Inequality and the China Trade

Paul Krugman writes about trade, China, and inequality. He says that in the late 1990s he, like most economists, thought the impact of trade on inequality was small. He says that the same models that showed a small impact then, point in the direction of a much larger impact in more recent years (and in the years soon to come), essentially because (to oversimplify a bit) China is so giant. Brad DeLong has his doubts:

But if those goods were to be produced here in the United States, they would be produced with higher-skil labor and with lots of capital. The key question is how has the shift in economic activity created by expanding trade affected the demand for different kinds of labor and capital here in the United States. We have had:

• A shrinkage in export and import-competing manufacturing.
• A tremendous expansion in construction.
• An expansion in consumer services.

I don't see how those shifts significantly reduce the demand for factor of production "labor" and enhance the demand for factor of production "capital." Construction employs lots of capital--but so does tradeable manufacturing. I want to see Leontief input-output matrices for the U.S. before I upweight trade and downweight education, collapsing unions, migration, changing norms, monetary policy, and other factors as more likely to be responsible for the lion's share of the increase in U.S. inequality over the past generation and a half.

I'm no economist, but DeLong and Krugman are partly talking past each other here. DeLong says the lion's share of growth in inequality "over the past generation and a half" are due to non-trading factors. Krugman, however, is talking about very recent events and events that are likely to occur in the near future.

To argue with guys with PhDs, if I were answering a question on the AP Macro test about opening up large-scale trade with a country featuring a huge labor force, I would say it should permit a faster rate of non-inflationary growth than previously existed. 21st century America has seen steady GDP growth but not anything remarkably faster than what took place previously. At the same time, we've seen little-if-any wage growth for most people. This is all related and the economy "should" (in some sense) be growing much faster.

Comments

"DeLong and Krugman are partly talking past each other here."

That's an overly kind way of putting it. It'd be more accurate to say that DeLong is badly missing the point.

And beyond the matter of DeLong arguing a different issue than Krugman is posing, I'd say that PhD or not, DeLong is getting a bunch of things wrong in his post.

-----

As always more trade and a more robust government safety net are the answer to whatever question is posed.

Posted by Petey | June 16, 2007 9:49 AM

The argument for free trade is about increasing the total size of the pie, not how it is distributed.

Even if you had more free trade, if you raised the tax rates on those making say more than 200k/yr to 70% and cranked up the estate tax to the point where no one would care about Paris Hilton, then inequality would fall even as free trade increased.

Posted by paul | June 16, 2007 10:06 AM

Dani Rodrik has another view:

"I think Lawrence is right to the extent that the skill premium has stopped rising since 2000, and therefore the type of approach that Cline and others used and which Krugman thinks would yield larger estimates of trade's contribution today, would not actually explain current inequality (which derives from the rise in incomes at the very top and from the increase in the profit share). But there are other (non-Stolper-Samuelson) models, based on bargaining for example, that could."

Posted by Shane Taylor | June 16, 2007 10:37 AM

Income and more importantly asset inequality in the US is driven first by the ever accelerating inflation of asset prices.

It's been 34 years since 73 and the post war trend of rising wages ended. It's long past time that to recognize the post war period was not a norm that can be regained. It was a one shot deal.

There is an ongoing tsunami of wealth creation that is being distributed very unequally. This 'wealth' is being created by financial means thru the ever increasing prices of most forms of financial assets. This is being enabled by an equal explosion of credit and money. Not the credit of the old days extended by banks and not the money of M1 and M2 which is what the real world economy runs on. Rather it is money and credit created by the Wall Street centered financial system. It's the M3 money, the money the Fed no longer reports since they have no control over it.

The explosion of hedge funds are a good place to look when envisoning this process. Already wealthy people are pouring trillions into these speculative funds. These trillions chase rising prices in various asset classes creating a self fullfilling cycle. Occasionally there is a bump like the ongong sub prime mortgage thing.

The main point is that asset prices are not increasing in value, they are inflating due to the said explosion in the M3 type money supply flooding the entire world. It isn't hard work or genius, just good old inflation. Us normal schmucks putter along and do OK in the M1 word but rarely better, and the top two deciles are doing super. Here to unequal distribution reigns. The top .1% is astoundingly richer, The 80th percentile is just doing good enough to get that McMansion and the Beemers.

