{ "schema_version": "welfare-voi-screen-v0.1", "generated_at": "2026-09-12", "status": "AI-assisted screening expansion; not an official Unjournal evaluation, cross-cause ranking, or funding recommendation", "title": "Expanded screening under the updated welfare and value-of-information protocol", "scope": "Eighteen papers selected from the public prioritization dashboard, three in each of six contrasting areas and excluding the six-paper pilot", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "purpose": "Add balanced comparison cases without manufacturing cross-cause welfare numbers.", "method": { "causal_chain": [ "research", "belief_or_tool_change", "decision_change", "natural_outcomes", "welfare" ], "provenance_tags": [ "source_estimate", "analyst_assumption", "unknown" ], "funding_formula": "relevant budget x research-caused allocation-share change x incremental outcome per dollar versus the displaced alternative x welfare per outcome, minus costs and harms", "coverage_formula": "eligible population x research-caused coverage change x incremental outcome per covered unit x duration x welfare weight, minus costs and harms", "research_tool_formula": "downstream decisions reached x probability the tool changes evidence x probability evidence changes the decision x outcome difference x welfare weight", "scenario_rule": "Low, central, and high entries are sensitivity scenarios, fixed source inputs, or source intervals as explicitly labeled; they are not automatically confidence intervals.", "aggregation_rule": "Do not produce a common score or rank when essential causal terms are unknown. Preserve natural outcomes and conditional calculations.", "income": { "utility": "CRRA", "eta_sensitivity": [ 1, 1.5, 2 ], "reference_consumption_ppp_per_person_year": 1000, "rule": "Calculate only when beneficiary consumption is available; aggregate GDP, transfers, and income indices are insufficient." }, "health_bridge": { "healthy_year_equivalents_per_consumption_doubling_year": 0.4347826086956522, "status": "analyst convention inherited from the six-paper pilot; not used to force a health conversion in this screen" }, "animal": { "source": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "rule": "Use 5th/50th/95th welfare ranges as sensitivity inputs; values already include probability-of-sentience and subjective-experience-rate adjustments." }, "catastrophic_risk": { "rule": "Keep present-generation and long-future scenarios separate. Conditional model probabilities are not overall credences." }, "evaluation_counterfactual": "Separate a paper's welfare relevance from the incremental value of commissioning or updating an Unjournal evaluation." }, "summary": { "paper_count": 18, "cause_area_counts": { "LMIC development and health": 3, "High-income health and education": 3, "Animal welfare": 3, "AI and global development": 3, "Methods and research tools": 3, "Catastrophic and pandemic risk": 3 }, "welfare_estimates_completed": 0, "reason_no_welfare_totals": "Every case lacks at least one decisive multiplicative input, usually research-attributable decision change. The report preserves those gaps rather than replacing them with generic priors.", "comparison_rule": "Screening signals organize follow-up work within cases; they are not a cross-cause rank or score." }, "headline_result": "All eighteen papers are missing at least one decisive multiplicative input, usually a research-attributable decision change. The screen therefore reports natural outcomes, conditional arithmetic, and unknowns without assigning a numerical welfare total or cross-cause rank.", "cross_case_assessment": { "defensible_total_welfare_estimates": 0, "paper_screens": 18, "qualitative_signals_are_not_ranks": true, "implication_for_current_0_to_10_scores": "The expanded screen does not recover or validate the existing ordinal impact ratings. It exposes the evidence needed before any rating can be described as a calculated welfare estimate." }, "framework_sources": [ "https://www.givewell.org/international/technical/programs/givedirectly-cash-for-poverty-relief-program", "https://www.givewell.org/how-we-work/our-criteria/cost-effectiveness/moral-weights", "https://rethinkpriorities.org/research-area/an-introduction-to-the-moral-weight-project/", "https://rethinkpriorities.org/research-area/welfare-range-estimates/" ], "papers": [ { "paper_id": 6, "title": "Better Together? The Impact of Cash Transfers and Couples' Financial Planning on Household Dynamics and Economic Outcomes", "url": "https://www.nber.org/papers/w35478", "case_role": "expanded_screening", "cause_area": "LMIC development and health", "pathway_type": "coverage", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "strong_candidate_for_full_model", "paper_welfare_relevance": "direct_but_mixed", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "A recent Liberian RCT offers a concrete cash-plus design decision and reports both an economic gain and an intimate-partner-violence warning. The missing harm magnitude prevents a net welfare calculation.", "causal_chain": { "research_output": "RCT evidence comparing cash plus facilitated couples' planning with cash alone and pure control in rural Liberia.", "belief_or_tool_change": "Change implementer beliefs about whether a couples-planning add-on improves investment and when it creates safeguarding risks.", "decision_change": "Add, target, redesign, or reject couples' planning modules in cash-transfer programs.", "natural_outcome": "Economic-outcome index change and intimate-partner-violence outcomes among high-conflict-risk households.", "welfare_bridge": "Consumption gains under eta sensitivity, minus health and suffering from IPV; neither side should be monetized and counted again." }, "quantitative_inputs": [ { "name": "economic_outcome_effect", "description": "Effect of planning add-on relative to cash only", "unit": "standard deviations", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w35478", "low": 0.16, "central": 0.16, "high": 0.16, "value": 0.16, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "ipv_effect", "description": "Additional intimate partner violence in the highest ex-ante conflict-risk group", "unit": "incidents, prevalence points, or severity-duration units", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "The public abstract reports direction but not magnitude." }, { "name": "research_caused_program_coverage_change", "description": "Additional or better-targeted coverage caused by this paper", "unit": "households", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "Named partners do not establish marginal research influence." }, { "name": "program_incremental_cost", "description": "Cost of facilitated planning relative to cash only", "unit": "2026 PPP dollars per household", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "newly or better-targeted households x incremental economic outcome, with IPV harms on a separate ledger", "scenarios": [ { "scenario": "paper-reported effect", "expression": "1 covered household x 0.16 SD", "result": 0.16, "unit": "economic-outcome SD per additionally covered household", "status": "paper_reported_effect", "note": "This is not a welfare unit and cannot be aggregated without the index scale, duration, costs, and IPV magnitude." } ] }, "natural_outcome_estimate": { "status": "paper_reported_effect_only", "low": 0.16, "central": 0.16, "high": 0.16, "unit": "economic-outcome SD per covered household", "harms": "direction reported; magnitude unknown" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "Program reach attributable to the paper, the economic index-to-consumption bridge, persistence, costs, and the IPV severity-duration effect are missing." }, "income_eta_sensitivity": { "status": "not_estimated", "etas": [ 1.0, 1.5, 2.0 ], "formula": "u_eta(c + delta_c) - u_eta(c), where u_1(c)=ln(c/c_ref) and u_eta(c)=((c/c_ref)^(1-eta)-1)/(1-eta)", "reason": "The paper reports an index effect rather than beneficiary-level consumption, and it does not provide a public distribution of baseline consumption in the source abstract.", "required_inputs": [ "beneficiary-level baseline real consumption on a common PPP basis", "the consumption change attributable to the research-caused decision", "duration and beneficiary distribution", "a fixed reference consumption for cross-scenario normalization" ], "interpretation": "For the same 1% consumption gain at one-quarter of a reference income versus at the reference income, the relative welfare weights are 1x, 2x, and 4x for eta=1, 1.5, and 2. This is a generic curvature illustration, not a paper estimate." }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "high_for_new_evaluation", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "The paper is new, the decision is live, and evaluation could materially test the benefit-harm tradeoff, subgroup reliability, and safeguards before scale-up.", "qualitative_assessment": "The paper is new, the decision is live, and evaluation could materially test the benefit-harm tradeoff, subgroup reliability, and safeguards before scale-up.", "decisive_unknowns": [ "magnitude and precision of the IPV effect", "component and implementation costs", "durability and cash-equivalent interpretation of the 0.16 SD index", "whether partner decisions remain open" ] }, "uncertainty_and_double_counting": [ "Do not net the 0.16 SD economic index against an unmeasured IPV outcome.", "Do not count productive investment, income, consumption, and the same economic index as independent gains." ], "sources": [ "https://www.nber.org/papers/w35478" ], "citations": [ { "label": "NBER paper page and abstract", "url": "https://www.nber.org/papers/w35478", "claim": "setting, comparison arms, 0.16 SD result, and direction of IPV harm" } ] }, { "paper_id": 529, "title": "Female Micro-Entrepreneurship and Child Development", "url": "https://www.nber.org/papers/w35710", "case_role": "expanded_screening", "cause_area": "LMIC development and health", "pathway_type": "coverage", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "strong_candidate_for_full_model", "paper_welfare_relevance": "direct", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The Malawi RCT reports short-run child-development gains in a severely disadvantaged population. Costs, component attribution, persistence, and the link from index gains to lifetime welfare remain decisive.", "causal_chain": { "research_output": "RCT evidence on a child-focused ultra-poor graduation program that includes early-childhood-development training.", "belief_or_tool_change": "Update beliefs about adding a child-development focus to graduation programs.", "decision_change": "Fund, redesign, or scale child-focused graduation programs rather than a standard bundle or alternative child intervention.", "natural_outcome": "Child cognitive and non-cognitive development at one year; anthropometrics.", "welfare_bridge": "Translate durable child-development and household-consumption changes separately; avoid assuming an SD gain is a lifetime income gain." }, "quantitative_inputs": [ { "name": "child_development_effect", "description": "Reported range of child-development effects within one year", "unit": "standard deviations", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w35710", "low": 0.16, "central": 0.21, "high": 0.26, "value": null, "note": "The midpoint is an arithmetic summary of the reported 0.16-0.26 SD range, not the paper's preferred single estimate." }, { "name": "trial_households", "description": "RCT sample size", "unit": "households", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w35710", "low": 1872, "central": 1872, "high": 1872, "value": 1872, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "incremental_cost_of_child_focus", "description": "Cost above a standard graduation program", "unit": "2026 PPP dollars per household", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "research_caused_coverage_change", "description": "Additional coverage or redesign attributable to the paper", "unit": "children", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "effect_persistence", "description": "Duration of child-development effect", "unit": "years", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "additional children reached x child-development SD effect x persistence, before any welfare conversion", "scenarios": [ { "scenario": "per additionally covered child", "expression": "1 x [0.16, 0.21, 0.26]", "result": [ 0.16, 0.21, 0.26 ], "unit": "child-development SD at one year", "status": "paper_reported_range", "note": "" } ] }, "natural_outcome_estimate": { "status": "paper_reported_effect_only", "low": 0.16, "central": 0.21, "high": 0.26, "unit": "child-development SD at one year" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "The welfare implications depend on persistence, construct validity, costs, affected-child counts, later health and income, and the value of displaced programs." }, "income_eta_sensitivity": { "status": "not_estimated", "etas": [ 1.0, 1.5, 2.0 ], "formula": "u_eta(c + delta_c) - u_eta(c), where u_1(c)=ln(c/c_ref) and u_eta(c)=((c/c_ref)^(1-eta)-1)/(1-eta)", "reason": "The source abstract does not report consumption changes or the consumption distribution, and the child-development index cannot be put into CRRA utility directly.", "required_inputs": [ "beneficiary-level baseline real consumption on a common PPP basis", "the consumption change attributable to the research-caused decision", "duration and beneficiary distribution", "a fixed reference consumption for cross-scenario normalization" ], "interpretation": "For the same 1% consumption gain at one-quarter of a reference income versus at the reference income, the relative welfare weights are 1x, 2x, and 4x for eta=1, 1.5, and 2. This is a generic curvature illustration, not a paper estimate." }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "high_for_new_evaluation", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "Because the working paper is very recent and the design could affect poverty-program portfolios, evaluation could clarify durability, cost-effectiveness, and what the child-focused component adds.", "qualitative_assessment": "Because the working paper is