Displacement model

Downstream

Downstream is an open research model that connects published studies to estimate the effects of worker displacement on income, health, family stability, and children’s future earnings. Adjust a scenario and trace each result through its evidence, assumptions, and math.

Scenario

Workers displaced

Modeled worker-equivalent exposure: 2.517

Logarithmic scale above 100. Enter any exact count.

ZIP 28362

2.517 modeled worker equivalents from the source page

Uses national study effects and baselines for this worker-equivalent count.

Source

See the studies and default assumptions.

Illustrative reference calculations

Parent income · first year

$20,480

Income lost · study-year dollars

Parameter range: $19,400–$21,570

Study & calculation

Child lifetime earnings lost

$1,205,380

2024 dollars · all assumed children

Parameter range: $313,080–$2,035,050

Study & calculation

Excess deaths · 20 years

0.07239

Additional deaths · male reference cohort

Parameter range: 0.02835–0.1261

Study & calculation

Population applicability is not established. These figures are ineligible for public headlines. Two children per worker. US baselines. Mortality uses a male reference cohort. Ranges vary model parameters; they are not confidence intervals.

Place baseline comparison

The main figures keep the shared national reference for comparison across entities: two children per worker and a 20-year mortality follow-up for the male study cohort.

Modeled worker-equivalent exposure: 2.517

This place uses the shared national reference. No different local baseline replaces it.

Requested source: zip 28362

The effects of displacement

Trace each result from its study coefficient to its baseline and calculation.

Evidence from studiesBaseline calculation

2.517Workers displaced

Workers × study loss$59,832 study income baseline

2 × $2.59mChildren × SSA lifetime earnings

20-year survival differenceCDC: 494.4 deaths / 100k / yearSullivan & von Wachter · US men

Family income

$20,480

Income lost in the layoff year.

Child earnings

$1,205,380

Lifetime earnings lost for the assumed children.

Health

0.07239

Additional deaths over 20 years in the male reference cohort.

Family income

$20,480 →

Hilger (2016) · US fathers. Multiply the direct $8,138.09 loss by the worker count.

Child earnings

$1,205,380 →

Oreopoulos et al. (2008) · Canadian sons. Workers × 2 children × 9.24% × $2.59m SSA lifetime earnings.

Health

0.07239 →

Sullivan & von Wachter · US men. Apply the five odds ratios to the CDC male baseline, then compare 20-year survival.

Hover or tap a coefficient or strand to inspect its study.Tap a study effect to inspect its source. Select a result to see how its study effect and baseline produce the estimate.

Ask the model

How does a father’s job loss affect his child’s future marriage?

The model does not yet have a son-specific estimate.

The parent-divorce step is measured. The next step uses a study of daughters. A study of sons is needed to complete your question.

01 · ScenarioFather loses his job

An H-1B replacement scenario enters here as job loss. The studies measure job loss, not an H-1B-specific effect.

02 · Study effect

Rege, Telle & Votruba (2007) · Norwegian married couples · plant closures in 1995–2000, followed through 2003.

03 · Exposed families6.9% → 7.63%

The model adds 0.73 percentage points to the share of marriages dissolved.

04 · Child outcomeSon coefficient needed

The published 1.88× coefficient covers daughters. The model leaves the son result unestimated.

Study composition, exact math & sources

Model the parental dissolution share after job loss.

s′ = 1 − (1 − 0.069) ^ 1.11 = 0.07629319

Count only the additional disrupted families.

Δs = s′ − 0.069 = 0.00729319

Apply the daughters’ hazard difference to that extra share.

HR = 1 + Δs × (1.88 − 1) = 1.00641801

Daughter reference projection: 0.64% higher hazard. Parameter range: 0.19%–2.13% higher hazard.

The model treats the parent rate ratio as a hazard multiplier over the study window. It then applies the 1.88× daughters’ coefficient only to additional disrupted families. It does not multiply 1.11 × 1.88.

A hazard describes the rate at which marriages end over time. It is not a lifetime probability of divorce. An absolute probability needs a matching baseline and follow-up period.

The child study measures family-history associations. Combining the two studies is a model assumption. The daughters’ result does not transfer to a son. Worker counts do not change this per-child relative estimate.

Model assumption · The Norwegian study reports an 11% higher probability of dissolution by 2003, not an estimated hazard ratio. Charles and Stephens find a divorce response to layoffs, with no significant response to plant closings.

Range: Selected model band across contexts, not a reported confidence interval.

Model assumption · The 0.069 value is the observed dissolution share among Norwegian couples with husbands at stable plants, from 1995 to 2003. The 0.072 endpoint is the closing-plant group’s observed mean, not an uncertainty limit.

Range: Two observed cohort means; not a confidence interval or annual baseline.

Model conversion · The inspected 1986 draft’s Appendix Table A1 gives 0.63 (SE 0.09) for white women in a background-adjusted proportional-hazard model. Parent absence includes divorce, separation, never-married parents and neither parent, not divorce alone.

Range: Exponentiated coefficient ± 1.96 standard errors, rounded.

Rege, Telle, Votruba (2007)McLanahan, Bumpass (1986)

The Austrian study finds a smaller increase against stable-firm controls and no increase against mass-layoff-firm controls. Study design changes the result.

Inside the calculation

Worked example

Family income

Back to chart

2.517 × $8,138 $20,480

Parent income · first year · study-year dollars

Changing the worker input scales this total. Study effects and baselines stay fixed.

Exact math, assumptions & evidence

Model estimate: $20,484

Worker input: 2.517 worker equivalents
National reference · per worker: $8,138.09 · study-year dollars
Calculation: Worker input × national reference per worker

2.517 × $8,138.09 = $20,483.57

Direct study loss: $8,138.09 per father. Baseline: $59,832 annual income. The parameter file rounds the retained-income ratio to 0.8640. Study-year dollars.

Range = workers × $59,832 × (1 − retained-income coefficient). Coefficient support: 0.8568–0.8712. The point uses the direct study loss of $8,138.09.

Study: Hilger (2016)
Population: US fathers laid off from their jobs
Parameter: −$8,138.09 first-year income effect; Hilger Table 3
Baseline: $59,832 pre-event mean; Hilger Table 3
Main limit: study-year dollars; not comparable with the 2024-dollar child figure beside it

Scope: US laid-off fathers; first-year income in the study's own dollars

Source estimate · 1 - 8138.09/59832; coefficient +/-1.96 SE; pre-event mean held fixed. Published AEJ: Applied Economics 8(3), 2016, Table 3, printed p.265 / PDF p.19

Range: 1 - 8138.09/59832; coefficient +/-1.96 SE; pre-event mean held fixed.

Inspect every outcome and calculation

Worked example

Child earnings

Back to chart

2.517 × $478,894 $1,205,380

Child lifetime earnings lost · 2024 dollars

2 children × 9.24% earnings loss × $2.59m SSA male lifetime benchmark.

Changing the worker input scales this total. Study effects and baselines stay fixed.

Exact math, assumptions & evidence

Model estimate: $1,205,376

Worker input: 2.517 worker equivalents
National reference · per worker: $478,894.0464 · 2024 dollars
Calculation: Worker input × national reference per worker

2.517 × 2 × (1 − 0.9076) × $2,591,418.00 = $1,205,376.31

Two children per worker. SSA male lifetime earnings benchmark in 2024 dollars.

Range $313,085–$2,035,051

Study: Oreopoulos, Page & Stevens (2008)
Population: children of Canadian fathers in firm closures
Parameter: −9.24% annual earnings per displaced father
Baseline: US male lifetime-earnings benchmark (SSA, 2024 dollars)
Main limit: Canadian firm-closure setting; transport to US displacement is an applicability decision, not a measurement

Scope: Illustrative reference calculation; Canadian father–son firm closures

Model conversion · The 0.9076 factor exponentiates the −0.097 log-earnings coefficient for sons of displaced Canadian fathers. A lifetime dollar loss also requires the model’s duration and earnings baseline.

Range: Exponentiated coefficient ± 1.96 standard errors (SE 0.037).

Inspect every outcome and calculation

Worked example

Health

Back to chart

2.517 × 0.02876 0.07239

Excess deaths · 20 years · deaths

Changing the worker input scales this total. Study effects and baselines stay fixed.

Exact math, assumptions & evidence

Model estimate: 0.07239

Worker input: 2.517 worker equivalents
National reference · per worker: 0.028758728902534 · deaths
Calculation: Worker input × national reference per worker

2.517 × (S₀ − S₁) = 0.07238572065; S₀ = (1 − m)²⁰; S₁ = ∏ᵧ(1 − qᵧ); qᵧ = rᵧm / (1 − m + rᵧm)

CDC male age 45–54 annual rate: 0.004944. Annual probability m = 1 − exp(−0.004944) = 0.004931798548. Odds ratios rᵧ: 2.6696 in year 1; 1.9838 in year 2; 1.7316 in years 3–4; 1.2712 in years 5–6; 1.135 in years 7–20.

Range 0.02835–0.1261

Study: Sullivan & von Wachter (2007 working paper)
Population: high-seniority men in Pennsylvania mass layoffs
Parameter: mortality odds 2.67× in the displacement year, declining phases after
Baseline: US men aged 45–54 all-cause mortality (CDC WONDER D76, 2015–2019)
Main limit: a male reference-cohort result; a mixed population does not inherit it without a reviewed transport rule

Scope: Illustrative reference calculation; high-seniority male Pennsylvania displacement

Model conversion · The values exponentiate log-odds coefficients in Table 5, column 3 of the 2007 working paper. They are mortality odds ratios for high-seniority Pennsylvania men after mass layoffs. They do not isolate an earnings-loss mechanism.

Range: Exponentiated coefficient ± 1.96 standard errors; source offsets and cohort apply.

Inspect every outcome and calculation

Scenario assumptions, sensitivity and source records

Outcome statistics use the displayed exposure. Baselines, coefficients, annual family schedules and component records retain their source normalization.

Source-profile mortality point (deaths): 0.0723857206477
No sustained excess from year 7 (deaths): 0.0516572760096

Declared structural comparison; not a confidence interval or a model probability.

