In the 1970s and 1980s, crime rates in the United States kept rising without pause. Media outlets flooded the public with grim warnings that America might collapse under crime, or that every major U.S. city would turn into a crime city. But when the 1990s arrived, crime rates—contrary to expectations—began to fall sharply. The media offered a variety of explanations for this decline. Policies such as ramped-up policing in the “war on crime,” aggressive use of the death penalty, crime-prevention campaigns, and gun control were highlighted as the most important factors.
Humans have a cognitive bias: when we see two events happen in sequence, we instinctively interpret them as cause and effect. If we observe police numbers rising while crime rates fall, most people assume one caused the other. But according to economists’ statistical analyses, none of the crime-fighting policies actually played a decisive role in the drop in crime rates.
Now let’s try answering a few quick questions. If more ice cream sales are linked with more drowning deaths, does that mean eating ice cream makes swimming dangerous? If a big-box store’s sales jump whenever its decorations are red, does that mean repainting everything red will boost profits? If students who take more private tutoring tend to have lower grades, should we conclude that tutoring makes students worse? If companies led by women tend to show lower profits and stock prices, should we assume women are worse managers than men?
These examples teach us why distinguishing correlation from causation is so important. Ice cream sales and drowning deaths rise and fall together—that’s correlation. But it would be wrong to conclude that more ice cream causes more drownings. Hot summer weather increases both ice cream consumption and swimming activity. The common factor—summer weather—explains the correlation, but there’s no direct causal link between ice cream and drowning. The same goes for red decorations and sales: Christmas season brings red decorations and higher shopping, but it’s the season, not the color, that matters.
The relationship between tutoring and grades is a case of reverse causation: students with lower grades tend to get more tutoring. What about women CEOs and firm performance? Researchers looked more closely at when women are appointed as executives and found that women are often chosen when companies are already struggling. This phenomenon is called the “glass cliff.” While the “glass ceiling” describes barriers to women’s promotion, the “glass cliff” refers to women being appointed only in situations that are already destined for failure.
Always remember: what we see is usually correlation, not causation. If we draw conclusions based only on what’s visible, we often get it wrong. Instead of taking things at face value, we need counterfactual thinking—that is, asking questions like: “What if that hadn’t happened?” or “Could another hidden factor explain this?”
For instance, suppose the government cut taxes last year, and this year tax revenue actually went up. A lawmaker might argue that tax cuts increase revenue. Would you nod in agreement? At this point, all we know is a correlation. To ask whether there’s causation, we’d need to ask: “What would have happened to revenue if taxes hadn’t been cut?” or “Could another factor explain the increase?” Perhaps the economy was booming, which naturally raised tax revenue regardless of the tax cut.
Here’s a simple example. Imagine that East Korea collected 1 trillion won in taxes last year. Early this year, it cut taxes, and revenue rose to 1.1 trillion. Politicians in East Korea start claiming that tax cuts boost revenue. Now imagine a parallel country, West Korea, identical in every respect—except it didn’t cut taxes. Its revenue rose to 1.2 trillion.
What can we conclude by comparing them? The economic boom raised revenues by 200 billion. But the tax cut reduced revenues by 100 billion. That’s why East Korea ended up with 1.1 trillion. What East Koreans see is “revenue rose by 100 billion after the tax cut.” But that’s just correlation. To identify causation, we must ask the counterfactual: “What if there had been no tax cut?”—and compare it with West Korea.
Understanding that correlation doesn’t imply causation changes how we interpret even the simplest graph. Suppose someone shows you an upward-sloping line comparing last year’s and this year’s revenues, claiming it’s clear evidence that tax cuts increased revenue. At that moment, you should raise your hand and ask: “Compared to what?”
So far, we’ve been talking about one country’s tax policy. But what if we compare across multiple countries? Suppose we find that, among many countries, those that cut taxes saw their revenues rise, while those that didn’t saw little change. Does this prove tax cuts increase revenue?
Not necessarily. Even across countries, the same problem remains. To really identify causation, we’d need something like a randomized controlled trial. Imagine randomly assigning some countries to cut taxes and others to raise them. Because the assignment is random, other factors influencing revenue are balanced out. Only then could differences in revenue between the groups be interpreted as causation.
The problem with simply comparing countries is that they weren’t chosen at random. Suppose fast-growing economies tended to cut taxes while slow-growing economies didn’t. Then the observed differences might be due to growth, not tax policy.