very recent and the design could affect poverty-program portfolios, evaluation could clarify durability, cost-effectiveness, and what the child-focused component adds.", "decisive_unknowns": [ "incremental cost and implementation burden", "longer-run follow-up", "separation of ECD training from the wider program bundle", "allocation alternative" ] }, "uncertainty_and_double_counting": [ "Do not count child-development scores, future income, and parental economic effects as separate benefits without a mediation model.", "The null anthropometric result is not evidence of no health effect and should not be converted to zero welfare." ], "sources": [ "https://www.nber.org/papers/w35710" ], "citations": [ { "label": "NBER paper page and abstract", "url": "https://www.nber.org/papers/w35710", "claim": "Malawi setting, sample, effect range, and null anthropometric finding" } ] }, { "paper_id": 345, "title": "Water Treatment and Child Mortality: A Meta-analysis and Cost-effectiveness Analysis", "url": "https://bfi.uchicago.edu/wp-content/uploads/2023/01/BFI_WP_2022-26-2.pdf", "case_role": "expanded_screening", "cause_area": "LMIC development and health", "pathway_type": "coverage", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "use_existing_evaluation_for_model_update", "paper_welfare_relevance": "direct", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The mortality outcome is unusually close to welfare, but study heterogeneity and the distinction between intervention effect and research-caused adoption dominate the paper-level contribution.", "causal_chain": { "research_output": "Bayesian meta-analysis of randomized water-quality interventions and a cost-effectiveness analysis.", "belief_or_tool_change": "Update beliefs about the mortality effect and cost-effectiveness of water treatment.", "decision_change": "Expand, target, or redesign chlorination and other water-treatment programs.", "natural_outcome": "Under-five deaths or mortality odds among covered children.", "welfare_bridge": "Life-years or health-year equivalents, with separate household cost savings if any and no monetized-health double count." }, "quantitative_inputs": [ { "name": "pooled_mortality_odds_ratio", "description": "Bayesian pooled odds ratio for all-cause child mortality", "unit": "odds ratio", "provenance_tag": "source_estimate", "source_url": "https://statmodeling.stat.columbia.edu/2023/01/25/water-treatment-and-child-mortality-a-meta-analysis-and-cost-effectiveness-analysis/", "low": 0.49, "central": 0.7, "high": 0.93, "value": null, "note": "95% interval as reported by a paper coauthor." }, { "name": "new_study_predictive_odds_ratio", "description": "Predictive odds ratio for a new study", "unit": "odds ratio", "provenance_tag": "source_estimate", "source_url": "https://statmodeling.stat.columbia.edu/2023/01/25/water-treatment-and-child-mortality-a-meta-analysis-and-cost-effectiveness-analysis/", "low": 0.29, "central": 0.75, "high": 1.5, "value": null, "note": "The wide predictive interval reflects heterogeneity and crosses no effect." }, { "name": "meta_analysis_children", "description": "Approximate number of children represented in 15 usable trial estimates", "unit": "children", "provenance_tag": "source_estimate", "source_url": "https://statmodeling.stat.columbia.edu/2023/01/25/water-treatment-and-child-mortality-a-meta-analysis-and-cost-effectiveness-analysis/", "low": 25000, "central": 25000, "high": 25000, "value": 25000, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "research_caused_additional_coverage", "description": "Water-treatment coverage attributable to this research beyond existing evidence", "unit": "child-years", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "local_baseline_mortality", "description": "Baseline child mortality in each target setting", "unit": "deaths per child-year", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "covered child-years x baseline mortality risk x relative-risk change", "scenarios": [ { "scenario": "odds-ratio evidence", "expression": "OR interval [0.49, 0.70, 0.93]; absolute deaths require baseline risk", "result": null, "unit": "deaths averted", "status": "not_estimable_from_public_screen", "note": "" } ] }, "natural_outcome_estimate": { "status": "paper_reported_relative_effect_only", "low": 0.49, "central": 0.7, "high": 0.93, "unit": "mortality odds ratio" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "Absolute deaths require baseline risk and coverage, while the paper's welfare contribution also requires the marginal decision change caused by the research and the counterfactual intervention portfolio." }, "income_eta_sensitivity": { "status": "not_applicable" }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "low_for_duplicate_evaluation_high_for_follow_up_use", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "Unjournal already commissioned two evaluations and an author response. The next value is in using those evaluations to parameterize an adoption model or checking a materially revised analysis, not commissioning a duplicate review.", "qualitative_assessment": "Unjournal already commissioned two evaluations and an author response. The next value is in using those evaluations to parameterize an adoption model or checking a materially revised analysis, not commissioning a duplicate review.", "decisive_unknowns": [ "how the published evaluation changed the paper or implementer beliefs", "country-specific absolute mortality risks", "cost and delivery evidence by intervention type", "incremental uptake caused by the research and evaluation" ] }, "uncertainty_and_double_counting": [ "Odds ratios should not be applied as risk ratios at high baseline risk.", "Do not count deaths averted and their monetized value as separate benefits." ], "sources": [ "https://statmodeling.stat.columbia.edu/2023/01/25/water-treatment-and-child-mortality-a-meta-analysis-and-cost-effectiveness-analysis/", "https://unjournal.pubpub.org/pub/evalsumwatertreatment/release/4" ], "citations": [ { "label": "Paper coauthor's statistical summary", "url": "https://statmodeling.stat.columbia.edu/2023/01/25/water-treatment-and-child-mortality-a-meta-analysis-and-cost-effectiveness-analysis/", "claim": "study count, sample, pooled and predictive odds ratios" }, { "label": "Unjournal evaluation summary", "url": "https://unjournal.pubpub.org/pub/evalsumwatertreatment/release/4", "claim": "existing evaluations, concerns about review transparency, interpretation, and cost analysis" } ] }, { "paper_id": 351, "title": "How Effective Is (More) Money? Randomizing Unconditional Cash Transfer Amounts in the US", "url": "https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4154000", "case_role": "expanded_screening", "cause_area": "High-income health and education", "pathway_type": "coverage", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "use_existing_evaluation_as_high_income_negative_control", "paper_welfare_relevance": "direct_but_geographically_limited", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "This is useful as a high-income negative/control case: the trial found spending responses but no positive prespecified survey outcomes. It should not receive a high global-welfare score merely because cash is easy to quantify.", "causal_chain": { "research_output": "RCT of one-time $500 and $2,000 transfers to Americans in poverty.", "belief_or_tool_change": "Update beliefs about short-run survey outcomes from one-time transfers in a wealthy-country pandemic context.", "decision_change": "Alter the size, timing, targeting, or use of one-time emergency cash programs.", "natural_outcome": "Expenditure and financial, psychological, cognitive, and physical-health survey outcomes through 15 weeks.", "welfare_bridge": "Consumption utility at beneficiary baseline consumption, plus measured health or distress outcomes without double counting spending itself as welfare." }, "quantitative_inputs": [ { "name": "trial_sample", "description": "Randomized sample", "unit": "participants", "provenance_tag": "source_estimate", "source_url": "https://www.globalprioritiesinstitute.org/wp-content/uploads/Ania-Jaroszewicz-Oliver-P.-Hauser-Jon-M.-Jachimowicz-and-Julian-Jamison-How-effective-is-more-money-Randomizing-unconditional-cash-transfer-amounts-in-the-US.pdf", "low": 5243, "central": 5243, "high": 5243, "value": 5243, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "small_transfer", "description": "Small one-time unconditional transfer", "unit": "nominal US dollars", "provenance_tag": "source_estimate", "source_url": "https://www.globalprioritiesinstitute.org/wp-content/uploads/Ania-Jaroszewicz-Oliver-P.-Hauser-Jon-M.-Jachimowicz-and-Julian-Jamison-How-effective-is-more-money-Randomizing-unconditional-cash-transfer-amounts-in-the-US.pdf", "low": 500, "central": 500, "high": 500, "value": 500, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "large_transfer", "description": "Large one-time unconditional transfer", "unit": "nominal US dollars", "provenance_tag": "source_estimate", "source_url": "https://www.globalprioritiesinstitute.org/wp-content/uploads/Ania-Jaroszewicz-Oliver-P.-Hauser-Jon-M.-Jachimowicz-and-Julian-Jamison-How-effective-is-more-money-Randomizing-unconditional-cash-transfer-amounts-in-the-US.pdf", "low": 2000, "central": 2000, "high": 2000, "value": 2000, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "research_caused_policy_change", "description": "Change in eligible recipients or transfer design caused by the paper", "unit": "recipient-program equivalents", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "net_long_run_outcome", "description": "Outcome after the 15-week measurement window", "unit": "consumption, health, or well-being units", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "research-caused recipient change x net outcome per recipient", "scenarios": [ { "scenario": "reported short-run survey result", "expression": "positive effect on prespecified survey outcomes", "result": null, "unit": "survey outcome units", "status": "no_positive_effect_detected", "note": "A null or negative estimate is not proof of zero welfare, particularly outside the measured window." } ] }, "natural_outcome_estimate": { "status": "no_positive_effect_detected_on_prespecified_surveys", "low": null, "central": null, "high": null, "unit": "multiple survey outcomes through 15 weeks" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "The public screen lacks outcome coefficients, attrition-adjusted estimates, longer-run effects, baseline consumption, and a policy-change attributable to the paper." }, "income_eta_sensitivity": { "status": "not_estimated", "etas": [ 1.0, 1.5, 2.0 ], "formula": "u_eta(c + delta_c) - u_eta(c), where u_1(c)=ln(c/c_ref) and u_eta(c)=((c/c_ref)^(1-eta)-1)/(1-eta)", "reason": "The trial concerns consumption and cash, but a welfare calculation needs participant-level baseline consumption and outcome effects, not only nominal transfer amounts.", "required_inputs": [ "beneficiary-level baseline real consumption on a common PPP basis", "the consumption change attributable to the research-caused decision", "duration and beneficiary distribution", "a fixed reference consumption for cross-scenario normalization" ], "interpretation": "For the same 1% consumption gain at one-quarter of a reference income versus at the reference income, the relative welfare weights are 1x, 2x, and 4x for eta=1, 1.5, and 2. This is a generic curvature illustration, not a paper estimate." }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "low_for_duplicate_evaluation", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "Unjournal already organized two evaluations. Additional value would come from a later follow-up or a synthesis comparing crisis-era US cash with LMIC evidence, rather than another general review.", "qualitative_assessment": "Unjournal already organized two evaluations. Additional value would come from a later follow-up or a synthesis comparing crisis-era US cash with LMIC evidence, rather than another general review.", "decisive_unknowns": [ "long-run follow-up", "pandemic-context external validity", "which policy alternative is displaced", "effect of the existing evaluations on users" ] }, "uncertainty_and_double_counting": [ "Do not equate an observed spending increase with a welfare gain.", "Do not treat absence of statistical evidence as a precise zero.", "The US-poverty label does not supply a global consumption percentile." ], "sources": [ "https://www.globalprioritiesinstitute.org/wp-content/uploads/Ania-Jaroszewicz-Oliver-P.-Hauser-Jon-M.-Jachimowicz-and-Julian-Jamison-How-effective-is-more-money-Randomizing-unconditional-cash-transfer-amounts-in-the-US.pdf", "https://unjournal.pubpub.org/pub/evalsumrandomizingcashtransfer/release/3" ], "citations": [ { "label": "Paper PDF", "url": "https://www.globalprioritiesinstitute.org/wp-content/uploads/Ania-Jaroszewicz-Oliver-P.-Hauser-Jon-M.