Might Facebook boost wages?

From Luis Armona at Stanford:

In this paper, I estimate the causal effect of increased exposure to online social networks during college on future labor market outcomes.

Using quasi-random variation from Facebook’s entry to college campuses during its infancy, I exploit a natural experiment to determine the relationship between online social network access and future earnings.

I find a positive effect on wages from Facebook access during college. This positive effect is largest in magnitude for female students, and students from lower-middle class families.

I provide evidence that this positive effect from Facebook access comes through the channel of increased social ties to former classmates, which in turn leads to strengthened employment networks between college alumni.

My estimates imply that access to Facebook for 4 years of college causes a 2.7 percentile increase in a cohort’s average earnings, relative to the earnings of other individuals born in the same year. This translates to an average nominal wage increase of $3,000-$5,000 in 2014.

To be clear, some of that could be a wage distribution effect.  Still, this paper points to the possibility of some very real networking and matching gains from the use of Facebook, and perhaps those gains do not favor traditional elites.

For the pointer I thank the excellent Kevin Lewis.


It would be interesting to estimate the effect (on public welfare/ output/ or surplus) of the increased "branching factor", from all social media? And from mass communication? (We used to be 6 steps from anyone else on the planet, on average, but now we're 2.3 (iicr)). Which "the zeitgeist", it seems to me, does everything it can to work against.)

What was the effect of "the telegraph", "phones", "mobile phones", (internet access,) "social media"? A revision and review would be really interesting.

Many thanks for the great site/service. M.

'I exploit a natural experiment to determine the relationship between online social network access and future earnings'

Or, one could just look at the relationship between the various institutions and future earnings - one would expect the relationship to map fairly well. Including among female students, and students from lower-middle class families.

What Clockwork_Prior said. I was a college freshman when facebook first made its appearance and so I know that facebook's entry/exit cannot be treated as a quasi-random with respect to earnings. Facebook began at harvard, then expanded to other ivy league schools + places like stanford/MIT/CMU, before expanding into a larger set of universities.

Presuming the author is using a differences-in-differences research design, the estimates would be biased as they would essentially be calculating averaging earnings difference between Elite schools and non elite schools. If the sample is just restricted to the period where schools were simply elite, the problem still exist because facebook originated at Harvard and this becomes a comparison of Harvard earnings v.s. other schools.

"causes a 2.7 percentile increase"
Did he actually prove causation or just correlation?

He did difference in difference, interpret as desired. His assumption is

“conditional on the university exposed to the treatment (Facebook access), any particular cohort within that university is randomly chosen to participate in the introduction of the Facebook social network.”

Unemployed coal miners and farmers hurt by Trump's trade war really need to get on Facebook instead of using the Federal government as their own personal piggy bank. Lazy.

Everyone's got their hand out, hun.

Where's mine? I paid for the bank bailouts, the auto bailouts, the farmer bailouts, the China tariffs, the corporate tax cuts, the elimination of SALT, etc. Oh I see, because my guy is in office and has an 'R' after his name so it is ok to go full kleptocracy. Doesn't make it right.

All those deplatformed antisemites are giving up good money.

In his immediately prior blog post, Cowen links a study that claims privacy compliance regulations are costing tech firms billions, billions that could be invested in something productive. Now, in this blog post, Cowen links a study that claims time spent on social media isn't wasted but actually benefits the users via higher compensation due to the networking benefits of social media. Indeed, according to the prior blog post, the extra hassle to the users caused by the privacy compliance regulations is discouraging people from using social media, which, based on this blog post, will cost users dearly in lower compensation because of the loss of networking. I suppose the lesson is that privacy is way overrated and surveillance is your friend.

Privacy will be considered quaint in future. In my view no one actually cares about privacy per se; they only care about the consequences, social and otherwise, of their secrets being know. Personally I think the world CAN be much more interesting with much less privacy (e.g. today compared to when I was born) but for the the world to be good for living requires strong norms that transparency will not be “abused” which can result in a conformist authoritarian equilibrium.

Does spending time on Marginal Revolution boost my wages too?

But the wage increase, if there is one at all, does not come for free. If you spend an hour a day on social media cultivating your networking contacts, your best-case $5000 means you were earning minimum wage.

I am not surprised that improving communications and ties among college cohorts improves job search; there is a great deal of literature on the strength of weak ties.

But, you have to wonder about the magnitude of the effect: other college students at other schools were using Facebook at the same time, so their search and network opportunities must have increased as well.

Now, it could also be that the Facebook users job prospects were at the expense of non-Facebook users.

And, how does it compare to the effect of Linkd-In users.

Good stuff. I'm not entirely sold on the low impact of regulation, but my median view has certainly changed after this.

Question: How much has aggregate demand for education shifted over this time? Could the increasing value of credentialism significantly effect the price?

In this paper, I exploit the vanity and virtue signalling of Silicon Valley tech firms in order to obtain a high paying job from one...

Because there are actual returns to getting your face out there in front of people to network. If you want a top-paying Silicon Valley job, you can't signal your way out of knowing algorithms and data structures. It is an absolute requirement.

returns to conformity

The author fails to consider the opportunity cost. If those Facebook users spent the same amount of time expanding their personal
networks through more traditional methods I believe they would see truly stellar results.

I am very senior engineer and spend much of my time mentoring. My strongest recommendation to my protégés is to “network more, and track your progress to ensure you do it”.

The article is provocative and a good opening salvo, but nowhere near conclusive. It seems much more likely that the "Facebook effect" is much more about selection, simply revealing latent traits like extraversion that make one a good employee. I would put very little stock in a causal relationship.

Comments for this post are closed