With this in mind, we can better understand the U.S. crime rate story. Police staffing increased because crime was high. So we can’t just say “more police means more crime.” And just because crime later fell after police numbers rose doesn’t prove the two were cause and effect—there could be hidden factors.
According to research in economics, one of the most important hidden factors was the U.S. Supreme Court’s 1973 Roe v. Wade decision, which legalized abortion. Unwanted births are statistically more likely to lead to future criminals; legal abortion broke that link. For example, children born to single mothers in poor neighborhoods face a higher risk of turning to crime later. Legalized abortion meant fewer such children were born, which in turn reduced crime. Still, whether abortion legalization truly caused the crime drop remains debated. Understanding that debate requires some advanced statistics and economics.
Another factor to remember is that, in social science, causes are rarely single and isolated. They’re usually multiple and interacting. Another economist studying crime found that stricter regulation of lead—a toxic heavy metal—also played a major role. Childhood lead exposure not only damaged health but also harmed intelligence and behavior, contributing to higher crime rates.
So far, research suggests that both legalized abortion and reduced lead exposure were among the most significant contributors to the 1990s crime drop. Causes and effects are not always one-to-one; they often emerge from multiple interacting factors.
One frequently debated topic in discussions about markets and government is the comparison between American capitalism and Nordic capitalism. American capitalism minimizes taxes and government intervention, stressing the role of free markets. Nordic capitalism, by contrast, levies high taxes to fund expansive social welfare systems. Which is better?
Whenever I hear this debate, two things come to mind. First, as explained earlier, marginal thinking requires us to move beyond black-and-white reasoning. Second, correlation does not imply causation. Too often, people rely on a few examples to make sweeping claims. For instance, supporters of Nordic capitalism argue that generous welfare drives growth. Supporters of American capitalism claim that low taxes drive growth. When a recession hits, newspapers rush to declare causes: if Nordic growth slows, welfare programs are blamed; if U.S. growth slows, unregulated capitalism is blamed. These claims are just correlations, not clear causal evidence.
Human cognitive bias craves simple answers and quick conclusions from just a case or two. But economic reasoning requires us to separate correlation from causation, and to recognize that multiple causes may interact. Complex problems don’t have simple answers.
The influence of economics is immense. Beyond its traditional domains—finance, taxation, trade, industrial growth, and regulation—economic logic and methodologies now permeate nearly every aspect of life, including education, healthcare, family, and personal well-being. In recent years, economics has also encouraged people to apply "thinking like an economist" to everyday decision-making. However, even this mindset is subject to the core economic principle that “every decision involves trade-offs.” Even thinking like an economist comes with an opportunity cost.
Ariel Rubinstein, a professor of economics at Tel Aviv University, conducted an experiment with 764 college students, assigning them the role of a corporate vice president tasked with deciding how many employees to lay off for business reasons. The company currently employs 196 workers. If 96 employees are laid off, the company earns $2 million in profit; if 52 are laid off, the profit is $1.6 million; if 20 are laid off, the profit is $1 million; and if no one is laid off, the profit is $400,000. The maximum profit is achieved by laying off 96 employees, while fewer layoffs result in lower profits.
Participants faced a difficult choice between maximizing profits and considering the impact of layoffs on employees' well-being. Their decisions ultimately reflected their core values. The results showed clear differences by academic background: 13% of philosophy majors, 16% of math majors, 27% of law majors, 33% of MBA students, and 46% of economics majors chose to lay off 96 employees in favor of profit maximization. Economics students were the most inclined to prioritize profit maximization, even more so than MBA students, who are trained in corporate decision-making.
In another experiment, participants were presented with a mathematical formula linking the number of layoffs to company profits. When participants could simply plug the layoff numbers into an equation, the results were the same as in the previous experiment. In other words, the only difference was that the information was presented in the form of a mathematical formula instead of raw numbers. This experiment tested whether the use of economic modeling, which heavily relies on mathematical formulas, affects decision-making.
When profits were presented as raw numbers, 31% of all students, regardless of major, chose to lay off 96 employees. However, when the same information was presented as a mathematical formula, a staggering 75% chose to lay off 96 employees. The introduction of a mathematical model made participants significantly less sensitive to the plight of the laid-off workers, shifting their decision-making drastically toward profit maximization. More strikingly, the differences between majors nearly disappeared. Even philosophy students, who had previously considered the workers’ situations more carefully, made decisions indistinguishable from economics students when reasoning through mathematical formulas.
