Tim Lee's post on power laws and inequality is interesting, but seems to me to be quite wrongheaded. The putative existence of power laws just can't be taken as evidence that income inequality is somehow inevitable and "simply a natural characteristic of complex social systems" that we have to live with, for a number of reasons.

First - unless there is some research that Tim isn't citing to here, we don't know that the distribution of income in the US is a power law. There is an unfortunate tendency to describe all long tailed distributions as power laws. They aren't. On this topic and others, see in particular Cosma Shalizi's classic post (and the associated paper), So You Think You Have a Power Law: Well Isn't That Special.

Second, even if we did know that US income was distributed according to a power law, we still wouldn't know all that much. There is a very wide variety of mechanisms that can produce power laws. Without further investigation as to which mechanism is in play (which can't be discerned just from noodling the observed distribution), we may have a Nature publication, if the editors are dozing off again, but we haven't gotten very much closer to an understanding of what is causing the observed data.

Third, we have no reason whatsoever to believe that

Not only is there no practical way to affect an equal distribution of blog traffic, but it pretty clearly wouldn't be helpful even if it were feasible.

We are not living in the world of Harrison Bergeron - there is a rather large range of possibilities between the highly unequal distribution of attention that we see, and the complete flat distribution that Tim suggests as an alternative. Moreover, there is no ex ante reason to believe that the distribution of links, traffic and attention in the blogosphere is at all optimal. Long tailed distributions may emerge from 'rich get richer' phenomena, in which people who received a somewhat higher number of links early on, continue to accumulate more attention and links thereafter, even if they are qualitatively inferior to others who didn't receive this advantage, or who entered the game later. There is good reason to believe that the blogosphere is characterized by rich-get-richer effects. Bloggers (like me) who got involved early get more links, traffic and attention than bloggers who got involved late, even if some of the bloggers who got involved late are almost certainly 'better.'

Second, Tim is likely right that it's very hard indeed to see how to rearrange traffic patterns among bloggers. It isn't as hard (although it certainly isn't easy) to dampen income inequality if we want to - many European countries have done it far better than the US.

Fourth, even if all the above objections were wrong, and income inequality in the US did emerge from some 'natural' phenomenon, this does not at all invalidate the case for addressing it. Paul Krugman was writing about these issues back in the 1980s and 1990s (his book Geography and Trade is an excellent and under-read primer on skewed distributions in the economy. In this period, Krugman was convinced that income inequality largely stemmed from technological changes, which had led to a 'Sierra Madre' economy in which many headed off to find gold in the desert, but few found it. As he notes here, this didn't invalidate the case for addressing it.

How does one live with a Sierra Madre economy? The United States and Europe have in effect put forward differing solutions. In the United States, especially since Ronald Reagan, we have simply accepted gross inequity as the inevitable consequence of free markets; everyone heads off into the desert in search of gold (that is, gets a job), but a growing number receive very little for their efforts. ... Nobody has yet found a way to raise taxes or provide benefits without reducing incentives. The best that we can hope for is a society that shows more sense of community than the United States, but that does so in ways that exacts less toll on incentives than the European welfare states. ... we must provide generous social insurance, but structure it so that it helps the working poor as well as the unemployed. ... the financing of social insurance should, as far as possible, avoid curbing employment prospects for the less-skilled.

Krugman has since, of course, changed his mind about the mechanisms generating income inequality. Based on new empirical evidence, he now believes that income inequality is largely a product of politics, rather than of 'natural' economic and technological processes. This suggests a different - and more direct - set of solutions. Still, the point I want to make is that even if you did believe that income inequality comes from non-political processes, you don't have to abdicate responsibility for dealing with it. I'm not sure why Tim thinks otherwise. I can understand how, when you see a new set of theoretical arguments that seem to support your prior beliefs, you might be eager to take them up, but I think that Tim has unfortunately gotten the wrong end of the stick.

Update: Tim responds here. Two points in return. First - if all that he is claiming is that power-law type phenomena suggest that it is possible that unequal distributions of income may arise naturally, then they don't tell us very much at all (we have known this for a long time. The key question is whether the existing pattern of inequality in the US has in fact arisen naturally. For what it's worth, some experts, like Paul Krugman, who used to believe that it had, have now changed their minds given new evidence. This is a debate that can only be resolved through looking at actual empirical evidence regarding the real world mechanisms that are actually producing inequality - general talk about power laws doesn't tell us much, if anything at all.

Second, Tim greatly underestimates the variety of reasons why we might be worried about inequality. Distributional fairness is important - but more important still are the ways in which economic inequality perpetuate political inequality (through differential access to political influence), are causally linked to patterns of crime, are linked to gross differences in health outcomes and so on.