Compensating differentials is a simple yet powerful economic model explaining how wages are determined. The idea dates back to Adam Smith, who noted that wages are not just based on productivity or hours worked.

The wages of labor vary with the ease or hardship, the cleanliness or dirtiness, the honorableness or dishonorableness of the employment.

Factors like whether a job is honorable, fun, or clean also play a significant role. For example, butchers earn about 10% more than bakers, despite being less likely to be college-educated. The reason is that being a butcher involves blood and guts, making it a less pleasant job than baking. Many people enjoy baking as a hobby, as evidenced by shows like 'The Great British Baking Show,' but a 'Great British Butchery Show' is unlikely to have the same appeal.

This dynamic affects the labor market. If the wages for butchers and bakers were the same, everyone would want to be a baker. This surplus of bakers would drive their wages down, while the shortage of butchers would push their wages up. The market reaches equilibrium when the wage difference is large enough to compensate people for the unpleasantness of being a butcher. The extra money is enough to convince some people to choose that job despite preferring baking.

Another example is forest conservationists, who earn less than agricultural inspectors despite similar education levels. This is because people are drawn to the idea of working in a forest and helping the environment. The concept of compensating differentials helps explain wage differences that are not driven by skill or education levels, which is important for understanding economic policy.

Tyler shared a family anecdote about his sister who worked as a butcher. She found the two hardest parts of the job were carrying heavy meat trays and the constant cold. These conditions led to serious health problems, often forcing people to quit the profession by their 40s. This illustrates the non-monetary downsides of certain jobs.

Alex agrees with the general point but challenges the idea that job choices are simple trade-offs based on compensating differentials. He argues that segmentation is often the governing principle. For example, tenured professors are paid more and have better job security than adjuncts. It's not a trade-off; one job is simply better across multiple dimensions. This happens when a common factor influences all qualities of a job.

He offers another example to illustrate his point: choosing a brain surgeon. If you found a surgeon with a dirty office, you wouldn't assume their surgical skills must be excellent to compensate for the uncleanliness. Instead, people expect good things to go together, like a great surgeon and a clean operating room. It's not a choice between a bad surgeon in a clean room and a great one in a dirty room.

This leads to a common confusion. In the real world, pleasant jobs like being a professor tend to pay more than unpleasant jobs like being a garbage collector. This seems to contradict the economic principle that more pleasant jobs should have lower wages, all else being equal. The key is the 'all else being equal' condition. The skills required to be a professor are rarer and more in demand than the skills needed to be a garbage collector. This difference in skill, and its market value, is the primary driver of the wage difference, overriding the simple compensation for the job's pleasantness or unpleasantness.

There are two ways to think about the relationship between job quality and pay. The first is Adam Smith's idea of compensating differentials, which suggests unpleasant jobs pay more to compensate for their poor conditions. For example, a garbage collector might earn more than a coffee shop clerk, all else being equal. The second idea is that all good things often come together. This tends to happen in industries with increasing returns. A hospital with the best brain surgeons is also likely to have the cleanest rooms and the best customer service.

These two ideas are not necessarily inconsistent. One way to reconcile them is to think of a worker as both selling labor to a firm and buying amenities like safety or pleasant working conditions from that same firm. A highly skilled worker with a high wage also has a high income. Just as they might use that income to buy a safe car like a Volvo, they also use it to "buy" safety at work. This transaction isn't explicit. Instead, a firm might offer the ability to work from home, but at a slightly lower wage. Because high-income workers can afford to buy more of these amenities, good things like high pay and pleasant conditions tend to cluster together.

Both effects can operate at the same time. Tyler notes that restaurants with beautiful patrons often have above-average food, but not the absolute best food. Patrons are making a trade-off, accepting slightly lower food quality for the atmosphere, which is a compensating differential. At the same time, the restaurant must offer good food to attract a high-income clientele in the first place, showing how good things can still cluster.

These trade-offs are significant. Research from Kip Viscuzzi estimates that workplace amenities in the United States are worth about 18% of GDP. This means if we eliminated these amenities, wages would be 18% higher, but jobs would be less pleasant and less safe. This concept also applies to policy discussions like the gender wage gap. After controlling for education and skill, a portion of the remaining gap can be explained by differing preferences for job amenities. For instance, Claudia Goldin has pointed out that men are often more willing to take jobs with inflexible hours, a trade-off that typically commands higher pay.

Differences in annual earnings between men and women can sometimes be explained by 'compensating differentials' rather than market discrimination. For instance, in a study of train and bus drivers with equal hourly pay, male drivers earned more annually because they took on significantly more overtime shifts, which pay time and a half. Similarly, male Uber drivers earn about 7% more on average because they tend to drive slightly faster. These choices, not employer bias, lead to different outcomes. Alex notes that these choices might involve trade-offs, like increased risk, so a higher wage isn't necessarily a better outcome overall.

Tyler presents an alternative model where 'all good things come together,' particularly in high-skill sectors like tech, law, and academia. In these fields, talented people cluster, creating highly productive environments. These workplaces often feature not only high pay and nice amenities but also long, inflexible hours. Instead of a simple trade-off, this creates a division. People who value flexibility, perhaps women on average, may end up in a less advantageous network, effectively getting the 'short end of the stick.' They don't just get compensated with flexibility; they miss out on the compounding benefits of being in the top-tier environment.

