents, working with multiple connections. Either way, arbitrage opportunities can’t last for ever, and are likely to be resolved either by intervention or inflation
The positive economics of all this are interesting enough. But how about policy analysis? Who benefits and who loses from this kind of trade, and do the benefits outweigh the costs?
It’s pretty easy to produce a model where no-one gains and all the ‘real’ inhabitants of Norrath (that is, people playing the game for pleasure) lose. Consider a model with two groups, Norrathians who play the game for pleasure, and Terran arbitrageurs, who are good at collecting salable items, but incur disutility from the work involved in doing so. As usual in economics, assume that the two groups are homogeneous. Suppose that each individual Norrathian enjoys the game more when they have more items than others, but that a uniform increase in the stock of items makes no-one better off. Consider the starting point before the arrival of the first Terrans. Each Norrathian is willing to pay a positive sum (in Terran dollars) for items, and we’ll suppose that this is initially higher than the cost to Terrans of collecting them. So, when the first Terrans arrive, trade begins. The Terrans are better off, but the Norrathians are worse off. Since they all expand their holding of items uniformly, they gain no more pleasure from the game than before. But they are now paying the Terrans for items. The problem here is that buying an item outside the game creates a negative externality for all players.
Now, since there are profit opportunities for Terrans, more and more keep arriving. New entrants keep entering the market until the Norrath-Terra exchange rate drives the return for Terrans down to the ordinary Terran wage, at which point the entire system is in equilibrium. Terrans are no better off, and Norrathians are strictly worse off. From a social point of view, encompassing both Terra and Norrath, the work done by the Terrans in acquiring items is entirely wasted. Hence, a welfare improvement can be realised by prohibiting trade.
I think this is basically the correct story, at least in terms of mainstream neoclassical economics. But there are some counterpoints to consider. As The Economist points out, one reason people pay to get items is that the early stages of the game (the ones you play when you don’t have many items) are less interesting than the later ones. This is a design flaw and extra-game trade is a warning about this flaw. I think a Hayekian could make more of this point than I have.
A second, more neoclassical, point is that, if Norrathians are heterogeneous, there may be real gains from trade. Time-poor cash-rich players may prefer to sweep past the obstacles, using money to smooth their path, while others may be happy to defray some of their costs by selling surplus items. I’m not sure if this works (why can’t the game owner capture all these rents and lower the average entry price), but it seems like an interesting argument.
I think there are deeper issues here, relating to the conflict between the kind of market rationality displayed by the Terrans in my model and the collaborative innovation needed to create virtual worlds, but that’s a story for another post.
fn1. I haven’t looked into the actual mechanics of this. But as long as trade is allowed on Norrath, it’s hard to prevent side payments being made on Earth.
{ 34 comments }
novalis 01.28.05 at 6:26 am
A SF story about the sociopolitical implications of this sort of trading
Sebastian Holsclaw 01.28.05 at 7:15 am
I haven’t played this particular game, but of the games I’ve played this wouldn’t be a good description. Much of the world is not so interactive between players as to make different levels of items that damaging if you don’t have them. (The one exception would be the player v. player areas).
Motoko 01.28.05 at 7:40 am
(Perhaps you mean crossposted at johnquiggin.com?)
Walt Pohl 01.28.05 at 8:25 am
The fact that there is a sweatshop in Mexico where people play video games for money may be the most disturbing piece of information I have ever heard in my entire life.
Sebastian Holsclaw 01.28.05 at 8:43 am
If I got three dollars per hour for playing video games….well I wouldn’t be rich, but I’d be getting paid for doing something I like.
Idiot/Savant 01.28.05 at 9:01 am
I’m just waiting for an MMORPG owner to start supplementing its revenue by explicitly creating in-game items for sale for real money. They’d probably want to do it secretly (to avoid “ruining the game”), but there’s definitely an opportunity for them to profit there.
Tom Scudder 01.28.05 at 10:53 am
There already are several games out there that operate off of a “pay one US dollar for N game dollars, which you then use to buy in-game goodies like weapons or whatever” model. One of them, Project Entropia, made the news lately when the management sold off a virtual island for $26,500. Not REALLY really $26,500, since it was $26,500 worth of in-game currency, and the purchaser probably would have had some difficulty directly withdrawing all of that money all at once.
