Any discussion of income inequality has to confront facts about the size and composition and history of the middle class. A simple economic model offers a good start. Suppose that something called ‘human capital’ is randomly endowed to people in something called a ‘normal’ distribution, the more familiar ‘bell curve.’ In this case, a small share of people is endowed with low levels of human capital, and an equally small share has high human capital. Most of us are clustered around the ‘average.’ This is a compelling assumption, since height, IQ and many other human characteristics follow this pattern.

Now suppose that income is strongly correlated with this thing called ‘human capital.’ In that case, there will be a large middle class, and much smaller shares of relatively rich and relatively poor people. This looks a great deal like income distribution in the USA, circa 1960.

The model is not very realistic, most especially because there is not one thing called ‘human capital’ but many dimensions of skills and education. To accommodate this, now suppose there is a second type of human capital, call it human capital ‘B’, which is likewise randomly distributed to all people, but is not related to the first type of human capital. The result of this is to flatten the ‘bell curve.’ There will be a smaller, but still large middle class, and more people at the ‘tails’ of the distribution.

If income remains tied to overall human capital, each time we add a new type of unrelated human capital, we get a smaller middle class, and more people who are very poor and very rich. Eventually, the distribution of income becomes perfectly uniform, with an equal number of high, low and middle income workers.

This is much more like the way the world actually works. Such human capital endowments as intelligence, judgement, industry, social adaptability and leadership ability do not usually fall on the same person, but are distributed across the population.

This simple model suggests the middle class of my youth was a passing phenomenon. So, too, would the unusual dynamics of labor markets a half century ago. From about 1945 to the mid-1970s, wages in the United States were heavily influenced by a very large, heavily monopolized manufacturing sector. Think General Motors. The historically unusual wage setting process in that industry heavily valued a single dimension of human capital (union loyalty). Maybe a third of all workers experienced this arrangement.

One result of this was a large middle class, which many of us view nostalgically. Another result was the miserable failure of the auto industry in the face of both domestic and international competition. The presence of structured wage setting of a monopolized industry might well have created a middle class, which largely did not exist prior to World War II. It also led to its demise.

Among the challenges before us is figuring out how our institutions, policies and political rhetoric adapt to a world where a large middle class may be nothing more than an historical relic.