Bread and Circuses THE REPLACEMENT OF AMERICAN COMMUNITY LIFE

Lyman Stone APRIL 2021

AMERICAN ENTERPRISE INSTITUTE Bread and Circuses THE REPLACEMENT OF AMERICAN COMMUNITY LIFE

Lyman Stone APRIL 2021

AMERICAN ENTERPRISE INSTITUTE Contents

EXECUTIVE SUMMARY ...... 1

I. INTRODUCTION...... 2 What Is Associational Life?...... 3 Why Does Associational Life Matter?...... 7

II. THE HISTORY OF ASSOCIATIONAL LIFE...... 10 The Bond of Brotherhood: The Fall, Rise, and Fall Again of Fraternal Life...... 10 The Young Republic: Voter Participation over Time ...... 13 Hard Work: Unions and Labor Activism ...... 17 The Home Front: Veterans’ Organizations ...... 20

III. BREAD AND CIRCUSES: ARE SPORTS A FORM OF SOCIAL CAPITAL?...... 23 Football vs. Boy Scouts: The Trade-Offs in Youth Lives...... 34 School Years: Training Ground for Adult Community Life ...... 36 Extended Adolescence and College Sports...... 38

IV. MEASURING ASSOCIATIONAL LIFE...... 40 Paper Trails: Measuring Organizations over Time ...... 40 Seize the Day: Changes in How Americans Spend Their Time...... 44 Complicating Factors: The General Social Survey and Reported Social Lives ...... 46 How We Talk: The Long Arc of Associational Life ...... 51

V. CONCLUSION...... 54 When It Counts: Community Health and COVID-19...... 54 What Can Be Done?...... 60

ABOUT THE AUTHOR...... 62

NOTES...... 63

ii Executive Summary

n recent years, policymakers and cultural com- However, most of the change in associational life can I mentators have identified “social capital” as a key be attributed to essentially one factor: technological factor in determining individual and societal well- improvements leading to a higher standard of living. A being. But social capital is often hard to identify. wealthier society provides more benefits via the state Debates rage about what should count as social cap- instead of private organizations, even as the invention ital, how to measure it, what causes people or places of radio and television displaced many traditional to have more or less of it, and whether and how it information networks. Piggybacking on mass media, affects important social outcomes of interest. This popular sports have conquered the American calen- report focuses on one dimension of social capital as dar, displacing numerous other activities. This history particularly important and more concrete: associa- cannot be undone. tional life. How Americans socialize and spend time But government policies do matter. Greater school together, the kinds of organizations they form, and choice allowing educational pluralism could connect how they talk about life together are all relatively young Americans to broader associational oppor- tractable measures of one kind of social capital that tunities and more diverse avenues for social capi- may be important. tal accumulation. Alternatively, legally safeguarding Tracing the history of associational life from Amer- “coordinated leisure” times such as evenings, week- ica’s founding reveals that not all associations are ends, and holidays by adding more federal holidays, created equal. Some popular associations provide protecting workers from unpredictable work schedul- undeniably positive benefits for their participants ing, and reinstituting Sunday closure laws could also and society (such as churches or labor unions), while help. Furthermore, zoning rules that make it effec- others have proven deeply destructive (such as the tively impossible for social and civic associations to Ku Klux Klan). Still others, such as the once-popular construct facilities close to clusters of members could Knights of Pythias, seem to make little difference also be repealed. despite their large member counts. Thus, the link Nonetheless, while there is a role for government, between associational life and social capital is com- it’s ultimately up to everyday Americans to decide plicated. Intangible social relationships and ties that what kind of lives they want to live. The future of provide benefits for society arise from some forms of associational life in America will depend on what associational life, but not others. Americans really want. If they want modern bread Numerous factors have driven changes in asso- and circuses, they will get it. Rebuilding lost associa- ciational life over time, so no single cause can be tional life will require a critical mass of Americans to identified and no single policy response can revive make costly personal choices to reinvest in their com- associational life and corresponding social capital. munities and relationships.

1 I. Introduction

ccording to Google Ngram, a massive searchable These questions ranged from extremely simple A database of published texts in various languages, to (as can be seen) absurd, but they reveal inter- the term “social capital” has risen from appearing once esting differences: A majority of respondents said per nine million published words in 1980 to once per social capital does not have to be “pro-social” or 300,000 today. A concept that was virtually unknown “laudable” (i.e., it doesn’t have to be socially produc- in recent memory is today twice as commonly cited tive—which is unusual, since economic capital is by as related terms such as “community development.” definition part of a production function). Many peo- Ideas such as social capital (or related “civil society,” ple (including experts on the subject such as Scott which shows a highly correlated rise) have become a Winship, then the senior-most staffer at the JEC major part of the political vocabulary of American life and prior head of the Social Capital Project) think of over the past 30 years. social capital as any social resource a person might The Joint Economic Committee (JEC) houses the use to achieve their goals. Others disagree, suggest- Social Capital Project, which has held hearings and ing that social capital should be defined as resources published papers on topics ranging from “deaths of that are socially beneficial. despair” to educational reform to housing policy. Even more strangely, while 92 percent of respon- There is a Social Capital Research consultancy ser- dents rejected that “having social capital means some- vice.1 The concept has come into its own in a polit- one owes me something,” 59 percent of respondents ically and economically powerful way. And yet, the agreed that social capital essentially means “reciproc- term means many things to many people. ity” in relationships. Effectively synonymous questions In an informal poll of my Twitter followers (who asked mere minutes apart to virtually the same group are likely to be policy-informed individuals familiar of people had polar opposite responses. And perhaps with the term social capital), across 26 hypothetical most jarringly of all, respondents were evenly split activities offered, only one scenario secured a major- about how to describe social capital. About half said ity of respondents saying it was “definitely” build- social capital is something individuals have, while about ing social capital.2 That scenario was: “You and a half said social capital is something communities have. group of people you met alone gather in person over This exercise was not scientific, but it proves a a few months to work on a shared interest project basic point: People don’t know what social capital together.” And yet in a follow-up question, most means. Even among people who use the term, it has respondents still thought that project was build- numerous different meanings. Virtually every paper ing social capital, even if the project was murder- about social capital spends its first sections simply ing a mutual enemy. A quarter of respondents said trying to define the term. It is so popular and vogu- praying alone in a closet builds social capital, and ish a label that every researcher wants to connect 61 percent said watching a popular Netflix series their work to it. As will be shown later in this report, alone at home builds social capital. On the other this has sometimes resulted in confusing outcomes, hand, majorities said interacting with someone on with social capital having well-demonstrated opposite Twitter did not build social capital. And, interest- effects on societal outcomes of interest. ingly, majorities said that exacting vigilante justice This report certainly is interested in what research- against a serial killer did not build social capital. ers talk about regarding social capital. But having

2 BREAD AND CIRCUSES LYMAN STONE

reviewed the extant literature and explored the term’s At first glance, this looks like a standard story of utility, I will depart in this report from prior research social capital. But on closer examination, Tocqueville in an important way: It is not about social capi- doesn’t present associations as a kind of social capital. tal, because the term doesn’t have a stable meaning. His associations include prisons (a state function), inns Rather, this is a report about “associational life”—the (a for-profit business), and even parties. Tocqueville various ways Americans spend time together, espe- was not describing social capital; he was drawing a con- cially for purposes that are not strictly related to earn- trast between sources of legitimacy in different societ- ing money and paying bills. ies. In England, it came from a royal endorsement or a Instead of conceiving of the entire web of relation- lord’s grant; in France, the imprimatur of the state; but ships and institutions in which people’s lives occur in America, the legitimizing force and the entrepreneur- as one thing—social capital—this report focuses ial energy behind an endeavor (even a for-profit or gov- on a more specific and concrete question: How has ernmental venture) came from associations. American participation in organized community life Thus, for example, whereas William Shakespeare’s changed over time, and what effects has that partici- theater company was called Lord Chamberlain’s Men, pation had? This is a report about community build- in America, the patron is replaced by the association; ing, social life, church, sports, television, and the 1792 the first theatrical company in America was called yellow fever outbreak in . It’s about our simply the American Company. Tocqueville did not lives together and how the rise of some forms of asso- argue that America had more theaters and inns than ciational life (especially telegenic activities such as Europe did (it did not), but rather that the theaters sports and politics) displaced other forms (such as and inns depended on implicitly republican asso- church and neighborhoods). ciations for their origin and legitimacy, rather than monarchical grants or state charters. A century later, historian Arthur Schlesinger Jr. What Is Associational Life? called America a “nation of joiners,” a term that has stuck since, so much so that it is often incorrectly When Alexis de Tocqueville visited America in the attributed to Tocqueville himself.4 When Schlesinger 1830s, he described Americans as exceptionally prone coined the term in the 1940s, he was writing in a coun- to “association” with one another. try swarming in clubs, societies, federations, and any number of other little polities voluntarily organized Americans of all ages, all conditions, all minds con- by Americans for innumerable purposes, as described stantly unite. Not only do they have commercial later in this report. and industrial associations in which all take part, The intensity of associational life that Schlesinger but they also have a thousand other kinds: religious, saw did not exist when Tocqueville was writing and moral, grave, futile, very general and very particular, does not exist today. Nor was Tocqueville even writ- immense and very small; Americans use associations ing about the kinds of associations Schlesinger had in to give fêtes, to found seminaries, to build inns, to mind. Whereas the associational life of the 1940s was raise churches, to distribute books, to send mission- dominated by mutual-benefit societies, fraternities, aries to the antipodes; in this manner they create and social clubs, the associations Tocqueville wrote hospitals, prisons, schools. Finally, if it is a question about were as likely to be for-profit corporations, of bringing to light a truth or developing a sentiment educational institutions, or even governmental func- with the support of a great example, they associate. tions. The associational life Tocqueville described is Everywhere that, at the head of a new undertaking, not what modern researchers measure when they talk you see the government in France and a great lord in about social capital. England, count on it that you will perceive an associ- Nonetheless, from Tocqueville’s description of ation in the .3 America, and especially Schlesinger’s reinterpretation

3 BREAD AND CIRCUSES LYMAN STONE

of it, scholars of social capital have built a narrative Thus, it is extremely challenging to measure of American culture and life together. “Tocquevillian” even concrete associations and organizations, let behaviors of people acting as joiners to form clubs alone social capital writ large, across any particu- and societies are portrayed as fundamentally Amer- larly long period, as how people associate with each ican. The story of social capital as an American idea other changes across time and often defies easy begins here. quantification. But as noted above, social capital as an idea can be Contemporary research on social capital inevi- nebulous. Does it refer to concretely defined commu- tably stems from the kinds of questions raised (and nities with formal membership or just a general feel- partially answered) in Robert Putnam’s book Bowl- ing of trust and goodwill toward others? Does it refer ing Alone.5 Assembling an impressive range of sta- to intact families or thriving schools? Does it refer to tistics, Putnam argued that associational life in individuals who pay their taxes and fill out their cen- America had declined markedly since the middle of the sus forms on time, or is it about people who check on 20th century, with the extraordinary decline in bowl- sickly neighbors? ing leagues providing the title. Figure 1 shows the Different measures of social capital can even membership statistics of the American Bowling Con- contradict each other. People who have a great gress from its inception to the present day. deal of trust in and commitment to civic and gov- This sharp decline in associational behaviors is ernmental institutions might be less interested in the supposed sign of declining social capital in Put- mutual-benefit societies and private charity. Those nam’s research. Americans were joining more and with strong extended family networks might be less more associations until the middle of the 20th cen- interested in activities detached from their kinship tury, and then, quite suddenly, they weren’t. Looking networks. Thus, not only might different kinds of at 34 chapter-based associations such as the Grange, social capital tell us different things about society, Boy Scouts, and the NAACP, Putnam assembles a but they may be rivalrous. standardized measure of associational participation Finally, social capital may reveal itself in different reaching back to 1900, which I and other researchers ways at different times, making it difficult to track have extended to the latest available data, as shown across time. Perhaps one of the best examples is the in Figure 2.6 Grand Army of the Republic (GAR), the Union vet- Based on Putnam’s data, participation in these erans’ organization formed after the Civil War. The chapter-based associations is as low as it was in 1900. GAR was an influential organization that provided This would seem to be an extraordinary decline. a wide range of activities and benefits for members But there is some reason to be skeptical of these and their communities, forming one of the largest claims. A large number of the organizations Putnam social organizations in American history up to that studies were founded in the late 19th or early 20th time. But if membership in social organizations were centuries, and virtually all of them were organizations tracked using a list of major organizations as they that survived in some form until the 1980s or 1990s appeared in either 1860 or 1940, the GAR wouldn’t for Putnam to study. Furthermore, organizations show up at all. It was founded after the Civil War, founded in the 1980s or 1990s were not included, not and its last members died in the 1930s. It existed for least because knowing which ones would be large a brief moment. But groups of veterans organized enough to merit study wasn’t obvious. together in other ways: Revolutionary War officers In other words, Putnam’s research, while impres- formed the Society of the Cincinnati, while veterans sive, ends up looking at an extremely narrow slice of of the Spanish-American War formed the Veterans of American associational life: the organizations that Foreign Wars. Many forms of associational life have had many members between 1900 and 1960. It is been ephemeral, existing only briefly, serving a spe- striking that historians have labeled the period from cific purpose for a specific time before fading away. 1880 to 1930 “the Golden Age of Fraternalism.”7 For

4 BREAD AND CIRCUSES LYMAN STONE

Figure 1. The Rise and Fall of Bowling

10,000,000 ess 9,000,000 8,000,000 7,000,000 6,000,000 5,000,000 4,000,000 American Bowling Congr 3,000,000 2,000,000 1,000,000

Members of the 0 191 0 191 6 1898 1904 1922 1928 1934 1940 1946 1952 1958 1964 1970 1976 1982 1988 1994 2012 2018 2000 2006

Source: US Census Bureau, Bicentennial Edition: Historical Statistics of the United States, Colonial Times to 1970, September 1975, https://www.census.gov/library/publications/1975/compendia/hist_stats_colonial-1970.html; US Census Bureau, Statistical Abstract of the United States 2010, 2009, https://www.census.gov/library/publications/2009/compendia/statab/129ed.html; and United States Bowling Congress, website, https://www.bowl.com/.

Figure 2. Average Membership Rate in 32 National Chapter-Based Associations, 1900–2017

1.0

0.5 Among

opulation 0.0

–0.5 ant Recruiting P dized Membership Rate

Rel ev –1.0 Standar

–1.5 191 2 191 6 1900 1904 1908 1920 1924 1928 1932 1936 1940 1944 1948 1952 1956 1960 1964 1968 1972 1976 1980 1984 1988 1992 1996 2012 2016 2000 2004 2008

Source: Robert Putnam, Bowling Alone: The Collapse and Revival of American Community (New York: Simon & Schuster, 2000), 54, 426, 450; and Taylor Mann, in discussion with the author, July 5, 2019.

5 BREAD AND CIRCUSES LYMAN STONE

reasons that historians still do not fully understand, effect of burning black churches to the ground. The Americans during that period founded an astonish- KKK’s purpose was to terrorize black, Jewish, immi- ing number of mutual-benefit, fraternal, auxiliary, grant, and Catholic Americans. Likewise, while few and club-based societies. It was the age of “the club” clubs were formed for the explicit purpose of terror, as a place where men in particular would spend their many membership-based organizations nonethe- time. By 1905–10, it is estimated that total member- less had as a selling point not the bonds they would ship in these societies reached 10 million or more, and build between members but whom they would keep membership in an officially traceable subset of these out. Especially in more diverse areas without strictly groups for an almanac produced in 1896 totaled well enforced racial segregation (such as cities in the Mid- over five million, and nine million in 1906.8 west and Northeast), “members-only” associations While it is tempting to see this efflorescence of could achieve for their members what Jim Crow laws fraternalism as a sign of robust social capital, histo- achieved in southern states. rians are divided on the question. Many see these While state-by-state data are not available for many groups as institutions of social division.9 One of the national membership associations in historic periods, latest-arriving such “national membership clubs” is data from the world almanac of 1896 show that mem- one Putnam does not include in his estimates, per- bership for many social fraternities was much higher haps because it would problematize accounts of in the North.12 For example, the Independent Order of their inherent value: the Ku Klux Klan (KKK), which Odd Fellows, a social fraternity that was whites-only claimed to have five million members in the 1920s and remains de facto segregated today, could claim (though almost certainly had fewer than that).10 The three times the membership rate among white males vast majority of these membership-based organiza- in northern states as in southern states. Similar differ- tions restricted membership to individuals of cer- ences held for the Knights of Pythias and, of course, tain races, ancestries, convictions, or classes, and, of the GAR. course, sex-based limitations were nearly universal. But even the relatively liberal Masons, for whom Scholars of social capital are aware of this issue and the best data are available, were affected. While slav- have often tried to address it by delineating between ery existed, Masonic membership was nearly twice two different types of social capital: “bridging” cap- as common among northern white men as among ital, which helps different types of people connect, southern white men. There were black Masonic and “bonding” capital, which helps similar types of lodges, but they were relatively small compared to the people deepen their commitments to each other. It longer-established white lodges. After the Civil War, is sometimes argued that bonding capital can go too however, the geographic trends flipped: Freemasonry far. For instance, the KKK was in some sense a form expanded everywhere, but it expanded fastest among of exaggerated bonding capital whereby white peo- southern whites seeking new forms of solidarity with ple bonded so strongly that it made bridging to oth- each other. While some traditionally white Masonic ers harder. A study of firewood collection in rural lodges began to integrate in the late 1980s, many China found that individuals with strong local bonds southern states still had not integrated their lodges as avoided over-collecting local firewood, but peo- of the 2000s.13 ple with stronger bridging ties outside the commu- And indeed, many associations explicitly recapitu- nity overconsumed firewood.11 But if high bridging lated deep social tensions. The iconic association of social capital destroys local socially shared resources, the Golden Age of Fraternalism was, as mentioned, should it even be considered social capital at all? the GAR, which was soon after met by the rival United This contrast between bridging and bonding also Confederate Veterans. Two of the larger social clubs seems facile in the case of the KKK or other racist of the era existed as exclusive clubs for Civil War vet- associations. The KKK’s purpose was not to bond the erans; it’s difficult to call this either bonding or bridg- members in it closer together with an incidental side ing social capital as some scholars might suggest. It’s

6 BREAD AND CIRCUSES LYMAN STONE

better described as a society that did not fully heal. membership as his key benchmark for social capital, Relatedly, academic research has found that regions of but by now it should be clear that this is not necessar- Weimar Germany with more membership in secular ily appropriate. social organizations gave more support to the Nazis.14 This report takes a general approach to associa- In this context, explicit membership restrictions tional life. It tracks numerous different indicators based on race, nationality, or other traits may not sim- of how Americans organize their social lives. Where ply be about bridging or bonding capital. When the possible, it notes shortcomings in each data source. Sons of Norway mutual-benefit society limits itself But the overall conclusion is somewhat different to Norwegians for their insurance-pooling product, than what prior research has found. While associa- that may be bonding capital. But when a club that is tional life in America today may be in decline, that generally available to men of good conduct arbitrarily decline is far more gradual than has previously been excludes black men of good conduct, it’s not social suggested. New forms of associational life have capital at all. It’s social corrosion. Likewise, when sprung up. However, those new forms may have the General Federation of Women’s Clubs (GFCW) different implications for how Americans live and was formed as a specific federation of white wom- cooperate together. en’s clubs, it’s not because white women were some natural constituency being bonded closer together by the GFCW; it’s because associational life recapitu- Why Does Associational Life Matter? lated, and perhaps even deepened, existing fissures in a deeply divided society. The 21st century has seen an explosion in interest in This kind of social capital will sometimes still be ancestry driven by an aging population interested in its described as social capital. Recall many of my sur- own legacy and increasing access to digitized records. vey respondents argued that even activities that are As Americans dig into their past, they lean on various destructive or bad, such as organizing a murder, could sources to understand what life was like: census doc- be seen as building social capital provided they are uments, immigration records, and family documents, done together. But this seems to divorce the term but also church records of baptisms and weddings, “social capital” from its economic origins; to be social newspaper clippings, club and social memberships, capital, surely something must be not just social but and school enrollment and graduation records. actually capital (i.e., a net-value-enhancing part of a Americans understand their own histories largely society’s production function). If it doesn’t create through the lens of how past individuals interacted net-positive value for the average member of society, with what experts today would call “civil society,” it probably shouldn’t be labeled social capital. Social the network of organizations, societies, and affilia- “currency” may be a better term; it is just the “coin of tions that people use to organize their public lives. At the realm” lubricating exchange, allowing individuals the extreme, historians even use these data for other exchange to achieve their possibly destructive goals, purposes. The 1793 yellow fever outbreak in Philadel- not a productive asset in an economically meaning- phia, Pennsylvania, one of the deadliest epidemic out- ful sense. breaks in modern, recorded history, is remembered But while membership in organizations is more largely because Philadelphia churches kept detailed concrete than social capital is, it isn’t the sum total records of vital events, implicitly tracking the course of associational life and certainly not associational of the epidemic.15 life that might be useful for society, since member- Formal associations have the unique feature of ship could be driven by deep suspicion and hostility leaving a paper trail for future researchers. But this to others or could create new hostility to others. Like- isn’t just an academic question; historic paper trails wise, knowing what organizations to track in which and records influence how we see and engage with periods can be difficult. Putnam uses organizational the world today. Indeed, when I asked my Twitter

