The Spence and Hlatshwayo paper is now on-line

The Browser informs us (pdf behind that link), bravo to them, here is the abstract:

This paper examines the evolving structure of the American economy, specifically, the trends in employment, value added, and value added per employee from 1990 to 2008. These trends are closely connected with complementary trends in the size and structure of the global economy, particularly in the major emerging economies. Employing historical time series data from the Bureau of Labor Statistics and the Bureau of Economic Analysis, U.S. industries are separated into internationally tradable and nontradable components, allowing for employment and value-added trends at both the industry and the aggregate level to be examined. Value added grew across the economy, but almost all of the incremental employment increase of 27.3 million jobs was on the nontradable side. On the nontradable side, government and health care are the largest employers and provided the largest increments (an additional 10.4 million jobs) over the past two decades. There are obvious questions about whether those trends can continue; without fast job creation in the nontradable sector, the United States would already have faced a major employment challenge. The trends in value added per employee are consistent with the adverse movements in the distribution of U.S. income over the past twenty years, particularly the subdued income growth in the middle of the income range. The tradable side of the economy is shifting up the value-added chain with lower and middle components of these chains moving abroad, especially to the rapidly growing emerging markets. The latter themselves are moving rapidly up the value-added chains, and higher-paying jobs may therefore leave the United States, following the migration pattern of lower-paying ones. The evolution of the U.S. economy supports the notion of there being a long-term structural challenge with respect to the quantity and quality of employment opportunities in the United States. A related set of challenges concerns the income distribution; almost all incremental employment has occurred in the nontradable sector, which has experienced much slower growth in value added per employee. Because that number is highly correlated with income, it goes a long way to explain the stagnation of wages across large segments of the workforce.

A few points:

1. p.10 offers interesting remarks about China, namely that China is approaching a “middle income range” where economic growth commonly slows.

2. This paper has some of the best disaggregated information for those who are not convinced by simpler calculations of median income growth slowdown.

3. pp.33-34 offer a good summary of results and also a good explanation of current structural unemployment which does not fall prey to the usual criticisms offered by the blogosphere Keynesians, who on this issue remain behind the curve.

4. p.37 has good, short remarks on Germany and (now switching to my words) why it is wrong to dismiss their recent successes.

5. The co-author, Sandile Hlatshwayo, is at the Stern School of Business, NYU.  He, she or a namesake is taking a honeymoon poll.

Overall, this is one of the most important papers of the year and perhaps the most important paper so far on “economic malaise” issues.  It is also a useful corrective to the political conspiracy theories of changes in the income distribution (if you are wondering, Spence at least would not count as a right-winger, I cannot speak to Hlatshwayo).

As for The Browser, it is better than I ever expected a web site to be.

Addendum: Arnold Kling comments.



(That explains the last two sentences of the abstract).

Tyler, first we are at the end of March. Too early to talk about the best of 2011.
Second, and more importantly, the distinction between tradable and non-tradable goods was introduced in the 1960s to analyze devaluation in small economies and since then important theoretical and empirical research has been done based on it. That distinction was different from the old Keynesian one between importable and exportable goods in the analysis of devaluation.
Third, although I have yet to read the paper, it is my impression that the analysis of the extraordinary changes in China and India in the past 25 years must be approached from the viewpoint of a global economy in which there are government barriers to market integration both within large nations and between nations, particularly to the integration of labor and financial markets. Those changes are not marginal --we are talking about integrating around one third of the world population into the world economy. Therefore, the conclusions of any "partial" analysis (that is, one focused on a few countries or a few markets) have to be regarded at best as tentative.

So the bottom line is, it looks like healthcare and government work are in your employment future. But, if we cut government spending, we can move to China to get a job in the tradeable side of the market. We will cut government spending because the upper income level, of which I am a part, will be unwilling to be taxed because that would reduce their consumption of goods from Japan, Taiwain, Korea and China.

Makes sense to me.

Of course, the alternative solution to all of this is find some way we can cause people to borrow more, build more and bigger homes, and imagine their wealth is increasing as others join the bubble bandwagon. And, maybe we can do this will at the same time giving ourselves a tax cut with borrowed money.

