Benjamin Friedman headshot in front of bookcaseOctober 2015. GrowthPolicy staff members Devjani Roy and Scott Leland recently interviewed Harvard Professor Benjamin Friedman, focusing on the three key questions which motivate the GrowthPolicy website. Below is an edited version of Professor Friedman's comments. Click here for more interviews like this one.


Where will the jobs come from? | What should we do about inequality? | How to prevent the next financial crisis?

SECTION 1: Where will the jobs come from?

Key factors affecting jobs:

1. Technology displacing human labor

2. Offshoring

3. Low-paid immigrant labor replacing domestic workers

Predictions for the future:

1. Offshoring of jobs will accelerate.

2. Large supply of low-skilled immigrant labor will continue to push domestic wages downward.

3. Unlike in the past, greater education and training won’t be adequate to improve weak labor market.

For Americans, finding decent job opportunities in the coming two decades is going to be a great challenge. I’m not talking about the recovery from the most recent business recession, but a longer horizon. There are three reasons for being pessimistic. One is technology: More and more of what we used to do with humans doesn’t take humans anymore. This trend of technology supplanting human labor input is not only going to continue but will also accelerate sharply. I’m confident that, in the fullness of time, we will come up with new products, new services, new wants, and, therefore, new things for humans to do, but that’s going to take a while. Second is the offshoring of jobs. Many jobs don’t require U.S. resident labor in order to produce goods or even services that are consumed by U.S. consumers. It’s not just a matter of call centers and low-end labor. There are all sorts of things like radiology, legal services, and accounting work. Things that used to be considered not just middle but upper-middle and even high-end labor can now be performed remotely and, therefore, outside the United States. I think it’s clear that this trend also is going to accelerate. We already have U.S.-trained radiologists in other parts of the world reading U.S.-taken X-rays. But an increasing array of medical services is now performed remotely. In many parts of the country, we’re even starting to move in the direction of remote medical diagnosis by doctors who are not in the same room or even the same town as their patients. But once we do that, the doctors don’t have to be in the same country either. These trends are only going to continue and accelerate. The third piece is that even for many of the kinds of labor that have to be performed on-site, increasingly it isn’t U.S. labor that’s doing the job anyway—low-end jobs are being taken by low-skilled immigrants. This impacts wages. It’s not that the people who are running the landscaping companies here in Cambridge, for example, are discriminating against Americans. The wages they pay are too low to attract American workers. If the immigrant labor weren’t there, lots of people would still have their lawns mowed and their trees cut and wouldn’t do it themselves. But the wage would then rise to whatever point would be necessary to attract American labor. That’s not the situation we have. If we put together these three factors: the technology that supplants labor altogether; technology that allows even labor-intensive tasks to be performed remotely; and the presence of immigrant labor for many of the jobs that have to be done on-site, I’m very pessimistic about where we’re going to generate the jobs. This is not an unfamiliar thought in economics. It goes back hundreds of years. Over time, technology has always generated new products and activities that allow the economy, empirically the service sector, to absorb the freed-up labor. I believe that in time that will happen again. But I think there’s going to be a very difficult transition. We will have a weak labor market for quite some time to come. Unfortunately, the usual economist’s mantra of more training, more education, just isn’t going to solve the problem.

SECTION 2: What should we do about inequality?

Reasons for growing inequality:

1. Weak economic returns to labor versus capital, especially in last two decades.

2. Taxation policies more favorable towards capital than labor.

3. Widening wage gap.

4. Internationalization of labor.

5. Technological changes.

6. Immigration policy tilted toward low-skilled workers.


1. Change taxation policies in favor of labor versus capital.

2. Raise the minimum wage.

3. Shift immigration policy to a focus on encouraging high-skilled immigrants.

4. Reform educational system, starting with pre-school education.

5. Provide firms with incentives to train their own workers.

There are two dimensions to the inequality problem in the U.S. One is the weak demand for labor and, therefore, weak economic returns to labor as a whole. As a result, the returns to our incremental production aren’t accruing to labor. They’re mostly accruing to the providers of other factors of production, and that would mostly mean capital. This is the main thrust of the book by Thomas Pikkety (2013) that got so much attention recently (Footnote 1). This is a relatively new phenomenon. It’s only within the last twenty years in the U.S. that the division of the total product between returns to labor and returns to capital has started to shift. For many years, the division was largely constant and for that reason, the most typical models we use in economics are ones that imply a fixed division of the returns to labor and the returns to capital. Economists have been using these mathematical models for well over half a century going back to the 1940s. You might ask, “Why did nobody challenge that?” The answer is that every time people would try to mount such a challenge, real-world data demonstrated that the ratio was pretty constant. But in the last twenty years, it isn’t constant anymore. It’s moving in the direction of more of the returns going to capital. I think that’s going to continue. Now second, in addition, and compounding this problem, are the widening wage differentials even within the category of returns paid to labor. Economists have offered a whole host of hypotheses to explain why returns paid to labor have been becoming more unequal not just over the past twenty years but now over almost fifty years. These hypotheses are almost all correct. They’re not equally responsible for the outcome, but they’re not mutually exclusive and they all play a role in growing inequality. One hypothesis is about changing technology and, therefore, the increasing rewards that the labor market provides for some jobs and the shrinking rewards to others, depending on where technologically driven demand is increasing and where demand is falling. A second is internationalization. More and more American labor is competitive with labor abroad, including large numbers of people especially in China and India who are being brought into the modern economy in ways that they weren’t before, in a way that is competitive with our labor. A third is immigration. We have an immigration policy in the United States that is skewed toward low-skilled workers. For that reason, the inflows of low-skilled workers widen the wage inequalities. If our immigration policy were skewed toward higher-skilled workers, immigration would narrow the wage inequalities, but that’s not the immigration system we currently have. Then there are other miscellaneous reasons as well, including matters of corporate governance (why are the CEOs paid so much more than they used to be?), the erosion of union power, and the erosion of the minimum wage in inflation-adjusted terms. Now as to what to do about inequality, there are lots of potential ways of addressing inequality at the pre-tax level. Let’s first start with wage inequality and then back up to the labor versus capital inequality. Depending upon what we think is the most important explanation for why wage inequality is widening, there are then policy implications that follow. If one attaches a lot of importance to the shrinking real value of the minimum wage, the obvious answer is to raise the minimum wage. If one attaches a lot of importance to the shrinking role of unions in our labor force, the obvious answer is to take steps to enable the union sector to rebuild itself. There are also ways of shifting the skill bias in our immigration system, or at least neutralizing the bias we now have in the direction of low-skilled immigrants. If one thinks a lot of the problem is due to changing technology increasing the demand for some kinds of labor and reducing the demand for others, the obvious implication is for a reform of the education system. I mean not just the Harvards of the world but more so the secondary schools, and even before that, all the way down to pre-school. We can also provide incentives for firms to do their own training of the labor that they’ve got on the payroll. You might ask, “Why do firms need incentives to train their own labor?” The answer is that if a firm trains a worker, the next week the worker might go off and work for somebody else. So by training a worker who is able to go work for somebody else, the firm is providing a quasi-public service. Then there is the matter of post-tax and post-benefits inequality. The tax system plays some role, but the benefits system is important as well because workers at the low end of the wage scale receive subsidized medical care, subsidized food, subsidized housing, and so on. Studies show, not surprisingly, that the inequality in what people consume has increased much less rapidly than the inequality of what people earn. As for the changing distribution between labor and capital, it’s hard to know how to affect that division at the pre-tax level. There, what we’re really talking about is again taxation. But now, in addition to merely taxing high earners versus low earners, what matters is taxation of labor versus capital. We have many favorable forms of taxation for capital income today. Capital gains are taxed on a more favorable basis than ordinary labor income. We also have all sorts of depletion allowances and special exemptions. Of course there’s a purpose for this. These tax policies were all meant to encourage physical capital formation, based on the theory that having more physical capital enhances productivity, raises wages, and, therefore, raises incomes for more than just the owners of the capital. But if that turns out not to be true, then it’s not obvious why it’s in the public interest to have these policies in place.

