Policy Innovation Blog

What 'Civilization' taught me about technological progress

Stian Westlake - 15.06.2012

Or: How do you find the low-hanging fruit in an invisible orchard?

Sid Meier's Civilization video games have eaten hours upon hours of my life, many of them between 1am and 4am. So I was delighted to see this week that someone had dragged out a game of Civilization 2 for 10 years (the world became "a hellish nightmare of suffering and devastation" by 3000AD - ha). But it got me thinking about how much this addictive god-game had shaped the way I think about technology and how it evolves.

A quick word of explanation for anyone who doesn't know Civilization (what have you been doing with your life?). It is a strategy game in which you play the leader of a world empire. Among other things, you're in charge of directing technological research, and by investing in science you can develop technologies from pottery to nuclear fusion. Here's where it gets interesting: you can only discover certain technologies after you've discovered other ones: so for your civilisation to "learn" nuclear fission, you need first to learn physics, which requires you to "invent" the scientific method. All the technologies you can discover can be grouped into a sort of tree, showing how each one leads to the next. You can progress at slightly different rates along different branches of the tree, but the linkages make it hard for progress to happen in very different ways.

When I think about technological progress in my day job, I find that the Civ technology tree has, er, rooted itself in my mind. Perhaps it's no surprise: we know that playing games is one of the most effective ways to learn, so it makes sense that the economic rationale of a game like Civilization might inform your thinking if you play it a lot. (And boy, did I play it a lot.) In any case, I find the idea that there is a set of technological opportunities that are available to us, that are out there in the ether to be discovered based on what we know currently, very intuitive. It accords with Steven Johnson's reflections on simultaneous discovery and the "adjacent possible". And with Brian Arthur's idea of technologies as things that build on other technologies.

This may all seem very abstract. But it has some important implications for how we think about the big political issue of the day: economic growth. Clearly technological progress is a big driver of economic growth. But it's always had an odd position in economists' theories of growth. For a long time, economists treated it as a mysterious exogenous variable, the factor left over once the effects of labour and capital had been accounted for. Starting in the 1980s, economists like Romer and Aghion began to bake innovation into their models. For the most part this happened by treating research as another input into the growth recipe: spending money on research led to steady increases in productivity and output, just as investing in new machines or more skilful workers would. (This article by Richard Nelson provides a good overview.)

All this assumes that the relationship between research and technology is smooth, or at least definable. In the words of one economist, it treats the relationship between technology and growth as like that of yeast and dough, not like that of mushrooms sprouting in a forest. This has informed the current debate on the so-called "Great Stagnation", the idea that technological progress is slowing and dragging economic growth down with it. Tyler Cowen and others who've expressed scepticism in the prospects for technology-driven growth (such as Chad Jones) are essentially arguing that we're gradually exhausting the technologies available to us. In Civilization language, we're climbing the technology tree and with each step up, the inventions get costlier and less marginally useful.

This is a grim picture in itself, and has been disputed by people like MIT's Erik Brynjolfsson and Andy McAfee, or Ryan Avent of the Economist (warning: article contains taco-copters). Perhaps our civilisation's latest invention, IT and the internet, actually makes it cheaper to devise new inventions. Perhaps existing inventions can be combined to enhance their power. Perhaps entrepreneurship matters as much as tech. All this gives us some cause for optimism: maybe we aren't facing a great stagnation, or the technologies we've devised happen to give us a way out.

But there's an even more troubling possibility in these uncertain times. As any Civilization player knows, not all technologies on the technology tree are equal. Not all are as easy to invent as others, or as valuable. Inventing steel has all sorts of economic and military benefits; inventing alchemy is less useful. But the progression is not predictable: some cheap technologies are very useful, some expensive ones are not.

We see this in real life too: it currently seems extremely easy to invent useful and valuable social media technologies, but in the early 2000s, lots of early attempts failed. In the 1950s and 1960s, clinical sciences developed wonder drug after wonder drug, but before 1950 and after 1980 the productivity of clinical science dried up. And in real life, you can't cheat by peeking at the technology tree to see what technologies are coming up next. In the early 2000s, it was thought that biotech and genomics would yield huge benefits. But so far they've mainly produced new and more fiendish problems to solve.

This complicates the idea of the Great Stagnation. What if economic growth does depend on technological progress and on investment in research, but in an essentially unpredictable way? The economist Michael Mandel speculated that economic growth has slowed not because of a permanent tech slowdown, but because a few great developments (cloud computing, biotech) had taken a few more years to develop than we'd expected.

Maybe the problem is not that we've eaten the low-hanging fruit, it's that the orchard is invisible. It's a sobering thought for anyone trying to understand future prospects for growth.