I live in Baltimore. A few days ago, I was walking along Roland Park Avenue after leaving the library when I passed one of those small public book boxes people install beside the road. It was full of books. Anyone could open it, take one, replace it later, or leave another book for someone else.
I stopped and looked through it. I found a James Patterson novel I wanted and took it home.
My immediate thought was: this is what a high-trust society looks like.
Nobody was guarding the box. There was no librarian checking identification, no camera pointed directly at it, no deposit, no membership application, no terms of service. Somebody had spent money or time building it. Other people kept it stocked. The neighborhood apparently expected strangers to use it without destroying it.
The box works because enough people believe that other people will behave reasonably. People can take books without taking every book. They can open the door without breaking it. They can leave something useful instead of dumping trash. They can place a small shared resource in public and expect it to remain there.
Then I corrected myself.
This was not evidence that America is a high-trust society. It may not even have been evidence that Baltimore is a high-trust city. It was evidence that this particular neighborhood could sustain this particular arrangement.
Move the same box several miles away and the outcome might be different. It could survive. It could disappear. It could become a place for dumping things. Someone might break its door. Someone might take all the books. The question would depend on the street, the surrounding institutions, how many people feel ownership over the space, how frequently disorder is corrected, and whether residents expect other residents to protect a shared resource.
That made me return to an idea I have been circling for some time: perhaps there are no high-trust countries. There are high-trust communities, neighborhoods, institutions, families, professional groups and social circles. A country receives the label because enough of these smaller systems exist, overlap and reinforce one another.
That claim is probably too absolute. Survey data still shows large and persistent differences between countries. Nordic countries regularly report some of the highest levels of generalized social trust in the world. In some surveys, more than 60 percent of people in Norway, Sweden and Denmark say that most people can be trusted. Japan also seems to retain forms of public trust and social order that are hard to imagine in much of the United States. Other countries report generalized trust in the single digits.
Countries clearly matter. Still, the national label hides too much.
The experience of trust is local. A person can live in a country described as high-trust and spend most of life inside an abusive family, a corrupt workplace or a dangerous neighborhood. Another person can live in a country with weak institutions and still belong to a community where doors remain unlocked, children move freely between homes and people lend money without contracts.
The United States makes this mismatch especially visible. Two neighborhoods separated by ten minutes can operate according to different assumptions about human behavior.
In one neighborhood, residents leave books, furniture, plants or children’s toys outside for strangers. Packages remain on porches. People jog after dark. Parents allow children to walk to a friend’s house. A stranger asking for directions is treated as a person needing directions.
In another neighborhood, a package left outside is understood as an opportunity. People avoid certain blocks at night. Residents install bars, cameras, alarms and motion-sensitive lights. A stranger approaching may be treated as a possible threat before he says anything.
Both neighborhoods exist within the same city, under the same state laws and inside the same country. They do not live in the same trust environment.
What the bookshelf is really measuring
The public bookshelf looks simple, but it contains several expectations.
The person who installs it expects that the structure will survive.
The person taking the book expects it was placed there sincerely and is safe to handle.
Neighbors expect that people using the box will not create noise or disorder around it.
The box is therefore a small piece of trust infrastructure. It converts a series of private possessions into a shared public resource without needing a formal organization to manage every exchange.
It also produces evidence. Each day the box remains intact tells residents that their expectations were reasonable. Every returned book and new donation becomes another tiny vote for cooperation. People see that strangers can contribute to something without direct supervision.
Trust grows through accumulated evidence like this.
A neighborhood does not become high-trust because everyone inside it is naturally virtuous. People observe repeated signs that cooperation is normal, exploitation is limited and disorder will be corrected. They adjust their behavior accordingly.
Someone sees the book box survive and decides to install a community pantry. Someone else leaves a stroller outside a store. A parent allows a child a little more independence.
A small act of trust can generate more trust because people learn that trusting behavior is not foolish.
The opposite process also occurs.
A package disappears. A store allows customers to use a restroom, then stops after repeated theft or drug use. A driver stops helping stranded motorists after hearing stories about robberies staged beside the road.
Each event creates evidence too.
People learn that openness carries a cost. They become more cautious. Stores place ordinary goods behind glass. Tenants demand written records. Drivers install dashboard cameras. Everyone collects evidence in advance because they expect that another person may later deny what happened.
The environment slowly becomes less pleasant, even for people who have never violated anyone’s trust.
This is one reason the description “low-trust society” should not be treated as an insult directed at the character of every person living there. Distrust can be an adaptation to conditions. In some environments, trusting strangers too quickly is dangerous. The person who locks everything, documents every conversation and refuses informal arrangements may simply have learned the local rules.
