It started with Ed Kilgore (now at New York Magazine)
Then original political animal Kevin Drum discussed it at motherjones.com and I threw a political cow
Now Nancy LeTourneau wrote "Let’s Talk About Welfare Reform"
She writes a lot of interesting and reasonable things (it is probably best to just click the link). She also copied Kevin Drum's graph *and* his crazy claim that a 50% increase is small. I will just steal the graph and link to Drum.
This will be my usual comment. I have written often on the topic in this comment box (because Ed Kilgore wrote something about welfare reform). The green curve Drum highlighted in the graph you posted suggests massive damage due to the 1996 bill -- the percentage of households with children living with less than $2 per capita per day has increased by 50% (=0.5 percentage points) since 1996. This is a huge increase indicating a huge number of additional people suffering third world poverty (when as is correct SNAP is included).
The curve looks roughly horizontal because the number was and is so much lower than the other numbers (say excluding food stamps/SNAP) that a huge proportional change looks tiny.
Similarly, if you had put the percentage of people murdered in a given year it would look perfectly flat (also in 2001). The graph shows that, correctly assessed, the welfare reform bill was no huge deal if and only if you consider 9/11 no biggie.
A 50% increase in poverty so severe we consider it horrible in the third world is not a minor change.
I object to somethign LeTourneau wrote "Understanding all of that, we can then talk about what an effective anti-poverty agenda would actually look like. Rather than assume that it means going back to the old AFDC model, it would include two things:" To assume that critics of the 1996 bill propose exactly repealing it and ask if that the perfect policy is to set up a straw man. You decide to ascribe one exact proposal to us (without citing anyone) and contrast it with open ended proposals. Also arguing agings "assuming" is always and invariably setting up a straw man. I don't think such rhetoric is effective and I think it lowers the level of debate.
Finally, on the comparison, I think (really guess based on near total ignorance) the recent evidence suggests that just giving cash is more efficient than giving advice via social workers. I'd consider "housing first" and the homeless and long term inter-generational effects of food stamps. The amount of money involved (say the amount to eliminate the 1.5% rate of extreme severe third world poverty) is tiny compared to the US budget.