Interview with Claudia Goldin

"When I was younger, the idea of peering through a microscope and discovering an entire world was fascinating," says Claudia Goldin, recalling a time when microbiology seemed her certain destiny. "The things you might find there could have an enormous impact."

In college, though, she was captivated by economics, conveyed with zeal by Alfred Kahn. In her doctoral studies at the University of Chicago, she studied the economics of social forces with Gary Becker and learned the power of economic history from another mentor, Robert Fogel.

Goldin made her own mark in economics with a pathbreaking study on urban slavery. She then devoted herself to the labor history of American women, another neglected area in economics. Recently, she's focused on educational movements and their influence on the U.S. economy. She has contributed, as well, to the economic literature on American immigration history, labor markets, technological change, inequality, the impact of war and many other topics. Throughout this career, her hallmarks have been original thought, intense investigation and a clarity of expression rare among economists.

Goldin has garnered numerous professional honors: the first woman to receive tenure in Harvard's economics department, a fellow in the American Academy of Arts and Sciences, etc. But her motivation lies in the work, in the eureka moments. Goldin's passion is raw data. She relishes digging through long-buried library files, dusty school records, obscure newspaper articles—always insisting that first-hand research is critical to true understanding.

Indeed, what's apparent from a conversation with Goldin is that the process of discovery fascinates her now just as it did in junior high. Microbiology is the poorer for it, but Goldin followed a parallel path—discerning unseen worlds through the lens of economics, finding things that have had enormous impact.

The Economics of Slavery

Region: Your career began with work on the economics of slavery. At the time, slavery was becoming a very divisive issue among economic historians, including your adviser at the University of Chicago, [Nobel Laureate] Robert Fogel. Have the major economic questions about slavery been resolved? And what is your view about reparations, an issue that Brown University is currently deliberating?

Goldin: I first became interested in the subject of slavery because my roommate in graduate school was a historian. She had gotten to know many radical historians in Chicago and also Gene Genovese, who was quite famous for his work on the South and on slavery. So our little apartment was a salon through which many of these people would come. What got me interested was a book written by a Chicago historian, Richard Wade, called Slavery in the Cities. And when I wrote the paper [on the subject] for Bob's course, he became very excited about it. He encouraged me to continue the work, but it wasn't until after I wrote the paper that I realized what he and Stan [Engerman] were working on.

It is now 30 years since Bob and Stan published their book, Time on the Cross, and more than 35 years since they began their work in the area. Prior to their work there were many other economic historians who contributed to our understanding of the economics of slavery, such as John Meyer and Alfred Conrad, let alone many historians.

I do believe we have a much better idea of the role that slavery played in the Southern economy. As economic historians we emphasize issues that have to do with economic aggregates, so we hold aside the subtleties of the human experience. One of the big issues having to do with the South and slavery concerns income per capita. A great paradox emerged concerning the fact that in the antebellum period, income per capita in the South (not accounting for distribution of that income, of course) was high and growing. By 1860, the South as a region on a per capita basis had income exceeding that in the burgeoning Midwest (which would have been called the West). Not higher than the North, which was industrializing—we would expect the North to have had a higher income per capita—but compared with other U.S. agricultural regions, the South was not poor and not stagnant.

So the paradox is that after the Civil War, the South did become poor. It was not stagnant, but it was relatively poor. Something big changed between 1860 and, say, 1880 and the question is: What happened? The work on the antebellum period and on slavery—and Bob and Stan's work in particular, I think—answered this question.

Not surprisingly, what happened was that in one period you had slavery, you had forced labor, and you had a system of production in which there were economies of scale for various reasons. And then you emancipate slaves, and the economies of scale that came with forced labor vanished. So in that sense economic historians made enormous gains in understanding the role that slavery played in terms of the differences in the wider economies of the North and the South.

Much of the literature on economic convergence among the U.S. states concerns the catching up of certain areas, especially the South. But we first have to understand why the South fell apart after the Civil War. One large part of the answer has to do with the ending of forced labor, with Emancipation. The high income in the antebellum South came because slaves were used in a particular form of agricultural production called the "gang system," which was brutal and harsh.

Historians might cringe at the notion that I could speak about slavery without talking about bondage and the human condition, about distributional issues and the family, about separation from loved ones and what has been termed "social death"—the notion that slaves didn't have a social identity—and also that I haven't said anything about distinctions between house slaves and plantation slaves, nor have I said anything about being "sold South."

In defense of economic historians, they have studied variations within slavery and, as such, have brought to light many different aspects of slavery as an economic system. My book on urban slavery, for example, concerns various aspects of slavery and its tremendous flexibility. Slaves were hired in cities and often hired out their own time. So large plantations, small plantations, small farms, the upper South, the lower South, the cities, the rural areas, industry in the cities, all of these areas give us a somewhat different view of what slavery was: slavery as a social system and as a work system, the horrors of slavery, the lesser horrors of slavery. So I think that we have gone a long way.

The Brown discussion and all questions about reparations are extremely difficult. Who gained and who lost? One of the large groups that gained includes anyone who bought cotton textiles. So, it's very complex. I'm not saying that it shouldn't be raised because I think we should always keep in our minds that there are labor systems that we rule out and there are social relations that we rule out, but I don't see that resolving the reparations issue would be a simple thing.

The Pollution Theory of Discrimination

Region: You've developed a theory of discrimination that you consider a hybrid of the discrimination model developed by Gary Becker, on the one hand, and that of Ken Arrow and Edmund Phelps on the other, and you suggest that your model can more fully explain discrimination against women in the workplace. Could you give us a brief statement of this "pollution theory" and what it adds to existing theory?

Goldin: It was only after I wrote a first draft and presented it that I realized the model was a hybrid. When I began the model, I thought of it as a reaction to our usual views of what we mean by discrimination. We generally mean that there is some bias or some prejudice. Various economists, and in particular Gary Becker, have interpreted discrimination as a distaste of one group for another, meaning a desire for distance. This interpretation may be correct for discrimination between races, religions and ethnicities, and in Gary's able hands this very simple assumption about what discrimination is goes a long way. His model has always been simple and brilliant.

