Thread by @lymanstoneky

It's #NBERday and we've got some goodies!

First up:

Q: Do work requirements help welfare recipients get into jobs?

A: Probably not.
#NBERday
nber.org/papers/w28877

The big problem we have in studying welfare programs is that surveys do a crap job of identifying welfare recipients, and so we end up exploiting spatial variation in rollout and looking at "plausibly eligible" groups. #NBERday

This design is interesting but it has a big problem:

"Plausibly eligible" people may have variation over time creating all sorts of biases, and also they may not even be the "intention-to-treat" group let alone the actually-treated. #NBERday

So for example, single mothers without college degrees are a commonly studied group. But that's not *actually* the eligibility criteria for any actual program, so we get a very fuzzy match on program effects. #NBERday

This paper leverages the fact that Virginia suspended their SNAP work requirement rule for able-bodied adults 18-49 without dependents from 2009-2013, and phased it back in gradually 2013-2016.

And they use the ***actual administrative data of Virginia***.
#NBERday

So they aren't looking at "maybe welfare eligible people." They're looking at "able bodied prime-age adults without kids who actually got SNAP then were subjected to a work requirement." #NBERday

Because the requirement doesn't apply to those 50 and older, they have a regression discontinuity in there as well as the time series treatment. So this is two very compelling models in one place! #NBERday

Here's their first finding. It shows the share of these able-bodied adults who were claiming JUST BEFORE the work requirement who were also claiming in a specific period.

As you can see, similar trends pre-work rule, then BIG decline in younger groups after. #NBERday

Then they look at work. For this, they look at whether the person appeared in Virginia's UI system after the work requirements were rolled out.

Basically, work requirements did not boost employment much at all. #NBERday

Methodologically, this is one of the better papers out there on work requirements. And what it's telling us is that work requirements do not appear to motivate very much work, on the margin. #NBERday

At least among able-bodied adults without dependents, telling them, "Instead of just giving you a benefit, we will give you a benefit tied to work" does not appear to motivate very much work. Though granted SNAP also has an income phase-out in play here! #NBERday

(cc @swinshi )

Note that this is not the same as the substititon/income effects of e.g. a cash transfer, because a work requirement isn't *quite* the same beast as removing an income effect. #NBERday

The interaction of that wedge and the phase-in/phase-out rules makes calculating the income effect a bit more complicated.

But it does ***seem to suggest*** that giving people SNAP might not have a *huge* income effect discouraging work. #NBERday

This in turn suggests that a lot of the work effects of SNAP likely run through substitution effects. i.e. having/not having SNAP doesn't motivate work/nonwork, but getting more/less SNAP may motivate working more hours or fewer hours. #NBERday

Okay I said there were several but actually that's gonna be all for today; got some other stuff to do!

#NBErday