7 BREAD AND CIRCUSES LYMAN STONE

followers, “If you find a packet of letters written by a key areas of material and economic interest, facilitat- great-grandparent you only met as a child but never ing other charitable and voluntaristic collaborations knew, and reading it deepens your sense of connec- as well. And of course, every group, association, chap- tion to them, have you built social capital?” more than ter, and assembly is an opportunity for politicians to a third answered in the affirmative.16 stump and fundraise. Whether Lincoln’s now-famous Beyond their paper legacies, established organiza- lyceum address or Cooper Union speech or Alexander tions build structures that can last for centuries. Cities Stephens’ cornerstone speech, American politics used across America are littered with old “Pythian castles,” to happen in the meeting halls and common rooms clubhouses of a now-defunct social club from the of clubs, societies, churches, and other associations. 19th century. They can also create institutions that Social and civic institutions depend on each other. outlive them. Freemasons made up an astonishingly Nonetheless, some overarching trends can be large share of the Founding Fathers, and Freemasonic observed. For example, the rise of the modern brotherhood was consciously engaged in promoting bureaucratic and welfare state is clearly connected “liberal revolution” through the 19th century.17 The to the decline of many institutions it replaced. More- constitutional institutions we have today reflect, to a over, the military conflicts of the 19th and 20th cen- nontrivial extent, the values, interests, and bonds of turies had a formative role in creating one particularly brotherhood of institutions that, even if they tech- important and dominant form of American associa- nically survive today, have dramatically reduced in tional life: sports. Specific state policies related to prominence. secularization weakened churches, even as anti-union But even as interest in these historic associa- economic policies and an economic sectoral shift tions has waxed, their living descendants today, toward service-sector work instead of manufacturing actual churches, social clubs, and organizations weakened unions.18 have waned. Membership and participation in vir- The rise of increasingly individualistic consump- tually every form of measurable association have tion habits, as expressed most clearly in the explo- declined, including churches, unions, and social and sions of leisure travel and social media, is a durable charitable clubs. Even once-robust American institu- force driven by deep changes in economic structure. tions such as the Boy and Girl Scouts seem to have Finally, to a considerable extent, the decline in formal hit their peak and are now in noticeable decline. The associational life mirrors changes in work, family, and decline in associational life extends far beyond these education that are unlikely to be reversed. More years membership-based organizations. However, signs of of schooling, delayed marriage, two-earner house- the decline in interpersonal connection can be spot- holds, and below-replacement-rate fertility have all ted in changing linguistic patterns, social surveys, and created challenges for formal membership-based even data on time use. organizations. The causes of this widespread decline in asso- Changes in associational life in America have con- ciational life are multifarious and interconnected. sequences. As Americans spend less time together, we Seemingly unrelated associations may be interde- have also become more suspicious and less trusting pendent, and when one declines, others suffer too. of one another, especially of civic and public institu- Scout troops may be organized by churches, and thus tions. As a result, our society has had a hard time con- when churches decline, the energy to organize the fronting the challenges and obstacles that, throughout scout troops vanishes. Wars created many veterans, human history, have always arisen. who often hired their brothers in arms, formed clubs Most dramatically, the COVID-19 pandemic has and leagues, and “won the peace” together. No one highlighted the dysfunctional character of Ameri- wants another world war, yet losing this strengthen- can civil society and associational life. Several aca- ing force clearly affected the fabric of American soci- demic papers have found that different measures of ety. Union membership can bind together families in social capital, cohesion, and associational life have

8 BREAD AND CIRCUSES LYMAN STONE

all affected the spread ofCOVID-19 .19 This report as zoning and planning principles, the structure of demonstrates that although associations such as welfare and social-relief spending, labor policy, and churches have repeatedly been singled out as super- federal holidays all interact with associational life. spreaders, in reality, areas with denser networks of However, the core drivers of declining associational associational life have performed better than other life extend beyond purely political questions. Resolv- places have. ing them will require Americans to make conscious Public policy has a role to play in reconstructing personal choices to reinvest in their communities in a vibrant associational life in America. Policies such costly and significant way.

9 II. The History of Associational Life

he bulk of this report concerns itself with the officers of a certain rank), and innumerable others of Thistory of American associational life. However, the associations Tocqueville would notice all entered rather than moving strictly chronologically, each sec- the scene. By the early 1800s, many other lodge- and tion explores a different kind or measure of associa- chapter-based societies had sprung up, such as the tional life, tracking how it has changed throughout Independent Order of Odd Fellows, which by 1860 American history. Ranging from social fraternities and had over 200,000 members.21 veterans’ associations to youth sports and pronoun But in general, Americans after the revolution did usage, it explores changes in how Americans experi- not focus their energies on social clubs and fraternal ence and think about our life together. societies. The Society of the Cincinnati faced pub- lic attacks due to worries that such an officer’s club might conspire against the young republic.22 That The Bond of Brotherhood: The Fall, Rise, the Society would be open only to the male descen- and Fall Again of Fraternal Life dants of the original officers entrenched social wor- ries that closed social clubs would inevitably form the Of the 29 groups Putnam analyzed for his long-term basis of exclusive aristocracies. Strikingly, no general associational data, seven are fraternal societies. With organization of veterans of the American Revolution their stated purpose of promoting social bonding and was ever formed, despite numerous near and actual their frequent involvement in philanthropic activities, insurrections by those veterans, often to press Con- fraternal societies are a particularly interesting case of gress for better benefits. Such a cause could easily an association that might build useful social capital. have lent itself to the formation of a wartime veterans Furthermore, while prior research has only charted association. membership in these groups back to 1900, statistics And while an astonishing share of American men, exist for many such groups well into the 19th and even and especially American men associated with the 18th centuries. cause of liberty, were Freemasons, the share of adults In the decades leading up to the American Revo- who were Masons almost certainly declined between lution, fraternal societies could claim perhaps 10 per- the 1780s and 1860, given that sporadically reported cent or more of adult Americans as members. This membership statistics show little growth and found- figure involves a great deal of extrapolation to reach, ing of new lodges slowed down, even as the general but even if only Freemasons are counted, for whom population grew rapidly. The Society of the Cincin- at least some concrete data on the founding of new nati nearly collapsed from lack of members in the lodges exist, there were probably more than 50,000 1850s, as eligible heirs of revolutionary officers simply “master Masons” by the end of the Revolutionary War, declined to join.23 amounting to over 6 percent of free white males claim- The American suspicion of such clubs and soci- ing membership in just this one fraternal society.20 eties, despite many of the republic’s founders being The end of the revolution saw a flurry of civic members of them, culminated in the formation of activity with many new societies, especially related the first major third party in American history: the to political causes, being formed. Nascent political Anti-Masonic Party. Much as the public feared that parties, America’s first formal veterans’ association the Society of the Cincinnati would give rise to an (the Society of the Cincinnati, exclusively for white aristocracy, the secretive rituals of Freemasonry

10 BREAD AND CIRCUSES LYMAN STONE

fueled public paranoia that antidemocratic forces The extent to which stories like the Morgan Affair were working to destroy the young republic. That just seem to involve an uncanny number of Freemasons a decade earlier Britain had invaded the country spe- can at first seem conspiratorial; the Freemasons must cifically for that purpose and torched the capital likely have been a wily conspiracy getting into high places made those fears quite potent. of power, so the reasoning goes, because look how But these fears about secret societies, and espe- many important people were Masons. But this rea- cially Freemasons, came to a head in an event known soning misses a vital fact: By 1860, about one of every as the Morgan Affair in 1826. William Morgan was a 100 adult males in America was a master Mason. former Freemason who, upon some financial diffi- Freemasonry was not in any meaningful sense a culties, announced he would write a book exposing secret society; it was a widespread social group that Freemasonic secrets. Shortly after this, Morgan was many ambitious men joined as a way to network and repeatedly arrested on disputed charges by local offi- advance, often after they had become successful. Free- cials (who were Freemasons), then kidnapped from masonry was twice as common in American society in jail (by Freemasons) and never seen again. While the early 19th century as Eastern Orthodox believers murder was never proven, it was widely presumed or Libertarian Party membership are today, and by the (though Freemasons claimed they paid him to move 1880s, it was as common as red hair or green eyes. to Canada under an assumed name). The scandal (and Thus even as elites may have joined the Masons to the book, which was eventually published) seemed to further their careers, the organization shrank steadily validate the worst fears about Freemasonry. as a share of all American men. It appealed to the The resulting furor was disastrous for Freema- ambitious but aroused the suspicions (and perhaps sonry in America, with popular protests against even jealousies) of wider society. Fraternal life in Freemasonry springing up in New York and other America reached a low ebb in the early 19th century states. Capitalizing on the moment, anti-Masonic on the heels of the success of the anti-Masonic move- sentiment took a strident anti-elitist tone but, par- ment. Figure 3 shows the long-run trend in member- adoxically, also fused with anti-Jacksonian senti- ship in fraternal societies in America as a share of the ment, since Andrew Jackson was a Freemason. As a adult population. result, the Anti-Masonic Party was formed, merging But after the Civil War, fraternal life began to anti-Jacksonianism and a unique strand of anti-elitist rise again, driven by an explosion in membership in populism. long-standing organizations such as the Freemasons The Anti-Masonic Party won five congressional and the Odd Fellows, but also by brand-new organi- seats in 1828, 17 in 1830, and 25 in 1832, before declin- zations such as the Knights of Pythias, Order of the ing and, by 1840, merging with the Whig Party. Per- Eastern Star, Order of the Elks, and Shriners. From haps even more impressively, the Anti-Masonic the 1860s onward, fraternal societies simply become candidate (a former Freemason) in 1832 won almost too numerous to count. One “cyclopædia” of such 8 percent of the popular vote and seven electoral groups in 1896 counted over 600 in the United States, votes (this despite, oddly enough, his refusal to ever each with its own network of chapters and lodges.25 actually condemn Freemasonry in public). Of course, However, historic estimates of total membership at about 100,000 votes, the Anti-Masonic Party’s can- exist for 1896, 1906, and 1916, collected by contempo- didate received fewer than twice as many votes as the rary experts who contacted numerous organizations number of Masons at the time (likely around 60,000). and produced estimates for others. These estimates As the anti-Masonic movement declined, its mem- count over five million members in 1896, over nine bers mostly gravitated toward the nascent National million in 1906, and over 16 million in 1916.26 Note Republican and Whig parties, supplying them with that nine million memberships in 1906 do not mean a much-needed popular movement to combat that nine million Americans were members. Many Jacksonianism.24 individuals were members of numerous societies and

11 BREAD AND CIRCUSES LYMAN STONE

Figure 3. The Fall, Rise, and Fall Again of Fraternal Societies

35%

30%

25% pulation

20% Adult Po 15% e of

10% as a Shar 5% otal Memberships in Fraternal Societies T 0% 1771 181 6 1861 191 5 1951 1762 1780 1789 1798 1807 1825 1834 1843 1852 1870 1879 1888 1897 1906 1924 1933 1942 1960 1969 1978 1987 1996 2014 2005

Source: General Social Survey, 1972–2004; World, “The World Almanac and Encyclopedia 1896,” January 1896, https://archive.org/ details/worldalmanacency1896newy/page/n5/mode/2up; William D. Moore and Mark A. Tabbert, Secret Societies in America: Foundational Studies of Fraternalism (New Orleans, LA: Cornerstone Book Publishers, 2011); and author’s estimates and imputations. clubs. Thus, fraternal memberships amounting to accounts of men with memberships in half a dozen about a third of the adult population in the mid-1920s clubs are common. In one example, a well-known does not imply that three in every 10 adult Americans businessman in Detroit was a Mason, a member of were members of a club. the related order the Nobles of the Mystic Shrine, an First of all, most such groups were limited to men Odd Fellow, a member of the Elks and Mooses, and (and many to white men), so membership among part of the Knights of Pythias. He also participated white men was certainly even higher. Data from four in a yacht club, a musical club, and a German gym- large organizations overwhelmingly composed of or nastics organization.28 Local notables often joined restricted to white men in 1906 (white Masonic lodges, numerous such societies. Thus, while total fraternal Odd Fellows, Pythians, and the GAR) show that mem- and social memberships were sky-high compared to bership rates in some places could be extremely high. population, the actual share of the population that These four groups alone counted total membership was a member of any fraternal society was almost equal to 30 percent of the white, adult male popu- certainly lower. lation in Maine, 26 percent in Colorado, and 24 per- But it isn’t even that simple. Many fraternal organi- cent in New Hampshire.27 And yet these groups made zations had related “auxiliaries” for the wives of male up about only a third of total reported membership members. That is, even if children and women were in such societies. It’s likely that in some states, social not members of a fraternal society, they might none- and fraternal group memberships totaled two-thirds theless have considered themselves closely connected or more of the white male population. to it. So while the share of individuals who were mem- Secondly, local businessmen, politicians, and bers of social fraternities was certainly lower than other notables had strong reasons to join clubs: to implied by the total number of memberships across make connections useful to their trades. Historical all groups, the community-wide footprint of those

12 BREAD AND CIRCUSES LYMAN STONE

social fraternities was probably bigger than their raw club existed as a space to participate in shared cul- membership counts were. ture, get the news, hear other people’s opinions about Whatever the case, the 1920s really did mark a that news, and socialize. But the advent of radio and Golden Age of Fraternalism. Membership in these television radically altered how Americans enter- societies plummeted during the Great Depression as tained themselves. Academic research in Indonesia the dues became prohibitively expensive. Many soci- has found that improvements in signal strength for eties collapsed entirely; others vanished in all but radio and TV reduce various measures of social cap- name. Societies such as the Knights of Pythias, Royal ital and community participation.29 Further research Arcanum, the Order of Red Men, and the Knights has found that the expansion of broadband internet in of the Maccabees all claimed tens or hundreds of the United Kingdom is closely connected to declining thousands of members in the late 19th and early participation in organized social and civic groups.30 20th centuries. But during the Great Depression, they It’s reasonable to suppose similar forces have been all experienced extraordinary decline. at work in American history. The rise of competitive The Knights of Pythias still exists, though with sports, discussed at length below, likewise doubtless fewer than 20,000 members today, versus almost replaced many club loyalties and affiliations. 900,000 in 1920. Royal Arcanum exists today almost And while this may seem to imply that television exclusively as a life insurance company. Because of and sports killed associational life, the reality may be their extraordinarily precipitous decline, they aren’t even more pessimistic: If television can replace the included in Putnam’s assessment of organizational service a group provides to its members, then that membership. Even long-standing organizations such group might never have been doing very much to begin as the Freemasons have seen long-term declines in with. Indeed, one of the few detailed, empirical studies membership. Today, total membership in fraternal of how fraternal societies in the Golden Age engaged societies amounts to just 5 percent of adults in Amer- with wider society (a study of the Knights of Pythias in ica, a lower figure than at any time since the Civil War, Buffalo, New York) found that the Pythians were seg- though not quite as low as the period of the Anti- regated along ethnic and class lines as much as wider Masonic Party. society and left no documentable record of almost any This is at once the story of extraordinary decline social or charitable engagement other than lobbying told by Putnam and something else. Fraternal mem- for regulatory privileges for their organization.31 The bership is as low today as it was in the 1830s, when reality is that many of these organizations were never Tocqueville visited America and thought Americans really doing very much. to be the most association-obsessed people he had ever seen. The decline of fraternal societies may be a decline in valuable social capital, but past periods with The Young Republic: Voter Participation few such societies were hardly devoid of vibrant asso- over Time ciational life. Tocqueville was impressed with Amer- ican associational life not because he encountered When Tocqueville saw Americans forming societ- the Odd Fellows or the Society of the Cincinnati but ies, a key kind of social life he observed was polit- because he saw inns, churches, road maintenance soci- ical. Compared to his native France, Americans eties, and missionaries. While fraternal membership had an incredibly high rate of political participa- may be an important form of associational life in some tion. The French parliamentary election of 1831 had circumstances, its rise and fall are not definitive. There just 125,000 votes cast due to extremely narrow are other vital forms of American social life. rules about the electoral franchise, equivalent to Moreover, many fraternal organizations them- about 0.4 percent of the population. On the other selves have a quite specific account of their own hand, 1.1 million Americans voted in the 1828 pres- declines: mass media. To a considerable extent, the idential elections, making up over 9 percent of the

13 BREAD AND CIRCUSES LYMAN STONE

Figure 4. The Rise of Political Participation

80%

70%

60% opulation 50% of P

40%

30%

20% otes Cast as a Sha re V 10%

0% 2011 1801 181 5 1871 191 3 1941 1780 1787 1794 1808 1822 1829 1836 1843 1850 1857 1864 1878 1885 1892 1899 1906 1920 1927 1934 1948 1955 1962 1969 1976 1983 1990 1997 2018 2004 USA Germany France Denmark United Kingdom Norway Netherlands Sweden Italy Spain Switzerland Ireland Australia Mexico Portugal

Source: Author’s estimates based on official election tallies and estimated national populations. total population, even though women, children, and political participation than almost anywhere else in enslaved people could not vote at all. the world. Of course, elections in the 1790s in France had This high participation was not universally higher participation, as would elections in France beloved. Franchise expansions in the 1830s remov- after the 1850s, due to the oscillations between revo- ing property requirements to vote led to the rise of lution and reaction in 18th- and 19th-century France. Jacksonian populism, which many elites regarded But France was virtually the only country to have with disdain. The writings of the Founding Fathers higher electoral participation than the United States are riven with remarks critical of the “mob” and even at any point in the 19th century. From the 1820s of democracy. George Washington himself famously onward, more Americans voted, as a share of the total opposed formal political organization into parties. population, than voted even in relatively liberal and The creation of large-scale, organized, participatory democratic societies such as Switzerland. The Franco- mass politics was neither inevitable nor necessarily American duopoly on high electoral participation desired by all parties. South Carolina would not allow would not be broken until 1903, when high Australian individual voters to vote in US presidential elections participation would vault the country into the high until after the Civil War. But to virtually any outside rungs of electoral participation. Figure 4 shows votes observer, one of America’s uniquely exciting traits cast in a selection of countries as a share of their total was that a log-splitting country lawyer like Abraham populations over a long time horizon. Lincoln could become president by making compel- In the 19th century, Americans really did have a ling appeals to a voting public. remarkably vibrant political society. Despite the hor- But America’s exceptional position changed in the rific institution of slavery rendering a large share early 20th century, especially in the key years around of Americans disenfranchised, we still had broader World War I. The extraordinary pressures the war

14 BREAD AND CIRCUSES LYMAN STONE

Figure 5. The Golden Age of Voter Engagement: 1840–1916

90%

80%

70%

60% ormal Franchise 50%

esidential Elections 40%

30% e of Estimated F 20% otes Cast in Pr V 10% as a Shar

0% 181 6 1861 191 5 1951 1789 1798 1807 1825 1834 1843 1852 1870 1879 1888 1897 1906 1924 1933 1942 1960 1969 1978 1987 1996 2014 2005

Note: Franchise is estimated based on changes in federal rules about eligibility based on sex, age, race, and slave or free status, with adjustments for Civil War or Reconstruction geographic ineligibility; a limited set of state-specific rules related to sex in the late 19th and early 20th century; and broad estimates of state-specific stringency of property requirements before the Civil War. Source: Official presidential popular vote tallies.

placed on belligerent and even neutral powers forced in play; there have also been changes in rates of par- virtually every country to make various kinds of con- ticipation among enfranchised individuals. Figure 5 cessions to deepen the political commitments of their shows the number of votes cast in presidential elec- respective publics. Expanding voting eligibility was tions divided by the estimated number of individuals part of this effort, and thus the wartime years and their who were technically eligible to vote. immediate aftermath saw massive increases in elec- American electoral participation, even among toral participation in almost every country that had those formally eligible to vote, was not especially any electoral system at all, except for France. Subse- high before the Jacksonian period. And it was not quently, over the 20th century, electoral participation until the 1840 election, which saw the advent of mod- has continued to rise in most countries but has been ern campaigning, that truly exceptional participa- relatively stagnant in the US. Indeed, US political par- tion rates were achieved. The term “stump speech,” ticipation actually fell in the late 19th century as Jim while invented in the 1820s, was popularized during Crow laws reversed the gains in electoral access won the 1840 election as part of William Henry Harrison’s by black Americans during Reconstruction and new campaign brand, portraying the actually well-to-do naturalization and immigration laws made it more Harrison as a “log cabin and hard cider” man, a tough difficult for recent arrivals to become citizens. Indian fighter, and a populist leader. Extensions and restrictions of the franchise by From 1840 until 1908, more than 60 percent of eli- altering limitations based on property, age, sex, or race gible voters actually voted in each election, and usu- have certainly played a key role in driving changes in ally more than 70 percent, except for the election of electoral participation. But they aren’t the only factor 1864, when wartime disruptions reduced turnout.

15 BREAD AND CIRCUSES LYMAN STONE

Figure 6. American Disengagement

Votes Cast in Recent Presidential or Parliamentary Election as a Share of the Voting-Age Citizen Population 0% 20% 40% 60% 80% 100% Nigeria Slovakia Japan South Africa Botswana USA Estonia Malaysia India Czechia Taiwan Canada Ghana Spain Ireland Italy Georgia (Country) Austria Brazil Hungary France South Korea Germany Norway United Kingdom Turkey Netherlands Sweden New Zealand Australia Belgium Source: Citizen voting-age population estimated using IPUMS International population estimates from national censuses with data on population by age and citizenship and using national statistical offices with the same data for countries not covered in IPUMS Interna- tional. Vote counts are taken from the presidential or parliamentary election occurring nearest the data for which a population estimate is available for the most recent such population-election pair.