But it didn't seem to be an accurate description of discrimination by men against women, or vice versa. Sex and maybe other groups are different because there doesn't seem to be a desire for distance. Men and women marry and they have children together. Every man has a mother and many have sisters and daughters. There doesn't appear to be a desire for distance, so how could there be a distaste for women by men? It has to be something else.

But more important than that, I wanted to model discrimination from some first principles, and not take it as a distaste, as if preferences are just given, a distaste for people of a different color or religion, for instance. Such models have astounding implications, you can go a long way with them, but it wasn't where I wanted to begin.

I began, instead, with the question: Why have men and women throughout history referred to certain jobs as women's jobs and other jobs as men's jobs? Take milking cows, for example. In the Midwest, men would say that's women's work. But oddly enough, in the East it was considered men's work. You would denigrate a man in the Midwest by saying, "you're doing women's work" if he were milking a cow. And you would do the same to a woman in the East by saying, "oh, you're doing men's work. There's something the matter with you."

I wanted to find the roots of this. I began with the possibility that men and women were actually identical in the distribution of their abilities—although the model, of course, could begin with differences—but that men didn't have experience working alongside women. It's a historical model in the sense that men and women begin in different spheres. Women begin in the home sphere and men begin in the outside or market sphere, so people didn't know how to discern the individual ability of women. They know only about the median.

And this informational asymmetry gave rise to a desire by certain men—and it turns out to be those who are above the median in the female distribution of some characteristic—to restrict entry into their occupations so as to preserve their prestige or their status. The model is considerably more complicated than this, but that's the main starting point.

Region: And why do you call it a "pollution" model?

Goldin: Because they feel that the entry of women into their occupations would pollute their prestige or status in that occupation. Very simply, some external group is the arbiter of prestige and status. Let's take an example of firemen, and let's say we begin not that long ago when there were no women who were firemen—which is why they're called firemen. And to become a fireman you have to take a test, lifting a very heavy hose and running up many flights of stairs.

And every night, the firemen get off from work and go to the local bar. Everyone slaps them on the back and says what great brawny guys they are and what a great occupation they are in, and everybody knows that to be a fireman requires certain brawny traits and lots of courage.

But nobody knows when there's a technological shock to this occupation. And in this case it might be that fire hoses become really light or the local fire department changes the test. There are information asymmetries.

But they do note that for this "brawny" characteristic, the median woman is much lower. So if we observe a woman entering the occupation and we don't know how to judge women, we're going to assume that her skills are those of the median woman. Or it may be that we can observe something having to do with her muscles and that may up it a little bit. But chances are we're going to assume that some technological shock has happened to this occupation. And so her entry into the occupation is going to pollute it. Then when they go to the bar, people will say, "oh you've got a woman in the firehouse; now firefighting has become women's work." That's where the pollution comes in.

Women in Economics

Region: You've been very active with the Committee on the Status of Women in the Economics Profession [CSWEP]. Its most recent report indicates some "serious shortfalls" in the status of women economists at Ph.D.-granting institutions but also shows considerable progress in overall long-term trends. Does the pollution theory help to explain what's going on in economics? And are you concerned about the status of women in the field?

Goldin: Good questions. The implications of the pollution theory are that in the absence of credentializing, you'll get segregated jobs. But if you have credentializing, then you will get more integrated jobs, because if we could prove without a shadow of a doubt that that woman took the test and the test was the same type of test that men took, then they'll be accepted into the job. They won't pollute. So everybody knows (or thinks) that a doctor is a doctor, a surgeon is a surgeon, a lawyer passes a particular bar, so that's a good credential, and someone who gets a Ph.D. has passed some barrier and is credentialized. So, I don't think that the pollution theory can explain what's going on in economics.

Am I concerned about what's happening in economics? The answer is yes, for three reasons. The first is that of new Ph.D.s today, 29 percent are women, up from 7.6 percent when I received mine. That's progress, but it's still fairly low. Here at Harvard College, even though the entering class is now about 50 percent female, the fraction of economics concentrators who are women is much lower, about 30 percent. And economics is the largest concentration in the university.

One concern is that women are somehow turned off from the field. If you look at Ec 10, our beginning class, it looks like a cross-section of the frosh class. Women take Ec 10 but then they don't concentrate in economics. This could have something to do with what they learn in Ec 10. Or it could be that they want to know some economics but they don't want to concentrate in economics. So I'm concerned about why women are turned off from the field, either before they enter universities or while they're at universities.

The second concern is that 35 years ago, economics was like law and medicine. So 35 years ago 10 percent of entering med students were women and in law school it had just increased from 5 percent in the late '50s to about 7 percent by 1970.

So 35 years ago economics was just like law and medicine. Today the fraction female in law and medicine is almost 50 percent. It's almost at parity. Certainly in law, a little bit less in medicine, I think. But economics—rather than looking like law school—is beginning to look more like computer science. And that is a concern.

Region: Does that reflect the fact that economics has become increasingly quantitative?

Goldin: I don't think so, because women have increased their numbers in math and have always been inclined to do math—not getting Ph.D.s in math but being undergraduate math/statistics majors. So I think it's something else. I don't want to hazard a guess as to what, but it's not something that's changed over time because the point is, it's not changing. It's not as if we started high and now we've gone low.

Another concern—which is an equivalent concern for many other fields in academia—is that even though we're producing 29 percent female Ph.D.s, if we go back 10 or 15 years, we were still producing about 29 percent then. In other words, we started at 7 percent, we went up to 20-something percent and we've been wiggling around that level for quite some time. It's reached a plateau.

I remember 10 or 15 years ago we would speak of the fact that the tenure ranks were not 29 percent, even though we were producing 29 percent women in Ph.D. classes. The tenure ranks were like 5 percent or 8 percent—the exact numbers are in the CSWEP report—and the answer would always be, "oh, it's a pipeline issue," that once we get enough women in this pipeline, the tenure ranks will begin to increase. But we're not there 10 or 15 years later. That issue is one that we also see in other fields. We produce academics, they then enter the pipeline and the pipeline leaks. In economics, you can leak out to private enterprise, to government, to jobs for which the biological clock and the tenure clock are not running together.

And even those who don't leak out at that point tend to leak out relative to their male counterparts at every other step. They leak out from top-ranked institutions to lesser-ranked institutions. They leak out from top-ranked institutions where they are trying to get tenure to nonacademic positions. So at every step they're losing out.

