Contribution
From the introduction alone: a framework for reading people’s encounters with the state as costs that policy imposes and could remove. The book develops it in ten chapters on voting, abortion, the ACA, Medicare, SNAP, Medicaid, the EITC and Social Security, ending in a proposal for an “evidence-based approach”; none of those was read.
Key insight
A right or benefit on paper is not one in practice: the costs of finding out about it, proving eligibility and bearing the stigma of asking decide who actually gets it, and those costs are set by choices that can be changed without touching eligibility. “Administrative burdens are the fine print of the social contract between citizens and their government” (p. 2).
Assumptions
The authors state four normative assumptions (pp. 13–14): programs should be simple, accessible and respectful; burdens should be minimized consistent with values such as cost and program integrity; judgments of burden should be evidence based; and costs on those with limited resources deserve special attention. They grant that some burdens are legitimate, and sort cases into three: burdens that deter fraud at no cost to participation, burdens that buy some value at the cost of access, and burdens with no tangible relation to their stated goal, where “limiting citizen access … is the tacit goal” (pp. 12–13).
Key results
As the introduction reports them, from the literature, not shown there:
- Three costs (p. 2): learning costs, compliance costs, psychological costs (stress, loss of autonomy, stigma). Burdens can be reduced (accepting a person’s word) or shifted to the state (eligibility workers checking administrative databases).
- Take-up (pp. 6–7): near 100 percent for Social Security and Medicare; 40–60 percent for SSI, about 65 for SNAP, 30–60 for Unemployment Insurance, 50–70 for Medicaid, about 80 for the EITC; AFDC 77–86 and its successor TANF 42–52.
- Constructed (pp. 8–11): the State Department’s use of documentation requirements to keep German Jewish refugees out in the 1930s, with quotas two-thirds unfilled; return-free tax filing technically feasible for an estimated 40 percent of taxpayers and resisted by Intuit’s lobbying.
- Capacity as well as will (pp. 11–12): states opposed to the ACA used procedure to deter enrollment, but low capacity also created burdens in governments that supported it. “Opportunities to shift administrative burdens away from individuals depends on a mixture of political will and administrative capacity”, sometimes supplied by nonprofit or private actors.
Concepts
- Administrative burden (p. 2): the learning, compliance and psychological costs a person meets when the state regulates private behaviour or structures access to public services.
- Burden shifting (p. 2, p. 10): moving a cost from the person to the state, for example by using data the state already holds.
Connections
The introduction cites the take-up literature and the field experiments on FAFSA help and college information packets (p. 7) without discussing them. Its pairing of political will with administrative capacity is the pair Hillebrecht’s review (LIT-1607) is said to use for compliance with court rulings; that work was not reachable.
Bearing on the record
- CASE-172 and TERM-055. The exchange’s use of the book holds on the introduction: it says burdens can restrict access to a benefit or a right with no change to formal eligibility, so that what is lost is the practical ability to use it. That is TERM-055’s operational availability carried from a resource to a right: present, known, and not deployable before the deadline. Whether the chapters give a measure of that loss (take-up is the introduction’s only one) is not known from this reading.
No instruction for machine-learning practice; nothing for the anthology.
Limitations
- Only the 14 pages of the introduction read, of a book whose index ends after p. 337; no chapter’s argument or evidence was checked.
- An inconsistent figure (p. 3): “A recent analysis estimated the compliance costs of filing to be about 19.5 trillion) is roughly $230 billion. One of the numbers is wrong; the source in note 5 was not checked.
Open questions
- Do the chapters measure burden other than by take-up, and does any separate the effect of a burden from that of eligibility?