Nuts, we did that already.

Any other ideas?

Bill, if you don't like your alternatives, vote with your feet. My first ancestors did it over 50,000 years ago and since then several generations of my ancestors did the same. I have followed the tradition: born in Argentina, I have lived in a few countries, now I'm in Chile, but I don't know where I'll live tomorrow. There is always at least one other to go.

Give in the the inevitable that the existence of a few billion poor Asians, and their moral right to have a decent job and consume the worlds resources just like us, means a convergence of incomes between the first and third world that many people in the first world won't like. There is no policy platform we can impose that will hold back billions of people from attaining a basic standard of living.

Without having actually read the paper, my main problem with the overview given here is: what is the message? Is it that all American companies will relocate abroad and we become Liberia? Should we close our borders and call it a day?

I am not kidding. The tone in these articles make it sound like we are at the edge of a total disaster. We are still the largest manufacturer in the world by a large margin. We are still home to the most inovative Tech companies in the world. To me this kind of doom and gloom sound like more politics than economics. Yes, we currently have high unemployment for US standards but Europe has been living with this for decades and they are doing just fine. Maybe the growth on other sectors was not that great exactly because our government was subsidizing education and health to levels we shouldn't?

Either globalization works or it doesn't. Of course jobs that produce lower value products will move to other countries. But if we are saying that all jobs (including high value) are moving away then something is really wrong and I have not yet heard one clear explanation of what that is.

Tyler, I've just read the most recent column that MS and SH have published based on their paper. See

Let me comment on each paragraph of this column.

1. "The global economy is at a crossroads as the major emerging markets (and developing countries more broadly) become systemically important, both for macroeconomic and financial stability and in their impact on other economies, including the advanced countries."
Comment: This is an understatement. The Global Market Economy (GME) included over 3 billion people in 1985 and now includes well over 6 billion. Any analysis of what happened to GME in the past 25 years must be analyzed as part of the process of absorbing such a huge number of people living under quite different political and legal systems. By far this has been the largest increase in any market economy in the history of humankind, so large that none has been able yet to propose a reasonable approach to explain how GME is being transformed.

2. "Consider, for example, what has occurred over the past 20 years in the United States. Some parts of the tradable sector (finance, insurance, and computer systems design) grew in value added and employment, while others (electronics and cars) grew in value added but declined in employment, as lower value-added jobs moved offshore. The net effect was negligible employment growth in the tradable sector."
Comment: The authors focus on the U.S. economy to argue that there has been a reallocation of resources. Indeed we know from Econ 101 that any large change in the number of people will lead to a resource reallocation. They choose to focus on the distinction between tradable and nontradable goods, one that since the 1960s has been used often to analyze small open economies but with little success. Actually the authors miss the most relevant change in relative prices that has taken place in the past 20 years: the large increase in the price of commodities in relation to manufactures and services --you should remember Raul Prebisch's theory of a secular decline in the relative price of commodities. In addition, they fail to explain why such a large increase in the price of commodities has not elicited a large increase in their supply and how the U.S. economy has been adjusting to this increase in the price of commodities (thus they ignore government's role in preventing an increase in the supply of U.S. energy).
BTW: to a large degree the discussion about inflation in the past 10 years has failed to take into account the inflationary effect of that change in relative prices (read about the structural school of Latin American inflation).

3. "The US economy did not have a conspicuous unemployment problem until the crisis of 2008 because the non-tradable sector absorbed the bulk of the expanding labor force. That pace of employment growth now appears unsustainable. Government and health care alone accounted for almost 40% of the net increment in employment in the entire economy from 1990 to 2008. Fiscal weakness, a resetting of real-estate values, and lower consumption all point to the potential for long-term structural unemployment."
Comment: They fail to explain why particular "non-tradable" sectors grew between 1990 and 2008. GME transformation was not the only determinant of the labor reallocation in that period. There were many other determinants, in particular many policies and regulations. We don't know the relative importance of all determinants but they want to give the impression that they know that the increase of employment in "non-tradable" sectors was not sustainable before the crisis and, regardless of the causes of the 2008 crisis, these sectors will hardly recover soon their peak levels and most important they will not grow. Indeed this view fails to take into account the many things that the U.S. government can do to eliminate distortions and restrictions and therefore to provide incentives for some activities to expand.

