Welfare Reform, 1996: How Dependency Beat Poverty

Accessible text version of the Week 9 Monday slides — same content in linear order, with no slide-scaling container. Asynchronous week. Reading: Béland and Waddan (2012), “Welfare Reform, 1996,” ch. 1 of The Politics of Policy Change.

POSC 459 — Social Welfare Politics and Policy Week 9 · Monday, October 19 · Asynchronous: no class meeting

This deck walks you through one reading: Béland and Waddan, “Welfare Reform, 1996.” Page numbers on each slide are from that chapter.

This week, no meetings

Wednesday’s deck posts Wednesday. Questions? Email me. We meet again Monday 10/26.

How to use this deck

1. Each slide starts with a plain sentence saying what happened, then the details, then a shaded box saying why it matters for the chapter’s argument.
2. If you read only the shaded boxes, you’ll have the argument. If you read the whole deck, you’ll have the evidence.
3. Slides marked → paper hold claims, with page numbers, you can use in Discussion Paper 3.
4. Three glossary slides come next. Come back to them whenever a term is unfamiliar.
5. The last three slides set up the film and the paper.

The chapter answers one question: why could Congress abolish welfare (AFDC) in 1996 when it has never been able to cut Social Security or Medicare?

Glossary 1: the programs and laws

Glossary 1, continued: earlier attempts and tools

Glossary 2: the ideas the authors use to explain it

The law, in one slide

What happened: In August 1996 President Bill Clinton signed PRWORA. It abolished AFDC, a program that had existed since 1935, and replaced it with TANF.

Why it matters: In 1996, 4.5 million families were on AFDC, nearly all single mothers with children. This was the only time a major federal benefit program was repealed outright. Social Security and Medicare have been trimmed and adjusted. AFDC was ended.

pp. 24–25, 61–62

The puzzle

A Democratic president signed rules that “would have been denounced as . . . right-wing extremism a quarter century earlier.”

p. 24

Why welfare and not Social Security? Why 1996 and not 1970, 1977, or 1988? How did a Democrat end up signing a Republican bill?

The authors’ answer, in advance

Béland and Waddan explain 1996 with four things working together. Keep them in mind; every slide feeds one of them.

1. A weak policy legacy. AFDC was never “earned,” so nobody powerful defended it.
2. Issue ownership moved. Voters came to trust Republicans, not Democrats, on welfare.
3. The idea changed. The problem stopped being poverty (too little money) and became dependency (welfare harms people).
4. The institutions finally lined up. Earlier reforms died at veto points. In 1996 a disciplined Republican Congress met a Democratic president who had promised reform and couldn’t afford to veto it.

Why it matters: The same four things explain why Social Security survived. Strong legacy, Democratic ownership, no reframing, and veto points that held.

pp. 68–71

1935: AFDC is born the weaker sibling

What happened: The 1935 Social Security Act created both Social Security and AFDC. They were built on different principles from the first day.

• Social Security was paid for by workers’ payroll taxes, so recipients could say they had earned it. AFDC was paid from general taxes to people in need. No one could claim to have earned it.
• AFDC was assigned to the Social Security Board, “a male dominated agency” that favored insurance programs and neglected aid to the poor.
• Southern members of Congress stripped out national standards. States decided who was eligible and how much they got. Most Black mothers in the South were excluded.
• Eligibility turned on “worthiness”: widows yes, others investigated, including midnight raids to check for a man in the house.

Why it matters: This is the policy legacy. From birth, AFDC lacked the “I earned it” claim that protects Social Security, and its rules varied by state. Both facts come back in 1996.

pp. 26–27

The 1960s: the rolls triple, and no politician takes credit

What happened: AFDC grew from a small program into an entitlement, mostly because courts and agencies opened the doors, not because Congress voted to.

• Applicants approved: one in three (1960) → 90% (1971). Cost: $1 billion (1960) → $6.2 billion (1971).
• The Supreme Court struck down state rules that denied aid to newcomers, to mothers with a “substitute father,” or after a “man in the house” raid.
• Nobody in Congress campaigned for this expansion. Nobody claimed credit.

Why it matters: Because unelected judges and bureaucrats expanded welfare, conservatives could later say its defenders were “out of touch with the values of most Americans.” The expansion gave AFDC more recipients but no political champions. Weak legacy again.

pp. 29–30

The road not taken: Nixon’s guaranteed income

What happened: In 1969, Republican president Richard Nixon proposed the Family Assistance Plan (FAP), which would have guaranteed every family a minimum income. It nearly passed.