SECTION 3: How should we prevent the next financial crisis?


1. Capital reform for leveraged financial institutions, starting with regulations for greater capital requirements.

2. Require contingent liabilities, subject to determination by an independent regulatory body deciding whether or not the financial institution is undercapitalized.

3. Accounting reform, including well-defined stipulations for calculating the capital ratio of a financial institution.

4. Segmentation and shrinking of the banking system to moving away from the “too big too fail” subsidy.

I would impose much more onerous capital requirements on leveraged financial institutions. They ought to be required to have much more capital than they do now, as under the proposed Basel standards (Footnote 2). By the end of this decade, banks will be forced to have more capital if the Basel standards go through. But I would go further than that. I would also have a significant part of those additional capital standards be in the form of contingent liabilities—meaning, debt obligations that are convertible into equity—in a circumstance not as we have now, in which the owner has the option to convert, but where some independent entity like a regulator decides that the firm is undercapitalized. This will presumably make it more expensive for many of these institutions to do business, both because they will have to carry more capital and because they will have to pay more for these contingent liabilities. In effect, they will have to buy, from the market, insurance that they now get free from the taxpayers. This will cause some contraction in the banking system, which I think will be a good thing. Capital requirement reform is also about accounting reform. Consider the case of Citibank. The reason Citibank went broke in the recent crisis is not that the capital ratio being assessed was too small, but that what was supposed to be included in the ratio wasn’t appropriately defined. The issue is as much about what’s in and what’s out as it is about picking the right number. Second, I would move much further than the new Volcker rule in forcing banks to separate their activities (Footnote 3). I thought repealing the Glass–Steagall Act was a poor idea at the time (Footnote 4). My sense today is that that was correct, and I would go back, at the very least, to Glass–Steagall. All of this would move in the direction of getting us out from under the current “too big to fail” system. This would cause a shrinkage in the banking system because the banks without the “too big to fail” subsidy would not be profitable. Moody’s gives separate ratings for large banks: on a stand-alone basis and also with the “too big to fail” implicit guarantee. By using these ratings, it’s possible to figure out what the profits of the bank would have been if they did not enjoy the “too big to fail” guarantee. It turns out that most of the banks are not profitable without the guarantee. They would therefore have to shrink and find ways to be profitable. I think this would be a good thing. There are further directions as well, having to do with things like money market funds. Currently the main function of U.S. money market funds is to lend to the broker-dealer operations of foreign banks. It doesn’t seem to me that this should be a priority for U.S. policy. It doesn’t matter to us very much whether Deutsche Bank or UBS has a big broker-dealer operation or not. Finally, there are other areas in which we could improve matters, like disclosure. In short, there’s lots to be done.

FOOTNOTES [1] See Thomas Pikkety, Capital in the Twenty-First Century (Harvard University Press, 2013) who demonstrates that concentration of wealth and, consequently, systemic economic inequality occurs when the rate of return on capital is greater than the rate of economic growth in a society.
[2] The Basel Standards, commonly known as Basel I (1998), Basel II (2004), and Basel III (2010), are a set of recommendations published by the Basel Committee on Banking Supervision stipulating capital requirements for financial institutions in order to promote safe lending practices and efficient functioning of financial markets.
[3] The “Volcker Rule,” proposed originally by former United States Federal Reserve Chairman Paul Volcker, restricts commercial banks from engaging in speculative activity financed by customer deposits.
[4] The Glass-Steagall Act, more formally termed the “Banking Act of 1933,” imposed restrictions on the securities activities of commercial banks, particularly affiliations between commercial banks and securities firms. Several economists hold the opinion that repealing two key provisions of the Glass-Steagall Act (via the Gramm-Leach-Bliley Act of 1999) may have indirectly precipitated the 2007-2008 financial crisis.