Specialization and Trade

A Re-introduction to Economics

Arnold Kling

• Publisher: Libertarianism.org Press
• Year: 2016
• Print ISBN: 9781944424152
• eBook ISBN: 9781944424169
• Topic: Economics
• Subtopic: Macroeconomics
• Index: Economics

Since the end of the second World War, economics professors and classroom textbooks have been telling us that the economy is one big machine that can be effectively regulated by economic experts and tuned by government agencies like the Federal Reserve Board. It turns out they were wrong. Their equations do not hold up. Their policies have not produced the promised results. Their interpretations of economic events -- as reported by the media -- are often of-the-mark, and unconvincing.

A key alternative to the one big machine mindset is to recognize how the economy is instead an evolutionary system, with constantly-changing patterns of specialization and trade. This book introduces you to this powerful approach for understanding economic performance. By putting specialization at the center of economic analysis, Arnold Kling provides you with new ways to think about issues like sustainability, financial instability, job creation, and inflation. In short, he removes stiff, narrow perspectives and instead provides a full, multi-dimensional perspective on a continually evolving system.

1

Filling in Frameworks

Believe those who seek the truth; doubt those who find it.

—André Gide⁸

In 2000, Ursinus College, a small undergraduate institution in Pennsylvania, raised its tuition by more than 17 percent. Subsequently, the number of applicants and acceptances for its freshman class rose.⁹ That outcome appears to violate the law of demand, which says that demand goes down as the price goes up.¹⁰ Has the law of demand been found to be false?

Neither I nor any other economist would be willing to concede that the law of demand fails to hold. Instead, we would look for factors that might account for the Ursinus College application anomaly. For example, we might ask whether competing colleges raised their prices as much as or more than Ursinus. Perhaps Ursinus College had a successful basketball season, or another factor raised its profile among high school students. Perhaps a new government program increased the subsidies for college students.

In fact, the article on Ursinus College mentions that it also raised its level of student aid by close to 20 percent, and that the majority of students paid less than half of the full price. One can argue that, rather than defying the law of demand, Ursinus was using it. The college was taking advantage of what economists call price discrimination, charging a high price to those students willing to pay while luring the more price-sensitive students with generous aid.

The point is that the law of demand holds only when “all other things are equal.” However, in the real world, other things are almost never equal. In the case of Ursinus College, financial aid policies were not being held equal.

In physics or engineering, when you leave out a factor (such as friction), you do so because you can show that in the context of your analysis that factor will not be important. In economics, we typically cannot do that, because we do not control the environment in which we undertake a study. Many important causal factors will be operating at once, and although we might hope or pretend that none of them matter, we often have no basis for ruling out their importance.

When economists seek to explain phenomena, we usually confront a long list of possibly influential factors. Unlike physicists or engineers, we cannot demonstrate that factors are unimportant in order to justify ignoring them. Instead, we are subject to what is known as confirmation bias. That is, we tend to selectively cite observations that confirm our views, ignoring other factors that might be at work. However, when observations appear to confound our views, we seek out and cite those other factors. If its applications had fallen when Ursinus College raised tuition, we would not have looked for other explanations. However, when demand increased, we are inclined to examine other factors.

Is economics a science? Some people, mostly economists, believe that it is. On the other hand, other people, mostly noneconomists, are skeptical or even scornful of what economists teach.

I think that both camps are guilty of underestimating the challenge of arriving at economic understanding. Those economists who claim the mantle of science are guilty of hubris. Noneconomists who think that their own intuition is superior to economic reasoning are dangerously misguided.

Imagine that you had a scale to measure the carefulness with which someone reasons about a subject. Let that scale run from 1 to 5, with 1 representing the most careless sort of reasoning, filled with superstition and personal biases, and 5 representing scientific reasoning, based on mathematical logic and experimental observation. Where does economics fit in?

I believe that good economics is at least a 6! That is, good reasoning in economics requires more careful thinking than good physical science—for two reasons. First, more causal factors are at work in economics than in physical science. Second, although physical relationships are relatively stable, the economy evolves rapidly, including evolution in response to government’s attempts at regulation.

A key component of the scientific method is making statements that are verifiable. A proposition can be verified only if it can be tested against a standard of truth. Putting a proposition up against a standard of truth means taking the chance that the statement can be falsified. Thus, scientific propositions must have the potential to be falsified. This philosophy of scientific inquiry is called “falsificationism.”