• A “negative income tax”: a family of four with no income would get $1,600 a year, shrinking gradually to zero at $3,920. Aid depended on income, not on why you were poor.
• Passed the House 243–155 in 1970. Died in the Senate Finance Committee, 10–6.
• Killed from both ends. Conservatives (Reagan, Agnew) called it a giveaway. The National Welfare Rights Organization, speaking for recipients, called $1,600 an insult and demanded $5,500.
• It would have remade the South. Mississippi paid $46 a month; a quarter of Louisiana would have qualified. One Georgia congressman: “There’s not going to be anybody left to roll these wheelbarrows.”

Why it matters: This shows the veto point at work: a bill with majority support can die in one committee. And it shows how far the debate moved. In 1970 a Republican proposed guaranteeing income. By 1996 both parties demanded work requirements.

pp. 31–35

Why the dead plan still matters: worthy vs. unworthy

What happened: Nixon tried again in 1971 (HR 1), and Carter tried in 1977 (the PBJI). Both died. What Congress passed instead was the EITC (1975), a benefit only for people who work.

• FAP sorted people by income. It didn’t ask why you were poor. AFDC sorted people by worthiness (widow? working? married?).
• Liberals at the time dismissed FAP as “nothing new, nothing startling.” The authors say it was both. Liberals assumed welfare policy was theirs to set and missed what was on offer.
• The EITC helps the working poor and gives nothing to the non-working poor. It cost more than AFDC by 1996, and it reinforced the deserving/undeserving line instead of erasing it.
• 1972 Republican platform: “We flatly oppose . . . a government-guaranteed income.” The Democrats attacked Nixon’s “forced work requirements.” By 1996 both parties were for forced work requirements.

Why it matters: The lasting American choice was to help the poor who work and judge the poor who don’t. Every later reform built on that line. 1996 was its sharpest expression.

pp. 34–36

Drift: the program stood still while the country moved

What happened: AFDC was designed in 1935 so poor mothers could stay home with their children. By the 1980s most mothers worked, and the poor had become mostly female-headed families.

Also: by the 1980s most AFDC mothers were not widows but divorced, separated, or never married.

Why it matters: This is drift. Nobody changed AFDC, but as society changed, a benefit that let mothers stay home came to look like paying people not to work. Heclo: “a program that stays the same while the society around it is changing can actually amount to a transformed policy.”

pp. 37–38, 58

The battle of ideas, 1981–1987: the problem gets a new name

What happened: A series of conservative writers argued that welfare itself caused poverty, by rewarding not working and not marrying. Their ideas moved from the fringe to the mainstream in six years.

Why it matters: This is the idea shift. The question changed from “how much do the poor need?” to “what is welfare doing to the poor?” PRWORA ended up looking like Mead’s plan (require work, keep government involved), not Murray’s (abolish everything). The 1987 Seminar “uncannily anticipated” the 1996 law.

pp. 38–41

The liberal response: “an uncertain trumpet”

What happened: Liberals disputed the conservative story but never produced a winning counter-story. Their own leading expert ended up proposing time limits.

• William Julius Wilson (1987): the cause of welfare growth is the disappearance of jobs from inner cities, not the values of the poor. Widely praised, but he admitted liberals “lacked a convincing rebuttal” to conservatives.
• David Ellwood, Poor Support (1988): a liberal who proposed time limits on cash welfare, paired with a guaranteed government job for anyone who couldn’t find work.
• Ellwood joined the Clinton White House and helped write Clinton’s 1994 bill. When the final 1996 law kept his time limit and dropped his job guarantee, he resigned and opposed it.

Why it matters: When the opposition’s own expert accepts your premise (that welfare should end at some point), the argument is half over. Liberals ended up debating how to limit welfare, not whether.

pp. 41–43

The Reagan years: cuts, speeches, and a halfway reform

What happened: Reagan cut AFDC in 1981, attacked “dependency” in speeches for eight years, and signed a compromise reform in 1988 that kept AFDC but added work rules.

• 1981 (OBRA). Tighter income and asset limits: about 400,000 families dropped, 300,000 more had benefits cut. Working while on AFDC stopped paying after four months, so fewer recipients worked. That shrank the program’s support further.
• Speeches. 1986: the “spider’s web of dependency.” 1988: welfare “created a poverty trap . . . Dependency has become the one enduring heirloom.”
• 1988 Family Support Act (FSA). States must put 20% of recipients in a training program (JOBS) by 1995. Two-parent families covered everywhere. Twelve months of Medicaid after leaving welfare. Cost: $3 billion over five years. Passed with big bipartisan majorities.

