A large and vibrant middle class is important to every nation. It contributes to economic growth, as well as to social and political stability.
The middle class helps mitigate class warfare. Marx (1948) believed that economic history was a class struggle between haves and have-nots, and that the have-nots would eventually band together and overthrow the capitalist system. What Marx missed was that a middle class might arise and serve as a buffer between the poor and the wealthy.
In addition, a middle class helps democracy flourish. This view goes back to Aristotle (1932, Book IV), who noted that political communities with a large middle class would likely be well administered and would not be dominated by either of the income extremes. More recently, Thurow (1984) argued that "a healthy middle class is necessary to have a healthy democracy" because social unrest usually increases when incomes and people become polarized. Barro (1999) has provided empirical support for this view--countries are more likely to be democratic the higher the share of income going to middle-class families.
Attaining a middle-class living standard also carries with it feelings of success and personal accomplishment. Malthus (1803, 594) noted in his second Essay on Population, "Our best grounded expectations of an increase in the happiness of the mass of human society are founded in the prospect of an increase in the relative proportions of the middle parts." Much recent literature documents the fact that people do care about their relative standing and that relative standing is correlated with subjective assessments of well-being (Frank 1999; Layard 2005).
Finally, a healthy middle class may be necessary for good macroeconomic performance. Henry George (1931, 510ff.) noted that sharp divisions in economic status place an obstacle in the face of development. Separation reinforces differences and prejudices; animosities thus grow, and economies are likely to decline rather than grow. Income inequalities may also reduce consumption, and therefore effective demand and economic growth. For Keynesian or demand reasons, we may get both fewer jobs and lower wages, thus leading to a smaller middle class (Brown 2004; Pressman 1997).
Because the middle class is important in so many ways, its decline must be a matter of concern. In the United States, during the 1980s many voices expressed such concern. Kuttner (1983) and Steinberg (1983) first drew attention to the fact that the traditional American middle-class family was disappearing, then economic think tanks (Lawrence 1984) and Congress (Belous, LeGrande and Cashell 1985) began to study this issue.
However, in the 1990s, as growth rates increased and unemployment rates declined in the United States, concern about the fate of the middle class waned. This was especially true during the late 1990s, when a plentiful supply of jobs and rising real wages seemed to make middle-class lifestyles accessible to more families.
Did the 1990s economic boom halt or reverse a middle class decline in the United States? Is a shrinking middle class strictly a U.S. phenomenon with domestic causes, or is it a problem that has plagued most of the world economy? If the middle class shrank in the United States and in other countries, what caused this to occur?
This paper explores these issues. The second section describes previous attempts at defining the middle class. Several possible causes of the declining middle class are set forth in the third section--changing demographic factors, structural economic changes, macroeconomic conditions, and a more Keynesian explanation, which depends on the importance of government spending. The fourth section measures the size of the middle class for a number of countries using the Luxembourg Income Study (LIS). Using the LIS, the fifth section empirically examines some of the explanations for a shrinking middle class. Finally, we conclude with some policy implications.
Who Is Middle Class?
One key issue in the debate over the shrinking middle class is how to measure it. Because theory does not and cannot tell us who counts as middle class (unless we arbitrarily say the middle class is always the middle third of the income distribution, in which case the middle class can never decline because it is, by definition, always the middle one-third), any definition we choose is going to be arbitrary. But this arbitrariness does not mean that we cannot and should not attempt to define and measure the middle class.
Frank Levy and Richard Michel (1983, 205f.) note that "middle class" has three meanings. One meaning is more sociological than economic, and concerns attitudes and behaviors--having a good education, a stable career, and resorting to reasoning rather than violence to settle arguments. Taking this approach, sociologists generally try to measure the middle class by a set of behavioral characteristics and socio-economic criteria. The middle class on this view are people who have achieved a certain educational level, whose jobs have a certain level of social status, and who have a particular set of values and attitudes (Coleman and Rainwater 1978).
A second meaning of middle class is an economic one. By this definition, being middle class means having a middle-class standard of living or having an income level that is somewhere in the middle of the income distribution. Virtually all empirical studies in economics have taken this route and have examined either the percentage of income going to the middle income quintiles or middle 60 percent of the population (Levy 1988), or some income space around the median level of income (Thurow 1985; 1987; Blackburn and Bloom 1985; Horrigan and Haugen 1988; Davis and Huston 1992). For example, Thurow (1985) defines the middle class as any household with 75 to 125 percent of median household income, a definition used in a large number of empirical studies and employed in the empirical work below.
A third approach is to ask people themselves what constitutes a middle-class income. One problem with this approach is that most people think of themselves as middle class, even if they are nearly poor or quite well-to-do. Surveys in the early 1980s found that Americans classify families with incomes of between $15,000 and $100,000 as being middle class (Rose 1983, 38-9). These incomes are roughly equal to $30,000 and $200,000 today, and at the time of these studies would have put households into the middle class if their incomes fell between the 45th and 99th income percentiles.
Possible Causes of the Declining Middle Class
Despite differing definitions of the middle class, different units of analysis (family income versus household income), and different methods of analysis, a majority of previous studies have found that the middle class did decline in the United States during the late 1970s and early 1980s (see Belous, LeGrande and Cashell 1985 for a summary).
In general, four factors have been advanced as explanations for the declining middle class: (1) demographic factors, (2) structural or microeconomic factors such as the loss of middle-class manufacturing jobs and the decline of labor unions, (3) macroeconomic factors such as unemployment resulting from the business cycle, and (4) changes in public policy.
A number of studies have cited demographic factors as the main culprit in the declining percentage of middle class families in the United States (Bradbury 1986). Unfortunately, there has been little agreement over the important demographic factors at work, and even less agreement that demographic changes are in fact responsible for shrinking the middle class.
Rising divorce rates is one frequently cited demographic culprit (Thurow 1984). When couples divorce, the economic condition of men is usually unaffected while the economic condition of women frequently deteriorates. Since women earn considerably less than men do (for a variety of reasons), and since they are likely to have to support themselves and their children, it is much more difficult for them to achieve middle-class status (Belous, LeGrande and Cashell 1985). Following divorce, the married-couple, middle-class family frequently becomes a middle-class household headed by a man and a lower-class household headed by a woman.
A second possible demographic cause of middle-class decline is the changing age structure of the population. It is well known that incomes rise with age and experience. This follows directly from human capital theory (Becker 1993; Mincer 1974). Lacking experience, young workers tend to have lower earnings. They also have greater wage dispersion (Mincer 1974). However, as people age, their incomes rise and income distribution flattens for each age group. Lawrence (1984) and Levy (1988, 94-9) argue that middle-class decline resulted from demographic forces such as the influx of baby boomers into the labor market. In the 1970s and 1980s, young and inexperienced workers came to form a relatively large portion of household heads because of the baby boom, and these households did not earn enough to make it into the middle class.
The growth of two-earner families may also have contributed to the declining middle class (Blackburn and Bloom 1985; Lerman 1996; Rosenthal 1985). Two individuals with middle-class earnings, who pool their income as a family, are likely to have an upper-class standard of living--the yuppie couples of the 1980s and 1990s. This will reduce the size of the middle class while the upper class expands in size.
Changes in the earnings distribution of women may have made this even worse. In the 1960s and early 1970s, the earnings of women were distributed rather equally and this seemed to increase the equality of family earnings (Danziger 1980). But women's earnings have become more unequal since the late 1970s, and there has been a greater tendency for high-income women to marry high-income men (Burtless 1996).
Other studies of the...