Surekha has a theory. She has held it for years, articulated it with impressive consistency every time we drive through town, and I am increasingly forced to admit she is almost certainly right. The fancier the car, the worse the driver. Not worse in the mechanical sense — worse in the human sense. More likely to cut you off, less likely to yield, more likely to blow through a crosswalk while making eye contact with you and continuing anyway. BMW and Tesla drivers in particular, she points out. And more than a few Silicon Valley acquaintances — people we knew before they “made it” — who seem to have undergone a subtle but unmistakable personality transplant somewhere between their Series A and their liquidity event.
I used to chalk this up to confirmation bias. Then I stumbled upon the science — and what it told me was not what I expected. This is not mainly a story about bad people. It is a story about a set of forces that make certain behaviors feel entirely justified, even to decent people. That conclusion took six disciplines and some uncomfortable reading to reach. It is, I think, both more disturbing and more useful than the simpler moral version. It is an important story that is increasingly playing out in the AI world today.
Before diving in, one clarification on what follows. This is not an argument that wealthy people are bad people — the evidence does not say that, and neither do I. What the evidence does suggest, across psychology, sociology, neuroscience, physics, and economics, is something more structural: wealth and power systematically reduce the constraints — psychological, social, and institutional — on self-interested behavior, unless actively and deliberately counteracted. The pattern operates through three overlapping mechanisms. Selection: the traits that help certain people rise — confidence, risk tolerance, competitive drive — overlap with traits that, unchecked, shade into entitlement. Transformation: power and wealth, once acquired, appear to alter cognition and values over time. Incentives: the system then rewards and amplifies these tendencies rather than correcting them. No individual bad actor is required. The design does most of the work.
What They Found on the Streets of Berkeley
In 2012, psychologists Paul Piff, Daniel Stancato, Stéphane Côté, Rodolfo Mendoza-Denton, and Dacher Keltner at UC Berkeley and the University of Toronto published a landmark paper in the Proceedings of the National Academy of Sciences (PNAS) titled “Higher Social Class Predicts Increased Unethical Behavior.” They dispatched graduate students to Bay Area intersections with a simple task: note the vehicle, watch the driver.
The findings were striking. Higher-status car drivers were significantly more likely to cut off other vehicles at a four-way stop — in violation of the California Vehicle Code — and in a separate crosswalk study, nearly 35% of drivers failed to yield to a waiting pedestrian, with the failure rate substantially elevated among luxury vehicles. All drivers were coded as having seen the pedestrian.
Across seven studies, upper-class participants were more likely to lie in a negotiation, take candy designated for children, cheat at a dice game to win a prize, and endorse unethical workplace conduct. The paper identifies a key psychological mediator: upper-class individuals tended to hold more favorable attitudes toward greed — viewing it as more justified, beneficial, and morally acceptable. And in a telling final study, Piff found that when lower-class participants were primed to think about the benefits of greed, their unethical behavior rose to match the upper class entirely — suggesting the gap is largely about values held, not fixed character, though this specific finding has proven difficult to replicate consistently.
Piff’s findings have been influential, and — to be very upfront — also genuinely contested. The paper carries a minor statistical correction (2012), and in a 2012 letter in PNAS, statistician Gregory Francis (2012) argued that the consistency of results across seven studies was itself suspicious — a pattern more consistent with publication bias than with a clean underlying effect. Subsequent replication attempts have been mixed: a 2023 naturalistic study at an Austrian crosswalk found no significant relationship between vehicle status and yielding behavior — a direct failure to replicate the traffic finding. Most comprehensively, a 2025 preregistered study by Batruch and colleagues tested 35 hypotheses from the Piff body of work across 33,536 participants in four countries and replicated approximately half the effects — with the mechanisms most tied to greed attitudes and self-versus-other orientation receiving the least consistent support.
The evidence points directionally toward the thesis, but with more uncertainty than the original paper’s results suggested.
The Follow-On: Narcissism, Empathy, and “Having Less, Giving More”
The findings deepened with subsequent work. A 2014 finding by Piff (that has not yet faced the same replication scrutiny as the 2012 paper) found that higher social class is directly associated with narcissism and entitlement — upper-class participants were literally more likely to spend time looking at themselves in a mirror during a laboratory task. Neuroscientists found a biological correlation: high socioeconomic status is associated with diminished neural empathic responses when viewing faces expressing pain. The cruel irony: wealthy participants self-reported being more empathic. Their conscious self-image and their brain’s actual response had diverged. And Piff’s earlier work (“Having Less, Giving More” (2010)) had already established the inverse: lower-class individuals are more generous, more trusting, more prosocial — because interdependence, not independence, has been their lived condition.
