Without geographic analysis, demographers often overlook keen insights into fertility trends. The same caution applies to economists. Here's Harvard economist Edward Glaeser overlooking keen insights into Boston's talent demography in a 2010 op-ed:
MASSACHUSETTS’ GREATEST natural resource is its stock of 535,000 college and graduate school students. Human capital brings the ideas and entrepreneurship needed for regional success, yet too many of our students leave, including the entrepreneurs who created Facebook. Retaining talent requires us to fight the regulations that make entrepreneurship too rare and housing too expensive, but the state should also aim at winning students’ hearts while they are still in school. ...
... The high cost of housing is the bad part of college life. Dormitories can be more expensive than apartments, but undergraduates who choose to live in normal neighborhoods can create plenty of conflict with other residents. The natural solution is to build more dedicated college space, but that’s financially impossible for many educational institutions.
One vision is to explore private interest in building a student-city somewhere in Greater Boston. Would a consortium of private developers and colleges be interested in erecting large amounts of dormitory space if they could also put in connected retail space and bypass local land use controls? If a collection of builders were willing to deliver dormitories, then they would also have an incentive to make the experience pleasant. A collective student-city would give students a sense of place and lead to more regional identity.
Glaeser is famous for offering up deregulated housing construction as a panacea to the demographic decline in the Frost Belt. If the rent wasn't too damn high, then college graduates wouldn't leave in droves. Glaeser believes that such policy drives migration. Tight zoning in Boston pushes people to zoning-liberal Houston, Texas.
Glaeser is wrong for two reasons. First, "research suggests that the availability of jobs (not housing costs, not amenities) is the most important fact in retaining recent college graduates." I've read at least a dozen of Glaeser's papers. He doesn't have a good command of migration theory. He also conflates population change with net migration. I don't put much stock into his policy recommendations. Second, Glaeser doesn't understand the geographic complexities of Boston's labor market:
Boston’s colleges and universities constitute an “export” cluster. The region attracts students from all over the country and the world and, after several years, exports college graduates to the national and global labor market. This process provides a large in-flow of dollars and jobs, all related to the global market rather than to Boston’s economy. That is, Boston’s higher education sector is driven by worldwide student recruitment and not by Boston’s labor market needs. The number of students recruited, the number who graduate, and the set of skills they graduate with do not necessarily conform to the needs of Boston’s labor market.
Glaeser has antiquated ideas about the role of Boston's higher education cluster in the regional economy. The universities don't exist to satiate the demand of local employers. Instead, the likes of Harvard and MIT staff the global labor market. If you want to join the lucrative ranks of the global labor market, then you will want to attend Harvard or MIT. Boston has a tremendous competitive advantage in the higher education marketplace, one that would intrinsically inform a lower retention rate of college graduates.
If an individual wants the greatest return for her talent, she should strive to be a member of the global labor market and will likely leave the area. Many universities, as a matter of mission statement, provide a path to the global labor market. Yet the host community expects local institutions of higher education to skill-up local labor for local industry. As far as students are concerned, that's a lousy return on a tuition investment. People develop, not places. Workforce development shouldn't be place-centric.