Thread by @arpitrage

Does auto-enrollment into 401(k) retirement accounts increase savings?

Surprisingly, a common finding is *not always.* A look at this literature:

1/

Mardrian and Shea (2001) launched the literature on automatic enrollment. Just from the summary table, you see almost 50 percent of workers got 401(k)s in the cohort defaulted to automatic enrollment; relative to previous cohorts.

Seems valuable!

ssc.wisc.edu/~scholz/Teachi…

2/

But they caution:

"Simulation results reported in a previous version of this paper ... show that default savings behavior under automatic enrollment may actually lead to lower total 401(k) savings after only a few years relative to more traditional 401(k) plans"

3/

You can see better why in follow-up work with Choi, Liaison, Madrian, and Metrick (2004)

scholar.harvard.edu/laibson/public…

4/

There is a large gap in participation rates in 401(k) plans based on whether or not you are defaulted into these retirement plans, with some catch-up over time.

5/

However, look at the savings *rates.* The default contribution rate was 2%. It looks like a lot of people saw their savings rate nudged *down* as a result of automatic enrollment. That is, they would have saved more on their own, but stuck with the 2% default chosen for them.

6/

See a fuller discussion of this in Bubb and Pildes (2014):

cdn.harvardlawreview.org/wp-content/upl…

7/

Which brings us to Taha Choukhmane's JMP, who focuses on cumulative savings

tahachoukhmane.com/wp-content/upl…

8/

First, he corrects for job-switchers. So, you could have dismissed some of the "catch-up" effect on auto-enrollees by saying that's limited to sample of people that job stayers

But people on auto-enrollment *lower* participation rate by 13% when switching to non-AE employer

9/

Second, he looks at cumulative savings rates over time. For median and richer employees -- there is substantial savings catch up over time. People who join at firms without AE don't sign up initially -- but save at higher rates when they do to the point they catch up

10/

Auto-enrollment does seem like it induces more lifetime savings for poorer workers, and his model calibrates this in more detail. But savings rates seem (to me) to be surprisingly well-captured by a neoclassical model, and less sensitive to nudges, for most workers.

fin/