Not until Jim Crow laws had been firmly entrenched, But American political participation, at least as reducing suffrage among black men, did electoral par- measured by voting behavior, has flatlined for many ticipation fall appreciably. Virtually the whole decline decades. Electoral participation as a share of the pop- in electoral participation in the late 19th and early ulation and of the enfranchised population has been 20th century is driven by declines in the South. stagnant at very low levels. Figure 6 shows votes cast Voter participation fell again after the franchise was as a share of the citizen voting-age population for the extended to women but soon recovered as women US and a selection of other countries in recent presi- speedily picked up the habit of voting. Electoral par- dential or parliamentary elections. ticipation has essentially been capped at around While there are developed countries with lower 60 percent of eligible voters for the past century. political participation among adult citizens than the In other words, Americans in the 19th century were United States does (such as Japan), the US’s rate of living when the franchise was rapidly widening and participation is quite low. Americans are simply less participation among the enfranchised was high. Fur- involved in self-government than people in many thermore, political participation in America was far other countries are. more broadly accessible than in almost any other coun- This wasn’t always the case. Americans once led try. This was the nation of joiners Tocqueville saw. the way in political participation, and indeed enfran- chised Americans in the past had higher participation

16 BREAD AND CIRCUSES LYMAN STONE

than they do today. Our organized participation in But these initiatives remained relatively small-scale political life together is lower than in our past and and limited to specific cities or trades. Major, nation- low compared internationally. However, this decline wide unions did not began to appear until after the occurred between 1890 and 1920 and resulted pri- Civil War. The accelerating pace of industrialization marily from the targeted disenfranchisement of black in part brought on by the Civil War expanded the Americans. Research by the JEC has found that his- ranks of the native and immigrant working classes toric patterns of slavery predict weak social capital even as fraternal associations were beginning to rise today as well.32 Just as many membership organiza- in popularity. tions entrenched existing social divisions, low US The link between fraternal associations and unions political participation originated in the racialized pol- was often quite direct. A cobbler named Uriah Ste- icies of the Jim Crow era and, to a lesser extent, in phens founded the Knights of Labor originally as a increasing restrictions on immigration and natural- Catholic workingmen’s social fraternity, with rituals ization after 1900. modeled on other groups. (He was also a Freemason and a member of the Knights of Pythias and the Odd Fellows.) But the organization grew to be more than Hard Work: Unions and Labor Activism a social group and eventually coordinated strikes and labor activities around the country. Its membership Although the country Tocqueville saw was an over- peaked at substantially more than half a million mem- whelmingly agricultural society, the industrial rev- bers in the 1880s.33 olution was brewing. The Erie Canal was completed But the various early unionization efforts failed, in 1821, and numerous other canal and road projects often bloodily. Southern planters, northern industri- began to knit the country together. In 1827, the first alists, and Appalachian coal barons alike responded railroad in America opened. Soon, agricultural goods to strikes with lethal violence. The Great Railroad from the western states went east, cotton from the Strike of 1877 ended in over 100 workers being killed. South went abroad, and manufactured goods from In 1885, a Wisconsin militia fired into a crowd of pro- the Atlantic seaboard hubs moved west. A national testors, killing 15. In 1887, perhaps as many as 100 economic system, powered by a massive amount of sugarcane workers were massacred in Thibodaux, human labor (both free and enslaved), had formed. Louisiana, in a strike organized by the Knights of With it came the first stirrings of labor activism. Labor. While union activists also engaged in violence, The earliest known union in America was the Fed- the total casualty count is exceptionally lopsided. eral Society of Journeymen Cordwainers, which One of the only incidents in which workers did not began in 1794. Formed in Philadelphia, it was a har- make up a majority of casualties was the 1886 Hay- binger of the numerous craft and trade unions that market affair, when a bomb thrown at a labor protest would spring up in American cities over the next killed seven police officers. 70 years. Philadelphia in particular would remain a Partly due to successful state-sponsored or state- hotbed of union activism. In 1826, a more general permitted private violence against union organizers, union would form and lobby for a shortened work- unionization rates remained low throughout the 19th ing day. That union would then spin off into the century. There were brief episodes of elevated union- Working Men’s Party and contest several local elec- ization associated with emergent attempts at national tions. Similar parties emerged in New York, Boston, unionization such as the Knights of Labor, but none and elsewhere. None would achieve much success, of them lasted. It wasn’t until the 20th century that but they were signs of things to come. By the 1850s, unions could claim more than one in 20 nonfarm, these local initiatives were going national. The first nonsupervisory workers. Improving communication nationally organized union was the National (later, and transportation helped facilitate the rise in union- International) Typographical Union. ization, as did growing rates of urbanization and

17 BREAD AND CIRCUSES LYMAN STONE

Figure 7. The Rise and Fall of Unions

35%

30%

25%

20%

15%

10%

5% Share of Workers Who Are Union Members

0% 1871 191 3 1941 2011 1850 1857 1864 1878 1885 1892 1899 1906 1920 1927 1934 1948 1955 1962 1969 1976 1983 1990 1997 2018 2004

Consistent Historic Method OECD–Reported Official

Note: “Consistent historic method” uses survey-based estimates of the number of nonfarm workers in primarily nonsupervisory occupa- tions as the population baseline and survey- or directly reported total union membership to estimate unions. This difference in approach accounts for the difference in level and annual changes. Source: Consistent historic method is the author’s calculation from historic sources and membership estimates. OECD-reported union density is the share of nonsupervisory workers identified as being unionized under the OECD’s definition. The OECD’s baseline popula- tion is smaller than the population used in the consistent historic method, explaining most of the difference between them.

industrialization. More politicians saw urban indus- But since the end of World War II, unionization trial workers as an important constituency. rates have plummeted. Massive strikes in 1945 and Simultaneously, a relatively new organization, 1946 provoked public anger at the major unions, lead- the American Federation of Labor (AFL), rose and ing to pro-union candidates losing heavily in 1946. As advanced a new model of labor activism. The AFL a result, in 1947, new laws rolling back pro-union leg- specifically eschewed radical politics, distanced itself islation were enacted.34 from socialism, and indeed avoided creating any kind Since then, the legal status granted to labor unions of “labor party.” Rather, it worked with prolabor poli- has varied over time, but unions have generally been ticians in all major parties, creating a “nonpartisan” or afforded fewer and fewer rights and privileges. This even “capitalist” brand for unions. This made outreach has almost certainly reduced unionization. For exam- to new industries easier, and the AFL grew quickly. ple, academic research has found that state right-to- Over time, increasingly labor-favorable laws along- work laws do reduce unionization rates.35 One of side the burgeoning American manufacturing econ- the only sectors with persistently high unionization omy led to a boom in unionization. As shown in rates is the public sector. Public-sector unions have Figure 7, unionization rates peaked at one-third of all generally been excluded from labor reforms for var- nonfarm, nonsupervisory workers in America in 1945. ious reasons, not least their extraordinary political

18 BREAD AND CIRCUSES LYMAN STONE

Figure 8. The Global Decline in Unionization

United States All-Country Average Countries withComplete-Data Average

40%

20%

0% 1960 1962 1964 1966 1968 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 –20%

–40% Initial Data , Usually 1960 –60% vs .

ent Change in OECD–Estimated Union Density –80% rc Pe

–100%

Source: OECD unionization density rates by country, indexed to 1960. Indexing and data averaging are author’s calculations.

clout due to being the implementers of policy and developed country, even in countries with formerly often the largest employers in many states. extremely high unionization rates, such as Sweden Unionization rates fell for other reasons too. The and Norway. In the average country for which the economy changed. Automation and offshoring have Organisation for Economic Co-operation and Devel- reduced manufacturing in America, opment (OECD) has data, unionization rates are while the increasing skill and complexity involved in 35 percent below where they were in the first data the manufacturing that remains reduced the evident point for each country (1960 in most cases), as shown demand for unionization. Skilled workers with good in Figure 8. “outside options” and hard-to-replace skills don’t These countries have all experienced the same sec- benefit as much from their union dues. The upskilling toral transition as the United States, with traditional of the American economy was bad news for unions. manufacturing being replaced by service-sector and Meanwhile, employment growth in the service sector higher-skilled labor. has been much more robust. There is no prima facie But there’s yet another factor at work: declining reason why fast-food workers couldn’t unionize (and demand. Unions served a vital function advocating for some have), but America’s unions have struggled to workers for many years, providing workers a vehicle to break into the service sector. get shorter working hours, better pay, and better work- This isn’t just a problem for American unions. ing conditions. But today, in many countries, working Unionization rates are in decline in almost every hours, pay, and conditions are regulated by law for all

19 BREAD AND CIRCUSES LYMAN STONE

workers. Much of the direct, employer-based advo- membership to officers certainly contributed to this cacy unions previously engaged in is now addressed small total membership. And indeed, the first Civil War through statute. In some sense, unions are a victim veterans society, the Military Order of the Loyal Legion, of their own success. Relatively few workers face the was also limited to officers and their descendants.37 degree of economic precarity that workers faced in the It was not until the formation in 1866 of the GAR, 1880s or 1920s. Much as the socialization, informa- the first mass veterans’ organization, that veterans’ tion, and entertainment provided by fraternities was societies came into their own.38 The GAR operated replaced by radio and TV, many of the labor protec- at first as a wing of the Republican Party. But after tions provided by unions have been replaced by regu- Reconstruction, the GAR transformed itself into a latory bodies, offshoring, technological advancement, social, fraternal, and mutual-aid society. It ultimately and economic sectoral change. claimed more than 400,000 members at its peak in 1890. Not to be outdone, Confederate veterans formed their own group, though it was never as large The Home Front: Veterans’ Organizations as the GAR. But even as GAR membership began to decline in The complexities of associational life in America can the 1890s as Civil War veterans began to die, a be seen clearly in the history of veterans’ organiza- new group, the Veterans of Foreign Wars (VFW), tions. The first veterans’ organization in America was was formed by soldiers returning from the the Society of the Cincinnati. This group was founded Spanish-American War. The VFW specialized in exclusively for Revolutionary War officers and their helping veterans readjust to life at home after hav- male descendants and immediately attracted vocif- ing served abroad, a new need for American soldiers erous criticism. Elite officers (including Washington in the late 1890s, as previous wars had been fought himself) forming a hereditary club for mutual aid and either on US soil or nearby. In another new develop- support sounded distinctly un-American to many in ment, the VFW was not exclusive to veterans of a spe- the new country who felt it would in time create an cific war. It was envisioned as being for veterans of aristocracy that threatened the republic. any foreign war. The Society of the Cincinnati struggled to survive. Then, after World War I, the American Legion was Facing withering public criticism, high membership formed, specifically to help ensure that veterans were dues, difficulties in organizing together given slow not drawn toward political radicalism, as had occurred transportation and communication, and tight mem- in many other countries after various wars. The Ameri- bership requirements, it struggled to maintain mem- can Legion thus from an early day was deeply concerned bership. By the 1850s, the Society of the Cincinnati with opposition to the revolutionary movements of the had nearly collapsed. early 20th century, especially socialism.39 Yet it did not collapse. Revisions to the member- Since the end of World War I, with variation during ship policy and a renewed interest in veterans’ orga- periods of active conflict, when estimating the num- nizations led to its revival and survival to the present ber of eligible veterans is complicated, membership day. Moreover, a flurry of other societies formed. The in veterans’ organizations as a share of all veterans Aztec Club formed for officers participating in the has not seen any radical changes. There have been Mexican-American War in the late 1840s, and then, local peaks and troughs, but veterans today are only belatedly, in the 1850s, the General Society of the War slightly less likely to be members of a veterans’ orga- of 1812 formed to advocate for the interests of war- nization than at any other time since 1920, as shown time pension beneficiaries.36 in Figure 9. But these organizations were all small and never And yet, veterans’ organizations have seen their managed to include, combined, more than 3 percent membership fall. This is not because veterans today of veterans in their membership. Their limitation of are far less likely to join veterans’ groups, but because

20 BREAD AND CIRCUSES LYMAN STONE

Figure 9. Stable Participation in Veterans’ Groups

40%

35%

30%

25%

20%

15%

of Veterans’ Groups 10%

5%

Share of Veterans Who Are Members 0% 1811 1901 1910 1919 1991 1784 1793 1802 1820 1829 1838 1847 1856 1865 1874 1883 1892 1928 1937 1946 1955 1964 1973 1982 2000 2009

Note: Some veterans may be members of multiple organizations. Source: Membership in veterans’ organizations estimated from archival reports by respective organizations and historic accounts, with membership imputed between missing periods. The underlying population of veterans was directly estimated from decennial census data where available. For periods before 1900 without direct data, the number of veterans was estimated by modeling the relationship between military personnel in a given year, male life expectancy, and veteran populations in later years, for the years for which data are available, extrapolated backward. the number of veterans is in decline. Figure 10 shows them, but rather a change in government policy toward the estimated number of veterans in America over fewer and smaller wars fought by a smaller, more pro- time. fessional military. This policy change makes a lot of In absolute numbers (and even more so as a share sense, but it is yet another case in which changes in a of the total population), the number of veterans is in measure of social capital turn out to be closely related steep decline. Veterans of the last wars of mass mobi- to changes in government policy. Moreover, many lization are aging and dying, while the modern mili- early veterans’ groups formed to advocate for veter- tary operates with less manpower. ans to receive various benefits. But today, extensive Thus, the decline in associational life connected benefits for returning soldiers are provided by statute. to veterans’ organizations is not particularly extreme, This advocacy role still exists, but to a considerable and their overall decline in numbers is not due to a extent, veterans can receive support services without sharp decline in the extent to which veterans join these organizations as well.

21 BREAD AND CIRCUSES LYMAN STONE

Figure 10. Smaller Wars, Fewer Veterans

30,000,000

25,000,000

oups 20,000,000

15,000,000

10,000,000 eterans or Members of Gr V 5,000,000

0 1814 1914 1784 1794 1804 1824 1834 1844 1854 1864 1874 1884 1894 1904 1924 1934 1944 1954 1964 1974 1984 1994 2017 2004

Membership in Veterans Groups Estimated Veterans

Note: Some veterans may be members of multiple organizations. Source: Membership in veterans’ organizations estimated from archival reports by respective organizations and historic accounts, with membership imputed between missing periods. The underlying population of veterans was directly estimated from decennial census data where available. For periods before 1900 without direct data, the number of veterans was estimated by modeling the relationship between military personnel in a given year, male life expectancy, and veteran populations in later years, for the years for which data are available, extrapolated backward.

22 III. Bread and Circuses: Are Sports a Form of Social Capital?

owling Alone concerns itself with social capital in the two states and claims that the prime difference B many forms but is named for a specific kind of between them is slavery. He makes clear that he sees association: a sporting association. Reverence for slavery as the reason that industry was located on the sporting culture is nearly a third rail in American “right bank” of the Ohio River (i.e., Ohio, not Ken- society, and at times, sports have become a political tucky) and why Kentucky was so much less populous football as well, such as during desegregation with than Ohio was. In every case, he contrasts the two Jackie Robinson or in recent years in the case of Colin societies and identifies the adverse impacts of slav- Kaepernick. Sports historians have long noticed that ery on life in Kentucky. Coming to how slavery has American commentators have (in most cases, without affected the two states’ cultural and attitudinal traits, any empirical evidence) claimed that sports make good he writes: citizens. Sports supposedly teach teamwork, respect for constituted authorities, and numerous other val- But the Kentuckian scorns not only labor, but all the ues important to a democracy.40 Antebellum reform- undertakings which labor promotes; as he lives in ers, Theodore Roosevelt, , and innumerable an idle independence, his tastes are those of an idle other Americans have all argued that, in sports, Ameri- man; money loses a portion of its value in his eyes; cans learn important values and build social capital. he covets wealth much less than pleasure and excite- Putnam’s own work began with sports. Before ment; and the energy which his neighbor devotes Bowling Alone, his research found that Italian soccer to gain, turns with him to a passionate love of field clubs were strongly predictive of less corrupt, more sports and military exercises; he delights in violent stable, and more democratic institutions in Italy.41 bodily exertion, he is familiar with the use of arms, The idea that sports are an important part of the and is accustomed from a very early age to expose his fabric of American social life is so prevalent that a life in single combat.44 2013 Atlantic article claimed that Tocqueville would have of course played high school football and that When Tocqueville wrote about American asso- his comments about American associational life must ciational life thriving and supporting democracy, he apply to sports.42 Nothing could be further from the didn’t mean baseball. He saw sports to be a special truth. Tocqueville mentions sports as we think of habit of slaveholding societies that were uniquely them today only twice. In one case, he notes that one likely to produce violent citizens who contributed lit- of America’s most distinctive traits is the strict obser- tle to the nation’s industry. vance of the Sabbath on Sunday, that many states This negative view of sports was not rare. Asked “hung chains across streets with churches on them,” about the best form of exercise, Thomas Jefferson and that almost all recreational activities, including said: sports and fishing, were banned.43 But he makes no value judgment about them in those statements. I advise the gun. While this gives a moderate exercise The more striking case is in a comparison made to the body, it gives boldness, enterprize, and inde- between Kentucky and Ohio. Tocqueville contrasts pendance to the mind. Games played with the ball

23 BREAD AND CIRCUSES LYMAN STONE

Figure 11. The Rise of Sports

0.016%

0.014% e) 0.012%

0.010% American English ds in

r 0.008% Wo 0.006%

0.004% Documents (Not Case-Sensitiv e of Published 0.002% Shar 0.000% 1814 1821 1891 1912 1919 1961 1800 1807 1828 1835 1842 1849 1856 1863 1870 1877 1884 1898 1905 1926 1933 1940 1947 1954 1968 1975 1982 1989 1996 2010 2017 2003

“Team” “Game” “Sport” “Ball” “Athlete” or “Athletic” Specifically Named Sports

Source: Google Ngram.

and others of that nature, are too violent for the body activities and pickup games or aristocratic pastimes and stamp no character on the mind. Let your gun such as horse racing. therefore be the constant companion of your walks.45 Quantifying how the American interest in sports has changed over time is challenging. There are no Skepticism of “games and spectacles” is a long- statistics about the frequency with which Americans standing staple of Western philosophical writing. played ball games in most of the 19th century. None- Early Christians strongly condemned entertainments theless, several indicators are suggestive. Figure 11 like horse racing, for example.46 Moreover, many shows the frequency of a few sporting-related words early Americans were steeped in classical history and in published American texts since 1800. were thus acutely aware of the lure of “bread and cir- As can be seen, the 20th century saw a significant cuses” as tools for distracting the public from their increase in almost all sports-related words. In the better judgment. The Roman poet Juvenal decried past, many of these words even had different mean- “bread and circuses” as superficial appeasement of ings. “Sport” in the past often meant making fun the masses: tools used by the government to distract of someone, a rarer use today, while “game” could the citizenry from its corruption and unaccountabil- refer to wild animal meat, a larger share of Ameri- ity. This is not to say that early Americans didn’t play can diets in the 19th century than today. (Tocque- games (John Adams in particular played “bat and ville used both words in those senses in various ball,” probably the English game of rounders), but it places.) This broadly suggests that sports have risen suggests that organized sports were not an important in prominence in American culture. But it is far from part of American social life in the early republic. Such decisive, given that these words might be used in sports as did exist were either disorganized children’s various ways.

24 BREAD AND CIRCUSES LYMAN STONE

Figure 12. Major Leagues Have Gotten More Popular over Time

700

eople 600

ting Leagues 500

400

our Major Spor 300 ootball Games per 1,000 P 200

100 Attendence at F otal and Collegiate F T 0 191 0 191 5 2019 1890 1895 1900 1905 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2010 2015 2000 2005

Note: Attendance across leagues is not always comparable. Some leagues include tournaments in attendance figures, while others do not. Leagues vary widely in number of games played, arena size, and attendance per game. Nonetheless, consistent treatment over time can be obtained for each league, yielding a total estimate that is a reliable indicator of the overall time trend in commercial sporting event attendance. Source: Historical Statistics of the United States and league-specific statistics where available. Includes the NFL and its precursor leagues, the NHL, all MLB and current precursor leagues, all NCAA football, NBA and precursor leagues, and all Major League Soccer leagues.