Alex pushes back, suggesting that feminism can sometimes undervalue the non-pecuniary benefits of childcare and time at home. He questions the assumption that the high-wage, long-hours track is inherently better, framing the 'workaholic' model as a potential mistake men make.

The men are the workaholics. And workaholics is probably not good either. And the men might be better off if they spent more time at home with the kids and getting those non pecuniary benefits. So one of the things I think feminism has a problem it has had is actually to be too male oriented to think too much, that the wage is the only thing which matters. These compensating differentials, they also matter.

The conversation broadens to a general economic trend. As more industries become tech-like, the clustering of highly skilled people—an 'increasing returns' effect—may become more common. This is related to Michael Kremer's O-ring theory, where in high-stakes work, there's a strong incentive to work only with others who are unlikely to make mistakes. This could lead to a greater divide between those inside and outside these elite, high-performing clusters.

In some sectors, wages are very high because you need to work with people who do not make mistakes. This idea is used to explain why adjunct professors might not be trusted with hiring decisions, out of fear they might make a mistake. However, technology can reduce the cost of such mistakes. For example, at McDonald's, cashiers use keyboards with pictures of the food items, like a Big Mac button, which minimizes errors.

Fixed costs in economics can ruin or validate good theories. An example illustrates this. Typically, a restaurant with a great view might not have the best food. But in a small town, like one in southern Nova Scotia, the situation can be different. For a quality restaurant to survive at all, it might need to have both the best food and the best view. This combination is necessary to attract enough customers to cover the high fixed costs of offering quality seafood. In this case, all good things must come together for the business to be viable.

This highlights a limitation of a common economic principle. While economists often advise thinking on the margin, the presence of significant fixed costs makes this difficult. If a business needs to generate a certain amount of revenue just to survive, it has to focus on more than just marginal decisions.

Despite the high upfront expenditures on data, energy, and training for AI, it is not a given that we are moving into a world dominated by high fixed costs where standard neoclassical economics no longer apply. While the range of possibilities for the future of AI is very wide, from explosive growth to doomsday scenarios, the economic question of fixed versus marginal costs can be analyzed more directly.

The initial fixed costs of training AI models may seem large, but they might not be significant when compared to the size of the total addressable market. The true market for AI is the global labor market, which constitutes most of the world's GDP. Relative to this massive market, the training costs could be quite small. This suggests that marginal costs will likely continue to dominate pricing in the long run. For example, while Waymo's fixed costs are substantial now, in a decade, its business will likely be driven by marginal costs.

This can be framed as a "time path issue." The high initial expenditures are fixed costs that, over time, become sunk costs that are forgotten. In the very long run, the economy could become even more neoclassical, with everything ultimately priced at the marginal cost of inputs like energy.

The concept of compensating differentials explains how prices and value adjust across many different areas. For example, housing prices are often lower in high-crime areas. If crime decreases, housing prices tend to rise. Similarly, people are willing to pay more to live in places with better weather. This demonstrates a key economic lesson: the economy always operates and equilibrates on multiple margins, not just one.

This principle can be illustrated with the "Happy Meal fallacy." Imagine a politician mandates that every burger must be sold with fries and a drink to ensure everyone gets a complete meal. In reality, the price of the meal increases. This policy harms the very people it intended to help—those who previously chose not to buy the fries and drink because they didn't value them enough. Now, they are forced to purchase them. Meanwhile, those who were already buying the full meal are no better off. This kind of intervention can actually make people worse off by ignoring the different margins on which people make choices.

This idea also applies to debates around the minimum wage. While a higher minimum wage might be mandated, other factors like working conditions could worsen in response. This doesn't mean the minimum wage has no effect, but it highlights that there are always multiple margins for economic equilibration.

Tyler first learned about selective incentives from Mancur Olson's 1965 book, "The Logic of Collective Action." The theory explains how public goods get provided. For example, a local chieftain might give private favors to people who contribute to the public good, which helps internalize some of the gains from collective action. The hosts knew Olson, whom they describe as a friendly man who likely would have won a Nobel Prize had he not died early.

Alex explains that Olson's book, along with his later work "The Rise and Decline of Nations," was highly influential. The core idea is that lobbying the government is a public good for an industry. Any law passed benefits everyone in that industry, even those who didn't contribute to the lobbying effort. This leads to a key insight: the bigger an industry is, the harder it is for it to effectively lobby for its own interests due to the free-rider problem.

When agriculture was a big share of the economy, agriculture got no subsidies. But as agriculture has become a smaller share of the economy, it's become much more heavily regulated and subsidized. And that's because of two reasons. One is because you can't subsidize a big share of the economy... And the second reason is that the fewer, the more big players there are in agriculture... they're able to lobby the government.

This is illustrated by the history of agriculture. When it was a large economic sector with many small family farms, collective action was difficult, and there were no subsidies. As agriculture became a smaller sector dominated by a few large firms like Conagra and Archer Daniels Midland, it became easier to coordinate and lobby for subsidies.

