Insert Coins to Continue
Can a video game company, a Greek economist, and the largest barter market in history stop the recession—and save capitalism?
Tim Fernholz is only a man.Liz Meyer is illustrator who is attracted to overly colorful things.
Imagine a game. Millions of people play every day, and at the end of the year, they report their scores. Usually the big players win and the little players lose, but generally everyone ends up better off than they were before. But one day, it turns out everyone was losing—something’s wrong with the rules of the game.
The game is called Europe, and understanding the rules is Yanis Varoufakis’ job: He is an economist at the University of Athens.
“I have to tell you that it’s very bleak,” Varoufakis says of his native Greece, his accented English bolstered by a plummy British undercurrent. “It’s an implosion, it’s free-fall, it's an economy that has entered a coma.”
The Euro crisis is a complicated tale of greed, optimism, and mistakes made, in Athens and around the world. Before the crisis, Greek citizens may have thought they were playing by Greek rules, but they were actually playing a game within a game. They must also play by the rules of the European Zone, with whom they recently began sharing a currency, and a global recession affects them, too. Most of these rules are not written down.
Varoufakis was an adviser to a Greek politician who became prime minister in 2009. The new government discovered that banks around Europe and the world had lent the country far more money than it could ever hope to repay. Similar problems turned up in Spain, Portugal, Italy, and Ireland: Richer nations began putting the screws on the strapped ones, demanding they adopt programs of huge cuts to wages, budgets, and pensions in exchange for enough time to pay back what they owe. Soon, the whole continent faced a recession, and the rules of the game were up for debate.
Varoufakis rejected the program his old boss signed Greece up for, and joined a group of dissident economists in a “modest proposal” to save Greece and the Euro—essentially, a new set of rules for the game. Their solution is designed to share both pain and gain within existing institutions, and it has largely been ignored amid the squabbling of Europe’s political leaders.
But due to Greece's unusual prominence in the Euro debate—“It’s like Delaware bringing the United States down,” he says—Varoufakis’ blog about the crisis attracted the attention of an American video game company, Valve Software. Gabe Newell, the company’s president, emailed Varoufakis and asked him to stop by the company’s headquarters in Bellevue, Washington.
“Here at my company we were discussing an issue of linking economies in two virtual environments (creating a shared currency) and wrestling with some of the thornier problems of balance of payments,” Newell wrote, “when it occurred to me, ‘This is Germany and Greece.’ ”
Valve had accidentally invited its players—tens of millions of real people—into a kind of virtual European Union, and now its staff worried that the company had set the stage for European-style consequences for itself and its customer-citizens. If the real Europe didn't want Varoufakis’ advice, perhaps he'd take on the challenges of a fake one?
The last four years have been a trying time for economists. In 2009, motivated by the global economic crisis, people were interested anew in economics as a way of talking about our problems. But none of the proposed solutions have succeeded fully in practice, often because of a lack of political will. Meanwhile, the profession appears to lack a vigorous response to the questions raised by tumultuous recent events. Varoufakis calls the lack of progress “the great disappointment of my working life.”
Part of the problem is that an economy is an act of faith. It isn’t something you can touch or see or feel, although we often talk about it that way. Economics is the study of a social phenomenon—the way people make things, trade them, and use them for maximum benefit. Because economic activity is so important, we measure our economic exchanges more carefully than most things, collecting data about prices, wages, and employment. Economists can use math to describe the human behavior behind this data as though it were physics, their discipline’s advantage above all other social sciences.
But economics is not a natural science, and, like social scientists, economists base their ideas on assumptions. They cannot perform economy-wide experiments again and again to test their theories: You can't deploy a stimulus program in one place and austerity in another and do nothing at all in a third because people would react to the experiment and thus ruin it. There are natural experiments in economics—similar nations that try different approaches—but those are few and far between.
So economists do their work by building and studying models, as though they were learning how to fix a carburetor by playing with Matchbox cars. The most famous model is supply and demand—the first thing taught in Econ 101 is how prices are affected by the supply of a thing and the desire for it. You can even make a nifty chart. By combining simple ideas, economists construct models for how the economy should behave.
Computers gave economists powerful new ways to manage data and create more complex models—algorithms tell us how long Social Security will last, or how many jobs a government program will create. But no matter how good the model, it can’t begin to encompass the complexity of the real economy. For all the data, no one really knows exactly how many cars were sold last year or exactly how housing prices changed. And it’s difficult to simulate how people will behave, especially lots of them. Humans have a pesky tendency to make decisions based on more than just price levels, and when economists come out with new ideas, people learn them and behave differently, throwing theories out of whack.
That's why there's such a broad range of error in economic forecasts.
If the real Europe didn't want his advice, perhaps a fake one would.
Small economic experiments do work; one useful trend is getting real people to participate in studies that illuminate the behavioral psychology underlying economic decision-making. A few dozen or few hundred people can participate in these exercises for a limited period of time, but there’s no way to convince millions of people to spend hours making economic choices and consent to have their actions analyzed by academics.