Across Professions: Bankers, Boardrooms, and the Valley
The pattern appears across professional domains, though with important caveats about causation. In most cases, what we are observing is a combination of selection — certain personalities are drawn to, and selected into, positions of wealth and power — and transformation — what those positions then do to people over time.
A landmark 2014 Nature study by Ernst Fehr and colleagues at the University of Zurich ran a coin-flip experiment with bank employees. When bankers were primed to think about their professional identity before the task, their self-reported “winning” flips climbed to a statistically improbable 58.2% — against 51.6% when they thought about home and family first. The same experiment on workers in other industries showed no such effect. The banking culture, once activated, appeared to cue dishonest behavior. A 2015 survey of 1,200+ financial professionals found 47% believed their competitors had acted illegally or unethically to gain an edge — a perception that, research suggests, itself lowers ethical standards.
In corporate leadership, Paul Babiak, Craig Neumann, and Robert Hare (2010) studied a sample of corporate professionals and found a psychopathy rate roughly three to four times the general population estimate, though the authors are careful to note that their sample was not randomly drawn. A 2020 meta-analysis of CEO narcissism (Ormonde Cragun, Kari Olsen, and Patrick Wright) found that narcissistic CEOs are associated with more earnings manipulation and more toxic work environments — and that narcissism also predicts attaining leadership positions in the first place. The traits that create problems at the top are often the ones that got people there. A 2023 Frontiers in Psychology study (Michael Olumekor, Muhammad Mohiuddin, Zhan Su) specifically analyzed the Silicon Valley model and found it structurally conducive to what one investigative reporter termed “Machiavellian narcissists,” pointing to the founders of WeWork, Uber, and FTX as cases in point.
Four Forces. No Villain Required.
Before going further, it is worth pausing to name what the evidence actually shows — because the pattern across all these studies is not what most people expect when they start asking why wealthy and powerful people behave the way they do.
There are four forces at work, and they operate simultaneously.
Structure makes the wealthy feel genuinely deserving. Through education, networks, and the accumulated weight of inherited advantage, wealthy people come to experience their privilege as merit. They are not lying when they feel deserving. The system has shaped them, from childhood, to experience it that way. This is Bourdieu’s insight (Distinction (1984)), and it is the most insidious part: there is no moment of conscious self-deception.
Psychology gradually shifts values. Exposure to wealth and elevated status makes greed feel progressively more justified, narrows the circle of moral concern, and raises narcissism — not through a single corrupting event, but through slow conditioning. The person does not decide to care less. Caring less becomes the path of least cognitive resistance.
Neuroscience rewires the hardware. Power suppresses the brain’s empathy circuitry. Hubris Syndrome (David Owen and Jonathan Davidson, Brain (2009)) develops in long-term power-holders. The brain is being run in a different environment, and it adapts accordingly. This is not metaphor — it shows up in brain scans and magnetic stimulation experiments.
Economics makes giving feel like genuine loss. This is the one that surprised me most. Loss aversion (Kahneman & Tversky (1979)) means the psychological pain of handing over $5M is roughly twice as heavy as the pleasure of receiving $5M would be. The endowment effect means money already held feels worth more than money not yet received. So even someone who could easily afford the gift experiences it as a real sacrifice — because the brain is not doing arithmetic, it is doing something far more primitive.
None of these forces require malice. None require a conscious decision to be worse. They are the predictable output of being wealthy and powerful over time, in a system that reinforces rather than corrects these tendencies. That is what makes this a design problem, not a villain problem. And that distinction matters — both for diagnosis and for any hope of remedy.
What the Sociologists Knew First — and Longest
The psychologists are, in some ways, catching up to what sociologists have been saying for over a century.
Thorstein Veblen, in The Theory of the Leisure Class (1899), saw the wealthy class as defined not by productive contribution but by visible exemption from it — and as an institutional brake on social change, its vast investment in existing arrangements making it resistant to anything that would redistribute power. Wright Mills, in The Power Elite (1956), added the structural argument: corporate, military, and political elites form an interlocking, self-referential governing class that circulates among itself and remains largely unaccountable to the public. When three billionaires control more wealth than the bottom half of the country, Mills’s analysis reads less like theory and more like journalism.
Robert Merton’s Matthew Effect (1968) named the compounding mechanism: initial small advantages tend to beget further advantages, requiring no conspiracy, only arithmetic and time. Pierre Bourdieu’s Distinction (1984) delivered the most counterintuitive insight: none of this requires conscious effort. The habitus — the dispositions, tastes, and behaviors that upper-class children absorb so thoroughly they feel like personality — transmits cultural and social capital invisibly across generations. The wealthy child who navigates elite institutions with an almost instinctive ease is not more talented; she is more prepared, in ways that feel, from inside the experience, indistinguishable from merit. Bourdieu called the mechanism by which this domination is accepted as natural “symbolic violence” — and it operates, most durably, precisely because its subjects do not recognize it as violence at all.