{
  "displayed_workers": 2.517,
  "source_normalization_workers": 1,
  "applicability": {
    "excess_deaths": {
      "status": "illustrative_reference_calculation",
      "eligible_for_public_headline": false,
      "target_population": null,
      "decision": null,
      "transport_assumptions": [
        "high-tenure male mass-layoff cohort, ages 45–54",
        "Pennsylvania early-1980s response transported to target place and calendar time",
        "fixed baseline age rate over follow-up; cohort aging is not modeled"
      ],
      "reason": "Target cohort applicability and transport have not been established for this outcome; display as an illustrative reference calculation."
    },
    "child_lifetime_earnings_lost_usd": {
      "status": "illustrative_reference_calculation",
      "eligible_for_public_headline": false,
      "target_population": null,
      "decision": null,
      "transport_assumptions": [
        "Canadian father-son firm-closure estimate transported to the target families",
        "US male synthetic career earnings convert a relative adult earnings effect to dollars",
        "lifetime dollars are undiscounted; this is not an annual loss"
      ],
      "reason": "Target cohort applicability and transport have not been established for this outcome; display as an illustrative reference calculation."
    }
  },
  "target_population": null,
  "projection_eligibility": {
    "child": {
      "status": "direct_estimate",
      "eligible_for_validated_direct_contrast": true,
      "reason": "direct displacement-to-child adult earnings estimate; transport still requires an applicability decision"
    },
    "grandchild": {
      "status": "structural_projection",
      "eligible_for_validated_direct_contrast": false,
      "reason": "combines a direct child estimate with an intergenerational persistence assumption; not an identified displacement intervention effect"
    },
    "greatgrandchild": {
      "status": "source_withdrawn",
      "eligible_for_validated_direct_contrast": false,
      "reason": "source withdrawn for grandchild_earnings->greatgrandchild_earnings: The source measures earnings for three generations and education for four. It does not estimate .55 causal gap transmission to a fourth earnings generation. Retaining a projection would require an explicit uncalibrated assumption, not a source-identified default."
    }
  },
  "child_applicability": {
    "source": "oreopoulos2008",
    "source_age_at_displacement": [
      10,
      14
    ],
    "source_child_sex": "Male",
    "source_parent_sex": "Male",
    "source_geography": "Canada",
    "source_event": "father firm closure 1980-1982",
    "target_child_profiles": null,
    "incompatible_children": [],
    "unknown_children": [],
    "source_income_quartile": null,
    "income_rank_definition": null,
    "eligible_for_reviewed_transport": false,
    "status": "unknown",
    "limitations": [
      "No response curve across child ages is identified by this narrow cohort.",
      "Income groups refer to father's average pre-displacement income in 1978-1979; target rank mapping requires a transport argument.",
      "A subgroup's nonsignificant coefficient is retained with its uncertainty, not replaced with zero."
    ]
  },
  "child_valuation": null,
  "assumptions": {
    "mortality_method": "odds_survival",
    "mortality_timing": "Source-offset profile: displacement, +1, +2–3, +4–5, and +6+; displacement is follow-up year 1, so offset +6 begins in follow-up year 7. Annual baseline rates use a constant-hazard rate-to-risk conversion; odds ratios modify annual risk, and survival compounds across phases. The selected baseline age rate is held fixed over the follow-up horizon.",
    "mortality_contract": {
      "method": "odds_survival",
      "timing": "source_profile",
      "baseline_conversion": "1-exp(-annual_rate)",
      "phases": [
        {
          "phase": "displacement",
          "start_offset": 0,
          "duration_years": 1,
          "first_follow_up_year": 1
        },
        {
          "phase": "offset_1",
          "start_offset": 1,
          "duration_years": 1,
          "first_follow_up_year": 2
        },
        {
          "phase": "offsets_2_3",
          "start_offset": 2,
          "duration_years": 2,
          "first_follow_up_year": 3
        },
        {
          "phase": "offsets_4_5",
          "start_offset": 4,
          "duration_years": 2,
          "first_follow_up_year": 5
        },
        {
          "phase": "offset_6_plus",
          "start_offset": 6,
          "duration_years": null,
          "first_follow_up_year": 7
        }
      ],
      "description": "Source-offset profile: displacement, +1, +2–3, +4–5, and +6+; displacement is follow-up year 1, so offset +6 begins in follow-up year 7. Annual baseline rates use a constant-hazard rate-to-risk conversion; odds ratios modify annual risk, and survival compounds across phases. The selected baseline age rate is held fixed over the follow-up horizon."
    },
    "place_application": "initial_only",
    "place_counterfactual": "national child-dollar result is the reference"
  },
  "exposure_provenance": {
    "source": "one-worker normalized reference input",
    "method": "scale by the stored warehouse worker-equivalent exposure at read time"
  },
  "place": null,
  "place_mix": [],
  "input_content_sha256": "a771d456f24b73ab4a27e51e83ec07640a0d93114d4ffc968aae925e700d843d",
  "engine_version": "0.4.1",
  "bundle_sha256": "d84162591c94d06c8148c4834ae705ff3f9a015c29955d95d26749834cd136ed",
  "exposure_input_sha256": "8b4c78f6529f0fbefd8fc3154c9e9b976065d910b4c6e9d55690e36d5c3cdeaa",
  "input_hash": null,
  "engine_sha256": "d84162591c94d06c8148c4834ae705ff3f9a015c29955d95d26749834cd136ed",
  "adapter_sha256": "5fe4b51a9cebfb247cfcd74c734a2b6cb0eae44fce1073917b01455525c6ae52",
  "outcomes": {
    "excess_deaths": {
      "point": 0.07238572064767808,
      "low": 0.02834717302873282,
      "high": 0.12609063771424764,
      "unit": "deaths",
      "baseline": {
        "value": 0.004944,
        "annual_probability": 0.004931798548274205,
        "unit": "deaths_per_person_year",
        "citation": "cdc_wonder D76 (Underlying Cause of Death, 1999-2020), Male 45-54 years, pooled 2015-2019: 514314 deaths / 104024440 person-years = 494.4 per 100k; pulled 2026-09-06 via scraper.cli wonder-mortality into warehouse wonder_mortality",
        "population": "US prime-age men (45-54)",
        "notes": "Pre-pandemic window: excludes 2020 (COVID-distorted; 2020 crude rate 601.5/100k). Single-year 2019 = 490.0/100k. Ages 45-54 match the Sullivan & von Wachter prime-age counterfactual population",
        "sex": "Male",
        "age": "45-54 years",
        "years": "2015-2019",
        "cause": "All causes",
        "geography": "US"
      },
      "steps": [
        {
          "link": "earnings_shock->mortality_peak",
          "phase": "displacement",
          "point": 2.6698,
          "low": 2.1144,
          "high": 3.3711,
          "citation": "sullivan2009",
          "tier": "canonical",
          "population_scope": "US high-seniority displaced men born 1930-1959",
          "evidence_role": "conditional"
        },
        {
          "link": "earnings_shock->mortality_offset_1",
          "phase": "offset_1",
          "point": 1.9838,
          "low": 1.3697,
          "high": 2.8732,
          "citation": "sullivan2009",
          "tier": "canonical",
          "population_scope": "US high-seniority displaced men born 1930-1959",
          "evidence_role": "conditional"
        },
        {
          "link": "earnings_shock->mortality_offsets_2_3",
          "phase": "offsets_2_3",
          "point": 1.7315,
          "low": 1.3238,
          "high": 2.2649,
          "citation": "sullivan2009",
          "tier": "canonical",
          "population_scope": "US high-seniority displaced men born 1930-1959",
          "evidence_role": "conditional"
        },
        {
          "link": "earnings_shock->mortality_offsets_4_5",
          "phase": "offsets_4_5",
          "point": 1.2712,
          "low": 0.97,
          "high": 1.6661,
          "citation": "sullivan2009",
          "tier": "canonical",
          "population_scope": "US high-seniority displaced men born 1930-1959",
          "evidence_role": "conditional"
        },
        {
          "link": "earnings_shock->mortality_sustained",
          "phase": "offset_6_plus",
          "point": 1.1354,
          "low": 1.0335,
          "high": 1.2474,
          "citation": "sullivan2009",
          "tier": "canonical",
          "population_scope": "US high-seniority displaced men born 1930-1959",
          "evidence_role": "conditional"
        }
      ],
      "components": {
        "follow_up_years": 20,
        "method": "odds_survival",
        "timing": "source_profile"
      },
      "applicability": {
        "status": "illustrative_reference_calculation",
        "eligible_for_public_headline": false,
        "target_population": null,
        "decision": null,
        "transport_assumptions": [
          "high-tenure male mass-layoff cohort, ages 45–54",
          "Pennsylvania early-1980s response transported to target place and calendar time",
          "fixed baseline age rate over follow-up; cohort aging is not modeled"
        ],
        "reason": "Target cohort applicability and transport have not been established for this outcome; display as an illustrative reference calculation."
      },
      "interpretation": "Conditional modeled expected difference; parameter bands are support envelopes, not a confidence interval.",
      "persistence_sensitivity": {
        "available": true,
        "status": "declared_structural_sensitivity",
        "eligible_for_public_headline": false,
        "unit": "deaths",
        "variants": {
          "source_profile": {
            "point": 0.07238572064767808,
            "low": 0.02834717302873282,
            "high": 0.12609063771424764,
            "evidence_status": "source_profile_with_transport_assumptions",
            "citations": [
              "sullivan2009"
            ]
          },
          "no_sustained_excess": {
            "point": 0.05165727600957149,
            "low": 0.023133408129215802,
            "high": 0.08900433094168766,
            "evidence_status": "structural_assumption",
            "motivating_citations": [
              "eliasonstorrie2009"
            ],
            "assumption": "Retain source odds through offset +5; set offset +6 onward OR=1.",
            "first_changed_follow_up_year": 7,
            "fixed_late_odds_ratio": 1
          }
        },
        "source_minus_no_sustained_point": 0.02072844463810658,
        "structural_point_range": {
          "low": 0.05165727600957149,
          "high": 0.07238572064767808,
          "interpretation": "Range of point estimates across these two assumptions; not a confidence interval or a bound on the true effect."
        },
        "shared_inputs": [
          "exposure",
          "resolved mortality baseline",
          "follow-up horizon",
          "source odds through offset +5"
        ],
        "limitations": [
          "No probabilities or model-average weights are assigned to the structures.",
          "A nonsignificant long-run estimate does not establish an exactly zero effect.",
          "The Swedish study differs in population, design, effect scale and follow-up; this hybrid does not reproduce its estimates or substitute its hazard ratios for odds ratios.",
          "The cutoff follows the source profile's offset +6 phase, not the Swedish study's windows; persistence beyond its 12-year follow-up is an assumption in either direction.",
          "Baseline aging and target-population transport remain unresolved in both structures."
        ]
      }
    },
    "child_lifetime_earnings_lost_usd": {
      "point": 1205376.3147888004,
      "low": 313084.7570880003,
      "high": 2035050.9210720002,
      "steps": [
        {
          "link": "displacement->child_earnings",
          "kind": "direct",
          "citation": "oreopoulos2008",
          "tier": "EXACT",
          "param_point": 0.9076,
          "param_low": 0.844,
          "param_high": 0.976,
          "value_point": 0.9076,
          "value_low": 0.844,
          "value_high": 0.976,
          "causal_role": "direct_displacement_estimate",
          "population_scope": "Canada 39k father-son pairs firm closures",
          "evidence_role": "conditional"
        }
      ],
      "unit": "usd_2024",
      "baseline": {
        "value": 2591418,
        "unit": "usd_2024",
        "citation": "SSA Annual Statistical Supplement 2025, Table 4.B6 (CWHS/Master Earnings File, 1 percent sample): male median earnings by age, 2023 (preliminary) — synthetic career ages 20-64 = 5x(24218+42561+52485+60423+64962+67506+68160+64854)+2x60666+3x56666 = 2517175 usd_2023; CPI-U annual averages 304.702->313.689 (BLS series CUUR0000SA0, pulled 2026-09-06) -> 2591418 usd_2024; warehouse ssa_median_earnings (pulled via ssa-earnings-download)",
        "population": "US male median lifetime earnings",
        "notes": "Derivation declared: the cross-sectional age profile summed over a 45-year working life (ages 20-64) is a synthetic cohort — the sum of age-band medians approximates rather than equals the median of individual lifetime sums; median-of-workers semantics excludes non-working years; 2023 medians are preliminary (SSA replaces them with final data in a later edition)"
      },
      "applicability": {
        "status": "illustrative_reference_calculation",
        "eligible_for_public_headline": false,
        "target_population": null,
        "decision": null,
        "transport_assumptions": [
          "Canadian father-son firm-closure estimate transported to the target families",
          "US male synthetic career earnings convert a relative adult earnings effect to dollars",
          "lifetime dollars are undiscounted; this is not an annual loss"
        ],
        "reason": "Target cohort applicability and transport have not been established for this outcome; display as an illustrative reference calculation."
      },
      "interpretation": "Conditional modeled expected difference; parameter bands are support envelopes, not a confidence interval."
    }
  }
}

All reference outcomes · 29 calculations

Cohort totals scale with the worker input. Study rates and effects keep their source populations. Structural projections connect studies through stated assumptions. Open any result to inspect the math.

Percentages show relative change from the study baseline. Effects with no baseline stay in percentage points. Open a row for the baseline, full calculation, and source values. P marks an illustrative percentile on a normal reference distribution.

Current input: 2.517 worker equivalents

29 of 29 outcomes

Earnings 7

Child lifetime earnings lostScaled total · 2024 dollars · 2 children per worker$1,205,380

What this means: The model applies the child earnings effect to two children per worker and the US male lifetime earnings benchmark. This is a conditional total, not an observed loss.
Comparison → model: $0 → $1,205,380
Exact result: $1,205,376.31
Calculation: 2.517 × 2 × (1 − 0.9076) × $2,591,418.00 = $1,205,376.31
Inputs and units: Two children per worker. SSA male lifetime earnings benchmark in 2024 dollars.
Study population: Canada 39k father-son pairs firm closures
Attributed studies: Oreopoulos, Page, Stevens (2008)

Range: Exponentiated coefficient ± 1.96 standard errors (SE 0.037).

Parents’ income lost · layoff yearScaled total · study-year dollars$20,480

What this means: The model applies the verified $8,138.09 first-year income loss from Hilger Table 3 to the stated worker-equivalent exposure. The pre-event mean is $59,832. Study-year dollars remain separate from 2024-dollar child earnings. Population transport is not established.
Comparison → model: $0 → $20,480
Exact result: $20,483.57
Calculation: 2.517 × $8,138.09 = $20,483.57
Inputs and units: Direct study loss: $8,138.09 per father. Baseline: $59,832 annual income. The parameter file rounds the retained-income ratio to 0.8640. Study-year dollars.
Study population: US laid-off fathers, annual income in the layoff event window (base $59,832)
Attributed studies: Hilger (2016)

Source estimate · 1 - 8138.09/59832; coefficient +/-1.96 SE; pre-event mean held fixed. Published AEJ: Applied Economics 8(3), 2016, Table 3, printed p.265 / PDF p.19

Range: 1 - 8138.09/59832; coefficient +/-1.96 SE; pre-event mean held fixed.

Workers’ long-run earningsCompared with no job loss−20%

What this means: Workers retain 80% of comparison earnings in the model. This row shows the earnings loss, not the share of workers who lose a job.
Comparison → model: 100% → 80%
Exact result: 80%
Parameter support: Coefficient support: 0.75 – 0.85
Calculation: 0.80 × comparison earnings
Inputs and units: Long-run earnings relative to the same worker without displacement. No worker lifetime-dollar total uses this coefficient.
Study population: US high-tenure displaced men
Attributed studies: Jacobson, LaLonde, Sullivan (1992)Oreopoulos, Page, Stevens (2008)Davis, von Wachter (2011)

Model assumption · The 0.80 retained-earnings factor is a cross-study selection. Jacobson reports about 25% losses; Davis and von Wachter describe losses of 15–20% after 20 years. These are different cohorts and follow-up periods.

Range: Cross-study model band, not a confidence interval.

Children’s adult earningsAfter a father loses his job−9.2%

What this means: Children retain 90.76% of comparison adult earnings. This direct study estimate also supplies the child lifetime-dollar calculation above.
Comparison → model: 100% → 90.76%
Exact result: 90.76%
Parameter support: Coefficient support: 0.844 – 0.976
Calculation: 0.9076 × comparison earnings; loss = 1 − 0.9076 = 9.24%
Inputs and units: Direct displacement-to-child estimate. The model does not multiply it by the worker earnings loss.
Study population: Canada 39k father-son pairs firm closures
Attributed studies: Oreopoulos, Page, Stevens (2008)

Range: Exponentiated coefficient ± 1.96 standard errors (SE 0.037).

Lifetime earnings lost per grandchildModel projection · one descendant$131,700

What this means: The model projects part of the child earnings loss into the next generation. The amount is for one grandchild. It is not a total for this employer.
Comparison → model: $0 → $131,700
Exact result: $131,695.86
Parameter support: Coefficient support: 0.4 – 0.6
Calculation: g₂ = 1 − 0.55 × (1 − 0.9076) = 0.94918; loss = (1 − g₂) × $2,591,418.00
Inputs and units: One descendant. The same national earnings benchmark applies. Descendant counts are not assumed.
Study population: US IGE literature
Attributed studies: Solon (1992)Corak (2013)Chetty, Hendren, Kline, Saez (2014)

Range: Exponentiated coefficient ± 1.96 standard errors (SE 0.037).

Model assumption · The 0.55 factor is a structural persistence assumption. Solon and Corak provide intergenerational-earnings context; Chetty’s baseline IGE is 0.344, distinct from its rank–rank slope. These associations do not measure transmission of a displacement-induced earnings change.

Range: Selected literature band, not a confidence interval.

Lifetime earnings lost per great-grandchildEstimate withdrawnSource unverified · estimate withdrawnEstimate withdrawn

What this means: This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.
Comparison → model: Absolute baseline not supplied
Exact result: Estimate withdrawn
Calculation: This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.
Inputs and units: This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.
Study population: No verified population and outcome match for this numerical input.
Attributed studies

Range: Exponentiated coefficient ± 1.96 standard errors (SE 0.037).

Range: Selected literature band, not a confidence interval.

Source unverified · estimate withdrawn · The source measures earnings for three generations and education for four. It does not estimate .55 causal gap transmission to a fourth earnings generation. Retaining a projection would require an explicit uncalibrated assumption, not a source-identified default.

Range: Assumed persistence band; source support unresolved.

