[twoodfin]

I'm pretty skeptical of the study on which this essay bases its premise that economic mobility in the U.S. is limited compared to, say, France. The most detailed information I can find online is the executive summary[1], which has to handwave that it found Italy with a strongly negative correlation between parents' and children's economic outcomes! That suggests to me that their methodology is not particularly robust.

I've found similar attempts to establish parent/child economic correlation equally suspicious when, for example, they measure income correlation with percentiles rather than in adjusted dollars (it's much easier to be "mobile" if there's only $10K in income separating the 40th and 60th percentiles!)

[1] http://www.pewtrusts.org/~/media/legacy/uploadedfiles/pcs_as...

[TheOtherHobbes]

The same results have been reported by many other studies. The facts are clear - for the poor the US is a land of economic handicap, not a land of opportunity.

There will always be individual exceptions to this, but they'll be single data points. Put crudely, for every heart-warming success you see interviewed in Forbes, there will be millions of failures no one hears about.

The most useful picture is broad-based and statistical, and studies like this one:

http://www.equality-of-opportunity.org/

show that mobility depends as much on where you were born as to whom you were born.

Broadly, inequality has exploded since the 1980s and in areas with limited social capital - including good free education - it's now more difficult than ever to work your way up from the bottom.

But this is balanced by increased opportunities in other areas - mostly affluent, mostly urban - which have created a halo effect for the poorer communities around them.

So average mobility has remained approximately constant, but only because bad areas have been balanced by good areas.

Meanwhile average mobility in the US continues to be worse than mobility in other countries.

[twoodfin]

Meanwhile average mobility in the US continues to be worse than mobility in other countries.

What's the best study you can find that demonstrates this? I haven't seen any that don't appear to succumb to the flaws I mentioned elsewhere in this thread, such as measuring relative mobility instead of absolute mobility.

[TheOtherHobbes]

We should probably define terms first.

Which definitions of absolute and relative are you using? (There are a few, and they're not identical.)

[twoodfin]

I would define absolute income mobility in terms of some globally comparable good. Dollars, PPP adjusted or not, for example.

Relative mobility to me means mobility within some particular income distribution: Moving from, say, an average household income in the first quintile to an average household income in the fifth quintile.

I prefer absolute mobility as a measure of economic mobility for the same reason that any rational actor would prefer to win a lottery for the difference between the income of an average bottom 20% household and an average top 20% household in the U.S.[1] vs. Sweden[2]. According to OECD, that's $75K vs. $37K.

[1] http://www.oecdbetterlifeindex.org/countries/united-states/

[2] http://www.oecdbetterlifeindex.org/countries/sweden/

[happyscrappy]

Do you have a citation or not?

[tehrania]

That's the whole point of economic mobility, to measure it against income and inequality.

[twoodfin]

No, that's not the point. A society with almost no income inequality, where everyone makes almost exactly the same amount, would have nearly perfect "mobility" as measured by the correlation between the percentile of parent and child earnings (a few random dollars would move you a lot!). But it wouldn't be a "mobile" society in the sense we usually mean, where the poor can become rich (and vice versa) because there would be no poor and no rich!

[kcorbitt]

It depends on what you're trying to measure by "economic mobility" though. I can think of valid reasons to want to know the percentile change between a person and their parents, irrespective of the absolute value difference, and reasons to be interested in the different-but-related measure of absolute change in income/quality of life.

It's generally better to start with the question "what do I want to know?" and then looking for a suitable statistical framework that will answer that question, rather than starting with a ready-made statistical analysis and then trying to use it to answer arbitrary questions.

[twoodfin]

I can think of valid reasons to want to know the percentile change between a person and their parents, irrespective of the absolute value difference, and reasons to be interested in the different-but-related measure of absolute change in income/quality of life.

Sure, but when comparing nation A to nation B, it's generally more interesting to look at either the absolute dollar difference or some relative quality-of-life measure (you could call the latter "class", I guess). Folks moving around within a tightly distributed income distribution don't really gain or lose much of either, relative to someone in a society with larger disparities.

[watwut]

Mobility is about moving between classes of society. So, the society you just described would be rightfully described as having high social mobility. What it probably does not have is high average wealth.

[twoodfin]

So, the society you just described would be rightfully described as having high social mobility.

But it would just be economic Brownian motion. Social mobility isn't (I hope!) valued because it shows that a society produces random outcomes, but rather that it enables people to achieve to the limit of their talents.

[watwut]

Social mobility is defined somehow and that somehow is not "the absolute total measure of fairness of any society". And speaking about virtually any country in the world, including North Korea, they all have lower class and upper class. Your worry about totally classless society is purely theoretical.

If we would live in classless society, social mobility would be less important measure.

Plus, you define achieving the limit of the talent in purely financial form. If I am genius mathematician, achieving limits of the talent mean doing great (potentially useful) math somewhere at the university. That does not translate to neither "high class" nor "rich".