Figure 5. Changes in average household net worth, 1983–1998. Sources: Wolff 1998; www.inequality.org.

      People in the middle simply don’t own much stock. Because their pensions were, for the most part, defined-benefit plans rather than defined-contribution plans, they did not benefit significantly from the stock market boom of the 1990s.

      As with income, the real growth in wealth came predominantly at the top. As shown in figure 5, for example, the bottom 40 percent of households actually experienced a significant decline in net worth between 1983 and 1998, a period during which the top 1 percent saw its wealth grow by more than 40 percent.

      But it was within the top 1 percent that the most spectacular changes in net worth occurred. For the past several decades, Forbes has published a list of the estimated net worth of the four hundred richest Americans. In 1982 there were only thirteen billionaires on this list, five of them children of the Texas oil baron H. L. Hunt. By 1996 there were 179 billionaires on the Forbes list, and by 2005 there were 374. Together the Forbes four hundred are now worth more than a trillion dollars, nearly one-eighth the national income of China, a country with one billion people.

      There were some 7.5 million American households with a net worth of at least a million dollars in 2004, more than 20 percent more than there had been just the year before.¹ If a net worth of a million dollars has become almost commonplace, a net worth of five million dollars still counts as real money. There were 740,000 such families in 2004, 37 percent more than there had been a year earlier.² Wealth at that level was once rare. Twenty-five years ago, if people worth more than five million dollars happened to find themselves in Ithaca on business, there would almost certainly have been an article about them in the Ithaca Journal. But now that almost three of every thousand people have a net worth that high, such events have become altogether unremarkable.

      My point is not that the creation of these big fortunes is by itself a bad thing; I cite these figures merely to present a rough picture of how the distributions of income and wealth were evolving in the United States at the end of the twentieth century. In contrast to the robber barons of the nineteenth century, most of the people who have amassed today’s big fortunes did so without having to crush labor unions with armies of hired thugs. And although there are obvious exceptions, most of today’s wealthy did not become rich by stealing money from others who had a rightful claim to it. Rather, they invented valuable new products and services and sold them to the public.

      But whatever processes may have been involved, the result has been that the distributions of income and wealth have become much more concentrated during the last several decades. Those in the middle of the income and wealth distributions have lost ground relative to those at the top, despite the absolute increases in their income and wealth. For them, to return to my first thought experiment, the United States has become much more like World A and much less like World B.

      CHAPTER THREE

      Inequality, Happiness, and Health

      With evidence on recent trends in income and wealth inequality in hand, we are now in a position to attempt to answer the question before us: Does rising inequality harm the middle class? One way to approach it is to try to answer a closely related question: Does growing inequality make the middle class less happy? Although few economists would pose such a question, I am persuaded that much can be learned from an attempt to answer it. As a first step, we require a workable measure of happiness. So I will briefly survey some highlights from the large literature devoted to the measurement of human happiness and well-being.

      Psychologists and other behavioral scientists have for several decades been trying to measure what they call subjective well-being.¹ I once asked Ed Diener, a pioneering researcher in this area, why he and his colleagues use that term instead of just saying happiness. He said, “Well, it is happiness that we’re studying, but we’d be much less likely to get NSF grants if we called it that. They wouldn’t think it was scientific enough.” But even though subjective well-being and happiness have much in common, the two concepts are not identical. Subjective well-being depends in part on how one feels at a given moment, but it also entails considered judgments about the overall quality of one’s life.

      In practice, one of the principal measures of subjective well-being comes in the form of responses to surveys that ask people to classify themselves into one of three categories: very happy, fairly happy, or not happy. Other surveys ask subjects to respond on a ten-point scale to questions like “All things considered, how satisfied are you with your life these days?”

      Economists often voice strong misgivings about such measures. What can we possibly learn, they wonder, by posing questions like that? This is the United States, after all, and most people know they are supposed to be happy here. But although some people may overstate their personal happiness levels on that account, a significant proportion are willing to admit to being only fairly happy. And a minority—a small minority, but still a sizable number—classify themselves as not happy.

      What is more, people are consistent in their responses. Their answer to a given question is likely to be the same eight months from now as it is today. If we try to measure happiness in other, less direct, ways, the results accord closely with the results we get when we just ask people whether they are happy. Suppose, for example, that a subject is asked to indicate on a five-point scale the extent to which he agrees with statements like “When good things happen to me, it strongly affects me,” or “I often do things for no other reason than that they might be fun.” (Five points indicates “strongly agree” and one point indicates “strongly disagree.”) People who strongly agree with such statements are likely to have classified themselves as happy in response to an overall happiness question. Conversely, those who strongly disagree are likely to have classified themselves as unhappy.²