Suicide rates are on the rise in the twenty-first century, and New York Times columnist Ross Douthat and I have been exchanging thoughts on exactly why. On Sunday, Douthat attributed the rise to cultural factors such as “the retreat from family obligations" and "from civic and religious participation.” I counter-rgued that although many of these trends are troubling, the decline of traditional institutions, and particularly marriage and religion, isn’t clearly responsible for increasing suicide rates. Douthat replied that my piece “did not prove that social belonging doesn’t play a role in suicide rates.” He’s right, but I wasn't taking issue with his entire argument: I never questioned the relationship between suicide and belongingness, just whether traditional institutions, like marriage and religion, guard against suicide. The data suggests otherwise—that long-term economic trends have the greatest influence on national suicide rates.
Douthat makes the assumption that relatively religious or traditional communities have lower suicide rates. But suicide rates aren't much higher than they were in the mid-20th century, despite imbalanced economic growth and the decline of the traditional family and of religious affiliation. The state-level data also complicates that convenient cultural explanation. Suicide rates are highest in the West, where marriage rates are quite high, and depression is highest in the Bible Belt. That doesn’t mean that religion or marriage don't, or can’t, play a role in reducing suicide. Who knows: Maybe suicide rates would be even higher in the Deep South if religion weren’t comforting a poor, unhealthy, and depressed population. Perhaps New York would be suicide-free without rampant secularism.
Part of why this discussion is difficult is the danger of an ecological fallacy, where group data can mislead about individual behavior. Douthat, for instance, cited a study on religion and suicide that found that active members of the Church of Jesus Christ of Latter-day Saints who live in Utah are far less likely to commit suicide than non-members or inactive members. But, paradoxically, Utah had the tenth-highest suicide rate in 2010, closer to the sinful state of Nevada's than the national average. If, in fact, religious people are less likely to commit suicide, then why would Utah have both high levels of religious participation and a high suicide rate? Perhaps the explanation relates to Utah's low population density or high gun ownership, which correlate with higher suicide rates. If Utah’s gun ownership rate was as low as New York’s, maybe the suicide rate would be extremely low. Or maybe religious communities are alienating to non-members: Non-Mormons in Utah have a very high suicide rate, even though their LDS counterparts are slightly less likely to commit suicide than the national average.
Regardless of the explanation, it is hardly clear that traditional communities with nuclear families and high church attendance reduce the risk of suicide. That’s relevant to the question raised by Douthat’s piece: whether the decline of traditional communities and social relationships ensures that suicide rates will continue to increase. Moreover, if we’re interested in what types of communities—and what type of country—reduces suicide, then the geography is quite relevant, perhaps even more relevant than recent trends. After all, there’s a 30 percent difference between suicide rates among middle-aged men today and a decade ago, but there’s a 300 percent difference in suicide rates between Wyoming and New York.
Now consider the county-level data on suicide between 2002 and 2006, which shows, for instance, that minorities don’t commit suicide at anything near the same rate as whites:
Notice the arch of light tones across the lowland South, where African Americans represent a large share of the population, or along the Rio Grande and across southern California, where Hispanics represent a large share of the population. Whites commit suicide at a high rate in most rural areas, but there are a few exceptions, like rural Yankeedom. Suicide rates are highest in the interior West—even in urban areas like Denver, Salt Lake City, or Phoenix. Major eastern metropolitan areas show far lower suicide rates. If America is facing a “suicide epidemic,” as Tony Dokoupil claimed this week in Newsweek, then the West is facing a severe crisis, and we should probably figure out why if we're interested in determining what causes suicide.
But talk of a “suicide epidemic" borders on sensationalism. Dokoupil’s piece is generally sound, but he contends that we live in a dystopic era of “unprecedented despair,” and that “we have never been more burdened by sadness.” Suicide rates have, indeed, climbed since 1999, a year that represented the all-time low for suicide in the United States, but today’s suicide rates are roughly comparable to those between 1940 and 1990, and below the rates of the 1930s.
That trendline corresponds with long-term trends in the national economy, adding credibility to the economic explanation for the suicide surge. During the Great Depression, suicide rates were higher than they are today, and by the end of the depression had tumbled to a rate slightly lower than today's levels, where they more or less held until economic growth slowed in the late 1960s. Then, suicide rates increased to a peak—higher than today's rate—in the late-'70s before steadily declining to an all-time low in 1999, at the height of the tech boom. Over the last decade of disappointing growth, suicide rates have ticked back up. A study on the relationship between long-term unemployment and suicide rates offers additional evidence for the economic explanation.
That said, Dokoupil’s piece highlights evidence consistent with the cultural explanation. Julie Phillips of Rutgers University found that post-Baby Boomer cohorts have higher suicide rates.1 If cultural shifts since 1960 are, in fact, driving up suicide rates, then we would expect suicide rates to keep climbing, as Douthat fears—especially once Boomers reach old age. But although this is an “important trend to watch,” as Phillips put it to me in an email, “it’s also important not to be too alarmist.” Post-Boomer cohorts appear only modestly more likely to commit suicide (it’s tough to detect in Phillips' charts).
After considering these and other data over the last few days, I lean toward the view that long-term economic health is the largest factor influencing changes in the national suicide rate. Phillips' analysis suggests that generational change may prove to be quite important, too. But it's far less clear what type of community or society is most likely to have lower suicide rates, even if the critical variable is not the prevalence of traditional institutions. After all, the suicide rate is within its historical bounds, despite an unprecedented decline in marriage and religion.
Unfortunately for conservatives, the community most likely to have lower suicide rates appears to be one without guns. Non-firearm suicide does not correlate with gun ownership, but firearm suicide does. It may also help explain lower suicide rates in minority communities with high levels of economic distress, but lower levels of gun ownership. The possibility that gun ownership influences suicide rates suggests that we should focus on depression or other maladies to judge societal dysfunction, not suicide.
After controlling for age and time effects.