The following are empirical measures of trust/social cohesion in a sampling of different countries.
Countries cluster into stable institutional-trust equilibria observable through compliance behavior, enforcement intensity, and institutional design assumptions. The U.S. now clusters with medium-low trust systems rather than high-trust Northern European or East Asian ones.
| Code | Indicator | What it measures |
|---|---|---|
| TC | Tax compliance efficiency | Revenue collected vs. owed relative to enforcement effort |
| IE | Institutional impartiality | Predictability and neutrality of courts and bureaucracy |
| RC | Rule compliance under low monitoring | Compliance where cheating is easy |
| CE | Contract enforcement reliability | Whether formal contracts are trusted over networks |
| ID | Institutional design assumption | Whether systems presume honesty or evasion |
The data used for the TC metric are from IMF and OECD tax-gap studies.as well as Piketty, Saez, Zucman for administrative tax data.
The data used for the IE metric are derived from World Bank Worldwide Governance Indicators and the EU Justice Scoreboard.
The data used for the RC metric are derived from Large-N lost-wallet experiments, fare-evasion and honor-system studies, Fare-evasion and honor-system studies and bureaucratic self-reporting accuracy studies.
The data used for the CE metric are derived from World Bank Doing Business, comparative commercial-law studies, and firm-level contracting behavior.
The JD metric is the least empirical. It is derived from documentation burden, audit frequency, means testing intensity, and fraud-prevention cost ratios. Short version: the Nordic and Japanese systems assume honesty; the U. S. and Italian systems assume evasion.
| Country | TC | IE | RC | CE | ID | Composite |
|---|---|---|---|---|---|---|
| Denmark | 9.5 | 9.5 | 9.5 | 9.0 | 9.5 | 9.4 |
| Sweden | 9.2 | 9.3 | 9.2 | 8.8 | 9.2 | 9.1 |
| Germany | 9.0 | 9.2 | 9.0 | 9.0 | 8.8 | 9.0 |
| Japan | 8.8 | 8.7 | 9.3 | 8.5 | 9.0 | 8.9 |
| Netherlands | 8.7 | 8.8 | 8.7 | 8.6 | 8.8 | 8.7 |
| Canada | 8.2 | 8.3 | 8.0 | 8.2 | 8.0 | 8.1 |
| United Kingdom | 7.5 | 7.8 | 7.2 | 7.8 | 7.2 | 7.5 |
| France | 7.3 | 7.4 | 6.8 | 7.5 | 6.8 | 7.2 |
| Spain | 6.8 | 6.8 | 6.5 | 6.9 | 6.5 | 6.7 |
| Italy | 6.2 | 6.0 | 6.0 | 6.5 | 5.8 | 6.1 |
| United States | 6.0 | 6.2 | 5.5 | 7.0 | 5.2 | 6.0 |
| Greece | 4.5 | 4.8 | 4.2 | 5.0 | 4.0 | 4.5 |
| Mexico | 4.2 | 4.5 | 4.0 | 4.8 | 3.8 | 4.3 |
| Brazil | 4.0 | 4.3 | 3.8 | 4.5 | 3.5 | 4.0 |
| India | 3.5 | 3.8 | 3.2 | 4.0 | 3.0 | 3.5 |
When institutional trust is measured via revealed compliance behavior, enforcement efficiency, and system design assumptions rather than opinion surveys, the United States clusters with medium-low trust countries such as Italy, well below Germany, the Nordics, and Japan, and well above Greece or Brazil.
Denmark doesn’t just have generous welfare policies. It has a high-trust equilibrium that makes them administratively cheap and fraud-resistant.
We don’t need to speculate about whether Americans would abuse a large, fast-moving welfare state. We just ran that experiment during COVID. Fraud did not occur at the margins; it occurred at industrial scale.
One criticism I have with the study is they compared US with EU states, instead of the EU itself.
Fundamentally, trust / social cohesion works differently at the continental scale — the incentives for graft over a population of 10 million looks different then 500 million. Not to mention its easier to have common bonds of ethnicity, language, religious affiliation or cultural practices with 10 million then 500 million. And you can’t just average the EU state scores, because I think the EU commission seems much more heavy on verify before trust.
Even within US states; there are small states like Vermont which resemble in character the Scandavian countries in social cohesion, and reflected in how their state government is run.
Considering the widespread governmental corruption, trillions stolen every year, the very heavy politicization of the courts (at every level) and of our educational system, and the open alignment of the Democrat party with violent criminals and the Republican with a particular hostile foreign government, the US should rank much lower, in the India-Mexico range.
I find the CATO public corruption index helpful for this assessment. The variation by state really is huge. The worst states, Louisiana, Montana, South Dakota, Kentucky, West Virginia are about 10 times as bad as the least corrupt, New Hampshire, Utah, Oregon, Colorado, Minnesota.
https://www.cato.org/blog/public-corruption-state
Steve
I have a hard time believing some of the big states (lets say ones with more then 10 electoral seats) weren’t in the list of most corrupt.
Then I saw CATO’s benchmark is the number of convictions per capita; which isn’t that useful.
The problem of trust and / or corruption in this country is not of the blatently illegal; but what is legal but immoral.
I agree on the immoral part, however i still think the CATO list is useful. These will be cases that are widely publicized and known by the public. People who frequently see their public officials being prosecuted would be less likely to trust them than people in states who rarely see people being prosecuted. It’s not a direct measure of corruption but it’s probably not a bad proxy. (I think the article specifically notes that Montana and S Dakota had been having a crackdown on corruption so it may have a bit of bias.)
Steve
CATO is grading on the curve. Any serious measure of government corruption would put Louisiana and Illinois neck-and-neck. Which other states have convicted the man who was chairman of the state’s Democratic or Republic Party for 40 years of corruption? Sadly, the actual statistic that would be dispositive is how much actual corruption isn’t treated as a crime?
Also I note that the least corrupt states are more homogeneous demographically than the most corrupt ones. That supports the point of my post. As a country we’re more similar to low-trust societies like Mexico than high-trust ones like Denmark.