Posts tagged: social trust

US Social Trust Has Fallen 23 Points Since 1964

Sorry for the hiatus. I keep doing podcasts and articles on other sites for Trust in a Polarized Age, and I don’t want to irritate my readers on social media with too much content. When I get done talking about the book so often, I’ll return to my regularly scheduled blogging.

Lots of people talk about the US decline in social trust. I do all the time. The data I use comes from the General Social Survey, which shows a drop of 46% in 1972 when it began to 31.5% in 2018, and the drop is fairly graduate. But the World Values Survey, taken every 4-6 years in the US, shows a 7 point decline between 1981 to 2017, and there’s no WVS survey between 1981 and 1998 that asks the standard trust question (“Can most people be trusted or can you never be too careful in dealing with people?”) in the US. Also, people sometimes say the WVS is done poorly. So I mostly ignore the WVS and go with the GSS.

However, foolishly, I didn’t realize the awesomely done American National Election Survey (ANES) asked the standard trust question,  starting in 1964. And it has a different pattern. Trust is around 55% in 1964, around 46% in 1972 (matching the GSS), and there’s a gradual decline, but for some reason 2000 and 2002 have social trust at 1964 levels. Then in 2008, trust plummets to just under 30%. Then they stopped asking the yes-no question and went to a five point scale in 2012 and 2016, and there are difficulties translating the 5-point scale (people can be trusted all the time, some of the time, none of the time, etc) into simple percentages. But it looks like trust remains quite low and maybe falls a bit further.

One reason it is cool to have both the GSS and the ANES is because the US was asking the trust question long before researchers were asking it all around the world. Some countries are just getting in the game in the 2000s, and no one but the US seems to have data before 1980.

I don’t know what’s going on, so I just plotted all the years we have data from any of the three surveys, and when I had multiple data points from one year, I just averaged them. From what I can tell, this is the only chart I’ve seen that plots social trust with an ANES-GSS-WVS average. And it is rather striking: a 23-point decline.

Now, other countries have great, big declines in social trust, but this is typically due to a major institutional transition from dictatorship to democracy (Spain, Chile, Romania, and Poland), or big increases, which typically follow periods of institutional stability (Germany and adjacent countries like Denmark, the Netherlands, and Switzerland). But the US hasn’t had anything like that level of institutional instability. So the US is an international outlier because we’ve had a sharp drop in social trust despite institutional stability. I think part of the reason for this is that political polarization, trust in government, and interpersonal trust are in a complex causal feedback loop, as I argue in the introduction to Trust in a Polarized Age.

So, here’s what I think is the most data-rich graph of where we stand (data). It is quite bad. Falls in social trust have big costs.

The Goal of Police Reform Must Be To Restore Trust Between Police and Citizens

One of the only promising political developments as of late is the possibility of policing reform, in some way to change the incentives that the police face in order to reduce police brutality. I will say, however, that I worry that people aren’t thinking far enough in advance of what a desirable new institutional equilibrium looks like. What sort of enduring relationship do we want there to be between police and the citizens they are sworn to protect?

The relationship we want is mutual trust. We want to be able to trust police to enforce the law (or at least most laws, besides grossly unjust and absurd ones), and to otherwise abide by ordinary moral norms. And we want police to be able to trust most people so they don’t react in suspicious and harmful ways without cause. That’s absolutely key: we want both groups to be able to trust each other. High trust and trustworthiness is a kind of equilibrium, and everyone is more at ease and can focus more on positive projects and forms of life when trust is high.

The worry I have about the current discussion is that the cited goals don’t seem to be to restore trust but to destroy it and make it harder than ever to restore. Violent protests and continued policy brutality are leading to fewer police and police less willing to do their jobs. This means crime will increase because of police inaction, and police brutality may not be reduced as a result. What we want is more cops on the beat, not fewer, since police presence seems to have a clear negative effect on crime. But we also want the behavior of police to change to resemble that more like in Western Europe, with less militarization, fewer violent weapons, and less use of physical force. We want police to police well, and we should want to be able to trust them to police and police well.

The first step in any reform, then, is to ensure that the police are given the right incentives to be trustworthy, in particular by acting within the norms of ordinary moral behavior and the public’s moral expectations about permissible police use of force. Ending qualified immunity could be a step in this direction because the penalties for acting outside of the moral norm would increase. The second step is to reform policing so that police are taught to be more trusting, and less suspicion and desiring to dominate others. With a large public reform in this direction, coupled with benefits for police who are more trusting and trustworthy would be a big step in the right direction. But policy that is punitive, that leads the police to feel despised and untrusted, and so with little incentive to reform, is likely to produce worse policy outcomes.

