Thread by @swinshi

Some thoughts on this week's WaPo op-ed on welfare reform by @sanfordschram @rfording and @jbsoss . Per usual, let me stipulate that I think welfare reform was a success because it reduced poverty, which is the thing we all profess to care about. /1

There are 6 points the authors make. #1 is that caseload reduction is the wrong metric for success. I agree with that, and most conservative reformers would agree with that too. (I know most of the most prominent ones personally.) But... /2

The authors punt big-time on this when they say that a better metric would be what happened to earnings from employment. I'm not sure they want to go here, but at any rate, they don't. They leave this dangling as if we just don't know the answer. We do. /3

In Table C-11 of this report (fas.org/sgp/crs/misc/R…), the Congressional Research Service finds that the poverty rate among single mothers in 2013, taking ONLY earnings into account, was only a bit higher than the OFFICIAL poverty rate in 1996 (which includes cash transfers) /4

The poverty rate including all other non-transfer income (but no transfers) was as low in 2013 as it was in 1996 INCLUDING all cash transfer income. I've yet to see anyone do anything but sputter when confronted with these stats, which are from a nonpartisan Congressional body /5

And that was in 2013! By 2016, child poverty was at an all-time low, acc to the liberal Center on Budget & Policy Priorities, and the child poverty rate PRE-TAX & -TRANSFER was only a few percentage points above the POST-TAX & -TRANSFER rate in 1996 (including noncash xfers) /6

I show that by 2014, poverty among the children of single mothers was at an all-time low after accounting for those noncash transfers (not including health care), tax credits, and changes in cohabitation. See Line 4 of Fig. 3 manhattan-institute.org/html/poverty-a… /7

But I digress. The point is that if welfare reform only reduced caseloads, without reducing poverty, it would be a failure. That didn't happen though. Welfare reform reduced poverty, period. /8 forbes.com/sites/scottwin…

Oh, & CBPP has found that all but the bottom 10 percent of single-mother families had higher income in 2012 than in 1995. /9 manhattan-institute.org/html/yes-96-we…

Here's the direct link: cbpp.org/research/famil…. They also found that increases in private income among single moms was greater than declines in welfare 1995-2005 for all but the bottom 20% of single moms. /10

Imma come back to that bottom 10/20%. For now, let's turn to the authors' 2nd point: labor force participation was rising among single mothers well before welfare reform. /11

The authors show the LFP rate, but the EPOP also began rising among single moms before welfare reform. Because the unemployed are in the labor force, the EPOP is the more relevant measure. /12

However, the single-mother employment rate doesn't appear to have been rising very fast in the decade before WR. See Figure 2 of brookings.edu/wp-content/upl… /12

From 1987 to 1996 (9 years), employment among never-married moms rose by 10 points, but from 1996 to 1999 (3 years) it rose by 15 points. /14

Like the LFP rate, the EPOP fell somewhat after the 1990s, but here's the thing: if you want to argue that the 1990s rise was about the economy, that's an argument about business cycles.... /15

You wld expect, then, that when the economy turns south, employment would return to pre-reform levels. But it never did. At its trough, in 2011, it was still about 10 points higher than in 1996. It never ceases to amaze me how people explain a secular rise with a cyclical arg /16

Point #3 of theirs is that the expansion of the EITC in 1993 "enticed more single mothers into getting jobs." That's an ambiguous phrasing--it's not clear whether they're saying that it was more important than welfare reform, or just that it was important. /17

Let me say that the EITC expansion was really important! But so was welfare reform. I should also note that the research has really only looked at the 1990s boom. We don't have good evidence since 2000. /18

The most often cited study, by Grogger (2003) has been mischaracterized a bit, but it shows that welfare reforms were about as important as the improvement in the labor market, but less important than the EITC expansion, in increasing employment among single moms. /19

I could believe that. But I also believe that the EITC expansion and welfare reform made each other more effective than either would have been on its own. The exercise of ranking which is more important is kind of fruitless. /20

The authors of the op-ed, though, have given the reader no reason to think that welfare reform was unimportant. OK, I'm counting slowly to 10, because Point #4 invokes extreme poverty.... /21

They say that welfare reform "may have" increased extreme poverty, citing Edin & Shaefer and some other research I critique in manhattan-institute.org/html/poverty-a…. Let me be clear: the extreme poverty research is garbage. Why? /22

Here are 3 reasons (though I could list more). First, better measures of income than used by Edin-Shaefer indicate much lower levels and a much smaller rise in extreme poverty. And even those probably mostly reflect bad income data. /23

You don't have to believe my CPS-based research (though you really should!). If you aren't convinced by Bruce Meyer's recent work, I don't know what to tell you aei.org/events/reexami… /24

2nd reason extreme poverty research is garbage: EVEN IF you believe the cash-income $2-a-day numbers (you shouldnt), extreme poverty begins rising 20 years before welfare reform! No one has explained to me how the 1996 reform could have initiated an increase starting in '70s /25

3rd reason extreme pov research is garbage: EVEN IF you believe the cash-income #s, extreme pov supposedly rose among groups unaffected by welfare reform, including childless households, elderly households, married college graduates. How did welfare reform do this? /26

(Incidentally, Edin & Shaefer recently replicated my results to show that extreme pov "rises" even in the CPS after mid-90s. Despite having provided comments on my paper draft, they failed to note in this work that extreme pov began rising in the 1970s & among other groups) /27

I could go on and on about extreme poverty, but you're not going to read my report or listen to Meyer's presentation, so I'll just drop it. Point #5 of the authors' is that welfare reform led to racial discrimination in some states. /28

I don't know this lit, and the authors seem to cite only their own work. I could certainly believe it. What I do know is that black child poverty is at an all-time low. Even by the official measure, it was as low in 2016 as at the end of the 1990s boom /29 datacenter.kidscount.org/data/tables/44…

Wrapping up...Point #6 is that 2/3 of adult Medicaid recipients (and most SNAP recipients) already work. Well...technically 60% of non-elderly, non-disabled adults that had Medicaid AT SOME POINT DURING THE PREVIOUS YEAR worked AT SOME POINT DURING THE PREVIOUS YEAR. Oh. /30

Here's the link: kff.org/medicaid/issue…. These academic authors cite...Vox, which cites KFF, which includes methods details at the end of a long post. /31

And we don't know how many of them could work more weeks in a year or more hours in a week. Nor do we know how many of the rest could work--we just know why they say they didn't work the (entire) previous year. /32

Key to work requirements, I've argued ad nauseaum, are the policy details--in particular, the generosity of exemptions. It's shoving your head in the sand to simply "oppose work requirements." There are certainly reasons to go slow, but opponents don't "know" what will happen /33

...any more than they "knew" what would happen in 1996. Most importantly, opponents of work requirements should really grapple harder with the evidence that welfare reform reduced poverty. Your case is not strong. And if you're wrong, you're not helping poor people. /fin

Doh-- initially forgot the link to their op-ed: washingtonpost.com/amphtml/news/m…