4. "One response is to assert that market outcomes always make everyone better off in the long run. But that is not supported by theory or experience. In the US, for example, while many goods and services are less expensive than they would be if the country were walled off from the global economy, we cannot assume that these cost savings necessarily compensate for diminished employment opportunities. People might trade away cheaper goods for assurances that a wide range of productive and rewarding employment options would be available, now and in the future."
Comment: I cannot believe that a Nobel Prize Winner has written this paragraph. Let me just say that I strongly believe that the history of humankind, before and after Adam Smith, fully supports the view that the elimination of artificial barriers to trade leads to a much larger number of opportunities for everyone. Actually the GME transformation is a good example of how China and the other new members of this GME are benefiting from their elimination of artificial barriers. The problem of U.S.A. and other old countries is how to adjust and unfortunately their political systems are failing to provide reasonable policies.

Also without having read it, my biggest problem with these big "structural problems" papers is that, save a troubling increase in inequality caused by institutional factors and two bubbles, the US economy did pretty friggin' well from 1990-2008. The current problems are cyclical.

And, I do suspect that as America continues to get richer, and cell phones and computers continue to get cheaper with rapid productivity advances, we'll continue to spend more on health care. Why wouldn't we? Having good roads, good schools, and good public transportation networks are probably also "luxury" goods from society's perspective...


5. "A second response is to acknowledge the distributional implications, but to accept them as the price of efficiency and openness. According to this view, the alternative – not having an efficient market system operating in a relatively open global economy – would be far worse."
Comment: They wanted to point out that the GME transformation would have had significant implications for income distribution in the U.S. economy, but the para. doesn't add anything else to the previous para. In particular, they ignore other determinants of changes in income distribution during the past 25 years. My impression is that GME transformation may have had some important direct effect on labor allocation, but its direct effect on income distribution would have been much smaller (policies and regulations would have been as important as GME transformation for labor allocation, and more important for income distribution). To answer these questions a lot of research has yet to be done.

6. "There probably are real choices to be made between income levels and distribution, on the one hand, and the range of employment opportunities on the other. It is not realistic to define the challenge as resisting or overriding the powerful market forces operating in the global economy. Rather, the challenge is how best to shift incentives at the margin in order to improve the distributional effects."
Comment: I agree that policymakers may face a trade-off between employment opportunities and distributional effects but earlier paragraphs don't provide any guide about these trade-offs, especially once they acknowledge that the effects of alternative policies will be marginal because of the overriding forces of GME transformation, a process that is far from complete and to which all governments are trying to adjust.

7. "There are several dimensions on which action can be taken."
Comment: This sentence is the first one of the seventh para. and of the second half of the column. This second half discusses the authors' proposal of light liberal policies. They suggest policies that amount to what U.S. local governments usually do to revive downtown areas but the authors are not optimist about their employment effects and concerned about their distributional effects. Because of their limited optimism, they feel pressed to warn that if their recommendations are not followed, there will be strong political pressure to close the economy. I've been living too long in Latin America to know well their admonitions and why we should not pay attention to them. Just let me remember what Raul Prebisch did when he had the opportunity to advise the Argentine government in 1956 --he ignored everything he had written about the urgency of implementing import substitution policies and recommended to eliminate all artificial barriers to production. He understood that Argentina had yet to adjust to the deep changes of the 1930s and 1940s and there was little that the government could do to accelerate it.

Um... is Sandile Hlatshwayo a guy? Might want to check on that Tyler...


Just a reminder that comparing raw wage rates between two regions or countries is fraught with difficulty

First of all, a comparison of wage rates means nothing without taking relative net productivity into consideration.

Secondly, even if the wage comparison is made between wages in a common currency, the wages, like all other prices, must tend to be higher in regions with
higher money supply per capita.

Regards, Don

Thanks to Uncle Sam 4.1 million jobs were created between 1990 and 2008. Bad news for anarchists and anti-government folks.