For the most part, statements that qualify as scientific propositions are falsifiable. They are either mathematical proofs, which can be falsified by showing a flaw in their internal logic, or else hypotheses about what we observe in the world, which can be falsified through careful observations and experiments.

According to that scheme, a belief that cannot be falsified either by logic or by evidence is nothing but dogma. Dogmatic beliefs cannot be falsified, but that is only because you hold onto your dogma regardless of any arguments that can be raised against it.

Reasonable beliefs should not be false, of course, but they should be subject to testing against logic or observation. To put the case for falsificationism another way, one would say that any proposition that cannot be falsified is by the same token a proposition that cannot be verified.

If you hold onto a belief so dogmatically that no evidence could change your mind, then that belief is not falsifiable. Nonfalsifiable dogma is the worst sort of belief. Reasonable people can settle differences of opinion regarding falsifiable statements. Not so with dogma. If that is the case, then scientific argument becomes pointless. That is why scientists prefer to deal in propositions that are falsifiable.

However, not all scientific beliefs are falsifiable. A few key beliefs, called paradigms by Thomas Kuhn,¹¹ and which I will call “frameworks of interpretation,” are so fundamental to how scientists view their subject that they are almost beyond question. For example, Darwin’s theory of evolution is a fundamental framework of interpretation in biology. Biologists no longer ask whether Darwinian evolution can explain phenomena. Instead, they talk about how the theory can be adapted to provide explanations.

A framework of interpretation cannot be falsified. However, many frameworks suffer from anomalies. In evolution, for example, some phenomena, such as a peacock’s large tail, would appear to reduce survivability. To address that anomaly, biologists have suggested that the large tail signals strength and attracts potential mates, thereby actually tending to increase the survivability of that characteristic.

The difference between a falsifiable proposition and an interpretive framework is that it takes only one anomaly to reject a falsifiable proposition. A single clear-cut logical flaw serves to falsify a logical proposition or mathematical proof. A single conclusive experiment serves to falsify an empirical hypothesis. However, a single anomaly does not lead someone to abandon an interpretive framework. (Keep that in mind the next time you see someone claim that “this one chart” provides definitive proof for or against a particular economic viewpoint.) An anomaly makes scientists uneasy, but they look for ways to address the anomaly without abandoning their interpretive framework.

Up to a point, scientists will stick with an interpretive framework in spite of anomalies. However, if enough anomalies accumulate that scientists become uncomfortable with a framework, and they find that an alternative framework addresses the anomalies and is compatible with existing knowledge, then they will switch to the new framework. That switch is what Kuhn calls a scientific revolution.

In general, I shy away from using the term “social science,” because I do not think that economists can aspire to the same level of falsifiability as physicists. I believe that the difference between social science and natural science boils down to this:

In natural science, there are relatively many falsifiable propositions and relatively few attractive interpretive frameworks. In the social sciences, there are relatively many attractive interpretive frameworks and relatively few falsifiable propositions.

The reason that there are relatively few falsifiable propositions in the context of social phenomena is that many causal factors exist, and decisive experiments are rarely possible. Social phenomena are characterized by high causal density, to borrow a term from James Manzi.¹²

As a result, economics is closer to history than to physics. If a historian wants to examine the causes of the decline of Rome, or the decline of empires in general, he or she will provide an interpretive framework. That framework cannot be falsified, but readers can compare it with other frameworks and make judgments about its plausibility.

For example, consider the phenomenon of the comparative salaries of men and women. Economists interpret salaries using the framework of human capital. That is, workers bring to the market different levels of ability, training, and experience, and those attributes determine what they are able to earn. Sociologists use a framework that emphasizes group identity, status, and power, with men the more dominant group and women the more oppressed group.

If a study were to suggest that women earn less than men, even when controlling for years of education and other indicators of human capital, then that would be an anomaly for the economists. If a study were to suggest that most of the lowest-paying occupations are occupied predominantly by men, then that would be an anomaly for the sociologists. However, such observations will not prove decisive. By invoking other factors to explain anomalous results, each side can remain unmoved. Economists will not abandon their human capital framework, nor will sociologists abandon their group-status framework.

What economists call “models” are interpretive frameworks. They are presented mathematically, with proofs that connect initial assumptions to ultimate predictions. However, the predictions are not falsifiable. The models’ predictions hold only when other things are equal, and other things are never equal.

