Moffitt 2015 — The Deserving Poor, the Family, and the U.S. Welfare System

Summary

Moffitt's 2014 Population Association of America (PAA) Presidential Address documents that U.S. means-tested welfare spending grew continuously and substantially from 1970 to 2007 — it was not cut. But the money was redistributed across demographic groups in three major shifts: toward the elderly and disabled, toward married-parent families and away from single-parent families, and toward the near-poor and away from the deeply poor. These redistributions trace to centuries-old Anglo-American "deserving poor" distinctions that politically privileged workers, married adults, and those with children over nonworkers, single parents, and the childless. The result is a welfare state that has largely abandoned its original constituency — the deeply poor, especially single mothers — while expanding dramatically for other groups.

Key Claims

• Aggregate spending grew, not fell: total real per-capita means-tested spending rose +60% (1970–1975), +25% (1975–1986), +72% (1986–2007); the "shredded safety net" narrative is factually wrong.
• Three redistributions, 1983–2004 (Ben-Shalom, Moffitt, Scholz 2012, Survey of Income and Program Participation [SIPP] data):
  1. By demographic group: elderly gained +$208/mo (+19%), disabled +$74/mo (+6%), non-elderly nondisabled +$20/mo (+13%).
  2. By family structure: single-parent families lost 20% of benefits; married-parent families gained 68%; childless families unchanged (+7%).
  3. By poverty depth: deeply poor (< 50% poverty line) lost substantially — single-parent deep-poor −35%, married-parent deep-poor −31%; near-poor and nonpoor gained 73–138%.
• Mechanism — program mix shift: Aid to Families with Dependent Children/Temporary Assistance for Needy Families (AFDC/TANF) collapsed to 1/4 of its 1995 level by 2007 (lower in real terms than in 1970). The Earned Income Tax Credit (EITC) grew +274% (1988–1998); the Child Tax Credit (CTC) created 1997 — both require earned income, structurally excluding the nonworking deepest poor. Supplemental Security Income (SSI) grew +80% (1990–1995) but serves only aged/blind/disabled. Medicaid grew +210% (1986–2007).
• Deserving poor thesis: the redistributions match a centuries-old Anglo-American ideological typology in which the "deserving poor" are those who work, are married or widowed, and have children. The "undeserving poor" are nonworkers, single parents, and the childless. These categories drive political support and withdrawal for programs.
• AFDC caseload composition change: the share of AFDC cases headed by widows fell from 59% (1942) to near zero; the share headed by never-married mothers rose to 58% (1992). This shift — not fiscal pressures — is the most important driver of political support for the 1996 welfare reform (the Personal Responsibility and Work Opportunity Reconciliation Act, PRWORA).
• Welfare dependency stigma operates independently of income: the political narrative stigmatizes receipt of welfare benefits per se, not just poverty, making even moderate-income recipients vulnerable to political attack.
• Work/family disincentives are too small to explain the redistributions: labor supply disincentive effects from AFDC/EITC are small in magnitude and would not change the redistribution findings; work disincentives may have actually decreased over this period as benefits for nonworkers fell.
• Great Recession 2007–2012 temporarily equalized the distribution: the Supplemental Nutrition Assistance Program (SNAP), unemployment insurance (UI), EITC, and CTC all expanded, producing a more even distribution across family types and income levels. But the expansion was mostly phased out; the long-run trend is expected to reassert.
• Medicaid complicates the picture: excluded from the SIPP-based calculations because families do not know the government's cost on their behalf; if included alongside employer-provided health insurance, many distributional shifts shrink substantially. The Affordable Care Act (ACA) would benefit the shallow poor, not the deepest poor.
• Policy implication: Support for elderly, disabled, and working poor is appropriate and should not be reduced. The problem is the collapse in support for the deepest poor and single mothers, who face real barriers to work (learning disabilities, physical and mental illness, domestic violence, substance abuse, criminal histories) that make the EITC's earnings requirement effectively exclusionary.

Concepts Introduced or Extended

• Deserving Poor
• Nonmarital Fertility
• DI Growth Decomposition

Entities Mentioned

• Robert Moffitt

Quotes

"The U.S. welfare system has indeed changed dramatically over the last four decades, but not by becoming smaller. Instead, it has changed by altering who receives benefits. The elderly, the disabled, working families with children, and married-couple families have gained. Single-parent families and the nonworking poor have lost."

"The deserving-undeserving distinction was part of the Anglo-Saxon tradition from the beginning, going back to the Elizabethan Poor Laws in the early 17th century."

My Take

This is primarily a descriptive accounting paper, not a causal identification paper — there are no instruments or quasi-experiments. Its value is synthetic: Moffitt assembles SIPP micro-data and program administrative statistics to document a pattern that was widely asserted but never rigorously established. The "deserving poor" framing is the paper's main theoretical contribution — it provides a parsimonious historical account for why political coalitions supported different programs differently. The most striking empirical result is the collapse of AFDC/TANF specifically: a program that paid benefits to the poorest families with children fell to below its 1970 level in real terms. The Medicaid exclusion is a legitimate limitation — the estimates would look different if in-kind health benefits were valued at cost — but the basic story holds for cash and near-cash programs. The Great Recession section is the weakest part: it reads as anecdotal and is based on limited data. The paper raises but does not answer the normative question of whether the shift toward the "deserving" groups was welfare-improving overall.