Part of trust is believing that trustees are acting for moral reasons, so we do not want reforms merely aimed at beating cops down and appealing solely to their self-interest. We should also aim at policy that encourages police to act on their conscience, rather than penalizing them for doing so. If we think police are only behaving out of fear, then they will misbehave when they think they can get away with it. But incentives for acting morally can generate more stable behavior because the desired behavior comes from within.

It turns out that trust in the legal system is closely correlated with social trust, though we don’t know why just yet. My sense is that police are often seen as exemplary community members, and so when trust in police falls, trust in most people can fall, at least among younger people who are still deciding how trusting to be. We very much don’t want this to happen, since social trust has enormous benefits. So another reason to support police reform is the potentially positive effect on social trust.

So, when formulating police policy, please keep the end goal in mind: we want to trust police to enforce the law morally. Our goal should not be to punish the police or destroy them, but to reorient their incentives, discourage bad behavior, and encourage good, trustworthy behavior. As trustworthiness rises, trust can rise as well, creating a mutually reinforcing cycle. The present attitudes towards the police, however, seem destined to reduce trust in the police and reduce police presence, which will hurt everyone. We need policing reform, but it must be guided by the goal of building community trust.


Why are Trump Supporters Demonstrating Against the Economic Lockdowns?

We are starting to see more polarization of opinion about the lockdowns, with Republicans returning to their original skepticism about the dangers of the virus. If a Democrat were President, this would be unsurprising. But Trump is in office, and people tend to approve and trust government more when their favored presidential candidate is in power. So what’s going on? Here are some hypotheses. I’m not satisfied with most of them.

1. Trump supporters trust the government and public health officials less than non-Trump supporters, and so are less likely to believe in official recommendations. Problem: there’s lots of skepticism about public health in minority communities, especially in the black community (and with good reason in some cases). But we don’t see protests there.

2. Trump supporters are disproportionately bearing the economic costs of the lockdowns, given that they tend to hold jobs that do not require college degrees. Problem: this is also a feature of many minority communities, and minority communities are getting harder hit by fatalities than rural whites.

3. Trump supporters sense Trump’s displeasure with the lockdowns, and hypothesize that he’s being misled by experts. This is where all the #fireFauci stuff is coming from. In general, Trump is sending mixed messages, at best, and people in his tribe are responding accordingly. Problem: I’d expect more ambivalence in their views if this were true.

4. Trump supporters are deeply anti-elitist in general, moreso than people on the left, and since elites are supporting lockdowns, Trump supporters are opposing them. Problem: there are anti-elitist leftists. Heard of Berniebros?

5. Trump supporters are seeing others members of the red tribe protesting, but the protests are being driven by political groups looking to expose weaknesses in Democratic state governors and create a groundswell of support for GOP candidates. Trump supporters then infer that their group is generally skeptical of the lockdowns and concerned with the economic costs and act accordingly. Problem: But why are people so eager to agree with the protestors in the first place? The resentment seems genuine, not like astroturf.

Hypotheses I’m more satisfied with:

6. Conservative and libertarian intellectual and policy elites chafe more at the greatly expanded power of government and the restrictiveness of the lockdowns, and have been challenging a lot of the flawed models and data shaping elite opinion. This is trickling down to grassroots people on the right through right-wing media.


7. Distrust of Mass Media: Trump supporters disproportionately distrust mass media, which is to say they don’t really trust it at all. But non-Trump supporting anti-elitists, vaccine-skeptics, etc. tend to trust what the mass media tells them. Since mass media is largely conveying a pro-lockdown message, Trump supporters are inclined to disbelieve them or even believe the opposite.

Anything I’m missing?

The Queen Builds Trust

Monarchies tend to be more trusting than non-monarchies. Why? One hypothesis is that societies with non-partisan leaders, who are “above the fray,” have the unique ability to remind everyone of the common interests of the tribe/nation. That’s one reason Queen Elizabeth’s amazing message to the British people is so effective. She speaks to all of the British people about their common concerns, about how they need one another, while taking no partisan stances.

It helps that, as she points out, she addressed the nation during World War II, as a teenager. This was a time when the Brits say “everyone did their bit,” activities that I think played a major trust-building role during and after the war.

The Queen also thanks people in whom people already trust a great deal, like nurses and doctors.

Further, she emphasizes how many people are observing important social norms and succeeding in tackling the virus. And she stresses that people will want to look upon their own actions favorably in the future, again stressing the importance and motivation for following important norms, like staying home, washing one’s hands, and so on.

In short, the Queen builds trust by being a long-time non-partisan, tribe-unifying, trust-reminding, norm-cueing, and compliance-motivating high-status trendsetter. From a trust-building perspective, she’s got it all!

Trump cannot be this kind of leader, not just temperamentally, but because he polarizes public opinion perhaps more than any figure in American politics in historical memory. The difference in leadership ability could not be more stark.