However, there is other evidence as well. Atten- for which no data exist, and thus the rise in atten- dance statistics are available for the “big four” major dance is likely overstated. national sports leagues and their precursor orga- But the corresponding increase in sports-related nizations back to the early 20th and, in the case of words in published documents suggests that the baseball, late 19th centuries. Estimates of National basic trend is correct, not least because still other Collegiate Athletic Association (NCAA) football atten- data sources support this trend as well. Reliable dance are also available across a long time horizon. data of high school sports are available back to the Figure 12 shows estimated attendance at National 1970s, and before that, extensive historical accounts Football League, National Hockey League, National of youth sports with general estimates of their scale Basketball Association, Major League Baseball, Major can be used to assess the range of youth participa- League Soccer, and NCAA football games from 1890 tion in sports leagues. This measure would probably to today per 1,000 Americans. not be influenced by nationalized transportation and Major league and NCAA football attendance rose communication, because the vast majority of youth persistently from 1890 until the mid-2000s. This sporting events are exclusively local affairs, with only is partly a product of nationalized culture, due to a small, elite group of youth athletes competing on a improved transportation, communication, and pop- national scale even today. Figure 13 shows that youth ularization driven by radio and TV. In other words, participation in organized sports has also grown dra- some of the increased attendance at national sporting matically over time. events was probably just replacing prior local leagues

25 BREAD AND CIRCUSES LYMAN STONE

Figure 13. More and More Kids Are Playing Sports

35%

30%

25%

20%

Athletes as a Sha re 15%

of High School–Age Kids 10% High School

5%

0% 191 0 191 6 1880 1886 1892 1898 1904 1922 1928 1934 1940 1946 1952 1958 1964 1970 1976 1982 1988 1994 2012 2018 2000 2006

Note: Only sports occurring in an organized context were included; pickup games or informal sports activities are not included. Source: Estimates of high school–age children were taken from IPUMS USA and IPUMS Current Population Survey–derived counts of people age 13–18, with interpolation between census years before Current Population Survey availability. The number of high school athletes 1971–2018 was taken from the High School Athletics Participation Survey. Pre-1971 figures were estimated based on historical accounts of league foundings and tournament sizes and frequencies.

In the late 19th century, an extremely small share compared to other countries. While data for other of high school–age kids participated in organized countries are not available for 1965, the gap between sports leagues. But that share steadily increased over the US and other countries has grown over time, the 20th century and then leaped upward in the 1970s and if comparable data for the 1960s did exist, based after Title IX led to numerous new athletic opportu- on the slope of increase observed in other coun- nities for high school girls. Since then, the share of tries, there likely would have been no gap at all. high schoolers participating in sports has continued That is, America’s unique sports obsession isn’t a its long rise. long-standing feature of the republic. It developed Moreover, American adults play more sports and in the 1960s at the earliest. exercise more than in the past, and American men American women are more likely to participate spend more time exercising or playing sports than in sports than their international peers are, though men did in any other country for which data are the gap is smaller. As a result, the gender sports gap available. Figures 14–16 show the number of min- is larger in America than in any other country except utes per week men and women age 18–59 reported Spain. American men are exceptionally sports focused. playing sports or doing exercise in time use surveys Some of this excess may be due to idiosyncrasies in in the US and other countries for which comparable reporting. When outdoor activities such as walking or data are available. gardening are included, American men become much Americans are as or more sporting as ever, and less exceptional, and thus it’s important not to read we are unusually interested in sports and exercise too much into one specific data point. But even with

26 BREAD AND CIRCUSES LYMAN STONE

Figure 14. American Men Spend Much More Time on Sports Than Men Elsewhere Do

200 180 160 140 120 100 80 60 of Sports and Exercise 40 Estimated Minutes per Week 20 0 1971 1981 1991 2011 1965 1967 1969 1973 1975 1977 1979 1983 1985 1987 1989 1993 1995 1997 1999 2001 2013 2015 2017 2019 2003 2005 2007 2009 Austria Bulgaria Canada Finland France Hungary Israel Italy South Korea Netherlands Spain United Kingdown US South Africa

Source: International data points and US values before 2000 were queried from the IPUMS Multinational Time Use Study (MTUS). Values after 2000 were queried from the IPUMS American Time Use Survey (ATUS). In both cases, the “sports and exercise” time use category was used for men and women age 18–59.

Figure 15. American Women Spend Somewhat More Time on Sports Than Women Elsewhere Do

200 180 160 eek

W 140 cise

er 120 100 80 ts and Ex 60 40 of Spor

Estimated Minutes per 20 0 2011 1971 1981 1991 2013 2015 2017 2019 1965 1967 1969 1973 1975 1977 1979 1983 1985 1987 1989 1993 1995 1997 1999 2001 2003 2005 2007 2009 Austria Bulgaria Canada Finland France Hungary Israel Italy South Korea Netherlands Spain United Kingdown US South Africa

Source: International data points and US values before 2000 were queried from IPUMS MTUS. Values after 2000 were queried from the IPUMS ATUS. In both cases, the “sports and exercise” time use category was used for men and women age 18–59.

27 BREAD AND CIRCUSES LYMAN STONE

Figure 16. The Gender Gap in Sports and Exercise Is Bigger in America Than in Other Countries

120

100

eek 80 W cise er 60

ts and Ex 40

of Spor 20 Estimated Minutes per 0 2011 1971 1981 1991 2013 2015 2017 2019 1965 1967 1969 1973 1975 1977 1979 1983 1985 1987 1989 1993 1995 1997 1999 2001 –20 2003 2005 2007 2009 Austria Bulgaria Canada Finland France Hungary Israel Italy South Korea Netherlands Spain United Kingdown US South Africa

Source: Difference in estimates provided in Figures 15 and 16.

the broadest possible treatment, Americans spend coincides with the decline in American membership far more time on sports and exercise today than in in other organizations. This coincidence shows up in the mid-20th century and more than average among other data as well. The General Social Survey (GSS) countries for which data exist. has periodically surveyed Americans about their The rise of sports can also be seen in national eco- membership in various kinds of groups, including nomic statistics. Membership in sports clubs and sports clubs. Comparing sports clubs to other kinds money spent at spectator sports has risen as a share of organizations discussed thus far is illustrative. (See of all personal consumption spending from 0.4 per- Figure 18.) cent in 1959 to 0.6 percent today, as shown in Figure 17, While all these organizations have seen a decline even as sports-related employment has risen from from 1994 to 2004, the latest data available, sports under 100 sports workers per million adults in the club membership is about the same in 2004 as it was mid-19th century to 1,800 today. As recently as the in 1974, the earliest data available. But membership in 1990s, there were fewer than 1,000 sports workers per fraternal societies, church groups, veterans’ groups, million American adults. and labor unions are all down considerably. The current rise in sports’ economic popularity Putnam cites the decline of the American Bowling is even more recent than other metrics; it begins in Association as a case of declining social capital. But just the 1980s. There is nothing time-honored about the extraordinary rise of sports in the 20th century the current scale and prominence of sports in Amer- suggests that this decline is probably the exception, ican society. not the rule. In reality, sports have grown enormously And indeed, far from building social capital, this in popularity, concurrent with a widely decried increase in sports and exercise activity actually decline in associational life.

28 BREAD AND CIRCUSES LYMAN STONE

Figure 17. Americans Are Spending More Money on Sports

1.4% 2000 1800 1.2% ts Clubs, 1600 Adults ts and

Spor 1.0% 1400 ts, 1200 0.8% 1000

0.6% ees per Million y

ation Equipment 800

0.4% 600 Rec re ting Facilities , and Spor 400 ts Emplo 0.2% Spor 200 Spor

Spending on Spectator Spor 0.0% 0 191 0 191 6 1850 1856 1862 1868 1874 1880 1886 1892 1898 1904 1922 1928 1934 1940 1946 1952 1958 1964 1970 1976 1982 1988 1994 2012 2018 2000 2006 Spending on Spectator Sports Sports Employment

Note: Spending was divided by US Bureau of Economic Analysis (BEA)–estimated personal income. Source: Spending on spectator sports from BEA Personal Consumption Expenditure data for the categories of spectator sports, mem- bership clubs and participant sports centers, sporting equipment, supplies, guns, and ammunition. Employment estimates are taken from queries of IPUMS USA’s repository of Census and American Community Survey data. For both sources, IPUMS USA’s assigned 1950 occupational codes for athletes, sports instructors, and other related jobs were used to identify sports employees. The number of adults used as a baseline was also queried in IPUMS USA.

Figure 18. Sports Clubs Have Been Resilient

6%

4%

2%

0% 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 –2%

–4%

–6%

–8%

–10% Change in Share of Adults Who Are Members of These Groups vs. 197 4 –12%

–14%

Church Group Sports Club Labor Union Veteran Group Fraternal Group

Source: General Social Survey.

29 BREAD AND CIRCUSES LYMAN STONE

And indeed, the supposed ability of sports to cre- Likewise, a German study found that adoles- ate useful social capital is empirically contested. Aca- cent sports clubs enhanced later-in-life community demic research into the ethnic composition of sports involvement, but only for those who participated in clubs in the Netherlands found that effects on migrant formal sports clubs. What mattered wasn’t playing and native attitudes toward each other were hardly soccer per se, but being part of an organization.54 An changed by exposure to one another and indeed that experimental intervention in a group of previously migrants in co-ethnic clubs had the lowest social trust. inactive women found that increased activity led to The supposed universal language of sports turned greater community participation but that not all sports out not to be associated with any improvement in would have this effect. Specific,high-collaboration across-group trust.47 A study of Norwegian civil soci- team sports (such as basketball or soccer) might be ety found that members of sports clubs were uniquely more effective than other sports (such as boxing or unlikely to have other organizational memberships, running).55 Another experimental intervention, this suggesting that sporting activities were not very effec- time involving cricket in India, found that sports par- tive at embedding participants in a wider social net- ticipation improved cross-caste attitudes, but it con- work (especially compared to other activities).48 tained no study of behaviors off the field, away from A study of sports clubs in Japan found that the experiment.56 improved social capital only occurred in sports clubs A study finding that sports improve behavior at with particularly advantageous physical locations.49 schools contained an important caveat: Sports only An experiment involving allocating Iraqis of dif- partly offset the worsening behavior patterns created ferent faiths to soccer leagues found that individu- by the large school sizes necessary to maintain large als who played on teams with people of other faiths sports programs.57 That is to say, while sports might reported more positive attitudes toward groups but improve social capital for kids, going to a school big that these effects “did not generalize tonon-soccer enough to support sports programs may have an contexts.”50 Reviewing the literature up to that point, opposite and larger effect. one researcher described the supposed relation- Most strikingly of all, a study of online gamers ship between sports and social capital as “ill-defined found similar effects: Online gaming guilds can cre- interventions with hard to follow outcomes.”51 And ate similar kinds of social capital as sports clubs.58 of course, there is ample research on the connection But while the researchers in this study thought this between sporting events and excessive alcohol con- conclusion to be bullish for online gaming, it may be sumption, pointing to an often poorly managed side seen to be bearish for sports generally. Football might effect of sports.52 be only a little bit better at creating social trust than Sports has its defenders, of course. An experimen- World of Warcraft is. tal study of a sports-related intervention in Poland Finally, Tocqueville’s association of southern sport found that, among women over age 41 with no pre- mindedness with violence (“field sports and military vious sporting activity, there was some increase in exercises”) may have been eminently rational. Atten- membership in other organizations, voluntary activ- dance at major league sporting events is extremely ities, and other reasonable measures of social cap- strongly correlated with changes in government ital. Older men also had some benefit.53 But among spending over time and, especially, with major wars, those under age 41, effects were small and inconsis- as shown in Figure 19. tent across groups. Furthermore, the social capital– The Spanish-American War, World War I, and related benefits of sports were closely connected World War II are all clearly associated with increases to health-related benefits. Making people health- in government spending followed by increases ier makes them more likely to get out of the house in sports attendance. This is consistent with aca- and participate in society, but competitive sports are demic research, which finds that wars create sharp hardly the only path to a healthier life. increases in certain forms of social capital, especially

30 BREAD AND CIRCUSES LYMAN STONE

Figure 19. Wartime Spending Hikes Boost Spectator Sports

700 60%

600 50%

500 40% 400 30% 300 20% 200

100 10% Government Spending as a Percent of GDP

and Collegiate Football Games per 1,000 Peopl e 0 0% Total Attendence at Four Major Sporting Leagues 1914 1890 1896 1902 1908 1920 1926 1932 1938 1944 1950 1956 1962 1968 1974 1980 1986 1992 1998 2010 2016 2004

Spectator Sports Government Spending as a Percentage of GDP

Source: Sports attendance as shown in Figure 12. Government spending as a share of gross domestic product from the US Bureau of Economic Analysis.

“parochial” or what might be called “tribalistic” social international cooperation and camaraderie, the rise of capital: rooting against outsiders.59 Recessionary the Olympics in the late 19th and early 20th centuries spending increases, such as during the Great Depres- was itself occurring in an environment of explosively sion and Great Recession, don’t seem to have the increasing nationalism, most famously in Nazi Ger- same effects. But the fact that the increase in sports many. Today, one need only watch how countries com- attendance stopped soon after government spending pete to host the Olympics (or watch how authoritarian reached its modern plateau is striking. countries such as China use the opening ceremonies to Bread and circuses have always gone together. As showcase ideas of national greatness) to see the real- the state’s role in American society has ballooned, ity of international sports: They provide opportunities especially through the increasing militarization of for costly public displays and reaffirmations of national American life, sports have risen with it. The society difference and competition, not just avenues for coop- Tocqueville toured had little participation in sports eration. (And, practically speaking, international sports and virtually no standing army. The spread of mass, leagues are riven with massive corruption and graft, as popular sports in other countries has also corre- in the case of Qatar’s 2022 FIFA World Cup bid; such sponded with the rise of more interventionist, wel- social capital as they build is more like organized crime farist, and nationalist states. than philanthropy.) The connection between sports and nationalism is Academic research exploiting the qualification particularly striking. Opinion polls from Gallup show dynamics of soccer tournaments has confirmed this that the consistently most-popular sport in Amer- effect. Looking at African countries, “barely” qualifying ica is not football, baseball, or basketball, but simply for a major international soccer tournament had vari- “Olympic sports” (Figure 20). While rhetorically about ous effects that researchers regard as positive: reduced

31 BREAD AND CIRCUSES LYMAN STONE

Figure 20. More Americans Are Fans of the Olympics Than Are Fans of Sports at All

80

70 e Fans

Ar 60

Who 50

40 Adults en Spo rt

30 of Giv

American 20

e of 10 Shar 0 7 2011 2001 2010 2012 2013 2014 2015 2016 2017 2018 2019 200 2000 2002 2003 2004 2005 2006 2008 2009

Consider Self a Sports Fan (Yes) Professional Baseball Professional Football Professional Basketball Professional Ice Hockey College Football College Basketball Auto Racing Professional Soccer Olympic Sports Professional Golf Professional Tennis Figure Skating Professional Wrestling

Source: Gallup.

ethnic identification, reduced violence, and increased But this outcome depends on the context. If the national identification.60 In a continental context division in a country is between a party that values where most violence is interethnic rather than inter- national pride, patriotism, and American exceptional- state, reducing ethnic identity salience and increasing ism and another party that values international coop- national identity salience may indeed reduce conflict. eration, pluralism, and a reckoning with America’s But in continents with long-established states, and past sins, sports may deepen the divide. If a society is where political discourses are strongly polarized across divided by educationally associated classes and sports questions of nationalistic versus pluralistic modes of have strong associations with educational institu- identity, the function of sports in promoting national tions, they may make interclass conflicts worse. If a identity might have bad effects. society is increasingly divided by geographic regions, In other words, national sports destroy one form geographically identified sports may not facilitate of social affiliation (ethnic, community, local, or reli- cooperation, but deepen a sense of tribal difference. gious identity) and elevate another (national iden- Surveys in the United States confirm the national- tity). If the affiliation being suppressed is indeed a istic functions of sports. GSS questions in 1994, 2004, source of violence and instability, this may be good. and 2014 all found that 80 percent or more of Ameri- Racial integration in the US, especially in interna- cans say they are proud of America’s performance in tional sports giving a majority-white country the sports. In most years, Americans were more proud of opportunity to celebrate from and feel a kinship with their sporting achievements than the country’s artis- the accomplishments of black athletes, probably tic or literary achievements, and pride in the country’s helped alleviate racial tensions. history or its armed forces only reached a bit higher at

32 BREAD AND CIRCUSES LYMAN STONE

85–90 percent. Meanwhile, about only 50 or 60 per- in 1987, the GSS asked members of 15 different kinds cent of Americans are proud of the country’s record of groups if that group “tries to solve individual or of equality and fairness (exact value depends on the community problems.” For fraternal groups, 84 per- year). National pride is apparently more bolstered by cent of members said yes. For professional societ- basketball than by voting rights. Given that Americans ies, 69 percent. For labor unions, 66 percent. Only spend more time on sports than people in most other 36 percent of members of sports clubs said their clubs countries do and this relationship between sports and tried to do anything to solve individual or community national pride, it is perhaps unsurprising that Ameri- problems, the lowest response of any group category cans are also, according to surveys, among the most besides “hobby clubs.” patriotic people in the world.61 This is striking, because there is a mythology of The link between sports and the welfare state is sports as a problem-solving activity. Movies about also well demonstrated in prior research. Academic people overcoming life hurdles or community prob- research explicitly links sports participation and the lems through sports are a staple of American enter- interventionist role of the state. Research on Danish tainment. And yet when push comes to shove, sports sports clubs specifically identifies such clubs as part club members acknowledge that their associations of the welfare state.62 A study of Canadian sports cul- don’t actually do very much. As much as the defend- ture outlined the “broad impact” of the welfare state ers of “sports as social capital” might want to argue on sports culture and highlighted various direct inter- that sports have large beneficial effects, most people ventions by the Canadian government to support who play sports basically just like to play sports and and promote sports.63 A study of six Western Euro- don’t look to receive any greater benefit. pean countries outlines the specific links between the While sports may build a kind of national-scale welfare state and the sporting system in each coun- social capital (i.e., nationalism), in American and try.64 Modern governments actively promote sports, global history the record on nationalism and social not only for nationalistic reasons but also as benign, capital is at best spotty. For every case in which politically nonthreatening entertainments to accom- nationalist sentiment was annealing (say, the US pany the welfare state. Bread and circuses is not only hockey victory over the USSR in the 1980 Olympics), an ancient Roman phenomenon but also the actual there is another in which sports nationalism was a function of modern sports. harbinger of terrible nationalist violence to come In sum, the relationship between sports and social (such as the 1936 Olympics held in Germany on the capital is complicated. While being a member of a eve of the Holocaust or the 2008 Olympics just a few sports club probably does help build social capital, it years before China’s concentration camps in Xinji- is far less effective than almost any other form of asso- ang began to be set up). ciational life. More broadly, sports culture as it exists For every team brought closer together by mutual today is a relatively recent arrival, truly appearing as a struggle, there is a team resentful of its defeat. The distinctively American trait since the 1960s. This focus idea that sports would have an annealing effect on on sports is flatly contrary to the vision of republi- American society is challenged by the fact that some can citizenship articulated by many of the Founding of the most contentious culture war–style debates Fathers, who had inherited a critique of “games” from of today revolve around sports: who can play, what Christian and classical thinkers of the past. sports mean, what speech is permissible around That the rise in sports culture has coincided with sporting events, and so forth. Whatever uniting effect a weakening in other forms of social capital and an sports might have in some other context doesn’t hold expansion of the state is suggestive as well. Sports may in 21st-century America. be more about the government’s desire to direct pub- But the key point is not that sports are bad. The lic sentiment toward trivialities than building effec- aristocratic sensibilities of a Jefferson or Tocque- tive associations that are useful for society. Strikingly, ville viewing the relatively plebeian pastimes of the

33 BREAD AND CIRCUSES LYMAN STONE

Figure 21. Youth Sports Are Rising, While Other Groups Fall

20%

18%

16%

14%

12%

10%

8%

6% Share of Kids Age 5–1 8 4% Participating in Each Activity Group 2%

0% 1910 1915 1880 1885 1890 1895 1900 1905 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2010 2015 2000 2005

Sports Scouts 4-H and Junior Chamber of Commerce

Source: Youth sports participation as in Figure 13, but with a denominator for people age 5–18 instead of 13–18. Data on Scouting, 4-H, and Junior Chamber of Commerce membership as in Figure 2.

common folk, such as ball games, with disdain need Football vs. Boy Scouts: The Trade-Offs in not be given any special credence. Sports are just not Youth Lives especially likely to create benefits for sporting societ- ies. The great attention and energy devoted to sports High school kids are playing more sports. But that may yield a great deal of enjoyment for participants, doesn’t mean they’re more invested in associational but they are no more likely to yield wider social ben- life generally. Indeed, many youth organizations have efits than any other form of personal entertainment struggled in recent years. Figure 21 shows the share of will, with others or alone. kids participating in Boy Scouts, Girl Scouts, 4-H, the To the extent sports compete with other forms Junior Chamber of Commerce, and high school sports. of personal entertainment, this is no knock against As can be seen, the “golden age” of youth organi- them. But to the extent that sports compete with zations came in the 1960s and early 1970s. Since then, other, possibly very socially beneficial activities, while sports have continued to rise, scouting and 4-H this lack of wider social benefit could be worrisome. have declined. There is some direct research showing that sports These data stretch over a long time horizon but compete with other forms of associations, such as cover only a limited set of organizations. Luckily, the church.65 But to truly demonstrate the fundamen- Survey of Income and Program Participation (SIPP) tal competition between sports and other, possibly has periodically asked kids about their extracurricular more social capital–building forms of associational activities as well, placing them in three major groups: life, we have to look at the role of sports in American clubs, lessons, and sports. Figure 22 shows that clubs childhood. have fallen significantly since 1998 (the earliest data