Here’s the Terra Nova discussion and the slashdot discussion.
yabonn 01.28.05 at 2:03 pm
Yay, mmog economy post!
Two hypothesis :
Suppose that each individual Norrathian enjoys the game more when they have more items than others,
vs
Time-poor cash-rich players may prefer to sweep past the obstacles, using money to smooth their path
I think the second one explains better.
The desire to show off is a strong drive in these games, but mainly for the more hardcore players. People are usually more worried about skills, tricks, and being decent at what they do – play your part in a typical hunt group.
So the uber sword of ass kickiness is good, but usually only as good as its stats. Better even if the buddy has it too, making hunting (usual occupation of these games) easier.
The second hypothesis works fine with that. I see the ebaying (may be just me) as a outgrowth of the way the u.s. economy is tilted towards local services. Paying someone to do something i find annoying : cut the grass, paint the house, get that sword. Quality time. My time is worth more that that, etc.
More generally, i think “real life” trade is bound to pop up each time a virtual world implements time consuming items.
I didn’t like the castronova article at his release, iirc, becaus it took the norrath economy seriously.
C’mon. Money and equipment pours in continually, from the “loot” – your reward for huntings things.
Problem is that there’s no destruction of money. To avoid inflation, there should be money “sieves” : paying your money to a character controlled by the server you play on, effectively making it dissapear.
In eq, there are no real of these sieves : food is cheap and equipment is usually looted. Equipment itself doesn’t decay : another possible sieve removed.
So money pours in endlessly, and the only way to avoid general inflation is to do what the authors do in eq : add new items all the time. Prices are nuts for the latest version, and lower levels can still buy some second best at decent prices.
As the player base doesn’t expand that fast, as the money stays in, as the best gear is replaced by the next generation best gear, the prices tend to go nuttier over time for the high end stuff.
A good way to track low level inflation would be to check what the price of the bone stack is. Bone stacks are reagents for a spell : one of the rare player-produced decaying things.
About two years ago, they sold for about 1pp. Anyone has the current price?
Heard they added some “show off” type sieves too for high levels.
Conchis 01.28.05 at 2:13 pm
“Now, since there are profit opportunities for Terrans, more and more keep arriving. New entrants keep entering the market until the Norrath-Terra exchange rate drives the return for Terrans down to the ordinary Terran wage, at which point the entire system is in equilibrium. Terrans are no better off…”
Um, isn’t Terran employment higher?
Alex R 01.28.05 at 2:24 pm
Walt — Disturbed?
asg 01.28.05 at 3:25 pm
Others have pointed out the flawed assumption in John’s story (which was very enjoyable):
Suppose that each individual Norrathian enjoys the game more when they have more items than others, but that a uniform increase in the stock of items makes no-one better off.
I think the enjoyment of the game comes from speed of advancement, which is not zero-sum. So a uniform increase in the stock of items does indeed make people better off.
However, despite this, John is correct to say that extra-game item selling creates a negative externality. This is because “farming” — the practice of killing monsters in areas lower level than you are solely to obtain gear which can be sold, rather than to advance through the game — reduces opportunities for advancement for the people who are playing for pleasure. If Jack needs to kill 25 orcs to advance a level, but Jill is hanging out in the orc area killing them all for their gear, Jack’s advancement is slowed down, regardless of the quality of his own gear (since having good gear doesn’t matter if the enemies you have to kill to advance are not available, having already been killed).
Incidentally, EQ’s major competitor, World of Warcraft, has problems along these lines too (“farming” is becoming visible on several of the servers). But WOW does have the important advantage that the low level areas are just as interesting as the high level ones.
Peter 01.28.05 at 3:47 pm
yabonn, the price for bone chips is pretty much the same, but the number of lower level chars harvesting them for sale to higher players has dropped drastically. Higher level necromancers (the ones who use bone chips like peanuts) also have an ability to turn any kind of meat into bone chips (and can do them in stacks of up to 20 at a time).
