I’ve been thinking a lot about decentralization lately, in both a technical and political context. Writing about the technical side is difficult because the examples I’ve been thinking about involve stuff at work that I can’t talk about, so I’m going to concentrate on the political context.

The first example of political decentralization is military and foreign policy. I was in a discussion about “American hegemony” yesterday, and it made me think of an article in last January’s Atlantic magazine. The specific article was by Benjamin Schwartz and Christopher Layne and was entitled A New Grand Strategy (I can’t link to it because the Atlantic doesn’t seem to make old articles available for free on the web any more). Its thesis was that the US is trying too hard to maintain its status as “sole superpower” when it should be better off pursuing a more multilateral strategy. We are, the authors state, keeping our allies in an “infantile” military state, assuming the role of global policemen on their behalf – and against their wishes – because the alternative would be the development of their own military capability to a point where it might rival our own. This worries us even though they’re our allies, because we can’t bring ourselves to trust them. Yeah, I know, the right-wingers are thinking of France and saying of course we can’t trust them. Grow up. They might not be cooperating as much as we’d like, but the idea that they’d actually attack us is flat-out ridiculous; we can trust them that much. Wars happen when one side thinks it can win. If they are one of many approximately-equal powers, where all are committed to act together against aggressors, nobody will think they can win. We can afford to let others “responsible” nations develop their military capabilities, and scale back our own. Not only would we save a ton of money that could be used to address domestic problems (or be put back in citizens’ pockets), but we’d also be more secure if we weren’t the sole scapegoat for every dysfunctional society or regime on the planet.

The second example of political decentralization is economic policy. Centralization of political power has the same drawbacks as centralization of military power, or centralization of control in a computer system. OK, I couldn’t stay away from the technical stuff forever. ;-) Here are some of those drawbacks:

  • In a centralized system, a failure or flaw in the central component can have a devastating effect on the entire system.
  • A centralized system’s capabilities tend to be limited by the capabilities of its central component.

Concentration of power is bad. So what? Here’s the key: wealth can be translated into power. It can buy information, and access, and visibility, and time (not necessarily one’s own) to pursue power; some think it can simply buy politicians. Therefore, concentration of wealth is also bad. The problem is that concentration of wealth is a natural phenomenon. There’s a big difference between “earning” interest and paying interest, between being pursued by merchants and being avoided by merchants, between hiring an assistant and being an assistant. Wealth naturally tends to become concentrated, and that’s a danger to everyone.

The solution, of course, is not to concentrate power somewhere else – e.g. the government. The real solution is to create a balance between sources of power, just as in the military example. The government is the only entity capable of countering the problems of concentrated wealth; it, in turn, can and must be restrained by the people from the arbitrary exercise of its own power; the wealthy, in their turn, exert significant influence on the people. It’s a three-part system, similar in some ways to the three-part “separation of powers” that underlies the US constitution – checks and balances, so that nobody can “win” and cause the game to be over.

I could go into a lot more detail, of course. In particular, there’s a huge tangle of follow-on ideas about taxation and its use to promote or retard the accumulation of wealth, but I’m deliberately not getting into that right now. The idea I’m trying to get at is that, in the political realm as much as in computers, multiple entities able to exert influence on one another can result in a system more stable than one in which one entity tries to control all the rest.