The Schumpeter hotel: income inequality and social mobility

In one  of his rare discussions of inequality, Joseph Schumpeter illustrated in a metaphor the difference between the inequality we observe at a moment in time and social (or inter-generational) mobility. Suppose, Schumpeter writes, that there is a multiple-story-high hotel with higher floors containing fewer people and having much nicer rooms. At any given moment, there would be lots of people on the ground floor living in cramped small rooms, and just a few people in the nice and comfortable top-floor rooms with a view. But then let the guests move around and change the rooms every night. This  is what, Schumpeter said, social mobility will do: at every given moment of time there are rich and poor but as we extend the time period, today’s rich are yesterday’s poor and vice versa. The guests from the ground floors (or at least their children) have made it to the top, those from the top might have tumbled down to the bottom.

Now, Schumpeter’s metaphor was for a long time a metaphor for US inequality too. It was granted that in the 20^th, and even in the 19^th, century US income inequality might have been greater than inequality in Europe, but it was also held that US society was much more fluid, less class-bound and that there was greater social mobility. (That view of course conveniently overlooked the huge racial divide in the US.) In other words, inequality was the price that America paid for high social mobility.

This was a reassuring  picture consonant with the idea of the American dream.  But was it  true? We actually never knew it, beyond anecdotal evidence of migrants’ lives, since no consistent empirical studies of inter-generational mobility existed until very recently. But before I go to their findings,  I would like to focus in  very simple terms  on the relationship between inequality and social mobility.

Consider a diagram classifying societies according to social mobility and inequality. From the US example mentioned above, we  would place US in the <high inequality, high mobility> quadrant. It is easy to imagine <high inequality, low mobility> societies: feudal societies will be one extreme example but all societies with high income differences and entrenched power of the elites would fall into that category too. So, let us write in Latin America or Pakistan in the SE quadrant. It is also relatively easy to imagine what countries we would place in <low inequality, high mobility> quadrant: probably Nordic countries with strong public education that allows high inter-generational mobility while redistribution of current income ensures low inequality.

It is much more difficult to find examples of <low inequality, low mobility> societies. It seems somewhat natural to think that if a society exhibits low inequality, it will be hard to keep sons and daughters of people who have just a slightly lower income (than another group) permanently below incomes of sons and daughters of that latter group. One can even wonder what mobility in these cases really means: if incomes between people and classes differ by an infinitesimal amount and your children remain richer by that infinitesimal amount above mine, I am not sure that that kind of lack of social mobility really matters much. Perhaps some guild-like societies where occupations cannot be freely chosen but income differences between the occupations are small could be placed in that category. Communist societies had some aspects that could make them (weak) candidates to be placed in North-East quadrant. 

	High social mobility	Low social mobility
Low income inequality	Nordic countries	Guild-like societies Communist ?
High income inequality	US (American dream)	Latin America Pakistan

So now that we have organized our thinking, let us consider the empirical evidence. Most famously, it comes from the recent work of Miles Corak, building on previous studies by Gary Solon, Blunden, Gregg and Macmillan, Björklund and Jäntti and others.  What these authors find is that there is a strong correlation between current and inter-generational inequality, or in other words, between inequality and low social mobility: the more unequal the society the less likely is the next generation to move upwards (or conversely, the less likely is the decline of the rich). So in terms of our simple diagram, Corak finds that societies are aligned along the diagonal: there are no outliers, whether the societies exhibiting the American dream or the guild-like ones.

The implication of that finding which was dubbed by Alan Krueger the Great Gatsby curve is that there is no American exceptionalism. The comforting picture of high inequality which does not impede mobility between generations turns out to be false. US does not behave any differently than other societies with high inequality. High income inequality today reinforces income differences between the generations and makes social mobility more difficult to achieve. This is also the point of my recent paper with Roy van der Weide. We use US micro data from 1960 to 2010 to show that poor people in US states with higher initial inequality experienced lower income growth in subsequent periods).

This important finding that the actually existing societies (as opposed to the dreamed up ones) are aligned along the diagonal of our table has two important implications:  (1) American exceptionalism in the matters of income distribution does not have a basis in reality, and (2) we can use, with a good degree of confidence, the easily available data on current inequality as predictors of social mobility. Thus one cannot argue that societies with high inequality in incomes are societies with high equality of opportunity. On the contrary, observed high income inequality today implies low equality of opportunity.