There are also some hefty “plat sinks” in terms of tradeskills. But not everybody does them. Some of the folks I play with have millions of plat (I however, being a tradeskiller have close to none).
Walt, there are a number of east asian and east european companies making a living selling plat. Some of them would have people (or maybe bots) at certain high cash/low risk places 24/7. Company rakes in $3/hour/employee, pays employees 1/6 that, makes a reasonable profit.
Sometimes there are pricing screwups in the game that yield cash income, those get removed as quickly as they come to the attention of SOE.
The introduction of worlds of warcraft, halo 2, and eq 2 has reduced the number of players in eq drastically.
On-line games really don’t want to have a cash value for things, and I think project entropia is going to get burned badly because of it. If “Item X” has a real-world value, what taxes do I owe on it? By looting it, have I received taxable income? Who is responsible for collecting/paying those taxes? If the item disappears due to a bug in the game, where is my $1000 doohickey? If I get banned from the game, where is my $1000 that my doohickey is worth? You introduced a new expansion, my $1000 doohickey is now worth $100. I want my $900. My evil ex-(wife, husband, child, parent, cat) logged into my account and sold my $1000 doohickey, I want it back, or my $1000. You came out with a new expansion, my $1000 doohickey is now worth $100, but my state charges personal property taxes on the value of the $1000, not on the new value. My $1000 doohickey was unbalancing the game, so you nerfed it, now it is worth $100, I want my missing $900.
If you want to see sad:
http://terranova.blogs.com/terra_nova/2005/01/buy_the_farm.html
I found it fascinating that folks would spend time writing AI-scripts (one meaning of “bot”, another is a second computer tha you use to play a 2nd thru 14th additional character at a time) to automate their characters. Some would do it to automate earning sellable cash, some to do it to automate unpleasant tasks (ok, I can see that), and in the “Buy the Farm” post on terranova, some wrote bots to hunt other bots. Bots to find and hunt other bots? That is almost like the movie Tron.
Edward Castronova 01.28.05 at 3:58 pm
Castronov_A_!
yabonn 01.28.05 at 4:23 pm
Hey i got the name right! Online economy still sounds like tulip bulbs to me too :)
Thanks for info, peter.
Asg is right about farming, i’d add that another effect of it is to create monopolies on certain items : if guild Uberer_than_thou farms all the mobs giving gizmos, it has an effective monopoly on it.
Wow seemingly has been thinking about all this. I still haven’t played it but they seem to have made the items either decaying or nodrop (nodrop meaning you can’t exchange them anymore once equiped).
yabonn 01.28.05 at 4:26 pm
Wow = World of Warcraft. Sorry.
asg 01.28.05 at 4:32 pm
A little bit. In WOW items gained as quest rewards are “soulbound” meaning that as soon as you get it it can only be equipped by you (so it is useless to anyone else except as essence bait — and maybe not even then; I know the auction house won’t accept soulbound items at all). Items gained from random drops are not soulbound, though, and hence can be farmed.
Marc 01.28.05 at 5:10 pm
There is one signnificant negative consequence for games that isn’t mentioned. Usually someone who has achieved high status has done so by plowing through a certain amount of content – and learned how to actually play the game. Most online games require co-operative effort, and the content is designed to require skill and planning at the high end. People not only sell items – they sell characters as well. In this case you can skip the learning curve, and powerful items will permit you to trivially blow through things that people had to actually figure out without them.
The net effect is that you can have people who are running powerful characters and have absolutely no clue about how to play the game. This has a real negative impact on the community. There is also a more subtle cultural phenomenon.
“High status” in game is set by what your character has achieved; if people can simply “buy” the equivalent of high status with real world cash then it removes the one virtual indicator of meaningful success. This is the fundamental objection that most players have of real-world sales: it removes the only meaningful way to “win” the games.
Marc
perianwyr 01.28.05 at 6:37 pm
The game Ultima Online recently introduced the ability for people to take characters from one server to another for $20 per transit. This caused some really amusing things to happen.
if guild Uberer_than_thou farms all the mobs giving gizmos, it has an effective monopoly on it.
