David Rose on the Moral Foundations of Economic Behavior
What this covers
David Rose, an economist at the University of Missouri, St. Louis, makes the case that large-scale economic prosperity rests on a particular moral architecture. In conversation with Russ Roberts, Rose argues that strangers can only trust each other enough to specialize and trade at scale if they hold a specific kind of moral restraint: not the avoidance of harm, but a principled refusal to commit wrong acts regardless of consequences. The central claim is that [golden opportunity](jargon:golden opportunity) situations—where private knowledge allows cheating undetected—cannot be policed by reputation or incentives alone, and thus depend entirely on moral conviction that transcends rational calculation.
The conversation ranges across the structure of trustworthiness itself. Rose distinguishes principled moral restraint from harm-based guilt, showing why the latter fails in large groups where harm diffuses and empathy dissolves. He argues that incentive compatibility produces only calculative assurance, not genuine trust, and that the greater good rationalization problem—where people justify cheating as serving a larger good—destroys trust because trading partners cannot know when they will be betrayed. The discussion draws on Adam Smith and Robert Frank on moral taste, on Hayek's Hayekian Knowledge Problem reconfigured for firm hierarchies, and on empirical shifts in American student ethics over 25 years. Rose closes by addressing whether moral decline in advanced societies matters immediately or merely draws down cultural capital, and whether institutions imported into less developed countries fail because moral beliefs supporting trust are absent rather than because the institutions themselves are flawed.
Rose argues that durable prosperity depends on a specific moral architecture—principled, duty-based moral restraint that ranks 'do no wrong' above any greater-good rationalization—because only such unconditional trustworthiness lets strangers trust each other enough to specialize at large scale.
- Large-scale specialization creates local-knowledge 'golden opportunities' to cheat that incentives and reputation cannot police
- Harm-based restraint fails in large groups because diffused harm dissolves the empathy that triggers guilt
- Trust collapses unless people refuse opportunism even when no one is harmed and even for a 'greater good'
Genuine trust requires principled moral restraint structured with lexical primacy over moral exhortations.
- When someone tells you to do something you know is wrong because a greater good can be achieved, you should recognize this as the language of a charlatan and refuse; the practical moral rules are simple—don't commit negative moral acts even if no one is harmed, because if you steal you are a thief regardless of harm, and reject greater-good rationalizations as manipulation.
“you need to realize that that is the language of a charlatan, that that is inappropriate, that you are being sucked in. We don't do things like that.”
- Even a person with principled moral restraint may feel compelled to commit a negative moral act if doing so enables a larger positive moral good (the greater good rationalization problem); this destroys trust because you cannot trust someone who might cheat you to help a third party they value more, so genuine trust requires the partner to not even think in those utilitarian terms.
“I have to believe that you are the kind of person who would say: Just because you could cheat me a little and help your nephew a lot, you don't do that sort of thing.”
- Duty-based moral restraint requires not just a list of moral values but a logical structure giving lexical primacy to obeying moral prohibitions over obeying moral exhortations—obedience to positive exhortations only counts morally once prohibitions are satisfied—so that a person will never trade off a greater good against committing opportunism; this is what makes them genuinely trustworthy.
“lexical primacy of the obedience of moral prohibitions over the obedience of moral exhortations solves the greater good problem.”
- Following Robert Frank, the only way to be trustworthy in a genuine golden opportunity is for trustworthiness to be based on moral taste—something pre-rational and antecedent to rational calculation—because any rational, self-interested basis for trustworthiness will fail when cheating cannot be detected.
“the only way to bust out of that is for trustworthiness to be based on moral taste. If it's in any way an exercise in rational behavior, it's not going to work for a golden opportunity.”
- Principled moral restraint—refusing a negative moral act not because of the harm it does but because it is wrong in and of itself—solves the empathy problem; the candy-bar thought experiment shows people will not steal even when guaranteed to get away with it, demonstrating that genuine restraint is not contingent on detection or harm.
“Principled moral restraint is the idea that I'm not going to do this particular negative moral action not because of the harm that it does but because I believe it's wrong in and of itself.”
- Solving the greater good rationalization problem requires both principled moral restraint and lexical primacy together to produce duty-based moral restraint, because without principled restraint the enabling negative act will not even be regarded as wrong when no one is harmed, and without lexical primacy a moral person may still trade it for a greater good.
“You have to have both principled moral restraint and the greater good rationalization to solve the greater good rationalization problem at the same time to produce a condition of duty-based moral restraint”
Harm-based and incentive-compatible trust fail in large groups and golden-opportunity situations.
- If the only reason a person refrains from opportunism is the harm it does to others (harm-based moral restraint), then when harm is diffused across a large group so that no identifiable person is hurt, guilt is not triggered and the person will still behave opportunistically—a problem that worsens as group size grows because guilt depends on empathy with a harmed person.
