Romesh Vaitilingam interviews Lawrence Katz for Vox

January 2009

Transcription of an VoxEU audio interview [http://www.voxeu.org/index.php?q=node/3566

Romesh Vaitilingam: Welcome to Vox Talks, a series of audio interviews with leading economists from around the world. My name is Romesh Vaitilingam, and today's interview is with Professor Lawrence Katz of Harvard University. Larry and I met at the American Economic Association's Annual Meetings in San Francisco in January 2009 where we spoke about his new book, co-authored with Claudia Goldin called "The Race between Education and Technology."

I began by asking Larry what he meant by the race between education and technology.

Lawrence Katz: What we mean by it is that we think, to a substantial extent, that the wage structure, or wage inequality, is greatly affected by what we would call the demand for skills, or the demand for different types of workers, and the supply of skills or the supply of different types of workers.

The actual phrase here - ‘the race between education and technology’ - actually goes back to Jan Tinbergen, who shared the first Nobel Prize in economics, and his classic work in the 1970s trying to look at, imagining something: that the demand for skills is driven by changes in technology, which today we'd call skill-biased technological change, and the supply of skills is driven by changes in accessed education, and he argued you could explain a lot of variations in educational wage differentials and inequality by thinking of this race between demand and supply, or between technology and education.

We have found that that's a very powerful framework. Our book focuses on the long-run history of wage inequality, educational attainment, and technological change in the US, and thinking of these things in simple economic supply and demand, but trying to get inside the black box of the way technological change works and how accessed education works take you a substantial distance towards explaining why, for example, in the first half of the 20th century there was a dramatic narrowing of wage inequality, a narrowing of educational wage differentials when the US did a tremendous job of leading the world in expanding access to education in a very decentralized manner, creating almost universal access to high school when much of the world was not yet fully through common, or elementary, or grade school.

And then how rapid educational expansion combined with rapid technological change meant a fairly stable wage structure and shared prosperity from World War II to the early 70s, and how the period since we've continued to see rapid technological change increasing the demand for skills, but we have not seen the rate of growth of educational attainment in the US keep up either with technology or with its past patterns and then trying to understand.

One of the key points of our book is: lots of people argue that things like computerization, international trade, outsourcing have been driving forces towards increased wage inequality, saying the United States and in many European countries in the last 25 years, and we don't dispute that those have been important factors. But they're only one half of the story.

They generate a demand for skills. One has to also think about what's happening on the supply side. The striking thing about the U. is that the demand for skills increase from these things isn't that different. If you think of the early 20th century and electrification, electricity was as big a change in reorganization of work in the early 20th century as computerization…

Romesh: OK, so the internet is not something special like that.

Lawrence: It's very interesting, but it's not completely different than many other previous technological changes that led to dramatic changes in the scale of enterprises - the lack of the need for labourers to haul things around now that you have assembly lines dramatically change the demand for skills, increasing the need for a white collar workforce. Then the big difference was not college, but was getting a high-school degree to become a white collar worker or more skilled blue collar worker.

In that period we have measures looking, for example, at the US manufacturing sector that indicate increasing for skill was about as rapid, say, in the 1920s, 10s, as it was in the 80s, 90s and 2000, yet you have narrowing wage inequality in that period, much of it, rather than widening. And the big difference is rapid access in what Paul Douglas, a great labour economist in the 1920s and teens called the decline of non-competing groups by creating greater access to education. In the U.S. it was creating universal access to high school.

We've expanded access to college in recent period, but nothing like we did early in the century. For people born from 1875 to, say, 1950 in the US, every generation, basically, had two more years of schooling than their parents.

People today are more educated than their parents. A 25-year-old today, but more like a half a year. So education continues to expand, but not nearly as rapid as the demand for skills. And you actually see that about 60% of the growth of US wage inequality in the last 25 years can actually, just in an accounting sense, be explained by the increased growing gap in wages between those with more education and less education.

Putting those into an economic framework and then trying to get inside the black box of what things drive technology and the demand for skills, and what things affect the supply of skills, have been key.

The other key point is educational attainment, there are several factors that affect those supply of skills. One is access to education, the cost and availability of it, which affects education of each cohort. The second, obviously, is cohort size. If, on average, young people are more educated than older people, when there's a big cohort, that will have a bigger impact. That's why skills are smaller.

And the third is immigration, which changes the skill mix. What we find is that this slow-down in educational attainment in the U.S. is not… immigration is not…a lot of people have blamed immigration. It's definitely not the driving force. In fact, when you properly aggregate things, immigrants to the US are incredibly highly educated relative to the US natives at the top end, and less educated at the bottom. When you aggregate them up into the main skill groups, college and non-college, they're pretty much neutral, and not an important component.

