A First Look at Economics
5. How to produce economic knowledge

Maria has heard from several friends, recent graduates, about struggling to find their first jobs. People with good grades and strong resumes are spending months applying for jobs but getting few interviews. Maria naturally wants to know what the job market is going to be like when she graduates. Important question. How can she find out?

The "unemployment rate" is a pretty good indicator of how hard it is to find a job. It's a somewhat strange word, "unemployment," meaning the percentage of the workforce that's between jobs and looking, or leaving school or home full-time, and looking. On the first Friday of July, 2006, the Bureau of Labor Statistics reported that 4.6 percent of the U.S. labor force was unemployed. Some European nations developed unemployment statistics a little before the United States did, and it's easy to find them there, too, for Amsterdam, for Milan, or for nearly any city in Europe.

Maria wants to know what the unemployment rate is going to be in three years, when she graduates. That requires a macro-prediction, very, very difficult to do. A bit like predicting what the weather will be three years from now. Don't believe a tipster on TV who tells you what the interest rate will be in six months, say, or the unemployment rate in three years.

But if she knows why the unemployment rate is what it is, she can at least be assured that it's not going to shoot up suddenly to 50%, or drop tomorrow to zero. Again, like the weather. If you know, say, that hurricanes lose their force over land you will not be driven by hysteria on the Weather Channel to worry about a hurricane in Cuba doing damage in Minneapolis. The biggest change in human understanding since the invention of writing was the invention of modern science, in the 1600s and 1700s. The scientists said, "Hurricanes have natural causes, and are not God's punishment." Likewise, say the economic scientists, "Unemployment has stable causes, and does not leap up and down erratically. Like the weather, it may be unpredictable in fine detail, but the wider patterns are predictable."

Klamer: Maria, suppose you're suddenly given a job as an economist assigned to find out why unemployment is what it is today.

Maria: Terrible thought!

McCloskey: Or not: there's free entry to the profession, unlike doctoring since about 1900, or lawyering since about the same time, anyone can claim to be a qualified economist!

Klamer: True enough, Deirdre. But, Maria. . . .

Maria: Well, I'd go around and ask people why they are unemployed, and ask the companies why they aren't hiring more workers.

Klamer: A natural and sensible answer. You would do as a journalist would do covering a story. An economic scientist had better be doing some of that kind of journalistic conversing with the employed, the unemployed, the bosses, the employment agencies, and the rest, or she'll miss a lot of the facts. But she would worry whether the unemployed person actually knows why he was let go. He might think that the boss in, say, his lamp assembling factory didn't like him. But the actual reason may have been a drop in lamp sales. And the boss at Almed Lamp might think that the reason she had to let people go is that Better Lamp and Industrial Lighting Company, Inc. was cutting into her business, when actually a rise in imported lamps from Thailand was the root cause. . . and a rise in American borrowing abroad the root cause of that.

Maria: Wheels within wheels!

Klamer: Fair enough. But I promise that if you work hard and diligently in this course you will be able to say a lot about the links between those "root causes."

Rodney: Instead of interviewing, journalism-style, the people involved she could ask an economic scientist.

Klamer: An excellent idea: since we don't have legal protection for the profession of "economist," we economists need people to have that attitude! But how does the economist know?

Explaining (or predicting) by theorizing

Maria's idea for gathering information is not unreasonable. But mere data collection is not always the most efficient way to learn, and it may even lead a researcher far off the story. A journalist has to have some notion, a theory of information, about who to ask. You have to have a theory of how a phone book is organized before you search in it. If, idiotically, you think the only way to approach finding Connor Nolan in the Chicago directory is to start reading the book at page 1 until you hit his name, you're going to spend a long, long afternoon reading. Since in fact you have the earlier knowledge---you can call it a "theory"---that a phone book is organized L, M, N, O, P, and among the Ns it goes Na, Nb, and so forth, you get Connor's name in ten seconds.

The point is true of some very fancy scientific work. Einstein didn't hit upon the theory of relativity by wading through masses of data. On the contrary, he came to it by thinking - logically and creatively - about problems. For example, the riding on the moving train versus standing and watching the moving train problem of perception. Similarly, in the introduction to his famous 1859 book, The Origin of Species, the English naturalist Charles Darwin, claimed that he discovered "natural selection" by using what most people in the 19th century thought of as the scientific method: collecting facts and then generalizing from them. But his Autobiography reveals that he actually got the theoretical idea that allowed him to find and organize the data from reading. . . an economics book!-an Essay on Population (1798), by Reverend T. R. Malthus.

Economists work things out in their heads, on pieces of paper, on blackboards, and even on luncheon napkins. It is one way to seek knowledge of the economy. We call it the theoretical strategy.

A theoretical strategy involves the construction or application of theories.

A theoretical statement takes the general form "If X, then Y." If the phone book I'm looking at is rationally organized, then I should be able to find Connor's entry in 10 seconds. If God does not play dice, then light will bend around the sun. If pigeons compete with others for their food and procreation, then show-off feathers in the males should be big and bright. After the "if" come certain conditions and simplifying assumptions; they describe a hypothetical situation. After then "then" comes the conclusion in the world. An economic theory could read, for example, "If the market for baseball tickets works as expected and if the demand for tickets exceeds supply, then the price of baseball tickets will go up." It has, actually.

Theoretical reasoning is how we think. "If I were rich and as generous as I am now, then I would give all my money to the homeless." With a theoretical strategy you specify a general principle ("I am generous"), add specific conditions ("I am rich"), and derive a prediction of an unknown event ("I'll give all my money to the homeless"). Thus by sheer thinking you try to determine what will happen. In physics it's called a "thought experiment." Economists settle for the word "theorizing." You can call it "thinking."