“if that's the only reason why you won't behave opportunistically is because of the harm that's done, then the problem is, if you are in a situation where you think nobody is going to be harmed by your opportunism, you'll still be opportunistic.”
- Trust modeled purely as incentive compatibility (enlightened self-interest, where being trustworthy pays off through reputation) is not genuine trust but mere assurance or 'calculative trust,' and it breaks down precisely in golden-opportunity situations where cheating cannot be detected.
“the only way to bust out of that is for trustworthiness to be based on moral taste. If it's in any way an exercise in rational behavior, it's not going to work for a golden opportunity.”
- Essentially all economic trust experiments are framed in a small-group context (typically fewer than 25 people), which actuates our hardwired small-group mental models and therefore tells us nothing about large-group trust; this is fixable but no one has tried, limiting what such experiments reveal about the moral foundation of large-scale markets.
“Very rare that an economic experiment in a trust game involves more than 25 people. So we are already in a small group situation. That frames the issue in a small group way and therefore actuates all of our small group mental models.”
- Apparent over-generosity in experimental games may reflect not only moral guideposts but also a hardwired and prudent aversion to false-positive golden opportunities: when a seemingly easy opportunity is offered, mature, careful people suspect a hidden further level to the game, so being incredulous about apparent golden opportunities is the mark of maturity.
“a person has an aversion to false positive golden opportunities. It seems to me that we are hardwired, that we should be hardwired, to be suspicious of what appear to be golden opportunities.”
Measurable trust decline in the United States reflects shift from principled to mechanical moral restraint.
- Robert Putnam and Eric Uslaner have documented a broad reduction in measured levels of trust and trustworthiness in the United States from 1950 to the present, with a clear downward slope.
“Robert Putnam has documented a pretty-much across-the-board reduction in measured levels of trust... From 1950 until the present, it's pretty grim.”
- Persistent moral sanctioning of cheaters (e.g., in baseball's steroid scandal) does not refute moral decline, because it conflates two different brain functions: conviction born of deep belief versus a mechanical habit of mind; mechanical rule-following is observationally equivalent to self-aware unyielding conviction but is fragile.
“You are conflating conviction born of a deep belief with a habit of mind.”
- Studies show student cheating has risen dramatically over 25 years, and crucially the justification has changed: where students once admitted cheating was wrong but felt compelled, a shockingly high proportion now ask 'what's the big deal?' and argue nobody was hurt—indicating a shift toward the absence of principled moral restraint rather than mere rule-breaking.
“in the past when you asked students why they cheated... they almost never excused the cheating; they never downplayed the moral import of it. They would say it was wrong but they had to do it.”
- The risky lending behavior during the financial crisis reflected the breakdown of principled moral restraint: people knew the loans were wrong but rationalized them because the harm was diffuse ('nobody's dying') and they told themselves they were doing good by putting a person in a house, with the con working only as long as house prices kept rising.
“Look at the kinds of loan behavior that was going on in the financial crisis... Many people knew what was going on was wrong. And they just shouldn't be doing it. But they thought, well, nobody's dying because of this.”
Moral beliefs precede and enable institutions; cultural capital persists before decline accelerates.
- Living by the moral foundation requires not twenty hours of schooling but many years of continuous reinforcement to build the character and moral conviction behind a belief, even though the beliefs themselves are simple—echoing the pragmatist insight that grandmother's rules of thumb, held without articulable justification, are part of why a culture succeeds.
“It requires many years of continuous reinforcement in order to build the character to produce the moral conviction behind a belief, but the beliefs themselves are pretty simple.”
- Following Adam Smith's 'there's a lot of ruin in a nation' and Charles Murray's argument about Scandinavia, a society moving in the wrong direction on moral beliefs can continue to prosper for a long time because it draws down a large accumulated stock of cultural capital before approaching the margin where decline bites.
“they've built up a huge pile of cultural capital. They are going to have to make a lot of withdrawals from their cultural account before you get close to the margin.”
- Kant did not bring about the moral changes embodied in the categorical imperative but rather gave rigorous voice to changes in moral beliefs already underway in the culture, codifying and making rigorous what was already emerging.
“I think what Kant was doing is he was giving a rigorous voice to changes in moral beliefs that were already underway.”
- Guilt is the mechanism through which moral restraint operates, and it is activated by moral values; what matters is not merely having moral values but how they are structured, because poorly structured values fail to trigger guilt in the right circumstances.
“Guilt is the mechanism through which all of this works; and the question is how do you put guilt to work? You put guilt to work by having moral values that actuate it.”
- Even societies where people return lost wallets cannot sustain an environment of trust unless moral beliefs have the right structure (principled, duty-based restraint), because people will not keep extending trust if they are continuously punished for doing so—if extending trust is not rational, people stop, like refusing to be a sucker.
