How Democracy 3 is simulating African politics

Somewhere in a backstage area of the London Science Museum there exists a strange, hulking and almost forgotten machine called MONIAC. For years it mouldered in a basement at the London School of Economics, but in its early Fifties heyday it was the most advanced economic simulation tool in Britain. Using a system of plastic tanks and tubes, it modelled the flow of money as a flow of water, controlled by a series of floats, valves, and counterweights. You could adjust the valve for tax rates, and watch how the flow changed, or tweak government spending and observe the result; a computer before computers were widespread

The MONIAC. Photo by Marcin Wichary.

The MONIAC. Photo by Marcin Wichary.

Today, the children of MONIAC exist in every central bank and every major financial institution, running the financial prediction models on which our economy increasingly depends. But I think there is an heir with equal claim in the realm of consumer entertainment technology. Its name is Democracy 3.

Democracy is a “political simulation” series created by indie developer Cliff Harris. The best, 2013’s Democracy 3, has been sold to 60 British schools. Under its jaunty edutainment surface lurks a dizzying spaghetti monster of interconnected statistics representing everything from GDP to racial tension via the rate of legal drug consumption. Policy adjustments cascade down through this network and return in unexpected ways: cutting labour laws might boost productivity but create a stress epidemic, while healthy eating plans can provoke a general strike by raising the price of food. 2,000 simulated voters judge your performance, and you can’t please all of them all the time.

This week, though, Democracy is expanding. Democracy 3: Africa is an “expandalone” remake adding ten new nations including Egypt, Nigeria, Kenya and Botswana, as well as the new mechanics necessary to model them. By the time you read this, it will probably be out via Steam and direct download. For dual-spec political/gaming nerds like me, this is exciting.

But how do you take a simulation designed for first world countries and adapt it to a continent with very different problems? What’s more, how do you task the player with ‘fixing’ African nations when the real world debate about what’s wrong with them is so bitterly divided? And how do two self-confessed white guys from the UK grapple with these issues of corruption and colonialism without reinforcing the misrepresentations which have dogged Africa for centuries? In short, how do you ethically simulate perhaps the most historically oppressed continent on the planet?

Enter the mega-matrix

“I didn’t know an enormous amount about Africa,” Cliff Harris told me when we meet at the Game Developers Conference. “Most people don’t either. They only know about it from TV reports of either terrorism or drought. It’s the setting nobody knows they want.”

Cliff has a complicated relationship with politics. He studied economics at LSE – MONIAC’s alma mater – and worked in IT for major banks before becoming a developer. He loves the serious debate between rival ideologies, but despises the multimedia circus which surrounds it. “Things are so bad, I honestly favour some sort of minimum awareness testing on the part of voters to qualify,” he once wrote on his blog. D3 is his ideal model of politics, with purely rational voters who don’t fall for spin. His business card bears a quotation from Franklin Roosevelt about education being democracy’s “real safeguard”, and his aim is to make at least a few voters a little wiser.

So why Africa? Announcing the game last November, he wrote: “When do you ever see Africa in a video game? I have no memories of it ever being anything but a destination where pirate bases or criminal gangs roam. Gaming seems to have a very distorted view of Africa, just like Hollywood does.” Since then, he says, little has changed. “If we were just talking about one little country, you could kind of think, ‘ah well, whatever’. But it’s an entire continent which has been ignored by gaming.” Beyond that, he considers it an interesting set of problems.

Actually implementing those problems fell to Jeff Sheen, a former derivatives trader at Morgan Stanley who developed the game with Cliff as his producer. “I knew that my own perception of African politics…was a function of the lens I was looking through,” Jeff told me in a phone interview. “That was the lens of western media, and I found it sporadic and often sensationalist. So I established from the very beginning of the project that I wanted to throw all of that out and start from a blank slate – assume that I knew nothing.”

To fix that, Jeff did what he’d done in his high finance days: drilled into the data. He drew together statistics and economic indicators from multiple sources (so as to “shatter the western lens”), and spoke to experts both inside and outside the continent: aid workers, NGO members, academics, think tank analysts, government officials. All this went into a “mega matrix” of linked spreadsheets which formed the basis, “as objectively as possible”, for each national model. Then came the real work – establishing, as Jeff puts it, “the threads of systemic commonality required to respectfully and accurately represent their economies, their societies.”

