In Data We Trust

The thing I care most about is trust; more than social capital, more than even happiness. Like both, it is hard to measure but I think it's worth trying. I think a lack of trust is the reason we lead the world in tax avoidance.

The institute Our World in Data has some, well, data.


Unsurprisingly poverty seems to make a difference; not much trust among poor countries. Those Nordic countries continue to beat everyone in the rankings we all care about. But Saudi Arabia (50%) and China (62%) surprised me.

I don't like the question "'most people can be trusted' agree or disagree" as I find it too vague. Someone in rural West Virginia might trust every person in her community, but not trust Mexicans, Muslims, or Silicon Valley liberals. Trust doesn't necessarily scale, so I think the question should be more specific: "Do you trust your neighbors?" vs. "Do you trust most Americans?"

(edit: I looked at the social capital index data, which gave a state by state percentage response of U.S. citizens who trust all or most of their neighbors. The average state's response was almost 60%. Of course that treats each state equally, so I looked the most populous ones, including California, Texas, New York, and Florida. They were all between 45%-49%. Less than the average of 60% but still higher than the Our World in Data figure of 38%. So framing the question matters.)

I am also discouraged by self reported data. I'd be interested to see methods that measure behavior, like how likely a citizen is to return a lost wallet. But this is the data we have, so let's dive in.

It's discouraging that most countries score poorly in trust. Only 9 countries score above 50% , so let's focus on those and look for patterns. In order:

1. Norway 73%
2. Netherlands 66%
3. Sweden 63%
4. China 62%
5. Finland 57%
6. New Zealand 56%
7. Australia 54%
8. Vietnam 51%
9. Saudi Arabia 50%

My hypothesis is that two factors cause trust in developed countries: size and homogeneity. I think trust doesn't scales to the size of bigger countries and when it does, unfortunately, it's because people look and think the same. Pew research shows that rural Americans are more trusting that urban Americans, supporting my "trust doesn't scale" bias.


Here is a chart of ethnic and cultural diversity. The darker the country, the more diverse it is.


I would have expected the U.S. to be one of the most diverse countries. One of the metrics looks at things like languages spoken, so it might more heavily favor different cultures within an existing country—especially if the borders were drawn by English colonists—and not immigration patterns. This map shows net migration by country (where blue is positive, orange is negative, and green is stable), but it doesn't tell us much other than people leave poor countries and come to rich countries.

We know that China, India and the U.S. are the most populous countries. We can now see that India and the U.S. have high(er) diversity and low trust, at least among advanced economies. However, China has low diversity and high trust.

The Nordic countries are small and have low diversity and high trust. Saudi Arabia has similar diversity to the U.S. but much higher trust, so I don't know what to make of that. Canada is not far off from scoring high in both rankings. Indonesia is very diverse and has 41% trust, so that's something.

None of our nine high-trust countries fall into the top two tiers of high diversity, although Saudi Arabia is close. So diversity might discourage trust.

What about the opposite? Here are the countries in the bottom tier of diversity not already mentioned that are developed economies and I can actually read on the map (trust % next to name):

• Poland 22%
• Germany 42%
• Italy 28%
• Japan 35%
• South Korea 29%

So they aren't doing great for trust, but Germany is pretty close to 50% and the other four are still scoring better than most of the world. Ranking by GDP, Japan, Italy, South Korea, Poland and Germany are all in the top 25, so I don't think a poor economy is causing their distrust. 

(For what it's worth, China (2), Australia (14), Netherlands (17), Saudi Arabia (20), Sweden (23), Norway (29), and Finland (42), Vietnam (46), and New Zealand (50) are all top 50 GDP countries.)

Income Inequality

I think wealth matters to a point, after that, if there is more income inequality it will decrease trust. Here is a map of income inequality. The darker the blue, the more equal incomes are. Lighter blue is more unequal.

No developed countries score above 50 for income inequality. In the 40-45 range you see the U.S., Russia, and China, the latter being the only place with high trust.

Starting in the lowest range of income inequality (20-30), you get all the Nordic countries plus some poor areas, like Kazakhstan. Next tier (30-35) you start to see Canada, Australia, South Korea, Vietnam, and most of western Europe.

Unfortunately there is no data for New Zealand, Japan, or Saudi Arabia, but so far this looks like a good predictor for trust. This paper seems to think so. This one, however, finds that "Once we control for general wealth, both types of inequality (real and perceived) do not seem to matter much for trust in Western countries."

U.S. Trust

The social capital index has two measures of trust. There is the Social Support Index, which measures, among other things, the percent of people who trust all or most of their neighbors. There is also the Institutional Health Index, which measures confidence in corporations, media, and public schools, as well as voting rate and census participation.

How do these measures relate to income inequality? Here is a U.S. map, the darker the green the more income inequality.

Let's look at top five and bottom five along with their social support ranking in parentheses, with 1 being the best and 50 being the worst:

Most unequal: DC(44), New York(49), Louisiana(47), Connecticut (30), California (50).
Most equal: Utah(1), Alaska (18), New Hampshire(11), Wyoming(16), Hawaii (37).

So Connecticut isn't as terrible as the other four, but still in the bottom half for social support/trust. And Hawaii is the only outlier among the most equal income states, with the other four all in the top 40% for social support/trust.

I felt compelled to look at all states and found that I could download the social capital data and just look at the "trust in neighbors" questions. So here is a chart comparing each state's trust to their income inequality score (I manually drew the trend line, so it might not be perfect).

There are a few outliers like Delaware, Maryland, Nevada, Connecticut, and Rhode Island, but otherwise a pretty strong relationship. So trust in neighbors seems to have a strong correlation with income inequality in the U.S.

I also isolated the data from the Institutional Health Index about trust in corporations to do the right thing. Here is how that compares with income inequality, or what they call gini coefficient.

I'm no statistician, but that looks pretty solid.

Religion

Our World in Data suggests that religious activity can also cause trust. I wonder if it's simply participation in social infrastructure, meaning routine face-to-face interactions with community members. (It's probably too much to expect secular people to start going to a church so they can be more trusting, but they might participate in a civic community organization if the right incentives are in place.)

The below chart shows that both trust and civic engagement are correlated with "peaceful collective decision making."

I tried to chart data from the social capital index that compared each state's trust in neighbors to religious adherents per 1000 residents, but it didn't show any correlation.

Trust in Government

Pew has some data about trust in our government, which tends to be determined by party and depend on the which party controls the White House; Dems are trusting in the Obama years but not W or Trump years; Repubs the opposite.




I'd love to see them ask about the layers of government. "How much do you trust your mayor? Your state senator? Your U.S. senator? Your Governor?" etc. In other words, how does trust in the government scale and how much does partisanship affect it at each level?

Time: Not on our Side

According to Putnam, the largest determinant of social capital was which generation you were born into. As such, trust has been declining over time.

I wish the above image went back further to the social-capital heights of the 1950s. Either way, you have to wonder if the decline in trust is generational values or sorting, more people moving into like-minded communities that become distrustful of their outgroup. And if they are sorting to like-minded communities, wouldn't that increase their trust of neighbors?

I don't think we will ever have Nordic-style social assistance programs unless we can build our trust. Otherwise the billionaires we rely on to fund those programs will keep finding ways to hide their taxable income or continue to hire lobbyist to kill such legislation. And as a diverse, multicultural society, we have to work even harder to build that trust.

I know this is my bias speaking, but I believe that nothing kills trust faster than coercion.