Region: One of the pipeline concerns highlighted by CSWEP is that while women hold 28 percent of tenure-track economics professorships at liberal arts schools, they hold only 11 percent at the top 10
Ph.D.-granting institutions, and that may reflect what you say about looking for a different quality of life.

Goldin: I think that's it, quality of life. As women are more valued as faculty members, universities, just like law firms, might have to reexamine the touchy issues of whether work effort should be front-loaded. The biological clock and the tenure clock are now ticking on exactly the same schedule. Even if you are in the incredibly fast lane, it is almost impossible to have both career and family on the same clock. You go to graduate school when you're, say, 22 or 23, you get your Ph.D. by the time you're 28. Then you go to your first job and in six or seven years you have tenure. Well, that puts you in your mid-30s.

The president of Harvard, for example, is extremely concerned about this issue, because every young academic knows that when they enter Harvard, Yale or Princeton or other highly selective institutions, for them to stay they're going to have to put out 150 percent. And gosh, even having a dog takes a lot of time. Thank goodness for doggy day care. That's where my dog, Prairie, is right now.

Why Economics?

REGION: Well, the next logical question is, why did you choose economics?

GOLDIN: For as long as I can remember, I was going to be what I would call a scientist. In my family, there were only two subjects. There was science and there was math—it would probably go in the order of math and then science. Those were subjects, and anything else was fluff.

REGION: You went to the Bronx High School of Science?

GOLDIN: That’s right. As a teenager I thought I knew everything—in part because I went to Bronx Science, in part because I was a New Yorker and in part because I was just me. But then I went to college, Cornell University, and instantly—I remember the moment it happened—instantly, I discovered how little I knew. But I wasn’t crestfallen, I wasn’t discouraged. In fact if anything, I felt exhilarated. It was as if someone had opened a door and I saw a new world. A gigantic room. A room filled with beautiful art objects and delicious foods and I was just going to have all of it.

So I discovered this gigantic world of ideas and of learning and I just couldn’t let it pass me by and I couldn’t pass by it. So I became what I might call a liberal arts junkie. And I just tasted everything and looked at everything. And some of the aspects of the world of ideas I admired but realized I wasn’t going to be very good at. I remember writing an essay once in literature. Somehow I decided that I could write and that I had brilliant thoughts and ideas, and you know your professors and lecturers and instructors, they encourage you because it’s their field and so they want to encourage you.

But every now and then I would get comments on my essays that whatever it was that I was discussing was not a math equation and that I would use words like “equate,” for example. And so my use of language was somewhat limited from my more mathy/science background.

And there were other fields I knew I wasn’t going to be very good at. For example, I decided that anthropology was fascinating, and I remember one of the few times I ever went to the office hours of any faculty member I went to see an anthropology professor whose course I was taking to tell him how brilliant this work was and how excited I was about it.

I was reading the classics of anthropology and the professor was so excited that he had in his midst someone who was going to graduate work in anthropology. Then he laid out very carefully all the obscure-language courses I would have to take if I was going to be a great scholar in anthropology, and foreign languages were never fields that I enjoyed. So I realized in some fields I wasn’t comfortable and in some fields I wasn’t very good.

And then I discovered economics. Economics was not something that was given in the high schools of my day—in the high school I attended or in any high schools that I knew of in New York City. And I was very fortunate when I was at Cornell to have the great Fred Kahn, Alfred Kahn, as my industrial organization (IO) professor and to write my senior thesis under his tutelage—although he’d probably say that he never once saw me during the time I wrote my senior thesis. I went off and did it myself.

REGION: What was it on?

GOLDIN: I have no idea how I got onto the subject but I was fascinated by it. It was on communications satellites, which had just begun, so the question was how do you regulate communication satellite rates, and entry?

REGION: Impressive. It seems ahead of its time.

GOLDIN: It may have been ahead of its time, but I was absolutely fascinated with IO and regulation. And part of it was Fred. He has always been a great showman. He was in many Gilbert and Sullivan productions. He was a great showman, and he was also in love with the work that he was doing. He later went into public service and did something that few people have ever done: abolished the agency [the Civil Aeronautics Board] that they were asked to head.

When I left Cornell, I was going to do something that sounds mundane today: go to law school. But women didn’t go to law school when I graduated from college, so it was a little odd. And somehow I hit upon the idea of going to Chicago. Even New Jersey was a remote state as far as I was concerned so going to Chicago was quite adventurous.

I knew little about the university, but in my readings in IO I must have learned that the great IO scholars, such as George Stigler, were at the University of Chicago. And then I decided that maybe I should do work in economics. Fred was not happy with my decision to go to Chicago, but I think he was happy that I was going to do economics and that I wanted to do IO.

So I went to Chicago. In some sense, then, going to the University of Chicago was accidental, but it probably was the best decision I could have made. When I went to Cornell, the room that I entered was filled with paintings and good food. But Chicago was a castle, a completely different universe. I walked in and realized, once again, that I knew nothing. Now I knew absolutely nothing. But at the same time there were brilliant people who were going to help me, and so that’s how I became an economist.

I should just add that when I entered Chicago I entered to do IO and I did a field in IO. And then Gary [Becker] arrived, and once again I realized that the world of economics was much larger than I had thought. Gary was doing brilliant work on many different issues that I would call the economics of social forces. And then, to make things even better, I met Bob Fogel. It was those three. I did a field in IO, a field in labor, and then I wrote in economic history. So Bob mesmerized me with economic history, and that combined my liberal arts junkie taste with my more rigorous math sensibilities.

Working Women

REGION: Married women roughly tripled their hours in the workforce between 1950 and 1990, while men and single women barely changed their participation. Several standing theories try to explain the trend through declines in wage discrimination, liberation from housework by labor-saving technology, etc. What is your understanding of this phenomenon?

GOLDIN: This is the big question. The subject of the female labor force is extremely complicated—filled with large, long-term changes ranging from fertility to education to technological factors such as housework. We see very similar changes in many emerging nations.

When you have a complicated problem such as this one, it’s very useful to organize one’s thoughts along the lines of simple supply and demand. Is it supply shifts, and what do we put on the supply side? Or is it demand shifts, and what exactly are we putting on the demand side?

