Sustainable Growth: Robert J. Gordon

Back to Previous Page

ASSA/AEA Annual Meetings 2012:
Panel Discussion:


Chicago, IL
January, 8 2012


Robert J. Gordon:

Thank you. I’m so glad to be here. I’ve thought a lot about economic growth, a very little about sustainable growth, so I had to look up the definition. The definition that I obtained I’ll quote you in a minute, but basically is that growth must be harnessed to the capability of natural resources to accommodate it. That is, the growth is the galloping horse, and we have to reign it in .

But my view is the US has the opposite problem. Our growth is slowing down, inexorably, for a number of reasons. And it’s my purpose here to tell you what some of them are. And we need to spur that horse. We need to get the growth going, not harness it or restrain it. And I’m not talking about the lack of growth in the context of the financial crisis; I’m talking about the way the world appeared as of 2007, before we had any crisis at all.

Viewed from 2007, the US faced two sets of fundamental problems: One is that our historical record of growth in real GDP per capita of 2% a year-- The actual number which my undergraduates all learn how to calculate with their calculators using natural logs, is that for 1929 to 2007, real GDP per capita grew at exactly 2.18%--more than 2%. But much of this rode on the back of the great inventions of the late 19th century which were fully exploited by 1970 and many of which could only happen once.

Even if we pretend that innovation is not going to slow down, which I know is a thought anathema to some of my friends who believe that innovation is everywhere, even if innovation proceeds apace, the US economy faces six headwinds, only one of which has to do with our current financial crisis, and the other five were evident and could have been analyzed, and were analyzed, in 2007.

The entire idea of sustainable growth is a clash between economic growth that has an inexorable momentum of its own and needs to be harnessed to avoid exhausting resources. Instead, in red over here, growth is faltering on its own. It needs to be spurred, not harnessed back. So here’s our perspective. Now this is an unusual chart. It may be the longest run-time series chart you’ve ever seen. It goes back to the year 1300, and it shows you the rate of advance of the frontier. That is, the blue line up until 1900 shows you the growth rate of per capital income over long intervals in the UK, and then we switch from blue to red, and we see what the growth rate of per capita income is in the United States, has been, in intervals that are designed to span over business cycles and to connect years which we agree were roughly normal economic activity.

That last down-spike on the right step of the red line represents my forecast for the next 20 years, from 2007 to 2027, and that was made back in 2007. And unfortunately, at the moment, we are currently 8% below that path. So if you think that’s illegitimate, that’s a forecast, I shouldn’t include that in the record, in fact it’s too optimistic from today’s point of view unless there’s a growth spurt in the economy that nobody is predicting.

Now just a brief couple of words about the causes of that enormous, what I call the big wave, the acceleration of economic growth and then the steady slow-down, which I predict is going to continue or even get worse. I’m giving a talk tomorrow night for EPS, and I’ll talk a lot more about the underpinnings of this process. But just to summarize it briefly: We had three great industrial revolutions. We had the first industrial revolution of steam and railroads that started around 1770 [1870?]. It occurred mainly in the UK and then helped to transform the American economy. We’ll see in a minute how much progress we’d already made by 1870. Then we had the Great Industrial Revolution, the second one, with all those inventions that were concentrated between 1870 and 1900. And they were simply more important for human welfare than those of the third industrial revolution, the electronics revolution, which is not what happened in 1995 with the invention of the Internet; it’s the transformation of human effort by computers that began in 1960.

If we think of all the ways that the third industrial revolution, contributions of electronics to replace human labor, have happened already, many of them 20, 30, 40 years ago: We had in the ‘60s conversion from dull, clerical, repetitive typing to computerized bank statements and telephone bills. We had the first credit cards. My American Express card is still stamped 1968. We had industrial robots. The very first one was introduced by General Motors in 1961. We had memory typewriters that eliminated repetitive retyping of legal briefs and even academic papers. Then we moved from memory typewriters to personal computers, to word processing, and way back in the 1980s we had—in fact in the 1970s—we had the ATM machines that helped to replace many bank tellers, bar code scanning that helped to make retail checkouts more efficient, and then our libraries in universities and public libraries were completely transformed in the 1990s by replacing card catalogues by flat screens that had all the information electronically. Think about how much that increased the efficiency of those people at the auto parts department [at] your car dealer that were constantly every day having to replace pages in their old enormous printed catalogues.

But how important was that? If you look at these three lines, we have a red line which is multi-[?] or total [?] productivity growth from 1948 until now. Then we have a blue line which shows you the trend—the left scale is logs. So this shows you the trend from 1948 to ’73, and in my interpretation, that period, the golden age after World War II, was the final set of applications of the great inventions of the late 19th century; for instance, the interstate highway system. Then we have the green line, which is the 1973 to ’95 trend. If you compare the blue trend and the green trends, the difference in logs is 58%. And our actual position, with all of that benefit of computers in the late 1990s, followed by, as you can see, an evident slow-down after that, we’ve only made up 11% out of the 58% gap by which the shortfall can be measured. That’s because many of the inventions were one time only.