For example, consider the very common equation Y = f (K,L), which says that output is a function of the amount of capital and the amount of labor. One obvious prediction is that more of either factor will tend to increase output.

That production function is used to interpret data in various contexts, including making comparisons of labor productivity. For example, suppose that Alan’s lawn service can mow more lawns per worker than Bob’s lawn service. The first variable that an economist will look for to explain the difference is the number of lawn-mowing machines per worker at each firm. If Alan’s service does not use more lawn-mowing machines per worker than Bob’s, then the economist will look at the quality of the mowing machines at the two firms. If that does not explain the difference, then the economist will fall back on “better management” or some other factor. The less closely that the explanation can be tied to capital, the more anomalous the result will be.

Economists actually try to use the production function to explain productivity differences between entire countries or to explain the historical path of productivity within a country. However, that approach requires taking a weighted average of many different types of outputs and treating the weighted average as if it were a single type of output. Similarly, economists must construct measures of aggregate capital and aggregate labor by taking weighted averages of many different types of each. Many other factors affect aggregate productivity, including endowments of natural resources, government policies, and the diffusion of knowledge. Not surprisingly, in empirical studies, many anomalies can and do crop up, so that the issue of what causes productivity to differ across countries or to change over time remains highly controversial.

Another challenge for economics is that the economy evolves. Consider some of the factors in the relationships between aggregate output, total labor input, and total capital input. Imagine trying to compare the U.S. economy today with that of 50 years ago. We have to take into account major changes, including the following:

• Many fewer people are in the labor force with less than a high school education, and many more people have at least some college education.

• The share of output in agriculture and manufacturing has fallen, whereas the share of output that consists of services has risen.

• Some outputs today, such as smartphones and heart transplant surgeries, cannot be compared with outputs of 50 years ago.

• The share of workers directly involved in production has fallen. The share of workers who are developing organizational capacity has risen.

• The share of computers in total capital has been rising. The cost of this particular type of capital equipment has plummeted sharply, and its characteristics have changed radically, making it difficult to measure reliably how the value of investment in computers has changed over time.

The economy also evolves as new business models, new production processes, and new institutions emerge to solve problems. The “market failures” identified in economic models are only a small fraction of the imperfections that exist at any one time in the economy. Businesses and other organizations are constantly working on solutions to those problems.

Nobel Laureate George Akerlof famously provided an interpretive framework for the used-car market in which high-quality used cars would be kept off the market, because buyers would have to assume, in the absence of other information, that all used cars were “lemons.”¹³ However, that framework assumes that no market adaptation exists to address the problem. The information problem in the used-car market can be addressed in a variety of ways. For example, mechanics can inspect used cars before consumers purchase them. Sellers can offer warranties on the cars. Decades after Akerlof ’s article was published, a national used-car dealer called CarMax emerged with a business model based on a reputation for selling high-quality used cars. Other services emerged to make the repair and service records of used cars transparent to buyers.

Markets also adapt in response to our attempts to regulate them. For example, economists have pointed out that the way in which physicians are compensated in the United States, with billing based on procedures, distorts the incentives of doctors so that they tend to perform too many procedures that have high costs and low benefits. However, if that system were changed so that doctors were compensated only according to the number of patients that they see, then we would likely have the opposite problem: to bill for as many patients as possible, doctors would try to avoid doing time-consuming procedures. If doctors were compensated on the basis of patient outcomes, then they would select patients who were likely to have good outcomes, avoiding some of the most difficult patients.

Because of causal density and evolution, economists cannot be certain of the reliability of our assumptions. Thus, any interpretive framework may be inappropriate, depending on circumstances.

Economic models contain many unverifiable assumptions in a context in which plausible alternatives exist. Consequently, when we observe, say, contrary to the expectations derived from a model, a decrease in the price of milk, or an increase in the overall unemployment rate, w...

Table of contents

• Cover Page
• Title Page
• Copyright Page
• CONTENTS
• Foreword
• Acknowledgments
• Introduction
• 1. Filling in Frameworks
• 2. Machine as Metaphor
• 3. Instructions and Incentives
• 4. Choices and Commands
• 5. Specialization and Sustainability
• 6. Trade and Trust
• 7. Finance and Fluctuations
• 8. Policy in Practice
• 9. Macroeconomics and Misgivings
• 10. Concluding Contemplation
• Appendix: How Housing and Mortgage Policy Worked in Practice
• Endnotes