There's been an interesting back-and-forth between Ryan Avent and Alec MacGillis over urban development-guru Richard Florida. Defending Florida, Avent most clearly explains his thinking in this section of his first response to MacGillis's American Prospect article criticizing Florida.

But what makes a successful clump changes over time. The economics that underpinned the older manufacturing economy supported clumps that don’t necessarily make economic sense today. With declines in transportation and communication costs, it became affordable to move plants away from expensive city land, and that’s precisely what many businesses did. In cities that were also home to a substantial knowledge economy sector, this ultimately proved to be a boon. By outsourcing their manufacturing (and later, their back office) components, firms could reduce the overhead on the offices of those who still needed to be in the city, improving margins (and making more room in the city for others who needed to be there, thus increasing the return to everyone of being in the city).

The result is a world where the key to urban success is a critical mass of workers with high levels of what economists like to call human capital. And because there are returns to scale at work, there is an element of the zero sum here. Or to put it another way, the world where every big city has its own fair share of talent is not a stable equilibrium; it will decay into a world with haves and have nots. And indeed, that’s what we have seen in recent decades. Educational levels in cities one hundred years ago strongly predict educational levels in cities today. And cities with high shares of college graduates have absorbed more than their share of new college graduates in recent decades.

The main problem with this argument is that it fails to account for the role of foreign competition. Transportation and communication breakthroughs not only enabled companies to disperse their operations all over the world; they enabled foreign companies to import cheaply and compete with U.S.-produced goods. Starting in the 1960s, U.S. manufacturers lost substantial market share in a variety of goods to foreign competitors, and many economists attribute the economic stagnation of the 1970s to the inability of U.S. industries to compete with foreign companies. In the 1980s, with the aid of deregulated markets, companies were able to return to profitability only through mergers and extensive outsourcing to low-wage countries.

Human capital levels are important, but the bigger issue is why they are important. Of course, the rise of high-tech clusters and Richard Florida's creative enclaves are part of the story. But so is the role of the global dispersion of production, which has given rise to new forms of agglomeration. Global production networks need to be coordinated, and these jobs typically require high levels of education and specialized knowledge, and therefore are subject to agglomeration in a few select cities, creating jobs in business and financial services. While this process has helped revive some American cities, it also creates a need for a massive industry engaged in speculative finance.

The financial crisis has shown just how unstable this arrangement is, and perhaps it has also demonstrated that the returns to this system of production are declining. After all, at some point, industry will drive labor costs as low as they will go.

Finally, the increasing importance of human capital does not necessarily mean that high-wage manufacturing in America is dead. One option is to take a "clusters" approach to regional development, which the Brookings Institutions' Metropolitan Policy Program is promoting. Others include adjusting the terms of trade, spreading best practices and innovations through manufacturing extension programs, reducing workforce shortages and upgrading existing workforce skills.