But there are also serious measurement issues to consider. Women worked in the past just as they work today, but they didn’t always work in paid labor markets. If we go back a hundred years, the reported labor force participation rate of white married women was low. But if we went into the homes of these women, they would be working all the time—unless they had household help, unless they were relatively well-to-do. It isn’t that they were just doing household labor. They were running boarding houses. They were taking in laundry, sewing on piece-rates. But they weren’t telling the census taker that they were doing it full time either because it wasn’t full time or there was a social stigma associated with such labor.

So, many of us have thought long and hard about this problem of measurement and have redone some of the census statistics. The further back you go in time, the more inaccurate the statistics are, and the more the household and the place of business were the same. The farm is the most important example, but also small businesses. When you run a store, you often live in back of the store. So the woman of the household was probably working the cash register, taking care of the books, ordering things at the same time that she was taking care of the kids and home laundry. But at some point in the 20th century, certainly by 1950, we have much better estimates.

Then the question is, what explains the changes from then on? As I’ve pointed out in other work, it’s probably the case that we have more of a U-shaped female labor force participation. Early on it’s high, then it decreases, and then it increases again. And after the midpoint, maybe in the ’20s or the ’30s, somewhere in there, it begins to increase again. That’s if we include the nonmarket work. But I don’t mean here that I’m expanding the concept of gross national product to include household services such as taking care of the children, washing home laundry or cleaning your own house. This is just the fact that the households and the market are so intertwined that we’re doing a very bad job in figuring out how much labor is being supplied by all family members.

If we jump over that difficult question of measurement, and just think about supply and demand issues, the starting point that I always use is Jacob Mincer’s. In one of his most famous papers written in the 1960s, Mincer was trying to understand changes in female labor force participation from the 1940s to the 1960s.

His working hypothesis was that we have a supply function of female labor and we have a demand function, and the demand function is shifting across a relatively stable supply function. The demand shifts were much greater than the supply shifts. What is the information we have available to us to assess this? Well, we have the fact that the real wages of women were rising at the same time that female labor force participation was increasing.

Behind the labor supply function are several variables and parameters of which the substitution and income effects are among the most important for study. Individuals make decisions about whether the value of their time in some other activity—let’s say in home production or in leisure—is greater or less than their value in the market. And as this real wage rises, more and more are going to get pulled into the labor force.

Mincer realized that there is a little bit of a problem here because if the wages of women are rising, then it’s most likely that the wages of men are rising, and if that’s true, then we have to consider the fact that there’s also an income effect. That’s how he viewed this: That the increase in female labor force participation is going to depend largely on whether the substitution effect is greater than the income effect.

In the period he was looking at, voilà, the impact of the substitution effect was monumentally greater than the income effect, and therefore he could go a long way in explaining this increase without considering any of these enormous complications.

Now this was a simple and an elegant analysis that worked for the period Mincer was looking at. Yet it is not correct for many periods in U.S. history. One of the problems is that on the supply side we have both the income elasticity and the substitution elasticity, and these elasticities aren’t necessarily stable over time.

So, let’s go back to the simple supply and demand model and incorporate the complications and allow for these varying substitution and income elasticities. So then we have to say, okay, what do we put on the demand side and what do we put on the supply side?

On the demand side, we can think of sectoral change—whether there are increases in demand for manufactured industrial goods, which generally demand relatively more men than women, or services, which generally demand (relative to manufacturing) more women. There’s the possibility of declining discrimination against women.

On the supply side, there are factors such as fertility, household production technologies, all of the things that alter the reservation price. There’s also the possibility of social stigma regarding women in the labor force. So when jobs were dirty, when they were more brawny than brainy, when they were jobs we would consider to be jobs in The Jungle vs. the jobs of Dilbert, then the husbands of women who were in the labor market, would generally be viewed as lazy, slothful, and not very caring.

But when jobs began to switch as they did in the 1920s to white-collar work, when there was a huge increase in clerical employment and service-sector employment, then women who were in the labor force were indistinguishable, those who were from lower-income households and those who were from medium- or higher-income households. So the social stigma began to disappear and the social stigma is going to affect the analysis by a substantial amount.

Let’s consider changes in the income effect. If women have a sense that when they enter a marriage it will last forever, then the income that their husband brings home is their income, the wealth that they have is their wealth. If they’re in a home with a high-income earning husband, or if their wealth is very high, they can both take a lot of time off.

But if instead we’re in a world in which marriage is more uncertain, then wives will want to build up their human capital. Therefore the income effect might not be as strong in a world in which marriage is more uncertain or life itself is more uncertain—if men are dying in wars, for example.

My point quite simply is that we can go a long way by taking a very basic model in which we have a demand side and a supply side, and we have some sense of what’s happening to the substitution and income elasticities, and make some gross assumption about demand elasticity. Then we know what’s happening to quantities and wages, and we can back out of that whether supply or demand is doing more of the work, without even looking at the specific demand-side or supply-side shift factors. Then we can look back and say, well, now let’s examine these demand (or supply) side factors more.

The back-of-the-envelope calculations in my book long ago were that from 1890 to 1930 supply-side factors were most important, whereas from 1940 to 1960 demand-side factors were most important. Well, that was the period Mincer was looking at, when this demand curve was shifting out over a very elastic female labor supply function. In the earlier period when the income effect was much stronger you have a far steeper supply function. So you can send that demand curve out to Neptune and it’s not going to have a very great effect on quantities. Prices are going to change but quantities aren’t.

And from 1960 to 1980, supply factors and demand factors are sort of sharing the load. What about from 1980 to 2000, which I didn’t do because I finished the book [Understanding the Gender Gap: An Economic History of American Women] in the late 1980s. I think it’s probably the case that demand factors have regained more importance, but I must confess that I haven’t looked at it. The important thing is to understand what the substitution elasticity is and what the income elasticity is.

So those are my two long answers for your very good question.

Are Women Opting Out?

Region: Last October, the New York Times Magazine published an article, "The Opt-Out Revolution," describing a trend among highly educated, professionally accomplished women who leave the upper tiers of the labor force in order to raise their children. According to the Times, it generated "record-breaking mail."

Is this a trend you've observed in the data? If so, what does it say about women, reproductive choice and the labor market?

