Myth: Income inequality is not the cause of this nation's social problems.

Fact: Studies by Harvard and Berkeley prove the correlation, and strongly suggest causation.


Poverty is correlated to higher mortality rates for all ages. Harvard and Berkeley conducted two studies that also found that income inequality, as opposed to absolute standard of living, is responsible for higher death rates, as well as numerous other social problems like crime, welfare, and poor educational outcomes. Conservatives point out that correlation is not causation; it could be that people suffering from these health and social problems naturally earn lower incomes. But economic fluctuations and income inequality happen too rapidly, drastically and locally to be attributed to personality changes, which often occur in the same populations. It is much more reasonable to attribute these economic changes to changes in economic policy.


Health researchers have long known that the poor have higher death rates and greater health problems than the rich. (1) Countless studies have proven this; the following one yields a typical result. In 1986, researchers studied two groups of men between the ages of 25 and 64: those that made less than $9,000 a year, and those that made more than $25,000. They found that poor white men had 6.7 times the death rate of rich white men, and poor black men had 5.4 times the death rate of rich black men. (2) One can also see this correlation in the improving living standards and life expectancies all around the world in the last 200 years. Better science, technology, safety, public education and prosperity have made us all healthier.

But why, even in the same society, would health and life expectancy be linked to income class? Perhaps the most obvious answer is that the poor cannot afford the same health care, especially preventative health care. But in fact there are hundreds of reasons why the poor have higher rates of death, disease and injury. The poor live and work in more toxic environments; they have less adequate diets; they are exposed to greater dangers and risks (both human and non-human); they cannot afford the safety features or creature comforts that make living safer or easier; they suffer higher and more negative forms of stress in trying to make ends meet (or even survive); and they have less access to education about things that would prolong their lives.

That, at least, is the mainstream view -- many on the far right have a much different interpretation. Correlation is not causation, they point out. Poverty does not cause poor health; rather, poor health might cause poverty. That is, people in poor health would probably tend to be society's least productive, and therefore least-paid, workers. Or yet a third factor could cause both poor health and poverty. For example, the same lack of education that ill-informs the poor on health matters also locks them out of the best paying jobs. Or laziness, substance abuse and other moral shortcomings could have a depressing effect on both incomes and health.

Which view is correct? Again, consider the simultaneous rise in living standards and life expectancies all around the world in the last 200 years. It would be rather bizarre to attribute this to a sudden change in individual human morals, habits and work ethics. Why would this occur simultaneously and coincidentally in mostly the 10 billion people of the last two centuries? This sudden upward trend virtually demands a social explanation; it is not the result of isolated individual behavior. Intrinsic human nature has not changed for countless thousands of years. What has changed -- and changed dramatically, especially in the last two centuries -- is society and its institutions, ranging from scientific to economic to political systems. Thus, it is far more reasonable to attribute health and economic advances to social causes and social programs, not individual personality changes.

To better determine the arrow of causality, let's take a closer look at how income inequality is linked to health and other social problems.

The Harvard and Berkeley studies of income inequality

If it is true that poverty kills, then this should place income inequality in a completely different light. A society that allows high levels of inequality could be accused of killing its own citizens. Conservatives defend against this charge by pointing out that the living standards of the poor have been continually rising, so it shouldn't matter if the rich are growing even richer by comparison. After all, this is not a zero-sum economy. Even a poor person with a small slice of the pie will benefit if the entire pie grows. This introduces two important concepts to this debate: relative poverty (or the slice of the pie, as measured in percent) and absolute poverty (or the actual size of the slice itself, compared to nothing else).

However, it turns out that relative poverty matters a great deal after all.

In 1996, Harvard and Berkeley published separate studies that examined income inequality in all 50 states. (3) According to Bruce Kennedy, the lead researcher of the Harvard study, "The size of the gap between the wealthy and less well-off, as distinct from the absolute standard of living enjoyed by the poor, appears to be related to mortality." (4) Both studies found that states with higher income inequality have all the following social problems:

The correlation between income inequality and mortality rates for all ages was significant. (Berkeley found a correlation of 0.62, with P -- the chance that the correlation could be zero accidentally -- being less than 0.001; Harvard found a correlation of 0.54, with P less than 0.05.) (5)

Both studies found that each state's average or median income did not predict its mortality rate. But inequality turned out to be a significant predictor, and remained so even after accounting for such possible confounding factors as smoking and drinking rates, household size and household income. The last one is especially important, because it means that the death rate is correlated not only to absolute poverty, but relative poverty as well.