If you haven’t watched the video, take a few minutes and watch it.

Is *Anything* Rotten in the State of Denmark? It Has Become Far More Trusting. We Don’t Know Why.

Denmark is the only major developed country where social trust has substantially increased since measurement began: from 47% in 1979 to 76% in 2008, converging with the high levels of trust in Norway and Sweden. Here’s something else: Germany borders Denmark and shares many cultural and institutional features with it, but German social trust in 2008 was 39%. So what could account for a difference over over 35 points? This 2012 paper suggests that the difference is due to greatly differing degrees of political stability. Denmark simply saw far less political instability during this time period.

My theory is that social trust is grounded in the observation of compliance with social norms.* If political instability generates more observable norm violations, or undermines stable norms, then the capacity to learn to trust other is greatly limited.

So here’s a thought: perhaps Denmark would ordinary have social trust levels at Scandinavian levels but trust was suppressed through somewhat less political stability and because of its proximity to Germany, whose recent institutional history is quite chaotic. Once Germany stabilized, Danish trust could converge with other countries.

This may suggest that German social trust should gradually approach Scandinavian social trust, and it might be at similar levels if not for Nazism, division under communism, and a difficult reintegration period. We don’t see that in the data yet, but that might be because social trust in East Germany remains low.

*The authors seem to agree. They say that the “stable ‘underlying rock’ of social trust in any society, including a political stability secured by formal institutions that are firmly embedded in shared norm” (355)).



My Recent Podcasts on Trust, Polarization, Liberalism, and Must Politics Be War?

In case you’re driving around and have a chance to listen to some podcasts, I recorded two in the last few weeks on The Philosophy Guy and The Curious Task. We talk about the book, but also a whole range of related issues, like the nature and sources of political polarization, the relationship between liberalism and ideology, the contribution to mistrust by the GOP, the relationship between classical liberalism and the cause of avoiding ideology, along with the challenges creating social trust. They’re fun. I cut a bit more loose than usual on both.



My Next Book – A Liberal Democratic Peace: Creating Trust in Polarized Times

I am pleased to announce that my next book, A Liberal Democratic Peace: Creating Trust in Polarized Times, will be published late this year with Oxford University Press. It is the data-driven sequel to Must Politics Be War? I argue in this book that specific liberal rights practices can not only be morally justified, but create social and political trust in the real world with real people. I focus primarily on freedom of associations, markets, social insurance, quality of governance, and democratic elections.
I wrote MPBW primarily for philosophers and political theorists, but this book is aimed more broadly at political scientists, economists, and policy people. It has philosophical argumentation, but I mark out where it begins and ends so that non-philosophical readers can profit from my overall argument. This book will also be much cheaper, under $30, so if you’re not a philosopher or political theorist who wants to learn about how we can build trust in diverse societies, then this book is for you.
Here’s another part of my pitch. I think that mistrust and polarization are in a causal feedback loop, and so those of you who are interested in addressing polarization may find the book of interest as well. If there are laws and policies that can increase trust, then perhaps we can contain the more destructive aspects of political polarization.

Can Nationalism Promote Trust?

I’m fond of the claim that liberal institutions can and do create social and political trust, but sometimes I wonder whether nationalism can too. Well, it turns out there’s now some evidence in favor of the latter claim, from a recent paper from Christian Bjørnskov, Martin Rode, and Miguel Ángel Borrella Mas. The process of nation-building surrounding secession in Catalonia probably increased social trust.

Here’s the abstract:

Consequences of social trust are comparatively well studied, while its societal determinants are often subject to debate. This paper studies both in the context of Catalan attempts to secede from Spain: First, we test if Catalonia enjoys higher levels of social capital that it is prevented from capitalizing on. Second, the paper examines whether secessionist movements create animosity and political divisions within society that undermine trust. Employing the eight available waves of the European Social Survey for Spain, we show that social trust levels are not higher in Catalonia than in the rest of the country. However, we find indications of a significant regional increase after secession became a real option in 2014. We argue that this finding is a likely result of the mental process of nation building, indicating that the formation of social trust may best be thought of as a stable punctuated equilibrium.

The authors argue that “Catalan social trust has not declined as a result of the secessionist conflict, as argued by the unionist side of the discussion, but has rather increased significantly after 2014,” a difference equivalent to the difference between trust in the Netherlands and Sweden. The authors, of course, don’t argue for nationalism or deliberate nation-building, but we do have at least a bit of evidence that when states and societies deliberately try to build new identities, or perhaps to rediscover old ones, that can increase social trust. This may also buttress the possibility that large-scale immigration will be trust-decreasing insofar as it undermines shared identity. I’m not happy about either of these results, as they make trouble for my thesis, but I’m obligated to report it.