34 BREAD AND CIRCUSES LYMAN STONE

Figure 22. Sport Are Rising, and Clubs Are Falling

45%

40%

35%

30%

25%

20%

15%

10%

5% Share of Kids Age 6–17 Participating

0% in Identified Kinds of Extracurricular Activities 7 2011 1998 1999 2001 2010 2012 2013 2014 200 2000 2002 2003 2004 2005 2006 2008 2009 Sports Clubs Lessons

Source: Brian Knop and Julie Siebens, “A Child’s Day: Parental Interaction, School Engagement, and Extracurricular Activities: 2014,”US Census Bureau, November 2018, https://www.census.gov/content/dam/Census/library/publications/2018/demo/P70-159.pdf.

available in the SIPP), while lessons have risen mod- It may be surprising that sports have experienced estly and sports have risen by a lot. such a boom while other extracurriculars have not. Notably, all activities declined in 2006–09 and This is not the narrative usually told about children’s 2009–11, perhaps because participation in these activ- lives today. Rather, one is more likely to hear parents ities is often costly for families and these years saw complaining that kids don’t do Little League anymore a large recession and slow recovery. But the trend in and instead get after-school tutoring or play video both lessons and sports during non-recession years games. Public discourse is dominated by publicized is upward, with sports having a much more dra- conflicts among free-range“ parents,” “helicopter par- matic upward trend. Sports are becoming more and ents,” and “tiger parents,” each acting as a proxy for more popular among kids even as clubs are declining attitudes toward extra lessons, extracurriculars, and in popularity. Other organizations are affected too: playing outside. But these narratives are overblown. Religious periodicals are packed with discussions of American kids’ evenings aren’t gobbled up by more how sports practices are increasingly eating away at and more lessons and fancy extracurriculars. Their church youth groups.66 evenings are being eaten by soccer and cross-country. In the case of children’s activities, the trade-off is Sports claim a larger share of children’s lives than clear. There isn’t a serious dispute that sports have at any other point in documented American history, displaced other forms of associational life, because it’s while other extracurriculars are in speedy decline. obvious: Kids have a limited set of time, their parents Kids are forgoing church youth group, service and vol- make a lot of their scheduling choices, and time spent unteer activities, competitive speech and debate, and on sports has risen directly at the expense of other any number of other activities to play more sports. activities, including activities that might have helped Meanwhile, the much-heralded rise of intensive par- kids acquire more valuable skills or relationships. enting has not led to a large increase in after-school

35 BREAD AND CIRCUSES LYMAN STONE

Figure 23. Public Education Is Dominating Childhood

160

140

Attending 120 ar

Ye 100

80 Age 5–18 Spent 60 ys Kids Public Schools per 40

rage Da 20 Ave 0 191 2 191 8 1840 1846 1852 1858 1864 1870 1876 1882 1888 1894 1900 1906 1924 1930 1936 1942 1948 1954 1960 1966 1972 1978 1984 1990 1996 2014 2002 2008

Source: Author’s calculations based on historic educational statistics from Thomas D. Snyder, ed., 120 Years of American Education: A Statistical Portrait, National Center for Education Statistics, January 1993, https://nces.ed.gov/pubs93/93442.pdf; and updated annual data from the National Center for Education Statistics.

lessons and tutorials. To the extent parents are push- schools. Children are spending more time than ever at ing their kids to overcommit their time, it’s due to school, especially public schools, as Figure 23 shows. ever-escalating commitments to sports, not orches- Much of this is driven by increasing rates of school tra, French lessons, Future Farmers of America (FFA), enrollment for kids. This is an important form of asso- or math club. ciational life. Schools can serve as hubs of social life, What is happening to children is happening to creating numerous avenues for children and adults all of America. Community organizations are being to engage with their communities and develop useful squeezed out by sports, even as taxpayer dollars are nonacademic skills. Thus, school expansion can serve used to subsidize an ever-increasing number of glam- to boost other associations by providing them with a orous sporting arenas or expensive high school sports network of connections and often physical spaces to complexes.67 While there’s nothing wrong with peo- meet and conduct their activities. ple of any age playing sports, the consequences for a But schools, and especially public schools as they community of a stripped civic life can be dire in the exist today, can also impede certain forms of commu- long run. nity life. While schools facilitate many different extra- curricular activities, there are unique reasons why increasing exposure to public schools might work in School Years: Training Ground for Adult favor of sports and against other forms of associa- Community Life tional life. First, in recent decades, debates about the role of Drawing the precise link between the growth of the religion and morality in public schools have led to administrative state, sports, and social capital can be many contentious court cases. Public schools today challenging. But one factor binds them all together: shy away from appearing to make any explicit

36 BREAD AND CIRCUSES LYMAN STONE

statements about anything other than lowest- promote a lowest-common-denominator public value: common-denominator moral questions. This may physical health. be appropriate for taxpayer-funded institutions, but Most club-based activities will simply struggle many clubs and organizations exist with an explicit to hold a child’s attention for as long as a sport can, ethos or creed. Most clubs try to solve individual and because they are less physically active. But while this community problems, which means they implicitly makes sports an easier sell for increasingly busy par- take a side on numerous social questions. The explic- ents, it may not make sports a better program. Chil- itly creedal character of many organizations (such dren often benefit from doing things they don’t enjoy, as the Girl Scouts, which believes in empowering such as doing math problem sets or practicing the women—a good thing, of course, but nonetheless a piano. creedal statement) is a poor fit for overtlynon-creedal Finally, despite rhetoric to the contrary, sports do schooling environments. not have an ethos in the way that, say, the Kiwanis Religious clubs and organizations face this chal- Club or the Reserve Officer Training Corps do. Very lenge most directly, but even nonreligious organi- few athletes will be expelled from their teams for zations can face it passively. Any strongly worded insufficiently manifesting the creed of the group. Very statement about a club’s aims, values, or beliefs is few sports clubs at any level have strict disciplinary almost certain to alienate some subgroup of the stu- standards off the field. Teamwork is rhetorically dent population, and thus even if administrators per- important, but in local sports, most participating ath- mit the group to exist, they are unlikely to advance it letes are amateurs, leaving the small number of highly as a model for all students to emulate. disciplined players to dominate as local “hometown Second, schools establish hierarchies that commu- heroes.” And while sports may encourage coopera- nicate to children what things are valued, honorable, tion, as noted earlier, the evidence that team sports and to be emulated. It is probably beyond dispute that actually break down boundaries and increase the almost all schools that have sports teams treat vic- social trust necessary to achieve important social torious student-athletes as extremely valued. Many tasks is contested and ambiguous. high schools, and of course colleges, actively recruit The only ethos ultimately represented in sports is student-athletes in a way that is comparatively rare the jersey itself, which is also why sports have thrived for students who excel at speech and debate or who as schools claim more of children’s lives. They are are prizewinning FFA members. That school admin- emblematic of the school community in a way 4-H istrators and teachers (and booster clubs that donate never will be, if for no other reason than that play- generously to school sports programs) communicate ers literally wear the schools’ emblems. Students in to students that athletes are especially highly val- FFA wear FFA jackets. Boy Scouts wear Boy Scout ued almost certainly influences students’ choice of uniforms. Kids doing Mathletes wear whatever their activities. competition’s dress code allows. But student-athletes Third, many schools use extracurricular activi- wear the standards and colors of their schools, and ties to achieve specific, nonacademic goals, such as they depend for their popularity on the school’s exis- extending the length of the school day for parents with tence. If the school goes away, the Girl Scouts will just long working hours or encouraging student physical meet somewhere else. The basketball team, however, health. Whereas an academic team might be seen as will vanish. a fun way for academically high-performing students Thus, as schools have claimed a growing share to compete, the extensive time commitment involved of American children’s lives, those activities most in sports may be seen as valuable in itself by both closely connected to school-based identity have cor- school administrators and parents. (Baseball practice respondingly grown in popularity, while activities that may be cheaper than a babysitter, and baseball parents could connect children across school lines have suf- donate to the booster club.) Sports are also a way to fered. Bonding capital inside a school community is

37 BREAD AND CIRCUSES LYMAN STONE

Figure 24. Young Adulthood Is Increasingly School Focused

35%

30%

25%

20%

15%

10% Enrolled in School

5%

Share of Adults Age 18–35 Who Are 0% 1861 1910 1917 1931 1840 1847 1854 1868 1875 1882 1889 1896 1903 1924 1938 1945 1952 1959 1966 1973 1980 1987 1994 2001 2015 2008

Source: Query of IPUMS USA decennial census and American Community Survey data by enrollment status; and IPUMS National Histor- ical Geographic Information System school enrollees in 1840.

perhaps enhanced (although students not involved in did, and the immediate aftermath of World War II, in athletics might dispute the idea that the glorification 1950, saw enrollments decline.68 of school sports enhances the quality of school com- Regardless of what caused the increase in student munity life), but bridging capital beyond it is incon- enrollments initially, today a large share of young trovertibly lost. adults spend many years in school. This is, again, an additional form of social capital. Universities form vital networks for their students, spin off import- Extended Adolescence and College ant community organizations in their surrounding Sports regions, and teach useful nonacademic and academic skills. There is no denying that the increase in univer- The story of school-based identities does not end in sity enrollment represents increasing participation in childhood. Whereas in the past Americans completed associational life in America. their transition to adulthood around the time of their And yet it is qualitatively different. Universities high school graduation, today the “learner stage” advertise themselves to students primarily in how of life extends further. Figure 24 shows the share of their services will benefit their lives individually, adults age 18–35 who are enrolled in school. not how students can come and join a movement The GI Bill of Rights jump-started college educa- for the benefit of the community (although univer- tion in America and led to a massive increase in the sities are increasingly trying to market themselves in share of American young adults enrolled in school. But these terms). Most parents sending their kids to col- while the GI Bill might explain some of the increase, lege hope they get good jobs on the other end; if they the reality is that nonveterans (and women) saw also occasionally donate to alumni charitable proj- nearly as large of increases in enrollment as veterans ects, that’s neither here nor there. The institution’s

38 BREAD AND CIRCUSES LYMAN STONE

purpose is basically internalized economic returns, sports for many American institutions. If football is not service to wider society or community for its political, then nonpartisan institutions will eventually own sake. abandon football. As a result, the primary form of community expres- Whether this ought to be the case is not the point; sion at many colleges (especially large public univer- the point is simply that the deeply conflicted and hos- sities, though to a much lesser extent at small, often tile responses by many fans to attempts to use profes- ideologically defined private universities) is athletic sional sports as a platform to advocate for social justice competition. College sports are an enormous busi- reveal the weakness of sports as a form of social cap- ness, and their related revenues and donations are ital. Extremely modest efforts by individual athletes often so lucrative that they can subsidize other uni- to make personal political statements are strongly versity activities. But what makes college sports so opposed by large numbers of fans, not just for the con- easy to rally around is that they are in some sense tent of the message, but on the grounds that moral or uncontroversial. A student or alumnus of any polit- political messages have no place in sports at all. Strik- ical persuasion can cheer for their team to beat the ingly, while the past few years have seen more politi- other team. But it is this very lack of an ethos that cal statements by athletes, attendance at major league makes sports, as noted above, a less useful form of games has fallen, and youth sports participation even associational life. Sports are broadly appealing but ticked slightly downward. As sports begin to politicize, lack a strong ability to persuade fans and participants some Americans are turning away from them. to make unrelated costly commitments. None of this implies that sports are bad. The This illuminates current conflicts related to Amer- benefits of sports for physical health are widely ican sports. The appeal of sports to many Ameri- demonstrated, and sports-based interventions have cans, and especially to the educational institutions sometimes been shown to be useful in achieving vari- that undergird and promote sports as a crucial part ous socially desirable outcomes, including, as just one of American life, is precisely that they have no overt example, reducing crime rates.69 It is unclear, how- or obvious meaning or significance. Their vapidity ever, if these interventions work because they focus is their appeal. They are just a group of athletes on on sports or because they provide any kind of helpful some grass with a ball, demonstrating dexterity. Polit- resource to poorer communities. ical, ethnic, and religious differences can suppos- However, to understand how associational life edly be left at home, allowing the field to be a neutral has changed in America, it is vital to understand that and even ground for honest competition. Because of the major change that has occurred is not a decline this conception of what sports are, any attempt to in all forms of associational life, but a replacement use sports to achieve or speak about a political proj- of clubs, churches, unions, and other societies with ect (such as Black Lives Matter) “ruins” sports for state-provided services and regulations alongside many Americans and jeopardizes the neutral status of value-neutral entertainment. Bread and circuses.

39 IV. Measuring Associational Life

he broad change in American associational life While it does not have a category specifically for Taway from clubs, churches, and even voting community development organizations, those orga- toward sports is well-documented. It may decrease the nizations would be counted. Likewise, the National benefits of associational life for society. But focusing on Center for Charitable Statistics (NCCS) has collected specific kinds of associations can sometimes obscure and categorized data on charitable organizations and broader trends. The next section of this report looks nonprofits reaching back several years. at four different kinds of data about associational life These two datasets track different kinds of groups. across long time spans: administrative data on legally The NCCS data track all legal nonprofits (including registered organizations, surveys of time use, the GSS, nonprofit hospitals and universities) that report to and Google Ngram data related to associational life. the IRS in a given year, while the CBP data track iden- tifiable public service and charitable organizations with a physical presence in a given county. The two Paper Trails: Measuring Organizations sources give different but complementary glimpses of over Time charitable, philanthropic, and community organiza- tions, as shown in Figure 25. It may seem unnecessarily academic to bemoan the Both sources show that community and nonprofit role of sports in displacing marginally more effec- organizations were growing in prevalence at least as tive forms of associational life. But the exact shape of of the 1990s. But in the 2000s, the CBP data show a associational life may matter a great deal for quality decline, while the NCCS data show just a slight slack- of life. ening in growth. The best example of this comes from a study of CBP data can be broken into seven major catego- community nonprofit organizations and crime. Cit- ries: business associations; professional-membership ies that have had larger increases in the number of organizations; labor organizations (such as unions); substance abuse, community development, youth- civic, social, and fraternal associations; political orga- oriented, crime prevention, and workforce devel- nizations; religious organizations; charitable organi- opment nonprofit organizations have seen bigger zations; and other or uncategorized organizations. decreases in crime over the past few decades.70 The Figure 26 shows the number of organizations in each presence of nonprofit organizations is also predictive subset per million people in America at that time. of better-performing labor markets.71 In other words, However, religious organizations (not the primary the number of explicitly community-oriented organi- focus of this report) dominate the figure; thus Figure zations trying to solve community problems turns out 27 shows the same figures, but with religious organi- to be important. Nebulous social capital untethered zations excluded. from organizations with explicit philanthropic mis- Religious organizations rose steadily until the sions just doesn’t do the trick. 1990s, when they peaked and began to decline. Mean- Since the 1940s, the US government has conducted while, labor unions have fallen from the second-most- a regular survey known as the County Business Pat- prominent category of organization to the second terns (CBP) survey. This survey collects detailed to last. Beginning in the 1980s, labor unions saw a information on businesses around the country. It steep decline in their on-the-ground presence, a also covers nonprofit organizations of various kinds. decline largely offset by the rise of explicitly partisan

40 BREAD AND CIRCUSES LYMAN STONE

Figure 25. Two Different Stories of Civil Society Organizations

1,400

1,200

1,000

800

600

400 Entities per Million People 200

0 1961 1991 1946 1949 1952 1955 1958 1964 1967 1970 1973 1976 1979 1982 1985 1988 1994 1997 2012 2015 2018 2000 2003 2006 2009

CBP Nonprofit and Membership Organizations NCCS Reporting Public Charities

Source: Estimates based on Country Business Patterns 1947–2018 reports and National Center for Charitable Statistics Nonprofit Sector in Brief reports. Adjustments have been made to both datasets to account for changes in methodology across the relevant time series and ensure comparability over time.

Figure 26. Religious Establishments Dominate the Associational Landscape

700 eople 600

500

400

300

200

ofit Establishments per Million P 100

0 Nonpr 1961 1971 1981 1991 2011 1959 1963 1965 1967 1969 1973 1975 1977 1979 1983 1985 1987 1989 1993 1995 1997 1999 2001 2013 2015 2017 2003 2005 2007 2009 Business Associations Professional-Membership Organizations Labor Organizations Civic, Social, and Fraternal Associations Political Organizations Charitable Organizations Religious Organizations

Source: Estimates based on Country Business Patterns 1947–2018 reports. Adjustments have been made to account for changes in methodology across the relevant time series and ensure comparability over time.

41 BREAD AND CIRCUSES LYMAN STONE

Figure 27. Different Kinds of Associations Have Had Different Trends

250

200

150 ople

100 ofit Establishments 50 per Million Pe Nonpr 0 1961 1971 1981 1991 2011 1959 1963 1965 1967 1969 1973 1975 1977 1979 1983 1985 1987 1989 1993 1995 1997 1999 2001 201 3 201 5 201 7 2003 2005 2007 2009

Business Associations Professional-Membership Organizations Labor Organizations Civic, Social, and Fraternal Associations Political Organizations Charitable Organizations

Source: Estimates based on Country Business Patterns 1947–2018 reports. Adjustments have been made to account for changes in methodology across the relevant time series and ensure comparability over time.

organizations such as think tanks and taxpayer feder- Whether this development is laudable or worrisome ations. Whereas labor unions began as mass organi- is a matter of perspective. Many of the social organiza- zations that incidentally conducted lobbying, political tions being replaced reinforced and entrenched dif- organizations are the opposite. They exist to lobby and ferences among people of different classes and races. often struggle to motivate mass engagement. Certainly, Their replacement by organizations overtly interested think tanks do not engender the kind of large-scale in bringing about social benefits could be a welcome community participation that labor unions did. change. However, the extent to which these issue Surprisingly, civic, social, and fraternal organiza- advocacy groups actually have any community role or tions have been relatively stable, although this dis- involve a meaningfully large number of people is debat- guises some internal shifts. Fraternal and purely able. A shift from large fraternities reinforcing social social societies have declined, whereas organiza- differences to a narrower but more engaged group of tions interested in social justice and civic causes people in philanthropic organizations is a change, but such as preservationist clubs, wildlife protection not necessarily an improvement. societies, and human rights organizations have Business and professional associations rose grown. Detailed information about these subcatego- in prominence until around 2000 but have since ries of organizations is only available in recent years, declined, likely falling victim to the same forces out- so they can’t be compared in a long historical time lined for social groups and unions: Organization-based series. But this mirrors the trend in unions: Issue socialization is in decline. Rather than join the Rotary advocacy and organizations directly concerned with Club and conduct philanthropy through its socially politics are growing, while organizations founded oriented organization, businesspeople would rather on some preexisting camaraderie or fellowship are work directly with charitable organizations, which are in decline. rising in prominence.

42 BREAD AND CIRCUSES LYMAN STONE

Figure 28. Social and Charitable Nonprofits Are on the Rise

400

350

eople 300

250

200

150 ofits per Million P 100

Nonpr 50

0 2011 1995 1996 1997 1998 1999 2001 2010 2012 2013 2014 2015 2016 2000 2002 2003 2004 2005 2006 2007 2008 2009

Arts Environment and Animals Human Services Other Public and Social Benefit

Source: Estimates based on National Center for Charitable Statistics Nonprofit Sector in Brief reports. Adjustments have been made to account for changes in methodology across the relevant time series and ensure comparability over time.

NCCS data, on the other hand, show significant organizations with branches in many locations might growth across many categories, not least because they still be tracked as a single organization depending on include major growth sectors such as universities their exact legal structures. and hospitals, many of which are listed as nonprofit In other words, the CBP data are a better proxy organizations. Most of the organizations comparable for public or mass participation in organizations, to CBP’s nonprofit bodies would be classified in the whereas the NCCS data are a better proxy for inter- NCCS groups for environment and animals, human est and attention by philanthropic or public-minded services, arts, or other public and social benefits. Fig- individuals. The impulse toward philanthropy and ure 28 shows the prevalence of these organizations. public-spiritedness is as strong as ever in America or NCCS shows growth in all these bodies. The num- stronger, as shown by rapid growth in the number of ber of IRS-recognized public charities working in the active nonprofit organizations trying to improve com- public interest or for public benefit, far from declin- munities and strengthen American social capital. But ing, appears to have actually risen. the mass buy-in simply isn’t there. But while at first this may seem to contradict the American communities are littered with nonprofit CBP data, they are simple to reconcile. The cod- organizations trying (and often succeeding) to make ing system used in the CBP data primarily reflects their communities healthier, happier, safer, and more membership-based organizations and tracks these beautiful. But nonetheless, these communities are not organizations’ physical establishments. Thus, each seeing growth in places where Americans gather to American Legion post would be an organization actually build deep, significant relationships with each tracked separately by the CBP. But the NCCS data other. Why this is the case is unclear, but the obvious track tax-filing entities, including many that have lit- explanation from the prior section is that Americans tle or no physical presence anywhere. Furthermore, are distracted by other activities, especially sports.

43 BREAD AND CIRCUSES LYMAN STONE

Figure 29. Americans Are Spending Less Time Together

2.5

2.0

1.5

1.0

0.5 Hours Spent with Adults Who Aren’t 0 Members of the Same Household per Day 1971 1981 1991 2011 1965 1967 1969 1973 1975 1977 1979 1983 1985 1987 1989 1993 1995 1997 1999 2001 2013 2015 2017 2019 2003 2005 2007 2009

On Weekends On Weekdays

Note: American Heritage Time Use Study (AHTUS) samples define time spent with others in various ways. Only some AHTUS data points were used to ensure comparability, and adjustments were made to account for differences in time use categorizations. Source: Estimates from IPUMS AHTUS and IPUMS ATUS using “who” variables for non-household adults. All activity categories included.