Yes. There are items called powerscrolls which allow you to take a skill above its standard limit of 100 in intervals of 5, all the way up to 120 maximum. Each level is rarer than the last, with 120s of very popular skills being worth millions of in game gold (archery and magery being some of the most expensive, I’ve seen them for 4 million before.) There are even more powerful scrolls (stat scrolls) too.
A quick glance at the Atlantic shard’s listings on EBay (http://tinyurl.com/4747f) will tell you that a million gold is worth about $8 or so.
The powerscrolls are obtained by fighting extremely hard monster spawns (“champ spawns”) in regions designed specifically so that other players can kill you there. This leads to the uberguild effect- gigantic groups of players will set up surveillance on spawns and either do the ones they prefer (there are easier and harder ones, and you can get good scrolls from either) or send in their thugs to kill any hapless people who have decided to challenge the spawn (I’m one of those thugs. Isn’t life grand?)
These uberguilds have, in some cases, made the champion spawns unprofitable and pointless for people who are not members of their own uberguild (why start a spawn and take the time to advance it if a shitload of murderers are going to pile in and crush you?)
So, some enterprising players have used this transfer service to move advanced characters to other, less populated servers that don’t have established uberguilds so that they can do scroll spawns literally all day, and summon the stat-scroll bearing “Harrower” repeatedly. They then take the scrolls back to their home shard via one $20 transfer, and either use them or sell them once they can’t use any more.
This effect has, in turn led to giant roving uberguilds that just transfer lots of members at a time to dominate shards and war with other transferring uberguilds, who come and go as the fortunes of the shard change.
This has led to lots of extra revenue for UO’s publisher, and a strange kind of political subgame where alliances are sought and discarded based on whether the guild feels like paying to get the fuck out of Dodge.
Fun stuff, eh?
Andrew McManama-Smith 01.28.05 at 6:53 pm
I play that world of warcraft loads, and I have to say, that people can just make their own items. It takes me ten minutes to make an item that sells for 100 gold, and that gold I can sell on ebay for about $5.
I think you’d have to be pretty lazy to buy gold on ebay, since it’s so easy to make it yourself.
The other thing is that items are level-capped, ie low-level characters cannot use high-level items. and low-level items are really cheap, so you don’t get much out of buying gold until you get high-level, at which time it’s really easy to make items to sell for a lot of money…
So who would buy gold or items on ebay? on the laziest person imaginable (too lazy to play video games!!!) would do such a thing.
Andrew McManama-Smith 01.28.05 at 7:08 pm
Despite the comment above, I actually am literate and usually I write with good grammar and spelling.
paul 01.28.05 at 7:19 pm
One bit of the analysis I don’t understand. You (JQ) wrote:
Each Norrathian is willing to pay a positive sum (in Terran dollars) for items, and we’ll suppose that this is initially lower than the cost to Terrans of collecting them. So, when the first Terrans arrive, trade begins. The Terrans are better off, …
Why would Terrans come in and trade? How are they better off if the Norathians pay them less than it costs them…? I would think that this is a simple typo, except I don’t notice that anyone else has commented on it yet. What am I missing?
John Quiggin 01.28.05 at 7:40 pm
Paul, thanks for picking this up. I’m always making mistakes like this. Fixed now, I hope.
greglas 01.28.05 at 7:40 pm
As previous comments have noted, we’ve chatted this stuff up incessantly at Terra Nova. My quick comment would be that if this Jeremy Bethamesque approch to the legal regulation of markets raises eyebrows in relation to property rights on Terra Prime, it is going to be even more unstable and subject to critique when applied to Norrath, i.e. social activities involving games and play. But staying in the economic vein: you might want to consider is what do you do if your utilitarian calculus comes out in favor of external markets, but the property holder of the server comes out in favor of banning the market? Does the inefficiency of that decision provide a basis for utility trumping our normal presumptions of ownership?
greglas 01.28.05 at 7:42 pm
As previous comments have noted, we’ve chatted this stuff up incessantly at Terra Nova. My quick comment would be that if this Jeremy Bethamesque approch to the legal regulation of markets raises eyebrows in relation to property rights on Terra Prime, it is going to be even more unstable and subject to critique when applied to Norrath, i.e. social activities involving games and play. But staying in the economic vein: you might want to consider is what do you do if your utilitarian calculus comes out in favor of external markets, but the property holder of the server comes out in favor of banning the market? Does the inefficiency of that decision provide a basis for utility trumping our normal presumptions of ownership?