Charles Tilly (1998) labeled the active maintenance mechanism “opportunity hoarding” — the systematic exclusion of others from valuable resources through zoning, legacy admissions, unpaid internships, and credentialing. Richard Reeves, in Dream Hoarders (2017), extended the argument to the top 20%, not just the top 1%. And Richard Wilkinson and Kate Pickett, in The Spirit Level (2009), provided the broadest empirical frame: across 11 social indicators — health, mental illness, drug use, imprisonment, violence, trust, social mobility — outcomes are consistently worse in more unequal societies, not just for the poor but across the distribution. The causal story is debated and the cross-national methodology has been challenged, but the direction of the correlations is robust and has held through a 15-year reassessment published in 2024.
What Physics, Medicine, and Economics Confirm
What the psychologists and sociologists describe, in other words, is a human problem — one of values, culture, and institutional design. What the physicists, neuroscientists, and economists describe is something colder: a system that would produce these outcomes even if every individual in it were trying to be good. That is where the story gets genuinely unsettling.
Physicists studying wealth as a statistical mechanics problem — particularly Bruce Boghosian (2019) — have shown mathematically that in a free market with no redistribution, wealth flows inexorably upward. Boghosian calls the endpoint a phase transition to oligarchy: like water turning to ice, sudden, categorical, and without external force, irreversible. The math is not based on cheating or corruption — it models fair exchanges between equal agents. Fairness doesn’t save you. Thomas Piketty arrived at the same place empirically: when the return on capital (r) exceeds economic growth (g), wealth concentrates. This condition has been the historical norm for most of the last three centuries. (Note: the r > g thesis has been contested empirically — the Wikipedia entry covers the main critical responses.)
Medicine adds the neurological dimension. Sukhvinder Obhi and his colleagues at McMaster (2014) used transcranial magnetic stimulation to show that power physically suppresses the brain’s mirror neuron system — the neural substrate of empathy. Neurologist-turned-parliamentarian Lord David Owen proposed the “Hubris Syndrome” as a genuine acquired personality disorder in long-term power-holders: contempt for others, loss of contact with reality, messianic recklessness. He documented it in Lloyd George, Thatcher, George W. Bush, and Blair. Crucially, it typically abates when power is removed — suggesting a brain state induced by unchecked authority, not a fixed character flaw.
Behavioral economics, finally, explains why the wealthy fight so hard to remain so. Kahneman and Tversky’s loss aversion — the finding that potential losses loom psychologically roughly twice as large as equivalent gains — means that the very wealthy have an enormous cognitive stake in protecting what they have. Every dollar of potential loss in a redistribution scheme feels, in the brain, more painful than a dollar of gain would feel pleasurable. Kahneman, Knetsch, and Thaler’s (1991) endowment effect amplifies this: people systematically overvalue what they already possess. Together, these biases create a powerful, neurologically embedded resistance to any claim on accumulated wealth — even ones that, on a ledger, are trivially small.
The Story That Crystallized It All for me
A fellow entrepreneur — a man of impeccable integrity who had built a fine company and engineered a successful exit — was commiserating with me recently. Turns out that the equity structure had left early contributors/employees behind. He went to the board and the investors — all of whom were making a great deal of money — and made a modest ask: set aside a small fraction of the net gain to make these employees whole. He asked for something on the order of $5 million out of a much larger pool.
He got $50,000. One percent of his ask!
The people writing that check were not struggling. They were not being asked to sacrifice. They were being asked, in a moment of shared good fortune, to remember the people who helped create it. And the answer was, essentially, no.
This is not the first time I have heard of or personally experienced investor greed. Naturally, I have been trying to make sense of that answer ever since. Taken together, the science suggests it is not anomalous. It is — across psychology, sociology, neuroscience, physics, and economics — entirely predictable.
Is There a Way Out?
There is — and that matters. Research from Piff’s lab (2012) suggests that prompting reflection amongst upper-class participants on shared humanity can narrow the ethical gap. A 2015 study on the emotion of awe, also from Piff’s lab, found that experiences of genuine vastness — a redwood forest, a great cathedral, the overview effect of seeing Earth from orbit — measurably increased generosity and ethical behavior, particularly among the entitled. Owen’s Hubris Syndrome abates when power is constrained. Boghosian’s mathematical oligarchy can be arrested by redistribution. Bourdieu’s habitus can be disrupted by genuine exposure across class lines.
None of this is fate. But none of it happens without deliberate institutional design — tax structures, democratic constraints, norms of accountability — applied against the natural direction of the system.
Surekha was right about the BMW drivers. It took the scientists a century of work across six disciplines to explain why.
