Human input
The privacy-safe aggregate contains 2 current ratings: 1 team and 1 public. The human mean is 90.0/100. The team has not made a final prioritization decision.
Loading public input and its source labels…
AI assessment: why consider this paper?
This study bears on a fairly concrete program choice: should cash-transfer programs add facilitated financial planning for couples, and what safeguards would that require? The trial reports better economic outcomes alongside increased intimate partner violence for households at highest risk. An evaluation could focus on that tradeoff, the subgroup analysis, and how far the Liberia results are likely to travel.
AI-generated criterion ratings and reasoning
Scores come from AI prioritization; the accompanying explanations include AI-assisted source checks. These are provisional judgments, separate from human ratings and commissioned evaluations.
Decision relevance
8.5/10
The result bears directly on whether cash-transfer programs should add facilitated financial planning for couples, and on the safeguards needed for households at higher risk of intimate partner violence.
- Paper excerpt “improves economic outcomes by 0.16 standard deviation units” NBER paper abstract
- Decision context Liberia’s social-protection ministry and GiveDirectly Liberia helped implement the trial, placing potential program users close to the research. NBER paper record
Value of added scrutiny
9.0/10
The score reflects the value of independent scrutiny of the benefit–harm tradeoff and the high-risk subgroup result. The targeted search found no independent methodological review or replication.
- Paper excerpt “at the cost of additional intimate partner violence” NBER paper abstract
- Search result No independent policy white paper, practitioner guidance, replication, or extended critique citing the paper surfaced in the targeted search. Unjournal public-attention search
Timing
9.5/10
This is a July 2026 working paper. Evaluation could still affect revisions, interpretation of the subgroup harm, and advice to implementing partners before journal publication or wider program use.
- Publication fact The public record identifies it as NBER Working Paper 35478, issued in July 2026. NBER
Methodological potential
8.5/10
A randomized trial offers a strong basis for evaluation, but the sensitive IPV outcome and subgroup analysis require close checks of measurement, statistical power, multiple testing, attrition, and the pre-analysis plan.
- Paper excerpt “We use a randomized controlled trial” NBER paper abstract
- Methodological issue The reported harm is concentrated among households with the highest ex-ante conflict risk, making subgroup specification and power central to the assessment. Paper abstract and evaluation record
Prominence
9.0/10
The NBER working-paper venue, experienced development-economics authors, and direct involvement of major implementation and funding organizations make the findings likely to receive serious attention.
- Venue and network The paper is distributed by NBER and credits the Liberian ministry, GiveDirectly Liberia, World Bank staff, and several research and funding partners. NBER paper record
Likely influence
8.5/10
The result could affect cash-plus program design because the implementing ministry, GiveDirectly, and World Bank staff were connected to the trial. Actual downstream use has not yet been established. See public-attention evidence below.
- Institutional connection The paper credits the Liberian ministry and GiveDirectly Liberia as implementation partners and names World Bank support for the evaluation. NBER paper record
What the paper says
Source abstract · Full source abstract recovered from the NBER paper page; HTML entities and whitespace normalized.
Households' most consequential economic decisions are usually made jointly, yet behavioral interventions are typically targeted to individuals and optimized to overcome intra-personal constraints. We use a randomized controlled trial to study whether behavioral interventions targeted to couples can overcome interpersonal constraints related to intra-household decision-making. In rural Liberia, we layer a facilitated joint financial planning exercise atop a large unconditional cash transfer, assessing effects relative to cash only and a pure control. Relative to cash only, adding planning increases the amount of the transfer spent on productive investment and consequently improves economic outcomes by 0.16 standard deviation units. However, for households with the highest ex-ante conflict risk, these gains come at the cost of additional intimate partner violence.
Claims to check
- In rural Liberia, adding a facilitated joint financial planning exercise to a large unconditional cash transfer increased the share or amount of transfer resources spent on productive investment relative to cash only.
- The cash-plus-planning arm improved economic outcomes by 0.16 standard deviation units relative to the cash-only arm.
- For households with the highest ex-ante conflict risk, the planning add-on increased intimate partner violence, creating an important welfare and safeguarding tradeoff.
Methodological or theoretical issues flagged for evaluation
The evaluation would need access to the pre-analysis plan, survey instruments, randomization records, attrition and compliance details, and ideally code/data to assess the core RCT claims. The IPV result is especially important but difficult: IPV is sensitive and self-reported, may be affected by survey protocols or social desirability, and subgroup effects by ex-ante conflict risk may raise power and multiple-testing concerns. A useful Unjournal evaluation should also examine whether the economic index and productive-investment outcomes are policy-interpretable, whether the planning module's implementation cost is low enough to justify scale-up, and how to value economic gains against increased IPV risk.
Opus 5.5 re-evaluation (experimental)
Read with care. This is an experimental re-evaluation by a different model (Claude Opus 5.5, 2026-10-01). Across a calibration sample Opus scored about 15 points lower than GPT-5.5, so 15 points are added to compare it with GPT-5.5 scores. The calibration is a small sample (n=44) and is not human-validated, and the two models disagree about the order of papers within the top group. The AI score at the top of this page is the usual evaluation-priority score shifted by the difference between the adjusted Opus score and the earlier holistic score; the synthesis uses that shifted value, and the unshifted score is shown beside it. On the dashboard you can switch the Opus scores off. Read the methods note.
