Across various measures of social network participation and general trust, the Midlands, Yankeedom, and Left Coast top the charts, El Norte and Spanish Caribbean are at the rear.
By Colin Woodard
In a 1995 article, “Bowling Alone,” Harvard University political scientst Robert Putnam warned that American’s stock of social capital – the fabric of a community’s trust and cooperation – had been rapidly plummeting since the 1950s, harming our well-being, personal relationships, health, lifespan, and economic growth. In the three decades since, social trust, norms, and community cohesion have continued to come undone, paving the way for the rise of authoritarian populism and the undoing of the republic. The number of people saying they have trust in the American people to make judgements under our democratic system, for example, stands at just 53 percent, down from 83 percent in April 1974, when Gallup first started asking the question, shortly after our chaotic withdrawal from Vietnam and right in the midst of the Watergate scandal
The decrease in trust, which Putnam had documented via nearly half a million interviews, has had all sorts of effects. It hurts economic activity because, as Nobel Prize winning economist Kenneth Arrow observed, commercial transactions rely on it. (“It can be plausibly argued,” he once said, “that much of the economic backwardness in the world can be explained by the lack of mutual confidence.”) It undermines social mobility, as Harvard’s Raj Chetty and his colleagues have shown. And it creates the social alienation that, in the words of the Brooking Institution’s Isabel Sawhill, can lead to “school shootings, opioid addiction, police brutality, and racial strife.” As Putnam found in his field work in Italy in the 1970s, it has a profound effect on the functioning of democratic institutions, and when its present it can build on itself, and when absent it is very difficult to rebuild.
“In the established democracies, ironically,” Putnam wrote in his landmark article, “growing numbers of citizens are questioning the effectiveness of their public institutions at the very moment when liberal democracy has swept the battlefield, both ideologically and geopolitically.”
Here at Nationhood Lab, we’ve long been interested if there are regional differences in social capital and if they mirror those we see in life expectancy, gun violence, various health problems, social vulnerability, persistent poverty, intergenerational economic mobility, the prevenance of authoritarian personalities, and support for Trump and Trumpism. This winter, we examined the data using several approaches to gauging social capital at the county-level, the unit of analysis required by my American Nations model of North American regional cultures. The results were stark and in some ways were just as we’d anticipated, in others quite surprising.
After examining several county-level Social Capital Indexes created by other researchers – including one developed by the American Enterprise Institute’s Scott Winship and another from sociologist Dean Kyne and political scientist Daniel P. Aldrich – we chose the one created by economists Anil Rupasingha of the American University of Sharjah, Stephen J. Goetz of Pennsylvania State University, and the University of Kentucky’s David Freshwater. It calculates a county’s social capital based on the density of ten types of associational institutions in the county – including bowling leagues, churches, sports clubs, civic organizations, and labor unions – plus turnout in previous presidential elections, the census return rate, and the number of non-profit organizations. We then used the data produced for this model by University of Texas Rio Grande Valley political scientists Dongkyu Kim, Mi Son Kim, and Natasha Altena McNeely for the year 2014, the most recent for which they had all of the required metrics.
Here are the county-level and American Nations results. (For clarity, you can also see just the “national” results in the map at the top of this article.)
The county-level affinity to the American Nations regional borders isn’t as stark as we see in a lot of our data, in part because, as the UT-Rio Grande Valley researchers showed in a 2020 paper in the Social Sciences Journal, social capital measured in this way is negatively impacted by urbanization and, depressingly, racial diversity, which tend to ding metropolitan areas, regardless of region. But there’s still a clear county-level regional pattern, with the Midlands scoring well across its entire range, west of central Pennsylvania, Yankeedom strong in much of northern New England, the Western Reserve of Ohio, Wisconsin and Minnesota (but not in most of Michigan or Upstate New York) and El Norte performing badly almost everywhere save the Big Bend counties of South Texas. Far West is heterogenous, with strong scores in the upper Plains and northern mountain West, and quite poor scores in Utah (surprisingly), Nevada and northern Arizona.