The opportunity cost of "thinking like an economist" is the loss of empathy. Adam Smith, often regarded as the father of economics, described empathy as a fundamental human moral sentiment. In this light, it is deeply ironic that studying economics may lead to a decline in empathy.
Lawrence Summers, a renowned economist who served as U.S. Treasury Secretary, President of Harvard University, and Chief Economist of the World Bank, once said, "The economic logic of dumping toxic waste in developing countries is impeccable." Though his remark sparked outrage, most economists understand that his argument is grounded in efficiency. By the same logic, cutting welfare for low-income populations or overlooking poor working conditions and workplace safety might also be seen as "efficient" choices. Economists, when writing papers and policy reports, risk becoming mere replicas of Summers. Decision-making and social policies rooted in "thinking like an economist" may carry a bias toward "empathy loss," a reality we must always be mindful of.
A few years ago, I contributed a series of articles to a religious magazine interpreting the Bible from an economist’s perspective. One of my writings, which analyzed Jesus’ parable of the vineyard owner, provoked discomfort among some readers, primarily evangelical Christians. They accused me of interpreting the Bible from a socialist perspective, and in response to their complaints and considerable pressure, the publisher ultimately decided to discontinue the series.
The parable goes as follows: The vineyard owner hires workers early in the morning, at 9 a.m., at noon, at 3 p.m., and even at 5 p.m. Regardless of their hours worked, he pays all of them the same wage. The vineyard owner is often understood as representing God, but many biblical scholars see the story as a realistic depiction of the social conditions of the time. I focused on the power dynamics between the employer and the workers, questioning the imbalance in bargaining power and the unfair wage-setting process. Fundamentalist theologians dismissed my interpretation as leftist or Marxist. They were more concerned about applying a socialist perspective to the Bible than about differences in biblical interpretation itself.
Interestingly, this episode reveals how the public often misunderstands the market economy. Many believe that a market economy is a system where corporations and employers have the freedom to do whatever they want, including setting discriminatory wages. However, only an employer with monopoly power can set wages as they please and engage in discriminatory practices. The way the vineyard owner determines wages would not happen in a competitive market economy. A market economy is defined by competition—firms must compete in product markets by offering better goods at lower prices, and they must compete in labor markets by providing better working conditions and higher wages. If anything, those fundamental Christian should have criticized why I was interpreting the Bible in favor of capitalism.
While a market economy forces businesses to compete, those who called me a Marxist mistakenly equated a system where businesses have free rein with a market economy. This kind of misunderstanding is both deep-rooted and widespread. Pseudo-market advocates often treat "pro-market" and "pro-business" as synonyms, though they should often be used as opposites. Pro-business policies protect the interests of dominant firms, whereas pro-market policies create the very competition that dominant firms despise.
The Yoon administration in South Korea also made the mistake of confusing pro-market with pro-business. It proposed a labor reform policy allowing a maximum workweek of 69 hours, branding it as a pro-market measure. To determine whether a policy is truly pro-market or merely pro-business, one should ask: "Would a firm with monopsony power in the labor market push for a 69-hour workweek, or would such conditions emerge only in a competitive hiring environment?"
The conflation of markets and businesses is even more pronounced in labor market issues than in product markets. The lack of attention to competition policy in labor markets illustrates this well. The consumer welfare standard, which has guided U.S. antitrust policy, largely ignored labor market concentration. For instance, corporate mergers have traditionally been scrutinized for their impact on consumer prices, with little concern for their effects on wages and working conditions.
Recent studies show that monopsony power in labor markets has contributed to the declining share of labor income. Drawing on such research, U.S. antitrust authorities are now pushing to ban noncompete agreements. Currently, one in five American workers is required to sign a contract that prevents them from taking a job with a competing firm. These workers include not only corporate executives and tech professionals but also doctors, nurses, fast-food employees, and janitors. Some argue, "In a free market system, what's wrong with employers setting contract terms in exchange for higher pay?" FTC Chair Lina Khan responds, "Noncompete clauses suppress workers' wages." In other words, noncompete agreements are pro-business but anti-market.
