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