That behavior then reinforces the environment. People encounter suspicion and respond with suspicion. The honest person is treated as a possible liar. The customer is treated as a possible thief. The company is treated as something that will exploit workers unless forced to behave. Each side can point to previous experiences proving that its caution is justified.
America as a collection of trust environments
The data on American trust fits this local picture better than the simple idea of one national trust level.
The percentage of Americans saying that most people can be trusted fell from around 46 percent in 1972 to 34 percent in 2018. Pew found roughly the same 34 percent level in its 2023–24 work.
That national number already suggests decline. The differences inside the number are more revealing.
One Pew comparison found that 46 percent of people in San Francisco said most people can be trusted, while only 24 percent said the same in Riverside. Earlier research on trust in neighbors also found large gaps by income: about 67 percent among people earning at least $75,000, compared with roughly 37 percent among people earning below $30,000. There were similarly wide racial differences in whether people felt that most neighbors could be trusted.
These are not small variations around one common American experience. They describe people living with very different evidence about strangers.
Money is part of the explanation, although money alone cannot explain trust. Higher-income neighborhoods usually have more stable institutions, lower exposure to certain forms of disorder, greater political influence, better-maintained public space and more ability to remove people who violate local rules. Residents can also pay for private substitutes when public systems fail.
A clean park may reflect civic-minded residents. It may also reflect reliable sanitation, active code enforcement, private security, homeowners with time to complain, and a tax base capable of responding. Those conditions make cooperative behavior easier.
A neglected public space sends the opposite message. Trash remains because nobody expects rapid removal. More trash appears because the existing disorder signals that nobody is protecting the space. Residents who care may eventually stop trying because their effort is repeatedly erased.
This is why it would be too easy to stand beside the Roland Park book box, compare it with a poorer part of Baltimore and conclude that one group of people has better values.
The neighborhood itself helps produce the behavior.
People care for places that appear capable of being cared for. They defend arrangements they believe other people will also defend. They invest when they expect their investment to last.
Trust needs some protection from exploitation. A book box can absorb one stolen book. It may survive somebody leaving rubbish once. It cannot survive endless misuse with no repair. High-trust systems work because violations remain uncommon or because the community responds before violations become the dominant expectation.
That combination can look effortless from the outside. The calm neighborhood appears naturally calm. Its residents seem naturally trusting. The visible ease may rest on years of accumulated stability, selective membership, wealth, social pressure and effective enforcement.
The same is true at the national level. A high-trust country may look as though its citizens simply possess a stronger civic spirit. That spirit often sits on top of competent institutions, relatively low corruption, predictable public services and a widespread belief that other people are also following the rules.
Trust is easier when rule-following does not feel like unilateral disarmament.
The old days and the memory of trust
People often describe an earlier America as a more trusting place.
Children stayed outside all day. Neighbors watched one another’s homes. People left doors unlocked. A shopkeeper extended credit based on a handshake. Families knew the teachers, police officers and local business owners.
Some of that memory is probably true.
Social life was more local. People spent more of their lives around people they expected to encounter again. Reputation travelled through churches, schools, workplaces and extended families. A person who cheated someone could not easily disappear into a platform containing millions of anonymous users.
Repeated contact supports trust. People behave differently when they know today’s stranger may be tomorrow’s customer, neighbor, in-law or employer.
Smaller communities also had more informal enforcement. Shame, gossip, exclusion and damaged reputation could punish behavior that never entered a courtroom. A person could lose standing in the community without violating a written law.
The internet weakened some of those boundaries. Modern life allows people to move between jobs, cities, relationships, online identities and marketplaces. This freedom has enormous benefits. It also reduces the cost of burning one relationship and moving to the next.
Social media added another layer. People now witness a continuous stream of fraud, crime, political hostility, public humiliation, infidelity, workplace abuse and institutional failure. Events that would once have remained inside one family or town can now reach millions of people before breakfast.
The world did not suddenly become full of bad behavior when cameras arrived. The distribution system changed. Every scandal became local news everywhere.
A person can live on a safe street, know kind neighbors and still spend hours each day consuming evidence that society is collapsing. The mind does not neatly separate direct experience from repeated mediated experience. A hundred videos of theft can change how someone treats the next stranger entering a store, even when that person has never been robbed.
This may be part of why trust feels lower than local crime statistics alone can explain. People receive a national and global feed of reasons to be afraid while living inside a particular neighborhood.
The nostalgia still needs qualification.