At the beginning of my introductory economics course, I always ask my students a question: “Imagine that a large number of immigrants move to a city in search of jobs. How will wages in the relevant labor market change?” Almost everyone answers that wages will decrease. Even without formal training in economics, they already understand the basic market principle that an increase in supply leads to a decrease in price.
In 1980, approximately 125,000 Cuban immigrants arrived in Miami, Florida. This influx increased Miami’s labor force by about 7%. Professor David Card conducted an empirical analysis to examine whether this wave of Cuban immigration actually lowered wages. The results showed little to no change in wage levels. Furthermore, the employment levels of local workers remained unaffected. Does this mean that the supply and demand theory is incorrect?
The conclusion that “an increase in immigration leads to lower wages” is a misuse of supply and demand theory. Like all economic theories, supply and demand is a framework for thinking. To isolate the causal relationship between changes in supply or demand and price, the model must be simplified. This means that supply and demand theory operates under many implicit assumptions—assumptions that even those well-versed in the theory often overlook.
The first type of misuse stems from the assumption of “all else being equal.” Wages in Miami remained stable despite the increase in immigrant labor because labor demand also increased. Immigrants are not just workers; they are also consumers. Their everyday consumption creates additional jobs. When both the supply and demand curves for labor shift outward, the effect on wages is unclear.
A similar misuse frequently appears in real estate policy debates. People often assume that policies restricting housing demand will lower prices. However, such policies can also reduce supply—for instance, homeowners may postpone selling their properties due to dissatisfaction with current lower prices and anticipation of future increases. In this case, if supply decreases more than demand, prices may rise instead. Conversely, many believe that increasing the supply of housing will automatically lower housing prices. However, the substitution effect between housing markets in different locations is typically weak, meaning that an increase in supply may have only a limited impact on prices in high-demand areas.
The second type of misuse relates to the assumption of “a sufficiently competitive market.” Suppose a Walmart opens in a small town, increasing the demand for labor. One might expect wages to rise as a result. However, empirical studies show that wages tend to decline instead. This is because Walmart holds significant monopsony power in the labor market. When a company has substantial market power over labor, the standard supply and demand framework does not apply. This is also one reason why numerous studies on minimum wage policies fail to consistently show increased unemployment—many labor markets are not as competitive as the theory assumes.
The real estate market is also far from free from monopoly issues. The fact that homeowners can collude to inflate housing prices suggests that the market is not fully competitive. Legal scholar Eric Posner and economist Glen Weyl argue that when monopoly assessment criteria are applied, real estate qualifies as a monopolistic market. Homeowners can set prices significantly higher than the minimum they would be willing to accept. The monopolistic nature of the housing market may explain why supply and demand logic does not always hold in real estate.
The misuse of supply and demand theory originates in the way economics is taught. Economics courses focus on ensuring that students understand supply and demand theory. In fact, the question about how immigration affects wages is a standard problem found in major textbooks, with the expected answer being “wages decrease.” However, there are no questions in these textbooks that ask students to identify the ways in which supply and demand theory can be misapplied.
This semester, I am teaching game theory. Game theory is a theoretical framework for analyzing situations in which individuals’ choices interact with one another. While once considered a narrow field, it has now become a widely used analytical tool in economics. For Ph.D. students in economics, it is a required course.
In the first lecture, I introduce a well-known game: the Prisoner’s Dilemma. This game illustrates situations where cooperation leads to better outcomes for all, yet each individual has an incentive to betray the other for personal gain. As a result, both parties end up worse off. The Prisoner’s Dilemma effectively explains the challenges of cooperation.
Many students encountering game theory for the first time raise objections. Some claim they would not betray their counterpart, while others cite real-world examples of cooperation overcoming selfish motives. They argue that such counterexamples disprove game theory. Indeed, experimental research shows that many people do choose to cooperate, and in some cases, they even make self-sacrificial decisions.
While these concerns seem valid, they stem from a fundamental misunderstanding of game theory’s role. Game theory is not a universal law that produces fixed, immutable conclusions. Rather, it serves as a framework for posing insightful questions and exploring different possibilities. Its primary function is to clarify why and under what conditions different outcomes emerge. This principle extends beyond game theory to economic theory as a whole.
Declaring that an economic theory is absolutely correct or absolutely wrong reflects a misunderstanding of its purpose. The Prisoner’s Dilemma applies to certain situations but not to others. Assessing its relevance requires examining how well its underlying assumptions fit the given context, rather than merely evaluating the logical rigor of the model itself.