The second is that cohort size did play some role. The 1970s had a very big growth of the educated because there was such a big baby boom. There's then a baby bust. But that's second order relative to accessed education, the growing cost of college, problems in K-12 schools in the US, problems of inner-city families.

That's where the slowdown is. We could have had the same sort of technological change and advantages with a much more egalitarian outcome.

The one final point is we are not unicausal people that say this is everything. Clearly the interaction of the race between education and technology is a driving force, but it's not in a vacuum. You have to understand institutions to understand, you know, in a place with strong minimum wages and strong unions, the same supply and demand shifts can have very different impacts on wage structure or on quantities employed.

Institutions that affect people's incentives and ability to get education drive those things. So it's not reductionist that there's just this sort of ongoing exogenous technology and supply changes. One really has to think deeply about things like labour market institutions and educational institutions as well to understand these changes in technology.

We don't deny they're important, but we think you have to have this interaction between institutions in the race between technology and education. When we look at Europe, there wasn't as big a growth in inequality for, we think, two reasons. The technological change is very similar because educational access continued expanding, catching up and now passing the US.

25 years ago young Americans were by far the most educated of any group on the planet. Today the US is near the bottom of the OECD in high school graduation rates, and sort of in the middle in college graduation rates. Clearly, given we were the top in both 25 years ago, we've had slower growth.

The second is the same supply and demand changes show up much more on wages in the US, where they're much freer and more flexible to change, than in places with stronger government and union intervention in wage setting.

Romesh: Do you think in policy terms in response to inequality, there's going to be a change now with the new administration and the fallout from the financial crisis and the economic recession?

Lawrence: This election clearly showed there have been some changes in attitudes. One part that's not the focus of our book, but is clearly important, part of the growth of inequality, which is the top 0.1 percent, which has been a large source of the growth of inequality in the US over the last 25 years.

In particular, there're going to be some - when the 2008 numbers come out - some substantial declines in incomes there, even without any policy of responses. But I do think some of that, because of the current financial crisis, things like adding progressivity to the income tax system is probably going to be delayed. But I think over the long haul, certainly, president-elect Obama campaigned on wanting to at least restore to the Clinton administration level, the degree of progressivity of the tax system.

I think, in fact, the current need for a major economic recovery package could be a down payment on some important investments in things such as early childhood education expanding, which will be very important in the US, creating greater access to college, such as the US has had dramatic increases in cost of college. We do have financial aid programs, but they haven't kept up.

The share that states pay, for example, for public universities in the US, relative to tuition, has fallen dramatically in the last 25 years. The state budgets have run into problems. Federal financial aid has not kept up with the cost of college. That, by itself, is not the be-all and end-all. A big problem in the US is many students who aren't ready for college at all. But there are a fair fraction who have a tough time paying for it, and have to work. We see very low college graduation rates because of that.

So I think there should be large changes, a start on that, in the upcoming economic recovery package. Clearly, more generally there will be some attempts to offset some of the declines in demand for, say, less educated males, in particular, in things like construction and manufacturing, as well as women in some of the service sectors and retail that have done poorly.

So there will be some effects through infrastructure, and other investments on demand side. Some to grow the supply. I think the broader changes and progressive tax system are likely to be delayed because of the current crisis. But I think there is no doubt that one could go a bit more progressive on the US tax system at the top end, and have a fairly substantial impact.

The other thing which actually relates to some other work I've been doing is, in this period of growing inequality, there's also been a dramatic shift of talented workers into the financial sector. Which, a couple of years ago was generating something like 15, 16% of all the labour compensation in the US, which is just an enormous amount.

I do believe there's a lot of value added in things the finance sector do, but it's hard to imagine an economy where one-sixth of all labour compensation is to people working in the financial sector. I think there's going to be a big reorganization. In recent years something like 40% of all the graduates of places like Harvard and Princeton who go and take immediate jobs were moving into finance and consulting.

That's certainly not going to be true of this upcoming year. I think there is an opportunity to start reallocating talent to things that may be long-run more valuable, like science and biotechnology and research. In the US there's a tremendous opportunity to create a new cadre of very talented teachers.

In the Great Depression, for example, there was something we found in a book about a 60% increase in young people going into teachers’ colleges. I think teaching, being a sort of reasonably stable job, might be very attractive to a number of talented kids.

Linking some money in a stimulus recovery package to state and local governments to have their school districts be willing to hire people with less traditional credentials for teaching could generate a new cadre of wonderfully talented individuals moving into K-12 teaching, which would be a crucial area for improvements in the US education and human capital in the long run.

Romesh: Larry Katz, thank you very much.