The old high-trust town was not necessarily trusting toward everyone. Many communities had strong trust among insiders and deep suspicion toward outsiders. Racial minorities, immigrants, religious minorities, unmarried women, gay people, political dissenters and anyone marked as socially unusual could experience the same town very differently.
The unlocked door is evidence of trust inside a boundary. It does not tell us who was allowed through the door.
The useful distinction here is between bonding trust and bridging trust.
Bonding trust joins people who already see one another as members of the same group. Families, ethnic communities, churches, alumni groups and close neighborhoods can have very strong bonding trust.
Bridging trust extends cooperation across difference. It allows people to trust strangers, institutions and people outside their immediate group.
A society can possess strong bonding trust and weak bridging trust. People may sacrifice greatly for relatives while viewing everyone outside the family as a potential threat. A neighborhood may organize effectively while resisting newcomers. A religious group may care generously for its members while distrusting the rest of society.
The broadest form of a high-trust society requires people to extend some presumption of good faith beyond people they personally know.
That is harder.
The American puzzle: formal cooperation without confidence
Francis Fukuyama once classified the United States, Germany and Japan as historically high-trust societies. His argument focused on the ability of people to cooperate beyond the family. Trust made it possible to build large voluntary organizations, companies and civic institutions without relying entirely on kinship or state coercion.
By that standard, the United States clearly developed enormous capacities for cooperation. Americans built companies, charities, churches, professional associations, unions, universities and civic groups. Alexis de Tocqueville noticed the same tendency much earlier.
Yet contemporary America often feels deeply suspicious.
People distrust government, media, corporations, universities, police, medicine and one another. The target changes by political group, but the posture remains. Each side believes some major institution has been captured by dishonest or hostile people.
Trust in the federal government was above 70 percent in the late 1950s. By the late 1970s, it had fallen to around 30 percent. It reached 16 percent in 2022 and stood around 22 percent in 2024.
Confidence across major American institutions has also fallen over several decades. Churches, banks, public schools, medicine and higher education have all lost substantial trust.
At the same time, the system continues to function. Money moves. Contracts are enforced. Insurance claims are paid. Courts settle disputes. Ratings allow strangers to enter one another’s cars and homes.
This suggests that a society can maintain high levels of organized cooperation while interpersonal trust declines.
The cooperation becomes more formal, documented and expensive.
A handshake becomes a written agreement.
A neighborly ride becomes a platform transaction with identity checks, GPS tracking, ratings and insurance.
A landlord who once relied on a reference now requests a credit score, background check, deposit and proof of income.
None of these systems is irrational. Many protect people from real harm. Together they reveal how much cooperation now depends on verification.
Lawyers, contracts, cameras, insurance and surveillance function as substitutes for confidence in other people. They allow strangers to transact without needing to believe deeply in one another’s character.
The system says: you do not need to trust the other person because evidence will exist, liability will be assigned and payment will be compelled.
This is a major achievement. It is also expensive.
Fukuyama described low trust as raising transaction costs. More time, money and administration are needed to make cooperation safe. Every exchange carries procedures designed around the possibility of dishonesty.
The social cost is larger than the paperwork. The default relationship changes. People approach one another as potential risks to be managed. America becomes a country of terms and conditions.
When nobody can say what happened
Trust also depends on a basic expectation of empathy.
I do not need every stranger to care deeply about my life. I need to believe that most people recognize some limit on what they are willing to do to me for money, convenience or amusement.
That expectation is exactly what organized fraud attacks. Insurers and rideshare companies have documented schemes in several states where people deliberately stage collisions with rideshare vehicles to collect insurance payouts. In a federal lawsuit filed in New York, Uber alleges that people ordered rides, arranged for associates to cause minor collisions, and then submitted injury claims. The lawsuit identifies eight allegedly staged crashes. Several occurred within minutes of the rides beginning. The allegations have not yet been proven in court.
Think about what the scheme requires. The organizers must treat a stranger’s body, livelihood, vehicle and peace of mind as acceptable collateral. The driver is not a person to them. He is a prop in a transaction he does not know he is part of.
Now think about what the scheme produces, beyond the payout.
In places where accountability is weak and evidence is difficult to obtain, ordinary events become hard to interpret. A collision may have been an accident, recklessness, fraud or deliberate harm. A false statement may have been confusion, malice or an attempt to receive compensation.
When nobody can determine which explanation is true, the uncertainty remains attached to the victim.
That uncertainty does not stay contained inside one event.
A person harmed by an identifiable offender can direct distrust toward that offender. A person harmed by someone who disappears has no clear place to put it. Suspicion spreads toward the location, people who resemble the suspected offender, and strangers encountered under similar circumstances.