Like many economic assumptions, this turns out to be wrong.
I come running out of the tunnel, lugging a chain gun and struggling to keep up with my red-uniformed teammates—a medic, a man with a rocket launcher, and one in a hard hat setting up a robotic machine gun—as we converge on the bridge. There is a control point on our side, and our enemies are on their way to take it. I crouch by a concrete wall and open fire just as four blue-clad enemy soldiers come careening around the corner. Someone is screaming for the medic. I’m dead within seconds: A cowboy-hat-wearing, pipe-smoking man with a blue mustache has killed me with an arrow. His name is “Archimedes!?!” and the game informs me that he is my new nemesis. But where did he get that cool hat?
“Initially, Team Fortress II was a really good first-person shooter,” Kyle Davis tells me. “Now there are people that play the game, and play the game every day, that don’t actually play that game. For them, the fun part of the game is finding good trade deals and running services so other people can use them as a resource.”
Davis is head of psychology at Valve Software, although that title isn’t very informative. A better description of Davis’ work is that he is a software developer who focuses on managing the problems that arise when millions of people spontaneously form an economy. Davis came to Valve thinking he would be programming new features for games, but soon found himself acting as an electronic central banker.
Like many modern video games, Team Fortress II allows people to play together over the internet. At any given moment, some 60,000 users are playing a heavily armed version of king of the hill. Amid the carnage, players can find deadlier weapons, tougher armor, and vanity items that change their avatars’ appearances, or pay real money for them at in-game stores.
In 2010, Valve decided to conduct an experiment: Why not create a marketplace for exchanges between players? People started playing just to trade. Some did it for the money, but others just like the thrill of a good deal. The most expensive single item was sold for $1,400. One player traded an in-game hat for a real-world car.
But Team Fortress II isn't the company’s only game that supports an independent economy. The game runs on a platform called Steam, which sells and maintains many games made by Valve and other developers—now responsible for 70 percent of the games bought and downloaded on the web—and a social network with some 4 million users. Why not allow users who play more than one game to trade with each other over Steam? In 2011, Valve made it possible for a player to exchange a rocket launcher in Team Fortress II for an ogre club in Defense of the Ancients—if both players think that’s a good deal.
Much like real-world economic managers—central bankers and finance ministers and the U.S. Congress, god help us all—the folks at Valve sought a stable economy that protected players from seeing their wealth affected by economic crises like those in the real world. Inflation, as new items and currency devalued existing supplies, was a serious problem in early online games known as multi-user dungeons as far back as 1993. In modern games like World of Warcraft, hyperinflation is a real problem, leading developers to create fancy, non-essential items called “gold sucks” to take money out of the economy. Otherwise, a money supply that expands too fast can make it hard for new players to get started (if they can’t afford a healing potion, say) while also devaluing the accomplishments of veterans.
For Valve, this this is not just a question of maintaining users’ suspension of disbelief: Team Fortress II is free to play, generating revenue through cuts of users’ financial transactions, so the company’s economic fortunes are intertwined with those of the players, and it wants to expand its marketplace without encountering problems like hyperinflation. So far, Valve has avoided many problems because its games’ economies are relatively unsophisticated, but Varoufakis can foresee new features—a currency, for example—that would transform one of the largest barter economies in human history into something far more potent, with a greater potential for unexpected surprises.
The possibilities for learning in this economic sandbox seemed limitless.
Valve’s corporate culture is interest-driven: There are no bosses, and projects are completed when enough people decide they’re important enough to get done. People found the virtual economies they were building important enough to prioritize, but programmers teaching themselves economic theory was, Davis says, “not the most efficient method.” That's where Varoufakis came in.
He isn’t typical of his profession. While he has the mathematical bent that characterizes most modern economic thinkers, his worldview was shaped by his youth in the tumultuous political environment of Greece during a right-wing military junta’s reign from 1967 to 1974. He is contrarian and occasionally obscure: His first book failed because it was too mathematical for philosophers and too philosophical for mathematicians. “I managed to fall spectacularly between two stools,” he says. He has collaborated with his partner, artist Danae Stratou, on a project about the “political economy and aesthetic” of seven areas of division around the globe, including Palestine, Kashmir, and the U.S.-Mexico border.
Still, video games? Varoufakis balked initially, but he realized the enormous opportunity he was being offered: not just the ability to run experiments on a huge trading platform, but access to a massive quantity of data in the form of every item’s price and each transaction’s details, in real time. The possibilities for learning in this economic sandbox, and for trying new ideas, seemed limitless. “The future lies in simulation and experimentation. And if you can combine the two, you have a very powerful tool in your hands,” Varoufakis says. “That gives Steam an opportunity to consider different regimes for different kinds of exchange systems.”