Adult earnings · through lower birth weightModel projection · separate from direct child earnings−0.5%

What this means: The model links lower birth weight to later earnings. This separate projection is not added to the direct child earnings loss. It needs a matching infant population to give a total.
Comparison → model: 100% → 99.53%
Exact result: 99.53% of comparison earnings
Parameter support: Coefficient support: 0.0216 – 0.1784
Calculation: 0.9541 ^ (0.1) = 0.9953123424
Inputs and units: Structural dose-response connection. It stays separate from the direct child earnings estimate and all-cause mortality total.
Study population: Norway twin sample, twin fixed effects on ln(birth weight), cohorts 1967-1988 (military conscription + earnings registries); IZA DP 1864 Table 5 'Ln(Earnings) FT', all-twins FE column
Attributed studies: Black, Devereux, Salvanes (2005)

Model conversion · The factor exponentiates the −0.047 log-birth-weight coefficient in IZA 5213. The sample includes births after displacement, including exposure years before birth; it is not an estimate restricted to in-utero job loss.

Range: Exponentiated coefficient ± 1.96 standard errors (SE 0.023).

Source estimate · The 2005 working paper’s Table 5 reports 0.10 (SE 0.04) for log full-time earnings on log birth weight in a twin fixed-effects sample. This is not a causal bridge from parental displacement to adult earnings.

Range: Coefficient ± 1.96 standard errors.

Education 8

Children enrolled in college · ages 18–22US · father’s job loss−1.1%

What this means: Hilger Table 3 reports an enrollment change of −0.432 percentage points (SE 0.094), with a 40.66% pre-event mean. The model rounds the retained-enrollment ratio. Enrollment is not eventual graduation.
Comparison → model: 40.66% → 40.23%
Exact result: 40.23%
Parameter support: Coefficient support: 0.9848 – 0.9939
Calculation: 40.66% × 0.9894 = 40.23%; change = -0.43 percentage points
Inputs and units: Pinned model denominator 0.0785. The source labels it a regression constant, not a verified control-group mean. The model value is retained for review.
Study population: US children of ~7M laid-off fathers, ages 18-22 enrollment, 1999-2009 administrative data
Attributed studies: Hilger (2016)

Source estimate · 1 - 0.432/40.66, with coefficient +/-1.96 SE and the pre-event mean held fixed. Published AEJ: Applied Economics 8(3), 2016, Table 3, printed p.265 / PDF p.19

Range: 1 - 0.432/40.66, with coefficient +/-1.96 SE and the pre-event mean held fixed.

Children who repeat a school gradeUS · household head’s job loss+14.7%

What this means: The share repeating a grade rises from 5.5% to 6.31%, an increase of 0.81 percentage points. The study measures children ages 5–19, at least one year after job loss.
Comparison → model: 5.5% → 6.31%
Exact result: 6.31%
Parameter support: Coefficient support: 1.0119 – 1.2827
Calculation: 5.5% × 1.1473 = 6.31%; change = 0.81 percentage points
Inputs and units: Model denominator 0.055: the source’s sample average, not a control-group mean. The relative probability is a model conversion.
Study population: US children 5-19 (SIPP 1996/2001/2004 panels), head-of-household involuntary job loss (fired/discharged, employer sold/bankrupt, slack work), child fixed effects, N=167,449 child-years / 2,170 displaced children / 1,196 job losses
Attributed studies: Stevens, Schaller (2009)

Model conversion · The model divides the 0.0081 grade-repetition probability effect by the 0.055 sample average. This is a relative-probability conversion, not an estimated hazard ratio or a measured test-score effect.

Range: Source coefficient ± 1.96 standard errors, divided by a fixed sample average.

Children with no grade-9 resultsEstimate withdrawnSource mismatch · estimate withdrawnEstimate withdrawn

What this means: This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.
Comparison → model: Absolute baseline not supplied
Exact result: Estimate withdrawn
Calculation: This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.
Inputs and units: This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.
Study population: No verified population and outcome match for this numerical input.
Attributed studies

Source mismatch · estimate withdrawn · Retired identifier; use displacement_event->exam_noncompletion_change_pp. IZA DP 16367 Table 2 Panel A col.2: probability difference .0054 (SE .0027), expressed as .54 percentage points. Regression constant .0785 is not an admitted control mean; no hazard ratio, relative-risk conversion or displaced-worker total. Published-version reconciliation remains open.

Range: Coefficient uncertainty scaled by an unsupported denominator; not an estimated hazard ratio.

Children finishing high school by 19US · single mother’s job loss−4.9%

What this means: The completion rate falls from 76% to 72.3%, a decrease of 3.7 percentage points. The study’s uncertainty range also allows no decrease.
Comparison → model: 76% → 72.3%
Exact result: 72.3%
Parameter support: Coefficient support: -0.0801 – 0.0061
Calculation: 0.76 + (-0.037) = 0.723
Inputs and units: Study control mean. Children of displaced single mothers. This rate does not apply to all assumed children.
Study population: children of single mothers, US NLSY79 (NLSY79 women's children), MATERNAL job displacement at child age 0-17 (PSM kernel TT); outcome = HS completion by age 19, control mean 0.760
Attributed studies: Brand, Simon Thomas (2014)

Source estimate · Table 3’s kernel-matching estimate is −0.037 for high-school completion by age 19; its interval crosses zero. The sample is children of single mothers; selection on observed characteristics is not random assignment. The displayed baseline is the unadjusted comparison-group mean in Table 1; applying the matched effect to it is a model choice.

Range: Coefficient ± 1.96 standard errors (SE 0.022).

Children ever attending college by 21–22US · single mother’s job loss−14.8%

What this means: The share ever attending college falls from 42.7% to 36.4%, a decrease of 6.3 percentage points. This measures any attendance by this age, not current enrollment.
Comparison → model: 42.7% → 36.4%
Exact result: 36.4%
Parameter support: Coefficient support: -0.1081 – -0.0179
Calculation: 0.427 + (-0.063) = 0.364
Inputs and units: Study control mean. Children of displaced single mothers. This rate does not apply to all assumed children.
Study population: same scope as the hs_completion row; outcome = some college attendance by age 21-22, control mean 0.427
Attributed studies: Brand, Simon Thomas (2014)

Source estimate · Table 3’s kernel-matching estimate is −0.063 for college attendance in young adulthood. The sample is children of single mothers; selection on observed characteristics is not random assignment. The displayed baseline is the unadjusted comparison-group mean in Table 1; applying the matched effect to it is a model choice.

Range: Coefficient ± 1.96 standard errors (SE 0.023).

Children earning a bachelor’s degree by 25US · single mother’s job loss−21.3%

What this means: The completion rate falls from 16.9% to 13.3%, a decrease of 3.6 percentage points. This study concerns children of single mothers.
Comparison → model: 16.9% → 13.3%
Exact result: 13.3%
Parameter support: Coefficient support: -0.0674 – -0.0046
Calculation: 0.169 + (-0.036) = 0.133
Inputs and units: Study control mean. Children of displaced single mothers. This rate does not apply to all assumed children.
Study population: same scope as the hs_completion row; outcome = college completion (BA) by age 25, control mean 0.169
Attributed studies: Brand, Simon Thomas (2014)

Source estimate · Table 3’s kernel-matching estimate is −0.036 for college completion in young adulthood. The sample is children of single mothers; selection on observed characteristics is not random assignment. The displayed baseline is the unadjusted comparison-group mean in Table 1; applying the matched effect to it is a model choice.

Range: Coefficient ± 1.96 standard errors (SE 0.016).

Children’s mathematics grades · grade 9Denmark · teacher-assessed grades−0.0206 SDP_49.2

What this means: Teacher-assessed mathematics grades fall after a parent’s job loss. SD means standard deviation, a unit of score spread. This is a score change, not a percentage of children.
Percentile reference: The percentile starts at the comparison mean (50th percentile) and applies the SD change to a standard normal distribution. It assumes a normal score distribution with unchanged spread. It is an illustrative position, not an observed rank or a prediction for an individual.
Comparison → model: 0 SD → -0.0206 SD
Exact result: -0.0206 SD change
Parameter support: Coefficient support: -0.0402 – -0.001
Calculation: Study coefficient = -0.0206; support = [-0.0402, -0.001]
Inputs and units: Grade-9 mathematics score. The estimate is conditional on taking the exam.
Study population: children exposed to parental plant closure age 0-16 (average across exposure stages), Denmark, conditional on taking grade-9 exams; mathematics
Attributed studies: Bingley, Cappellari, Ovidi (2023)

Source estimate · Table 4 reports −0.0206 SD (SE 0.0100) for teacher-assessed mathematics grades in the Danish sample. The separate standardized-test estimate is smaller and not significant. These outcomes are not interchangeable.

Range: Coefficient ± 1.96 standard errors.

Grandchildren’s cognitive scoresModel projection · different tests−0.007828 SDP_49.7

What this means: The model links Danish child mathematics grades to a Norwegian father–son IQ relationship. The tests differ. This is a projection, not a measured job-loss effect on grandchildren.
Percentile reference: The percentile starts at the comparison mean (50th percentile) and applies the SD change to a standard normal distribution. It assumes a normal score distribution with unchanged spread. It is an illustrative position, not an observed rank or a prediction for an individual.
Comparison → model: 0 SD → -0.007828 SD
Exact result: -0.007828 SD
Parameter support: Coefficient support: 0.38 – 0.42
Calculation: -0.0206 × 0.38 = -0.007828 SD
Inputs and units: A structural link from Danish mathematics scores to a Norwegian father–son IQ persistence slope. The measures differ.
Study population: Norway male conscripts (IQ at age 18, father-son)
Attributed studies: Black, Devereux, Salvanes (2008)

Range: Coefficient ± 1.96 standard errors.

Model assumption · The source reports a 0.38 father–son IQ correlation in Norway. The authors give it no causal interpretation. Applying it to an induced achievement change is a structural assumption, not a measured intergenerational treatment effect.

Range: Selected literature band, not a confidence interval.

Health 11

Expected excess deaths · 20 yearsScaled total · high-seniority male reference cohort0.07239

What this means: The model compares 20-year survival with and without job loss. The difference gives expected additional deaths. It does not count observed deaths or establish that this employer caused them.
Comparison → model: 0.237 → 0.3094
Exact result: 0.07239
Calculation: 2.517 × (S₀ − S₁) = 0.07238572065; S₀ = (1 − m)²⁰; S₁ = ∏ᵧ(1 − qᵧ); qᵧ = rᵧm / (1 − m + rᵧm)
Inputs and units: CDC male age 45–54 annual rate: 0.004944. Annual probability m = 1 − exp(−0.004944) = 0.004931798548. Odds ratios rᵧ: 2.6696 in year 1; 1.9838 in year 2; 1.7316 in years 3–4; 1.2712 in years 5–6; 1.135 in years 7–20.
Study population: US high-seniority displaced men born 1930-1959
Attributed studies: Sullivan, von Wachter (2007)

Range: Exponentiated coefficient ± 1.96 standard errors; source offsets and cohort apply.

Adult children’s depression symptomsUS · ages 25–29 · mother’s job loss+0.025 score

What this means: The average symptom score rises from 0.208 to 0.233 on a 0–1 scale. Higher scores mean more symptoms. This study concerns children of single mothers. It does not measure the share diagnosed with depression.
Comparison → model: 0.208 → 0.233
Exact result: 0.233
Parameter support: Coefficient support: 0.0034 – 0.0466
Calculation: 0.208 + (0.025) = 0.233
Inputs and units: Study control mean on a 0–1 CESD symptom scale. This is an index, not a probability.
Study population: same scope as the hs_completion row; outcome = CESD 7-item depressive-symptom index (0-1) at ages 25-29, control mean 0.208
Attributed studies: Brand, Simon Thomas (2014)

Source estimate · Table 3’s kernel-matching estimate is +0.025 on the 0–1 CESD index at ages 25–29, not a standard-deviation change. The sample is children of single mothers; selection on observed characteristics is not random assignment. The displayed baseline is the unadjusted comparison-group mean in Table 1; applying the matched effect to it is a model choice.

Range: Coefficient ± 1.96 standard errors (SE 0.011).

Babies’ birth weightUS · father’s job loss−4.6%

What this means: Birth weight is 4.59% lower in the source study. A birth-weight baseline and a matching infant population are needed to convert this to grams or birth counts.
Comparison → model: 100% → 95.41%
Exact result: 95.41% of comparison birth weight
Parameter support: Coefficient support: 0.912 – 0.9981
Calculation: Study coefficient = 0.9541; support = [0.912, 0.9981]
Inputs and units: Birth-weight ratio. An infant cohort and birth-weight baseline are needed for grams or birth counts.
Study population: US PSID children 1985+, husband displacement, mother fixed effects
Attributed studies: Lindo (2010)

Range: Exponentiated coefficient ± 1.96 standard errors (SE 0.023).

Death rate · heart & circulatory diseaseDanish men · first 4 years+54%

What this means: The death hazard is 54% higher than in the comparison group. Hazard means the death rate among people still alive. This is not a death count or a 54% chance of dying.
Comparison → model: 1× → 1.54×
Exact result: 1.54 × comparison hazard
Parameter support: Coefficient support: 1.29 – 1.83
Calculation: Study coefficient = 1.54; support = [1.29, 1.83]
Inputs and units: Cause-specific study hazard ratio. A hazard ratio is not a probability. No cause-specific count is added to all-cause deaths.
Study population: Danish full-time male workers age 20-60 with strong labor-market attachment displaced by plant closures 1986-2002 (base years 1985-2001; propensity-score weighted matched controls; Danish full registers 1980-2006)
Attributed studies: Browning, Heinesen (2011)

Source estimate · The inspected 2011 report gives a cumulative hazard ratio over years 1–4 for Danish men with strong labor-market attachment after plant closure. It is not an annual hazard or an odds ratio.

Range: Reported 95% confidence interval for the specified cause and follow-up window.

Death rate · alcohol-related causesDanish men · first 4 years+62%

What this means: The alcohol-related death hazard is 62% higher than in the comparison group. Hazard means the death rate among people still alive. These deaths are not added to the all-cause total above.
Comparison → model: 1× → 1.62×
Exact result: 1.62 × comparison hazard
Parameter support: Coefficient support: 1.09 – 2.41
Calculation: Study coefficient = 1.62; support = [1.09, 2.41]
Inputs and units: Cause-specific study hazard ratio. No cause-specific count is added to all-cause deaths.
Study population: Danish full-time male workers age 20-60 with strong labor-market attachment displaced by plant closures 1986-2002 (base years 1985-2001; propensity-score weighted matched controls; Danish full registers 1980-2006)
Attributed studies: Browning, Heinesen (2011)

Range: Reported 95% confidence interval for the specified cause and follow-up window.