Dr. George Kaplan, the lead researcher of the Berkeley study, says: "People might assume that states with higher income inequality have more poor people, and we know that poor people have higher death rates. [But] the evidence in these two studies suggests that the increased death rates in those states are not due simply to their having more poor people. Income inequality seems to be increasing mortality rates among nonpoor people as well, and we are investigating that possibility." (6)

Elsewhere, Kaplan says: "This effect on health wasn't just happening to poor people; middle-class people were affected too. When we accounted for income differences, there was still a strong relationship between income inequality and mortality rates." (7)

A hypothetical example might best describe these findings. Imagine an equal society where everyone makes $30,000 a year, and enjoys a life expectancy of 70 years. Now let's add income inequality to this society, so that it continues to make as much collectively as it did before, but a third now make $20,000 a year, a third $30,000, and a third $40,000. We might expect the average life span to remain at 70 years, since the shorter life spans of the poorest group should be offset by the longer life spans of the richest group. But that is not the case; not only do the poor lose years, but the middle class as well. The new average life span of the whole group may be only 67, not 70. We'll explore some possible explanations for this unexpected result below.

Dr. Kaplan and his colleagues reported that "income inequality increased in all states except Alaska from 1980 to 1990." Over the decade, they said, mortality declined in all states (due to increases in the absolute standard of living), but those with greater income inequality showed smaller declines in mortality. (8) If this lends comfort to those who still believe in the "growing pie" argument, it should be pointed out that inequality still costs hundreds of thousands of lives per year. Let's consider just one of the many problems in the above list: coronary heart disease. The Harvard team concluded that if the U.S. reduced its Robin Hood index of income inequality from 30 to 25 percent (about where it is in England), deaths from coronary heart disease would be reduced by 25 percent. In 1993, the U.S. suffered 489,970 deaths from coronary heart disease; a quarter of that would have represented 122,493 lives saved for that year alone. (9)

The European evidence

The results of the Harvard and Berkeley studies cohere very nicely with the international evidence. The U.S. has the most unequal society of all rich nations, and by far. These other nations have higher taxes and better-funded social programs to alleviate poverty, and their progressive social policies have resulted in more equal societies. So who has the worst health statistics, mortality rates, crime rates, sedentary lifestyles, economic growth and other social problems? The U.S., usually by far. (10)

Some conservatives try to nullify this observation by playing the race card, by pointing out that blacks drag down the U.S. statistics. However, the U.S. still compares worse to other rich nations even when blacks are removed from the statistics completely.

What are the causal factors?

Again, correlation is not causation. Which causes which is a matter of debate. Liberals claim that income inequality causes higher health and social problems. Conservatives claim that people who suffer from these social problems naturally earn lower incomes. Yet another possibility is that both form a vicious circle. Let's examine each of these possibilities.

Liberals claim that in an interdependent, interlocking society, the weakening of one sector weakens us all. That would explain why states with greater inequality have greater health and social problems not only among the poor, but the middle class as well. Dr. Kaplan suggests that "income inequality affects all segments of the population because it affects rates of violence and disability, as well as public spending on police protection, education, welfare and health care." (11)

Also, it's likely that average statistics drop in states with greater inequality because losses among the poor are not fully compensated by gains among the rich. For example, there might be a point beyond which having more money does not improve one's health. If a person already has a full, comprehensive and top-quality health care plan, then acquiring still more money will not improve it. It will, however, deprive someone from the lower class of the health care they would have received otherwise.

The fact that higher death rates creep into the middle class tends to run counter to the conservative claim that social and health problems result in lower incomes. If they did, we would not see these problems growing among those with middle-class incomes.

