Thursday, May 26, 2005

Mobility again 

Mark Gisleson at Norwegianity, about whom I wrote a post wondering why he had a problem with something James Taranto had said (of which I approved), has decided to return fire. It's a long and rather tedious post, full of fury that the "culture of business" is "f****** over people" and that "The system has been gamed by the people running the show, and they�re not even trying to play fair." He refers to it as a fisking. Well, call it what you wish, I don't particularly care. Let me respond to substantive points and not play with his pyrotechnics. I think his points are threefold:

  1. The data I use is older and not telling what is happening right now;
  2. The data show that income mobility is declining; and
  3. It's all the Republicans' fault.
Quickly, my replies are 1) the data is hard to come by and doesn't change dramatically year to year, and it's hard to control for other factors that change over time; 2) we agree that mobility is less than twenty years ago, but that 3) the trend has been happening for so long that it's unlikely due to any specific policy of either party.

Over at Cafe Hayek, Russ Roberts has been running a series on income mobility.

The bottom line is that it is extremely difficult to measure mobility and opportunity. One way to avoid all of these structural changes that contaminate the numbers is to look at the same people over time, rather than looking at snapshots of different people over time.
This is a major problem with how numbers get used in the debate. The size of the survey you need to calculate the distributions are large -- in the States there are only two or three sources you could use, including the Census (full disclosure: I have no exposure to doing these studies here; I have worked with ones in Ukraine and Armenia, but not much and they're not comparable) -- and so mobility studies tend to be somewhat infrequent.

I used the chart that Mark apparently had trouble reading because it tries as best possible to look at people longitudinally. It says that 59% of people in the lowest quintile of the income distribution in 1969 moved up a quintile by 1994. Nearly 53% went up from the second lowest quintile to the middle. &c. The part that should catch your eye as much as that on the right hand side you see over 61% of those in the top quintile in 1969 moved down. If The Man was always protecting his own, how come 3 of 5 rich people in 1969 were relatively worse off by 1994?

Mark's second major point is that it's not that there is mobility, but that there's declining mobility. The evidence does support that; my statement is that it's still substantial (did I spell that right?) And the decline in mobility pre-dates the present Administration and, in my judgment, has declined for reasons that are not related to any government policy. Roberts links to a study from the New England Federal Reserve (link is to a non-technical summary; here's the more technical article if you wish to wade into deeper water.) Here's Roberts' summary of the results.

What percentage of families during the '70s, '80s and '90s stayed in the same quintile of the income distribution:

1970 36%
1980 37%
1990 40%

What percentage moved up or down one or more quintiles:

1970 38%
1980 39%
1990 39%

What percentage moved up or down two or more quintiles:

1970 26%
1980 24%
1990 21%

Roberts and I are not sure what to make of that. The study Roberts quotes thinks it has to do with the increasing number of females heading households. Female-headed households are simply poorer that families with two adults present. The median income of a family with two working spouses in 2000 was $74,000, up 48% over 1970, but the median income of a female-headed household $27,000 and up only 28% from 1970. Higher divorce rates and increasing numbers of kids born out of wedlock play a role in these data. (Interestingly, Roberts notes, the income of households headed by a male and no female adult present has grown even less than for female-headed households.) As well, immigration plays a huge role. Roberts is rather skeptical that it plays a role, but I think it might be more: Think of the new enterprises immigrants creating high-tech firms in California. Those weren't McJobs. Is it enough to increase wages overall? Hard to say, but see this. (Interestingly, I also found that African immigrant household median income is more than $9000 greater than that of African-American households; this makes me wonder if the gains in black per capita income are also influenced by immigration. I didn't look further into this yet.)

But ascribing the mobility numbers to some thought that "Bush did change everything" is simply missing the point. The trends we're looking at have been going on for quite some time, for a variety of reasons -- which we cannot reliably disentangle, but seem to have little to do with public policy. I don't have the luxury of easy socialism that Gisleson turns to with his jeremiad against businesses (and their Clinton bagman Bob Rubin, he says) to provide simple answers, but neither do I have an easy response to why mobility is declining. If Mark wishes to take comfort in that or launch another blog missile, expect nothing more from me.

It appears as well that it's not clear what a government could do other than maintain a healthy market economy. A recent article in The Economist wondered why India, a democracy with healthy electoral competition, had done so much more poorly reducing poverty than China, which has no electoral competition. Citing a World Bank study, it says
Why might democracy militate against poverty reduction in poor countries? Mr Varshney has two suggestions. First, democracies have a bias towards �direct� methods of tackling poverty, such as subsidies and hand-outs, which, in the long run, are less effective than �indirect� methods�ie, those that generate faster economic growth. In India, this seems undeniably true. Governments have built up whopping budget deficits, thanks largely to subsidies. Many farmers, for example, receive subsidised or free fuel, fertiliser, electricity and water. But little public money is spent on improvements that would do most to lift the growth rate: in infrastructure, primary education and basic health care. Everybody wants better roads, and nobody votes against them. But every politician promises to build them and hardly any do. Cutting subsidies, on the other hand, is a sure vote-loser.

Second, the poor are not necessarily a homogenous group. In a democratic system, they may organise themselves along lines other than economic class and �the shared identities of caste, ethnicity and religion are more likely to form historically enduring bonds�. If you are born poor, you may die rich. But your ethnic group is fixed. In India, with its myriad linguistic and caste-based groups, the upshot is a dispiriting beggar-thy-neighbour politics. Just as subsidies are easier to deliver than are roads and schools, so are affirmative-action schemes, giving jobs to members of specified castes.
That description is apt elsewhere. Attempts to use the political process to redistribute directly sacrifices resources that could be used productively. This rent-seeking behavior occurs in any democracy where the government puts out to auction its coercive power to parties that win elections. I'll argue with my Republican brethren that they are as guilty of this as Democrats -- this was one of the two things that originally drove me into the arms of the Libertarians -- but those that respect the power of markets yield benefits that cut across class, caste or ethnicity.