All Commentary
Friday, November 1, 1996


Some Computer Games Suggest a Society Can Be Planned

My daughter, at 13, has built a railroad across Europe. She has planned and governed a large metropolitan area. Recently she began a successful farm. Oh, and she’s conquered the world. She has done it all at our home computer—sometimes while I was waiting for my turn to play.

Welcome to the world of personal computer simulation games, an important segment of the PC software market. The games are fun. They’re involving. They can be so involving, in fact, that players can find themselves spending many hours at the computer and taking the games very seriously. In some ways the games are quite educational. But some simulation games, implicitly, embody an economic point of view that is miseducational—even dangerous—when it is taken seriously. Namely, they suggest that an entire society can or even must be planned from the top down by a clever strategist.

Three games—all members of Computer Gaming World’s Hall of Fame—have been our favorites in the past few years. In SimCity your objective is to plan a thriving city in detail. Civilization puts you in charge of founding and guiding an entire nation of cities from ancient to modern times. In Railroad Tycoon you build a rail network and schedule trains in pursuit of profit.


SimCity, created by programmer Will Wright and introduced by his Maxis software company in 1989, was the first of the best-selling simulation games. The game has won many awards, and Maxis has cashed in on its popularity with numerous other simulations, including SimFarm (which simulates a farming enterprise), SimLife (Darwinian evolution), SimIsle (island ecology), and SimAnt (an ant colony). Two years ago the company upgraded the original game to run on newer, more powerful PCs, especially improving the graphics, and reintroduced it as SimCity 2000.

In some ways SimCity is more of a virtual hobby kit than a game. There’s no rival player to beat or set objective to achieve. Rather, you decide where to zone a wilderness for commercial, residential, or industrial development. You decide where to lay down transportation, water, and power grids, and where to locate airports and police and fire stations. You set the property tax rate. You then watch how your city develops in response to your decisions. If conditions are good, lots of simulated citizens (Sims) move in. Houses, stores, and factories appear and grow denser. As your tax revenue allows, you can extend or revise the infrastructure. Good decisions—as indicated by high Sim incomes with low unemployment, crime, and pollution—give you a high popularity rating as mayor.

As much fun as it is to play around with SimCity, the simulation rests on some troubling premises. Not only is clever zoning (and its coordination with mass transit) the key to the simulated city’s success, but without zoning there would be zero development. The Sims won’t ever build houses, stores, or factories without a zoning go-ahead. The city must be planned. Of course, this feature of the program is practically necessary in order to give the player something to do (other than watch the simulation run itself). But a teenager who grew up playing SimCity or SimCity 2000 might not realize that top-down zoning is not a necessary feature of the real world, let alone a good idea. Real-world cities like Houston have developed well without zoning. In areas of an unzoned city where landowners want to have not-too-dissimilar neighbors, decentralized developers can and have incorporated contractual covenants into land titles to ensure the character of residential subdivisions or shopping districts or industrial parks.

I’ve tried to suppress my ideological misgivings about SimCity by thinking of my plaything as a proprietary city, where every Sim who moves in has signed a contract delegating zoning and other decisions to me, the city owner. But the game doesn’t quite allow that interpretation. Unfortunately a player (or at least a player like me, not a skilled hacker) can’t edit the screen text so that the Sims are paying dues or rents rather than taxes, or alter the scoring so that success is rated on real estate profits rather than on popularity as mayor. (As the game stands, absence of low-rent housing actually counts against your success rating.) And the simulated city, with a population easily reaching tens of thousands, is simply too big to be realistically owned by a single decision-maker. In real life, proprietary or contractually planned communities are generally much smaller than that, seldom larger than a village. This is mostly because tastes differ. Profit-seeking developers find that only a limited number of people are willing to pay cost-covering prices for a specific set of community features (like a golf course, a lake, and one-acre lots) in a given location.


The notion that one mind can see and wisely plan everything for a society—an intellectual error that F. A. Hayek has called the fatal conceit and others have called the synoptic delusion—is troubling enough when the society operates at the scale of a city. In Sid Meier’s Civilization, published by MicroProse, the conceit is magnified to empire scale. As civilization-builder you decide where to found cities, where to irrigate and build roads around them, what tax rate to impose, what city improvements and wonders of the world to build, how heavily to invest in research (needed to make scientific and cultural discoveries from the wheel to women’s suffrage), which discoveries to pursue first, and most importantly whether to expand the civilization by economic growth or by conquering neighboring civilizations (which are governed by the computer).

I don’t mean to be an ideological spoilsport about this. I love playing Civilization. So does my daughter, who has recently begun lobbying me to upgrade or replace our current computer so that we can run the more sophisticated new Civilization II. But players of the game should not forget that the advance or stagnation of an actual civilization does not entirely depend on the allocational skill of the head of state. In fact, the less the head of a real-world state has to allocate, the greater the country’s progress will be.

It is surprisingly easy to be seduced by the synoptic delusion built into the game. In an Internet discussion group, one correspondent suggested that Civilization teaches us that in the real world we should only elect experienced rulers, because we as game-players all find that our scores improve with experience. But that is because, as in SimCity, hands-on management is necessary for the simulated society to thrive. Complete laissez faire is not even an option, much less the highest payoff option. Civilization‘s cities will not found themselves, will not decide what improvements to build or technological advances to make, and will not establish trade with other cities by themselves. A player soon finds that the easiest way to garner a high score (which depends mostly on population size) is not by peaceful trade and growth, but by aggressive military conquest, for which a single-minded strategic plan is essential. Again, all this is understandable from the point of view of giving the player something interesting to do throughout the game. But it’s not the way the real world works, where trading and not war-making nations prosper. (Civilization II, so I have read, at least strikes a better balance between conquest and trade.)

Railroad Tycoon

My favorite game of the three is Railroad Tycoon, also authored by Sid Meier and sold by MicroProse, no doubt partly because no economic or ideological misgivings intrude. A railroad network, as shown by real-world experience with private rail companies, is something that can successfully be run according to a single planner’s decisions. In the game, as in the real world, that planner is an entrepreneur who decides which cities and industrial sites to serve, where and when to lay the track between them, which engines and cars to buy, and how to schedule trains. Profit is the scoring or feedback guide for when a decision is successful or unsuccessful.

With maximum profit—or, as in the game, maximum net worth of the firm—as the goal, market prices indispensably guide the decision-making. The game forces you to consider the prices of track, bridges, and various parcels of right-of-way (the land parcel prices even vary with changing phases of an irregular business cycle). You confront prices of buying and maintaining various engines. There is an interest rate, varying with the firm’s indebtedness and the business cycle, at which funds can be borrowed to finance the firm’s investments. And there are of course passenger and cargo revenues for various routes (known only after the fact, because they vary with cargo, load size, business cycle phase, distance, and realized speed).

Railroad Tycoon does not capture the real-world way in which prices are brought into equilibrium by competition (rate wars can arise in the game’s cutthroat competition mode, but their simulation is disappointingly crude). Nor does it hint at the way in which the profit motive operating within the price system guides the overall economy to the prosperous coordination of plans. But that’s okay. A successful real-world economy doesn’t need entrepreneurs to comprehend those system-wide processes, only to seek out profits. In contrast to the other simulation games, Railroad Tycoon puts the strategic decision-maker in a reasonable place, at the head of a single firm in a competitive marketplace rather than at the head of a whole society. It does not obscure the fact that the overall coordination of a complex real-world society does not and cannot rely on a single clever plan, but instead requires an impersonal market system for bringing into alignment the multitude of individual plans.