Perhaps the most tantalizing possibility is the chance to reconsider basic economic assumptions. When models reduce everything to prices, they fail to take into account real-world factors. Is your decision about how much to pay for orange juice based merely on supply and demand, or also on taste, brand preference, and availability? Economists call everything not taken into account by their models “externalities,” and one of their biggest challenges is incorporating those externalities into analyses. A virtual economy, Varoufakis says, is a “realm of externalities … The experience of playing these games is an interactive one, so the utility that you get from being part of the community, is by nature, by design, a collectivity. You cannot separate yourself and what you ‘consume’ from the actions of others; they influence not just through some price.”
The converse of this observation is what video-game economists call “the problem of fun”: The behavior of people playing a game recreationally might not be an accurate representation of serious decision-making. Edward Castronova, one of the first economists to study virtual economies in video games, noted this paradox in an early paper: Gamers are essentially paying to make their lives more difficult, he wrote—after all, a video game in which accomplishment was easy would be boring and never sell.
Fun might be a problem, but it’s not a big one. After all, economists generally believe people maximize their utility—their desires—which can result in some bizarre economic decision-making. Just before the 2010 holiday season, the Team Fortress II stores started selling “Secret Saxtons,” which players could use to deliver a random reward to another player in the game. These gifts were hugely popular, and the next year Valve began selling badges that displayed how many gifts players had given out in the previous year. One player had paid about $10,000 to give out 12,000 gifts. This is not an easy phenomenon to model if you’re looking only at prices, but neither are the gifts your mother gives you at Christmas or the used car a kind uncle sells you for a buck. And while demographic representation is a consideration—limited data on players of popular games shows they are largely male and in their 30s, though composition is changing all the time—people playing Team Fortress II probably behave more realistically than the ones who live in economists’ models.
“What constitutes a real economy?” Varoufakis asks. “In today’s financialized world ... the purchase of a digital item, on a basis of barter, is far more real than what our Wall Street colleagues are doing.”
In a typical experiment, Valve can set four different prices for the same item in different markets simultaneously, then track how that affects players’ trades. They must act fast: Players usually figure out the experiment within 30 minutes and try to game it—a fairly good simulation of how real financial markets behave. The company is transparent about the research, both out of a general desire to treat its users right and because its success comes from customers: Many of Valve’s hit games, including Team Fortress II, are based on modifications to old games that users hacked up just for fun. So the company emails users about its experiments, and Varoufakis blogs about his work on the Valve site.
“Any time something unusual happens in one of the games, people blame the economist and assume they're being experimented on,” Davis says. “Sometimes they're right.”
After Varoufakis posted about what he had learned from one of these experiments, a number of users mentioned they were not always using the barter economy to make deals. Sometimes they were just ditching items they didn’t need, or giving a friend a discount to be nice—the problem of fun all over again. These users argued that their behavior doesn't meet the definition of an economic exchange described by Adam Smith, the founding father of economics. That’s because Smith was wrong, Varoufakis replied. One more assumption down the tubes!
Smith believed money arose in barter economies because people needed it to make transactions easier to price. But Varoufakis observed that no single item on Steam has become a numeraire, or de facto currency, despite the increasing complexity of the market. This jibes with recent anthropological research, which suggests the real basis of trade is “cooperative non-market interactions”—cultural norms and personal relationships that do more to guide economic development than purely interest-driven exchanges. Valve’s early identification of this phenomenon is a return on the potential for games to open up old models to all the interesting wrinkles of human behavior. Maybe fun is the new money.
Valve's unique organization prepares employees for these kinds of insights. There is no hierarchy, and no job titles: Kyle Davis’ self-described role as head of psychology was a joke at the expense of one of his colleagues, a behavioral psychologist.
Employees are hired carefully, but once brought on board, they spend their time on the tasks they find most important, wheeling their desks into self-formed project teams after they figure out who wants to do what. You might imagine the lack of structure resulting in chaos, but Valve is no joke: a 400-plus-person company worth several billion dollars, it claims more profit per employee than Google or Apple. Varoufakis says Valve has imported the invisible hand of market competition into the organization of a business to allocate its most scarce resource: time. It’s amazing that all the other capitalists never thought of it.
There are drawbacks to this approach—it probably wouldn’t work in a manufacturing firm, and who knows how far it can scale—but Valve’s organization and the insights the company gleans from the game world stem from the same phenomenon: People using new technologies and organizing themselves spontaneously into economic units, based on cooperative non-market interactions. Studying them is a chance to re-examine first principles and shed old assumptions. It might even represent the post-crisis breakthrough we’re looking for.
If Valve does decide to unite its games under a single currency, it could experience the same problems that Europe has today—an up-and-coming game swamped by the dislocation that comes from partnering with an established giant, a Greece and a Germany. But if Varoufakis can prevent that fate in the virtual world, it will be difficult for his critics to accuse him of merely wanting bailouts for his country in the real one. It will show that he may have been onto something with his proposal for improving the rules of the game.
“Economists have a tendency to look around for new realms where they can demonstrate that their prejudices reflect facts,” he says. “We can be more honest, as economists, when we are studying virtual economies.”