Goldin: That article came out exactly when I was giving a lecture at the University of Pennsylvania for a conference that was first called "Mommies on the Fast Track" and then "Mommies and Daddies on the Fast Track" because of the idea that everyone is on the fast track and the kids are being left behind. Participants at the conference were outraged by the article. Absolutely outraged because of the absence of facts. Rather than being outraged, I'll just treat the article as a set of real-life stories.

The point is that it's just not a dominant trend at this time. But it is the case that family and career trade-offs are more and more important for both men and women. The point that was missed in the barrage of letters that was sent to the New York Times over that article is that we simply don't know if these women are actually opting out. They're too young. They could just be slowing down for a while. A project that Larry Katz [Harvard economist] and I will soon be embarking on is to get better longitudinal information on career-oriented women, mainly women who graduated from selective universities, who went on to get advanced degrees. What have been their career trajectories? And we don't have data sets right now that give us that information, with some exceptions such as certain specialty data sets like the young physicians survey.

Consider the lives of Sandra Day O'Connor and Ruth Bader Ginsburg, who are among the most high-profile women in America today. I think that there were moments in the careers of each of them when someone could have said "oh, they're opting out," though we wouldn't have called it that then. But they didn't opt out. Perhaps each knew that they were going to continue at some point, but they were slowed down by either family forces or, in O'Connor's case, by the fact that she found it nearly impossible to get anyone to hire her.

Region: You've pointed out in your work that American women have always had trouble combining career and family, and that each cohort deals with it a bit differently. The Times writer was looking at a very specific cohort.

Goldin: Yes, and also these women are still very young. So just because they have a three-year-old and they're still not back in the law firm doesn't mean that they're going to have a 10-year-old and not be back doing something. They may not be corporate lawyers—they may be local civic leaders, perhaps better than lawyers, but they're still lawyers.

Labor Markets in the 20^th Century

Region: In other work, you describe the enormous gains in wages and leisure, and significant changes in the composition of the labor force that took place during the 20th century. But you argue that the vast majority of these changes can be attributed to "fundamental advances in technology" rather than to labor union activity or government intervention.

That's a surprising conclusion, given the dramatic scale of union activity in the 20th century and the major pieces of legislation that were passed in that century—immigrant labor restrictions, the Civil Rights Act, Social Security legislation come to mind. Could you explain?

Goldin: The reason I made that statement is that almost all of these changes began before the great rise of private-sector unions and federal government labor regulations. Real wages have increased for a very long time, long before the rise of union activity and in sectors in which there were no unions. Real wage increases occur because workers are more productive, so it could be technological change or it could be education. Educational advances concern government but generally not the federal government. Local governments together with parents have been responsible for those changes.

What about hours declines? Unions, as you know, were quite insistent about hours reductions. But hours declined in manufacturing from the early part of the 19th century to the midpart of the 20th century both per day and in terms of days per week. The big rise of unions in the United States was from the mid-1930s to the early 1950s. Thus the decline in hours preceded the rise of unions, although increased union activity in the late 1910s and in the 1930s may have furthered the decrease in hours and the increase in leisure.

Union activity in the United States in the nonagricultural sector hovered around 5 percent to 10 percent of the labor force until the mid-1930s when it increased greatly and in the '40s when it increased again by a large amount, reaching a peak of about 35 percent. In the mid-1950s it began its steady decline. That is, almost as soon as it reached its peak, it began to decline.

Region: Private sector?

Goldin: Yes, but even if we fold in the public sector, it holds back some of the enormous decreases in the private-sector union activity but not by very much. So today's level of union activity in the private sector is about what it was in the early 20^th century.

By the 1920s, hours are down to a much much lower level than in the mid-19th century. The next decrease was due to the Depression. And after that decrease, hours don't move much. They increase during World War II for good reason since people wanted to work and wages were high, and Americans didn't know if wages were going to stay high. At the end of World War II hours in the United States are lower than in Europe, but then European hours go down and ours do not to the same extent.

So the second point is—the first one is real wage increases precede lots of fetters—hours declines also precede lots of fetters. The beginning of retirement precedes Social Security by many, many decades. And the change in wage structure precedes unions.

Now this doesn't mean that certain groups were not affected more than others. So African-Americans were clearly affected more by government action than I'm portraying for the entire labor force. Older Americans were greatly affected by the antipoverty aspect of the Social Security Act, and old-age insurance more so than retirement was immediately affected by Social Security itself.

In the piece you cited, I did mention that one of the ironies of thinking about fetters imposed by the government is how unimportant many of them are, and also how important are some that we often don't consider when we think about the labor force and labor market regulations. The two that we often omit, yet which are among the most important, are immigration restriction and education.

The High School Movement

 REGION: Tell us about the “high school movement,” if you would. What led to it and what consequences did it have for the American economy?

GOLDIN: Education is a hierarchical process. Very few people go to college before they go to kindergarten. And it’s also a hierarchical process in its production in the sense that if you want teachers to teach in elementary school, you have to train them in normal or teaching schools. So when America was founded it almost immediately established institutions of higher education. Harvard was founded in 1636 and opened in 1638. And William and Mary was next and then we get many of the private institutions. And the nation’s oldest public institution of higher education is the University of Georgia, chartered in 1785 but which opened sometime after.

So because education is hierarchical, you’d expect that in terms of the masses of people that you would have elementary or what we common school or grammar school education diffusing across the population first, then the high school and then college.

Now in certain countries where access to education is limited for various reasons, in which the public sector is not very involved and in which education has to be done by the private sector, countries that are sparsely settled and so on, you might not get this type of rapid diffusion. And so it’s the rapid diffusion that we call the “movement.”

In America what we call the common school movement occurred in the early to mid-19th century. And the high school movement—by which I mean the movement of young people from not having a high school education, not having graduated from high school, to the median 18-year-old being a high school graduate—is dated from around 1910 to 1940.

And given that foundation of the mass of Americans having a high school education, we could then embark on the third transformation of American education which would be higher education, a transformation that we’re still going through.

REGION: You point out that's very different than the European experience?

GOLDIN: In certain respects it was different from the European experience and the ways in which it was different are mainly those above the common or elementary school level.

