Nearly a century after James Truslow Adams coined the phrase, the “American dream” has become a staple of presidential campaign speeches. Kicking off her 2016 campaign, Hillary Clinton told supporters that “we need to do a better job of getting our economy growing again and producing results and renewing the American dream.” Marco Rubio lamented that “too many Americans are starting to doubt” that it is still possible to achieve the American dream, and Ted Cruz asked his supporters to “imagine a legal immigration system that welcomes and celebrates those who come to achieve the American dream.” Donald Trump claimed that “the American dream is dead” and Bernie Sanders quipped that for many “the American dream has become a nightmare.”

But the American dream is not just a pie-in-the-sky notion—it’s a scientifically testable proposition. The American dream, Adams wrote, “is not a dream of motor cars and high wages merely, but a dream of social order in which each man and each woman shall be able to attain to the fullest stature of which they are innately capable…regardless of the fortuitous circumstances of birth or position.” In the parlance of behavioral genetics—the scientific study of genetic influences on individual differences in behavior—Adams’ idea was that all Americans should have an equal opportunity to realize their genetic potential.    

A study just published in Psychological Science by psychologists Elliot Tucker-Drob and Timothy Bates reveals that this version of the American dream is in serious trouble. Tucker-Drob and Bates set out to evaluate evidence for the influence of genetic factors on IQ-type measures (aptitude and achievement) that predict success in school, work, and everyday life. Their specific question was how the contribution of genes to these measures would compare at low versus high levels of socioeconomic status (or SES), and whether the results would differ across countries. The results reveal, ironically, that the American dream is more of a reality for other countries than it is for America: genetic influences on IQ were uniform across levels of SES in Western Europe and Australia, but, in the United States, were much higher for the rich than for the poor.

For their analysis, Tucker-Drob and Bates identified 14 studies, with a total sample of nearly 25,000 pairs of twins. Each study compared identical twins to fraternal twins on an IQ measure, and also collected information on SES, including income, wealth, and education. Identical twins share 100% of their genes, whereas fraternal twins share, on average, only 50 percent of their genes. Thus, if variation across people in a trait is influenced by genes, identical twins will be more similar to each other on that trait than fraternal twins will be. Degree of genetic influence—or heritability—is determined by statistically comparing the identical twins and the fraternal twins. Tucker-Drob and Bates then aggregated the findings of the studies using meta-analysis, a statistical tool that quantitatively synthesizes the results of multiple studies.

The results were striking. For studies conducted in the U.S., the heritability of IQ was 61% at the 95th percentile for SES, compared to just 24% at the 5th percentile. (In more technical terms, there was a Gene x SES interaction, meaning that the genetic contribution to IQ varied as a function of SES.) By contrast, for studies conducted in Western Europe or Australia, heritability was statistically no different across levels of SES. IQ was as heritable for the poor as for the rich.

What might explain this finding? The new study doesn’t pinpoint the cause, but the leading hypothesis is that social policies in countries like Sweden, Australia, and Germany create living conditions that facilitate genetic influences on intellectual functioning. In these countries, people have relatively equal access to high-quality education and healthcare, and childhood poverty rates are low. Much like fertile soil allows plants to reach their maximum height, these conditions are hypothesized to promote the expression of any genetic differences in IQ that exist. By contrast, in the U.S., an estimated 33 million people still do not have health insurance, and there are dramatic differences across school districts in quality of education.

The results of Tucker-Drob and Bates’ analysis are as important as they are sobering. As economists have documented, the gap between the rich and poor in the United States is vast and widening. As a consequence, large numbers of Americans cannot afford even the basic necessities of life. Until they can, the American dream will likely remain a dream that these Americans have little hope of fulfilling.