Seize the Day: Changes in How This decline in time together has been almost pre- Americans Spend Their Time cisely offset by more time spent watching TV, lis- tening to the radio, reading, surfing the internet, The claim that Americans are withdrawing from and playing video games. Of course, the advent of relationships with each other may seem bold, but it smartphones has dramatically increased the oppor- finds support in detailed data collected about how tunity to consume various forms of media. Whereas Americans spend their time. Whereas Americans in media consumption in the 1960s only came to about the 1960s and 1970s likely spent an average of about 90 minutes on weekdays and 120 minutes on week- 90 minutes per weekday with non-household mem- ends, today the average American adult has over bers outside of work and school and about 120 min- 150 minutes of media consumption per weekday and utes per weekend day, those figures have fallen over 210 minutes on each weekend day. markedly. Today, the average American spends less While I include reading books and newspapers in than an hour of each weekday on nonwork activities this metric to account for the possibility of a change with other adults outside the home and about only in the medium of consumption from physical pub- 75 minutes per weekend day. Figure 29 shows data lications to digital publications, changes in reading, from historic time use surveys. internet usage, and video games account for a small Americans are spending less and less time together. share of the long-run change in time usage. Rather, This decline is most severe among younger Amer- the lion’s share in time use changes is driven by icans. Time spent with friends among people under ever-escalating television viewership, which in these age 30 is down almost 40 percent since 2005, but only data would include streaming services such as You- down about 15 percent among those over age 60. Tube and Netflix. Figure 30 shows US TV viewership

44 BREAD AND CIRCUSES LYMAN STONE

Figure 30. Everyone Is Watching More TV, but Americans Watch the Most

25

20

15

10

5

0 Estimated Hours of TV Watching per Week 1971 1981 1991 2011 1965 1967 1969 1973 1975 1977 1979 1983 1985 1987 1989 1993 1995 1997 1999 2001 2013 2015 2017 2019 2003 2005 2007 2009 Austria Bulgaria Canada Finland France Hungary Israel Italy South Korea Netherlands Spain United Kingdom US South Africa

Note: There was no age restriction on the sample. Source: IPUMS MTUS “watching TV” activity. International data points and US values before 2000 were queried from IPUMS MTUS. Val- ues after 2000 were queried from the IPUMS ATUS.

versus other countries for which comparable time of Americans watched the Super Bowl. But the fig- use data are available. ure has stayed close to 50 percent of all Americans in As with sports, the United States appears increas- most years since the mid-1980s. The most-watched ingly exceptional. While a few countries have some- regular TV shows have also changed over time from times had more TV viewership than us, in recent plotted shows and sitcoms in the 1950s and ’60s to years Americans have begun to watch far more televi- news to reality television to, in every year since 2011, sion than their peers in other countries do. NBC Sunday Night Football. It may seem contradictory to point to both sports Sports and television are an increasingly unified and television as major culprits in the decline of phenomenon. Americans are watching more and more American associational life. Americans are accus- TV, and a lot of what they are watching is sports. The tomed to a parent-centric narrative about young loss of sports TV during the COVID-19 pandemic was people either being too obsessed with video games a huge social shock to millions, and news about league and never going outside or too obsessed with sports reopenings and COVID-19 procedures was almost as and never doing their schoolwork. To present the high profile as the presidential campaign was. A larger problem as both too much sports and too much tele- and larger share of American free time away from vision can seem unusual. But they go together. Of the the screen is made up of sports and exercise, even as 30 most-watched TV broadcasts of all time in Amer- watching sports is claiming an ever-faster-growing ican history, 29 are Super Bowl broadcasts. (The share of American schedules. other is the M*A*S*H finale.) In 1968, just 25 percent

45 BREAD AND CIRCUSES LYMAN STONE

Figure 31. Reported Social Life Is Fairly Stable

450

400 350 300 250 200 150 100 50 Estimated Evenings Spent per Year with Friends, Relatives, or Neighbors 0 18 21 24 27 30 33 36 39 42 45 48 51 54 57 60 63 66 69 72 75 78 81 84 87 90 93 96 Age at Time of Survey

1890s 1900s 1910s 1920s 1930s 1940s Birth Decade: 1950s 1960s 1970s 1980s 1990s 2000s

Source: General Social Survey, 1972–2018.

Thus, when parents are pulled between the twin Complicating Factors: The General Social poles of having their kids join the football team or Survey and Reported Social Lives watch TV, in either case they are participating in the same cultural system, more or less invented after Some researchers have pushed back against the World War II and arguably since the 1960s, that estab- “declining social capital” narrative in one particularly lishes basically two primary ways for people to spend compelling way: by citing data from the GSS show- leisure time: play sports or watch sports (or play ing that the number of evenings respondents report sports-themed video games). spending with friends and families has not declined It is not simply nostalgic to think that Americans over time. Figure 31 shows the estimated number used to spend their time in other ways: in clubs, at of evenings per year spent with friends, relatives, or church, visiting each other at home, and so forth. The neighbors. Note that these questions are asked sep- best available data support the view that Americans arately, so if a person spent the evening with both today truly have different, less community-oriented friends and relatives, it would be double counted. As social lives. Associational life is dominated by orga- a result, in theory, the maximum possible value would nizations that don’t depend on mass participation or be over 1,000. In practice, however, estimates over extensive community involvement, while recreational 350 probably imply people spending part of almost time is spent participating in activities that build little every evening with others. social capital or simply watching those activities on This shows that while social life outside the home TV, at home, often alone. declines rapidly from age 18 until age 50, that trend isn’t new. Controlling for age, younger Americans actually appear to spend slightly more of their eve- nings with others.

46 BREAD AND CIRCUSES LYMAN STONE

Figure 32. Membership Rates in Most Groups Have Declined

6%

4%

2%

0% 1970s 1980s 1990s2000s –2%

–4%

–6%

of This Kind Group Since the 1970s –8% Change in Share of Adults Who Are Members –10% Church Group Professional Society Literary or Art Group Farm Organization Nationality Group School Fraternity Hobby Club School Service Youth Group Sports Club Labor Union Political Club Veteran Group Service Group Fraternal Group

Source: General Social Survey, 1974–2004.

How can it be that more evenings are spent with Finally, the increase in reported evenings spent others even as less time is spent with others? One with families has occurred almost exclusively among answer is that the increase is driven entirely by younger Americans and may be related to younger friends and family. Evenings spent with neighbors Americans increasingly living at home with their par- have indeed declined, even controlling for age. Mean- ents. It’s not clear that young adults spending more while, the question the GSS asks may be somewhat years at home with longer delays before the tran- ambiguous. “How often do you spend a social evening sition to residential independence, marriage, and with friends?” can be challenging to answer. During childbearing is a good thing. While intergenerational COVID-19, many individuals spent “social evenings” households might build social capital according to with friends over Zoom. Individuals might spend a some definitions of that phrase, few commentators “social evening” with their guild in World of Warcraft. would treat an increasing rate of young adult failure Should this count? What if it’s an in-person school to launch as a success for the community. During study group, but with friends? Is that social? An appar- the COVID-19 pandemic, evenings spent with family ently simple question can be difficult to interpret. presumably rose dramatically as people stayed home Indeed, the shift in time spent between friends and more, but this was not a sign of healthy communities. neighbors is suggestive: If socialization shifts from The GSS has other measures of social capital as neighborhood acquaintances to friends outside the well. From 1974 to 2004, the GSS asked respondents neighborhood, then the same number of evenings about their membership in various organizations. Fig- spent together would entail more time spent travel- ure 32 consolidates those responses by decade and ing to and from social activities and thus less actual shows how responses for each organizational type time spent together. have changed over time.

47 BREAD AND CIRCUSES LYMAN STONE

Figure 33. Americans Belong to Fewer Groups

70%

60%

50%

40%

30%

20%

Share of Adults Who Report 10% Belonging to Given Kind of Group 0% Political Party Union or Church or Sports, Leisure, Other Professional Religious Body or Cultural Group Organizations

2004 2014

Source: General Social Survey, 2004 and 2014.

For most kinds of groups, membership rates have But these data cut out in 2004. After 2004, the declined since the 1970s. For labor unions and fra- GSS switched its question wording to asking about ternal societies, self-reported membership is down “belonging” to groups rather than membership, and 40 percent. Membership in veterans’ groups and farm different types of groups were consolidated. Figure 33 organizations is down 35 percent. Church member- shows the share of the adult population that reported ship is down 20 percent, while school fraternities, belonging to each kind of organization. nationality groups, and (surprisingly) sports clubs Rates of belonging declined between 2004 and are all down 10 percent. School service groups, polit- 2014 for religious groups, political parties, and unions ical organizations, and youth groups are approxi- or professional associations. They fell by a relatively mately unchanged since the 1970s. A few groups, such small amount for sports, leisure, and cultural activi- as service clubs, hobby clubs, literary and art groups ties and rose for other groups. A follow-up question (including book clubs), and professional societies are about whether individuals actively participated in more prevalent than in the 1970s. But most of these these organizations turned up the same trends. groups declined between the 1990s and the last sur- In 2002 and 2018, the GSS asked a more detailed vey in 2004. Only literary and art groups, political question about organizational participation across a organizations, and youth groups grew in prevalence few organizational categories. Figure 34 shows the during that period. share of the population that participated in various Thus, self-reported rates of membership suggest organizational activities more than twice per year. that while individual interest-based activities such as Participation rates in all three major organizational professional societies or book and hobby clubs have categories have declined. Because categories used fared reasonably well, large collective organizations across these questions vary, for most groups they have struggled. can’t be synthesized into one coherent story across all periods. However, the category of “sports, culture, or

48 BREAD AND CIRCUSES LYMAN STONE

Figure 34. Fewer Americans Are Participating in Associations

40%

35%

30%

25% 20%

15%

10% More Than Once per Year Share of Adults Who Report 5%

Participating in Given Associations 0% Sports, Leisure, or Politics or Political Charitable or Religious Cultural Group Associations Voluntary Organizations 2002 2018

Source: General Social Survey, 2002 and 2018.

leisure organizations” can be reconstructed across all social attitudes. But Americans have become far less questions. Figure 35 shows how participation in these cooperative and trusting of each other over time. Data groups has changed over time. from the GSS show that whereas in 1972, Americans Exact indicators vary. But since the 1990s, every were about evenly divided about whether you “can available indicator of participation in what might be trust most people” or “can’t be too careful,” today the called social organizations has declined. While this “can’t be too careful” response is twice as common, as kind of organization might not always create extremely Figure 36 shows. robust and socially useful relationships, as discussed This shows up in other data as well. Gallup has in detail for sports, nonetheless the decline in formal surveyed American confidence in institutions since connections to or participation with these institutions the 1970s. The share of Americans who rate the aver- reflects an important change in American cultural life. age institution about which Gallup asked as trust- Thus, while one indicator of community life worthy has fallen over time, while the share who rate (evenings spent with others) measured by the GSS institutions as untrustworthy has risen (Figure 37). suggests that the decline of social capital may be over- Note that this decline in trust is not shared with stated, the other indicators all point to the view that, other countries. Academic research has found that at least since the 1990s, the intensity of American the decline in trust in the United States is uniquely participation in associational life has indeed fallen extreme.72 Thus, US-specific explanations for considerably. declining trust are important to identify, hence the Finally, one of the major goals of a tighter-knit emphasis this report has placed on the exceptional associational life is more amicable and cooperative rise in American interest in sports and television.

49 BREAD AND CIRCUSES LYMAN STONE

Figure 35. Social Organizations Are in Decline

70%

60%

50%

40%

30%

20%

10% Culture, and Leisure Organizations 0% 2 Share of Americans Associated with Sports, 74 76 78 19 19 19 199 1980 1982 1984 1986 1988 1990 1994 1996 1998 2010 2012 2014 2016 2018 2000 2002 2004 2006 2008

"Member of" Any Sports, Culture, Leisure "Belong to" Any Sports, Culture, Leisure "Belong to" Question Association (Different Wording) Any Sports, Culture, Leisure "Active in" Any Sports, Culture, Leisure More Than Twice-Annual "Participation in" Any Sports, Culture, Leisure

Source: General Social Survey, 1974–2018.

Figure 36. Americans Are Becoming Less Trusting

70%

60% 50%

40% 30%

Can Be Trusted 20%

10% Asking Whether Most People

Share of Responses to a Question 0% 7 3 75 72 78 2011 1981 19 19 19 198 199 1990 1996 1999 2014 2017 1984 2002 2005 2008

Most People Can Be Trusted It Depends You Can't Be Too Careful

Source: General Social Survey.

50 BREAD AND CIRCUSES LYMAN STONE

Figure 37. Americans Have Less Confidence in Institutions

60%

50%

40%

30%

20%

10% Share of Americans Who Have Little/ 0% No or Quite a Bit/Great Deal of Confidence in the Average Institution Surveyed in That Year 1981 1991 2011 1973 1975 1977 1979 1983 1985 1987 1989 1993 1995 1997 1999 2001 2013 2015 2017 2019 2003 2005 2007 2009 Little or No Confidence Quite a Bit or a Great Deal of Confidence

Note: The exact set of institutions about which respondents were asked varies over time. Almost all years include churches, the Supreme Court, Congress, organized labor, big business, public schools, newspapers, and the military. Most years since 1991 include the presi- dency, the media, banks, television, the police, the criminal justice system, and small businesses. Source: Gallup Historic Trends.

How We Talk: The Long Arc of The boom peaked in the 1950s, just as Schlesinger Associational Life was coining the term “nation of joiners.” Thus, the Schlesinger-type view of American associational life is Finally, changes in American associational life can be inseparable from this particular boom in associations, traced from 1750 to today using the Google Ngram and especially of public thought and writing about database. By tracking the frequency of words that those associations. The prevalence of words related often designate formal associations (club, society, to organizations has fallen by a quarter since then, league, guild, etc.), the extent to which public dis- consistent with the decline seen elsewhere. cussion included or referred to formal associations, But this long time series helps put that decline in a proxy for their overall prevalence in society, can perspective. Associations in America today remain be derived. Figure 38 shows the incidence of associ- about twice as prominent in American published ation label words per million published words in the texts as they were during the heyday of high voter par- preceding decade of Google Ngram’s publications. ticipation by enfranchised individuals and Tocquevil- There is a slight increase in association-related lian associationism and remain at levels similar to the words from 1750 to 1880. But beginning in 1880, public early days of fraternalism. The decline in associational writing about and interest in associations exploded, life has been, in terms of total scale and prominence, coincident with the Golden Age of Fraternalism dis- considerably overstated. cussed earlier. But this wasn’t just fraternities. Advo- However, there is a different textual indicator cacy groups, labor unions, political parties, and other that shows a more novel decline. Google Ngram can societies and associations experienced a boom in be used to track the relative frequency of individ- their rhetorical prominence. ual versus collective pronouns as a method to track

51 BREAD AND CIRCUSES LYMAN STONE

Figure 38. The Rise and Fall of Associational Words

1,800 1,600 1,400 ds per Million r 1,200 Wo 1,000 American English

s in 800 rd 600 Wo 400 200 Published 0 Associationally Referring 1750s 1760s 1770s 1780s 1790s 1800s 1810s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1910s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2000s 2010s

Note: Words included are group, club, association, team, community, society, alliance, league, guild, fellowship, lodge, circle, union, hall, encampment, fraternity, brotherhood, federation, and knights. Results are averaged by decade. Source: Google Ngram; and author’s calculations.

the extent to which published texts reflect individ- Individual pronouns have always been more com- ual or single-author perspectives versus collective or mon than plural pronouns have been, especially in multiple-author perspectives. A change in the ratio of fictional texts. But in the early 19th century, the gap collective to individual pronouns could reflect greater was comparatively small. From the 1830s to the 1860s, salience of publications reflecting an individual’s individualist writing became more common across voice or perspective and probably therefore greater all texts, and in the 1840s, a major individualist turn social salience of individualist speech and possibly occurred in fiction writing. attitudes. However, this relationship could be con- On the other hand, opinion pieces in newspapers founded if the kinds of publications produced over still often spoke of what “we” wanted the government time have changed. For instance, if legal texts or fic- to do. But the 1840s to the 1880s brought a burst of tion books tend to have different uses of singular or more individualistic phrasings, with relatively less plural pronouns and the relative publishing volume of usage of collective pronouns such as “we” and “us” legal texts has changed, a change in pronoun ratios and relatively more of “I” and “me.” might not indicate any actual change in public values. But in the latter half of the 19th century, collec- Thus, as a control, I tracked the pronoun ratio of tive pronouns made a comeback across texts and not only all published texts but also exclusively fic- ceased their decline in fictional texts. The balance tional texts, which should be free of bias due to genre of pronoun usage would remain about stable from compositional changes. Positive values indicate that 1920 to the 1970s, when fictional texts began to see individual pronouns were more common than plural a sharp increase in the relative usage of individual pronouns were. As Figure 39 shows, the indexes of pronouns. By the late 1980s, nonfiction texts had pronoun ratios have changed by similar amounts over begun to follow. Today, published texts have more time for fiction and for texts generally. individual pronouns relative to plural pronouns

52 BREAD AND CIRCUSES LYMAN STONE

Figure 39. The Rise of “I” and “Me”

18,000

16,000

14,000

12,000

10,000

8,000

Published Words 6,000

4,000 Collective Pronouns, per Million 2,000 Frequency of Individual Pronouns Minus 0 1810s 1910s 1800s 1820s 1830s 1840s 1850s 1860s 1870s 1880s 1890s 1900s 1920s 1930s 1940s 1950s 1960s 1970s 1980s 1990s 2010s 2000s

All American English Texts All English Fiction

Source: Google Ngram; and author’s calculations.

than at any time since the American founding. Sim- suggests that the United States is indeed experienc- ilar trends appear for second-person pronouns such ing a significant decline in community orientation. as “he” and “she” versus “they” or “him” and “her” People write about themselves, not their communi- versus “them.” ties. Americans spend less time together. They join Were this indicator the only evidence of a shift fewer associations, and the associations they do join away from associational life toward a more solo life- tend to be based around either a few very active indi- style, it would not be highly compelling. But with viduals without mass participation or mass activities all the other evidence presented above, it strongly with little social or public benefit.

53 V. Conclusion

merican associational life has changed many in need of additional care or assistance should be able A times in its history and taken on new and dif- to get it. ferent forms. These different kinds of associations The academic literature strongly supports that met different social needs and created different ben- dense networks of association can alleviate pub- efits and costs for wider society. But in the long run, lic health problems. Reaching back to the premod- organized public associations with some clear positive ern period, the medieval leprosy pandemic was only aim are a benefit to society, whatever their form. Thus, ended by religious orders establishing safe and sep- the decline in the density of these associations, and arated residential areas for lepers, allowing the dis- especially their replacement by narrower ideological ease to be slowly eliminated from Europe.73 More organizations or broad, low-stakes associations such recently, an abundant health literature has found that as sports, may create new difficulties for American individuals with stronger ties to membership-based society. The extent of this decline has been somewhat organizations and higher social trust are more likely overstated in prior research but is nonetheless real to get vaccinated, cooperate with public health and closely tied to a decline in community orientation. authorities, and experience less mortality in longi- tudinal studies.74 There is an extensive academic literature on the When It Counts: Community Health and role of social capital in creating “disaster-resilient” COVID-19 communities, although it is primarily driven by qual- itative, theoretical, and management case study This report has discussed changes in associational papers, not quantitative analysis for causal infer- life over time and identified many causes and effects ence. Sociologist Eric Klinenberg has argued that of those changes. But rather than offer an exhaustive neighborhood-level mortality during ’s 1995 account of the effects of increasing social disengage- heat wave was driven by differences in the density of ment across a litany of different areas, it may be more social institutions.75 Unfortunately, no study has tack- fruitful to engage with one particularly pressing prob- led whether more robust associational life improves lem: COVID-19. By exploring the role of associational community performance during an epidemic in a life in mediating the effects of theCOVID-19 pan- large-scale, quantitative way. demic, the broader benefits of robust associational However, enough data exist to produce a credible ties can be demonstrated. estimate of state-level social capital similar to con- Pandemics are extremely useful test cases for the temporary estimates used by academics for the 1918 social benefits arising from associational life. Most influenza pandemic, which killed over 60 million peo- epidemic diseases are spread through human contact, ple around the world and over 600,000 Americans.76 and thus managing novel epidemics requires rapid Data from the 1916 census of religious bodies can be changes to individual behavior, often coordinated used to identify the extent of church membership with others in a community. This is exactly the kind of (a key indicator of associational life), and data from situation in which having lots of associations should the 1910 census can be used to identify what share of be useful. Disseminating information should be easy, adults were employed in jobs directly related to asso- inducing cooperation with public health measures ciations. (There are specific occupational codes for should be relatively straightforward, and individuals various nonprofit association workers in 1910.)