The regional pattern – which calculates the score for the entire region by weighting each county score by its population — is a lot clearer, as few people live in the (often higher scoring) rural counties. In this model, three of our “aggressively communitarian” regions lead the pack. The Midlands has the highest reserves of social capital with an overall index score of -0.16 on their scale, followed by the Left Coast at -0.2 and Yankeedom at -0.26. The two regions founded by imperial Spain, El Norte and Spanish Caribbean, had very poor scores, -1.23 and 1.-01 respectively, as did New Netherland (-1.07) and First Nation, which was at the bottom of the list at -1.47. This is very similar to the spatial pattern we found for persistent poverty, which substantiates the Texas researcher’s finding that wealthy places tend to have more social capital.
Three southern regions that often have among the worst metrics in whatever phenomenon we analyze — Greater Appalachia, Deep South and New France – are in the middle of the pack for social capital, as is the Far West. Tidewater, a region that’s rapidly transformed from an apartheid society to one of the wealthiest and most progressive of the “nations” over the past 60 years, had the fourth best score, just behind Yankeedom at -0.34. Hawaii, despite being aggressively communitarian, fared worse that the southern regions and Far West, with a score of -0.99.
Looking at the results, we were curious if there were differences in social capital between rural counties, given the county map suggested rural places in the southern regions fared badly, while the Yankee and Midlands ones did well. Here are the results when only rural counties are considered, with rural defined as the category five and six counties in the National Center for Health Statistics ’s six-tiered classification scheme for 2013.
As you can see, the rural pattern mirrors the overall regional results, with Left Coast, Yankeedom and the Midlands having almost identical scores (+0.38 or .37) and El Norte and First Nation at the bottom. The biggest shift is the Far West, which has the fourth best social capital index for rural counties (+0.22), well ahead of the three southern regions, Spanish Caribbean, and Greater Polynesia.
We also compared the results for just the major core metropolitan counties – NCHS’s category 1 – defined as the core county of a metropolitan area of at least one million people or one of its neighbors if it has a population of at least 250,000 — a class of counties that presumably performs poorly under the index’s metrics. The basic regional pattern held again, though this time Tidewater jumped to the second-best position, probably on account of its biggest Cat 1 counties being in super affluent, highly educated, white collar northern Virginia and southern Maryland. The inner cities of the Left Coast – home to Seattle, Portland, and San Francisco – performed even better, with a score of -0.16, which is actually slightly better than its rural score and matches the Midlands’ overall index score. Social capital seems in ample supply in those West Coast cities.
There are many ways to define social capital, however. For instance, the index AEI uses – created by economic committee staffers serving under Senator Mike Lee of Utah – gives substantial weight to metrics of “family unity,” which incorporates the share of out-of-wedlock childbirths, older women who are unmarried, and the share of children in single-parent families, plus metrics that arguably are effects of low social capital, not a measure of its existence, like the violent crime rate and the share of people reporting confidence in corporations, the media and public schools. (Under this system, Sen. Lee’s home state of Utah jumped to the best ranking in the country, rather than one of the laggards, as in the Rupasingha et al index used above.)
Purists, following on Alexis de Tocqueville’s lead, might look to civic associations – be they churches, bowling alleys, environmental groups or fraternal organizations — as the core production sites of social capital. One way to get at this is what scholars call associational density, the number of relevant institutions per capita in a given area. The UT Rio Grande Valley researchers collected this information at a county-level for 2014, and we used it to calculate the density of all types of associations in each American Nations region overall. Here are the results:
You can see it generally mirrors the overall Social Capital Index scores we introduced at the beginning of the post, which isn’t surprising given this data is a subset of that the index uses. There are a couple of exceptions. Notice Left Coast actually doesn’t do very well by this metric, meaning its strengths lie in other measures, like voting and turning in one’s census forms. Greater Appalachia, meanwhile, surges to second place, slightly ahead of Yankeedom, and New France makes the top five. There are a lot of associations in these aggressively individualistic regions.