It isn't that the ordinary guys and gals are doing so bad, it's just that the top is doing so good. That's what makes these issues so devlish to discuss. Just rember the reason the top is doing so well is because they are inflating the prices of their assets, with a generous helping hand from evey branch of goverment up to and including especially the Federal Reserve.

Posted by rapier | June 16, 2007 11:00 AM

"It isn't that the ordinary guys and gals are doing so bad, it's just that the top is doing so good."

Hogwash. I think flat median wage growth over the past 35 years is indeed "doing so bad".

And we shan't even discuss the devastation that has struck the bottom quintile, shall we? After all, they don't read blogs.

Posted by Petey | June 16, 2007 11:05 AM

21st century America has seen steady GDP growth but not anything remarkably faster than what took place previously.

Then again, the 1980s and 1990s were also a time of rapidly increasing trade with the outside world (especially Asia), as countries like Thailand, Korea, Indonesia (remember those Nike factories?), and Malaysia joined the mix. In the 1990s, China was ramping up its trade with us as well, just not to the extent we've seen in the last 6 years.

Saying we haven't been growing as fast as we "should" be is very hard, since it's almost impossible to find a reference point for comparison.

Posted by Mr. Noah | June 16, 2007 11:18 AM

There are two things that confuse me about trade issues.

China has a different economic system than the United States, or at least I hope it does. You don't have the military owning busineese in the US, labor unions are repressed, the currency doesn't float, etc. Wouldn't the standard model of trade theory not be as accurate for trade between countries with such different economic systems? When Ricardo was writing, every European country had substantially the same economic system. China at the time did not participate in the international trade network, due to Chinese government policy.

Second, if country A allows labor union organizing, bans dropping toxic sludge in rivers and enforces it, and so on, and country B doesn't, and I move my business from A to B in order to export back to A, aren't I just trying to evade A's regulations? Why should government A be OK with this, or even encourage this? And if A values its own regulations so little as to allow this type of evasion, why doesn't it just change its regulatory regime to be more like B so at least more businesses don't relocate?

Posted by Ed | June 16, 2007 12:11 PM

I don't think there's any doubt that Krugman is a better economist than Brad DeLong; on trade, Krugman is a better economist than almost anyone else putting out an opinion into the blogosphere.

Having said that, it ought to be acknowledged that Krugman is reasoning from an a priori, theoretical model, which may be seriously oversimplified. Most economists are not as familiar with the actual patterns of trade and their implications, as a naive layman might reasonably expect. Krugman is taking a possibly oversimplified analysis, doing a back-of-the-envelope calculation and telling us what he expects.

DeLong is reasoning from a more plausible model, which includes some elements, which Krugman leaves out, to wit: the effects of trade on the returns to factors in the so-called non-tradeable sectors.

You can be sure that Krugman's model is logically consistent; only the correspondence with the real world is in doubt. It is not clear that DeLong's model is even going to be arithmetically bounded: it is hard to tell if he's even going to be able to make things add up -- he's waving his hands over a number of variables that might matter. DeLong's temperamental optimism can easily get the better of him, in the context of such a facile narrative analysis.

The most obvious point to make, in reply to them both, is to suggest that the median wage is tied directly to marginal productivity, and that wage growth has lagged because marginal labor productivity has lagged because growth in capital per worker has lagged. Marginal productivity is basically a function of the worker leveraging the power of his tools, of the productive capital per worker; the stock of productive capital per worker is not increasing.

From there, it is not so hard to get to a narrative analysis of how technological obsolescence and changing patterns of trade erode existing capital stock. Or how financial policies tilt returns on new manufacturing plant investment to China and make the New York financial services industry wildly profitable, but the Midwest rustbelt an manufacturing investment desert.

Economists of a lesser caliber than either Krugman or DeLong are prone to making certain characteristic errors of analytic judgement, that DeLong and Krugman are less likely to make. One is to ignore the dominating importance of increasing returns (the combination of economies of scale and a virtuous cycle of innovation) that make an economic powerhouse out of a Silicon Valley, a Hollywood (or, in an earlier era, a Detroit).

The other bias is to overemphasize allocative efficiency to the point of thinking that a populous country like China or India will or should "invest in its comparative advantage" in "labor-intensive" goods. This is utter nonsense, precisely because of the dominance of increasing returns in building income from international trade. A country like China is building a capital-intensive manufacturing sector, by sacrificing current consumption and the foreign exchange value of its currency, to make almost any investment in manufacturing financially profitable.