Death rate · injuries & suicideDanish men · first 4 years+53%

What this means: External causes include injuries, suicide and traffic deaths. The death hazard is 53% higher than in the comparison group. These deaths are not added to the all-cause total above.
Comparison → model: 1× → 1.53×
Exact result: 1.53 × comparison hazard
Parameter support: Coefficient support: 1.15 – 2.04
Calculation: Study coefficient = 1.53; support = [1.15, 2.04]
Inputs and units: Cause-specific study hazard ratio. No cause-specific count is added to all-cause deaths.
Study population: Danish full-time male workers age 20-60 with strong labor-market attachment displaced by plant closures 1986-2002 (base years 1985-2001; propensity-score weighted matched controls; Danish full registers 1980-2006)
Attributed studies: Browning, Heinesen (2011)

Range: Reported 95% confidence interval for the specified cause and follow-up window.

Partners’ mental-health scoresGermany · after spouse’s job loss−0.194 SDP_42.3

What this means: Partners’ mental-health scores fall about 11 months after job loss. Lower scores mean worse mental health. SD is a score-spread unit, not a percentage of partners.
Percentile reference: The percentile starts at the comparison mean (50th percentile) and applies the SD change to a standard normal distribution. It assumes a normal score distribution with unchanged spread. It is an illustrative position, not an observed rank or a prediction for an individual.
Comparison → model: 0 SD → -0.194 SD
Exact result: -0.194 SD change
Parameter support: Coefficient support: -0.327 – -0.061
Calculation: Study coefficient = -0.194; support = [-0.327, -0.061]
Inputs and units: MCS score change in standard-deviation units. Lower scores mean worse mental health.
Study population: spouses of displaced workers, German couples (SOEP 2002-2010), 109 displaced-worker couples vs >14,000 entropy-balanced controls; MCS normed mean 50 SD 10
Attributed studies: Marcus (2012)

Model conversion · Table 3’s spouse mental-health effect is −1.94 MCS points (SE 0.68). The source normalizes the score to SD 10; division by 10 gives −0.194 SD. The follow-up is about one year after unemployment following closure.

Range: Coefficient ± 1.96 standard errors, divided by 10.

Distress among unemployed adultsCompared with employed adults+0.51 SDP_69.5

What this means: Unemployed adults report more distress than employed adults. This study comparison does not establish that job loss causes the whole difference. SD is a score-spread unit.
Percentile reference: The percentile starts at the comparison mean (50th percentile) and applies the SD change to a standard normal distribution. It assumes a normal score distribution with unchanged spread. It is an illustrative position, not an observed rank or a prediction for an individual.
Comparison → model: 0 SD → 0.51 SD
Exact result: 0.51 SD difference
Parameter support: Coefficient support: 0.47 – 0.54
Calculation: Study coefficient = 0.51; support = [0.47, 0.54]
Inputs and units: Cross-sectional unemployed-versus-employed comparison. This is not an identified displacement effect.
Study population: involuntary unemployed vs employed adults (26 countries, 1950-2004); cross-sectional meta k=315 samples, N=209,379, outliers excluded
Attributed studies: Paul, Moser (2009)

Source estimate · Overall outlier-excluded cross-sectional association, not the factory-closure subset or a causal job-loss effect. Published JVB 74, 2009, Table 1 Overall (out. ex), printed p.271 / PDF p.8

Range: Overall outlier-excluded cross-sectional association, not the factory-closure subset or a causal job-loss effect.

Children reported with depression or anxietyUS · father’s job loss · ages 6–17+0.8 percentage points

What this means: The reported share rises by 0.8 percentage points. That is an additional 0.8 reports per 100 children in the study population, not a 0.8% relative increase.
Comparison → model: Absolute baseline not supplied
Exact result: 0.8 percentage-point change
Parameter support: Coefficient support: 0.0002 – 0.0158
Calculation: Study coefficient = 0.008; support = [0.0002, 0.0158]
Inputs and units: Paternal job-loss study. A compatible baseline and child cohort are needed for counts.
Study population: children ages 6-17, US MEPS, PATERNAL job loss (tenured-father sample, >=1y tenure), individual-FE LPM; outcome = child depression/anxiety report
Attributed studies: Schaller, Zerpa (2015)

Source estimate · Table 2’s probability effect is +0.008: 0.8 percentage points, not 0.008 percentage points. The inspected MEPS sample is children aged six and older with a tenured displaced father.

Range: Coefficient ± 1.96 standard errors (SE 0.004).

Adult diabetes odds · through lower birth weightModel projection · needs an infant cohort+4.6%

What this means: The model links lower birth weight to adult type-2 diabetes odds. Odds compare people with and without an outcome. This is not a disease probability or a count of future cases.
Comparison → model: 1× → 1.046322×
Exact result: 1.046322 × comparison
Parameter support: Coefficient support: -1.3834 – -0.5401
Calculation: 0.9541 ^ (-0.9637) = 1.046322014
Inputs and units: Structural dose-response connection. It stays separate from the direct child earnings estimate and all-cause mortality total.
Study population: adults, pooled random-effects dose-response, 32 studies / ~7.6M participants (Knop et al. JAHA 2018, 7(23):e008870, OR per +1 kg birth weight; read via PMC6405546 full text)
Attributed studies: Knop, Geng, Gorny, Ding, Li, Ley, Huang (2018)

Range: Exponentiated coefficient ± 1.96 standard errors (SE 0.023).

Model conversion · The meta-analysis reports odds ratio 0.78 [0.70, 0.87] per 1 kg. The model converts this using ln(OR) / ln(4.4/3.4). The 3.4 kg baseline and constant elasticity are assumptions; odds are not hazards.

Range: Transformed reported 95% odds-ratio limits, conditional on the model’s assumed baseline.

Adult heart-disease odds · through lower birth weightModel projection · needs an infant cohort+3.3%

What this means: The model links lower birth weight to adult cardiovascular disease odds. This is not a disease probability or a count of future cases. It is not added to the all-cause death total.
Comparison → model: 1× → 1.033408×
Exact result: 1.033408 × comparison
Parameter support: Coefficient support: -0.8173 – -0.585
Calculation: 0.9541 ^ (-0.6994) = 1.0334085
Inputs and units: Structural dose-response connection. It stays separate from the direct child earnings estimate and all-cause mortality total.
Study population: adults, pooled random-effects dose-response, 33 studies / ~5.9M participants (same source as the T2D row; OR per +1 kg birth weight)
Attributed studies: Knop, Geng, Gorny, Ding, Li, Ley, Huang (2018)

Range: Exponentiated coefficient ± 1.96 standard errors (SE 0.023).

Model conversion · The meta-analysis reports 16.5% lower odds, with a printed rounded odds ratio of 0.84 [0.81, 0.86] per 1 kg. The model converts this using ln(OR) / ln(4.4/3.4). The 3.4 kg baseline and constant elasticity are assumptions; odds are not hazards.

Range: Transformed reported 95% odds-ratio limits, conditional on the model’s assumed baseline.

Family 3

Marriages that endNorway · married workers+11%

What this means: Marriages are 11% more likely to dissolve over the study follow-up. This is a relative increase, not 11 extra divorces per 100 marriages.
Comparison → model: 1× → 1.11×
Exact result: 1.11 × comparison rate
Parameter support: Coefficient support: 1.05 – 1.25
Calculation: Study coefficient = 1.11; support = [1.05, 1.25]
Inputs and units: Married-worker study effect. The selected input does not specify a married cohort or a compatible dissolution baseline.
Study population: married displaced workers
Attributed studies: Rege, Telle, Votruba (2007)Charles, Stephens (2001)

Range: Selected model band across contexts, not a reported confidence interval.

Couples having a baby · layoff yearFinland · woman’s job loss−0.5 percentage points

What this means: The annual probability of a birth falls by 0.5 percentage points after a woman’s job loss. This study effect does not apply to male job loss.
Comparison → model: Absolute baseline not supplied
Exact result: -0.5 percentage-point change
Parameter support: Coefficient support: -0.0089 – -0.0011
Calculation: Study coefficient = -0.005; support = [-0.0089, -0.0011]
Inputs and units: Female-displacement study. A baseline birth probability and an eligible couple cohort are needed for counts.
Study population: couples, Finnish FLEED registers, WOMEN displaced by plant closures (base years 1991-1993) vs matched non-displaced; outcome = couple's probability of a birth in the year
Attributed studies: Huttunen, Kellokumpu (2012)

Source estimate · Table 3’s female plant-closure coefficient is −0.005 (SE 0.002), a 0.5 percentage-point probability change. The Finnish sample is women aged 20–40; it is not a change in total births for every worker.

Range: Coefficient ± 1.96 standard errors.

Adult daughters’ marriage breakdownModel projection · white US daughters1.006418× hazard

What this means: The model projects a small increase in daughters’ marriage-disruption rate through additional parental separations. Hazard means the rate of breakdown among marriages still intact. This is not a lifetime divorce probability. No estimate is available for sons.
Comparison → model: 1× → 1.006418×
Exact result: 1.006418 × comparison hazard
Parameter support: Coefficient support: 1.57 – 2.24
Calculation: s′ = 1 − (1 − 0.069) ^ 1.11; r′ = 1 + (s′ − 0.069) × (1.88 − 1)
Inputs and units: The model combines Norwegian parental dissolution data with a US association for white daughters. Only additional disrupted families transmit the gap. This is a relative hazard, not a divorce probability. A son-specific coefficient is not available.
Study population: US women (NSFG cycles I-III), white daughters who spent part of childhood with a divorced/separated/never-married parent vs two-parent families; proportional-hazard model of own marital disruption
Attributed studies: Rege, Telle, Votruba (2007)Charles, Stephens (2001)McLanahan, Bumpass (1986)

Range: Selected model band across contexts, not a reported confidence interval.

Range: Two observed cohort means; not a confidence interval or annual baseline.

Range: Exponentiated coefficient ± 1.96 standard errors, rounded.

Complete parameter ledger · 74 links

Every link in the pinned model appears below. The citation check distinguishes reported estimates, model conversions, assumptions and unresolved support. Range definitions vary by parameter. A separate exposure or missing baseline prevents an absolute prediction.

displacement → worker earnings0.8

Calculation: Outcome level = 0.8 × comparison level; relative loss = 1 − 0.8
Parameter support: 0.75 – 0.85
Study population: US high-tenure displaced men
Evidence role: conditional
Attributed studies: Jacobson, LaLonde, Sullivan (1992)Oreopoulos, Page, Stevens (2008)Davis, von Wachter (2011)
Citation check: Model assumption · The 0.80 retained-earnings factor is a cross-study selection. Jacobson reports about 25% losses; Davis and von Wachter describe losses of 15–20% after 20 years. These are different cohorts and follow-up periods.
Range: Cross-study model band, not a confidence interval.

earnings shock → mortality sustained1.1354

Calculation: q = 1.1354m / (1 − m + 1.1354m); m = annual baseline probability
Parameter support: 1.0335 – 1.2474
Study population: US high-seniority displaced men born 1930-1959
Evidence role: conditional
Attributed studies: Sullivan, von Wachter (2007)
Citation check: Model conversion · The values exponentiate log-odds coefficients in Table 5, column 3 of the 2007 working paper. They are mortality odds ratios for high-seniority Pennsylvania men after mass layoffs. They do not isolate an earnings-loss mechanism.
Range: Exponentiated coefficient ± 1.96 standard errors; source offsets and cohort apply.

earnings shock → mortality peak2.6698

Calculation: q = 2.6698m / (1 − m + 2.6698m); m = annual baseline probability
Parameter support: 2.1144 – 3.3711
Study population: US high-seniority displaced men born 1930-1959
Evidence role: conditional
Attributed studies: Sullivan, von Wachter (2007)
Citation check: Model conversion · The values exponentiate log-odds coefficients in Table 5, column 3 of the 2007 working paper. They are mortality odds ratios for high-seniority Pennsylvania men after mass layoffs. They do not isolate an earnings-loss mechanism.
Range: Exponentiated coefficient ± 1.96 standard errors; source offsets and cohort apply.

earnings shock → mortality offset 11.9838

Calculation: q = 1.9838m / (1 − m + 1.9838m); m = annual baseline probability
Parameter support: 1.3697 – 2.8732
Study population: US high-seniority displaced men born 1930-1959
Evidence role: conditional
Attributed studies: Sullivan, von Wachter (2007)
Citation check: Model conversion · The values exponentiate log-odds coefficients in Table 5, column 3 of the 2007 working paper. They are mortality odds ratios for high-seniority Pennsylvania men after mass layoffs. They do not isolate an earnings-loss mechanism.
Range: Exponentiated coefficient ± 1.96 standard errors; source offsets and cohort apply.

earnings shock → mortality offsets 2 31.7315

Calculation: q = 1.7315m / (1 − m + 1.7315m); m = annual baseline probability
Parameter support: 1.3238 – 2.2649
Study population: US high-seniority displaced men born 1930-1959
Evidence role: conditional
Attributed studies: Sullivan, von Wachter (2007)
Citation check: Model conversion · The values exponentiate log-odds coefficients in Table 5, column 3 of the 2007 working paper. They are mortality odds ratios for high-seniority Pennsylvania men after mass layoffs. They do not isolate an earnings-loss mechanism.
Range: Exponentiated coefficient ± 1.96 standard errors; source offsets and cohort apply.

earnings shock → mortality offsets 4 51.2712

Calculation: q = 1.2712m / (1 − m + 1.2712m); m = annual baseline probability
Parameter support: 0.97 – 1.6661
Study population: US high-seniority displaced men born 1930-1959
Evidence role: conditional
Attributed studies: Sullivan, von Wachter (2007)
Citation check: Model conversion · The values exponentiate log-odds coefficients in Table 5, column 3 of the 2007 working paper. They are mortality odds ratios for high-seniority Pennsylvania men after mass layoffs. They do not isolate an earnings-loss mechanism.
Range: Exponentiated coefficient ± 1.96 standard errors; source offsets and cohort apply.

displacement → divorce hazard1.11

Calculation: Outcome rate or hazard = 1.11 × its comparison rate or hazard; counts require a matching cohort and baseline
Parameter support: 1.05 – 1.25
Study population: married displaced workers
Evidence role: conditional
Attributed studies: Rege, Telle, Votruba (2007)Charles, Stephens (2001)
Citation check: Model assumption · The Norwegian study reports an 11% higher probability of dissolution by 2003, not an estimated hazard ratio. Charles and Stephens find a divorce response to layoffs, with no significant response to plant closings.
Range: Selected model band across contexts, not a reported confidence interval.

displacement → child earnings0.9076

Calculation: Outcome level = 0.9076 × comparison level; relative loss = 1 − 0.9076
Parameter support: 0.844 – 0.976
Study population: Canada 39k father-son pairs firm closures
Evidence role: conditional
Attributed studies: Oreopoulos, Page, Stevens (2008)
Citation check: Model conversion · The 0.9076 factor exponentiates the −0.097 log-earnings coefficient for sons of displaced Canadian fathers. A lifetime dollar loss also requires the model’s duration and earnings baseline.
Range: Exponentiated coefficient ± 1.96 standard errors (SE 0.037).

divorce → child earningsEstimate withdrawn

Calculation: Estimate withdrawn
Evidence role: boundary
Citation check: Source mismatch · estimate withdrawn · Unilateral-divorce-law ITT does not identify the earnings effect of an individual parental divorce. Its sex/cohort estimates cannot justify a common .95 multiplier.
Range: Model band; the cited source does not establish this parameter.