John Quiggin 01.28.05 at 7:47 pm
Sebastian and asg, if more items and easier progress make the game better for everyone then there’s a design flaw in the game as it stands.
Yabonn, to quote the immortal Pigou “assume full employment”. You can’t easiy do this kind of welfare analysis in a Keynesian framework where unproductive work can generate social benefits.
A preferred Keynesian response is for the Terran government to run productive public works program.
Thanks by the way, to all who’ve contributed to this thread. I’m learning a lot.
John Quiggin 01.28.05 at 8:02 pm
And yet another typo fixed. Sorry for misspelling your name, Edward.
asg 01.28.05 at 10:50 pm
John — Sorry, I wasn’t being clear. I didn’t mean “speed of advancement” as a direct “faster advancement = more fun” value. Rather, it’s a question of getting the advancement curve right. In WOW, getting to 10th level takes maybe three or four hours, but getting to 20th may take another 20 hours. That rate of advancement is ideally optimal for the majority of players; advancing too fast isn’t any fun because once you get some neat new powers, you want to spend some time actually using them before moving on to the advanced versions of those powers, etc. Advancing too slow is obviously no fun because that gets boring.
This line of thinking also leaves out the social elements; many people find it fun to help lower level characters out (many also find it fun to ruin the game for other people, but those are Nozick’s utility-monsters). So having too few low level characters harms the social experience a little too.
Dan Hunter 01.29.05 at 3:24 am
I didn’t get a chance to comment earlier today, and things have moved on. So I’ll just make a couple of observations based on John’s last comment:
>Sebastian and asg, if more items and easier
>progress make the game better for
>everyone then there’s a design flaw in the
>game as it stands.
If you talk with the game devs and the power gamers and even some of the casual gamers, you’ll find (I think it’s fair to say) that they won’t agree that the actions of the eBayers are neutral. Nor will they agree that the mechanisms in place are good for everyone. The power gamers in particular have a peculiarly hypocritical view that those who purchase virtual assets are the lowest scum of the virtual universe, unless of course they are purchasing assets from the power gamer. The game devs know which side their bread is buttered and so spend a lot of PR effort cozying up to power gamers and demonstrating how active they are at policing the worlds. They fairly often use the kind of argument presented above by John as justification, ie that there are negative externalities in the world from the transactional activity outside the world. But mostly that’s just a smokescreen.
Which leads me to my other thought, prompted by John’s observation:
>A preferred Keynesian response is for the
>Terran government to run productive public
>works program.
It is, I think, a mistake to think of these worlds as worlds. That is, the immediate response of economists and lawyers and policy makers is to assume that real world social phenomena are going to map cleanly to these virtual worlds, and that the usual economics/law/whatever can be applied here. This sometimes works and sometimes doesn’t. Perhaps the best example is the work that Ted Castronova has done about the nature of inflation in-world (usually called “mudflation” for reasons too arcane to worry about here), where he concludes that the real reason that mudflation exists is because it’s fun.
Now this is not exactly the standard account of inflation, but what is perhaps most interesting is that virtual world observations and theories derived from them may actually provide useful insight into real world phenomena. But for the moment, we have to be pretty careful about the degree to which we can compare real world apples with virtual oranges (and vice versa)
doug 01.29.05 at 7:26 am
actually not a totally accurate summation as players are not necessarily competing against each other but against the game iitself.
so if a player buys or gets a +10 sword rather than use a +1 sword he will gain levels faster and reach top level faster. which is one of the main game aims.
on the player vs player servers it is likely to be more competitive tho.
Tracy 01.29.05 at 1:04 pm
Assuming the supply of Terrans is not infinite, Terrans would wind up better off as a whole, including Terrans who never play the game at all. As Terrans are drawn to the virtual world, all other things being equal, wages would have to rise slightly for Terran occupations in the real world. (Supply of labourers is falling, while demand isn’t).