The mirror of that is that almost any investment in manufacturing in the U.S. is depressed.

Could U.S. government policy change the economic landscape in such a way that investment to increase the capital stock per worker earned a financial return, and therefore increase marginal labor productivity and wages?

Clinton did it, by reducing real interest rates, and triggering an investment boom.

Bush has stifled domestic investment in productive capital stock. Corporate profitability is very high, returns in the financial services industry are fantastic, but corporate investment in capacity is very low, and median wages, even with very low unemployment are barely growing enough to cover the increased cost of medical insurance.

The most irresponsible thing is for economists to pretend that the economy is meterological and mysterious and inevitable, and not something entirely synthetic and artificial. There is no such thing as the "natural" economy. The economy is the outcome of choices and policies. Whether it is intentional is entirely a matter of how smart we are about discerning the consequences of our choices.

Posted by Bruce Wilder | June 16, 2007 1:18 PM

"The most obvious point to make, in reply to them both, is to suggest that the median wage is tied directly to marginal productivity, and that wage growth has lagged because marginal labor productivity has lagged"

Marginal labor productivity has been flat for the last 35 years? News to me.

Posted by Petey | June 16, 2007 1:24 PM

"if country A allows labor union organizing, bans dropping toxic sludge in rivers and enforces it, and so on, and country B doesn't, and I move my business from A to B in order to export back to A, aren't I just trying to evade A's regulations? Why should government A be OK with this, or even encourage this? And if A values its own regulations so little as to allow this type of evasion, why doesn't it just change its regulatory regime to be more like B so at least more businesses don't relocate?"

Which is why all of our industry has moved to Chad. Oh wait, no it hasn't. As Ha-Joon Chang has pointed out, building up the institutions and rules that the rich nations have built up - labor standards and protections, intellectual property regimes, effective environmental regulatory regimes, etc. - are rather expensive and often only saw real progress in rich nations as those countries became rich, in part through trade. It is more likely to succeed to help poor nations become rich through trade that makes meeting these standards easier than to ask, say, Liberia, to catch up to us.

"A country like China is building a capital-intensive manufacturing sector"

Ever looked inside a productive Chinese factory? The "capital-intensive" manufacturing in China is actually more of a drain on the company that is there to prevent rioting. Under communism, state-owned heavy industry was placed in the Northeast away from Taiwan. As China opened up, the Southern areas closer to Southeast Asian ethnic Chinese networks, Taiwan and Hong Kong received FDI that took advantage of China's comparative advantage in labor. Kellee Tsai's research into informal financial networks in China centers around the financing of small businesses that rely largely on labor. The relatively more capital-intensive manufacturing in the North soaks up official loans from state-owned banks for the sole reason of keeping those workers employed and thus keeping the "floating population's" numbers down. These are bad loans because these enterprises usually cannot pay back the loans due to poor profits our constant losses, which is the time bomb in China's financial system. It's not like real estate speculation in Shanghai is what is the primary fuel of the Chinese economy.

Posted by Reality Man | June 16, 2007 1:47 PM

I think "China Trade" is simply a euphemism for a form of slavery. Some companies own a bunch of slaves in China, we all benefit from exploiting these slaves but some of us benefit much more than the others. How 'bout this for a model, simple enough?

Posted by abb1 | June 16, 2007 2:17 PM

Rather than treating us to relentless, platitudinous paeans to the glories of free trade by Thomas L Friedman the press might begin by covering these trade agreements as what they are: dirty deals no less poisoned by special interests than practically anything that comes out of the arsehole of Congress.

Monetary policy means nothing to most people; they only know the factories and the jobs are gone. The stink of politicians and the rich behaving badly does.

Posted by Linus | June 16, 2007 2:30 PM

Bruce, where exactly are you getting your capital per worker numbers? Are you just looking at an overhang from what I understand was the very rapid growth in capital per worker during the tech boom?

Posted by mq | June 17, 2007 1:49 AM

Petey:

[me:]"The most obvious point to make, in reply to them both, is to suggest that the median wage is tied directly to marginal productivity, and that wage growth has lagged because marginal labor productivity has lagged"

Marginal labor productivity has been flat for the last 35 years? News to me.

Maybe I am being prickly and grumpy, by reading that "news to me" as sarcasm. Probably, I am just feeling foolish and oversensitive about writing comments almost no one will read, where I appear to criticize economists, who are smarter than I am. But, dammit, I'm right and they're wrong. So there.