We could do the following things only once—and there’s a huge long list. We could only replace the horse once with the motor car and motor truck. You know that a horse cost as much to maintain every year as its capital cost, the cost of buying it. Imagine if your $30,000 car cost $30,000 every year to fuel and to repair. That’s an easy way to remember the enormous efficiency gain of replacing the horse--not to remember the 10 to 15 tons daily of manure distributed on urban streets per square mile in 1880.

We replaced the backbreaking labor of housewives who no longer had to carry water into the house or carry coal or wood. We invented running water and consumer appliances. We learned gradually to make the transition between the only form of changing the temperature inside the house, which was an open-hearth fireplace in 1870, to 72 year-round temperatures thanks to central heating and air conditioning. We learned to raise the speed of transportation from the speed of a horse in 1800 to 550 miles per hour on a jet plane, a Boeing 707, in 1958, more than 60 years ago, and we’re not flying any faster now than we did then.

Okay, so, we had a lot of progress, a lot of inventions. Let’s pretend that invention and innovation is going to proceed at the same pace as it has over the last 20 years, which I obviously doubt. The US economy faces six headwinds. That forecast you say with the last rightward step on the graph incorporates the first two of them. These are very uncontroversial. First of all, the demographic dividend is reversed. The reason our income per capita could grow as rapidly as it did between 1970 and 1995, despite dismal slow productivity growth, as you saw by that green [?] line, is because we had an increase in the number of hours of marketed work relative to the population thanks to the entry of women into the labor force and thanks to the arrival of the Baby Boom teenagers, who became adults and raised the number of people working compared to the total population.

But now of course the Baby Boomers are going to be retiring over the next 20 years, and so hours per work will grow more slowly than the population. That’s completely uncontroversial; it’s incorporated in everybody’s long-term growth forecast.

Then if you read particularly Claudia Golden and Larry Katz’s great book of two years ago, we have reached a plateau of educational achievement in the United States—not in most other places. We’re suffering at the college level, the secondary level, and the elementary school level for different reasons. At the college level, we have a cost disease in higher education leading to mounting student debt that distorts life choices and prevents able people from going to college, because they don’t want to saddle themselves with that much debt. At the secondary level, in international OECD-run tests of 37 countries, we ranked 21st in reading, 31st in math, and 33 in science, out of 37, including such countries as Mexico and Korea. We have a continuing achievement gap of black and Hispanic students. It isn’t getting worse, but it isn’t getting better. But the percentage of Hispanic students in our secondary schools is continually rising, and it’s driving down the national average.

But that’s just the beginning. That’s already incorporated in that 1.4% future growth. What about four additional headwinds?

In this group you would certainly accept the first one: inequality. Growth in median income is much slower than the statistical averages that we all look at on growth in average income per capita. Here are the facts from Emmanuel Saez’s website: Between 1993 and 2008, before the financial crisis really kicked in the growth in average real household income was 1.3% per year. Notice that’s already a fairly slow number. Growth in the bottom 99%--does this sound like Occupy Wall Street?—was 0.75%. So there is a gap there between the bottom 99 and the total of .55% per year. And I’m going to have to deduct that from my projection of 1.4 in order to give you some forecast of the future growth of the standard of living for the vast majority of our population. The top 1% gathered, earned, or captured, is a better word, 52% of income gains during that 15-year period.

Number four: Globalization linked with IT. That hurts the leading nation more than others. It creates a convergence process where IT and outsourcing and the availability of modern technology spreading around the world, and modern electronic appliances, allow radiologists in India to read American electronic X-rays, and lead to a convergence, which means US wages grow slower, and Indian and Chinese wages grow faster.

The fifth one is the environment. This is partly going to, to the extent that we deal with global warming by social measures such as increase in carbon taxes, that’s a payback for past growth. It’s also a burden on the United States to have to partially pay for the global warming that’s created by the rapid growth of the emerging nations, like China and India. The Chinese economy spews out more carbon than we do, and yet we’re being asked to pay, I think, more than our fair share.

Contrast that with 1901, when the United States was able to push growth forward, the environment be damned. Nobody cared about that. All the pictures of economic progress included huge black smoke coming out of industrial smokestacks.

And then finally something that does relate to the post-2007 period: We have an overhang of consumer and government debt. And any solution to that means that growth and consumption and disposable personal income is going to be slower in the future than actual personal income and GDP, because obviously we can only fix it by some combination of raising taxes and reducing transfers.

Now here, today, I’m going to try to propose a few quick solutions. And there’s my controversy that I’ll add, because I’m sure some of the people here will not agree with many or all of them.

First of all, to deal with the reversal of the demographic dividend, we’ve somehow got to raise the ration of working-age population to retired population. There’s an easy way. Everybody agrees about this. We need to completely eliminate any kind of limits on HB1 visas for high-skilled immigration. There was an idea that Steve Jobs proposed to President Obama right before he died at a dinner in Silicon Valley. That’s incorrect there. It should be “Staple a green card to every degree of any foreign citizen who completes a Masters or PhD in math, science, or engineering.” And I would broaden those subjects a little bit further. Anybody who wants to should be able to buy their way in to the United States. Canada has followed that policy for the last 30 years, and indeed Vancouver is more prosperous due to the influx of rich people from Hong Kong. We should eliminate visas for tourists for most countries. Why should a Chinese person who wants to come and visit Hawaii have to wait eleven months for an interview for a visa, much less the visa itself.