Why it matters: The FSA was “full of compromises,” but its conservative half lasted and its liberal half didn’t. It put the principle of required work into law. The authors call it “a staging post rather than the final destination.”

pp. 43–46

The 1990s “crisis”: public opinion and who owns the issue

What happened: Right after the 1988 reform, the rolls surged again. The public had turned hard against welfare, and voters now trusted Republicans on it.

• Polls, 1994: 89% want a two-year limit followed by work. 76% say a mother who has another child on welfare should get no increase. 75% say welfare makes people “too dependent.”
• Martin Gilens: after the mid-1960s the news media pictured welfare recipients as Black. “The symbolic power of welfare . . . stems in large measure from its racial undertones.” (Gilens is a Week 12 reading.)
• A Clinton budget official conceded the public “believe[s] that poor people should play by the same rules as the middle class. When you don’t work, you don’t get paid.”

Why it matters: Issue ownership had flipped. Democrats built welfare, but by 1992 “welfare” was a Republican issue, a word for a problem rather than a program. A Democrat who wanted to win had to run against it.

pp. 46–48 · → paper

October 23, 1991

“In a Clinton administration we’re going to put an end to welfare as we know it.”

Candidate Bill Clinton, repeated for a year. A Democratic pollster called the line “pure heroin.”

What he was doing: stealing the issue back. By attacking welfare, a Democrat could prove he was not an old-style liberal. It worked as a campaign line. It also committed him to a result.

pp. 48–49

Clinton’s own bill, 1994: big promise, small plan

What happened: Clinton’s administration finally wrote its own welfare bill, the Work and Responsibility Act (WRA). It was tougher than any Democratic bill before it, and it went nowhere.

• The idea: two years of cash aid, then a required community-service job. Teen mothers must stay in school. Mothers must name the father.
• But it had to be paid for without new taxes. The White House found $9.3 billion over five years, mostly by cutting aid to legal immigrants. Enough to move only 170,000 of 5 million adults into subsidized jobs by 1999.
• Inside fight: Ellwood wanted help to continue for anyone making a real effort. Bruce Reed, who coined “end welfare as we know it,” wanted a hard end date.
• Liberals attacked it in hearings; Newt Gingrich mocked it: “brilliant at describing a Ferrari, but his staff continues to deliver a Yugo.” The budget, NAFTA, and health care crowded it out. It never got a vote.

Why it matters: Clinton had promised to end welfare and had not. That left the issue open for Republicans to claim after they won Congress in November 1994. The authors’ point: a promise without a law is an invitation to the other side.

pp. 49–51

State waivers: building the new system inside the old one

What happened: Through the early 1990s, Washington let states run AFDC under their own experimental rules. By 1996 more than half the states were doing it, and the experiments became the evidence for national reform.

• Reagan, 1988: “a thousand sparks of genius in 50 states.” Clinton, 1993, promised to approve waivers even when he disagreed with them.
• Riverside County, California: told recipients to take “a job, any job,” now. It moved more people into work than the five counties that stressed education and training first. “Work first” became the national lesson.
• Wisconsin, under Gov. Tommy Thompson, went furthest: time limits, sanctions, a plan to end the entitlement at the state level. Thompson’s spokesman on block grants: “No longer would we have to go to Washington and kiss somebody’s ring.”

Why it matters: This is layering. Nobody repealed AFDC, but state experiments stacked new rules on top of it until time limits were already real somewhere. Next slide: what the experiments actually showed.

pp. 51–53

What the waiver experiments proved, and what they were said to prove

What happened: Every waiver came with a research requirement, so by 1996 there were dozens of studies. They did not agree.

• The results were “mixed and inconclusive.” Income gains were small. Many people who found work stayed poor enough to keep getting welfare.
• Liberals read the studies as: moving people to work is slow, so change gradually. Conservatives read them as: change works, so change dramatically.
• The studies that got attention were the ones that fit the conservative reading, Riverside above all.
• Wisconsin’s political situation made bold reform easy there. Whether it could be repeated nationally was never tested. It became the model anyway.