-Jachimowicz-and-Julian-Jamison-How-effective-is-more-money-Randomizing-unconditional-cash-transfer-amounts-in-the-US.pdf", "claim": "sample, transfer amounts, outcomes, and main result" }, { "label": "Unjournal evaluation summary", "url": "https://unjournal.pubpub.org/pub/evalsumrandomizingcashtransfer/release/3", "claim": "existing two evaluations and key limitations" } ] }, { "paper_id": 468, "title": "Legislating Longevity: State Public Health Laws and Mortality in the Early Twentieth Century", "url": "https://www.nber.org/papers/w35686", "case_role": "expanded_screening", "cause_area": "High-income health and education", "pathway_type": "research_tool", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "candidate_if_modern_transfer_path_is_specified", "paper_welfare_relevance": "indirect_historical", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The historical US mortality claim is large, but a current welfare pathway depends on which modern legal or institutional choices the paper can change and whether the design survives close scrutiny.", "causal_chain": { "research_output": "Historical law database, legislative-calendar IV, and topic classification linking state public-health laws to mortality.", "belief_or_tool_change": "Update beliefs about whether legal authority and regulation caused historical mortality decline.", "decision_change": "Change present investments in public-health legal capacity, WASH regulation, or communicable-disease authority.", "natural_outcome": "Cause-specific deaths under a modern policy that is genuinely analogous.", "welfare_bridge": "Healthy life-years, plus separately measured institutional costs and rights burdens." }, "quantitative_inputs": [ { "name": "share_of_historical_mortality_decline", "description": "Approximate mortality decline accounted for by the universe of public-health laws", "unit": "share", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w35686", "low": 0.667, "central": 0.667, "high": 0.667, "value": null, "note": "The paper says approximately two-thirds; this is not a causal confidence interval." }, { "name": "modern_policy_analogy", "description": "Share of a modern proposed policy effect informed by the historical setting", "unit": "share", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "research_caused_policy_change", "description": "Additional legal capacity or adoption caused by this paper", "unit": "jurisdiction-years", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "modern_mortality_effect", "description": "Mortality effect in target modern settings", "unit": "deaths per jurisdiction-year", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "target jurisdiction-years x research-caused legal change x modern incremental mortality effect", "scenarios": [ { "scenario": "historical attribution only", "expression": "approximately 2/3 of study-period mortality decline", "result": 0.667, "unit": "share of historical decline", "status": "paper_reported_claim", "note": "This does not transfer mechanically to current countries or laws." } ] }, "natural_outcome_estimate": { "status": "historical_aggregate_claim_only", "low": 0.667, "central": 0.667, "high": 0.667, "unit": "share of study-period mortality decline" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "No modern action, reachable population, transfer factor, or research-attributable adoption effect is specified, and the IV and aggregate attribution require detailed validation." }, "income_eta_sensitivity": { "status": "not_applicable" }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "moderate_if_identification_and_user_are_in_scope", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "A review could be valuable if it focuses on the legislative-calendar instrument and the broad two-thirds attribution and identifies a concrete current user. Without that link, the global-welfare case is mainly historical learning.", "qualitative_assessment": "A review could be valuable if it focuses on the legislative-calendar instrument and the broad two-thirds attribution and identifies a concrete current user. Without that link, the global-welfare case is mainly historical learning.", "decisive_unknowns": [ "validity of the legislative-calendar instrument", "enforcement versus enactment", "modern policy analogue", "current decision-maker whose action could change" ] }, "uncertainty_and_double_counting": [ "Do not treat an historical share of mortality decline as a forecast effect size.", "Avoid counting sanitation-driven mortality decline again through general income growth." ], "sources": [ "https://www.nber.org/papers/w35686" ], "citations": [ { "label": "NBER paper page and abstract", "url": "https://www.nber.org/papers/w35686", "claim": "historical setting, identification strategy, cause-specific pattern, and two-thirds claim" } ] }, { "paper_id": 17, "title": "Intergovernmental Grants to School Districts and Educational Outcomes During the COVID-19 Pandemic", "url": "https://www.nber.org/papers/w35447", "case_role": "expanded_screening", "cause_area": "High-income health and education", "pathway_type": "funding", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "useful_policy_null_with_limited_global_weight", "paper_welfare_relevance": "direct_but_geographically_limited", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The paper can inform US fiscal incidence and emergency-school-aid design, but the studied margin did not raise per-pupil spending or mitigate test-score declines. Its global relevance requires a transferable mechanism, not the headline $189.5 billion budget.", "causal_chain": { "research_output": "Difference-in-discontinuities evidence near a district poverty-share threshold in US ESSER allocation.", "belief_or_tool_change": "Update beliefs about whether marginal federal grants become school spending or local tax relief.", "decision_change": "Change grant formulas, maintenance-of-effort rules, or emergency education funding.", "natural_outcome": "Per-pupil spending, local revenue, test scores, enrollment, reopening, and political engagement.", "welfare_bridge": "Education and tax incidence by beneficiary consumption, after distinguishing real resource gains from transfers." }, "quantitative_inputs": [ { "name": "total_esser_appropriation", "description": "Total federal ESSER appropriation described by the paper", "unit": "nominal US dollars", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w35447", "low": 189500000000, "central": 189500000000, "high": 189500000000, "value": 189500000000, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "local_first_stage", "description": "Threshold-induced additional federal funding per district or pupil", "unit": "nominal US dollars", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "research_caused_formula_change", "description": "Policy allocation change attributable to the paper", "unit": "nominal US dollars", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "education_outcome_gain", "description": "Incremental learning or completion outcome under a redesigned grant", "unit": "test-score SD or completions", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "reallocated dollars x incremental outcome per dollar relative to the displaced use", "scenarios": [ { "scenario": "reported studied margin", "expression": "additional grant -> no increase in per-pupil expenditure -> no detected mitigation of test-score decline", "result": null, "unit": "test-score SD", "status": "no_effect_detected_on_test_scores", "note": "" } ] }, "natural_outcome_estimate": { "status": "no_effect_detected_on_test_score_decline", "low": null, "central": null, "high": null, "unit": "test-score outcomes near the threshold" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "The $189.5 billion program scale is not the paper's identified first stage and cannot be multiplied by a research-influence guess. Welfare incidence of tax relief and any enrollment/reopening benefit are not quantified here." }, "income_eta_sensitivity": { "status": "not_estimated", "etas": [ 1.0, 1.5, 2.0 ], "formula": "u_eta(c + delta_c) - u_eta(c), where u_1(c)=ln(c/c_ref) and u_eta(c)=((c/c_ref)^(1-eta)-1)/(1-eta)", "reason": "The fiscal incidence is unknown; a local tax dollar does not identify the final household or its baseline consumption.", "required_inputs": [ "beneficiary-level baseline real consumption on a common PPP basis", "the consumption change attributable to the research-caused decision", "duration and beneficiary distribution", "a fixed reference consumption for cross-scenario normalization" ], "interpretation": "For the same 1% consumption gain at one-quarter of a reference income versus at the reference income, the relative welfare weights are 1x, 2x, and 4x for eta=1, 1.5, and 2. This is a generic curvature illustration, not a paper estimate." }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "moderate_for_us_policy_low_for_global_welfare_model", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "Evaluation could test the discontinuity design and clarify why aid changed taxes rather than spending. The case for prioritizing it globally is limited unless the grant-incidence mechanism transfers to major LMIC education budgets.", "qualitative_assessment": "Evaluation could test the discontinuity design and clarify why aid changed taxes rather than spending. The case for prioritizing it globally is limited unless the grant-incidence mechanism transfers to major LMIC education budgets.", "decisive_unknowns": [ "identified funding increment at the cutoff", "precision of outcome estimates", "beneficiary incidence of tax relief", "transfer to other intergovernmental systems" ] }, "uncertainty_and_double_counting": [ "The total federal appropriation is context, not identified exposure.", "Tax reductions are transfers and should not be valued as gross social benefits without the financing counterfactual." ], "sources": [ "https://www.nber.org/papers/w35447" ], "citations": [ { "label": "NBER paper page and abstract", "url": "https://www.nber.org/papers/w35447", "claim": "program scale, design, pass-through, and education outcomes" } ] }, { "paper_id": 641, "title": "The Animal-Welfare Levy", "url": "https://academic.oup.com/jeea/advance-article-abstract/doi/10.1093/jeea/jvaf063/8383422", "case_role": "expanded_screening", "cause_area": "Animal welfare", "pathway_type": "funding", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "strong_framework_candidate_requires_behavioral_link", "paper_welfare_relevance": "direct_normative_model", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The levy paper directly puts farmed-animal welfare into a price instrument and highlights population-ethics dependence. Welfare impact still requires a research-caused policy change, species-specific quantity response, production pass-through, and animal-life counterfactual.", "causal_chain": { "research_output": "Pigouvian levy framework and calibrated species-specific meat taxes under alternative normative assumptions.", "belief_or_tool_change": "Give analysts and policymakers a way to include farmed-animal welfare externalities in tax design.", "decision_change": "Adopt or revise meat levies, subsidies, or accompanying welfare regulation.", "natural_outcome": "Changes in quantities of meat produced, animal lives created, and welfare conditions by species.", "welfare_bridge": "Animal-years x change in realized welfare as a fraction of the species range x RP adjusted range; population ethics kept explicit." }, "quantitative_inputs": [ { "name": "chickens_adjusted_welfare_range", "description": "Rethink Priorities probability-of-sentience- and subjective-experience-rate-adjusted welfare range for chickens", "unit": "human-range equivalents per unit time", "provenance_tag": "source_estimate", "source_url": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "low": 0.002, "central": 0.332, "high": 0.869, "value": null, "note": "The three values are RP's 5th, 50th, and 95th percentiles. They already include a sentience adjustment; do not apply one again. The median is a sensitivity case, not an expected value." }, { "name": "pigs_adjusted_welfare_range", "description": "Rethink Priorities probability-of-sentience- and subjective-experience-rate-adjusted welfare range for pigs", "unit": "human-range equivalents per unit time", "provenance_tag": "source_estimate", "source_url": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "low": 0.005, "central": 0.515, "high": 1.031, "value": null, "note": "The three values are RP's 5th, 50th, and 95th percentiles. They already include a sentience adjustment; do not apply one again. The median is a sensitivity case, not an expected value." }, { "name": "cattle_welfare_range", "description": "Relative welfare range for cattle", "unit": "human-range equivalents per unit time", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "The cited Rethink Priorities table does not report this species." }, { "name": "research_caused_levy_change", "description": "Tax or subsidy change attributable to this paper", "unit": "currency per kilogram by species", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "production_response", "description": "Change in animal production caused by the levy", "unit": "animal-years by species", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "life_worth_living_counterfactual", "description": "Welfare effect of preventing an animal life under the chosen population ethic", "unit": "species welfare-range years", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "animal-years affected x realized welfare-range change x adjusted relative welfare range", "scenarios": [ { "scenario": "RP range sensitivity only", "expression": "1 chicken-year x 100% of range x [0.002, 0.332, 0.869]", "result": [ 0.002, 0.332, 0.869 ], "unit": "human-range-equivalent years", "status": "bridge_only_not_policy_effect", "note": "This holds all policy and welfare-condition inputs at an artificial one and is not a paper impact estimate." } ] }, "natural_outcome_estimate": { "status": "policy_outcome_not_estimated", "low": null, "central": null, "high": null, "unit": "animal-years and welfare-condition changes" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "No research-attributable levy, demand or production response, welfare-condition change, or population-ethics choice is established by the public abstract." }, "income_eta_sensitivity": { "status": "not_applicable" }, "animal_welfare_sensitivity": { "status": "range_sensitivity_shown_policy_effect_not_estimated", "source": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "warning": "Values are probability-of-sentience- and subjective-experience-rate adjusted; median cases are not expected values." }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "moderate_after_publication", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "The framework is highly relevant, but the article is already published. A useful evaluation would stress-test calibration, political feasibility, population ethics, and links to observed demand rather than repeat a general conceptual review.", "qualitative_assessment": "The framework is highly relevant, but the article is already published. A useful evaluation would stress-test calibration, political feasibility, population ethics, and links to observed demand rather than repeat a general conceptual review.", "decisive_unknowns": [ "calibrated tax tables and assumptions", "demand and supply response by species", "policy window and