However, Tyler points out some complications. The U.S. still has around a million farmers, which seems too large a number for effective collective action according to Olson's model. He questions whether subsidies are driven by a few large corporations or by the political power of voters in places like the Iowa primary. Furthermore, other large groups, like school teachers, also receive indirect subsidies, which seems to challenge the idea that only small, concentrated groups can successfully lobby for benefits.

Mancur Olson's model suggests that small, well-organized groups effectively lobby for benefits at the expense of larger groups, a dynamic he described as "the small exploit the large." Over time, the accumulation of these special interests leads to "demosclerosis," where the economy is held down by countless tiny restraints, stifling economic growth.

However, Alex has revised his view. While backroom deals, like a tax code provision favoring Starbucks, certainly exist, he now argues that the primary force holding back economic growth is the voters themselves. The most significant barriers are not hidden deals but policies driven by public opinion.

It's more due to the voters. They actually don't want growth. They want to keep their neighborhoods the way they are. They don't want the new building going up. They're frightened of AI. They're frightened of corporations.

Many costly regulations have broad public support. Environmental and safety laws, for instance, are often passed due to public demand, not just industry lobbying. The decline of the nuclear power industry serves as a prime example. Its stagnation was largely caused by public fear, fueled by events like the Three Mile Island accident and movies like "The China Syndrome," rather than by competing industries.

The media amplifies these fears, focusing on dramatic but rare events while ignoring more common, deadlier risks. This creates a skewed perception of danger among the public.

Many, many, many more people die from worse air pollution and from coal mining accidents than have ever died from nuclear power.

Ultimately, the most harmful regulations are not the small, hidden favors for special interests. They are the large, publicly supported ones that operate in plain sight.

The effectiveness of using selective incentives to fund public goods, a model reminiscent of a village chieftain, diminishes as markets become more competitive. Tyler shares an anecdote about his father, who ran a chamber of commerce and successfully funded lobbying for local businesses by soliciting donations. In today's market, however, businesses would likely just hire a lobbying firm directly, bypassing the collective effort. The more entry and competition there is, the less important this collective model becomes.

This trend is visible in the news industry. Foreign news bureaus, a public good, were historically supported by cross-subsidies from newspaper advertising revenue. Journalists pursued stories they felt were important, which also enhanced the publication's brand and prestige. As advertising revenue from sources like classifieds dried up, these bureaus could no longer be supported and largely disappeared.

That has all gone away because the advertising revenues from the classifieds have gone away and now everything has to be supported on its own barrel.

A similar shift is happening in academia. The university system is built on cross-subsidies, such as research being funded by teaching revenue. However, this model is eroding. Increasingly, the best researchers do little to no teaching, while those who teach the most do not research, creating a division between the two functions.

This market-driven breakdown also challenges the theories of Nobel laureate Elinor Ostrom. Ostrom observed that many local communities successfully managed common resources like fisheries or irrigation systems for centuries, avoiding the tragedy of the commons. Her model relied on tight-knit communities where a leader, or 'chieftain,' could apply selective social incentives to enforce cooperation. However, with widespread urbanization and mobility, these stable, local communities are becoming rare. The social structures that made Ostrom's solutions work are fading away.

The Olson hypothesis, which deals with selective incentives in groups, was explored by thinkers like James Q. Wilson, Guy Clark, and Herbert Simon long before it was formally named. They were studying what were, for their time, some of the largest organizations.

There seems to be a U-shaped destiny for this theory. It was initially very useful for understanding small-scale collective action, such as managing irrigation in rural Nepal. Then came an intermediate stage where market dynamics were the primary focus. Now, as companies have become massive, the theory is relevant again. To understand a modern large corporation, you might have to think about its internal dynamics in the same way you would about local chieftains managing a water irrigation project.

Charter cities represent one attempt to solve the public goods problem by having a single large corporation internalize externalities and use land rents to support various services. This approach, along with creating special enterprise zones or homeowners associations, aims to better organize property rights to solve collective problems. This is one of two potential market-based solutions. The other, known as the Cosian solution, is to create more defined property rights in the externality itself, such as pollution.

Technology can facilitate these solutions. Better monitoring allows for charging individuals based on their specific impact, leading to tools like congestion pricing. An extreme example is China's use of surveillance where jaywalkers are automatically identified and fined, with the money taken directly from their accounts. This raises a concern: creating efficient markets for externalities might require building a surveillance society.

However, there is a trade-off to consider between different types of surveillance. The impersonal nature of mass surveillance might be preferable to the personal oversight of a local leader, like the 'village chieftain' in Mancur Olson's model.

In the Olson model, the village chieftain, so to speak, is surveilling you. And it's someone personal. They can have a grudge against you. They know where you live. The mass surveillance model, while highly problematic for other reasons, often political, but there's a sense in which no one cares.

While this impersonality has its own risks, as someone might eventually decide to care, the local chieftain always had to care, or at least pretend to. This presents a dilemma that societies may never fully escape. Ultimately, this illustrates that seemingly simple microeconomic problems are often quite complex, and any market contract involves many different margins of adjustment.