How to Improve Long-term Immigration Projections

In recent years, policy experts worldwide have come The rudimentary state of immigration projections to understand the importance of demographic projecis a cause for concern. Over the past few decades, net tions in their efforts to think strategically about longimmigration rates in most developed countries have term challenges, from national security to retirement surged, more than doubling in the United States and security. Much progress has been made in improvWestern Europe as a whole since the 1960s. This ing the fertility and longevity modules of the demosurge has occurred, moreover, during a period in graphic projection puzzle. Little progress, however, which both public opinion and immigration policy in has been made in dealing with cross-border migration most countries have grown increasingly restrictive. or (more specifi cally, from the point of view of most With undocumented or “illegal” entry growing faster developed countries) immigration. than any other type of immigration, policy experts Offi cial immigration projections, both in the are no longer confi dent that total immigration is still United States and abroad, remain largely ad hoc and subject to the effective control of national policy. judgmental. Some projection-making agencies simThe range of plausible assumptions regarding ply assume that net immigration will stay constant long-term immigration rates is therefore widening. at the current level throughout the projection period. Unbounded by any consensus projection method, this Most of the rest trend the current level until it reaches widening range can generate a similarly widening a “target” or “ultimate” level, which is typically based and often dramatic variety of long-term population on the historical average over some recent period. A outcomes. The spread between the “low” and “high” few agencies explicitly build their projections around immigration variants for the U.S. Census Bureau current national immigration policy. When describprojection for the national population in 2100, for ing how they make assumptions, most agencies offer example, is 417 million — from a total of 438 million little more than a vague reference to “expert opinion,” in the low variant to a total of 854 million in the high “national policy,” or “historical experience.” Few if variant (see Figure 1). This is a very signifi cant difany use assumptions that are justifi ed by any explicit ference from any policy perspective. reference to a theory of how or why immigration happens.