But the question is, what led to the high school movement? This was not just an urban movement. The movement included many parts of the country and extended from about 1910 to 1940. At the beginning of the period, only a small fraction of youth graduated from a school that we would consider to be a high school; whereas, at the end of the period, at the time of our entry into World War II, the median 18-year-old was a high school graduate.

The fraction graduating was a little bit higher for girls than for boys during much of the period, but by the ’40s the difference was small because of the Depression. In fact, the Depression catapulted some of the more industrial parts of America into the high school movement by withdrawing jobs for young people. Sometimes economic downturns are good for education. But the war was very bad for education because relative wages of the unskilled increased and kids in high school left in droves to go work.

This high school movement was well under way by 1900 in parts of the country like New England, Massachusetts in particular. Even in the 19th century, high schools were built in the towns of New England. But when the high school movement got under way in the nation, by the 1910s, it exploded in some areas that are even today very sparsely settled. Places like Iowa, Nebraska, Kansas, Oregon, Washington and California by and large. And so it jumped from New England to the central part of the country and out to the Western states. It looks sort of curious because it was not a movement only rooted in the older parts of America. It was not rooted only in places that we generally think of as wealthy and prosperous (although these places were rich in taxable property), and it was not rooted in places that were to be the receivers of population, although it did exist in California which gained enormously in population in the 20^th century.

So what ties all these places together? One thing is that all had relatively high wealth per capita that was taxable. Land in the prairie states was—in 1912, 1922 when we have data per capita—highly taxable wealth and it was extremely fertile, and thus valuable, agricultural land. Plus education is governed by very small communities, just about throughout America, and small communities responded to the desires of parents and civic leaders.

The Academy Movement

To understand why this happened, why it happened where it happened, one actually has to back up a bit and touch on something that we know only a little about. It’s something called the “academy movement.” Let’s say we’re in Illinois or upstate New York in the 1870s and in our area there’s an ungraded common school—maybe a graded grammar school if you’re living in the city—which allows us to go to 8th grade. But there’s clearly demand for clerks and other people with skills and knowledge: accounting, business, rudiments of science, people who are literate, who might know different languages. People who have more than an 8th grade education.

So what were parents to do? Well, in some parts of the country, common schools were educating kids beyond eighth grade because there was demand for it. But what could a teacher who’s teaching kids from 5 years old to 18 years old do with them? Throughout America there arose something which today we call the academy movement, but we know very little about it because it existed for such a brief moment in American history.

These academies were private, tuition based. Most were ephemeral institutions, save some, such as Phillips Exeter Academy. But others, such as (I’ll make it up) Miss Smith’s School for Proper Girls and Captain Joe’s Nautical Academy and Mary’s School of Clerical Skills, they came and went. Some were academic academies and others were vocational academies. There were hundreds all over America. And we have some information about them because the 1870 census and also 1860 and 1850 asked county agents for information about educational institutions of all types. But if you try to find other documents on the academies you’re hard pressed to do it. The data that we have on them are quite lacking.

They were sort of like the proprietary vocational institutions that we have today—like the ones that taught shorthand when I was a kid and teach computer skills today. They were fly-by-night. These were sometimes in the home of an educated person who took in students for, say, $15 a semester. Some of them might board there.

What does this demonstrate? It demonstrates beyond a shadow of a doubt that this was a grass-roots movement. This was not a top-down movement only. This was not a movement in which some New York school people, perhaps at Teachers College at Columbia University, were telling Americans what to do. These were Americans who district-by-district, school-by-school, family-by-family, decided what they wanted. This isn’t to say that there wasn’t any coordinating mechanism once the high school movement started to take off. This isn’t to say that the principals and supervisors weren’t meeting at state conventions and that the state leaders weren’t meeting at national conventions. We know that information traveled in that manner.

But Americans weren’t being duped into having their kids taught at this higher level and taught certain subjects at this higher level. What I’m arguing against is the educational literature from the far left, including various Marxist historians, and also that from the far right. Oddly enough, these two groups are saying the same thing. On the left they are claiming that the dastardly elites were dictating to the lower classes what they should learn so that they could hold them in line, so that they could Americanize immigrants, so that they could make them in their own mold.

And on the right they are claiming that the dastardly progressives, who may have had some potentially good ideas, rid our schools of standards and excellence and in its place put nonacademic subjects such as art, music, typing and shorthand. So these two groups are saying that all of what happened was top down, that people are like sheep who can be herded one way or another. They just pay their taxes and are totally apathetic about what their children are taught.

But I’m constantly coming across evidence to the contrary, evidence that is adding up to a sense that there was a grass-roots movement that spread across America. I have found statements of parents saying, “I want my children to learn about the wider world of business,” particularly after World War I which suddenly got Americans thinking internationally. And they would say, here we are in Nebraska. My kids aren’t going to stay in Nebraska. Farms are getting bigger, there’s no room for my kids to farm here. The real place for my kid is in Omaha, in the big city, maybe in another state farther West, and I want my child to have knowledge of this wider world.

REGION: So that demand leads to local interest in high school?

GOLDIN: Precisely, and it led them to say we want courses that are relevant. We don’t want our students to know only Latin as their foreign language. We want the live languages of Europe and the Western Hemisphere. And these communities throughout New England, the Midwest and particularly the prairie states, in tiny towns they built high schools. This was not—as I keep saying—this was not a movement that was contained to cities, it was not contained to places that were dense in population. These high schools often had 40 students, 60 students. And you might say, well, they weren’t real high schools. But they were real high schools. I’ve seen pictures of them. I went to Iowa and saw the actual buildings. Larry and I went to Ames, Iowa, to give a talk and the weekend before we gave the talk we took a tour of some tiny towns of south-central Iowa. I had a listing of the town populations in 1914 and what schools existed then and when they were built, and we went to some of the towns to see whether the buildings still existed. And by and large, they did with cornerstones proclaiming dates like 1910, 1911, 1912.

REGION: Few economists do that kind of field research.

GOLDIN: True. One spring vacation, we went to Nebraska, one we went to Kansas, another to Iowa and another to Wisconsin. Each of these places we went to for different reasons, but whenever people would say, where are you going, and we’d say, we’re going to Iowa or Kansas, we would get one question from everybody: Why? And we’d come back with these happy tales, and people would say, it’s really more interesting than I thought. Everyone else went to the Caribbean.

