Every once in a while, I run into a bit of economic dogma that just doesn't seem right. Mostly, I just let it slide. But others aren't so embarrassed. As when credentialed economist Tyler Cowen asks why the coordination problems that keep even small countries from running planned economies don't apply to the management of General Electric, which manages more than many small countries' GDPs. (Even Brad DeLong offers no easy answer). So maybe I ought to let a few of my own questions out.

So here's one. In the offshore outsourcing debate, it's a cliche from the free-trader side that innovation can be the mainstay of the future American economy, even as any routine job goes overseas. Here's Daniel Drezner making the case in a widely cited piece in Foreign Affairs:

One thing that has made offshore outsourcing possible is the standardization of such business tasks as data entry, accounting, and IT support. The parts of production that are more complex, interactive, or innovative -- including, but not limited to, marketing, research, and development -- are much more difficult to shift abroad. As an International Data Corporation analysis on trends in IT services concluded, "the activities that will migrate offshore are predominantly those that can be viewed as requiring low skill since process and repeatability are key underpinnings of the work. Innovation and deep business expertise will continue to be delivered predominantly onshore." Not coincidentally, these are also the tasks that generate high wages and large profits and drive the U.S. economy.

So, here's the dumb question: does reality really work like this?

To begin with, let me try to set a few terms for the debate. To me, at least, it isn't really about total numbers of jobs on-shore or off-shore. Economists often make that explicit when they're modeling the effects of trade -- they just assume full employment on all sides, to simplify things. Which really just amounts to saying that when things all settle out, someone will be willing to pay Americans some amount of money to do something, so I don't even have a problem with it. And that, in turn, means that the extensive arguments about total numbers of jobs in Drezner's piece don't really address my own concerns in the outsourcing debate, which are these: when things all settle out how much will Americans get paid to do what?

Drezner seems to offer a future in which outsourcing has left high-wage jobs in marketing and advanced R&D on shore, while routine work has gone overseas. But some of the most publicized outsourcing and offshoring is precisely of research and development. Business Week's "Innovation Economy" issue last fall had an article on "Scouring the Globe for Brainiacs" which starts out with a tour of Microsoft's new research lab in Beijing (where Bell Labs and IBM Research also have labs in place). It goes on to discuss hard-core R&D by native Indian firms like Bangalore-based ITTIAM Systems:

Founded in 2001 by seven veteran Texas Instruments executives, the 125-engineer company has found growing demand for its embedded software and systems designs for decoding highly compressed audio and video content on the MPEG4 format. ITTIAM's 50 clients in the U.S., Europe, and Asia have used its designs in everything from a hand-size $199 camcorder to a "digital-media album" that can store up to 130 hours of video and thousands of songs.

One customer is e.Digital Corp. The San Diego outfit develops multimedia appliances such as the digEplayer, a portable in-flight entertainment system used by 11 carriers, including Ryanair, Alaska Air Group Inc., (ALK ) and Hawaiian Airlines Inc (HA ). The machine, sold by Tacoma (Wash.)-based APS Inc., stores up to 30 highly compressed movies. Economy passengers rent the player for around $10 to watch what and when they please. "We're constantly scouring the world for high-performance technologies that are already out there," says e.Digital Senior Vice-President Robert Putman.

So, we have here an example of this San Diego-based company where a lot of the R&D jobs behind their new products have been effectively outsourced overseas. Which is an increasingly common pattern. Many American venture capitalists want the startups they invest in to have an outsourcing strategy even at a stage so early that R&D is the major expense.

A common dodge in arguments like this, by the way, is to say that only the boring, less innovative parts of the development process are going off shore (thus applying the conclusions of Drezner's IDC study to the R&D process itself). It's not clear to me why anyone believes that, other than pure chauvinism: if Indians in India can do it, whatever it is, it must not be all that hard. So, looking at the list of projects ongoing in IBM's New Delhi lab, we can discover that developing new analytic techniques for analyzing magnetic resonance brain imagery and applying open source technologies and practices to the problems of government are boring jobs that require no real innovation.

But let's grant the contention, for the sake of argument, that only "low skill" jobs are going overseas (and that computer programmers and M.D.s interpreting X-rays have low skill jobs, because those jobs are going overseas). We're then left with the problem that low skill jobs, particularly by such an inclusive definition of "low skill", are most of them. Not everyone wants to get a job in advanced R&D or marketing (or the unspecified other areas that Drezner points to). Not everyone could if they wanted to -- for many, it would take years of expensive education training that they can't afford. Particularly not if they're still paying down the loans on training for one of the IT jobs now going overseas, from back when the cliche was that it didn't matter that manufacturing was vanishing, because IT jobs were the future of the American economy.

(But the gurus have learned their lesson from that little episode. "You have to take the leap of faith that the economy will evolve and there will be this innovation economy that comes", opines outsourcing guru John McCarthy of Forrester Research. And is roundly roasted for it by Brad DeLong's commenters, because even if you share his faith that The Quasi-Deified Market will provide a new innovation economy, he offers no hint at all as to who it will hire).

Besides, in making that argument, pro-globalizers seem to forget where they started. The whole point of globalization is that anyone can provide any service to anyone else, anywhere. That includes innovative research and development as much as any other type of service -- as in the projects I mentioned above which IBM has outsourced to New Delhi. All you need to be productive as a computer science researcher is a net connection and a computer in an office, with a semi-reliable connection to a power grid. (At the power levels required by modern PCs, a commercial uninterruptable power supply will cover glitches that would really put a kink in a manufacturing operation). And the best Chinese and Indian researchers, trained in many cases in American universities, are as good as the best Americans. So, to restate my first dumb question about economics: why does anyone think that "innovation" will mean more to the economy of twenty-first century America than it did to post-World-War-II Britain, where the first jet airliner was built, the first modern computers were built (and the first company -- a chain of tea shops of all things -- made use of computers in daily work), the structure of DNA was discovered, and the economy as a whole didn't do much.

By the way, to answer one ad hominem that pro-globalization flacks frequently toss at their critics, I have absolutely no problem with Indians and Chinese trying to do the best for themselves and their economies. My problem is with Americans who suggest that we as a society should ignore obvious deleterious trends and, like McCarthy, have faith that the Market will correct them, and offer well-paid, intellectually rewarding jobs for everyone, without offering any convincing reason to believe that it will.

There is, however, an more sophisticated pro-globo argument -- that all of the problems that are associated in the popular American discourse with outsourcing really shouldn't be, and that what's actually going on is structural changes in the American economy which are hitting workers in the gut. The Brad DeLong piece I mentioned earlier will serve as a decent example, and I actually do buy that argument as far as it goes -- though I may have some stupid questions later about how far that is, and whether the solutions commonly proferred for those problems are realistic. But for now, I'm just trying to address the "innovation" argument, and stupidly asking why anyone buys it.