The individual event produces a more general social judgment.
This judgment can be unfair. A whole neighborhood may carry the cost of crimes committed by a small number of people. Members of a group may be treated as suspects because someone else escaped accountability. Innocent people inherit suspicion they did nothing to create.
That is one of the social costs of unresolved crime. The harm includes the damaged property, physical danger, lost income and fear. It also includes the distrust transferred to other people afterward.
Petty crime is often discussed mainly through the value of the stolen object. A broken window may cost a few hundred dollars. That accounting misses the social meaning. The offender tells the victim that his time and security do not matter. The victim must repair the damage, make reports, contact insurance, miss work or change routines. The person causing the harm often receives a small benefit while imposing a much larger cost.
The person who steals one item may believe the loss is absorbed by a large company. The company distributes the cost through prices, staffing decisions, store closures and security measures. The burden eventually returns to ordinary customers and workers.
And the defensive systems keep growing. The driver who hears about staged collisions installs a camera facing the passenger compartment. The platform suspends accounts first and investigates later, because passenger safety is real and it cannot know what happened inside the car. Other drivers hear the stories and enter the next ride with more caution.
The camera may establish facts after something goes wrong. It does not create the feeling of being trusted.
High-trust societies are sometimes described as places where people assume strangers are decent. That assumption depends on a belief that serious violations are unusual and that the system can identify and respond to them.
When serious violations repeatedly go unresolved, generalized trust becomes harder to maintain. People stop treating events as isolated exceptions. They begin treating them as information about the environment.
The public book box creates one kind of evidence: strangers can handle something fragile without destroying it.
Unresolved harm creates the reverse evidence: another person can endanger you, disappear and leave you unable to know what happened.
One expands the moral circle. The other contracts it.
Institutions can compensate for distrust
America may be better at compensating for low trust than it is at producing trust.
Insurance is a good example. A driver does not need to trust the stranger who damaged the car. The insurance contract exists so that payment does not depend on the offender’s empathy or willingness to take responsibility.
The legal system serves a similar function. Two parties can cooperate because courts may enforce the agreement. A customer does not need a personal relationship with a corporation because consumer law, payment disputes and reputational pressure offer some protection.
These systems allow a large diverse society to function. People can transact across distance, culture and anonymity.
The presence of strong contracts should therefore not be mistaken for strong interpersonal trust. Sometimes contracts reveal that trust is missing. The parties cooperate because they know there are consequences for betrayal.
This raises a question about how we classify societies.
Should a country count as high-trust because institutions reliably force people to keep promises? Or should high trust mean that people generally expect others to keep promises without needing constant enforcement?
The United States often performs better on the first measure than the second.
A person may distrust the government while trusting that a court will enforce a commercial contract. A customer may distrust a seller while trusting the credit-card dispute system. A driver may distrust another driver while trusting the insurance carrier to pay.
These are forms of institutional confidence. They allow life to continue. They do not produce warmth, solidarity or a presumption of good faith.
This may explain why Americans sometimes describe the country as highly functional and socially broken at the same time.
There is also a pattern in which institutions people still trust. Americans often trust smaller and more local institutions more than large distant ones. One Pew measure found 86 percent saying small businesses have a positive effect, compared with only 29 percent saying the same about large corporations.
People often trust a local shop owner, doctor, pastor, teacher or mechanic because the relationship is concrete. The person has a face. Reputation matters. Poor behavior can be confronted directly. The institution exists inside a social world.
Large systems feel different. Decisions arrive from people nobody can identify. Responsibility is distributed across departments. A customer speaks to an employee who has no authority. An algorithm imposes a decision.
Distance weakens accountability.
The Roland Park book box is perhaps the smallest possible institution. Nobody owns the whole experience. The community sustains it through repeated voluntary actions. It asks very little from each participant and produces visible evidence that other people are contributing too.
This scale may be important. Trust becomes harder as systems grow. Larger systems contain more strangers, more opportunities for exploitation and weaker reputational consequences. They respond by formalizing everything.
A high-trust country is one where trust survives scaling better. A low-trust country is one where cooperation quickly retreats into families, clans, private networks and heavily enforced transactions.
The United States appears somewhere in between. It can build enormous systems for coordinating strangers, but those systems increasingly rely on verification, legal enforcement and technological monitoring.
Is Japan still high-trust?
Japan remains the example that makes me hesitate before declaring that no high-trust countries exist.
There are public behaviors associated with trust: lost property being returned, children navigating cities independently, orderly public transportation, low levels of visible disorder and strong expectations around consideration for others.
