Increasing wage inequality in the U.S. is associated — mainly, if not entirely — with expanding wage gaps between companies rather than within them.

But what's causing those gaps to expand?

Are the high-paying companies doing something new and different, or are they simply outsourcing the jobs of low-wage workers? 

The outsourcing hypothesis would suggest that the widening wage gap is a red herring, because low-wage workers have merely been shuffled to another company. Indeed, news reports would suggest that domestic outsourcing — when a company has outside workers do certain jobs — has become a dominant feature of our economy. 

In his book "The Fissured Workplace," David Weil of Boston University says companies have outsourced janitorial work, security, payroll administration and other tasks that are "peripheral to their core business models," and he presents evidence of the phenomenon among such occupations. And yet when you look at analyses that attempt to measure the impact of domestic outsourcing on the economy as a whole, the results are underwhelming. 

Researchers at the Bureau of Economic Analysis, for example, found that, from 1997 to 2006, domestic outsourcing increased only modestly, to 12 percent of the economy from 10 percent.

To be sure, the increase is greater when the comparison begins in 1982 — back then, the share of gross domestic product was just 7 percent. Still, the 2006 share is presumably too low, and the increase over the past three decades not large enough, to allow the outsourcing story to be the sole explanation for the rise in between-firm wage inequality. 

Other evidence, from the Contingent Work Supplement — a survey that the Bureau of Labor Statistics conducted from the mid-1990s to the mid-2000s — indicates that, during that time, there was no major increase in the share of workers “without an explicit or implicit contract for long-term employment.” This doesn't cover all possible forms of outsourcing, however, and another analysis has noted that there is a lack of comprehensive data. 

These various imperfect data sources don't disprove the idea that there has been a big increase in domestic outsourcing, but they don't prove it either. The evidence also suggests that there's been an increase in the share of companies earning very high returns on capital, and it wouldn't be surprising if these returns partially spilled over into wages. That would be a dramatically different story — suggesting that wage inequality might be linked more to differences in company performance than was previously imagined. 

The only thing that's clear at this point is that we have work to do to figure out what's widening the wage gap.