Next week at the Canadian Economics Association meetings I'm discussing a provocative paper by Andriana Bellou on the The Impact of Internet Diffusion on Age at First Marriage.

The theory behind the model is straight from the e-harmony ads: "I decided to give eHarmony a try after several frustrating attempts to find someone to love, admire, cherish and respect for the rest of my life." People delay marriage because it takes time to find Mr. Right. The internet means women can find Mr. Right faster, so they marry younger.

So far, so plausible. What actually happened to the age at  first marriage around the time that the internet was introduced? Broadband access grew rapidly between 1999 and 2006. For comparison, here is some information I downloaded on the median age at first marriage:

The median age at first marriage steadily rose throughout the 70s and 80s, then began to level off in the mid- to late- 90s, just around the time the internet was introduced. Could it be because people are meeting partners on-line? It could. Or it could be because of abstinence education (when people decide to delay sex until marriage, they marry young), economic prosperity, a slow down in the growth of income inequality, or demographic change. How do we know it's the internet?

In her paper, Bellou takes advantage of the fact that there are 50 different states in the US, and broadband became available at different times in different states. If the internet is helping people find partners, the number of singles should drop when the number of internet connections grow. That's exactly what Bellou finds.

Perhaps there's something different about the states with lots of broadband (Washington, DC) as compared to states without much broadband (Montana)? Bellou controls for state characteristics like per capita income or the ratio of men to women in the state, does other things to control for the possibility that a third factor is driving both broadband adoption and singleness rates, and still finds the internet matters. It's pretty convincing.

But I'd still like to see more evidence. Some marriage markets are thicker than others. This ad paints a picture:

I am a SWJM professional, looking for petite SWJF who can make a candle light dinner, enjoys a midnight swim in my pool, watch the sunset, hold hands and likes to hike.

If you're in a niche marriage market – in the minority with respect to your religion, ethnicity, sexual orientation or… – it's harder to find the right person. So the internet should make more of a difference for those particular sub-groups of the population who might otherwise find it especially difficult to find someone.

Another thing I'd like to know more about is what is happening to men. Bellou finds that the probability of women 21-30 being single falls when broadband is introduced, but finds less an effect on the probability of singledom for men in that age group. Perhaps this is because she's looking at the wrong men – since women generally marry men who are older than they are, it would make sense to compare 21-30 year old women with 24-33 year old men.

Or perhaps something else is going on. One of the most successful internet dating sites is ashleymadison.com: "life's short, have an affair." Are more young women in relationships in part because more older men are getting divorced (and forming new relationships with younger women)?

The other social change that happened in the late 90s was the introduction of Viagra. Perhaps, like the ads suggest, Viagra allows older married couples to strengthen their relationships. Or perhaps it allows older men to initiate new relationships. Even if men in their 60s are starting new relationships with women in their 40s, there will be a knock-on effect on younger women as men in their 40s, with less women available in their own age group, start looking for women in their 30s, men in their 30s start looking for women in their 20s, and so on. 

The paper Bellou is presenting at the CEA meetings is one on a series of papers coming out of her doctoral research on internet diffusion. It's an exciting research agenda – and it's exciting to have a new and imaginative scholar join the Canadian economics community.