It’s harder and more controversial to deal with low-skilled immigration. I’m all for eliminating the heartless deportation and terrorization of illegal immigrants. I want to open a road to citizenship. And I think the overriding principle here is they add to society; they create businesses and, properly administered, it will not be a burden to society, partly because on average they’re younger than our current population.

Education: At the higher education level, it’s easy to think of ways of improving our student loan system, which is get the private sector out of it, have the government directly in control, make many loans income-contingent, which we were all talking about back in graduate school in the 1960s, and somehow that never happened. What to do about the higher education cost disease is a long discussion, and I don’t have time for it here.

For elementary and secondary schools, stop demonizing teachers. The problem starts inside the family at age zero. Listen to Jim Hechman. Operation Head Start is too small, and it’s not big enough. We need to get inside the low-income families at age six months or less to start offsetting the handicaps that children have there.

There was a great story in The Chicago Tribune on Wednesday. They did a very detailed survey of the behavior of classroom of working class versus middle class children. Middle class children feel entitled. When they don’t understand something, they ask the teacher. They’re constant asking the teacher and expect to be learning. The working class children sit there like stones, don’t ask. That’s something we have to change, and it’s not the fault of the teacher. Elementary and secondary school: Create more prizes to change the culture about sports versus math and science. Deemphasize sports. Replace soccer moms with math moms and dads—I emphasize dads.

Inequality—Most high incomes are rents. The baseball player making $10 million a year has his next best opportunity--if he couldn’t be a baseball player he could maybe drive a cab, or make maybe $100,000 a year. So a lesson to Henry George of 1899: Let’s tax rents. The supply is inelastic. But that guy as a baseball player, his supply is won. If we had a top income tax rate on baseball players and everybody else making more than a certain threshold of say 50% of 35%, he’s still going to be a baseball player. Labor supply is unchanged. All of these fantasies about economic growth being hindered just would not come true.

Raise tax rates on capital gains and dividends to the top bracket income tax rate and cure the Warren Buffet problem that he pays lower taxes than his secretary. Tax reform is a long issue. I listened to a very good panel on it at the AEA today. But I think a simple way of saying where we need to go is have a progressive consumption tax with a top tier income tax on top of it because rich people get a lot of power even if they don’t consume. And eliminate many tax expenditures, but slowly, not suddenly, to avoid effects on asset markets.

The hard one to deal with is number four, because wages around the world are inexorably converging; there’s no direct solution without protectionism, and that creates as much harm as good. So I think what we should do is try to make the US more like other countries to prevent the kind of distortions like US auto makers couldn’t compete with the Japanese because they were saddled with all these legacy costs and medical care. We need to have single payer insurance from the government, not have medical care tied to employment.

Finally, on the environment, every econ 101 student learns that subsidies create dead weight losses. When you can clearly identify an externality, like with public transport, eliminating or easing externalities and congestion, then public transport subsidies are okay, indeed essentially. But learn from Solyndra: Stop subsidizing solar and wind power and ethanol and electric cars. Fortunately, the congressional subsidies on ethanol were allowed to expire through our political gridlock. Murphy’s Law is perfectly exemplified by ethanol. What it’s done is raise the price of corn around the world and help to lower the standard of living of poor people in less developed countries. Combat global warming by encouraging conversion from coal and oil to natural gas and nuclear. Introduce carbon taxes gradually, not all at once, and let price incentives work. Gas prices in Europe in higher, and they really do drive smaller cars.

For the consumer debt overhang, a big disappointment of the Obama administration has been that they put all this money and thought into bailing out the big banks; but they did not put equal thought into restructuring mortgages and rules to stop foreclosures and force big financial institutions to take losses in repayment for being bailed out.

For the government—this is my last point—let’s listen to the balanced budget multiplier principle. We’ve got multipliers like food stamps that are, say, 1.6, and that comes from Alan Blinder and Mark [Zandy?]. We’ve got other multipliers at 0.4, like taxes on the rich. We can have a revenue-neutral shift from taxes on the rich—you know which party I’m [?], I don’t have to tell you—to more generous unemployment compensation or other redistributional activity without any effect on the budget. Entitlement reform—we all know that’s easy for Social Security. You have some mix of changing the indexation formula, gradually upping retirement ages to keep up with life expectancy, and some adjustment of benefits and the way wages are defined.

For medical care, forget Obama Care, forget the whole thing; we gotta go back to scratch. You have to eliminate through government incentives fee-per-service individual practice. You have to make medical coverage a right of citizenship, and end its tie to employment completely so that when people become unemployed, it’s bad enough; they shouldn’t lose their medical care coverage along the way. And you have to eliminate private insurance overhead. The last estimate I saw was $380 billion a year of the cost of private per-profit medical insurance overhead and bureaucracy, together with the extra employees that doctors have to hire to fight with the medical care bureaucracy.

I hope that’s enough controversial ideas for 15 minutes.


Economists for Peace and Security