This is not even a paper. It makes no attempt whatsoever to draw or build on the existing literature. The "empirical" work consists of displaying graphs of time series and fabulous polychromatic 3D cylinder bar charts and then offering casual observations and eyeball-metric extrapolations. There is no reference to or use of any theory of international trade, and some of the conclusions sound jarringly inconsistent with basic notions of trade, notably the claim that low wage jobs have been "leaving" the U.S. and now high wage jobs are going to follow, leading to a "major empoyment challenge." It would appear the authors are worried that the U.S. doesn't have comparative advantage in anything at all! I see nothing new and correct here.

What am I missing, exactly?

I hate the 3D bar charts. If there was ever a bad chart representation this is it. Bar charts are bad enough usually. Pure chartjunk. Have we learnt nothing from Tufte?

Why would such respected economists use a 3D bar chart? Why? That's what I expect from corporate MBA's.

Kling says:

"On the other hand, if you want a lot of education and health care, then even if your income has grown a lot, you may be barely treading water."

Yes and no. Some of the highest value health care services you might want (statin drugs, joint replacement, minimally invasive surgery) didn't exist before. And with education, it depends on what you have in mind. A lot of that 'free stuff on the internet' that Kling mentions isn't just cheap diversions -- it's material that used to be available only at a university and at a high cost. The cost of education has dropped precipitously -- it's the damn parchment that keeps getting more and more expensive. Education is not inherently expensive to provide -- quite the reverse, but credentialism/accreditation/rent-seeking are limiting lower-cost alternatives. (But not blocking entirely, of course -- there are many situations where you need to know X or how to do Y, but don't need credit-hours or certificates for X and Y, and the internet has already lowered costs dramatically in those cases).

Note, too, that until a couple of years ago, Kling would have mentioned housing as one of the exceptions along with education and health care. Which of the other two will be the next bubble to burst?

Can anyone explain the choice of 1990 to 2008 and why it is treated as a single period when the graphs show at least two periods with different trends, and possibly as many as four periods?

Having looked at job creation by sector, I'm aware the datasets are not continuous as census/bls/commerce have modified their data collection to provide insight into the questions investigated by the paper. But I think the authors could have gone back to 1980....

In any case, I'll assume that the authors were both lazy and processing the most consistent dataset available and not cherry picking the start-end points to make a point.

I would note that their data suggests that if the US invested more heavily in drill baby drill, employment would be higher, not because energy would be cheaper, but because production per employee in mining fell from 1990 to 2008, and fell most rapidly in the later years. Thus higher US oil production would have meant even lower productivity and thus a disproportionate increase in job creation.

But investing more heavily in wind and solar would show an increased productivity per worker trend as volume increased and economies of scale brought unit costs down.

Hi Tyler,
as an added piece of information to your remark nbr #4 (on Germany) I'd like to mention this recently published paper on Germany's recent economic "success".
It may put it into perspective and I'm not sure that the authors of the CGS working paper were aware of it.
Best regards
Germany's economy
Miracle, or delayed gratification?
Mar 17th 2011, 18:49 by R.A. | WASHINGTON

MICHAEL BURDA and Jennifer Hunt have an interesting paper as part of the Spring 2011
Brookings Papers on Economic Activity which looks at the "German Labor Market Miracle".
Germany experienced an even deeper fall in GDP in the 2008-2009 Great Recession than the
United States, yet German employment scarcely fell. We show that 35 percent of the missing
employment decline was due to weak hiring in the preceding expansion, principally caused by
manufacturing employers’ lack of confidence the boom would last. Another 10 percent may
be explained by wage moderation. We argue that a third element was the increased use of
working time accounts, which permit employers to avoid overtime premia if hours per worker
average to standard hours over a certain window. These accounts did not change adjustment
in hours per worker, but did provide disincentives for employers to lay off workers in the
Just over one-third of the "missing employment decline" may be due to Germany's lacklustre
expansion prior to the Great Recession. You can see that in, for instance, this chart from the
blogger Kantoos:
As you can see, nominal output rose from 2000 to 2006, along with the unemployment rate—
which was over 10% at the point that America's housing market began to implode.

The discussion of tradables reminded me of Mancur Olson's "Rise and Decline" on free trade rather than Baumol.

Comments for this post are closed