Original pinned record · excluded from estimates

This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.

{
  "id": "divorce->child_earnings",
  "from": "divorce_hazard",
  "to": "child_earnings",
  "tier": "canonical",
  "evidenceRole": "boundary",
  "citation": "gruber2004",
  "populationScope": "US divorce-law-reform cohorts",
  "point": 0.95,
  "low": 0.9,
  "high": 0.99,
  "notes": "adult children of easier-divorce states: lower earnings + attainment, higher own divorce; PARALLEL stream (never composed with the displacement path silently)"
}

family size → child earningsEstimate withdrawn

Calculation: Estimate withdrawn
Evidence role: boundary
Citation check: Source mismatch · estimate withdrawn · Outcome is education, with family-size effects becoming negligible after birth-order controls and twins IV. No .98 child-earnings multiplier follows; a null schooling finding cannot be converted to zero earnings harm.
Range: Model band; the cited source does not establish this parameter.

Original pinned record · excluded from estimates

This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.

{
  "id": "family_size->child_earnings",
  "from": "family_size",
  "to": "child_earnings",
  "tier": "canonical",
  "evidenceRole": "boundary",
  "citation": "black2005",
  "populationScope": "Norway admin",
  "point": 0.98,
  "low": 0.95,
  "high": 1,
  "notes": "per-child penalty for larger families / later birth order; PARALLEL stream"
}

child earnings → grandchild earnings0.55

Calculation: g′ = 1 − 0.55(1 − g)
Parameter support: 0.4 – 0.6
Study population: US IGE literature
Evidence role: structural
Attributed studies: Solon (1992)Corak (2013)Chetty, Hendren, Kline, Saez (2014)
Citation check: Model assumption · The 0.55 factor is a structural persistence assumption. Solon and Corak provide intergenerational-earnings context; Chetty’s baseline IGE is 0.344, distinct from its rank–rank slope. These associations do not measure transmission of a displacement-induced earnings change.
Range: Selected literature band, not a confidence interval.

grandchild earnings → greatgrandchild earningsEstimate withdrawn

Calculation: Estimate withdrawn
Evidence role: structural
Citation check: Source unverified · estimate withdrawn · The source measures earnings for three generations and education for four. It does not estimate .55 causal gap transmission to a fourth earnings generation. Retaining a projection would require an explicit uncalibrated assumption, not a source-identified default.
Range: Assumed persistence band; source support unresolved.

Original pinned record · excluded from estimates

This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.

{
  "id": "grandchild_earnings->greatgrandchild_earnings",
  "from": "grandchild_earnings",
  "to": "greatgrandchild_earnings",
  "tier": "canonical",
  "evidenceRole": "structural",
  "citation": "lindahl2015;adermon2018",
  "populationScope": "Sweden 3-generation admin",
  "point": 0.55,
  "low": 0.4,
  "high": 0.6,
  "notes": "same IGE band; weakest-identified layer - Lindahl 2015 shows persistence beyond the middle generation survives conditioning; BEHRMAN-TAUBMAN 1990 CROSS-CHECK (v1.14): PSID one-year-measure correlation 0.20 (elasticity ~0.07) vs correlations OVER 0.5 with longer-run income measures - the transitory-income attenuation Solon 1992 corrected; our multi-year IGE band [0.40,0.60] sits inside their long-run range and the one-year number documents why naive single-year IGEs bias low"
}

ipv exposure → daughter violence oddsEstimate withdrawn

Calculation: Estimate withdrawn
Evidence role: boundary
Citation check: Source unverified · estimate withdrawn · Retired identifier; use childhood_parental_violence_exposure->adult_partner_victimization_odds. Published Table 3, violence between parents, net of other risks: OR 2.68, reported 95% CI 1.49-4.82. Sex-adjusted pooled association, not a daughters-only causal estimate, not adult perpetration or injury, and not a household-incidence multiplier. No count conversion or daughter-chain composition.
Range: Unverified model band.

Original pinned record · excluded from estimates

This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.

{
  "id": "ipv_exposure->daughter_violence_odds",
  "from": "household_ipv",
  "to": "daughter_violence_odds",
  "tier": "canonical",
  "evidenceRole": "boundary",
  "citation": "widom1989;ehrensaft2003",
  "populationScope": "US prospective cohorts",
  "point": 2.5,
  "low": 2,
  "high": 3,
  "notes": "cycle-of-violence transmission; upstream household_ipv incidence link queued (aizer2010)"
}

displacement → local service jobs5

Calculation: Service jobs lost = net tradable jobs lost × class multiplier (1.6 manufacturing; 5.0 high-tech). Both inputs are required; this span is not an uncertainty band.
Parameter support: 1.6 – 5
Study population: US metros
Evidence role: conditional
Attributed studies: Moretti (2010)
Citation check: Model assumption · Moretti reports 4.9 local jobs per additional high-tech manufacturing job and 1.59 for manufacturing overall. The model rounds these to 5 and 1.6. Applying additions to losses requires a symmetry assumption and evidence of a net employment change.
Range: Different sector estimates, not a confidence interval.

school spending → child earnings1.07

Calculation: Full study exposure: earnings ratio = 1.07; exposure = +10% per-pupil spending sustained for 12 years.
Parameter support: 1.03 – 1.1
Study population: US school-finance reforms
Evidence role: boundary
Attributed studies: Jackson, Johnson, Persico (2015)
Citation check: Model assumption · The working-paper abstract reports 7.25% higher wages for a 10% spending increase over all 12 school years. The model rounds this to a 1.07 earnings factor.
Range: Assumed 1.03–1.10 band, not the paper’s confidence interval.

youth wages → youth crime-1

Calculation: Δ% crime = -1 × Δ% youth wages
Parameter support: -1.5 – -0.5
Study population: US counties / young men
Evidence role: boundary
Attributed studies: Gould, Weinberg, Mustard (2002)Grogger (1997)
Citation check: Model assumption · The −1 wage elasticity is a rounded model selection. Grogger describes crime-participation elasticities around −0.61 to −0.89; Gould studies local crime and low-skilled male labor-market conditions. Their outcomes and samples differ.
Range: Assumed rounding band, not a reported confidence interval.

wage ratio → household ipvEstimate withdrawn

Calculation: Estimate withdrawn
Evidence role: boundary
Citation check: Source mismatch · estimate withdrawn · Retired identifier; use female_male_wage_ratio_change->female_assault_log_change. Table 2 Panel A col.3: -.813 (SE .317). Change in log female assault hospitalizations per absolute unit change in potential female/male wage ratio. Input .01 means one ratio percentage point; not a relative wage multiplier or log-log elasticity. Hospitalization outcome is not all household IPV incidence.
Range: The coefficient’s SE is 0.317; its interval does not resolve the unit mismatch.

Original pinned record · excluded from estimates

This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.

{
  "id": "wage_ratio->household_ipv",
  "from": "wage_ratio",
  "to": "household_ipv_elasticity",
  "tier": "EXACT-results",
  "evidenceRole": "boundary",
  "citation": "aizer2010",
  "populationScope": "US counties 1990-2003, women's assault hospitalizations",
  "point": -0.813,
  "low": -1.45,
  "high": -0.18,
  "notes": "elasticity of ln(IPV) w.r.t. female/male wage ratio; AER Table 2 col 3 (lagged-DV spec), 95% CI from clustered SE 0.317. SIGN: male displacement RAISES the ratio and so LOWERS IPV against women (Aizer bargaining); wiring this into the daughter chain needs incidence weights for who is displaced - deliberately not composed here"
}

displacement → infant birth weight0.9541

Calculation: Outcome level = 0.9541 × comparison level; relative loss = 1 − 0.9541
Parameter support: 0.912 – 0.9981
Study population: US PSID children 1985+, husband displacement, mother fixed effects
Evidence role: conditional
Attributed studies: Lindo (2010)
Citation check: Model conversion · The factor exponentiates the −0.047 log-birth-weight coefficient in IZA 5213. The sample includes births after displacement, including exposure years before birth; it is not an estimate restricted to in-utero job loss.
Range: Exponentiated coefficient ± 1.96 standard errors (SE 0.023).

import shock → non displaced wage spillover-0.822

Calculation: Δ% wages = -0.822 × (import exposure dollars per worker / 1000)
Parameter support: -1.3042 – -0.3398
Study population: US commuting zones 1990-2007, noncollege workers OUTSIDE manufacturing, 2SLS state-clustered SE
Evidence role: boundary
Attributed studies: Autor, Dorn, Hanson (2012)
Citation check: Source estimate · The −0.822 estimate concerns average log weekly wages of non-college workers outside manufacturing, per $1,000 import exposure per worker. An area average is not a panel of verified continuously employed workers.
Range: Coefficient ± 1.96 standard errors (SE 0.246).

family income shock → child achievement sd0.061

Calculation: Δachievement SD = 0.061 × (income change dollars / 1000)
Parameter support: 0.0157 – 0.1063
Study population: US children (NLSY79 mothers) ages 5-14, EITC instrumental variables
Evidence role: conditional
Attributed studies: Dahl, Lochner (2008)
Citation check: Source estimate · Table 3 of the inspected working paper reports 0.0610 SD per $1,000 of family income in 2000 dollars. That source’s population, income measure and test outcome apply.
Range: Coefficient ± 1.96 standard errors (SE 0.0231).

family income shock → life expectancy years0.1333

Calculation: Δlife years = 0.1333 × (income change dollars / 1000). Local income–longevity association; not a causal death count.
Parameter support: 0.1167 – 0.15
Study population: US adults age 40, household earnings P15-P20 ($14k-$20k 2012 dollars), men+women averaged
Evidence role: boundary
Attributed studies: Chetty, Stepner, Abraham, Lin, Scuderi, Turner, Schrag, Cutler (2016)
Citation check: Model conversion · The model divides the midpoint of an associated 0.7–0.9-year longevity difference by a $6,000 income difference. Chetty’s result is an income–longevity association, not a causal response to an income shock.
Range: Converted example endpoints, not a confidence interval.

displacement → college enrollment0.9894

Calculation: Outcome rate or hazard = 0.9894 × its comparison rate or hazard; counts require a matching cohort and baseline
Parameter support: 0.9848 – 0.9939
Study population: US children of ~7M laid-off fathers, ages 18-22 enrollment, 1999-2009 administrative data
Evidence role: conditional
Attributed studies: Hilger (2016)
Citation check: Source estimate · 1 - 0.432/40.66, with coefficient +/-1.96 SE and the pre-event mean held fixed. Published AEJ: Applied Economics 8(3), 2016, Table 3, printed p.265 / PDF p.19
Range: 1 - 0.432/40.66, with coefficient +/-1.96 SE and the pre-event mean held fixed.

displacement → parental income shortrun0.864

Calculation: Outcome level = 0.864 × comparison level; relative loss = 1 − 0.864
Parameter support: 0.8568 – 0.8712
Study population: US laid-off fathers, annual income in the layoff event window (base $59,832)
Evidence role: conditional
Attributed studies: Hilger (2016)
Citation check: Source estimate · 1 - 8138.09/59832; coefficient +/-1.96 SE; pre-event mean held fixed. Published AEJ: Applied Economics 8(3), 2016, Table 3, printed p.265 / PDF p.19
Range: 1 - 8138.09/59832; coefficient +/-1.96 SE; pre-event mean held fixed.

unemployment rate → property crimeEstimate withdrawn

Calculation: Estimate withdrawn
Evidence role: boundary
Citation check: Source unverified · estimate withdrawn · Retired identifier; use unemployment_rate_change_pp->property_crime_log_change. Table 5 col.6: b=5.018 (SE 1.134) per unit unemployment fraction. Per 1pp change, log crime changes by .05018; percent response is 100*expm1(.05018*dose). The old 5.018 percent reading was a first-order approximation, not an exact percent effect.
Range: Source-table extraction unverified.

Original pinned record · excluded from estimates

This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.

{
  "id": "unemployment_rate->property_crime",
  "from": "unemployment_rate",
  "to": "property_crime_response",
  "tier": "EXACT",
  "evidenceRole": "boundary",
  "citation": "raphael2001",
  "populationScope": "US states 1971-97, N=1,293 state-years, 2SLS defense-contracts + oil-cost instruments",
  "point": 5.018,
  "low": 2.795,
  "high": 7.241,
  "notes": "queue-10 LANDED v1.13: +1pp state unemployment -> +5.018% property crime (SE 1.134), Table 5 \"All property crime\" linear+quadratic-trends 2SLS (95% CI = 1.96SE); overidentification PASSES for property crime in all 2SLS specs. OLS full-spec 1.635 (0.182); the paper summary quotes OLS 1.6-2.4% per pp. Violent crime: null/mixed (2SLS insig 2 of 3 specs, overID rejected in the third) - deliberately NOT landed. Lin 2008 second anchor still queued for the band. Boundary-applied: NEVER chained"
}

eitc exposure → adult earnings early564

Calculation: Δannual adult earnings dollars = 564 × (childhood EITC exposure dollars / 1000)
Parameter support: 83.996 – 1044.004
Study population: US PSID children, EITC exposure ages 13-18, PSID 1968-2013, state-clustered SE
Evidence role: boundary
Attributed studies: Bastian, Michelmore (2018)
Citation check: Source estimate · Coefficient +/-1.96 SE; exposure is maximum potential EITC eligibility, not observed income or actual payments. Publisher 2018 accepted preprint, Table 2 col.6, PDF p.32; 2013 dollars; ages 13-18 exposure, ages 22-27 earnings
Range: Coefficient +/-1.96 SE; exposure is maximum potential EITC eligibility, not observed income or actual payments.

male earnings → marital fertility1.24

Calculation: Δln(marital fertility) = 1.24 × Δln(male earnings)
Parameter support: 0.397 – 2.083
Study population: US PUMAs, fracking boom 1997-2012, women 18-34, IV (shale reserves), PUMA-clustered SE
Evidence role: boundary
Attributed studies: Kearney, Wilson (2017)
Citation check: Source estimate · The 1.24 elasticity is in Table 8, Panel B of the 2017 working paper: marital birth rates for women aged 18–34 during the fracking boom. Nonmarital births have the same elasticity. Reversing a positive-income shock is a model assumption.
Range: Coefficient ± 1.96 standard errors (SE 0.43), rounded.

import shock → gop win probabilityEstimate withdrawn

Calculation: Estimate withdrawn
Evidence role: boundary
Citation check: Source mismatch · estimate withdrawn · Retired identifier; use import_penetration_change_pp->gop_win_probability_change_pp. Table 4 col.6: 24.08 (SE 12.07) Republican win-probability percentage points per 1pp employment-weighted CZ Chinese import penetration change. Not per USD1000/worker and not a national election probability. Other vote-share columns and pre-2010 effects are not replaced by this result.
Range: The published coefficient interval cannot repair the exposure-unit mismatch.

