I thought it was just Jim DeLong, but it looks like there’s a full blown trend of libertarians opposed to a government procurement policy favoring open source software [OSS]. The most recent instance is a TechCentralStation piece by Sonia Arrison rather hysterically entitled “Source Socialism.” (I’m sure that will come as a surprise to hardcore libertarian and open source guru Eric Raymond! But then, according to the typically foaming-at-the-mouth Ayn Rand Institute, Lawrence Lessig is a Marxist too. Who knew?)

The logic here is so tortured that I frankly have to wonder (at the risk of resorting to a favorite ploy of lefties, the argumentum ad funderam) whether the Microsoft banner ad running alongside the article had as much to do with its conclusion as did impartial analysis. The setup is a sort of pointless jab at OSS advocate Bruce Perens. Arrison then flirts briefly with the notion that the software industry’s purpose is to provide programmers with an income — the kind of thinking usually reserved for legislators handing out pork — before conceding that this is beside the point. Finally, in the last third or so of the piece, we get something resembling an argument.

First, says Arrison, “Microsoft has market power because it creates products that satisfy technology needs at the right price.” If Linux (or whatever) were better, well then, by golly, the market would select it. Well, that’s true in most markets certainly, and libertarians are so used to advancing this line that you can all but pull a string on our sides and hear it repeated. But to just say that, without further argument, is to ignore the ways in which software markets (not to mention government procurement markets) are different. Operating systems are not like, say, sweaters or blenders. How “good” one is depends not just on the intrinsic characteristics of the product, but also on (among other things) how many other people are using it, on the extent to which it is “standard.” Government procurement is such a huge market that, paradoxically, it might prove to be the “better” system if and only if government chooses it. If the “better” OS is determined in large part by the choice government makes, then the choice itself is, in a sense, necessarily political. The question is not “which is better?” but “which will we make better”: an open system supported by a potentially unbounded community and under nobody’s control, or a closed one owned by a single firm and designed or modified exclusively for the benefit of that firm — which may or may not dovetail with the benefit of consumers? Yeah, I’ve read that Margolis and Leibowitz book too — I’m not really persuaded, and even if I were, it wouldn’t mean that path dependence (or “lock-in”) isn’t a real potential phenomenon. That is, even if (this is a stretch, but I’m trying to be discursively charitable here) Microsoft’s current dominance had nothing to do with path dependence and everything to do with the quality of their product, there’s no reason to think their current dominance mightn’t be parlayed into future path-dependence based dominance. I may not think government should “fix” such market outcomes via regulation, but it also doesn’t mean the state has to encourage the process.

Arrison suggests that government’s criteria should be cost and effectiveness, that it would be wrong to make a “political” decision to opt for open source. Well on those grounds — including security considerations — I suspect Linux might win out in any event. But actually, the threat of political influence is really the best reason to opt for OSS procurement. Opting for Microsoft in government not only places state systems at the mercy of a closed, proprietary system; it also gives one more large corporation a powerful incentive to lobby for the preservation of its sinecure. Consider, for example, that defense contractors who profit massively from the War on Drugs (not to mention the pharmaceutical, tobacco, and alcohol industries) are among the largest contributors to antidrug lobbying groups like the partnership for a Drug Free America. It is one thing to say the state should not play politics in the software market; it is irresponsibly naive to suggest that we can presume that outcome on the basis of our urging unless the institutional framework is set up to ameliorate some of the all-too-familiar public choice problems. The advantage of open source is that government systems need not rely on coders trained in Microsoft’s guarded and proprietary code; it can use its own in-house technicians, or contract out to any one of a number of firms specializing in Linux support. The more diffuse the benefits of government procurement, the less the incentives for untoward corporate influence on policy. I am as fond as anyone — indeed, more so than almost everyone — of the free play of markets. But like it or not, government is not just one more market participant like all others. It borders on obtuse to pretend that it is, or to expect it to act like one.

Pieces like this make me worry that we’re getting lazy. We have a set of pretty brutal arguments, which apply in a large number of contexts, to the effect that markets produce either near-optimal outcomes, or significantly better ones than government can manage. Maybe we’ve gotten so used to deploying these against people who don’t quite get the more elementary points that we’ve stopped thinking about how well they apply to, in this instance, government procurement markets, or markets with heavy positive externalities. That or some of us have grown addicted to Macrosalaries from Microsoft — something that, despite this suspicious record, I would prefer not to believe about people I consider, broadly speaking, ideological kindred.