Monday, August 29, 2011

America's Secret Manufacturing Success (And Why It Doesn't Matter)

The Center for Economic and Policy Research argues America needs more manufacturing: it's really the only way to pay for the goods we import. Services, they say, won't cut it. It's amazing how many mistakes they make in five short paragraphs. Here's the full text.
The NYT devoted a major Sunday magazine piece to this question. It never raised the most fundamental question, if we buy all our manufactured goods from someone else, how are we going to pay for them?

Our goods deficit is currently running at annual rate of around $800 billion or 5.3 percent of GDP. We have a surplus on services of around $170 billion a year, less than 1.2 percent of GDP. If we lost all our manufacturing, then the deficit on goods would increase by about $1.2 trillion to more than 13 percent of GDP.

What services do we think that we will export to make up this gap? We are rapidly losing ground in many areas. For example in software and computer services we are already a big net importer from India. It is hard to see how this gets reversed any time soon. We do earn a lot of patent licensing fees, but these fees will always be vulnerable to a tide of free trade sentiment. Besides, it is very hard to imagine them rising beyond a couple of percent of GDP as a maximum.

One of our biggest surplus areas is tourism. This raises the prospect that the anti-manufacturing crowd thinks that we are too sophisticated to work in factories, but not to clean toilets and make beds. There is nothing wrong with latter (I have done it as a summer job), but it's not what most folks would consider upscale employment.

The bottom line is that unless we think someone is going to hand us trillions of dollars worth of manufactured goods for nothing indefinitely, then there is zero doubt that America needs manufacturing. It also needs people writing on economic issues who know arithmetic.
OK, let's start with what they got right. First, yes the net exports on services is much smaller than that of manufacturing: $170 billion on services and $800 billion on exports. And as more services go abroad, that first number is likely to get a lot smaller.

But basic arithmetic doesn't get you very far here; it's much more complicated than it seems. Trade deficits are arbitrary distinctions between net exports and capital flows. Every dime that flows out of the country is a dime that flows in. Sometimes it's as an import but more often it's an investment. In other words, Americans pay for imported goods by supplying investment opportunities for foreigners. Despite all that's happened in the past few years, the US is still a hub of promising entrepreneurial activity. But more on that in a second.

This relates to what the CEPR foolishly suggests: that the trade deficit generates debt. Thus the "need" to develop manufacturing jobs because there is no debt. None. Zero. Nada. No debt is generated from the trade deficit. When I buy something from Japan, my debt does not increase (unless I borrow to buy that thing I bought). I owe Japan nothing.

Now it's true that few Americans work in manufacturing, but it's not because of trade as the CEPR implies. It's technology. Americans produce about 20% of the world manufacturing output: that's equal to Brazil, Russia, India, and China combined. And the reason why so few Americans produce so much is because each American is so productive. Technology allows one person do to the work of ten or a hundred which is why productivity is high while the number of jobs is low.

But even if you ignore all of that, trade deficit statistics suffer a fundamental accounting problem. A $600 iPhone imported from China increases the trade deficit to China by $600. But only a small fraction of that money goes to Chinese workers. Most of it (60%) goes to American workers for design, engineering, marketing, and profit. Adjusted for valued added, a new paper argues that the US's trade deficit with China (about $133 billion) isn't a deficit at all but a surplus at about $32.25 billion.

In other words, the US manufactures A LOT: they're called ideas.

Thursday, August 25, 2011

The Secret to Apple's Success

New York Times has a nice article over-viewing various patents which bear Steve Job's name (there are 313 total though I'm not sure the Times has them all indicated). Many relate to iconic Apple products. But many are just design patents including the casing for desktop computers, packaging for the iPhone, and the design layout for the iPod.

So Apple's secret seems to be claiming a monopoly on things people find cute. What's next? Patents on adorable kitten videos?

Monday, August 22, 2011

The Future of Higher Education

In about an hour, I'll have my first class as an assistant professor. Lo and behold I stumble upon two articles about the future of high education. The first is this article by econ prof James Miller who warns that increasing technological sophistication will eventually render flesh-and-blood professors obsolete. He advises new faculty to "get out while you can."

In contrast, the Chronicle of Higher Education reports a study of community college students who use online courses. They, in general, do worse than ones who go to a physical class. I can think of several sources of selection bias (those more likely to use online courses have less time to study, for example) but the basic theory that a lack of structure and community inhibit performance is believable as well.

I don't know what the future will hold but, in general, when technology makes things easier people favor to make up that difference with more leisure, not productivity.
Still, there's no doubt that, survive or die, online learning will be a critical skill for educators.

Interestingly, if both of these articles are to be taken seriously then this spells danger for the professors at the country's best schools since such are most likely to have (a) the funds needed to purchase the inevitably expensive super computers to run these automated classes and (b) the driven students who are the best position to take advantage of virtual learning. One must wonder how employers will read letters of recommendations from PROFESSORBOT-5000.

The Price of Patents

Google spent $400,000 per patent when it bought Motorola Mobility. Microsoft spent $510,204 per patent when it bought Novell. When Nortel Networks was bought, they got $750,000 per patent.

More here about the increasingly problematic patent war.

HT: Keith Pham