Original pinned record · excluded from estimates

This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.

{
  "id": "import_shock->gop_win_probability",
  "from": "import_shock_per_worker",
  "to": "gop_win_probability",
  "tier": "EXACT",
  "evidenceRole": "boundary",
  "citation": "autor2020polarization",
  "populationScope": "US county-congressional-district cells 2002-2010, N=3,772, 2SLS full controls",
  "point": 24.08,
  "low": 0.423,
  "high": 47.737,
  "notes": "queue-16 LANDED v1.15: change in the probability of electing a Republican to the House, in percentage points, per $1k/worker decadal import exposure = 24.08 (SE 12.07; t=2.00), Table 4 col 6, 2002-2010, N=3,772 county-district cells, 2SLS full controls (spec = Table 3 col 5), 95% CI = 1.96SE. Authors' own scaling: x 0.49 (p75-p25 exposure gap) = 11.8pp. HONEST SCOPE: the two-party VOTE-SHARE columns are NULL (all districts -1.08, SE 5.98; competitive +6.10, SE 4.93; solid-GOP -6.24, SE 3.93 - re-sorting, not uniform right shift); the win-probability effect emerges only from the 2010 mid-term (null 2002-2008, Feigenbaum-Hall consistency). Boundary-applied: NEVER chained"
}

male job loss rate → child maltreatment6

Calculation: Δ% maltreatment reports = 6 × Δmale job-loss percentage points
Parameter support: 3.65 – 8.35
Study population: California counties, child-maltreatment reports, county-year panel, county-clustered SE
Evidence role: boundary
Attributed studies: Lindo, Schaller, Hansen (2013)
Citation check: Model conversion · Table 3 gives a Poisson slope of 0.060 per percentage-point increase in California’s male mass-layoff rate. Six percent is the small-change approximation; the exact change is about 6.18%. The outcome is reports, not measured incidence.
Range: 100 × (coefficient ± 1.96 standard errors), rounded; SE 0.012.

displacement → grade retention hazard1.1473

Calculation: Outcome rate or hazard = 1.1473 × its comparison rate or hazard; counts require a matching cohort and baseline
Parameter support: 1.0119 – 1.2827
Study population: US children 5-19 (SIPP 1996/2001/2004 panels), head-of-household involuntary job loss (fired/discharged, employer sold/bankrupt, slack work), child fixed effects, N=167,449 child-years / 2,170 displaced children / 1,196 job losses
Evidence role: conditional
Attributed studies: Stevens, Schaller (2009)
Citation check: Model conversion · The model divides the 0.0081 grade-repetition probability effect by the 0.055 sample average. This is a relative-probability conversion, not an estimated hazard ratio or a measured test-score effect.
Range: Source coefficient ± 1.96 standard errors, divided by a fixed sample average.

eviction order → emergency shelter use4.778

Calculation: Outcome rate or hazard = 4.778 × its comparison rate or hazard; counts require a matching cohort and baseline
Parameter support: 1.076 – 8.48
Study population: US eviction-court cases (compliers), Cook County + NYC, heads 18-55, judge-leniency IV (first-stage F=934/288)
Evidence role: boundary
Attributed studies: Collinson, Humphries, Mader, Reed, Tannenbaum, van Dijk (2022)
Citation check: Model conversion · The model divides the first-year IV shelter-use effect of 0.034 by the rounded non-evicted mean of 0.009. It is a converted first-year ratio, not a reported universal homelessness effect.
Range: Source coefficient ± 1.96 standard errors (SE 0.017), scaled by a fixed rounded mean.

eviction order → eviction earnings responseEstimate withdrawn

Calculation: Estimate withdrawn
Evidence role: boundary
Citation check: Source mismatch · estimate withdrawn · Retired identifier; use eviction_order->quarterly_earnings_gap_q5_8. Table VI col.6 IV earnings difference -613 (SE 248), divided by matching col.4 non-evicted sample mean 4254, not year-1 mean 4300. This scaled index holds the descriptive comparison mean fixed; it is not a separately identified complier-counterfactual earnings ratio. Never composed into displacement earnings.
Range: The source SE is 248; the model’s denominator is from the wrong period.

Original pinned record · excluded from estimates

This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.

{
  "id": "eviction_order->eviction_earnings_response",
  "from": "eviction_order",
  "to": "eviction_earnings_response",
  "tier": "EXACT",
  "evidenceRole": "boundary",
  "citation": "collinson2024",
  "populationScope": "same design as the shelter row",
  "point": 0.8574,
  "low": 0.7444,
  "high": 0.9705,
  "notes": "queue-17 LANDED v1.21: evicted tenants' average QUARTERLY earnings, quarters 5-8 post-filing: IV -$613 (SE 248) on the $4,300 non-evicted mean = -14.3%; 95% CI = 1.96SE; stored as a gap multiplier: 1 - 613/4300 = 0.8574 (band 1 - CI/4300), 2016 USD in the derivation above. Year-1 window (quarters 1-4) is -$323 (SE 175), NOT significant - recorded, not landed (the stevens2011 timing pattern: the sustained window carries the identified effect). OLS -$269*** is SMALLER than IV - the authors read complier effects as larger than the average evicted tenant's. SUBGROUPS (Table VI): concentration among female (-$767, SE 295) and Black (-$931, SE 307) tenants in year 2; formal equality tests do not reject, so the full-sample point stands with the concentration noted. REFUSAL WIRING: the desmond2016 matched +11-22pp JOB-LOSS estimate is NOT encoded (CITING §1 class + collinson2024 refutes the size: employment IV -1.5pp (SE 2.1) year 1 / -1.8pp (SE 2.7) year 2, both null); see SPEC §10. PARALLEL STREAM: quarterly-earnings LEVEL response for the eviction-court population - NEVER composed into the worker_earnings displacement multiplier. Boundary-applied: NEVER chained"
}

foreclosure order → house price gap0.99

Calculation: Nearby house-price ratio = 0.99; loss = 1 − 0.99. Distance and property cohort are required.
Parameter support: 0.9875 – 0.9925
Study population: US all-type house transactions, Massachusetts 1987-2008 (1.78M sales; Warren Group), reference distance 0.05 miles
Evidence role: boundary
Attributed studies: Campbell, Giglio, Pathak (2009)
Citation check: Model assumption · The source describes an approximately 1% price reduction for neighboring sales within 0.05 mile of a foreclosure. The model selects 0.99 and adds a ±25% band on that effect.
Range: Assumed band, not a reported confidence interval.

household ipv → child internalizing sd0.48

Calculation: Outcome = baseline + (0.48); change is in standard-deviation units
Parameter support: 0.39 – 0.57
Study population: children/adolescents 18 or younger exposed to interparental physical violence (second-generation studies, 1990-2006); k=58 samples, N=7,602
Evidence role: boundary
Attributed studies: Evans, Davies, DiLillo (2008)
Citation check: Source estimate · Reported CI, not recomputed from rounded SE. Cross-study exposure association, not a causal intervention response. Published Aggression and Violent Behavior 13, 2008, printed p.135 / PDF p.5, Table 1 and sections 3.2.1-3.2.2
Range: Reported CI, not recomputed from rounded SE. Cross-study exposure association, not a causal intervention response.

household ipv → child externalizing sd0.47

Calculation: Outcome = baseline + (0.47); change is in standard-deviation units
Parameter support: 0.38 – 0.56
Study population: children/adolescents 18 or younger exposed to interparental physical violence (second-generation studies, 1990-2006); k=53 samples, N=7,200; pooled across gender - boys .46 / girls .23
Evidence role: boundary
Attributed studies: Evans, Davies, DiLillo (2008)
Citation check: Source estimate · Reported CI; exposure association with significant heterogeneity. Not a causal intervention response. Published Aggression and Violent Behavior 13, 2008, printed p.135 / PDF p.5, Table 1 and sections 3.2.1-3.2.2
Range: Reported CI; exposure association with significant heterogeneity. Not a causal intervention response.

local unemp shock → household ipv1.58

Calculation: Odds ratio = 1.58 ^ u; u = 12-month proportional unemployment increase (1 = doubling); q = odds ratio × q₀ / (1 − q₀ + odds ratio × q₀).
Parameter support: 1.49 – 1.58
Study population: mothers of the Fragile Families birth cohort (1998-2000 births in 20 large US cities), Waves 3-5 pooled person-waves (2001-2010), all-mothers sample including unpartnered; N=10,656 area models / 10,584 distress models; UR shock 1.0=flat, 2.0=doubling
Evidence role: boundary
Attributed studies: Schneider, Harknett, McLanahan (2016)
Citation check: Model assumption · The 1.58 point exponentiates the lagged-outcome combined violence/controlling-behavior coefficient, 0.454. Table 3’s logit and fixed-effects coefficients are 0.401 and 0.453. Physical violence alone is not significant.
Range: Range across specifications, not a confidence interval.

household hardship → household ipv2.14

Calculation: Prevalence ratio = 2.14; q = 2.14 × q₀. Source contrast: 15% versus 7%.
Parameter support: 1.61 – 2.68
Study population: mothers of the Fragile Families birth cohort (1998-2000 births in 20 large US cities), Waves 3-5 pooled person-waves (2001-2010), all-mothers sample including unpartnered; N=10,656 area models / 10,584 distress models; hardship = 1+ of 10 material-hardship items, past 12 months
Evidence role: boundary
Attributed studies: Schneider, Harknett, McLanahan (2016)
Citation check: Model conversion · The 2.14 ratio divides the paper’s predicted probabilities, 15% and 7%. It is not the reported logit odds ratio or an established causal hardship effect. The outcome includes controlling behavior.
Range: Assumed ±25% rounding band, not a confidence interval.

couple unemployment → household ipv1.3

Calculation: Prevalence ratio = 1.3; q = 1.3 × q₀. Source contrast: 13% versus 10%.
Parameter support: 1.04 – 1.63
Study population: mothers of the Fragile Families birth cohort (1998-2000 births in 20 large US cities), Waves 3-5 pooled person-waves (2001-2010), all-mothers sample including unpartnered; N=10,656 area models / 10,584 distress models; either partner unemployed and looking
Evidence role: boundary
Attributed studies: Schneider, Harknett, McLanahan (2016)
Citation check: Model conversion · The 1.30 ratio divides predicted probabilities, 13% and 10%. Exposure includes some noncoresiding unemployed mothers. The combined-outcome logit coefficient is 0.306; the bundle note’s 0.404 belongs to another specification.
Range: Assumed rounding band, not a confidence interval.

unemployment status → mental health sd0.51

Calculation: Outcome = baseline + (0.51); change is in standard-deviation units
Parameter support: 0.47 – 0.54
Study population: involuntary unemployed vs employed adults (26 countries, 1950-2004); cross-sectional meta k=315 samples, N=209,379, outliers excluded
Evidence role: boundary
Attributed studies: Paul, Moser (2009)
Citation check: Source estimate · Overall outlier-excluded cross-sectional association, not the factory-closure subset or a causal job-loss effect. Published JVB 74, 2009, Table 1 Overall (out. ex), printed p.271 / PDF p.8
Range: Overall outlier-excluded cross-sectional association, not the factory-closure subset or a causal job-loss effect.

import shock → radical right vote shareEstimate withdrawn

Calculation: Estimate withdrawn
Evidence role: boundary
Citation check: Source mismatch · estimate withdrawn · Retired identifier; use import_exposure_eur1000_per_worker_2y->radical_right_vote_share_change_pp. Accepted manuscript Table 1 col.10: .132 (SE .051) vote-share fraction per EUR1000/worker import increase over the two years before an election. Multiply output by 100 to obtain percentage points. No USD conversion, decadal-dose substitution or persistence assumption.
Range: Rounded coefficient ± 1.96 standard errors; the exposure-unit mismatch remains.

Original pinned record · excluded from estimates

This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.

