This post by Matthew Klein on the Financial Times‘ Alphaville (FT’s “free” blog) got me wondering about capital mobility, inequality, and job exodus.
The point is simple enough: crunch the numbers, a lot of manufacturing jobs were lost to foreign competition in the First World. Yet, this is also wrapped in a paradox: a lot of jobs (although not nearly as many) were gained in manufacturing, as US manufacturing exports shot up significantly since 1990s as well. In the usual economic terms, the net value created exceeds the net value lost, and all that.
It is tempting, especially following the argument put forward by Piketty and the Stolper-Samuelson logic, to think in terms of rewards to capital and labor in the “West” and the “Rest.” Basically, capital is plentiful in the West. Labor is plentiful in the Rest. West focuses on the capital intensive output, while the Rest focus on the Labor intensive output. Except, this does not work.
In this tweet exchange, Ed Crooks raises an excellent point. A lot of goods that do not seem obviously “capital intensive” ARE in fact capital intensive. Auto parts are as capital intensive, if not more so than jet engines. To be fair, this is a classic example (and a common mistake by non-econs–and I’d been non-econ long enough to have forgotten about it) about what it and what isn’t capital intensive: potato chips, for example, are a very capital intensive product, thanks largely to automation. (This was in context of “computer chips, not potato chips” exchange in 1990s debate over free trade.)
I wonder if the naive application of Stolper-Samuelson logic has the argument backwards: in the present world (and I think, in most cases in the real life generally), capital is mobile, goods are mobile, but labor is much less mobile than the others. If an industry is capital intensive AND the demand elasticity of the capital is high (i.e. it does not matter who works the machines, as long as somebody does), it can be set up anywhere. You can have workers making pennies a day churning out computer chips in Southern China (literally true), as long as there is no barrier to capital mobility, and sell the output in United States, as long as there is no barrier to goods mobility. To be fair, this is something that Piketty recognizes as odd in his book: he observes that a lot of gains to capital today seems to accrue from what he describes as “human capital,” rather than physical capital, as per Marx’s argument. Of course, “human capital” is something that just never made sense (at least to me) as an unambiguous concept. Human capital, as we understand it, is much more given to complementarity problems rather than the amorphous “monetary” capital, although it is perhaps analogous to the physical capital that has already been invested in. Much of the 10 years spent as a Cobol programmer, for example, would be a complete waste in an environment where Cobol is not used (i.e. everywhere these days). Certainly, the knowledge acquired about how to set up algorithms, general programming approach, and computer-science know-how, will be applicable beyond Cobol, but they will constitute only a relatively small fraction of the investment into the human capital. (I was reminded of this when someone casually wondered, “what kind of ‘data scientist’ isn’t very good at Python?” Umm, maybe the main qualification for a “data science” person is really knowing something about what to do with “data”? But this isn’t really how “data science” seems to work in practice.) The bottom line is that a lot of human capital investment is not as fungible as monetary capital.
Demand elasticity, at least, conceptually, captures the substitutability of different kinds of labor, given capital. Workers who make computer chips are easily substitutable. Workers who design computer chips are not. But there are far more workers who are (potentially) qualified to make computer chips than design them, not just in the West, but also the Rest. So computer chips are designed by a handful of people in the West, but are manufactured by many people in the Rest. The people who design the computer chips make a lot of money. The people who make computer chips get a decent amount of money. The “rest of the people” in the West get nothing–until they are willing to lower their cost to that of their counterparts in the “Rest.” Since all these are capital intensive activities, owners of the capital, wherever they are, get plenty also. So the story is a bit more nuanced than Piketty’s, perhaps? The bottom line is that the inequality bug hits the West far harder than the Rest as the capital becomes both more mobile and the returns thereto increase relative to “labor,” but with a peculiar twist that comports with “meritocracy,” in that those who offer less substitable, less commodifiable skill sets, for whatever reason (in technology, arts, sports, etc) are spared.
Additional trouble comes from that demand elasticity is a fickle thing. Often, new technologies, fashions, and fads quickly make older “skill sets” obsolete and make them as replaceable by anything, and this process is usually unpredictable. Remmeber: in 1999, the few people who still knew how to code in Cobol was highly in demand. Once 2000 hit, the only thing they had to sell were their skills in basic programming and computer science concepts, which, although hardly insignificant, were but a fraction of their human capital investment. This discourages deep investment in a particular kind of human capital: deep understanding in topics without obvious immediate economic applications. For example, becoming an expert in ancient Norse is irrelevant, unless you are so enamored with old languages and independently wealthy–as an economically useful “human capital,” knowledge of the Old Norse is less than nothing. Funny thing, since knowledge of Old Norse and other dead languages did help make Tolkien’s novels fashionable…but the popularity of Tolkien’s work hit crescendo long after he’d have appreciated (even if he might not have really “needed”) the money.
PS. One analogy, albeit an incomplete one, for my thinking is something like the following (it is not the same because, rather than the Rest of Worlders taking away jobs, it’d be the robots, paid for by capital, that’d take the jobs away in this example–but econ textbooks remind us that they are functionally the same): Google Translate makes translation easy, but some translation is much easier than others. Invest a lot of capital in Google Translate, then you can push translators of contracts, technical documents, and other easy-to-translate written matter out of jobs. Translating poetry, however, is technologically more difficult at a fundamental level and their translators will not be so easily substituted, but only if there is a significant demand for poetry in Russian, say, among the non-Russian speakers. If you will, the lack of substitutability is a necessary, but not a sufficient condition, and what among the not-so-easily substituted labor will become fashionable is, potentially, completely arbitrary.