Or, if we step outside strictly classical economics and include some involuntary unemployment in our model, employment might rise and unemployment fall, or both wages and employment might rise.
In this world, I wouldn’t expect any increase in wages due to gaming to be actually measurable, since millions of other things are happening at the same time. But your model only produces no gains for the Terrans by ignoring the question of what the Terrans would otherwise have been doing.
Whether the disutility to the Norrathians from the negative externalities outweighs the gains to the Terrans is not answerable from theory and is probably impossible to answer confidently in practice. However, given the increase in income to the Terrans, we cannot assume that prohibiting trade would be a Pareto-enhancing improvement.
Tracy 01.29.05 at 1:13 pm
Assuming that the supply of Terrans is not infinite, the trade will make Terrans better off, including Terrans that never work in the virtual economy. This is because as people are attracted to work in the virtual economy, there are less people available to work in the ‘real’ economy, without a corresponding drop in demand, so wages will rise.
Or, if we step away from strict classical economics and allow some involuntary unemployment amongst Terrans, instead of wages rising employment might drop, or both might rise to a lesser extent.
Your model only produces that prediction because it ignores the question of where the Terrans come from and what they would be doing if the virtual economy did not exist.
Therefore we cannot assume that a welfare improvement can be realised by prohibiting trade without some empirical work on the magnitude of the negative externalities and the rise in wages amongst Terrans.
In our world, to step away from hypothetical models, I don’t expect there to be any measurable impact on wages from the increased job opportunities in virtual economies, due to the massive number of things that are also changing at the same time and the error rates in what statistics are collected.
Tracy 01.29.05 at 1:18 pm
Sorry for double post. Got an error message on the first & thought I had to rewrite.
Tracy 01.29.05 at 5:32 pm
And I have just realised that in my responses above, I forgot to cover where the money the Norrathians spend on virtual assets comes from, since spending a dollar on the game means that you are not spending a dollar in the ‘real’ economy – although presumably the Terrans eventually will.
If the Terrans and the Norrathians differ only by whether they want to play in virtual worlds then there’s no net impact on Terran employment or wages (as the Terrans would have otherwise earned the same dollar another way). If there are barriers to other sorts of trade between the two groups, e.g. tarriffs or just living in different countries, so in the absence of the game less trade would have taken place due to economic inefficiences, then employment or wages would rise amongst the Terrans. I was thinking too much of Mexicans and Americans and forgot to generalise the case.
So if we assume negative externalities to increasing the supply of goods in the games, the case for stopping trade in virtual goods to benefit the public interest is stronger if both sides are in the same country than if they are in different countries.
And the case for banning is weaker again if you assume different utilities and disutilities for different players. There is nothing in the real world to stop someone from both enjoying playing the game and being happy to sell some goods, or to, while they would not pay to play the game, prefer playing the game for money to working in a shop for the same amount of money. If the Terrans sort themselves out into ones who least dislike playing the games and ones who least dislike working in the shop, there are definite gains from trade.
Which is why apparently Truman said he wanted only one-handed economists, because whenever he asked an economic advisor a question he got a response saying “on the one hand … on the other hand …”
Jeremy Neal Kelly 01.29.05 at 10:54 pm
> The positive economics of all this are interesting
> enough. But how about policy analysis? Who
> benefits and who loses from this kind of trade,
> and do the benefits outweigh the costs?
Unfortunately I’m traveling this weekend, so I don’t have much time to participate in this discussion. However, I recently finished a paper that addresses this very question. It can be viewed here:
http://www.anthemion.org/Words/words.html
> A second, more neoclassical, point is that, if
> Norrathians are heterogeneous, there may be
> real gains from trade. Time-poor cash-rich
> players may prefer to sweep past the obstacles,
> using money to smooth their path, while others
> may be happy to defray some of their costs by
> selling surplus items. I’m not sure if this works
> (why can’t the game owner capture all these
> rents and lower the average entry price), but it
> seems like an interesting argument.
The paper broadly follows these lines. It also offers a brief analysis of the owners’ ability to capture this trade surplus through the sale of game goods (I call this trade ‘fiat sale’). You might find it interesting.
Comments on this entry are closed.