54 BREAD AND CIRCUSES LYMAN STONE

Figure 40. Higher Social Capital Predicted a Smaller 1918 Influenza Pandemic Death Spike

1.5

1.0

0.5

0.0

–0.5 Index of Social Capital

2 –1.0 R = 0.3023 (Church Membership, Fraternal Associational Occupations, and Voter Turnout) –1.5 0% 5% 10% 15% 20%25% 30%35% 40% Increase in Average Annual Deaths, 1919–20 vs. 1918

Source: Voter participation rate calculated using 1910 and 1920 IPUMS USA estimates of the citizen voting-age populations by state, interpolating like 1916 eligible voter population; 1916 presidential vote totals used. Statistics on church membership by state were taken from the US Census Bureau’s 1916 Census of Religious Bodies, combining all religious bodies, divided by each state’s adult population. Fraternal membership totals for predominantly white segments of the Odd Fellows, Freemasons, and Knights of Pythias in 1914–16 were taken from World, “The World Almanac and Encyclopedia 1916,” 1915, https://archive.org/details/ worldalmanacency1916newy. Increases in mortality by state estimated from Historical Vital Statistics of the United States reports.

Data from the 1916 world almanac provide state- the correlation remains: Denser associational ties are level membership data for the Knights of Pythias, correlated with better epidemic outcomes. This find- Odd Fellows, and Freemasons, three of the largest ing is robust to dropping out any randomly selected fraternal societies. Finally, with data from the 1916 one or two states as well, suggesting it isn’t driven by election, in combination with data on what share of outliers. This analysis is insufficient to show causality, citizens in each state were of age to vote and could do but it does suggest that associational life has a special so under local suffrage rules (some states let women role during pandemics. vote in 1916, while others did not), voter turnout rates Recent research about COVID-19 is far more can be calculated. Figure 40 shows that an index of robust. One study found that European regions with these variables is indeed predictive of lower death higher electoral turnout, more blood banks, or higher rates during the 1918 influenza pandemic for the historic literacy rates (all indicators of social capital 27 states for which complete data for all indicators used in prior research) practiced more aggressive are available. social distancing earlier in the pandemic and as a In those states where people were more tightly result saw less spread of COVID-19.77 Another study bound together by associational ties, the 1918 pan- found that US counties with higher voter participa- demic was slightly less severe. Even dropping the tion have practiced more social distancing and com- voter participation data out of the calculation and plied with stay-at-home orders more readily, and only using data on church membership, fraternal it also found the same thing for individuals across society membership, and associational occupations, the US and Europe who had high trust in others.78

55 BREAD AND CIRCUSES LYMAN STONE

The same finding showed if voter participation was as mask wearing at different times. As more people replaced with a broad measure of social capital from began to wear masks, they may have ventured out of the JEC that includes organizational membership, their homes more without necessarily shirking any volunteerism, family structure, and local crime rates civic duty to protect others. (discussed more below). A third study conducted Finally, it remains unclear exactly which poli- by an independent team using various social capital cies were most effective in dealing with COVID-19, indicators at the county level also found that social but one of the putative benefits of robust civil soci- capital was strongly related to positive COVID-19 ety is that it enables debates about social choices to outcomes.79 reach efficient results much more quickly. With more However, there is some disagreement. One study associations providing avenues for public delibera- found that while indicators such as voter partici- tion, effective strategies can be discovered. For many pation predicted more compliance with social dis- Americans, decisions made in their churches were tancing, higher participation in community-based the front line of their civic participation in COVID-19 activities (such as churches) was actually associated mitigation, so those decision-making processes could with less compliance with official orders.80 That is to have exposed them to information they would not say, civic-mindedness had positive effects, but associ- otherwise have obtained. ational life did not. Finally, a particularly concerning The importance of diverse information has paper found that more social distancing was strongly been compellingly demonstrated in the context of associated with support for overtly xenophobic and COVID-19 itself. A recent study exploited variation racist political parties in Russia and that areas with in cable news scheduling, time zones, and sunset more interethnic tensions and violence practiced timing to identify effectively random variation in TV more social distancing.81 A reduced-form version of viewership.82 It identified random forces that caused the model also predicted US social distancing behav- either Tucker Carlson’s or Sean Hannity’s TV shows iors well. on Fox News Channel to get more viewers in a given Current evidence suggests that civic-mindedness county. Early in the pandemic, Hannity downplayed is strongly associated with more people taking mea- the severity of COVID-19 and encouraged his viewers sures to protect themselves and others. But the to live life as normal, while Carlson forcefully argued record on community participation is more mixed, that COVID-19 was a large problem getting insuffi- and extremely divided communities may perform just cient attention. As a result, counties randomly more as well as those with rich associational ties. exposed to Carlson’s show practiced more social dis- However, almost all these studies assess essentially tancing and had fewer COVID-19 cases. The research- two kinds of indicators: officially reported case counts ers also conducted a national survey of Carlson’s and and social distancing as measured by cell phone ana- Hannity’s viewers and confirmed their findings: Carl- lytics. Both measures have problems for measuring son’s viewers reported adopting precautionary behav- pandemic outcomes. Official case counts may vary iors earlier, even though Carlson’s and Hannity’s due to changes in testing or may have time delays. viewers are demographically and politically similar. In Furthermore, substantial controls are necessary to other words, information matters. establish an appropriate baseline. Areas closer to New To test whether dense associational life is bene- York would inevitably have more cases, for example. ficial, I used two methods, both designed to bypass Cell phone distancing data are also fraught, since dif- problems of imperfect data quality. The first uses esti- ferent areas took different approaches to tackling mated effectiveR-values for COVID-19 at the state COVID-19. In some areas, authorities did not strongly level. EffectiveR-values are a shorthand way to mea- encourage social distancing, so lower rates of social sure how many people the average infected person distancing could reflect compliance with authorities. will infect. Importantly, these values can be calculated Beyond this, different areas adopted measures such relatively reliably, even from incomplete data, and

56 BREAD AND CIRCUSES LYMAN STONE

have been tested using different calculation methods activities, meetings, and organizations and is a good by various researchers. proxy for what I term “associational life.” This is the As an added check, I calculated excess mortal- same indicator used by several other researchers. ity rates for each state and month, calculating what I also analyzed two other indicators from the JEC: factors influence changes inall-cause mortality. This social support, which measures mutual-aid behaviors measure is less directly related to COVID-19 and such as how often individuals check on their neigh- could be influenced by other factors, but it has the bors, and institutional health, which captures indica- benefit of being immune to biases arising from incon- tors such as voter participation and census response sistent testing or cause-of-death attribution. It is the rates and is similar to the civic-mindedness used in most reliable measure, even if it is less directly tied other research. I analyzed panel models of change to COVID-19. Furthermore, the lives lost is the actual over time of excess death and R-rates at the state level harm arising from COVID-19. It’s possible that areas using the control variables, the community health with denser associational ties could have more spread indicator on its own, and models using all three social of COVID-19 but fewer deaths if informal care or pro- capital indicators. tection of vulnerable members of society also varies. I As Figure 41 shows, I find that aone-standard - used data through January 2021 for both R-values and deviation increase in community health has large, excess death rates. beneficial effects reducing excess mortality rates, For both models, I added controls for state-level and smaller, less precisely estimated effects reduc- monthly temperature and temperature squared ing R-values. Notably, community health had a large (for reasons explained below), population density, effect reducingR-values in states where support Google-provided data on how many people stayed for Trump was low. In states that voted for Trump at home, the share of the presidential vote in 2020 more, community health was less predictive of lower in favor of Donald Trump, and the R-value or excess COVID-19 transmission. That this relationship with mortality value in neighboring states in the prior death rates was stronger than with COVID-19 trans- month. Stay-at-home behavior and population den- mission could also indicate that associational life sity are both proxies for the extent of contact occur- might have affected trends in other causes of death ring in a state, which could affect the spread of such as suicide, drug overdose, car accidents, and COVID-19 independently of differences in associa- elder deaths at home. tional life. The use of lagged values from neighboring Contrary to prior research, civic burden-sharing states reflects thatCOVID-19 can spread geographi- behaviors had no effect on the spread ofCOVID-19 or cally, and if neighboring states have an outbreak, it is excess mortality rates. Informal mutual-aid activities likely to spill over and drive higher infection spread such as checking on neighbors also have no significant and death rates over time. Including presidential vote effects. I have not shown effects for control variables shares may seem odd, but as noted in the research such as temperature or voting, but effects were gener- cited above, it’s important. Political attitudes influ- ally small. The effects of density are included because, enced how people responded to COVID-19. It’s for excess mortality, the effect is large. More dense possible that more Trump-supporting states could places have more deaths, as might be expected from have a systematically different relationship between a communicable disease, although they did not have associational life and responses to COVID-19, so I higher R-values. On the other hand, staying home allowed for this kind of effect using an interaction more was strongly associated with lower transmis- term between Trump’s vote share and indicators of sion of COVID-19 but with a much more modest, and social capital. less significant, effect on death rates. Neighboring To measure associational life, I used the commu- states having more transmission or higher death rates nity health indicator from the JEC. This indicator is in the month before significantly increased transmis- an index of state-level participation in community sion and death rates, suggesting that interstate travel

57 BREAD AND CIRCUSES LYMAN STONE

Figure 41. Factors Driving COVID-19 Spread and Excess Mortality

20%

10%

0%

–10%

–20%

–30% Deviation Increase in Given Variable R-Value Associated with One Standard Change in Monthly State-Level Effective –40% Associational Informal Mutual- Civic Burden Social Distancing Population Geographic Life (JEC Aid Activities Sharing (25 Percent Density Spillover Community (JEC Social (JEC Institutional Increase in (500 More (25 Percent Health Support Health Google People per Increase in Subindex) Subindex) Subindex) Measurement Square Mile) Neighboring- of People State Prior-Month Staying Home) Dependent Variable) Effect on Excess Mortality Rate Effect on R-Value

Note: Ninety-five percent confidence intervals are shown. Source: Estimates are derived from two generalized linear model (random/pooled effects) panel models with robust standard errors of state-level excess mortality rates using Centers for Disease Control provisional mortality data and R-values taken from Youyang Gu’s COVID-19 Projections project. Population density data are from the US Census Bureau. Monthly average temperature are from the National Oceanic and Atmospheric Administration. Social capital indicators are taken from JEC’s Social Capital Index. Presidential vote shares are taken from public election tallies.

is an important driver of COVID-19 transmission and induced his viewers to take precautions, the deluge of excess death rates. associations sending notices about pandemic proce- What is particularly striking about the strong, ben- dures caused Americans who were on their call lists eficial impact of vibrant associational life is that most or listservs to be inundated with warnings and advice associational activities during COVID-19 were sus- about COVID-19. People detached from associational pended. Bans on gatherings essentially canceled asso- life didn’t get these warnings. ciational life in America for months on end. And yet As an example, someone who exercised in a sports states with deeper associational commitments did club likely experienced clear warnings and disruptions better after controlling for basic proxies of their risk to their life early in the pandemic as that institution exposure such as climate, exposure to nearby out- took measures to mitigate risks to its members. But breaks, and social distancing behaviors. someone who exercised by jogging alone would have There are several reasons why this may be. First, experienced less such disruption and communication as associations communicated cancellations, sched- and so might not have adopted precautions as early. ule revisions, and safety protocols to members, they Second, even with in-person associational life may have helped persuade undecided individuals to canceled, associational connections may have made take COVID-19 seriously. Much as Carlson’s warnings social distancing more bearable and thus sustainable.

58 BREAD AND CIRCUSES LYMAN STONE

Churches continued to meet on Zoom. Small “pods” Researchers usually use indicators such as civic par- or “quarantine clubs” of people who shared preexist- ticipation or the density of associations. But the ben- ing deep connections rapidly formed in many areas. efits of vibrant andpublic-spirited social lives may Indeed, organizations facing financial ruin if they be better demonstrated in the unique circumstances missed out on months of activities leaped into high of 2020 than in the prior academic literature. Amer- gear, modernizing their digital footprint and innovat- icans with stronger ties to their communities were ing new ways to connect to members. better equipped to deal with the rigors of the pan- A survey conducted by the Pew Research Center’s demic. One wonders if, had COVID-19 occurred in Forum on Religion & Public Life in July 2020 demon- the 1960s, it might have been even better handled strates both these effects, at least for individuals with due to most people then having deeper associational religious associational ties. It found that 72 percent of ties. It’s difficult to say, but the decline in associa- Christians reported talking on the phone or by video tional life has certainly left many Americans in a at least weekly to friends or family, as did 73 percent uniquely vulnerable position. of Jews, while only 65 percent of unaffiliated Ameri- Worryingly, COVID-19 may lead to an accelerating cans reported doing so, and only 61 percent of atheists decline in associational life. After most natural disas- did.83 Religious people were also more likely to have ters, communities band together to recover. But after helped friends or neighbors during the pandemic and pandemics, they often do not. A study of Norwegian donated to charitable organizations. cooperative associations found that while harsh frosts Perhaps most striking, an enormous 76 percent led to more cooperatives being founded, severe flu of adults who attended religious services in the prior mortality led to fewer cooperatives.86 The research- month reported hearing sermons that discussed the ers suggest that, because the danger during an epi- importance of taking steps to limit the spread of the demic is “other people,” epidemics erode social trust coronavirus. For comparison, just 40 percent heard and nudge people toward isolation and detachment. any comments about voting, and just 31 percent This might be especially true in mandated lockdowns. heard any comments opposing restrictions on public Extensive study of the 1918 influenza pandemic in assembly.84 Perhaps unsurprisingly, a Gallup survey in America has found that people who experienced the December 2020 found that while self-assessed emo- pandemic and their descendants had significantly tional health declined sharply for most Americans in reduced social trust as a result.87 Much research has 2020, for those who regularly attended religious ser- shown that disasters lead people to be more cautious vices, emotional health actually improved.85 and risk averse in many areas of life.88 Religious identity is just one form of associational Thus, while strong associational ties may help life, but it clearly shows how formal associations can people and communities get through a crisis, disas- work to provide information, motivate pro-social ters themselves, and especially epidemics, often leave behavior, and make the isolation of the pandemic communities with weaker social ties than before. The more bearable. Many other forms of associational life same Pew data that show that religious people weath- surely have had similar effects, although many could ered the pandemic better also showed that only 37 per- not. It’s likely that members of sports clubs experi- cent of people have been attending church in person enced exceptionally severe disruptions to their lives or online during the pandemic, a huge decline from in the early months of the pandemic, for example, 45 percent in October 2019. (Attendance had previ- especially members of indoor sports and gyms. ously been falling about 1 or 2 percentage points per There is a mammoth literature demonstrating year.89) Beyond the human lives it claims, COVID-19 a litany of effects of social capital on various eco- will likely also spell the end of innumerable institu- nomic, social, and political outcomes of interest. tions, organizations, and habits of social life.

59 BREAD AND CIRCUSES LYMAN STONE

What Can Be Done? cooperation with formal associations, those organiza- tions are likely to persist. But when the government Solutions to declining associational life cover a wide co-opts formal associations and competes against range of policies, because different kinds of associa- them to provide similar goods and services, those tional life have declined for different reasons. Many associations will struggle. In that struggle, wider soci- forms of associations have died out because chil- ety is the main loser. dren in school are provided with a sports-centric set The cat is largely out of the bag on this issue, but of role models and opportunities, leaving the clubs there are some opportunities for improvement. For that could have connected them to lasting associa- example, governments could establish school choice tional ties bereft of new recruits. Others, such as labor programs, allowing more diversity in educational con- unions (or, not discussed in great detail in this report text and thus providing opportunities for civil society due to length, mutual-benefit societies), have strug- to take a more active hand in education and role mod- gled because their historic success has led to a more eling. This approach has been dubbed “educational prosperous society and more extensive government pluralism” and could be helped by an explicit effort by support and protection. state and local governments to identify a wide range Many social organizations have declined due to com- of organizations able to partner in education.93 Cur- petition from television and mass media. As I’ve shown rently, to the extent that social organizations have in prior research, specific state policies related to sec- a role in providing education, it is mostly through ularization have reduced religious activity over the church-sponsored schools. But in many countries, past century.90 Veterans’ organizations have declined nonreligious associations participate in school man- mostly because wars have become smaller and less agement, supervision, and operation. frequent. Meanwhile, the broad turn toward a greater The rise of television and mass media cannot be culture-wide focus on individual identity has created undone, and indeed it primarily reflects voluntary headwinds for almost all membership-based organiza- choices by individuals that in principle should be tions, even as forms of association that reinforce iden- respected. However, the rise of television could partly tity claims, such as political organizations, book clubs, reflect a lack of other options. New subdivisions and hobby clubs, have done relatively well. Thus, there rarely designate spaces for community centers in easy is no silver bullet for associational life. walking distance of residences. Whereas Pythian cas- Nonetheless, government choices matter. Research tles and Masonic temples dot the urban landscapes of has shown that Italian cities that briefly achieved yesteryear, modern subdivisions are stripped of visi- independence a thousand years ago still have denser ble signs of associational life. Revising zoning codes networks of associational life today, a finding that is to explicitly allow community organizations to be robust to numerous controls.91 Government struc- located in the communities they serve is an obvious tures can have extremely durable impacts on associ- reform that might enhance the benefits they provide ational life. and increase participation. Additionally, because it Numerous government policies can buttress a consists in removing a restriction on free contract, vibrant associational life. Government agencies can this approach is consistent with empowering individ- use public-private partnerships to develop social and ual choice, not simply trying to force people to join community services, for example. Much social wel- the Rotary Club. fare in Germany is delivered through religious orga- The decline of unions obviously connects to policy nizations.92 Likewise, labor unions have remained choices. But while it is tempting to think that the solu- stronger in countries that formally incorporate them tion is to undo right-to-work rules and return to the into the regulatory process, such as many Northern era of closed shops, this probably isn’t the case. Of the European countries. To the extent that state poli- four countries that the OECD identifies with rising or cies and priorities can be advanced through explicit stable unionization rates in recent years (Belgium,

60 BREAD AND CIRCUSES LYMAN STONE

Chile, Iceland, and South Korea), there is little com- of working schedule changes. Whether America will monality in their policy environments. Furthermore, see a resurgence in labor conflict or cooperation will the decline in unions isn’t just a policy story. Unions depend on how these emerging work trends are han- politically overreached, and closed-shop rules were dled. But in the end, labor will win. In every devel- unpopular. Any unionization revival will depend on oped country around the world, working hours have choice-consistent mechanisms, much as any revival been falling for decades. American working hours in participation in religious bodies cannot depend have fallen by less than almost any other country, and on a return to the public cultural norms of the 1940s. eventually American workers will demand their due.95 Nonetheless, there are things that can be done. The Managing this in an organized way through regular solidarity, mutual aid, and representation formerly work scheduling and formal holidays would be ben- provided by unions may be duplicable in voluntary eficial for the associational life that makes American (but federally supported) workers’ cooperatives.94 society so vibrant. One of the most striking measurements of declin- But while government policies certainly matter, ing associational life in America presented in this a considerable part of the work rebuilding Ameri- report was from time use data, which show that can associational life depends not on political lead- Americans spend far more time together on certain ers but everyday Americans. Associations don’t just days than others. On holidays and weekends, Ameri- benefit societies; they benefit individuals. Those with cans enjoy something more like the associational life robust communities have been less miserable during of yesteryear. Governments could strengthen associ- COVID-19. People with strong relationships of trust ational life by implementing more bank, school, and and support are are healthier, happier, and less likely government holidays and requiring employers to to die. grant their workers paid time off on a limited number But it takes work. The easy appeal to societal vapor- of those holidays. ware in the form of values-free entertainment in our Furthermore, as a growing number of workers face mass media and values-free life paths set up as admi- extremely irregular working hours and schedules, and rable in our schools will have to be widely resisted for as working from home becomes more common, gov- Americans to rebuild their communities. Quite liter- ernments are likely to either face a return of popu- ally, it takes time. Valuable time. On weekends and lar, activist unions locked in costly confrontations weeknights, with precious vacation time, Americans with businesses or be forced to adopt prudent regula- will simply have to choose what kind of lives they tions protecting workers from arbitrarily short notice want to live—isolated or together.

61 About the Author

Lyman Stone is an adjunct fellow at AEI, a research a PhD student in sociology at McGill University. He fellow at the Institute for Family Studies, the direc- is a former international economist at the US Depart- tor of research of the consulting firm Demographic ment of Agriculture, where he forecasted cotton mar- Intelligence, a Robert Novak Journalism Fellow, and ket conditions.