How to pick a side

The problem with a game like Democracy 3 is that it cannot help but make political statements. In the real world, questions about the relationship between different economic factors – such as whether high tax rates lead to lower tax revenues, or how much debt a government can take on – are hugely controversial, and yet D3 has to answer them in order for its model to work. Cliff insists he is a political magpie (“I’ve been called both a fascist stooge and a communist”), and D3 sort of bears that out: it feels left-wing on religion (which decreases with science subsidies) and government stimulus (Cliff admits there is a Keynesian bias), but conservative on budget deficits (ratings agencies downgrade you at the drop of a hat), and the importance of “wealth creators” (piss off the rich too much and watch your GDP suffer). But either way, the needle has to fall somewhere. That acquires additional freight when you’re Brits making a game about Africa.

Jeff is sanguine about this inevitable partisanship. “As a designer, you’re making calls on this sort of stuff all the time. If you wanted a data source that gave you a completely untainted relationship between, say, oil price and the opinions of the wealthy, it doesn’t exist. You just have to look at the trends in your research and make a call on it.” For him the answer is in faithfully synthesising diverse data sources, both western and African, to eliminate as much bias as possible, and find “atomic fundamentals” to which all of them testify.

The results include a new “infrastructure” stat which reflects the quality of your roads, railways, and power lines. As it improves your rural voters will start to migrate to the cities as – a transition you must manage carefully.  In a coup-stricken continent, “stability” determines your levels of foreign investment, while voters will always agitate for more “democracy”. Women are also a distinct group and gay rights a specific concern, reflecting what Cliff and Jeff say are higher levels of discrimination than in most of the first world. Then there’s corruption. As I try to reform the Egyptian government on Cliff’s GDC laptop, ShadowHand creator Jake Birkett interjects: “Did you hear that £14bn of oil revenues disappeared in Nigeria?” Cliff shakes his head with a mixture of despair and amusement. “It’s just madness, isn’t it?”

These new mechanics make their own statement: for better or worse, the game is saying that corruption and gender inequality are significant factors in Africa but not in western Europe. On the other hand, nothing has been removed from vanilla D3, and all its high-tech DLC packs are fully compatible with D3:A, meaning there’s no reason why, as Cliff puts it, “you can’t make Nigeria a superpower with a space program.” Other values are more specific: violent societies produce few liberals, development reduces the power of religion, and the game is “very unabashedly against” female genital mutilation, which has no positive effects except on conservative opinion. “That is to represent the only friction against eliminating it, which is that it’s a spiritual tradition inside the societies that practice it,” says Jeff.  Again, he says, he has the research to prove it.

Yet I wonder if there is a more subtle ideology at play beneath all this data. “Econometric” modelling a la MONIAC is now a crucial tool for banks and governments across the world. But these models have a murky legacy because their attractive but abstract theories are not always strongly connected to reality. They remind me of the games that Problem Attic creator Liz Ryerson calls “puzzle worlds” – perfect systems “without disharmony” which exist mainly to flatter our desire for totality. In the wake of the 2008 financial crisis, a group of economists publicly excoriated their own profession for “constructing models that, by design, disregard the key elements driving outcomes in real world markets.” The North American Free Trade Agreement which both Donald Trump and Bernie Sanders want to scrap was originally sold partly on the basis of wildly optimistic economic models. And while the models which drive D3 are simulating African economies, they are ultimately produced by academics and bureaucrats in the first world (even the pan-African Mo Ibrahim Foundation has its HQ in London). Might they not be just as disconnected and ideological?

“I see where you’re coming from,” says Jeff. “The simulation is based on procedural aspects that have been created by western institutions to model economic and political data. But these are well-established models that have been used to run analytics for decades.” On the other hand, he points out, Nigeria became Africa’s biggest economy in 2014 simply by revising the way it calculates GDP. “So if it’s biased, it’s biased by a primary source,” says Jeff. “If there’s a bias, it’s the bias of global consensus.”