REGION: What consequences did the high school movement have for the American economy?

GOLDIN: Well, it helps to contrast it with other countries. Why was America different? In what ways was America different? And then we can ask what consequences it had. One way of thinking about how it was different is that during World War II both America and Britain made a promise to their fighting forces and to the types of families that were providing the fighting forces. In England, the Labor Party passed a bill in 1944 that assured all British children they could have a secondary school education free of charge. At the same time, the United States passed a bill that assured all GIs they could continue to college, because we had already set down the foundation of high school graduation.

Even by the mid-1950s, the fraction of 15- to 18-year-olds in academic secondary schools in Europe was trivial. In not one European country out of 18 was the fraction greater than 25 percent; it was almost 80 percent in the United States. Northern Europeans frequently continued from elementary schools to apprenticeship programs and technical institutes. Scandinavians did, as did Germans. The British did to a somewhat lesser extent. But even including these levels of education, the European quantities were considerably lower than the American—in no country was the fraction greater than 40 percent. I’m referring here to quantity rather than quality because the quality differences are harder to measure and also because one must first have quantity before one can talk about quality.

The high school movement was a typically American endeavor. It was mass education. Americans want mass everything. Mass production, for example. We have been concerned with a mass group. We are a democratic nation. Of course, there are some who were left out: blacks in the South for example, even many blacks in the North. But we were concerned about getting the masses through school. The quality of education is another important issue, since we’ve got thousands of school districts and had virtually no strict, uniform standards at the state level.

Now what difference did the high school movement make? Well, the consequences were that it gave America a mass of relatively skilled, literate, trained workers. It gave America a cadre of men and women who were poised to enter college in the post-World War II era. That was important.

At the same time that the high school movement was just beginning, we still had mass immigration. Most factories in America were being manned by not terribly literate adult males who were from other countries. One of the other impulses behind the high school movement (not in Nebraska and Iowa but in bigger cities—Chicago, Philadelphia, New York, Baltimore, places that were receiving immigrants) was to distinguish your kids from all these other workers. Well, one way of doing so is get them education that’s going to catapult them into something that is a commercial job, a banking job, a job in the new industries. We know from data we have from the 1940s that the “newer” industries had a much higher fraction of workers who were more highly educated even if they were blue-collar workers.

REGION: You’ve written that service attendants at gas stations, a new industry at the time, tended to be well-educated in that era.

GOLDIN: Yes, and they had their brains fried on leaded gas. It’s a good thing that they were well-educated because they must have lost a lot of IQ points!

The "Human Capital Century"

Region: You refer to the 20th century as the "human capital century" and explain that several "virtues" relative to our educational system—local control, public funding and gender equity, among others—led to that outcome. You've also suggested that these virtues may no longer have such virtuous outcomes in the future.

Goldin: Right. There are several virtues of the American system. One virtue of the human capital century has to do with the openness of the educational system, with its being forgiving. The American school system is infinitely forgiving compared with the European. In Germany or Britain in the mid-20th century, for example, children took exams that determined whether they could continue with academic training. If you were 11 years old and you didn't feel well the day you took the exam, too bad, because you're not going to go to the academic high school. You're going to become a carpenter. Americans have a forgiving system so that if a kid didn't do well in one grade, perhaps he'll do well in the next grade. If he didn't do well in one school, perhaps he'll do well in another school.

There's a lot of competition, too. There's great virtue to having competition. We have competition among school districts. We have competition among colleges within states. The University of California system is the best example of this competition. The state universities compete with each other. And layered on top of that is a private system of higher education.

The virtues of the American system have been its openness, forgivingness, competition, public funding, nonsectarian character—although not without God, it never has been without God—gender neutrality and its small, fiscally independent districts, which give rise to competition. It's important that school districts be fiscally independent. France has lots of small districts, but their role is to pick the teacher or textbook or sometimes not even that. But they're not fiscally independent, they're not deciding on taxes. They're not deciding on whether schools should be built or most other important decisions.

Each of these aspects of what I would call a virtuous system, a system based on egalitarianism, has been recently attacked and often for good reasons. So we generated a lot of competition in America with good outcomes, but in the process we left a lot of people behind. We were very forgiving, but we didn't lay down strict standards. If you lay down strict standards, you have to enforce them, which means that you can't be very forgiving. So what do you do with the groups who are left behind? Americans refused to have this other apprenticeship program that would accept all the kids who would have been left behind.

Most of the virtues have raised potential problems. I mentioned the virtue of openness and forgivingness vs. having strict standards, for example. I think that that's a difficult issue. We want to have equality of opportunity. The basis of egalitarianism is to be open and to have equality of opportunity. But at the same time, we want there to be greater equality of inputs. So we want to give children the ability to go to nice schools with good teachers, but we want also to foster competition. These tend to conflict because if you have small fiscally independent districts and have a lot of competition, some of the districts are going to be poorer than others.

And then we say, well, we want equality of opportunity. But does that also mean that we want equality of inputs? Well, we certainly don't want the great divergences that we see today and we've seen for quite some time. But this is the type of conflict that we may be able to resolve.

We want equality of opportunity, but how do you get there without distorting incentives? I think that one of the biggest questions we have to face is how to achieve our goal of being egalitarian and still keep the incentives in place. Some of my colleagues work on these issues and have interesting solutions. The issue about how to raise standards without actually leaving some children behind is more difficult because we have to figure out what to do with the students who fail.

Investing in Early Childhood

REGION: In a chapter written with Brad DeLong and Lawrence Katz for a recent Brookings Institution book, Agenda for the Nation, you argue that sustaining economic growth calls for increased investment in human capital, especially among the young from minority and low-income families, and you write that “human-capital policies must target early childhood.” Why do you emphasize it?

GOLDIN: We emphasize it because the United States has had a real slowdown in human capital growth as measured by school years. We’re good at measuring years. And the place where the slowdown has occurred most is in the lower part of the income distribution.

What do I mean by a slowdown? There’s a nice graph in that chapter that shows the completed school attainment of cohorts from the late 19th century up to the 1970s—we’re showing years completed as of age 35 and also projecting forward. Over that period, years of education go up until the cohorts that were born in the 1940s, and then it levels off. It’s just now beginning to increase again because college entry and years in college have taken off again.