{
  "id": "import_shock->radical_right_vote_share",
  "from": "import_shock_per_worker",
  "to": "radical_right_vote_share",
  "tier": "EXACT",
  "evidenceRole": "boundary",
  "citation": "colantone2018",
  "populationScope": "15 Western European countries, 76 legislative elections 1988-2007, 198 NUTS-2 regions; district-election panel N=7,782 (IV), country-year FE, SE clustered NUTS-2 x year",
  "point": 13.2,
  "low": 3.2,
  "high": 23.2,
  "notes": "queue-15 LANDED v1.26: IV Table 1 col 10, radical-right vote share. Coefficient .132*** (SE .051) in vote-share FRACTIONS per 1.0 ($1k/worker change in Chinese import exposure over the past 2 years) = 13.2 pp per $1k/worker of exposure (SE 5.1), CI = 1.96 SE. AUTHORS' OWN SCALING: one SD of the 2-year shock (.133 = $133/worker) -> +1.7pp radical-right vote share, on an average RR share of 5% (SD 7%). IV (.132) > OLS (.041, SE .023): positive domestic demand shocks attenuate the OLS association (authors' interpretation). First stage: Chinese imports to the US as instrument, Kleibergen-Paap F = 19.2. Robustness (Table 3-4): historical-trajectory trend interactions (coefficients equal or larger), year-specific regional controls (larger), alternative exchange-rate IV, exports and FDI nulls, EU-import controls (EU15 negative on Nationalist Autarchy, EU12 positive on the radical right). NO-LEFT-WING discipline: protectionist left NULL (IV -.052, SE .047), protectionist left proper NULL (-.081, SE .042), liberal right NULL (-.017, SE .075) - NO polarization, which CONTRASTS with the Autor et al. 2020 US finding (declared on that row); pro-trade left NEGATIVE significant (-.134**, SE .054) - the electorate abandons mainstream social-democratic parties; protectionist right POSITIVE .278*** (SE .094). INDIVIDUAL-LEVEL (ESS waves 1-4, Table 5-6): confirms the district evidence; NO significant group heterogeneity - interactions all n.s., the response is SOCIOTROPIC (service and public-sector workers respond too; only students sheltered, n.s. overall) - declared so no surface reads this row as a displaced-manufacturers-only effect, and so no surface quotes paul2009's person-level unemployment distress as the transmission mechanism here. WINDOW DISCIPLINE: the regression's shock is the 2-year window change; the row stores the same coefficient expressed per $1k/worker of cumulative exposure (13.2pp), which assumes each window's shock persists to the next election - the authors' 1-SD scaling reads directly off this coefficient. Boundary-applied: NEVER chained into the earnings/IGE chains. Read via the owner-downloaded PDF (the published AJPS doi, not the SSRN preprint doi the staging bib carried)"
}

neighborhood exposure → child outcomes modifier0.037

Calculation: Annual exposure coefficient = 0.037; moving age and origin/destination income-rank difference are required.
Parameter support: 0.031 – 0.043
Study population: children of families moving across counties (US tax data, births 1980-88); county-level predictions of permanent residents' income ranks at age 24
Evidence role: structural
Attributed studies: Chetty, Hendren (2016)
Citation check: Source estimate · Appendix Table V reports 0.037 income-rank points per additional childhood year in a county with one rank-point better permanent-resident outcomes. Multiplying a displacement penalty by this place factor is a model assumption.
Range: Coefficient ± 1.96 standard errors (SE 0.003), rounded.

unconditional income → child education years1.127

Calculation: Δeducation years = 1.127 for the source's approximately $4,000 annual transfer per eligible adult.
Parameter support: 0.247 – 2.007
Study population: children in Eastern Band of Cherokee area households previously in poverty (Great Smoky Mountains Study, age cohorts 9/11 at wave 1), treated by ~$4,000/yr per adult member for four pre/post childhood survey waves
Evidence role: boundary
Attributed studies: Akee, Copeland, Keeler, Angold, Costello (2010)
Citation check: Source estimate · Table 4’s previously poor-household coefficient is 1.127 education years (SE 0.449), N = 438. It is the youngest-cohort interaction per eligible tribal-member parent, not per arbitrary household or $1,000.
Range: Coefficient ± 1.96 standard errors.

unconditional income → youth any crime-0.224

Calculation: Δarrest probability = -0.224 for the source transfer exposure; eligible child cohort required.
Parameter support: -0.377 – -0.071
Study population: same population as the education row (previously-poor eligible households; arrests from administrative records merged at the individual level)
Evidence role: boundary
Attributed studies: Akee, Copeland, Keeler, Angold, Costello (2010)
Citation check: Source estimate · Table 6 reports −0.224 probability points for any arrest at ages 16–17. Treatment is the youngest-cohort interaction per eligible tribal-member parent receiving casino transfers.
Range: Coefficient ± 1.96 standard errors (SE 0.078).

unconditional income → youth minor crime ever-0.179

Calculation: Δminor-offense probability = -0.179 for the source transfer exposure; eligible child cohort required.
Parameter support: -0.353 – -0.005
Study population: same population as the education row (previously-poor eligible households)
Evidence role: boundary
Attributed studies: Akee, Copeland, Keeler, Angold, Costello (2010)
Citation check: Source estimate · Table 6 reports −0.179 probability points for ever a minor arrest by age 21. It is a specified cohort/eligible-parent contrast, not a generic income elasticity.
Range: Coefficient ± 1.96 standard errors (SE 0.089).

youth crime conviction share → youth crime convicted0.061

Calculation: Δconviction probability = 0.061 × Δarea youth conviction-share percentage points
Parameter support: 0 – 0.123
Study population: male refugee children in Denmark assigned quasi-randomly to municipalities 1986-1998 (avg age at assignment 9; N=4,425)
Evidence role: boundary
Attributed studies: Damm, Dustmann (2013)
Citation check: Model conversion · The model divides a 0.043 probability effect by the neighborhood conviction-share SD of 0.70 percentage points. This concerns Danish refugee boys’ later convictions, not the volume of reported neighborhood crimes.
Range: Coefficient ± 1.96 standard errors (SE 0.022), scaled by 0.70; lower endpoint rounded to zero.

youth violent crime conviction share → youth crime convicted0.366

Calculation: Δconviction probability = 0.366 × Δarea violent-conviction-share percentage points
Parameter support: 0.143 – 0.589
Study population: same population as the overall-share row (male refugee children, quasi-random municipality assignment, N=4,425)
Evidence role: boundary
Attributed studies: Damm, Dustmann (2013)
Citation check: Model conversion · The model divides a 0.045 probability effect by the violent conviction-share SD of 0.123 percentage points. The source’s male refugee population and conviction-share exposure apply; girls’ estimates are not significant.
Range: Coefficient ± 1.96 standard errors (SE 0.014), scaled by 0.123.

displacement → circulatory mortality hazard1.54

Calculation: Outcome rate or hazard = 1.54 × its comparison rate or hazard; counts require a matching cohort and baseline
Parameter support: 1.29 – 1.83
Study population: Danish full-time male workers age 20-60 with strong labor-market attachment displaced by plant closures 1986-2002 (base years 1985-2001; propensity-score weighted matched controls; Danish full registers 1980-2006)
Evidence role: conditional
Attributed studies: Browning, Heinesen (2011)
Citation check: Source estimate · The inspected 2011 report gives a cumulative hazard ratio over years 1–4 for Danish men with strong labor-market attachment after plant closure. It is not an annual hazard or an odds ratio.
Range: Reported 95% confidence interval for the specified cause and follow-up window.

displacement → alcohol mortality hazard1.62

Calculation: Outcome rate or hazard = 1.62 × its comparison rate or hazard; counts require a matching cohort and baseline
Parameter support: 1.09 – 2.41
Study population: Danish full-time male workers age 20-60 with strong labor-market attachment displaced by plant closures 1986-2002 (base years 1985-2001; propensity-score weighted matched controls; Danish full registers 1980-2006)
Evidence role: conditional
Attributed studies: Browning, Heinesen (2011)
Citation check: Source estimate · The inspected 2011 report gives a cumulative hazard ratio over years 1–4 for Danish men with strong labor-market attachment after plant closure. It is not an annual hazard or an odds ratio.
Range: Reported 95% confidence interval for the specified cause and follow-up window.

displacement → external cause mortality hazard1.53

Calculation: Outcome rate or hazard = 1.53 × its comparison rate or hazard; counts require a matching cohort and baseline
Parameter support: 1.15 – 2.04
Study population: Danish full-time male workers age 20-60 with strong labor-market attachment displaced by plant closures 1986-2002 (base years 1985-2001; propensity-score weighted matched controls; Danish full registers 1980-2006)
Evidence role: conditional
Attributed studies: Browning, Heinesen (2011)
Citation check: Source estimate · The inspected 2011 report gives a cumulative hazard ratio over years 1–4 for Danish men with strong labor-market attachment after plant closure. It is not an annual hazard or an odds ratio.
Range: Reported 95% confidence interval for the specified cause and follow-up window.

displacement event → spouse mental health sd-0.194

Calculation: Outcome = baseline + (-0.194); change is in standard-deviation units
Parameter support: -0.327 – -0.061
Study population: spouses of displaced workers, German couples (SOEP 2002-2010), 109 displaced-worker couples vs >14,000 entropy-balanced controls; MCS normed mean 50 SD 10
Evidence role: conditional
Attributed studies: Marcus (2012)
Citation check: Model conversion · Table 3’s spouse mental-health effect is −1.94 MCS points (SE 0.68). The source normalizes the score to SD 10; division by 10 gives −0.194 SD. The follow-up is about one year after unemployment following closure.
Range: Coefficient ± 1.96 standard errors, divided by 10.

displacement event → child achievement sd-0.0206

Calculation: Outcome = baseline + (-0.0206); change is in standard-deviation units
Parameter support: -0.0402 – -0.001
Study population: children exposed to parental plant closure age 0-16 (average across exposure stages), Denmark, conditional on taking grade-9 exams; mathematics
Evidence role: conditional
Attributed studies: Bingley, Cappellari, Ovidi (2023)
Citation check: Source estimate · Table 4 reports −0.0206 SD (SE 0.0100) for teacher-assessed mathematics grades in the Danish sample. The separate standardized-test estimate is smaller and not significant. These outcomes are not interchangeable.
Range: Coefficient ± 1.96 standard errors.

displacement event → exam noncompletion hazardEstimate withdrawn

Calculation: Estimate withdrawn
Evidence role: conditional
Citation check: Source mismatch · estimate withdrawn · Retired identifier; use displacement_event->exam_noncompletion_change_pp. IZA DP 16367 Table 2 Panel A col.2: probability difference .0054 (SE .0027), expressed as .54 percentage points. Regression constant .0785 is not an admitted control mean; no hazard ratio, relative-risk conversion or displaced-worker total. Published-version reconciliation remains open.
Range: Coefficient uncertainty scaled by an unsupported denominator; not an estimated hazard ratio.

Original pinned record · excluded from estimates

This input is excluded from public estimates because its source, units or numerical extraction could not be substantiated. The original pinned record is retained below.

{
  "id": "displacement_event->exam_noncompletion_hazard",
  "from": "displacement_event",
  "to": "exam_noncompletion_hazard",
  "tier": "EXACT",
  "evidenceRole": "conditional",
  "citation": "bingley2026",
  "populationScope": "same children/scope as the child_achievement_sd row; outcome = obtains NO grade-9 exam scores or teacher assessments (control mean 7.85%)",
  "point": 1.0688,
  "low": 1.0014,
  "high": 1.1362,
  "notes": "queue-sweep LANDED v1.40, same read as the child_achievement_sd row. DERIVED RELATIVE (pp effect / control mean; stevens2011 derivation precedent): Table 2 col 2 (preferred) 'Treated X Exposure (age 0-16)' +0.0054 (SE 0.0027) on control mean 0.0785 -> 1 + 0.0054/0.0785 = 1.0688; band = 1.96 x SE/0.0785. AUTHORS' FRAME: explains 6-9% of observed non-completion (base ~7%; test-taking near-universal at 16). TIMING STRUCTURE, monotone by age at exposure: infancy (0-1) +0.99pp (SE 0.41) -> 1.126 [1.96SE 1.024..1.228]; early childhood (2-5) +0.76pp -> ~1.097; mid (6-11) +0.45pp n.s.; late (12-16) +0.30pp n.s.; exposure at age >=18 jointly zero (p=0.986 - identifying-assumption placebo). Completing exams BY AGE 16: infancy -1.44pp (SE 0.65). Concentrated in below-median-income families (no detectable effect above median); displaced father > mother, sons > daughters (qualitative). Conditional: Danish compulsory-school institution; left-tail outcome capturing the most vulnerable children"
}

displacement event → hs completion-0.037

Calculation: Outcome = baseline + (-0.037); change is in probability points
Parameter support: -0.0801 – 0.0061
Study population: children of single mothers, US NLSY79 (NLSY79 women's children), MATERNAL job displacement at child age 0-17 (PSM kernel TT); outcome = HS completion by age 19, control mean 0.760
Evidence role: conditional
Attributed studies: Brand, Simon Thomas (2014)
Citation check: Source estimate · Table 3’s kernel-matching estimate is −0.037 for high-school completion by age 19; its interval crosses zero. The sample is children of single mothers; selection on observed characteristics is not random assignment. The displayed baseline is the unadjusted comparison-group mean in Table 1; applying the matched effect to it is a model choice.
Range: Coefficient ± 1.96 standard errors (SE 0.022).

displacement event → college attendance-0.063

Calculation: Outcome = baseline + (-0.063); change is in probability points
Parameter support: -0.1081 – -0.0179
Study population: same scope as the hs_completion row; outcome = some college attendance by age 21-22, control mean 0.427
Evidence role: conditional
Attributed studies: Brand, Simon Thomas (2014)
Citation check: Source estimate · Table 3’s kernel-matching estimate is −0.063 for college attendance in young adulthood. The sample is children of single mothers; selection on observed characteristics is not random assignment. The displayed baseline is the unadjusted comparison-group mean in Table 1; applying the matched effect to it is a model choice.
Range: Coefficient ± 1.96 standard errors (SE 0.023).

displacement event → college completion-0.036

Calculation: Outcome = baseline + (-0.036); change is in probability points
Parameter support: -0.0674 – -0.0046
Study population: same scope as the hs_completion row; outcome = college completion (BA) by age 25, control mean 0.169
Evidence role: conditional
Attributed studies: Brand, Simon Thomas (2014)
Citation check: Source estimate · Table 3’s kernel-matching estimate is −0.036 for college completion in young adulthood. The sample is children of single mothers; selection on observed characteristics is not random assignment. The displayed baseline is the unadjusted comparison-group mean in Table 1; applying the matched effect to it is a model choice.
Range: Coefficient ± 1.96 standard errors (SE 0.016).

displacement event → adult depression cesd0.025

Calculation: Outcome = baseline + (0.025); change is in 0–1 index units
Parameter support: 0.0034 – 0.0466
Study population: same scope as the hs_completion row; outcome = CESD 7-item depressive-symptom index (0-1) at ages 25-29, control mean 0.208
Evidence role: conditional
Attributed studies: Brand, Simon Thomas (2014)
Citation check: Source estimate · Table 3’s kernel-matching estimate is +0.025 on the 0–1 CESD index at ages 25–29, not a standard-deviation change. The sample is children of single mothers; selection on observed characteristics is not random assignment. The displayed baseline is the unadjusted comparison-group mean in Table 1; applying the matched effect to it is a model choice.
Range: Coefficient ± 1.96 standard errors (SE 0.011).

displacement event → spouse participation elasticity-0.04

Calculation: Δln(wife labor-force participation) = -0.04 × Δln(husband earnings)
Parameter support: -0.07 – -0.03
Study population: married couples, Austrian ASOS registers 1990-2007, HUSBANDS displaced by firm closure/mass layoff vs non-event firm controls; wife's labor-force participation response
Evidence role: conditional
Attributed studies: Halla, Schmieder, Weber (2020)
Citation check: Model assumption · The inspected AEA manuscript’s Table 2 reports employment elasticity −0.043 (SE 0.011) for wives after husbands’ displacement. The model rounds to −0.04. This is an elasticity, not a percentage-point participation coefficient.
Range: Rounded subgroup band, not the main estimate’s confidence interval.