jurisdiction", "population-ethics treatment of lives not created" ] }, "uncertainty_and_double_counting": [ "RP ranges already include sentience probability and must not be discounted again.", "A reduced quantity of meat is not identical to animals spared; supply response and animal weights matter.", "Human health, climate, and animal-welfare co-benefits need an overlap model." ], "sources": [ "https://academic.oup.com/jeea/advance-article-abstract/doi/10.1093/jeea/jvaf063/8383422", "https://rethinkpriorities.org/research-area/welfare-range-estimates/" ], "citations": [ { "label": "Journal article abstract", "url": "https://academic.oup.com/jeea/advance-article-abstract/doi/10.1093/jeea/jvaf063/8383422", "claim": "levy logic, population ethics, and species pattern" }, { "label": "Rethink Priorities welfare ranges", "url": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "claim": "adjusted 5th, 50th, and 95th percentile ranges and their limitations" } ] }, { "paper_id": 196, "title": "Price elasticities of meat, fish and plant-based meat substitutes: evidence from store-level Dutch supermarket scanner data", "url": "https://ideas.repec.org/p/tin/wpaper/20240046.html", "case_role": "expanded_screening", "cause_area": "Animal welfare", "pathway_type": "funding", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "candidate_if_elasticity_identification_is_credible", "paper_welfare_relevance": "direct_policy_parameter", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The Dutch scanner study supplies price-response parameters for meat-tax debates, but observational price identification, product substitution, and the production response determine whether a price change improves animal welfare.", "causal_chain": { "research_output": "Demand-system estimates of own- and cross-price elasticities for animal products and plant-based substitutes.", "belief_or_tool_change": "Update tax-model and retailer beliefs about consumer response to species-specific price changes.", "decision_change": "Set meat taxes, subsidies, retailer pricing, or protein-transition interventions.", "natural_outcome": "Quantity demanded and ultimately produced by species.", "welfare_bridge": "Animal-years avoided or improved x welfare change x species range, net of substitution into other animal products." }, "quantitative_inputs": [ { "name": "pork_own_price_elasticity", "description": "Estimated pork own-price elasticity", "unit": "percent quantity change per 1% price change", "provenance_tag": "source_estimate", "source_url": "https://www.econstor.eu/bitstream/10419/300328/1/1895437938.pdf", "low": -2.1, "central": -2.1, "high": -2.1, "value": -2.1, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "pbms_own_price_elasticity", "description": "Estimated plant-based-meat-substitute own-price elasticity", "unit": "percent quantity change per 1% price change", "provenance_tag": "source_estimate", "source_url": "https://www.econstor.eu/bitstream/10419/300328/1/1895437938.pdf", "low": 1.52, "central": 1.52, "high": 1.52, "value": 1.52, "note": "The authors attribute the positive sign to increasing variety; it is not a conventional causal demand response. Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "pigs_adjusted_welfare_range", "description": "Rethink Priorities probability-of-sentience- and subjective-experience-rate-adjusted welfare range for pigs", "unit": "human-range equivalents per unit time", "provenance_tag": "source_estimate", "source_url": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "low": 0.005, "central": 0.515, "high": 1.031, "value": null, "note": "The three values are RP's 5th, 50th, and 95th percentiles. They already include a sentience adjustment; do not apply one again. The median is a sensitivity case, not an expected value." }, { "name": "chickens_adjusted_welfare_range", "description": "Rethink Priorities probability-of-sentience- and subjective-experience-rate-adjusted welfare range for chickens", "unit": "human-range equivalents per unit time", "provenance_tag": "source_estimate", "source_url": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "low": 0.002, "central": 0.332, "high": 0.869, "value": null, "note": "The three values are RP's 5th, 50th, and 95th percentiles. They already include a sentience adjustment; do not apply one again. The median is a sensitivity case, not an expected value." }, { "name": "carp_adjusted_welfare_range", "description": "Rethink Priorities probability-of-sentience- and subjective-experience-rate-adjusted welfare range for carp", "unit": "human-range equivalents per unit time", "provenance_tag": "source_estimate", "source_url": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "low": 0.0, "central": 0.089, "high": 0.568, "value": null, "note": "The three values are RP's 5th, 50th, and 95th percentiles. They already include a sentience adjustment; do not apply one again. The median is a sensitivity case, not an expected value." }, { "name": "salmon_adjusted_welfare_range", "description": "Rethink Priorities probability-of-sentience- and subjective-experience-rate-adjusted welfare range for salmon", "unit": "human-range equivalents per unit time", "provenance_tag": "source_estimate", "source_url": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "low": 0.0, "central": 0.056, "high": 0.513, "value": null, "note": "The three values are RP's 5th, 50th, and 95th percentiles. They already include a sentience adjustment; do not apply one again. The median is a sensitivity case, not an expected value." }, { "name": "illustrative_pork_price_change", "description": "Price change used only to translate the reported elasticity", "unit": "percent", "provenance_tag": "analyst_assumption", "source_url": null, "low": 10, "central": 10, "high": 10, "value": null, "note": "A 10% change is a transparent arithmetic illustration, not a proposed tax or forecast." }, { "name": "research_caused_price_change", "description": "Actual price change caused by this paper", "unit": "percent by product", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "retail_to_production_pass_through", "description": "Share of demand change that changes production and animal-years", "unit": "share", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "price change x own-price elasticity, as a first-order demand approximation", "scenarios": [ { "scenario": "illustrative pork translation", "expression": "10% x -2.1", "result": -21.0, "unit": "approximate percent pork quantity-demanded change", "status": "illustrative_arithmetic", "note": "This is not a policy, production, or welfare estimate and ignores nonlinearity and cross-price effects." } ] }, "natural_outcome_estimate": { "status": "illustrative_demand_translation", "low": -21.0, "central": -21.0, "high": -21.0, "unit": "percent pork quantity demanded for a hypothetical 10% price rise" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "The screen lacks a research-attributable price change, credible policy response, supply pass-through, animal counts, cross-product substitution totals, and welfare-condition counterfactuals." }, "income_eta_sensitivity": { "status": "not_applicable" }, "animal_welfare_sensitivity": { "status": "range_sensitivity_shown_policy_effect_not_estimated", "source": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "warning": "Fish products are a mix; carp and salmon ranges are illustrations of species variation, not a classification of Dutch fish purchases." }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "moderate_to_high_for_targeted_method_review", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "The parameter is decision-relevant, while observational scanner estimates can be badly confounded. A targeted evaluation should focus on price instruments, demand-system validity, and the unusual positive plant-substitute elasticity.", "qualitative_assessment": "The parameter is decision-relevant, while observational scanner estimates can be badly confounded. A targeted evaluation should focus on price instruments, demand-system validity, and the unusual positive plant-substitute elasticity.", "decisive_unknowns": [ "instrument validity", "price and product-variety trends", "cross-price matrix", "mapping retail kilograms to animal lives and production" ] }, "uncertainty_and_double_counting": [ "First-order elasticity arithmetic is local and need not hold for a large tax.", "Substitution from beef to chicken can worsen animal welfare despite reduced meat weight.", "RP ranges already include sentience adjustment." ], "sources": [ "https://www.econstor.eu/bitstream/10419/300328/1/1895437938.pdf", "https://rethinkpriorities.org/research-area/welfare-range-estimates/" ], "citations": [ { "label": "Working paper PDF", "url": "https://www.econstor.eu/bitstream/10419/300328/1/1895437938.pdf", "claim": "data, demand system, and headline elasticities" }, { "label": "Rethink Priorities welfare ranges", "url": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "claim": "species range inputs and limitations" } ] }, { "paper_id": 650, "title": "Monetizing Animal Welfare Impacts for Benefit–Cost Analysis", "url": "https://doi.org/10.1017/bca.2024.19", "case_role": "expanded_screening", "cause_area": "Animal welfare", "pathway_type": "research_tool", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "framework_input_not_intervention_evidence", "paper_welfare_relevance": "indirect_methodological", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "This paper is a measurement framework that may improve policy analysis; it is not evidence that any policy changes animal outcomes. Its impact depends on analyst adoption and on whether the bridge changes decisions.", "causal_chain": { "research_output": "Framework combining species welfare potential, quality-of-life effects, and animal counts in benefit-cost analysis.", "belief_or_tool_change": "Enable analysts to include nonhuman welfare in a common appraisal framework.", "decision_change": "Change regulatory impact analyses, benefit-cost rankings, or program design.", "natural_outcome": "Policies selected differently because animal welfare enters appraisal; resulting animal-years and welfare states.", "welfare_bridge": "The framework itself supplies a bridge, but its realized contribution requires decision uptake and compatible normative assumptions." }, "quantitative_inputs": [ { "name": "pigs_adjusted_welfare_range", "description": "Rethink Priorities probability-of-sentience- and subjective-experience-rate-adjusted welfare range for pigs", "unit": "human-range equivalents per unit time", "provenance_tag": "source_estimate", "source_url": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "low": 0.005, "central": 0.515, "high": 1.031, "value": null, "note": "The three values are RP's 5th, 50th, and 95th percentiles. They already include a sentience adjustment; do not apply one again. The median is a sensitivity case, not an expected value." }, { "name": "chickens_adjusted_welfare_range", "description": "Rethink Priorities probability-of-sentience- and subjective-experience-rate-adjusted welfare range for chickens", "unit": "human-range equivalents per unit time", "provenance_tag": "source_estimate", "source_url": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "low": 0.002, "central": 0.332, "high": 0.869, "value": null, "note": "The three values are RP's 5th, 50th, and 95th percentiles. They already include a sentience adjustment; do not apply one again. The median is a sensitivity case, not an expected value." }, { "name": "analyses_adopting_framework", "description": "Additional policy analyses using the framework because of the paper", "unit": "analyses", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "decision_change_probability", "description": "Probability adoption changes the selected policy", "unit": "share of adopting analyses", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "animal_outcome_difference", "description": "Welfare difference between selected and counterfactual policies", "unit": "species welfare-range years", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "analyses adopting x probability of changed policy x animal welfare difference", "scenarios": [ { "scenario": "framework only", "expression": "unknown x unknown x unknown", "result": null, "unit": "human-range-equivalent welfare years", "status": "not_estimable", "note": "" } ] }, "natural_outcome_estimate": { "status": "methodological_output_only", "low": null, "central": null, "high": null, "unit": "policy analyses and animal outcomes" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "Adoption, decision sensitivity, implementation, and animal outcomes are all unmeasured at this screening stage." }, "income_eta_sensitivity": { "status": "not_applicable" }, "animal_welfare_sensitivity": { "status": "range_sensitivity_available_no_policy_effect", "source": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "warning": "A common numeraire does not settle the underlying moral assumptions." }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "moderate_for_methods_and_uptake_review", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "The article is published and openly lays out a framework. Evaluation value lies in comparing it with alternatives, testing worked regulatory cases, and checking whether analysts can use it without hiding moral assumptions.", "qualitative_assessment": "The article is published and openly lays out a framework. Evaluation value lies in comparing it with alternatives, testing worked regulatory cases, and checking whether analysts can use it without hiding moral assumptions.", "decisive_unknowns": [ "documented uptake in regulatory analysis", "decisions that reverse when animals are included", "treatment of uncertainty and equity", "choice of numeraire and population ethics" ] }, "uncertainty_and_double_counting": [ "Do not treat monetization as creating welfare or as evidence of policy uptake.", "Do not count an animal QALY and its monetized value separately.", "RP ranges already include sentience adjustment." ], "sources": [ "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/4558DB0D80DA18D4DC25788E967C0ACD/S2194588824000198a.pdf/monetizing_animal_welfare_impacts_for_benefitcost_analysis.pdf", "https://rethinkpriorities.org/research-area/welfare-range-estimates/" ], "citations": [ { "label": "Open-access journal article", "url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/4558DB0D80DA18D4DC25788E967C0ACD/S2194588824000198a.pdf/monetizing_animal_welfare_impacts_for_benefitcost_analysis.pdf", "claim": "framework, intended common-scale use, and acknowledged challenges" }, { "label": "Rethink Priorities welfare ranges", "url": "https://rethinkpriorities.org/research-area/welfare-range-estimates/", "claim": "species range inputs and limitations" } ] }, { "paper_id": 690, "title": "Who Takes the Hit? The Uneven Impacts of Generative AI on Labor Demand Across Countries", "url": "https://thedocs.worldbank.org/en/doc/07f7fe676b8331172af31de2c3fc11e1-0050062026/original/Cross-country-Labor-Demand-in-the-age-of-GenAI-0727.pdf", "case_role": "expanded_screening", "cause_area": "AI and global development", "pathway_type": "research_tool", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "candidate_if_policy_response_and_external_validity_are_specified", "paper_welfare_relevance": "potentially_direct_but_early", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The 84-country evidence addresses distribution across countries rather than assuming a universal AI labor effect. Job postings are an intermediate outcome, and current muted LMIC effects may reflect delayed adoption rather than protection.", "causal_chain": { "research_output": "Difference-in-differences and event-study analysis of 555 million job postings in 84 countries after ChatGPT.", "belief_or_tool_change": "Update beliefs about where AI-related labor-demand changes are appearing and what adoption complements mediate them.", "decision_change": "Target skills, social protection, digital infrastructure, or service-export diversification.", "natural_outcome": "Job postings, employment, earnings, and transitions by occupation and country.", "welfare_bridge": "Consumption changes over affected workers' baseline distribution, plus non-income work effects; job postings are not themselves welfare." }, "quantitative_inputs": [ { "name": "job_postings", "description": "Job postings in the analysis", "unit": "postings", "provenance_tag": "source_estimate", "source_url": "https://papers.ssrn.com/sol3/Delivery.cfm/6868421.pdf?abstractid=6868421&mirid=1", "low": 555000000, "central": 555000000, "high": 555000000, "value": 555000000, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "countries", "description": "Countries in the analysis", "unit": "countries", "provenance_tag": "source_estimate", "source_url": "https://papers.ssrn.com/sol3/Delivery.cfm/6868421.pdf?abstractid=6868421&mirid=1", "low": 84, "central": 84, "high": 84, "value": 84, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "high_income_vulnerable_posting_change", "description": "Relative change in postings for above-median-vulnerability occupations in high-income countries", "unit": "percent", "provenance_tag": "source_estimate", "source_url": "https://papers.ssrn.com/sol3/Delivery.cfm/6868421.pdf?abstractid=6868421&mirid=1", "low": -5.8, "central": -5.8, "high": -5.8, "value": -5.8, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "employment_and_earnings_effect", "description": "Change in employment and earnings represented by the posting change", "unit": "worker-years and PPP consumption", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "research_caused_policy_response", "description": "Policy change attributable to the paper", "unit": "workers reached or budget share", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "workers affected by research-caused policy x incremental employment or earnings outcome", "scenarios": [ { "scenario": "paper-reported intermediate outcome", "expression": "-5.8% postings in a high-income-country vulnerability comparison", "result": -5.8, "unit": "percent job postings", "status": "paper_reported_effect", "note": "LMIC estimates were smaller and statistically insignificant; this is not a worker welfare effect." } ] }, "natural_outcome_estimate": { "status": "intermediate_labor_demand_effect", "low": -5.8, "central": -5.8, "high": -5.8, "unit": "percent relative job-posting change in high-income countries" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "The mapping from postings to employment, wages, worker transitions, and household consumption is absent, as is any policy response caused by the paper." }, "income_eta_sensitivity": { "status": "not_estimated", "etas": [ 1.0, 1.5, 2.0 ], "formula": "u_eta(c + delta_c) - u_eta(c), where u_1(c)=ln(c/c_ref) and u_eta(c)=((c/c_ref)^(1-eta)-1)/(1-eta)", "reason": "Country and occupational heterogeneity is central, but the paper screen does not supply the distribution of affected workers' baseline consumption or earnings changes.", "required_inputs": [ "beneficiary-level baseline real consumption on a common PPP basis", "the consumption change attributable to the research-caused decision", "duration and beneficiary distribution", "a fixed reference consumption for cross-scenario normalization" ], "interpretation": "For the same 1% consumption gain at one-quarter of a reference income versus at the reference income, the relative welfare weights are 1x, 2x, and 4x for eta=1, 1.5, and 2. This is a generic curvature illustration, not a paper estimate." }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "moderate_to_high_for_new_evaluation", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "The paper is recent and could enter development-policy advice quickly. Evaluation should test identification, pre-trends, exposure measures, platform coverage, and whether the LMIC result is delay, low exposure, or measurement.", "qualitative_assessment": "The paper is recent and could enter development-policy advice quickly. Evaluation should test identification, pre-trends, exposure measures, platform coverage, and whether the LMIC result is delay, low exposure, or measurement.", "decisive_unknowns": [ "pre-trends and concurrent macro shocks", "Lightcast coverage by country", "posting-to-employment mapping", "specific policy user and response" ] }, "uncertainty_and_double_counting": [ "Do not interpret a posting decline as an equal employment decline.", "Do not count avoided job losses and preserved earnings separately if earnings are the mechanism." ], "sources": [ "https://papers.ssrn.com/sol3/Delivery.cfm/6868421.pdf?abstractid=6868421&mirid=1" ], "citations": [ { "label": "SSRN paper page and abstract", "url": "https://papers.ssrn.com/sol3/Delivery.cfm/6868421.pdf?abstractid=6868421&mirid=1", "claim": "sample, design, and headline cross-country result" } ] }, { "paper_id": 689, "title": "Artificial Intelligence Interventions to Support Frontline Healthcare Workers in Low-Resource Settings: A Review", "url": "https://thedocs.worldbank.org/en/doc/ae79099bd924fb298c1922e640f5ac4d-0050062026/original/Artificial-Intelligence-Interventions.pdf", "case_role": "expanded_screening", "cause_area": "AI and global development", "pathway_type": "research_tool", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "strong_evidence_gap_map_not_effect_estimate", "paper_welfare_relevance": "direct_domain_indirect_effect", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The review appears useful because it shows how thin causal and deployment evidence is. That can prevent premature procurement, but the welfare effect depends on which decisions the evidence gap changes.", "causal_chain": { "research_output": "Systematic review of AI interventions supporting frontline health workers in low-resource settings.", "belief_or_tool_change": "Update funders and ministries about the small causal evidence base and missing costs, infrastructure, and patient outcomes.", "decision_change": "Delay weak procurement, fund trials, choose better-validated tools, or require deployment evidence.", "natural_outcome": "Patient health, care quality, worker time, errors, and costs under changed procurement or research portfolios.", "welfare_bridge": "Health-year effects and consumption/resource costs, with avoided ineffective spending measured relative to the next-best use." }, "quantitative_inputs": [ { "name": "studies_screened", "description": "AI-health studies screened", "unit": "studies", "provenance_tag": "source_estimate", "source_url": "https://thedocs.worldbank.org/en/doc/ae79099bd924fb298c1922e640f5ac4d-0050062026/original/Artificial-Intelligence-Interventions.pdf", "low": 9762, "central": 9762, "high": 9762, "value": 9762, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "studies_with_causal_evidence", "description": "Studies providing causal evidence for frontline-worker AI in low-resource settings", "unit": "studies", "provenance_tag": "source_estimate", "source_url": "https://thedocs.worldbank.org/en/doc/ae79099bd924fb298c1922e640f5ac4d-0050062026/original/Artificial-Intelligence-Interventions.pdf", "low": 17, "central": 17, "high": 17, "value": 17, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "decisions_changed", "description": "Procurement or research decisions changed by the review", "unit": "decisions or dollars", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "incremental_health_outcome", "description": "Patient-health difference between the changed and counterfactual decisions", "unit": "healthy life-years", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "causal-evidence studies / studies screened", "scenarios": [ { "scenario": "evidence-base proportion", "expression": "17 / 9,762 x 100", "result": 0.174, "unit": "percent of initially screened studies", "status": "descriptive_arithmetic", "note": "This is a review-flow diagnostic, not intervention effectiveness or welfare impact." } ] }, "natural_outcome_estimate": { "status": "evidence_gap_measure", "low": 0.174, "central": 0.174, "high": 0.174, "unit": "percent of initially screened studies with causal evidence" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "The review count does not show whether any AI intervention improves health, how large effects are, or whether the review changes a decision." }, "income_eta_sensitivity": { "status": "not_applicable" }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "high_for_new_evaluation_if_review_methods_are_available", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "A prompt methods evaluation could check search coverage, eligibility, causal-evidence classification, deployment-context coding, and whether the policy conclusions match the thin evidence.", "qualitative_assessment": "A prompt methods evaluation could check search coverage, eligibility, causal-evidence classification, deployment-context coding, and whether the policy conclusions match the thin evidence.", "decisive_unknowns": [ "review protocol and duplicate screening", "definition of low-resource setting and frontline worker", "risk-of-bias assessment", "specific procurement or research decisions likely to change" ] }, "uncertainty_and_double_counting": [ "Do not treat a low evidence share as evidence that AI is ineffective.", "Avoid counting avoided procurement spending as welfare unless its alternative use is specified." ], "sources": [ "https://thedocs.worldbank.org/en/doc/ae79099bd924fb298c1922e640f5ac4d-0050062026/original/Artificial-Intelligence-Interventions.pdf", "https://www.worldbank.org/en/publication/wdr2026/brief/world-development-report-2026-background-papers" ], "citations": [ { "label": "World Bank background paper PDF", "url": "https://thedocs.worldbank.org/en/doc/ae79099bd924fb298c1922e640f5ac4d-0050062026/original/Artificial-Intelligence-Interventions.pdf", "claim": "review flow, evidence gaps, and deployment questions" }, { "label": "World Development Report 2026 background-paper index", "url": "https://www.worldbank.org/en/publication/wdr2026/brief/world-development-report-2026-background-papers", "claim": "status as a WDR 2026 background paper" } ] }, { "paper_id": 694, "title": "AI Adoption in Developing Countries: Lessons from the Adoption of other Advanced Digital Technologies", "url": "https://thedocs.worldbank.org/en/doc/d531dfdd855a57988e33c464d780cae1-0050062026/original/AI-adoption-in-developing-countries.pdf", "case_role": "expanded_screening", "cause_area": "AI and global development", "pathway_type": "research_tool", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "candidate_if_causal_language_and_transfer_are_bounded", "paper_welfare_relevance": "indirect_policy_learning", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The paper may help prevent simplistic leapfrogging policies by emphasizing managerial, digital, and workforce complements. Descriptive adoption evidence and Romanian AI data do not provide an intervention effect.", "causal_chain": { "research_output": "Cross-country firm-technology surveys and new Romanian evidence on AI adoption and intensive use.", "belief_or_tool_change": "Update beliefs about adoption versus productive use and the complements required in developing-country firms.", "decision_change": "Shift support among access subsidies, management upgrading, training, digital infrastructure, and firm diagnostics.", "natural_outcome": "Technology use, productivity, employment, wages, and firm survival.", "welfare_bridge": "Worker and owner consumption changes by baseline income, plus consumer surplus and fiscal costs without overlap." }, "quantitative_inputs": [ { "name": "firm_survey_sample_lower_bound", "description": "Firm-technology survey coverage", "unit": "firms", "provenance_tag": "source_estimate", "source_url": "https://thedocs.worldbank.org/en/doc/d531dfdd855a57988e33c464d780cae1-0050062026/original/AI-adoption-in-developing-countries.pdf", "low": 20000, "central": 20000, "high": null, "value": null, "note": "The source summary says more than 20,000; 20,000 is a lower bound, not a central estimate." }, { "name": "policy_reallocation", "description": "Budget shift among AI-adoption complements caused by the paper", "unit": "2026 PPP dollars", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "incremental_productivity", "description": "Productivity gain relative to the displaced support option", "unit": "percent value added", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "distributional_incidence", "description": "Allocation of gains and losses across workers, owners, and consumers", "unit": "PPP consumption by group", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "reallocated support budget x incremental productivity or worker outcome per dollar x welfare incidence", "scenarios": [ { "scenario": "descriptive screen", "expression": "no causal program effect supplied", "result": null, "unit": "productivity or employment", "status": "not_estimable", "note": "" } ] }, "natural_outcome_estimate": { "status": "descriptive_adoption_patterns_only", "low": null, "central": null, "high": null, "unit": "firm adoption and use measures" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "The evidence does not identify a policy treatment effect, and Romania may not transfer to lower-income or informal firms. Research-caused budget change and incidence are unknown." }, "income_eta_sensitivity": { "status": "not_estimated", "etas": [ 1.0, 1.5, 2.0 ], "formula": "u_eta(c + delta_c) - u_eta(c), where u_1(c)=ln(c/c_ref) and u_eta(c)=((c/c_ref)^(1-eta)-1)/(1-eta)", "reason": "Firm-level productivity is not beneficiary consumption, and the screen lacks incidence across workers, owners, consumers, and countries.", "required_inputs": [ "beneficiary-level baseline real consumption on a common PPP basis", "the consumption change attributable to the research-caused decision", "duration and beneficiary distribution", "a fixed reference consumption for cross-scenario normalization" ], "interpretation": "For the same 1% consumption gain at one-quarter of a reference income versus at the reference income, the relative welfare weights are 1x, 2x, and 4x for eta=1, 1.5, and 2. This is a generic curvature illustration, not a paper estimate." }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "moderate_to_high_for_new_evaluation", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "The report is recent and positioned to inform the World Development Report. Evaluation could distinguish description from causation and test how much older digital-technology adoption predicts AI.", "qualitative_assessment": "The report is recent and positioned to inform the World Development Report. Evaluation could distinguish description from causation and test how much older digital-technology adoption predicts AI.", "decisive_unknowns": [ "Romania-to-LMIC external validity", "measurement of intensive use", "causal evidence for complements", "actual WDR or development-bank recommendations affected" ] }, "uncertainty_and_double_counting": [ "Do not treat adoption as productive use.", "Do not add productivity, wages, profits, and consumer surplus without an incidence accounting framework." ], "sources": [ "https://thedocs.worldbank.org/en/doc/d531dfdd855a57988e33c464d780cae1-0050062026/original/AI-adoption-in-developing-countries.pdf", "https://www.worldbank.org/en/publication/wdr2026/brief/world-development-report-2026-background-papers" ], "citations": [ { "label": "World Bank background paper PDF", "url": "https://thedocs.worldbank.org/en/doc/d531dfdd855a57988e33c464d780cae1-0050062026/original/AI-adoption-in-developing-countries.pdf", "claim": "survey coverage and adoption/use argument" }, { "label": "World Development Report 2026 background-paper index", "url": "https://www.worldbank.org/en/publication/wdr2026/brief/world-development-report-2026-background-papers", "claim": "status as a WDR 2026 background paper" } ] }, { "paper_id": 7, "title": "Estimating the Economic Effects of Federally Funded R&D", "url": "https://www.nber.org/papers/w35472", "case_role": "expanded_screening", "cause_area": "Methods and research tools", "pathway_type": "research_tool", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "strong_tool_pathway_with_indirect_global_incidence", "paper_welfare_relevance": "indirect_but_potentially_large", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The CBO framework can directly affect US budget scoring, but GDP is not welfare and the global incidence of resulting R&D spillovers is central. This is a good test of the tool pathway.", "causal_chain": { "research_output": "CBO analytical frameworks for estimating economic and budget effects of federal R&D funding.", "belief_or_tool_change": "Change official budget-scoring parameters and policymaker beliefs about R&D returns.", "decision_change": "Change federal R&D appropriations or how their fiscal effects are scored.", "natural_outcome": "Public and private R&D, potential GDP, innovations, health and technology spillovers, and fiscal effects.", "welfare_bridge": "Consumption and health effects by global beneficiary distribution, net of taxation or displaced spending and research harms." }, "quantitative_inputs": [ { "name": "private_rd_crowd_in", "description": "Additional nonfederal R&D associated with one federal nondefense R&D dollar over the relevant window", "unit": "nonfederal dollars per federal dollar", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w35472", "low": 0.25, "central": 0.25, "high": 0.25, "value": 0.25, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "annual_policy_example", "description": "Illustrative annual increase in federal nondefense R&D", "unit": "nominal US dollars per year", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w35472", "low": 30000000000, "central": 30000000000, "high": 30000000000, "value": 30000000000, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "present_value_gdp_per_dollar", "description": "Reported 30-year present-value GDP gain per federal dollar under deficit-financed and deficit-neutral scenarios", "unit": "PV GDP dollars per federal dollar", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w35472", "low": 3.57, "central": null, "high": 3.82, "value": null, "note": "The endpoints are financing scenarios, not a confidence interval." }, { "name": "research_caused_appropriation_change", "description": "R&D funding change caused by this paper or its evaluation", "unit": "nominal US dollars", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "global_welfare_incidence", "description": "Share and distribution of GDP, health, and technology gains reaching people globally", "unit": "PPP consumption and health outcomes", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "research-caused funding change x incremental outcome per dollar relative to displaced spending", "scenarios": [ { "scenario": "paper policy example nominal scale", "expression": "$30 billion/year x 10 years", "result": 300000000000, "unit": "nominal federal appropriation dollars", "status": "descriptive_arithmetic", "note": "This is not present value, welfare, or a research-attributable change." } ] }, "natural_outcome_estimate": { "status": "modeled_economic_return", "low": 3.57, "central": null, "high": 3.82, "unit": "PV GDP dollars per federal dollar across two financing scenarios" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "A paper-level welfare estimate needs the actual scoring or appropriation change caused, the counterfactual use of funds, the composition and timing of innovation, global spillovers, and distributional incidence." }, "income_eta_sensitivity": { "status": "not_estimated", "etas": [ 1.0, 1.5, 2.0 ], "formula": "u_eta(c + delta_c) - u_eta(c), where u_1(c)=ln(c/c_ref) and u_eta(c)=((c/c_ref)^(1-eta)-1)/(1-eta)", "reason": "Aggregate GDP cannot be converted to welfare without its distribution across countries and people and the consumption share of gains.", "required_inputs": [ "beneficiary-level baseline real consumption on a common PPP basis", "the consumption change attributable to the research-caused decision", "duration and beneficiary distribution", "a fixed reference consumption for cross-scenario normalization" ], "interpretation": "For the same 1% consumption gain at one-quarter of a reference income versus at the reference income, the relative welfare weights are 1x, 2x, and 4x for eta=1, 1.5, and 2. This is a generic curvature illustration, not a paper estimate." }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "high_for_targeted_model_review", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "CBO describes the framework as developmental and policy-facing. Review of parameter choices, lags, crowd-in, financing, and global spillovers could influence an institutional tool rather than only scholarly beliefs.", "qualitative_assessment": "CBO describes the framework as developmental and policy-facing. Review of parameter choices, lags, crowd-in, financing, and global spillovers could influence an institutional tool rather than only scholarly beliefs.", "decisive_unknowns": [ "which CBO estimates will use the framework", "sensitivity to R&D depreciation and lags", "counterfactual financing or displaced spending", "global versus US incidence" ] }, "uncertainty_and_double_counting": [ "GDP is not welfare.", "Do not count private R&D crowd-in and the GDP attributed to it as separate final benefits.", "Health and productivity effects of the same innovation can overlap." ], "sources": [ "https://www.nber.org/papers/w35472" ], "citations": [ { "label": "NBER/CBO paper page and abstract", "url": "https://www.nber.org/papers/w35472", "claim": "framework purpose and modeled inputs and outputs" } ] }, { "paper_id": 283, "title": "Benefit–Cost Analysis of Increased Funding for Agricultural Research and Development in the Global South", "url": "https://doi.org/10.1017/bca.2023.27", "case_role": "expanded_screening", "cause_area": "Methods and research tools", "pathway_type": "funding", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "strong_quantitative_case_requires_model_audit", "paper_welfare_relevance": "direct_model_based", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "This is the most numerically developed case in the expansion, with a $61 billion present-value cost and $2.1 trillion modeled benefit. Those are intervention estimates, not the value created by this paper; the missing attribution layer remains decisive.", "causal_chain": { "research_output": "Benefit-cost model of expanded agricultural R&D in the Global South using IFPRI's IMPACT model.", "belief_or_tool_change": "Update donor and government estimates of agricultural-R&D returns.", "decision_change": "Increase or reallocate agricultural research funding.", "natural_outcome": "Agricultural output, hunger prevalence, food prices, per-capita income, and economic surplus.", "welfare_bridge": "Consumption changes across poor consumers and producers under eta sensitivity, health effects of hunger, and fiscal opportunity costs with overlap removed." }, "quantitative_inputs": [ { "name": "pv_cost", "description": "Present value of additional agricultural R&D costs over 35 years at an 8% discount rate", "unit": "2023 US dollars", "provenance_tag": "source_estimate", "source_url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/5E4F7A33E8DBCD5326D3C164AB51F84C/S2194588823000271a.pdf/benefit-cost-analysis-of-increased-funding-for-agricultural-research-and-development-in-the-global-south.pdf", "low": 61000000000, "central": 61000000000, "high": 61000000000, "value": 61000000000, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "pv_net_economic_surplus", "description": "Present value of modeled consumer and producer surplus benefits", "unit": "2023 US dollars", "provenance_tag": "source_estimate", "source_url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/5E4F7A33E8DBCD5326D3C164AB51F84C/S2194588823000271a.pdf/benefit-cost-analysis-of-increased-funding-for-agricultural-research-and-development-in-the-global-south.pdf", "low": 2100000000000, "central": 2100000000000, "high": 2100000000000, "value": 2100000000000, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "benefit_cost_ratio", "description": "Central modeled benefit-cost ratio", "unit": "benefit dollars per cost dollar", "provenance_tag": "source_estimate", "source_url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/5E4F7A33E8DBCD5326D3C164AB51F84C/S2194588823000271a.pdf/benefit-cost-analysis-of-increased-funding-for-agricultural-research-and-development-in-the-global-south.pdf", "low": 33, "central": 33, "high": 33, "value": 33, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "output_change_after_35_years", "description": "Agricultural output change relative to continued historical funding trends", "unit": "percent", "provenance_tag": "source_estimate", "source_url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/5E4F7A33E8DBCD5326D3C164AB51F84C/S2194588823000271a.pdf/benefit-cost-analysis-of-increased-funding-for-agricultural-research-and-development-in-the-global-south.pdf", "low": 10, "central": 10, "high": 10, "value": 10, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "hunger_prevalence_change", "description": "Relative reduction in hunger prevalence after 35 years", "unit": "percent", "provenance_tag": "source_estimate", "source_url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/5E4F7A33E8DBCD5326D3C164AB51F84C/S2194588823000271a.pdf/benefit-cost-analysis-of-increased-funding-for-agricultural-research-and-development-in-the-global-south.pdf", "low": -35, "central": -35, "high": -35, "value": -35, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "food_price_change", "description": "Food-price change after 35 years", "unit": "percent", "provenance_tag": "source_estimate", "source_url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/5E4F7A33E8DBCD5326D3C164AB51F84C/S2194588823000271a.pdf/benefit-cost-analysis-of-increased-funding-for-agricultural-research-and-development-in-the-global-south.pdf", "low": -16, "central": -16, "high": -16, "value": -16, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "per_capita_income_change", "description": "Per-capita income change after 35 years", "unit": "percent", "provenance_tag": "source_estimate", "source_url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/5E4F7A33E8DBCD5326D3C164AB51F84C/S2194588823000271a.pdf/benefit-cost-analysis-of-increased-funding-for-agricultural-research-and-development-in-the-global-south.pdf", "low": 4, "central": 4, "high": 4, "value": 4, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "research_caused_funding_change", "description": "Additional