“People will not extend trust if they are continuously punished for doing so. If it's not rational to extend trust, you don't.”
- Empirical work measuring trust across the world shows that measured levels of trust co-vary well with economic performance and general quality of life, which—while not conclusive—is consistent with the claim that achieving trust pays off economically.
“measured levels of trust do co-vary well with economic performance and general quality of life in societies.”
Modern moral education teaches utilitarian reasoning unsuited to large-group society.
- Modern character and moral education that teaches children everything is a shade of gray and moral dilemmas are everywhere is effectively rank utilitarianism—a small-group morality applied to the modern world—which is persuasive because humans are hardwired for it, but it brings a knife to a gunfight by misapplying small-group reasoning to large-group society.
“It is just rank utilitarianism. It's a perfectly good approach to morality if you live in a very small group.”
- In 1870 the vast majority of Americans would have regarded taking money from some people to give to others as a self-evidently inappropriate use of government power, but over time this became the new normal and the conversation now begins far downstream of whether such redistribution is legitimate at all—an example of greater-good rationalization shifting moral baselines.
“the vast majority of Americans in 1870 would have said: You can't do that. That would have been self-evident.”
- When Jesse Jackson was caught diverting Rainbow Coalition funds, many people defended him by pointing to all the good he had done; Rose argues such a defense would have been universally laughed at in 1950, illustrating a shift in moral beliefs that now tolerates greater-good rationalization of overt fraud.
“many people said: Well, but you've got to look at the whole picture and all the good he's done, give him a break.”
Trust enabling large-scale specialization requires moral beliefs, not just institutional structures.
- Importing Western institutions into less developed countries often does nothing or gets subverted into vehicles of opportunism, because something beyond institutional apparatus—moral beliefs supporting trust—is missing; South American nations photocopy the U.S. Constitution and Argentina once had higher per capita income than the U.S., yet do not replicate U.S. outcomes.
“We have found that when you drop institutions into less developed countries very often they either do nothing or they are subverted and co-opted and become vehicles of opportunism themselves.”
- Being able to trust people you don't know is extremely rare in human history; 500 years ago genuine and generalized trust at the same time was essentially nonexistent, because nobody had the moral beliefs required to produce it.
“being able to trust people you don't know is something that has been extremely rare throughout human history. It's even rare today but if you go back 500 years or so, I would say it was completely rare.”
- Prosperity comes fundamentally from specialization: societies that can effectuate dramatic specialization through large-scale production reach productivity levels many orders of magnitude greater than others, and the question of distribution matters far less than having enough output to divide in the first place.
“Societies that are able to effectuate dramatic specialization through very large scale production are those that are going to have levels of productivity that are many orders of magnitude greater than other societies.”
Unobserved institutional losses from opportunism and localized knowledge waste compound over time.
- The true cost of unsolved opportunism is largely unobserved because it takes the form of economic organizations and institutions that never come into existence; the field's preoccupation with incentive-compatibility mechanisms reflects a survival bias, studying only the institutions that do exist to police common cheating.
“Most of the cost takes the form of economic organizations that don't exist, or institutions that don't exist.”
- Hayek's local knowledge problem, solved across society by prices, recurs inside firms: the larger and more complex a firm's production, the more likely individuals possess knowledge no one else has or can have, and failing to exploit it wastes efficiency just as ignoring market prices would.
“The larger a firm is and the more complex its production is, the more likely it is that there are people who know things that nobody else knows.”
- The localized knowledge problem in firms has both a stock dimension (knowledge embodied in employees who come and go) and a more daunting flow dimension emphasized by Hayek, where constantly changing daily problems are best answered by the person with localized knowledge of how that area works.
“The larger a firm is and the more complex its production is, the more likely it is that there are people who know things that nobody else knows.”
- Third-degree opportunism occurs when a person with local knowledge knows of a profitable-but-not-most-profitable action that superiors cannot know about, and chooses the merely good-enough action; because no one else can know the optimal counterfactual, the cheater can never be caught, making this an inescapable problem that worsens as firms grow.
“if an action is one that is profitable but not the most profitable is known to the person with local knowledge but not known to the others, and the person who possesses the local knowledge knows this, he might pick an action that is good enough but is not the best.”
Economic paradigms rise and fall; institutionalism now addresses what neoclassical economics missed.
- Following Thomas Kuhn, a new economic paradigm answers many questions early and is celebrated, but its usefulness eventually peters out; this happened to the neoclassical paradigm, prompting the revival of institutionalism, which in turn has also begun to lose steam.
“as Thomas Kuhn has pointed out, when you have a new paradigm, you always say that things are great; you start to answer a lot of questions; but over time you start to peter out.”