Friday, January 18, 2013
A Comment on What’s Causing the Rise in Inequality
I was at the ASSA session where Larry Mishel faced off against David Autor, and I came away thinking, like Jared Bernstein, that the EPI view of the world holds more water than the it’s-all-technology argument that Autor was defending. The timing issues, including the discontinuities in the wage structure traceable to the early 1980s, are important, for instance.
Nevertheless, I think the whole debate suffers from insularity. The critical technological developments of recent decades, especially digitization, computing and networking, have swept over the entire world. They have transformed work in every developed country and much of the developing world. Meanwhile, the dramatic rise in inequality, and the portions of the earnings distribution most impacted, differ tremendously. You simply don’t see the same change in profile in most of continental Europe or Japan that you find in the US data.
Understanding these differences is where explanation would begin, in my opinion. And here’s a hypothesis: analysts of income inequality in the US suffer from a dynamic, self-reinforcing lamppost effect. They begin with a model of the world in which only the individual characteristics that workers bring to the market should matter for wage determination. Then, to measure what’s taking place, they set up or utilize systems, like the CPS, to collect these individual data: your age, marital status, education, occupation and earnings. Armed with this information, they crunch and recrunch the numbers to see which aspects of worker characteristics play the most important role. The struggle to produce a convincing labor supply-based story generates demand for even more detailed individual-level observation on workers. No doubt big data will be brought to bear shortly on this topic.
But what if the critical drivers of the wage structure have to do with the way work is organized in production systems? I’m thinking here of decisions regarding how much autonomy workers can have at different levels within an organization, what monitoring and incentive mechanisms are adopted, how extensive are internal job ladders, etc. All of these are affected by technology, of course, but only as refracted through organizational strategy, governance systems, market structure and the like. We know about these things mostly through case studies because systematic data are not collected on them. Anyone who has compared work organization and management across the “varieties of capitalism” knows that these matters are crucial, but it is difficult to construct formal tests in the absence of large sample data. At best, matching workers to industry-level variables like age-adjusted average tenure and capital-labor ratios can generate proxies for what we really ought to measure directly but don’t. (I have a little experience with the use of these proxies and am tempted to do more work with them.)
As for the policy implications, I think the change-in-the-nature-and-structure-of-firms story has rather radical implications. How can we change how work gets done in America?
Addendum: This past week Gerald Davis’ book Managed by the Markets, has been the main assigned reading in the class I’m teaching. It is excellent in describing, lucidly and concretely, how financialization has changed the American corporation. This is the sort of account that could be written only by someone who has devoted a career to corporate finance and governance research. It is a bit less effective when it strays into political theory and cultural criticism and also too Anglo-Saxon-centric, but at least what it says still makes sense if you know a bit about how the rest of the world is evolving.