Thread by @san_muly

How helpful are socially connected parents for young workers’ who are entering the labor market? In my #JMP, I use administrative data from Israel to find out.

Website: sites.google.com/view/mulysan

#EconTwitter #EconJobMarket

A thread

1/9

Social connections could help by

(1) Alleviating frictions by improving information flows about job openings

(2) Increasing the probability of a match, conditional on having information

I build a matching model to separately identify these mechanisms

2/

For identification, I compare: (a) firms where a young person has an existing parental contact during their labor-market entry year, with (b) firms where the young person *used to* have a parental contact but left shortly before their labor-market entry year

3/

Data suggest parental links matter for young workers:

(1) workers are 3.7 times more likely to find employment in firms with (weak) parental connection than in otherwise similar firms

-->

4.1/

(2) workers' probability of starting at a particular firm discretely falls the year after their parental connection leaves to work somewhere else

4.2/

I then build a model in which matching takes place in two steps:

(1) Workers and firms meet randomly

(2) Workers and firms who have met choose their optimal match

Both the probability of meeting and match utility can vary as a function of connections

5/

I estimate the model using two distinct types of information: where do young entrants end up working, and how much they are paid

6/

Model shows that both mechanisms are important. Parental connections increase:

(1) The meeting probability by 115%

(2) The likelihood of being hired by 35% (conditional on a meeting)

7/

Counterfactuals show that parental connections matter for inequality.

The wage gap between Arabs and Jews decreases by 12% when equalizing the groups' connections

-->

8.1/

However, the gap Increases by 56% when prohibiting the hiring of connected workers.

These results are explained by the fact that Arabs rely more heavily on connections but have connections to lower-paying firms

8.2/

The full paper is here:

mulysan.github.io/San_JMP.pdf

Thanks for reading! Any comments are much appreciated.

9/9