Why it matters: The evidence was ambiguous; the political use of it was not. Rogers-Dillon: pilots “made time-limited welfare a reality and took away the argument that it could not be done.” Mead: “Without Wisconsin’s example, PRWORA would have been less radical.” For your paper, this is a claim about how evidence gets used, with a page number.

pp. 52–55 · → paper

Fifty programs before there were fifty programs

What happened: Even under one federal law, AFDC paid wildly different amounts depending on the state. In 1995 a family’s AFDC income, as a share of the poverty line:

Why it matters: The variation came first; block grants in 1996 made it official. When Campbell describes “fifty different worlds of social assistance” (Wednesday’s reading), this is where they come from.

p. 52

1995: the new Republican Congress writes the bill

What happened: Republicans won Congress in November 1994 with welfare reform in their Contract with America. In 1995 they passed a bill far to the right of Clinton’s, and Clinton endorsed the Senate version.

• The Contract’s bill: a five-year lifetime limit, 35 hours of work a week after two years.
• Pressure from the right: the Heritage Foundation’s Robert Rector (author of this week’s recommended reading) attacked the leadership’s first draft as not “real reform.” The pattern: denounce the moderate bill, write a tougher one, lobby until the tougher one is the standard.
• Block grants won over Republican governors: a fixed sum, fewer federal strings. Frozen grants looked fine because the rolls were already falling.
• One overreach: Gingrich’s talk of orphanages for the children of mothers who hit the limit. Democrats won that round, and he dropped it.

Why it matters: The bill was written by the people who had spent fifteen years renaming the problem. Block grants did the institutional work: once states got a fixed sum, there was no entitlement left to defend.

pp. 55–57

1995: the Democrats’ answer is almost the same bill

What happened: Rather than defend AFDC, House Democrats offered their own time-limited bill. It lost, and most Senate Democrats then voted for the Republican bill.

• House Democrats’ alternative (the Deal bill) had a four-year lifetime limit and let states adopt family caps. By any earlier Democratic standard, “an extraordinarily conservative measure.” It lost anyway.
• House Republican bill passed 234–199, March 1995.
• The Senate wrote a more moderate version (more child-care money, family cap as a state option, no ban on aid to teen mothers). Clinton endorsed it. It passed 87–12 in September 1995.
• After that vote the authors call it “clear that any final compromise bill . . . would be at least as conservative as the Senate bill.”

Why it matters: Hard time limits and the end of the entitlement were now the consensus of both parties. The only open question was how far right the final bill would land, and whether Republicans wanted to give Clinton a bill to sign or a bill to veto.

pp. 58–60

Two liberals hand Republicans a number

What happened: In 1994 Mary Jo Bane and David Ellwood, both liberal scholars working for Clinton, published a new way of counting how long people stay on welfare. Republicans used it.

All three numbers are true. Most people who ever use welfare use it briefly. But on any given day, most of the people on the rolls are in the middle of a long stay, because the long-stayers pile up.

Why it matters: Week 1 again: flow is not stock. “Welfare is temporary for most users” and “most current recipients are long-term” are both correct. Which one you quote depends on which question you’re asking. Republicans quoted 48%.

p. 58 · → paper

Two vetoes, then a signature

What happened: Clinton vetoed welfare reform twice, in December 1995 and January 1996. In August 1996, facing re-election, he signed the third bill.

• Veto 1: welfare was buried inside a huge budget bill. Clinton fought that bill over Medicare cuts and won the government-shutdown fight. Welfare wasn’t the point.
• Veto 2: the January bill also cut food stamps and school lunches.
• What changed by summer: Tommy Thompson took over the governors’ association. Bob Dole left the Senate to run for president. Republicans feared losing Congress and wanted a law to campaign on. So they sent a stand-alone welfare bill with nothing else in it to veto. > Why it matters: The first two vetoes were about other things in the bill, not about welfare. Republicans learned the lesson and removed everything else, which left Clinton nowhere to hide.

pp. 60–61

August 1996: the signature

What happened: On August 1, 1996, Clinton’s advisers met to decide. The policy staff said veto; the political staff said sign. He signed.

• Urging a veto: HHS Secretary Donna Shalala, chief of staff Leon Panetta, and most policy advisers.
• Urging a signature: Bruce Reed, the political advisers, and Hillary Clinton.
• Pollster Dick Morris told Clinton a veto would turn “a fifteen-point win into a three-point loss.” George Stephanopoulos: Clinton’s “heart urged a veto, while his head calculated the risk.”
• Once the president said he would sign, congressional Democrats saw no point standing to his left.