One sensible goal for a scientist is to be precise and explicit. So many economists insist on expressing their theories in mathematical equations or diagrams - to make their reasoning more precise and explicit. They construct theoretical models. Like a model car, an economic model is a simplified representation of reality, often expressed in mathematical terms. Here's one, just to show you what the math looks like: "If given D(P) and S(P), and if P is not prevented from moving, then P tends to the solution of D(P) = S(P)."

The model car is simplified because it is smaller than the real thing and is made of wood or plastic instead of metal and fiberglass. Yet automobile manufacturers can use it to do experiments, such as testing the car's aerodynamics. Similarly, economists can use an equation or a computer to "simulate" economic reality.

An economic model is a simplified and artificial representation of an economic system. It is usually specified in mathematical terms.

Economic models force you to spell out all the relevant assumptions and conditions. What are the properties of the system when the system is at rest? What if one of the conditions changes? What motivates the economic people in the system? If something crucial is left out, the model doesn't work. It "doesn't work" either in a logical sense---more unknowns to be determined than there are equations to determine them, say. Or in a factual, usability sense: it doesn't work to tell you correctly what to expect in the world.

You can experiment with a mode by changing the assumptions and conditions to see what happens. An economic model makes it possible to calculate the effects of government spending on employment levels. It's like figuring out the effects of a higher wind speed on the performance of a model car. And the model can be tested against the facts. For example, if the model implies that unemployment will fall if government spending goes up, we can investigate the truth of this prediction. Is it true? Only a confrontation with actual cases will tell.

So economics is partly philosophy and mathematics, as we said earlier, a careful thinking out of what we already believe (or assume) to be true, using models. Yet economics is also an empirical science like geology or history, as we also said. Economists keep trying to confront their theories with the facts. The empirical estimation and testing of economic models is no picnic. Economic reality is complex, and people-"economic agents" are foxy. It's the same in physics. The great American physicist Richard Feynman wrote that a merely mathematical approach to physics fails "because the actual physical situations in the real world are so complicated that it is necessary to have a much broader understanding of the equations."

Other ways to learn about the economy

Historical inquiry

In addition to abstract modeling, economists study past events in order to gain new insights into how economies work, or have worked. This strategy does not preclude the use of theory. In fact a whole field has arisen, called historical economics, in which economic theory is used to improve our understanding of history and vice versa. In 1993 Robert Fogel and Douglass North shared the Nobel Prize in Economic Sciences for inventing it. Fogel has for example changed the way economists think about the rise of the railway, the fall of slavery, and the decline of disease and mortality. After Adam Smith, most of the great economists have made some historical contribution. The core skill of the economic historian is the construction of plausible explanations about the economic past---stories informed by theory and measurement.


Maria suggested sensibly that we seek knowledge by asking people. The very numbers economists use are often based on asking people. To find out how many people are unemployed, officials from the Department of Labor, following up on a paper survey, literally knock on doors and ask people whether they have looked for work in the last week.

Surveys and interviews do not always tell us why something is happening, for reasons we just explored. But they certainly give us a vivid idea of what is happening. Many economists are suspicious of surveys and use them only sparingly. They note that people generally have nothing to gain by telling the truth in surveys. "What value do you put on your iPod?" So the answers people give should not be taken at face value. We don't find the economists' argument very convincing. After all, often the people have nothing to lose by telling the truth. Social researchers have shown that carefully designed surveys can produce important insights. When Ziliak was a social worker for the welfare department he found he could nearly rely on the oral statements made by 90 percent of his clients. There are signs that surveys are starting to come back into favor with economists, pioneered by the economic historian Richard Easterlin, of the University of Southern California, who has long used surveys of how happy people are.


Economists for a long time thought that they could not do controlled experiments to test their theories the way physicists or chemists could. It's true that we can't shut the government down to see how an economy would behave without police and politicians (though in summer 2006 the Governor of New Jersey gave it his best shot!). And when economists want to know the effects of a tax hike, they can't prevent other relevant factors from changing at the same time. They can't be entirely sure that the tax hike is responsible for lower spending or whether, say, the higher prices of imports caused lower spending. The physicists (it is said) have it easier, since they can conduct controlled experiments. (People often forget that many branches of physics, such as geophysics and astronomy, cannot do much in the way of experimentation, either.)

Since around 1970, however, a few economists have begun to do experiments--- though not on whole economies: that is left to former Communist countries like Poland or China. Inspired by the experiments psychologists have conducted for over a century, experimental economists take people or rats off the streets and study their reactions to carefully constructed situations. Experimental economists hope to learn how people-or, they say, a good substitute for people, rats-respond in actual situations. In one sense the situations are not true to life. After all, they are experiments, not actual business or market situations. But in another sense they are true enough. After all, if an economic theory is correct, it should be able to explain the behavior of people in experimental situations, too. Experimental economics got a boost in 2002 when one of its pioneers, Vernon L. Smith, Jr., was awarded the Nobel Prize.

Likewise, economists have recently revived the technique of "natural" experiments. A natural experiment in labor market adjustment occurs when influenza kills off ten percent of a large supply of labor, producing (sadly) an observable trace of "before" and "after" wages, as in India in 1919. Or when the State of New Jersey raises its minimum wage relative to neighboring Pennsylvania, other things equal, as it did in the early 1990s. What constitutes a useful natural experiment, you will not be surprised to hear, is a matter of some disagreement. But natural experiments, such as hyperinflation in Brazil (2000% per year), or the abolition of welfare to the poor, or enormous foreign aid in Ghana, or the Chinese Communists' shift to capitalism, or the U.S.-only embargo on trade with Cuba, are often the most persuasive economic arguments we have.