Survey data also places Japan above many countries in generalized trust, though the Nordic countries usually rank higher.
The question is what produces this behavior.
It may reflect interpersonal trust. It may also reflect strong norms, shame, social conformity, effective enforcement and fear of imposing on others. These mechanisms can coexist.
A society can be orderly because people trust one another. It can also be orderly because violating expectations carries a high social cost. The lived result may look similar from the outside.
This is another reason the simple scale can mislead. Trust, conformity, institutional competence, social cohesion and low crime are related, but they are not identical.
Every high-trust society has a machinery beneath the visible calm. The question is whether that machinery creates enough repeated evidence that ordinary cooperation is safe.
Which America you inhabit
Calling America a high-trust society now feels too broad.
A person’s answer may depend on where he lives, how much money he earns, what institutions he encounters, whether he belongs to the dominant group, and how much direct disorder he experiences.
The Roland Park resident sees the book box.
The store owner dealing with repeated theft sees locked shelves.
The delivery driver moving through dozens of neighborhoods sees the invisible borders between them.
The immigrant may trust American contracts more than contracts in his home country while finding Americans personally distant.
The Black resident may trust neighbors and family while distrusting police or medical institutions.
Each person answers the trust question from a different America. The national average compresses these worlds into one number.
The phrase “America is a low-trust society” can also become lazy. It can turn into a general complaint about crime, diversity, government or modern life. It can romanticize the past and treat poor communities as the cause of national decline.
The more useful claim concerns fragmentation.
America no longer offers one broadly shared trust environment. It contains high-trust islands surrounded by institutions and places operating through suspicion, documentation and defensive behavior.
Some people spend most of life on the islands. They can sincerely believe that society remains cooperative and safe.
Other people move constantly between trust environments. They know how quickly the rules change from one street, workplace or institution to another.
Trust as accumulated evidence
The thought I keep returning to is that trust is accumulated evidence.
People learn from what happens around them.
Did the package remain on the porch?
Did the employer keep the promise?
Did the police take the report seriously?
Did the public bookshelf survive?
Every answer changes the next interaction slightly.
High-trust communities generate enough positive answers that people can afford to remain open. Low-trust environments generate enough negative or unresolved answers that caution becomes rational.
Institutions influence which evidence people receive. A competent police department can prevent one offender from creating suspicion toward an entire neighborhood. A responsive city can repair vandalized public infrastructure before neglect becomes the dominant signal.
Accountability contains distrust.
When harm is identified, addressed and repaired, people can treat it as an exception. When it remains unresolved, people begin treating it as a feature of the environment.
This may be the bridge between interpersonal trust and institutional trust. Institutions do more than deliver services. They decide whether individual violations remain individual or become general lessons about society.
A provisional conclusion
I started with a public bookshelf in Roland Park and the feeling that I was looking at a high-trust society.
I still think that feeling was accurate. I was looking at trust made visible.
The mistake was the scale.
The box told me something about that block, the people using it and the institutions surrounding them. It did not tell me what every person in Baltimore experiences. It did not prove that the United States shares one culture of trust.
Perhaps the original statement should remain as a provocation: there are no high-trust countries, only high-trust communities.
The research prevents me from treating it as literally true. Country-level differences are real. Nordic societies, Japan and several other countries sustain generalized trust at levels far above the United States and many parts of the world.
The deeper idea still holds.
National trust is built from smaller places where people repeatedly encounter evidence that cooperation is safe, promises matter, violations are corrected and strangers remain inside one another’s moral circle.
When those conditions disappear, formal systems step in. Cameras, contracts, insurance, lawyers, ratings, deposits and background checks make cooperation possible. Society continues functioning, though each interaction becomes more guarded and expensive.
America may now be a country that can enforce cooperation more effectively than it can inspire confidence.
The public bookshelf survives in Roland Park because enough people have reason to believe that it will survive. Several miles away, someone may have equally strong reasons to lock everything down.
Both are rational responses to the evidence available.
That is the part the national label misses.
The boundary between a high-trust and low-trust society may run through the same city. It may run from one neighborhood to another, from the sidewalk into the workplace, or from a private relationship into a public institution.
It may even run through the same person.
Someone can trust neighbors and distrust government. Trust courts and distrust police. Trust a local business and distrust corporations. Trust contracts and distrust the people signing them.
The question may therefore be less about whether America is high-trust or low-trust.
The more revealing questions are: Where does trust still survive? What protects it? Who is included inside it? What happens after someone violates it? How far can it extend before cameras, contracts and coercion have to take over?
The little book box does not answer those questions.
It gives them a physical form.