Reading my own comment, I realize that "marginal productivity" is probably an unintended neologism. No one else uses such a term, which is too bad, really.

But, petey, the marginal product of labor is the wage. So, if real wages have stagnated, it follows, practically by definition, that marginal product has stagnated as well. And, this really should not be news to anyone. The real wage of the median wage-earner, after growing steadily from the end of World War II to 1973, has basically only fluctuated with the business cycle since. Maybe, there was a bit of growth late in the Clinton boom, but that's it.

Average productivity -- which is total output divided by labor hours worked -- has increased over that time. The difference between rising average productivity and stagnating marginal product is just an indirect way of saying that Capital is taking an increasing share of output.

And, while, even with a declining share of total income relative to Capital, Labor has, on average, scored some income gains, almost all of those gains have been at the top of the distribution. CEO's are paid a lot more, now, than they were in 1973.

You could argue that the marginal product of CEO's is soaring, and so, the average marginal product of labor is also rising a bit, but new technology has been biased in favor of the productivity of top management. I think that's crap, but that's a whole other argument. The median marginal product has not been changing, is my point.

mq: "Bruce, where exactly are you getting your capital per worker numbers? Are you just looking at an overhang from what I understand was the very rapid growth in capital per worker during the tech boom?"

In my argument, above, I was inferring capital stock per worker. It is a logical inference, given the relation between marginal product and capital employed.

Capital stock is only very imperfectly observable or measurable. Much of "capital" is intangible -- technical knowledge and training, general education, social structures, organizational procedures, brand recognition and reputation, and so on. And, even tangible capital is subject to obsolescence, which can be difficult to measure.

The tech boom, triggered in part by Clinton's policy of eliminating the government's fiscal deficit and its crowding out of private investment demand, did result in increased capital stock per worker. And, did result in a rare rise in the real median wage.

In fact, it is the rise in the real median wage, which allows me to feel comfortable inferring that there even was a net increase in capital stock per worker.

Obviously, the tech boom saw a lot of measured spending on new capital equipment, as well as the creation of new organizations and businesses, all of which may tend to increase the capital stock per worker.

But, the tech boom also saw a less well-measured increased obsolescence of the existing capital stock. The creation of cellphone networks and fiber optic internet backbone added to the capital stock, but the accompanying obsolescence of the copper wire telephone network decreased capital stock. It did not happen with perfect simultaneity, which is a prime reason that the economy shows a pattern of booms and busts -- for a short time, the explosion of the internet actually increased demand for plain old telephone service -- people got second numbers for internet connections, internet connection calls stretched out for hours and hours, placing a serious strain on the phone network (but also boosting the utilization and therefore the productivity of the old network). But, now, it is obvious that the whole copper-wire, switched line phone system will go away; the number of phone numbers in the U.S. actually declines at a rate of a half million numbers per month.

Anyway, I think what we can observe of capital spending tends to reinforce or confirm what we can logically infer. When capital spending increased in the Clinton boom, finally, we saw increasing real wages.

The Bush deficit and the low level of corporate capital investment appear to have combined to hold down wage growth even though the unemployment rate is pretty low.

Posted by Bruce Wilder | June 17, 2007 11:58 AM

Fair enough, Bruce.

Posted by Petey | June 17, 2007 12:06 PM

Re: the number of phone numbers in the U.S. actually declines at a rate of a half million numbers per month.

Unless you meant to say, "The number of landline phone numbers..." I don't think that's right. Not only is the population increasing, but the number of people with cell phones is still increasing. Before, each household had one phone number. Now, each person in a household has their own phone number. The number of phone numbers, counting cell numbers, should be increasing still.

Posted by JonF | June 17, 2007 8:24 PM

normally, i'd be unwilling to wade into a BDL/PK argument and insist that i can clarify things, but, heck, i am actually writing a book on trade and inequality, so, here goes.

BDL is (whether he knows it or not) thinking about the impact of the trade *deficit* on inequality. he's right that the deficit's impact on inequality is ambiguous. i tend to vote very slightly unequalizing, but, only because there are still smallish labor rents in manufacturing.

PK is talking about the long-run impact of a rising share of imports from LDCs (mostly China) on inequality. This is utterly unambigious - it is highly inequality-generating.

and that's really the whole disconnect. almost.

Posted by josh bivens | June 18, 2007 2:40 PM