The Agenda

A Few Thoughts on the Great Relocation Thesis

Noah Smith has written a post on what he calls the “Great Relocation.” I recommend reading it, in part because I’m going to skip summarizing his argument. I agree with most of Smith’s prescriptions, e.g., an increase in high-skilled immigration, promotion of urban density, investment in infrastructure, lowering trade barriers, etc. I’ll focus on disagreements.

(1) Smith refers to the “nonsensical anti-train animus” of conservatives. As a fan of rail, I think it is safe to say that I have no nonsensical anti-train animus. I am, however, wary of the nonsensical pro-train sentiments of some non-conservatives, who assign quasi-mystical powers to high-speed rail. As George Monbiot has observed, many of the environmental claims advanced on behalf of HSR are overblown, particularly when we factor in the carbon-intensive process of manufacturing rolling stock. I am not averse to sustainable HSR that requires limited public subsidy, but like Stephen Smith of Market Urbanism I tend to think that cost-effective investments in medium-speed rail are a more sensible first step. 

Moreover, I often think that nonsensical pro-train sentiments flow from a failure of imagination. If Stanford’s Sebastian Thrun succeeds in fostering the widespread adoption of self-driving cars, we could radically reduce the congestion and energy costs associated with personal vehicles. The CityCar concept advanced in Reinventing the Automobile could make dense urban areas far more energy efficient while medium-range inter-city distances could be traversed by “trains” of personal vehicles that are made available via sophisticated sharing platforms and that move at least as quickly as the Acela. The advantage of these pseudo-trains is drivers could stop and start their journeys at any time and at any given place on an extensive road network.

Rail has the great advantage of being able to haul heavy goods at a relatively low energy cost, which is why the rail freight business has proven so successful. But people are very light. The Federal Railroad Administration mandates that passenger trains be far heavier than is strictly necessary, which swells the costs of domestic passenger rail projects. It is also true that sophisticated collision detection systems will allow us to build lighter personal vehicles, thus reducing their energy costs as well.

Projecting today’s transportation technologies into the future is, in my view, a mistake. Smith references Cowen’s thesis (by way of Peter Thiel) that transportation technology has been stagnant in recent decades, which is true enough, particularly if we use speed as our sole metric. But my view is that the rise of “the mesh,” i.e., of sophisticated Internet- and GPS-enabled sharing platforms, represents a significant step forward in transportation, and that the advent of self-driving cars and “intelligent roads” will deliver impressive productivity gains.

(2) Smith presents a stylized picture, and I think it’s unfair to point out that it isn’t entirely in tune with reality — he’d happily acknowledge that. I’ll just note a few things:

(a) If we believe Michael Pettis, as I do, China’s investment-led growth miracle is extremely vulnerable and likely to experience a sharp slowdown long before China achieves anything close to parity with the rich economies of North America and Europe. As domestic consumption increases in China, we will see a shift to retail and other services in which productivity gains are far harder to achieve than in manufacturing. 

(b) A recent BCG report suggested that America has an emerging manufacturing advantage relative to China’s YRD, due to soaring congestion and compensation costs in the latter. This might be just a blip, but it might also reflect the fact that wages represent a smaller slice of the cost of manufactured goods and that because the income gap isn’t close to closing, proximity to affluent markets still counts for something. 

(c) I like Smith’s reference to “sustain points.” One way of conceiving of China’s post-1979 growth performance is that its growth performance in the years from 1945 to 1979 were severely constrained, by civil war and famine among many other disasters, both natural and man-made. The next thirty years saw an explosive catch-up, driven at least in part by a demographic dividend. The next thirty years, in contrast, might see a sharp slowdown due to a combination of global rebalancing (the hypertrophy of China’s manufacturing sector will end and the next inter-sectoral shift will be growth-dampening rather than growth-enhancing), premature demographic transition (i.e., China will grow old before it grows rich, a phenomenon Smith acknowledges), and the fallout from decades worth of bad debts associated with state-owned and state-favored enterprises. That is, China has been “overperforming” and a severe shock will force a severe decade(s)-long readjustment that will make America’s present discontent look positively mild. 

(3) I think my deeper disagreement is over whether manufacturing really matters. I think that Smith winds up his discussion of the forces driving the Great Relocation with the conclusion that perhaps we never should have been building cars and fridges in the first place. I won’t quite concede that point, but I’d be very happy to argue that manufacturing cars and fridges isn’t vitally important. It’s at least possible that Smith agrees with me, per the following:

In fact, under some scenarios (which are difficult to explain concisely), the old Core even gets slightly poorer while the “new Core” catches up. For a while, the negative effects of relocation trump the positive effects of progress.

That is, it’s possible that I just have a different sense of the time horizon as to when the positive effects of progress kick in.

(4) Here is a version of the Great Relocation thesis I might be willing to buy: lots of less-skilled and mid-skilled manufacturing jobs have been offshored, and this has made it somewhat harder for non-college workers to earn a middle-class income. So far, this is a very familiar story. The U.S. tradable sector has grown far more productive, but employment levels haven’t increased over the last two decades. The nontradable sector has increased employment levels, but not productivity levels. The U.S. has experienced huge productivity gains, but the benefits have accrued to (i) the owners of corporate capital and (ii) skilled workers in knowledge-intensive “digital organizations.”

The story I’m most interested in vis a vis the American future isn’t about the rise of China, India, etc. Those countries face so many grave challenges that I’m more inclined to wish them well than to have any anxiety or dread over their success. Rather, I’m interested in how we can make the most of the changing demographic composition of the U.S., and how we intend to mitigate various social problems, e.g., the persistent failure of U.S. schools to offer a high-quality education to students from disrupted households, Latinos, and African Americans. If we can find a way to unleash productivity advances in education and the health sector, as Arnold Kling and Nick Schulz remind us, we can more than make up for the loss of mid-century manufacturing jobs, virtually all of which will be lost to robots and fabs in the not-too-distant future. 


The Latest