In other words, just as the same technology used to build the hydrogen bomb was eventually redeployed to play Missile Command, here the technology of economic forecasting has been transplanted into a pleasant indie game. The children of MONIAC are emulating the children of ENIAC. In that sense, what we are seeing is a kind of final stage of economic modelling: one in which the model simply dispenses with the real world, curling back into itself under the guise of a fictional universe; econometrics finally freed from the contradictions of the Earth.

Not long after interviewing Jeff, I go for a drink with a friend who works for a global aid charity, soon to travel to Bangladesh. He is trained in economics too – a refugee from a blue-chip accounting firm. When I explain D3:A, he’s initially surprised, but then he starts giggling, which progresses quickly to fits of laughter which continue for half an hour. As someone who knows what gaps in the theory can look like from ground level, D3’s perfect Ouroboros circularity is just too much.

A fantasy of government

But of course D3:A is not really a closed loop. By naming its model nations after real African countries it hooks itself into real-world debates. By offering to teach western players something about these places it feeds back into their real lives.  

Cliff is painfully aware of the responsibility that implies. “We are very conscious of the fact that it’s basically two middle class white guys, neither of which have been to Africa, making a game about Africa. And we expect to be pulled up on a few things – that’s bound to happen, with people claiming ‘this just isn’t the case in our country.’ But I didn’t want to not do it out of fear, because that just reinforces yet another reason not to cover Africa. And frankly, if people say ‘here we are as African game developers, and we’re going to do a political game that actually says what’s going on’ – well, that’s great.”

Later, he emails: “The plan for D3:A has always been to represent the country exactly as it actually is, rather than through the lens of western media… hoping what we can do is just provide a fairly accurate simulation and let the player take from that what they will.”

Yet there is one story Democracy 3 can’t help tell. By its nature it will always endorse a myth about government which says that politics can be fixed if one very smart person is just given the information; that this information can be interpreted in a stable way towards coherent, lasting decisions. As Jeff puts it, D3 “idealises the dictatorial decision-making of a strong and popular leader who can implement their policies without the complication of politicking.” When you apply this managerial myth to western players in Africa, you get something more pernicious, something like The Last Samurai for economics. Elsewhere at GDC, games for change don Colleen Macklin spoke about how a game designed to teach players what it was like to be poor sometimes made them less empathetic, because they were able to ‘win’ it and wondered why others weren’t too. I can easily imagine naïve kids playing D3:A and deciding that all Africa needs is a sufficiently decisive white person with absolute power. 

That is emphatically not what either man wants to teach. “We’re trying our best not to be like that,” says Cliff. “People have learned from D3 that governing is hard and that it almost always involves compromise. Hopefully it has given people a lot of political empathy.” Jeff, likewise, doesn’t want anyone to come away from D:3A believing they know how to fix Senegal; the game is an abstraction, and real politics is much harder. Instead, he says, players should learn broader lessons about the vast range of different polities which exist in Africa and the very wide range of different paths available to them. “If you’re very resolutely basing your life, your theology, your worldview, on static rocks that you think are immutable, then you should reconsider that position,” he says. “Because everything is fluid.”

I don’t think the game can quite live up to that. Although D:3’s sheer complexity always threatens to dissolve into chaos, it can be mastered, and players who can do so will treat even Cliff’s planned electioneering DLC pack as merely a extra set of problems to compute. On the other hand, I get that this a commercial product, whose target audience want to play it because it lets them create and achieve things. Even if you could radically reform it, how would you sell it? Imagine the bullet points on the back of the box: “Experience the inscrutability of causation in a complex world! Feel the TOTAL futility of all directed action!”

Perhaps the way out is to remember MONIAC and think of D3:A as a metaphor, a teaching aid rather than a textbook. Hopefully people will genuinely learn things from it that they didn’t know about Zambia. But it’s just as important to learn a way of thinking – to understand the economic lens, and be able to deploy and debate it. If you’re asking whether some aspect of the simulation might be wrong, and wondering how it could be different, it has already more than done its job. The map is not the territory, the model is not the economy, and Democracy 3: Africa is not Africa. Get that straight and you’re at least on the path to using its powers for good.