College graduates of the new millennium are different than previous generations. Not just because they prefer Snapchat to email and have mountains of school loans, but also because of their choices of where to live.

In the past, several factors such as the proportion of a city’s workers who are college educated, job prospects, income levels, and city amenities have influenced college graduates’ decisions on where to live.

So, are college graduates looking for the same things as they did in the earlier decades? Or has a changing world affected what attracts educated workers?

These choices matter to city officials, who work hard to attract the next generation of skilled workers through the lure of exciting new entertainment districts or by becoming a hub of a trendy new industry.

As a regional economics researcher, studying the migration of highly skilled workers, I found that there have been changes in how college graduates choose where they go to work, which has implications for the health of the national economy.

Grads in the 1990s

Previous research found that in the 1990s – after controlling for lots of other city characteristics like population, income and amenities – the proportion of the population with a college degree increased more in cities that already had lots of college grads.

For example, small college towns like Boulder and Ann Arbor had some of the country’s highest proportions of populations holding a bachelor’s degree – around 42 percent of the population at the start of the 1990s.

These towns ranked first and eleventh in the population of college graduates. Over the next decade, they added 10.2 and 6.2 percentage points to their respective totals.

There are many reasons why highly educated workers might be attracted to towns with a high population of college graduates.

One, there is evidence that productivity from better educated workers can “spill over” to other workers, enabling them to earn higher wages. Two, better educated places tend to offer more cultural amenities and entertainment options, which college graduates value highly.

If this trend goes on long enough, it could lead to segregation according to education level across cities, as cities with better educated workers will continue to attract more educated workers. In fact, one prior study suggests this is exactly what happened in the three decades between 1970 to 2000.

We were interested in seeing whether these trends still held, or if different factors were now luring graduates. So what is this generation of graduates looking for?

Large cities attract grads

A lot has changed for colleges graduates of the new millennium.

After all, in the last decade there have been two recessions (including the Great Recession in 2008), a continued decline in overall interstate migration, and significant industry restructuring from the loss of so many manufacturing jobs. Other studies have highlighted the trend of college-age adults living with their parents longer, providing some evidence that the struggling economy was swaying behaviors.

We wanted to see if the same economic forces had also changed where college graduates were moving.

We used data from the U.S. Census on the share of population who are college graduates in 358 Metropolitan Statistical Areas in the United States for two separate decades: 1990-2000 and 2000-2010. We controlled for many other city characteristics like size, median income and natural amenities.

We found it was not necessarily the better educated cities that attracted graduates as in the 1990s, rather it was the large cities (holding city education levels constant) that attracted graduates post-2000.

Bigger cities usually have more diversified economies and therefore pose less risk of unemployment. It could be that graduates preferred places where they would increase their odds of landing any job during a time when the overall economy was struggling.

In previous decades graduates might have risked failure by chasing down that dream job in a city that was less diversified but more risky. Take Des Moines, Iowa, for example. It is a city of about 200,000 people that is highly specialized in the insurance industry. The percent of the population who were college grads increased six percent in the 1990s, but had slowed to less than four percent post-2000.

We believe, that because economy was so strong in the 1990s, graduates were willing to take those risks. So they went to cities with industries that employed highly educated people and were growing fast, but which also were more risky. If things didn’t work out, they believed they could always find a decent job somewhere else.

In the first decade post-2000, much of that perceived job security had eroded.

What can policymakers do

Our research is important because it further refines models of regional college graduate growth and updates our understanding of how the mix of factors influencing city college graduate growth has changed in the new millennium.

The link between education and economic growth is well-established and intuitive. Better educated workers are more productive in their jobs and more likely to start new businesses that create even more jobs. College graduates are more productive workers, have higher civic engagement, and help create new businesses.

Because of these things, our results have important policy implications for decision-makers aiming to lure talented workers to their cities. While city officials can do little to make their cities quickly grow in population, they can enact policies such as investing in education and supporting local entrepreneurs that create more stable labor markets.

Author Bio: Michael Betz is an Assistant Professor of Human Development and Family Science at The Ohio State University