Furthermore, the historical evidence does not support the suggestion that low incomes are the result of people suffering from health problems or personalities which are criminal, addictive, lazy or hostile to education. The main problem is that fluctuations in income inequality are too rapid, too drastic and too localized to be attributed to character changes and social habits in people, especially when they are the same people. For example, between 1980 and 1990, the Gini index of income inequality rose from .365 to .401. (12) For those familiar with the Gini index, this is a major jump. Is it really reasonable to believe that the personalities and morals of the poor worsened considerably in just one decade? The absurdity of this belief becomes all the more apparent when you consider the economic changes that occurred to the rich. During the 80s, the share of national income collected by the top 1 percent increased from 8 to 12-13 percent, an increase of over 50 percent. (13) Are we to believe that in just one decade, the personal and individual productivity, intelligence, education, health, virtue and work ethic of the top 1 percent grew over 50 percent?

It is much more parsimonious to attribute such rapid economic changes to the rapidly changing economic policies of the time. During the 80s, Reagan slashed the top tax rate for personal income from 70 to 28 percent, allowing income to concentrate among the wealthy. And in 1983, regressive payroll taxes were raised on the working poor. Furthermore, between 1980 and 1993, family welfare payments were reduced from $350 to $261 per month in constant dollars -- a 25 percent drop. (14) One doesn't have to look far to find the reasons for growing income inequality during the 80s.

It might be possible that income inequality and social problems form a vicious circle. No one wants to hire an uneducated person for a high-paying, high-skilled job. But the lack of a high income is what prevented that person from acquiring a college education in the first place. Still, there is nothing about this circle that is inevitable; it can be broken through social programs, like student loans and grants.

The argument that income inequality is largely caused by social policy, not individual merit, is one that has been endlessly developed in an abundant academic literature. And the implications should be troubling to every American: our society bears great responsibility for both its deplorable level of income inequality and poverty-related mortality.

Return to Overview


1. George Davey Smith and others, "Socioeconomic Differentials in Mortality Risk among Men Screened for the Multiple Risk Factor Intervention Trial: I. White Men," American Journal of Public Health Vol. 86, No. 4 (April, 1996), pp. 486-496; George Davey Smith and others, "Socioeconomic Differentials in Mortality Risk among Men Screened for the Multiple Risk Factor Intervention Trial: II. Black Men," American Journal of Public Health Vol. 86, No. 4 (April, 1996), pp. 497-504; Gopal K. Singh and Stella M. Yu, "US Childhood Mortality, 1950 through 1993: Trends and Socioeconomic Differentials," American Journal of Public Health Vol. 86, No. 4 (April, 1996), pp. 505-512; C. Wayne Sells and Robert Wm. Blum, "Morbidity and Mortality among US Adolescents: An Overview of Data and Trends," American Journal of Public Health Vol. 86, No. 4 (April, 1996), pp. 513-519.

2. Robert Pear, "Big Health Gap, Tied to Income, Is Found in U.S." The New York Times, July 8, 1993, pp. A1.

3. George A. Kaplan and others, "Inequality in income and mortality in the United States: analysis of mortality and potential pathways," British Medical Journal Vol. 312 (April 20, 1996), pp. 999-1003. Bruce P. Kennedy and others, "Income distribution and mortality: cross sectional ecological study of the Robin Hood index in the United States," British Medical Journal Vol. 312 (April 20, 1996), pp. 1004-1007.

4. Quoted in Robert Pear, "Researchers Link Income Inequality to Higher Mortality Rates," New York Times, Friday, April 19, 1996.

5. British Medical Journal.

6. Quoted in Pear, "Researchers Link Income Inequality to Higher Mortality Rates."

7. Quoted in Alison Bass, "Income inequality, mortality linked; Gap found to hurt wide segment in US," The Boston Globe, April 19, 1996, Friday, City Edition, p. 14.

8. Pear, "Researchers Link Income Inequality to Higher Mortality Rates."

9. American Heart Association, "Heart and Stroke Facts, 1996 Statistical Supplement."

10. For a comprehensive comparison of the U.S. to other rich nations on a wide range of issues, see 8Comparison.htm.

11. Quoted in Pear, "Researchers Link Income Inequality to Higher Mortality Rates."

12. U.S. Bureau of the Census, Current Population Reports, Series P60.

13. Kevin Phillips, Boiling Point (New York: HarperCollins, 1993), p. 112.

14. AFDC figures from U.S. Social Security Administration. Current dollars converted to constant 82-84 dollars from CPI-U.