agricultural R&D funding attributable to this paper", "unit": "2023 US dollars", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "PV benefits / PV costs", "scenarios": [ { "scenario": "arithmetic check", "expression": "$2.1 trillion / $61 billion", "result": 34.43, "unit": "benefit-cost ratio before source rounding/model details", "status": "reproduction_check", "note": "The paper reports 33; rounded headline inputs yield 34.43, so the exact model tables are needed for reconciliation." } ] }, "natural_outcome_estimate": { "status": "modeled_intervention_outcomes", "low": null, "central": 33, "high": null, "unit": "reported benefit-cost ratio; component outputs also reported" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "The intervention model is unusually complete, but the paper's incremental contribution requires a funding shift caused by the paper. Welfare weighting needs the country/household incidence and overlap between economic surplus, income, food prices, and hunger." }, "income_eta_sensitivity": { "status": "not_estimated", "etas": [ 1.0, 1.5, 2.0 ], "formula": "u_eta(c + delta_c) - u_eta(c), where u_1(c)=ln(c/c_ref) and u_eta(c)=((c/c_ref)^(1-eta)-1)/(1-eta)", "reason": "The paper reports aggregate per-capita income and surplus; the welfare conversion requires their joint distribution across poor consumers, producers, and countries.", "required_inputs": [ "beneficiary-level baseline real consumption on a common PPP basis", "the consumption change attributable to the research-caused decision", "duration and beneficiary distribution", "a fixed reference consumption for cross-scenario normalization" ], "interpretation": "For the same 1% consumption gain at one-quarter of a reference income versus at the reference income, the relative welfare weights are 1x, 2x, and 4x for eta=1, 1.5, and 2. This is a generic curvature illustration, not a paper estimate." }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "moderate_for_reassessment_lower_for_duplicate_general_review", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "The paper is published and its large BCR depends on model choices. A targeted audit could add value if funders actively use it, especially by reconciling costs, benefits, adoption lags, spillovers, and distribution. Its age lowers timing value.", "qualitative_assessment": "The paper is published and its large BCR depends on model choices. A targeted audit could add value if funders actively use it, especially by reconciling costs, benefits, adoption lags, spillovers, and distribution. Its age lowers timing value.", "decisive_unknowns": [ "documented funder reliance", "model cost and benefit tables", "R&D lags and probability of research success", "consumer/producer incidence by income", "research-attributable budget change" ] }, "uncertainty_and_double_counting": [ "Do not add the 4% income gain, 16% price reduction, 35% hunger reduction, and $2.1 trillion surplus as independent benefits.", "Do not mistake the intervention BCR for the paper's value of information.", "The current dashboard abstract misstates $2.1 trillion as costs; the primary paper says $61 billion costs and $2.1 trillion benefits." ], "sources": [ "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/5E4F7A33E8DBCD5326D3C164AB51F84C/S2194588823000271a.pdf/benefit-cost-analysis-of-increased-funding-for-agricultural-research-and-development-in-the-global-south.pdf", "https://copenhagenconsensus.com/halftime-sustainable-development-goals-2016-2030/agricultural-rd" ], "citations": [ { "label": "Open-access journal article", "url": "https://www.cambridge.org/core/services/aop-cambridge-core/content/view/5E4F7A33E8DBCD5326D3C164AB51F84C/S2194588823000271a.pdf/benefit-cost-analysis-of-increased-funding-for-agricultural-research-and-development-in-the-global-south.pdf", "claim": "costs, benefits, BCR, and modeled outcomes" }, { "label": "Copenhagen Consensus project page", "url": "https://copenhagenconsensus.com/halftime-sustainable-development-goals-2016-2030/agricultural-rd", "claim": "public funding framing and intended policy use" } ] }, { "paper_id": 294, "title": "Adjusting for Scale-Use Heterogeneity in Self-Reported Well-Being", "url": "https://www.nber.org/papers/w31728", "case_role": "expanded_screening", "cause_area": "Methods and research tools", "pathway_type": "research_tool", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "use_existing_evaluation_and_track_adoption", "paper_welfare_relevance": "indirect_measurement_tool", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The paper may improve welfare measurement and therefore many downstream decisions. Its impact is a tool pathway: adoption, changes in estimates, and changed decisions matter more than the MTurk sample size itself.", "causal_chain": { "research_output": "Calibration questions and econometric adjustments for dimensional and general scale-use heterogeneity in subjective well-being.", "belief_or_tool_change": "Change how researchers estimate group differences and treatment effects using self-reported well-being.", "decision_change": "Revise evaluations, cost-effectiveness estimates, or policies that depend on subjective well-being.", "natural_outcome": "Number and importance of estimates materially corrected, and downstream decisions changed.", "welfare_bridge": "Outcome difference between decisions with and without corrected measurement, not the statistical correction itself." }, "quantitative_inputs": [ { "name": "study_sample", "description": "Approximate calibration-question sample", "unit": "MTurk participants", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w31728", "low": 3350, "central": 3350, "high": 3350, "value": 3350, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "studies_adopting_method", "description": "Additional studies using a validated version of the method because of the paper or evaluation", "unit": "studies", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "decision_change_probability", "description": "Probability a corrected result changes a real funding or policy decision", "unit": "share", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "outcome_difference", "description": "Welfare-relevant difference between corrected and counterfactual decisions", "unit": "natural outcome units", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "studies adopting x probability estimate changes materially x probability decision changes x outcome difference", "scenarios": [ { "scenario": "tool pathway", "expression": "unknown x unknown x unknown x unknown", "result": null, "unit": "welfare-relevant outcomes", "status": "not_estimable", "note": "" } ] }, "natural_outcome_estimate": { "status": "measurement_method_only", "low": null, "central": null, "high": null, "unit": "corrected estimates and downstream decisions" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "Method adoption, external validation, the number of decision-relevant reversals, and the consequences of those reversals are unmeasured." }, "income_eta_sensitivity": { "status": "not_applicable" }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_applicable" }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "low_for_duplicate_evaluation_high_for_update_and_adoption_tracking", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "Unjournal has already organized an interim evaluation and an author response, and the authors report collecting a more representative US sample. The best next step is an update tied to the revised paper and evidence of use.", "qualitative_assessment": "Unjournal has already organized an interim evaluation and an author response, and the authors report collecting a more representative US sample. The best next step is an update tied to the revised paper and evidence of use.", "decisive_unknowns": [ "performance on representative and cross-country samples", "comparison with anchoring vignettes and alternatives", "frequency of changed substantive conclusions", "documented use in WELLBY decisions" ] }, "uncertainty_and_double_counting": [ "Do not convert a reduction in measurement bias directly into welfare.", "Avoid multiplying adoption probabilities twice when study uptake already conditions the number of decisions affected." ], "sources": [ "https://www.nber.org/papers/w31728", "https://unjournal.pubpub.org/pub/evalsumheterogenity/release/1" ], "citations": [ { "label": "NBER paper page and abstract", "url": "https://www.nber.org/papers/w31728", "claim": "method, sample, and initial applications" }, { "label": "Unjournal interim evaluation summary", "url": "https://unjournal.pubpub.org/pub/evalsumheterogenity/release/1", "claim": "existing evaluations, manager note, and author response" } ] }, { "paper_id": 448, "title": "Accelerating Vaccine Innovation for Emerging Infectious Diseases via Parallel Discovery", "url": "https://www.nber.org/papers/w30126", "case_role": "expanded_screening", "cause_area": "Catastrophic and pandemic risk", "pathway_type": "funding", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "use_existing_evaluation_for_model_revision", "paper_welfare_relevance": "direct_model_based", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The portfolio model identifies a financing problem for emerging-disease vaccines, but it estimates financial returns rather than health or catastrophe-risk reduction. Existing Unjournal evaluations already identify substantial credibility and usefulness concerns.", "causal_chain": { "research_output": "Simulation of a portfolio of preclinical mRNA vaccine candidates for emerging infectious diseases.", "belief_or_tool_change": "Update governments and investors about private financial viability and the need for public-private support.", "decision_change": "Create subsidies, advance commitments, public portfolios, or different trial strategies.", "natural_outcome": "Vaccine candidates advanced, time to availability, infections and deaths in outbreaks, and portfolio financial returns.", "welfare_bridge": "Expected health losses and catastrophe risk averted, net of program cost; financial NPV is not health welfare." }, "quantitative_inputs": [ { "name": "vaccine_candidates", "description": "Hypothetical preclinical mRNA vaccine candidates", "unit": "candidates", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w30126", "low": 120, "central": 120, "high": 120, "value": 120, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "target_diseases", "description": "Emerging infectious diseases targeted", "unit": "diseases", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w30126", "low": 11, "central": 11, "high": 11, "value": 11, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "annualized_return", "description": "Average annualized financial return in the simulation", "unit": "percent per year", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w30126", "low": -6.0, "central": -6.0, "high": -6.0, "value": -6.0, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "net_present_value", "description": "Portfolio financial net present value", "unit": "US dollars", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w30126", "low": -9500000000, "central": -9500000000, "high": -9500000000, "value": -9500000000, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "clinical_trial_cost_share", "description": "Share of total modeled investment attributed to clinical trials", "unit": "percent", "provenance_tag": "source_estimate", "source_url": "https://www.nber.org/papers/w30126", "low": 94, "central": 94, "high": 94, "value": 94, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "research_caused_financing_change", "description": "Public or private financing change caused by the paper or evaluation", "unit": "US dollars", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "vaccine_health_effect", "description": "Incremental infections, deaths, or pandemic risk averted", "unit": "cases, deaths, or risk", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "research-caused financing change x incremental vaccine progress per dollar x outbreak health effect", "scenarios": [ { "scenario": "paper financial model", "expression": "simulated portfolio", "result": -9500000000, "unit": "financial NPV US dollars", "status": "paper_reported_model_output", "note": "The negative NPV motivates public support but is not a welfare loss estimate." } ] }, "natural_outcome_estimate": { "status": "financial_model_outputs_only", "low": null, "central": -9500000000, "high": null, "unit": "financial NPV US dollars" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "The paper does not establish a research-caused financing change, time-to-vaccine effect, outbreak probability, cases or deaths averted, or the opportunity cost of public support." }, "income_eta_sensitivity": { "status": "not_applicable" }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_estimated", "formula": "research-attributable change in catastrophe risk x welfare loss avoided", "present_generation": { "status": "not_estimated", "required_inputs": [ "research-attributable risk change over a stated horizon", "present lives or life-years at stake", "implementation costs and backfire risk" ] }, "long_future": { "status": "not_estimated", "required_inputs": [ "research-attributable existential-risk change", "future population and welfare trajectory", "population ethics", "pure time preference, if any", "background survival and value-loss risks" ] }, "reason": "The paper models vaccine finance, not a research-attributable change in pandemic or extinction risk.", "separation_rule": "Present-generation and long-future outputs must remain separate; the latter must not be smuggled into an unexplained multiplier." }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "low_for_duplicate_evaluation_high_for_follow_up_model_revision", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "Unjournal already commissioned two evaluations whose summary reports agreement on importance but substantial doubts about credibility and usefulness. Further value lies in checking a revised model or tracing actual financing use.", "qualitative_assessment": "Unjournal already commissioned two evaluations whose summary reports