Research by a Boston Fed economist indicates that as many as 7 million “gig workers” aren’t being counted in the main survey that measures U.S. employment. Why? And how can policymakers get a more accurate view of their contributions to the economy?