Why it matters: The veto point that killed reform in 1970 (a Senate committee) and 1994 (the president’s own party) gave way because the president had promised reform and the bill was built so a veto would cost him the election. Institutions and ideas together.

pp. 62–63

Two senators from New York, same day

“This bill is not welfare reform, but welfare repeal. . . . Do not doubt that Social Security itself . . . will be next.” — Daniel Patrick Moynihan, Democrat

“To continue business as usual . . . would have been . . . a conspiracy to continue to keep our people on that narcotic.” — Alfonse D’Amato, Republican

Moynihan was wrong about Social Security. The authors spend the rest of the book explaining why.

p. 63

Did it work? Depends on which column you read

What happened: By 2000 supporters called reform an unqualified success. Critics pointed to who had been left out. Both used real numbers.

A 2005 law (the Deficit Reduction Act) made the work rules harder to meet, so states had even more reason to keep families off the rolls.

Why it matters: “Fewer people on welfare” and “fewer poor people” are different claims. Rank’s Section IV (Wednesday) asks the same question with more data. Your paper can use either column, but say which one.

pp. 63–68 · → paper

Nobody came back for it

What happened: Once passed, welfare reform became permanent. Neither party has tried to reverse it.

• 2000: Al Gore attacked a Democratic rival for voting against PRWORA.
• 2002–05: the only fight was whether to require 40 hours of work a week instead of 30. Senator Hillary Clinton proposed 37.
• 2008: Senator Obama called it “an imperfect reform” that erased the line between “the deserving poor and the undeserving welfare poor. . . . Now, everybody was poor, and everybody had to work.” Hillary Clinton: AFDC “did not work.”
• 2009: the stimulus added a $5 billion TANF emergency fund. Senate Republicans let it expire in 2010.
• Paul Krugman, a critic: the law caused “a great deal of hardship,” but it “took that issue off the table” and drained the race politics out of welfare.

Why it matters: Some Democrats had hoped ending welfare would let the public hear a liberal case for the poor. Soss and Schram (the grad reading) checked and found public opinion didn’t move. The law settled the argument by ending it.

pp. 64–68

The answer: why welfare, and not Social Security

The chapter in one sentence: AFDC fell because it was weakly defended, politically re-owned, intellectually renamed, and finally caught at a moment when the institutions that usually block reform all pointed the same way.

pp. 68–71

Carry this into the film

Two American Families follows two working-class households in Milwaukee, Wisconsin from 1991 to 2024. You just read that Wisconsin was the state that led welfare reform. These families lived through it.

As you watch, look for:

1. Work first. The reform’s rule was “a job, any job.” What do those jobs look like inside one family? What do they pay?
2. The 40%. The chapter says most families poor enough to qualify for TANF don’t get it. Do you see a moment where a family needs help and doesn’t get it, or doesn’t ask?
3. Drift. Who in the household works? Does the system notice?
4. Flow and stock. Is this family a short spell of hardship or a long one?
5. The 44%. Welfare rolls fell 44%. If these families left the rolls, does that count as success?

Two households are not national data. That gap is step 4 of your paper: which kind of evidence fits the claim you picked?

Discussion Paper 3, due Friday 10/23, 11:59 p.m.

About 600 words. One moment from the film, one claim from the readings. The full prompt is on Canvas.

1. One moment from the film: a specific scene, person, or decision I could find.
2. One claim from the readings about what reform did, with a page number. Claims from this chapter you could use:
   • Welfare rolls fell 44% between 1996 and 1999 (p. 63)
   • Families getting aid, of those eligible: 84% in 1995, 40% in 2005 (p. 67)
   • “Work first” moved more people into jobs than training did (p. 53)
   • On any given day, 48% of recipients are in a stay of ten years or more (p. 58)
   • In the 2008 recession, unemployment doubled but caseloads rose only 13% (p. 68)
3. Does the moment support or challenge the claim? Show how, with detail from the scene.
4. Which do you trust more on this point, the film or the reading, and why? Say which kind of evidence fits the kind of claim.

Wednesday’s deck adds claims from Campbell and Rank.

Wednesday’s deck

Campbell, chs. 4–5: how means-tested programs keep people poor, and why the rules differ in fifty states

Poorly Understood, Section IV: does welfare work?

Deck posts Wednesday. Film by Friday. Paper Friday night.

Email me with questions. Monday 10/26 we’re back in the room: health care reform and the ACA.