So the total years of education of an average 35-year-old today aren’t very much different from those of their fathers and mothers who are about 25 years older, whereas for cohorts in every year before that, there were increases year after year. Every year was better than the previous year. We now have reached a plateau of sorts, what we call a slowdown in human capital growth. Europe, by the way, hasn’t had this slowdown because it is going through the process of huge increases in education that we once went through.

It isn’t that we’ve reached some global maximum beyond which we can’t go. It’s that there’s a greater problem in the lower part of the income distribution. The rest of the distribution is doing OK, but that lower part is doing very poorly in terms of kids completing high school and sending kids on to college and completing college.

The high school completion experience has really been dreadful, and part of the problem is another of these issues of virtues and vices. We thought after World War II that the great virtue of the American system was its enormous flexibility. If you didn’t complete high school and you were out in the world, you could take the GED exam, and if you passed it we would award you sort of an honorary high school diploma because in fact you learned everything you would have learned in high school.

In fact, high stakes testing is backfiring in certain places. There are now kids in high school who aren’t doing well and who think their chances of passing the high stakes test are pretty low. They decide that because they’re borderline and would have to put out a tremendous amount of effort to pass this test, why put out any effort? So they drop out and in three years they’ll say, “well, I’ll just go back and get my GED.” It’s this problem of openness, being forgiving and flexible. One of the ironies is that even when you have standards, it can cut into the standards. So you have this bifurcation now. You have certain kids who are doing better because of the standards, and then you have others who are not.

What many educational scholars have asked is: At what moment do interventions work? And for many kids in low-income communities in which their parents haven’t had a college education or even finished high school, in many (not all) immigrant communities, early interventions do work.

Much of this has to do with parenting—an area that we know little about, oddly enough. What makes for good parenting? Is it the number of times you take your child to the library? Is it the number of hours you allow your child to watch television, as a negative influence? Is it the number of books you have in the household? Is it the amount of time a child spends with mommy and daddy, or one of them? We actually don’t know what makes for good outcomes later on. But we do have some good evidence concerning early childhood interventions that show that they do a tremendous amount of good. We have real experiments.

However, they’re not going to do any good if you stop there. You might say if I have an additional dollar to spend, spend it on early childhood education, but not if you’re going to let everything go, because you need to go the distance. And I think that’s one of the areas that hopefully one of my colleagues will work on and come up with some good answers about: How, at this moment, would you allocate funds across these areas? Can we figure out where the high returns are? Do you get just phenomenally more bang for the buck in early childhood if you keep these other things at some level, or do you have to just raise everything by the same proportion to get the eventual return?

Current Projects

Region: What are you working on now?

Goldin: Several things at once. The project that I'm working on right now is something completely different, and I hope will end soon because I want to get back to the education, technology and inequality book and to a new project that Larry and I are doing on men and women who are at highly selective institutions. We're going to follow them over time using some retrospective information on career and family outcomes. We're very excited about this project.

But before I do that, and before I finish the book, there's something else I'm involved in. [Harvard economist] Ed Glaeser and I are running a conference on corruption and reform using American history as an example of a country that was rife with corruption and then went through a series of reforms of some type and emerged as a country which was not free of corruption, but not as filled with corruption.

And so we assembled a group of superb researchers a couple of years ago, we had a planning conference, then a preconference and in a couple of weeks, we're having the conference. Economic historians and economists are contributing papers on a particular period or area that they know about—the Progressive Era, the New Deal, the antebellum period—case studies of what corruption looked like in America and what reforms worked or didn't.

I decided last summer that no one was working on the Fourth Estate, the media, and that Ed and I and Matt Gentzkow should do that. We begin the paper with a quotation from [Thomas] Carlyle. He was paraphrasing [Edmund] Burke and said something like, "in the reporters' gallery of Parliament there sat the Fourth Estate more important than all others because he who can print can hold the government in check."

As circulation expanded for various reasons, including the reduction in the price of newsprint, the press became less beholden to political parties and less biased. We have a nice model of the decline in bias, a general history of the rise of the independent press in the United States and of circulation. But what we needed were case studies, so we decided to study the Crédit Mobilier scandal, which occurred during the Grant administration, and the Teapot Dome affair in the Harding and Coolidge administrations. We picked about 20 newspapers, got the microfilms of these papers, searched over the periods of time when we knew the breaking stories occurred, figured out who was reporting the breaking stories and when.

We have 1,200 articles and they're all in PDF files. I have more core in my machine devoted to these files than anything else! But all of this gets distilled to, like, six graphs and two tables. I wanted to learn about the history of the press in the United States, and this exploration has been an enormously interesting learning experience.

The Importance of Economic History

Region: What are your thoughts on the importance of economic history within the discipline of economics?

Goldin: I can speak with respect to people in my department and people at our neighboring school, Massachusetts Institute of Technology [MIT]; it pleases me enormously that, almost to a person, economic history is treated with great importance when done well.

Why are they so interested in economic history? It's not because it is some sort of quaint scholarship. When I was visiting at the Institute for Advanced Study at Princeton, a tale was frequently told that the historical studies group was founded there because the mathematicians thought it was quaint and interesting, not that it was useful. Here economic history is important in capital letters; it is useful.

Economic history is the continuation into the past of what people are interested in today. It's accepted here that economic history is the study of the long term, that every good scholar and certainly every grand scholar and meaningful scholar in economics under- stands that minor league papers are written about what happened yesterday or the day before or the year before, the decade before. Economic history links together the different periods, eras and years in a seamless piece of cloth. You can't separate one from the other; you do so at your own peril.

Some of the work that's being done here—and I don't want to mention names because I might inadvertently leave someone out—but there are people in both my department and at MIT who are working on, I would say, deeply historical questions about why countries differ and why certain countries are poorer than other countries, why certain countries today seem to do better under different circumstances, whether they're less corrupt or have less regulation or have a better ability to produce human capital. This isn't necessarily something that happened yesterday or even 50 years ago. This is something that was put in place some time ago, and to understand that you have to be historically aware. You have to deal with the long term. So what economic history is to me is simply the study of the long term. I am no different an economist than other economists. It's just that my emphasis is more on the long term.

Region: Thank you very much.