Even experts and scholars frequently make similar mistakes. A recent and striking example is Harvard Law Professor Mark Ramseyer. In his paper "Contracting for Sex in the Pacific War," published in the International Review of Law and Economics, he used game theory to frame the Japanese military’s comfort women system as a rational contractual arrangement between prostitutes and brothel owners. He applied game theory indiscriminately to a situation where voluntary transactions did not exist.
In reality, economic theories are often produced and consumed in ways similar to Ramseyer’s approach. There is a widespread tendency in academia and the media to assume that if a theory is logically sound, it must also explain reality. For instance, the Chicago School of economics has frequently applied price theory indiscriminately to various forms of allegedly anticompetitive corporate behavior, consistently reaching the same conclusions. Similarly, some economists use a simplistic application of supply and demand theory to argue that minimum wage laws must necessarily lead to unemployment. Others cite general equilibrium theory as proof that government intervention is inefficient and that free markets are inherently superior.
Interestingly, the role and limitations of economic theory are most evident in legal disputes where economists serve as expert witnesses. When regulatory agencies and corporations clash over allegations of unfair competition, both sides hire distinguished economists to provide expert opinions. These experts often rely on different theories to support opposing arguments. However, even when they use the same economic model, they frequently apply it in ways that lead to sharply divergent conclusions. While they may agree on the theoretical foundations, they fiercely contest its practical application.
The controversy surrounding Mark Ramseyer’s work exposes the unfiltered reality of academia. Economics is often presented as a rigorous science, but in practice, both the development and application of economic theories are influenced by subjective judgments and biases.
Economic policy is often likened to the issue of growing and dividing a pie. Consider an economy consisting of two groups, A and B. Some policies increase the size of the pie, allowing everyone to receive a larger share than before, while others shrink the pie, leaving everyone with a smaller portion. There is little room for debate over these policies. Whether we ask A and B themselves or a third party, there is generally broad agreement on what constitutes a good or bad policy.
However, there are also policies that shrink the pie but allow A to take a larger share than before, forcing B to receive a significantly smaller portion. This situation often arises when A takes from B, which can be described as an "exploitative economy." Such cases are rare in an inclusive market economy where property rights, contractual freedom, and a fair legal system function effectively.
More commonly in a market economy, the pie grows, but not everyone receives a larger share. Most of the benefits go to A, while B suffers losses. It is understandable that A and B assess this situation differently. But how should a third party evaluate it? The question of how to weigh A’s gains against B’s losses helps illuminate the complexities of economic policy debates. It also provides important insights into the role and limitations of economists.
Traditionally, economists have offered a straightforward answer: they focus on the overall size of the pie while keeping their distance from the distribution of shares between A and B. By aggregating producer and consumer surplus into a single measure of "social welfare," they determine the success or failure of a policy. Economists would not object to being labeled as totalists, maximizers, or adherents of "the greatest happiness of the greatest number."
For a long time, economists supported free trade policies because they increased the size of the pie. Even if workers in certain industries lost their jobs, as long as the gains to other industries outweighed these losses, maximizers saw no problem. Only recently, as job losses in declining industries and their associated social issues have become more pronounced, have economists started paying attention to the broader implications of free trade.
The debate over antitrust policies for platform companies follows a similar pattern. Due to the nature of two-sided markets, one group of users may suffer while another benefits. For example, if Amazon copies the products of small businesses and sells them at lower prices, those businesses may be driven out of the market, but consumers enjoy cheaper goods. Economists who adhere to total surplus analysis and those who oppose it reach different conclusions about the anti-competitive behavior of platform firms.
Understanding how a policy changes the size of the pie and how it alters the shares of A and B is the core task of economists, requiring technical expertise. However, comparing A’s gains and B’s losses—and deciding how to evaluate them—is a different matter. This is where the role of economists becomes questionable, as such judgments are intertwined with value-based considerations.
I often pose the following question to students: "Would you support a policy in which half of you lose $1 each while the other half gain $2 each?" The majority say they would support it. Then, I adjust the question: "Now, half of you lose $10,000 each, while the other half gain $20,000 each." Support drops significantly. "What if one person gains $100,000 while 99 others each lose $100?"
When I present similar questions in different policy contexts—such as the adoption of new technologies, child labor in developing countries, or low-interest-rate policies—their responses vary widely, even for the same question.