Opus rationale (AI-generated)
This is an NBER working paper reporting an RCT in rural Liberia. It tests whether a cheap, facilitated joint financial planning session for couples, layered on a large unconditional cash transfer, changes how the money is used. It finds a 0.16 SD improvement in economic outcomes relative to cash alone, but more intimate partner violence among couples at high ex-ante conflict risk. That combination is squarely in Unjournal's core area (LMIC development, cash transfers, intra-household dynamics, gender-based violence). It bears directly on a live design choice for GiveDirectly, World Bank social protection teams and other cash implementers: whether to bundle 'cash-plus' couples programming, and how to screen or safeguard against backlash. As a fresh working paper with no evidence of independent public review, an evaluation is timely. Evaluators could probe whether the heterogeneity analysis was pre-registered, how conflict risk and IPV were measured (including survey ethics and reporting bias), how the economic index was built, and whether the cost-effectiveness case holds once the IPV harms are counted. The main concerns are single-site external validity and possible fragility of the subgroup result. These are exactly what a public evaluation could clarify before the finding shapes programme design.
Dashboard details and provenance
Full AI dashboard scoring rationale
This is a strong Unjournal candidate: an NBER working paper on a randomized cash-transfer intervention in rural Liberia, directly relevant to how governments and donors design social protection programs in low-income settings. The concrete decision is not whether cash transfers work in general, but whether agencies should add facilitated couples' financial planning to cash programs, and under what IPV-risk safeguards; that matters for the World Bank, Liberia's Ministry of Gender, Children, and Social Protection, GiveDirectly, IFPRI, IPA, FCDO, USAID, and other implementers of cash-plus programs. The paper reports an economically meaningful gain from the planning add-on, including more transfer spending on productive investment and a 0.16 SD improvement in economic outcomes relative to cash only, but also increased intimate partner violence among households at highest ex-ante conflict risk. Because this is a July 2026 NBER working paper rather than a peer-reviewed publication, and because partner organizations are directly positioned to use the findings, independent evaluation would add substantial value before journal publication or program scale-up. The main concern is not scope or relevance, but that evaluation needs to scrutinize the welfare tradeoff between economic gains and IPV harms, heterogeneity by conflict risk, and whether the findings generalize beyond this Liberian implementation context.
AI decision-relevance rationale
The study informs major social-protection design decisions: whether unconditional cash transfers should be paired with household or couples planning, whether such modules should be targeted only to lower-conflict households, and what IPV screening and referral safeguards are needed when cash changes intra-household bargaining. These are live decisions for the World Bank's Social Protection and Jobs teams, Liberia's cash-transfer program, GiveDirectly, and donors comparing simple cash to cash-plus programming. The welfare stakes are unusually concrete because the same intervention appears to raise productive investment and economic outcomes while creating harm for a high-risk subgroup.
AI timing assessment
Timing value is very high: the paper is an NBER working paper issued in July 2026 and has not yet gone through journal peer review. GiveDirectly's research listing describes the project as under analysis and registered with the AEA RCT Registry, so an Unjournal evaluation could still affect interpretation, replication materials, journal submission, and partner guidance before the result becomes a standard reference for cash-plus design.
Public attention and use
AI-assisted review of the linked public sources.
Several organizations helped run or fund the trial. We found little sign of independent discussion or use of the results since publication. The paper is only a few months old, so this mainly tells us that uptake is still too early to judge.
Institutional implementation and funding direct project context
The official NBER record credits Liberia’s Ministry of Gender, Children, and Social Protection and GiveDirectly Liberia as implementation partners. It also names World Bank staff who supported the evaluation and several funders. Potential users were therefore close to the project. We have not found evidence that they have adopted its conclusions.
Source: NBER · Relationship: official paper record
Research dissemination listing / discoverability
The paper is publicly available as NBER Working Paper 35478, is indexed in RePEc, and appears on Daniel Maggio’s research page. These improve discoverability but do not constitute substantive external scrutiny or use.
Source: RePEc · Relationship: bibliographic index
What the search did not establish
- No exact-title EA Forum or LessWrong discussion surfaced in the targeted search.
- No independent policy white paper, practitioner guidance, replication, or extended critique citing this 2026 paper surfaced.
- The most decision-relevant follow-up is whether implementing partners change cash-plus design or IPV safeguards in response to the heterogeneous harm result.
Targeted search checked 2026-09-01. Search scope: Targeted exact-title searches across the open web, institutional and author pages, news and public social-media results, EA Forum/LessWrong, and policy/white-paper contexts. Evidence records distinguish commissioning or report use from independent discussion, media attention, indexing, and post-publication policy use. A search miss is reported as uncertainty, not proof of absence. Entries with check_depth "quick" rest on roughly one to two searches and should be read as especially uncertain; entries checked on 2026-10-01 were added for the shortlist of papers without human ratings.