62 Notes

1. Social Capital Research & Training, website, https://www.socialcapitalresearch.com/. 2. Lyman Stone (@lymanstoneky), “If you watch a widely-watched TV show live, like the Super Bowl or M*A*S*H finale, are you building social capital?,” Twitter, August 14, 2020, 11:22 p.m., https://twitter.com/lymanstoneky/status/ 1294474497811243008?s=20. 3. Alexis de Tocqueville, Democracy in America (New York: George Dearborn and Co, 1838). 4. Arthur M. Schlesinger, “Biography of a Nation of Joiners,” American Historical Review 50, no. 1 (October 1944): 1–25, https://doi. org/10.1086/ahr/50.1.1. 5. Robert Putnam, Bowling Alone: The Collapse and Revival of American Community (New York: Simon & Schuster, 2000). 6. Putnam, Bowling Alone. 7. Harriett W. McBride, “The Golden Age of Fraternalism: 1870–1910,” Heredom 13 (2005): 1–31, http://phoenixmasonry.org/ Golden%20Age%20of%20Fraternalism.pdf. 8. World, “The World Almanac and Encyclopedia 1896,” January 1896, https://archive.org/details/worldalmanacency1896newy/page/ n5/mode/2up; and World, “The World Almanac and Encyclopedia 1906,” 1905, https://archive.org/details/worldalmanacency1906newy/ page/n63/mode/2up. 9. Jason Kaufman, For the Common Good? American Civic Life and the Golden Age of Fraternity (New York: Oxford University Press, 2002). 10. Cindy Alexander, “Statistics: Immigration in America, Ku Klux Klan Membership: 1840–1940,” Gilder Lehrman Institute of American History, https://www.gilderlehrman.org/history-resources/teaching-resource/statistics-immigration-america-ku-klux-klan- membership-1840-1940. 11. Kathy Baylis, Yazhen Gong, and Shun Wang, “Bridging vs. Bonding Social Capital and the Management of Common Pool Resources” (working paper, National Bureau of Economic Research, Cambridge, MA, July 2013), https://www.nber.org/papers/ w19195. 12. World, “The World Almanac and Encyclopedia 1896.” 13. Samuel Biagetti, “The Masons Are Still Segregated,” Killing the Buddha, October 2, 2009, https://killingthebuddha.com/ktblog/ the-masons-are-still-segregated/. 14. Shanker Satyanath, Nico Voigtländer, and Hans-Joachim Voth, “Bowling for Fascism: Social Capital and the Rise of the Nazi Party,” Journal of Political Economy 125, no. 2 (April 2017): 478–526, https://www.journals.uchicago.edu/doi/full/10.1086/690949. 15. Molly Caldwell Crosby, The American Plague: The Untold Story of Yellow Fever, the Epidemic That Shaped Our History (New York: Berkley, 2007). 16. Lyman Stone, (@lymanstoneky), Twitter. 17. Steven Bullock, Revolutionary Brotherhood: Freemasonry and the Transformation of the American Social Order, 1730–1840 (Chapel Hill, NC: Omohundro Institute of Early American History and University of North Carolina Press, 1998). 18. Lyman Stone, Promise and Peril: The History of American Religiosity and its Recent Decline, American Enterprise Institute, April 30, 2020, https://www.aei.org/research-products/report/promise-and-peril-the-history-of-american-religiosity-and-its-recent-decline/. 19. Patrick Brown, “More Research Confirms Importance of Social Capital in Studying Coronavirus,” US Congress Joint Economic Committee, June 25, 2020, https://www.jec.senate.gov/public/index.cfm/republicans/2020/6/more-research-confirms- importance-of-social-capital-in-studying-coronavirus. 20. Author’s analysis of Masonic membership statistics and dates of Masonic lodge foundings. 21. Independent Order of the Odd Fellows: The Sovereign Grand Lodge, “History of American Odd Fellowship,” https://odd-fellows. org/history/wildeys-odd-fellowship/. 22. Society of the Cincinnati, “The Society and Its Critics, 1784–1800,” https://www.societyofthecincinnati.org/about/history/critics.

63 BREAD AND CIRCUSES LYMAN STONE

23. Society of the Cincinnati, “Overview: A Short History of the Society of the Cincinnati,” https://www.societyofthecincinnati.org/ about/history. 24. Edward S. Mihalkanin, ed., American Statesmen: Secretaries of State from John Jay to Colin Powell (Westport, CT: Greenwood Press, 2004), 451. 25. Albert Clark Stevens, The Cyclopædia of Fraternities: A Compilation of Existing Authentic Information and the Results of Original Investigation as to More Than Six Hundred Secret Societies in the United States (Provo, UT: Repressed Publishing, 2015). 26. World, “The World Almanac and Encyclopedia 1896”; and World, “The World Almanac and Encyclopedia 1906.” 27. World, “The World Almanac and Encyclopedia 1906.” 28. Bill Loomis, “Clubbing in Days Past: When Fraternal Societies Ruled,” Detroit News, October 10, 2015, https://www.detroitnews. com/story/news/local/michigan-history/2015/10/10/fraternal-societies-detroit-history/73514852/. 29. Benjamin A. Olken, “Do Television and Radio Destroy Social Capital? Evidence from Indonesian Villages” (working paper, National Bureau of Economic Research, Cambridge, MA, October 2006), https://www.nber.org/papers/w12561. 30. Andrea Geraci et al., “Broadband Internet and Social Capital,” IZA Institute of Labor Economics, September 2018, http://ftp.iza. org/dp11855.pdf. 31. Jason Kaufman and David Weintraub, “Social-Capital Formation and American Fraternal Association: New Empirical Evidence,” Journal of Interdisciplinary History 35, no. 1 (Summer 2004): 1–36, https://www.jstor.org/stable/3656414?seq=25#metadata_info_ tab_contents. 32. Social Capital Project, US Congress Joint Economic Committee, “Social Capital, Slavery, and the Long Reach of History,” Jan- uary 28 2019, https://www.jec.senate.gov/public/index.cfm/republicans/2019/1/social-capital-slavery-and-the-long-reach-of-history. 33. Encyclopaedia Britanica, “Knights of Labor: American Labour Organization,” https://www.britannica.com/topic/Knights-of- Labor. 34. National Labor Relations Board, “1947 Taft-Hartley Substantive Provisions,” https://www.nlrb.gov/about-nlrb/who-we-are/ our-history/1947-taft-hartley-substantive-provisions. 35. Ozkan Eren and Serkan Ozbeklik, “What Do Right-to-Work Laws Do? Evidence from a Synthetic Control Method Analysis,” Journal of Policy Analysis and Management 35, no 1 (Winter 2016): 173–94, https://doi.org/10.1002/pam.21861. 36. Aztec Club of 1847: Military Society of the Mexican War, website, https://www.aztecclub.org/. 37. Military Order of the Loyal Legion of the United States, website, http://www.suvcw.org/mollus/mollus.htm. 38. Sons of Union Veterans of the Civil War National Headquarters, “Grand Army of the Republic History,” http://www.suvcw. org/?page_id=167. 39. Marquis James, A History of The American Legion (New York: William Green, 1923). 40. Mark Dyreson, “Maybe It’s Better to Bowl Alone: Sport, Community and Democracy in American Thought,” Culture, Sport, Soci- ety 4, no. 1 (Spring 2001): 19–30, https://doi.org/10.1080/713999807. 41. Robert D. Putnam, Robert Leonardi, and Raffaela Y. Nanetti,Making Democracy Work: Civic Traditions in Modern Italy (Prince- ton, NJ: Princeton University Press, 1993). 42. Emma Green, “Would Alexis de Tocqueville Have Joined a High School Football Team?,” Atlantic, November 18, 2013, https:// www.theatlantic.com/culture/archive/2013/11/would-alexis-de-tocqueville-have-joined-a-high-school-football-team/281562/. 43. Tocqueville, Democracy in America. 44. Tocqueville, Democracy in America. 45. Founders Online, “From Thomas Jefferson to Peter Carr, 19 August 1785,” https://founders.archives.gov/documents/Jefferson/ 01-08-02-0319. 46. Lyman Stone, “It’s Time to Finally Admit Professional Sports Are Bad for Society,” Federalist, May 21, 2018, https://thefederalist. com/2018/05/21/professional-sports-bad-society/. 47. Jan Janssens and Paul Verweel, “The Significance of Sports Clubs Within Multicultural Society. On the Accumulation of Social Capital by Migrants in Culturally ‘Mixed’ and ‘Separate’ Sports Clubs,” European Journal for Sport and Society 11, no. 1 (2014): 35–58, https://doi.org/10.1080/16138171.2014.11687932.

64 BREAD AND CIRCUSES LYMAN STONE

48. Ørnulf Seippel, “Sports in Civil Society: Networks, Social Capital and Influence,”European Sociological Review 24, no. 1 (February 2008): 69–80, https://doi.org/10.1093/esr/jcm035. 49. Isao Okayasu, Yukio Kawahara, and Haruo Nogawa, “The Relationship Between Community Sport Clubs and Social Capital in Japan: A Comparative Study Between the Comprehensive Community Sport Clubs and the Traditional Community Sports Clubs,” International Review for the Sociology of Sport 45, no. 2 (June 2010): 163–186, https://doi.org/10.1177/1012690210362027. 50. Salma Mousa, “Building Social Cohesion Between Christians and Muslims Through Soccer in Post-ISIS Iraq,” Science 369, no. 6505 (August 2020): 866–70, https://science.sciencemag.org/content/369/6505/866. 51. Fred Coalter, “Sports Clubs, Social Capital and Social Regeneration: ‘Ill-Defined Interventions with Hard to Follow Outcomes’?,” Sport in Society 10, no. 4 (June 2007): 537–59, https://doi.org/10.1080/17430430701388723. 52. Bosco C. Rowland et al., “Is Alcohol and Community Sport a Good Mix? Alcohol Management, Consumption and Social Capital in Community Sports Clubs,” Australian and New Zealand Journal of Public Health 39, no. 3 (June 2015): 210–15, https://doi.org/ 10.1111/1753-6405.12280. 53. Łukasz Skrok et al., “Impact of Sports Activity on Polish Adults: Self-Reported Health, Social Capital & Attitudes,” PLOS One 14, no. 12 (December 2019), https://doi.org/10.1371/journal.pone.0226812. 54. Ute Schüttoff et al., “Sports Participation and Social Capital Formation During Adolescence,”Social Science Quarterly 99, no. 2 (June 2018): 683–98, https://doi.org/10.1111/ssqu.12453. 55. L. Ottesen, R. S. Jeppensen, and B. R. Krustup, “The Development of Social Capital Through Football and Running: Studying an Intervention Program for Inactive Women,” Scandinavian Journal of Medicine & Science in Sports 20, no. 1 (April 2010): 118–31, https:// doi.org/10.1111/j.1600-0838.2010.01123.x. 56. Matthew Lowe, “Types of Contact: A Field Experiment on Collaborative and Adversarial Caste Integration” (working paper, Center for Economic Studies and Ifo Institute, Munich, Germany, 2020), https://www.econstor.eu/bitstream/10419/215091/1/cesifo1_ wp8089.pdf. 57. Laura Langbein and Roseana Bess, “Sports in School: Source of Amity or Antipathy?,” Social Science Quarterly 83, no. 2 (June 2002): 436–54, https://doi.org/10.1111/1540-6237.00093. 58. Sabine Trepte, Leonard Reinecke, and Keno Juechems, “The Social Side of Gaming: How Playing Online Computer Games Creates Online and Offline Social Support,”Computers in Human Behavior 28, no. 3 (May 2012): 832–39, https://doi.org/10.1016/j. chb.2011.12.003. 59. Michal Bauer et al., “Can War Foster Cooperation?” (working paper, National Bureau of Economic Research, Cambridge, MA, June 2016), https://www.nber.org/papers/w22312. 60. Emilio Depetris-Chauvin, Ruben Durante, and Filipe R. Campante, “Building Nations Through Shared Experiences: Evidence from African Football” (working paper, National Bureau of Economic Research, Cambridge, MA, May 2018), https://www.nber.org/ papers/w24666. 61. Thomas Colson, “Ranked: How Patriotic 19 World-Leading Economies Are,” Business Insider, November 23, 2016, https://www. businessinsider.com/yougov-19-most-patriotic-countries-in-the-world-2016-11. 62. Remco Hoekman, Jan-Willem van der Roest, and Hugo van der Poel, “From Welfare State to Participation Society? Austerity Measures and Local Sport Policy in the Netherlands,” International Journal of Sport Policy and Politics 10, no. 1 (2018): 131–46, https:// doi.org/10.1080/19406940.2017.1381636. 63. Jean Harvey, “Sport Policy and the Welfare State: An Outline of the Canadian Case,” Sociology of Sport Journal 5, no. 4 (1988): 315–29, https://doi.org/10.1123/ssj.5.4.315. 64. Klaus Heinemann, “Sport and the Welfare State in Europe,” European Journal of Sport Science 5, no. 4 (2005): 181–88, https://doi. org/10.1080/17461390500344347. 65. Ryan T. Cragun, John Stinespring, and Andrew Tillman, “Sunday Football or Church? A Case Study in Substitutes and Comple- ments,” Review of Religious Research 61 (2019): 169–87, https://link.springer.com/article/10.1007/s13644-019-00367-0. 66. David E. Prince et al., “When Church Gets Sidelined by Youth Sports,” Christianity Today, May 21, 2018, https://www. christianitytoday.com/pastors/2018/may-web-exclusives/when-church-gets-sidelined-youth-sports.html; and Amy Jacober, “Why Kids

65 BREAD AND CIRCUSES LYMAN STONE

Choose Sports over Youth Ministry,” Youth Specialties, August 28, 2019, https://blog.youthspecialties.com/why-kids-choose-sports- over-youth-ministry/. 67. Rick Paulas, “Sports Stadiums Are a Bad Deal for Cities,” Atlantic, November 21, 2018, https://www.theatlantic.com/technology/ archive/2018/11/sports-stadiums-can-be-bad-cities/576334/; Jon Sanders, “Publicly Funded Stadiums,” John Locke Foundation, January 2020, https://www.johnlocke.org/policy-position/publicly-funded-stadiums/; and Michael Wornow, “How Sports Teams Exploit City Budgets to Fund Stadiums,” Harvard Political Review, May 21, 2018. 68. Lyman Stone, Red, White, and Gray: Population Aging, Deaths of Despair, and the Institutional Stagnation of America, American Enterprise Institute, June 11, 2019, https://www.aei.org/research-products/report/red-white-gray-aging-despair-stagnation/. 69. Douglas Hartmann and Brooks Depro, “Rethinking Sports-Based Community Crime Prevention: A Preliminary Analysis of the Relationship Between Midnight Basketball and Urban Crime Rates,” Journal of Sport and Social Issues 30, no. 2 (May 2006): 180–96, https://journals.sagepub.com/doi/10.1177/0193723506286863. 70. Patrick Sharkey, Gerard Torrats-Espinosa, and Delaram Takyar, “Community and the Crime Decline: The Causal Effect of Local Nonprofits on Violent Crime,”American Sociological Review 82, no. 6 (2017): 1214–40, https://www.rootcausecoalition.org/wp- content/uploads/2018/08/Community-and-the-Crime-Decline-The-Causal-Effect-of-Local-Nonprofits-on-Violent-Crime.pdf. 71. Brian J. Asquith et al., “Social Capital and Labor Market Networks” (working paper, National Bureau of Economic Research, Cambridge, MA, January 2019), https://www.nber.org/papers/w23959. 72. Jonathan Rauh, “Is Trust in Government Really Declining? Evidence Using the Sequential Probability Ratio Test,” Acta Politica (April 2020), https://link.springer.com/article/10.1057/s41269-020-00163-7. 73. Jesper L. Boldsen, “Leprosy in Medieval Denmark—Osteological and Epidemiological Analyses,” Anthropologischer Anzeiger 67, no. 4 (December 2009): 407–25, https://doi.org/10.1127/0003-5548/2009/0031. 74. Ying-Chih Chuang et al., “Social Capital and Health-Protective Behavior Intentions in an Influenza Pandemic,”PLOS One 10, no. 4 (April 2015), https://doi.org/10.1371/journal.pone.0122970; Björn Rönnerstrand, “Social Capital and Immunisation Against the 2009 A(H1N1) Pandemic in Sweden,” Scandinavian Journal of Public Health 41, no. 8 (December 2013): 853–59, https://doi.org/ 10.1177/1403494813494975; Björn Rönnerstrand, “Social Capital and Immunization Against the 2009 A(H1N1) Pandemic in the American States,” Public Health 128, no. 8 (August 2014): 709–15, https://doi.org/10.1016/j.puhe.2014.05.015; Nobutoshi Nawa and Takeo Fujiwara, “Association Between Social Capital and Second Dose of Measles Vaccination in Japan: Results from the A-CHILD Study,” Vaccine 37, no. 6 (February 2019): 877–81, https://doi.org/10.1016/j.vaccine.2018.12.037; and Giuseppe Nicola Giordano, Jan Mewes, and Alexander Miething, “Trust and All-Cause Mortality: A Multilevel Study of US General Social Survey Data (1978–2010),” Journal of Epidemiology and Public Health 73, no. 1 (2019): 50–55, https://jech.bmj.com/content/73/1/50. 75. Eric Klinenberg, Heat Wave: A Social Autopsy of Disaster in Chicago (Chicago: University of Chicago Press, 2002). 76. Centers for Disease Control and Prevention, “1918 Pandemic (H1N1 Virus),” https://www.cdc.gov/flu/pandemic-resources/ 1918-pandemic-h1n1.html. 77. Alina Kristin Bartscher et al., “Social Capital and the Spread of COVID-19: Insights from European Countries,” IZA Institute of Labor Economics, May 2020, http://ftp.iza.org/dp13310.pdf. 78. John M. Barrios et al., “Civic Capital and Social Distancing During the Covid-19 Pandemic” (working paper, National Bureau of Economic Research, Cambridge, MA, June 2020), https://www.nber.org/papers/w27320. 79. Christos A. Makridis and Cary Wu, “Ties That Bind (and Social Distance): How Social Capital Helps Communities Weather the COVID-19 Pandemic,” Social Science Research Network, May 2020, https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3592180. 80. Wenzhi Ding et al., “Social Distancing and Social Capital: Why U.S. Counties Respond Differently to COVID-19” (working paper, National Bureau of Economic Research, Cambridge, MA, June 2020), https://www.nber.org/papers/w27393. 81. Georgy Egorov et al., “Divided We Stay Home: Social Distancing and Ethnic Diversity” (working paper, National Bureau of Eco- nomic Research, Cambridge, MA, May 2020), https://www.nber.org/papers/w27277. 82. Leonardo Bursztyn et al., “Misinformation During a Pandemic” (working paper, Becker Friedman Institute, Chicago, September 1, 2020), https://bfi.uchicago.edu/working-paper/2020-44/. 83. Pew Research Center Forum on Religion & Public Life, “2. U.S. Adults Regularly Turn to a Variety of Activities to Help Cope with

66 BREAD AND CIRCUSES LYMAN STONE

Coronavirus Outbreak,” August 7, 2020, https://www.pewforum.org/2020/08/07/u-s-adults-regularly-turn-to-a-variety-of-activities-to- help-cope-with-coronavirus-outbreak/. 84. Pew Research Center Forum on Religion & Public Life, “3. Online or in Person, Many Americans Have Heard Sermons Address- ing Timely Political and Social Events,” August 7, 2020, https://www.pewforum.org/2020/08/07/3-online-or-in-person-many-americans- have-heard-sermons-addressing-timely-political-and-social-events/. 85. Megan Brenan, “Americans’ Mental Health Ratings Sink to New Low,” Gallup, December 7, 2020, https://news.gallup.com/ poll/327311/americans-mental-health-ratings-sink-new-low.aspx. 86. Hayagreeva Rao and Henrich R. Greve, “Disasters and Community Resilience: Spanish Flu and the Formation of Retail Coopera- tives in Norway,” Academy of Management Journal 61, no. 1 (2018): 5–25, https://journals.aom.org/doi/full/10.5465/amj.2016.0054. 87. Arnstein Aassve et al., “Epidemics and Trust: The Case of the Spanish Flu,” Health Economics 30, no. 4 (April 2021): 840–57, https://onlinelibrary.wiley.com/doi/full/10.1002/hec.4218. 88. Qing Miao and David Popp, “Necessity as the Mother of Invention: Innovative Responses to Natural Disasters” (working paper, National Bureau of Economic Research, Cambridge, MA, July 2013), https://www.nber.org/papers/w19223; and W. Kip Viscusi and Rich�- ard J. Zeckhauser, “National Survey Evidence on Disasters and Relief: Risk Beliefs, Self-Interest, and Compassion” (working paper, National Bureau of Economic Research, Cambridge, MA, October 2006), https://www.nber.org/papers/w12582. 89. Stone, Promise and Peril. 90. Stone, Promise and Peril. 91. Luigi Guiso, Paola Sapienza, and Luigi Zingales, “Long Term Persistence” (working paper, National Bureau of Economic Research, Cambridge, MA, August 2008), https://www.nber.org/papers/w14278. 92. Ipek Göçmen, “The Role of Faith-Based Organizations in Social Welfare Systems: A Comparison of France, Germany, Sweden, and the United Kingdom,” Nonprofit and Voluntary Sector Quarterly 42, no. 3 (June 2013): 495–516, https://doi.org/10.1177/ 0899764013482046. 93. Social Capital Project, United States Congress Joint Economic Committee, Multiple Choice: Increasing Pluralism in the American Education System, December 10, 2019, https://www.jec.senate.gov/public/index.cfm/republicans/analysis?ID=BFFED882-7A9E-4462- AC0B-69ED4E67D663. 94. Social Capital Project, United States Congress Joint Economic Committee, The Wealth of Relations: Expanding Opportunity by Strengthening Families, Communities, and Civil Society, April 30, 2019, https://www.jec.senate.gov/public/index.cfm/republicans/2019/4/ the-wealth-of-relations#_ednref65. 95. Organisation for Economic Co-operation and Development, “Average Annual Hours Actually Worked per Worker,” https://stats. oecd.org/Index.aspx?DataSetCode=ANHRS.

© 2021 by the American Enterprise Institute for Public Policy Research. All rights reserved. The American Enterprise Institute (AEI) is a nonpartisan, nonprofit, 501(c)(3) educational organization and does not take institutional positions on any issues. The views expressed here are those of the author(s).

67