Knowing there’s a regional maldistribution of churches and religious attendance, we repeated this exercise, but excluded church groups from the analysis. This could, arguably, be relevant because churches are one type of association that can foster bonds between its members but not with others, what Putnam calls bonding capital rather than bridging capital. (This can be true of secular groups too, of course, like professional associations or social clubs, but the data doesn’t make for easy distinctions.) Here’s what we found for “secular only” associational density:
With churches excluded, Greater Appalachia falls out of the top tier, with scores resembling those of the Far West rather than Yankeedom or the Midlands. (Yankeedom, unsurprisingly, tops the list.) The New France enclave in southern Louisiana holds is own, though, while First Nation – which has the worst overall SCI score – surges to second place, topping even the Midlands. This suggests the Rupasingha model may have a hard time evaluating this indigenous cultural region, whose low social capital score is primarily driven by resident’s relative disinterest in participating in federal elections and census project. (Many county-level equivalent units in these parts of Alaska had census and voting percentages in the thirties and forties.)
If you’re electorally focused, maybe you’re curious about the density of just political organizations. Left Coast is, far and away, the richest region in this regard (with an index score of 0.16), overshadowing even D.C.-adjacent Tidewater (0.11), as well as the Far West (0.11), Yankeedom (0.1), and Midlands (0.09). The least “politically organized” regions are Deep South (0.054), Greater Polynesia (that’s Hawaii, again, 0.05), Spanish Caribbean (0.4) and First Nation, which the data set found had no political orgs at all.
And, finally, we checked if there were regional differences in the number of “non-movers,” people who reported they’d lived in their community for five years or more and, thus, had time to put down roots and build networks, again using data collected by the UT Rio Grade Valley team for 2009 and 2014. The Midlands and Greater Appalachia are the most rooted regions by this metric (with index scores of 1.19 and 1.18), followed by Tidewater, Yankeedom and the Deep South (all between 0.95 and 1.04.) The least rooted are the two Spanish legacy regions, El Norte (0.57) and Spanish Caribbean (0.6), with Greater Polynesia (0.65) and First Nation – which has a lot of non-indigenous transient oil and fisheries workers – at 0.69.
What does this all mean? To summarize, across the various approaches and indices, three out of the four big communitarian regions — Midlands, Yankeedom and Left Coast – have the strongest social capital indices, whether looking at the cores of big cities or rural counties, associations or overall metrics. El Norte and Spanish Caribbean – very different regions that share a Spanish imperial legacy – have uniformly weak social capital, likely exacerbated by both the high proportions of newcomers and the relative lack of associations to foster cooperation.
The big, aggressively individualistic regions – Deep South and Greater Appalachia — present a mixed picture, with social capital scores buoyed by very high numbers of religious associations and decent numbers of other types of social capital-generating organizations. Greater Appalachia, for instance, has very high associational density driven by a huge number of churches, but more middling scores in other respects. Far West, the only large “passively individualistic” region, scores a bit worse, with fewer churches and very low social ties in its city cores, where an unusually large proportion of the region’s population lives.
Communitarian New Netherland, the densest and most diverse region on the continent, performs badly all the way around, as does very low-density First Nation, whose results are suggestive of a closely-bonded indigenous society distrustful and disengaged from wider American society, federal politics, and possibly the Far Western-dominated state it lies within. Another small communitarian region, the Hawaiian Islands of Greater Polynesia, has poor reserves of social capital, with lots of newcomers and few associations to build trust. As in other analyses we’ve performed, Tidewater presents a strong departure from its legacy cultural pattern, with quite high social capital driven by having lots of associations — secular or otherwise — and unusually high scores within its major metro counties.
We’ll be further exploring social capital in the coming months, and hope to be able to gather the data to see if there have been any changes in the regional distribution of social capital over the past decade and how those changes correlate to various factors, including support for authoritarian political movements, ethnonationalist politics, and sweeping economic policy changes.
Thanks, as always, to our partners at the culturally-minded geospatial consultancy, Motivf, where Tova Perlman wrangled this data and Aimee Trehey created the maps you see herein.
— Colin Woodard, author of American Nations, Nations Apart and other books on U.S. history, geography, nationhood and politics, is the director of Nationhood Lab.