agreement on importance but substantial doubts about credibility and usefulness. Further value lies in checking a revised model or tracing actual financing use.", "decisive_unknowns": [ "response to existing evaluations", "probability-of-success and cost calibration", "link from financing to vaccine availability", "outbreak and health scenarios" ] }, "uncertainty_and_double_counting": [ "Do not count financial losses as welfare losses without identifying who bears them and the counterfactual.", "Do not add expected outbreak deaths and catastrophe risk if they describe overlapping states." ], "sources": [ "https://www.nber.org/papers/w30126", "https://unjournal.pubpub.org/pub/accelvax" ], "citations": [ { "label": "NBER paper page and abstract", "url": "https://www.nber.org/papers/w30126", "claim": "portfolio scale, financial returns, cost shares, and sensitivity" }, { "label": "Unjournal evaluation summary", "url": "https://unjournal.pubpub.org/pub/accelvax", "claim": "existing two evaluations and summarized credibility concerns" } ] }, { "paper_id": 359, "title": "Longer-run Economic Consequences of Pandemics", "url": "https://www.nber.org/papers/w26934", "case_role": "expanded_screening", "cause_area": "Catastrophic and pandemic risk", "pathway_type": "research_tool", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "low_timing_value_useful_context", "paper_welfare_relevance": "indirect_historical", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The historical asset-return evidence helps characterize long-run macroeconomic aftermaths, but it neither estimates pandemic prevention nor a present policy treatment. The paper is published and widely visible, which lowers incremental evaluation value.", "causal_chain": { "research_output": "Historical analysis of rates of return following major pandemics since the fourteenth century.", "belief_or_tool_change": "Update macroeconomic beliefs about persistent interest-rate and labor-scarcity effects after pandemics.", "decision_change": "Change fiscal, monetary, savings, or preparedness planning.", "natural_outcome": "Post-pandemic asset returns, wages, investment, consumption, and recovery policy.", "welfare_bridge": "Distributional consumption and health effects of changed policy, distinct from preventing a pandemic in the first place." }, "quantitative_inputs": [ { "name": "effect_duration", "description": "Duration and magnitude of depressed real returns by pandemic", "unit": "years and percentage points", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "research_caused_policy_change", "description": "Macroeconomic or preparedness decision change caused by the paper", "unit": "policy units", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "incremental_consumption_effect", "description": "Consumption difference from the changed policy", "unit": "PPP person-years", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "decisions changed x consumption or health difference relative to the without-paper policy", "scenarios": [ { "scenario": "public abstract", "expression": "persistent effects for about decades; no extractable numeric effect in this screen", "result": null, "unit": "real-return percentage points over time", "status": "not_estimable", "note": "" } ] }, "natural_outcome_estimate": { "status": "qualitative_long_run_macro_pattern", "low": null, "central": null, "high": null, "unit": "real return and wage trajectories" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "The screen lacks effect magnitudes and a current decision response. Historical comparability, selective pandemic records, and mechanism identification require detailed review." }, "income_eta_sensitivity": { "status": "not_estimated", "etas": [ 1.0, 1.5, 2.0 ], "formula": "u_eta(c + delta_c) - u_eta(c), where u_1(c)=ln(c/c_ref) and u_eta(c)=((c/c_ref)^(1-eta)-1)/(1-eta)", "reason": "Asset returns and wages have different incidence. Aggregate historical patterns do not identify consumption changes by baseline income.", "required_inputs": [ "beneficiary-level baseline real consumption on a common PPP basis", "the consumption change attributable to the research-caused decision", "duration and beneficiary distribution", "a fixed reference consumption for cross-scenario normalization" ], "interpretation": "For the same 1% consumption gain at one-quarter of a reference income versus at the reference income, the relative welfare weights are 1x, 2x, and 4x for eta=1, 1.5, and 2. This is a generic curvature illustration, not a paper estimate." }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_estimated", "formula": "research-attributable change in catastrophe risk x welfare loss avoided", "present_generation": { "status": "not_estimated", "required_inputs": [ "research-attributable risk change over a stated horizon", "present lives or life-years at stake", "implementation costs and backfire risk" ] }, "long_future": { "status": "not_estimated", "required_inputs": [ "research-attributable existential-risk change", "future population and welfare trajectory", "population ethics", "pure time preference, if any", "background survival and value-loss risks" ] }, "reason": "The paper concerns economic aftermath conditional on a pandemic, not an attributable reduction in pandemic probability.", "separation_rule": "Present-generation and long-future outputs must remain separate; the latter must not be smuggled into an unexplained multiplier." }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "low_for_new_general_evaluation", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "The 2020 working paper was published in 2022. A new evaluation is unlikely to add much unless tied to a specific preparedness or macro-policy model that still relies heavily on its estimates.", "qualitative_assessment": "The 2020 working paper was published in 2022. A new evaluation is unlikely to add much unless tied to a specific preparedness or macro-policy model that still relies heavily on its estimates.", "decisive_unknowns": [ "current operational reliance", "robustness to historical data revisions and event selection", "specific policy decision", "distribution of macroeconomic effects" ] }, "uncertainty_and_double_counting": [ "Do not treat a post-pandemic interest-rate effect as the value of preventing a pandemic.", "Avoid counting lower returns and higher wages independently without a general-equilibrium welfare model." ], "sources": [ "https://www.nber.org/papers/w26934" ], "citations": [ { "label": "NBER paper page and publication record", "url": "https://www.nber.org/papers/w26934", "claim": "historical question, qualitative result, and 2022 publication" } ] }, { "paper_id": 453, "title": "Long term cost-effectiveness of resilient foods for global catastrophes compared to artificial general intelligence safety", "url": "https://allfed.info/images/pdfs/Long_term_cost_effectiveness_of_resilien.pdf", "case_role": "expanded_screening", "cause_area": "Catastrophic and pandemic risk", "pathway_type": "funding", "assessment_level": "screening_level_lower_depth_than_six_paper_pilot", "screening_signal": "use_existing_evaluation_and_rebuild_core_model", "paper_welfare_relevance": "direct_model_based_highly_assumption_sensitive", "completeness": "causal_chain_screened_welfare_total_not_estimated", "screening_summary": "The paper performs the kind of explicit cross-cause calculation this project seeks, but the headline confidence comparisons are conditional on a contested model. Present-generation and long-future claims must remain separate.", "causal_chain": { "research_output": "Monte Carlo comparison of marginal resilient-food and AGI-safety funding under two parameterizations.", "belief_or_tool_change": "Update donors about the relative cost-effectiveness and timing of resilient-food preparedness.", "decision_change": "Reallocate marginal catastrophic-risk funding and accelerate resilient-food research.", "natural_outcome": "People fed or lives saved after agricultural catastrophes; probability of civilizational recovery; AGI-risk reduction forgone or gained.", "welfare_bridge": "Present-generation expected lives separately from long-future survival and flourishing under explicit population ethics." }, "quantitative_inputs": [ { "name": "survey_model_confidence_100m", "description": "Model confidence that resilient foods outperform AGI safety at the $100 millionth dollar", "unit": "probability", "provenance_tag": "source_estimate", "source_url": "https://www.sciencedirect.com/science/article/pii/S2212420922000176", "low": 0.84, "central": 0.84, "high": 0.84, "value": 0.84, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "survey_model_confidence_margin", "description": "Model confidence that resilient foods outperform AGI safety at the current margin", "unit": "probability", "provenance_tag": "source_estimate", "source_url": "https://www.sciencedirect.com/science/article/pii/S2212420922000176", "low": 0.98, "central": 0.98, "high": 0.98, "value": 0.98, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "author_model_confidence_100m", "description": "Author-parameterized model confidence at the $100 millionth resilient-food dollar", "unit": "probability", "provenance_tag": "source_estimate", "source_url": "https://www.sciencedirect.com/science/article/pii/S2212420922000176", "low": 0.93, "central": 0.93, "high": 0.93, "value": 0.93, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "author_model_confidence_margin", "description": "Author-parameterized model confidence at the current margin", "unit": "probability", "provenance_tag": "source_estimate", "source_url": "https://www.sciencedirect.com/science/article/pii/S2212420922000176", "low": 0.99, "central": 0.99, "high": 0.99, "value": 0.99, "note": "Repeated across low/central/high because this is a reported fixed input, not an uncertainty interval." }, { "name": "research_caused_funding_shift", "description": "Actual funding reallocation caused by the paper and existing evaluations", "unit": "US dollars", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" }, { "name": "joint_structural_uncertainty", "description": "Probability mass on models outside the paper's parameterized structure", "unit": "probability", "provenance_tag": "unknown", "source_url": null, "low": null, "central": null, "high": null, "value": null, "note": "" } ], "arithmetic": { "formula": "E[welfare with funding allocation informed by research - welfare without it]", "scenarios": [ { "scenario": "reported current-margin comparison", "expression": "model P(resilient foods more cost-effective than AGI safety)", "result": [ 0.98, 0.99 ], "unit": "probability under two model parameterizations", "status": "paper_reported_model_output", "note": "These are conditional model outputs, not credences that include full model uncertainty or estimates of the paper's own welfare contribution." } ] }, "natural_outcome_estimate": { "status": "conditional_cross_cause_model_outputs", "low": 0.98, "central": null, "high": 0.99, "unit": "model probability under two parameterizations at the current margin" }, "welfare_estimate": { "status": "not_estimated", "low": null, "central": null, "high": null, "unit": null, "reason": "The paper's probabilities are conditional on its model. Research-caused funding shifts, out-of-model uncertainty, overlap between risks, feasibility after catastrophe, and the value of the long future remain decisive." }, "income_eta_sensitivity": { "status": "not_applicable" }, "animal_welfare_sensitivity": { "status": "not_applicable" }, "catastrophic_risk_sensitivity": { "status": "not_estimated", "formula": "research-attributable change in catastrophe risk x welfare loss avoided", "present_generation": { "status": "not_estimated", "required_inputs": [ "research-attributable risk change over a stated horizon", "present lives or life-years at stake", "implementation costs and backfire risk" ] }, "long_future": { "status": "not_estimated", "required_inputs": [ "research-attributable existential-risk change", "future population and welfare trajectory", "population ethics", "pure time preference, if any", "background survival and value-loss risks" ] }, "reason": "The paper supplies a conditional model comparison, but this screen does not independently validate its risk changes or long-future welfare assumptions.", "separation_rule": "Present-generation and long-future outputs must remain separate; the latter must not be smuggled into an unexplained multiplier." }, "evaluation_value": { "status": "not_estimated", "qualitative_band": "low_for_duplicate_evaluation_high_for_model_update", "counterfactual": "paper available without a new Unjournal evaluation versus paper available with a new or updated Unjournal evaluation", "rationale": "Unjournal already evaluated the paper. The best next step is a revised open model using the evaluations, updated famine research, observed funding, and distinct present-generation and long-future outputs.", "qualitative_assessment": "Unjournal already evaluated the paper. The best next step is a revised open model using the evaluations, updated famine research, observed funding, and distinct present-generation and long-future outputs.", "decisive_unknowns": [ "changes made after existing evaluations", "updated catastrophe frequencies and food feasibility", "current resilient-food and AGI-safety funding curves", "model uncertainty beyond input distributions", "actual donor response" ] }, "uncertainty_and_double_counting": [ "Do not interpret 98-99% conditional model output as overall confidence.", "Do not add present-generation lives and long-future value without a stated population ethic.", "Avoid overlapping nuclear-war, famine, collapse, and extinction losses." ], "sources": [ "https://www.sciencedirect.com/science/article/pii/S2212420922000176", "https://unjournal.pubpub.org/pub/y2a1lbzv" ], "citations": [ { "label": "Published article abstract and highlights", "url": "https://www.sciencedirect.com/science/article/pii/S2212420922000176", "claim": "model structure, headline comparisons, and acknowledged model uncertainty" }, { "label": "Unjournal evaluation summary", "url": "https://unjournal.pubpub.org/pub/y2a1lbzv", "claim": "existing evaluation package" } ] } ] }