displacement event → annual birth response-0.005

Calculation: Outcome = baseline + (-0.005); change is in probability points
Parameter support: -0.0089 – -0.0011
Study population: couples, Finnish FLEED registers, WOMEN displaced by plant closures (base years 1991-1993) vs matched non-displaced; outcome = couple's probability of a birth in the year
Evidence role: conditional
Attributed studies: Huttunen, Kellokumpu (2012)
Citation check: Source estimate · Table 3’s female plant-closure coefficient is −0.005 (SE 0.002), a 0.5 percentage-point probability change. The Finnish sample is women aged 20–40; it is not a change in total births for every worker.
Range: Coefficient ± 1.96 standard errors.

displacement event → child depression anxiety0.008

Calculation: Outcome = baseline + (0.008); change is in probability points
Parameter support: 0.0002 – 0.0158
Study population: children ages 6-17, US MEPS, PATERNAL job loss (tenured-father sample, >=1y tenure), individual-FE LPM; outcome = child depression/anxiety report
Evidence role: conditional
Attributed studies: Schaller, Zerpa (2015)
Citation check: Source estimate · Table 2’s probability effect is +0.008: 0.8 percentage points, not 0.008 percentage points. The inspected MEPS sample is children aged six and older with a tenured displaced father.
Range: Coefficient ± 1.96 standard errors (SE 0.004).

child achievement sd → grandchild achievement sd0.38

Calculation: Δgrandchild SD = 0.38 × Δchild SD
Parameter support: 0.38 – 0.42
Study population: Norway male conscripts (IQ at age 18, father-son)
Evidence role: structural
Attributed studies: Black, Devereux, Salvanes (2008)
Citation check: Model assumption · The source reports a 0.38 father–son IQ correlation in Norway. The authors give it no causal interpretation. Applying it to an induced achievement change is a structural assumption, not a measured intergenerational treatment effect.
Range: Selected literature band, not a confidence interval.

child education years → grandchild education years0.296

Calculation: Δgrandchild education years = 0.296 × Δchild education years
Parameter support: 0.255 – 0.337
Study population: Sweden, Malmö four-generation register (G3 children, N=1,823)
Evidence role: structural
Attributed studies: Lindahl, Palme, Massih, Sj{\"o}gren (2012)
Citation check: Source estimate · Table 2 reports a 0.296 association between parent and child years of education. It is an observational regression coefficient, not the causal transmission of an induced schooling change.
Range: Coefficient ± 1.96 standard errors (SE 0.021), rounded.

infant birth weight → child earnings0.1

Calculation: y / y₀ = (x / x₀) ^ (0.1); x / x₀ is the named exposure ratio
Parameter support: 0.0216 – 0.1784
Study population: Norway twin sample, twin fixed effects on ln(birth weight), cohorts 1967-1988 (military conscription + earnings registries); IZA DP 1864 Table 5 'Ln(Earnings) FT', all-twins FE column
Evidence role: conditional
Attributed studies: Black, Devereux, Salvanes (2005)
Citation check: Source estimate · The 2005 working paper’s Table 5 reports 0.10 (SE 0.04) for log full-time earnings on log birth weight in a twin fixed-effects sample. This is not a causal bridge from parental displacement to adult earnings.
Range: Coefficient ± 1.96 standard errors.

infant birth weight → adult type2 diabetes hazard-0.9637

Calculation: y / y₀ = (x / x₀) ^ (-0.9637); x / x₀ is the named exposure ratio
Parameter support: -1.3834 – -0.5401
Study population: adults, pooled random-effects dose-response, 32 studies / ~7.6M participants (Knop et al. JAHA 2018, 7(23):e008870, OR per +1 kg birth weight; read via PMC6405546 full text)
Evidence role: conditional
Attributed studies: Knop, Geng, Gorny, Ding, Li, Ley, Huang (2018)
Citation check: Model conversion · The meta-analysis reports odds ratio 0.78 [0.70, 0.87] per 1 kg. The model converts this using ln(OR) / ln(4.4/3.4). The 3.4 kg baseline and constant elasticity are assumptions; odds are not hazards.
Range: Transformed reported 95% odds-ratio limits, conditional on the model’s assumed baseline.

infant birth weight → adult cardiovascular disease hazard-0.6994

Calculation: y / y₀ = (x / x₀) ^ (-0.6994); x / x₀ is the named exposure ratio
Parameter support: -0.8173 – -0.585
Study population: adults, pooled random-effects dose-response, 33 studies / ~5.9M participants (same source as the T2D row; OR per +1 kg birth weight)
Evidence role: conditional
Attributed studies: Knop, Geng, Gorny, Ding, Li, Ley, Huang (2018)
Citation check: Model conversion · The meta-analysis reports 16.5% lower odds, with a printed rounded odds ratio of 0.84 [0.81, 0.86] per 1 kg. The model converts this using ln(OR) / ln(4.4/3.4). The 3.4 kg baseline and constant elasticity are assumptions; odds are not hazards.
Range: Transformed reported 95% odds-ratio limits, conditional on the model’s assumed baseline.

divorce hazard → child divorce hazard1.88

Calculation: s′ = 1 − (1 − s)^r; child hazard ratio = 1 + (s′ − s) × (1.88 − 1)
Parameter support: 1.57 – 2.24
Study population: US women (NSFG cycles I-III), white daughters who spent part of childhood with a divorced/separated/never-married parent vs two-parent families; proportional-hazard model of own marital disruption
Evidence role: structural
Attributed studies: McLanahan, Bumpass (1986)
Citation check: Model conversion · The inspected 1986 draft’s Appendix Table A1 gives 0.63 (SE 0.09) for white women in a background-adjusted proportional-hazard model. Parent absence includes divorce, separation, never-married parents and neither parent, not divorce alone.
Range: Exponentiated coefficient ± 1.96 standard errors, rounded.

married cohort → parental dissolution share0.069

Calculation: Counterfactual dissolution share s = 0.069; use only for the matched married cohort.
Parameter support: 0.069 – 0.072
Study population: married workers at Norwegian plants 1995-2000 (the rege2007 entry population and follow-up window, dissolution measured by 2003)
Evidence role: conditional
Attributed studies: Rege, Telle, Votruba (2007)
Citation check: Model assumption · The 0.069 value is the observed dissolution share among Norwegian couples with husbands at stable plants, from 1995 to 2003. The 0.072 endpoint is the closing-plant group’s observed mean, not an uncertainty limit.
Range: Two observed cohort means; not a confidence interval or annual baseline.

undocumented share change pp → receiving violent crime log rate-0.12

Calculation: log(rate1/rate0) = -0.12 × share change in percentage points; rate ratio = exp(log-rate change). Requires the undocumented-resident share margin.
Parameter support: -0.18 – -0.06
Study population: US 50 states + DC 1990-2014; lagged percentage-point undocumented population share; Light-Miller 2018 Table 2 Model 4
Evidence role: conditional
Attributed studies: Light, Miller (2018)
Citation check: Source estimate · Table 2’s unweighted fixed-effects Model 5 gives −0.12 (SE 0.03) for log violent-crime rates per percentage-point undocumented population share. It is an observational association; the weighted model differs.
Range: Coefficient ± 1.96 standard errors, rounded to two decimals.

regular immigrant share ratio → receiving total crime rate0.105

Calculation: rate1/rate0 = (share1/share0)^0.105. Shares require both population denominators; this is a ten-year endpoint contrast.
Parameter support: -0.2616 – 0.4716
Study population: Italy 95 provinces 1991-2001; log valid residence permits / total province population; 2010 working-paper ten-year-difference IV
Evidence role: conditional
Attributed studies: Bianchi, Buonanno, Pinotti (2010)
Citation check: Source estimate · Table 5 of the 2010 manuscript gives IV 0.105 (SE 0.187) for log total-crime rate versus log valid residence permits / population. The interval includes zero. This is a ten-year endpoint-rate contrast in Italian provinces.
Range: Coefficient ± 1.96 standard errors, rounded; not a cumulative incident count.

unemployment rate change pp → property crime log change0.05018

Calculation: Table 5 col.6: b=5.018 (SE 1.134) per unit unemployment fraction. Per 1pp change, log crime changes by .05018; percent response is 100*expm1(.05018*dose). The old 5.018 percent reading was a first-order approximation, not an exact percent effect.
Parameter support: 0.0279536 – 0.0724064
Study population: US states 1971-97, 1293 state-years
Evidence role: boundary
Attributed studies: Raphael, Winter-Ebmer (2001)
Citation check: Source estimate · Table 5 col.6: b=5.018 (SE 1.134) per unit unemployment fraction. Per 1pp change, log crime changes by .05018; percent response is 100*expm1(.05018*dose). The old 5.018 percent reading was a first-order approximation, not an exact percent effect. Published JLE 44(1), 2001, Table 5 col.6, printed p.275 / PDF p.18; unemployment is a fraction
Range: Table 5 col.6: b=5.018 (SE 1.134) per unit unemployment fraction. Per 1pp change, log crime changes by .05018; percent response is 100*expm1(.05018*dose). The old 5.018 percent reading was a first-order approximation, not an exact percent effect.

female male wage ratio change → female assault log change-0.813

Calculation: Table 2 Panel A col.3: -.813 (SE .317). Change in log female assault hospitalizations per absolute unit change in potential female/male wage ratio. Input .01 means one ratio percentage point; not a relative wage multiplier or log-log elasticity. Hospitalization outcome is not all household IPV incidence.
Parameter support: -1.43432 – -0.19168
Study population: California county/race/year cells 1990-2003; female assault hospital admissions
Evidence role: boundary
Attributed studies: Aizer (2010)
Citation check: Source estimate · Table 2 Panel A col.3: -.813 (SE .317). Change in log female assault hospitalizations per absolute unit change in potential female/male wage ratio. Input .01 means one ratio percentage point; not a relative wage multiplier or log-log elasticity. Hospitalization outcome is not all household IPV incidence. Author manuscript in PMC4123456, Table 2 Panel A col.3; machine-readable primary full-text XML
Range: Table 2 Panel A col.3: -.813 (SE .317). Change in log female assault hospitalizations per absolute unit change in potential female/male wage ratio. Input .01 means one ratio percentage point; not a relative wage multiplier or log-log elasticity. Hospitalization outcome is not all household IPV incidence.

import penetration change pp → gop win probability change pp24.08

Calculation: Table 4 col.6: 24.08 (SE 12.07) Republican win-probability percentage points per 1pp employment-weighted CZ Chinese import penetration change. Not per USD1000/worker and not a national election probability. Other vote-share columns and pre-2010 effects are not replaced by this result.
Parameter support: 0.4228 – 47.7372
Study population: US county-congressional-district cells 2002-2010, 3772 observations
Evidence role: boundary
Attributed studies: Autor, Dorn, Hanson, Majlesi (2020)
Citation check: Source estimate · Table 4 col.6: 24.08 (SE 12.07) Republican win-probability percentage points per 1pp employment-weighted CZ Chinese import penetration change. Not per USD1000/worker and not a national election probability. Other vote-share columns and pre-2010 effects are not replaced by this result. Published AER 110(10), 2020, Table 4 col.6, printed p.3167 / PDF p.29; employment-weighted import penetration pp
Range: Table 4 col.6: 24.08 (SE 12.07) Republican win-probability percentage points per 1pp employment-weighted CZ Chinese import penetration change. Not per USD1000/worker and not a national election probability. Other vote-share columns and pre-2010 effects are not replaced by this result.

import exposure eur1000 per worker 2y → radical right vote share change pp13.2

Calculation: Accepted manuscript Table 1 col.10: .132 (SE .051) vote-share fraction per EUR1000/worker import increase over the two years before an election. Multiply output by 100 to obtain percentage points. No USD conversion, decadal-dose substitution or persistence assumption.
Parameter support: 3.204 – 23.196
Study population: 15 Western European countries; 76 elections 1988-2007; 7782 district-election observations
Evidence role: boundary
Attributed studies: Colantone, Stanig (2017)
Citation check: Source estimate · Accepted manuscript Table 1 col.10: .132 (SE .051) vote-share fraction per EUR1000/worker import increase over the two years before an election. Multiply output by 100 to obtain percentage points. No USD conversion, decadal-dose substitution or persistence assumption. Accepted AJPS manuscript, Table 1 col.10 / PDF p.9; exposure definition PDF pp.5-6; footnote 5: all monetary figures deflated to 2006 euros
Range: Accepted manuscript Table 1 col.10: .132 (SE .051) vote-share fraction per EUR1000/worker import increase over the two years before an election. Multiply output by 100 to obtain percentage points. No USD conversion, decadal-dose substitution or persistence assumption.

eviction order → quarterly earnings gap q5 80.8559

Calculation: Table VI col.6 IV earnings difference -613 (SE 248), divided by matching col.4 non-evicted sample mean 4254, not year-1 mean 4300. This scaled index holds the descriptive comparison mean fixed; it is not a separately identified complier-counterfactual earnings ratio. Never composed into displacement earnings.
Parameter support: 0.7416361071932299 – 0.9701645510108132
Study population: Cook County and NYC eviction court populations; quarterly earnings in quarters 5-8 after filing
Evidence role: boundary
Attributed studies: Collinson, Humphries, Mader, Reed, Tannenbaum, van Dijk (2022)
Citation check: Source estimate · Table VI col.6 IV earnings difference -613 (SE 248), divided by matching col.4 non-evicted sample mean 4254, not year-1 mean 4300. This scaled index holds the descriptive comparison mean fixed; it is not a separately identified complier-counterfactual earnings ratio. Never composed into displacement earnings. OUP accepted QJE manuscript, cover date September 30, 2022; download stamp September 26, 2023. Table VI cols.4,6, printed p.45 / PDF p.46; quarters 5-8
Range: Table VI col.6 IV earnings difference -613 (SE 248), divided by matching col.4 non-evicted sample mean 4254, not year-1 mean 4300. This scaled index holds the descriptive comparison mean fixed; it is not a separately identified complier-counterfactual earnings ratio. Never composed into displacement earnings.

displacement event → exam noncompletion change pp0.54

Calculation: IZA DP 16367 Table 2 Panel A col.2: probability difference .0054 (SE .0027), expressed as .54 percentage points. Regression constant .0785 is not an admitted control mean; no hazard ratio, relative-risk conversion or displaced-worker total. Published-version reconciliation remains open.
Parameter support: 0.0108 – 1.0692
Study population: Danish parental plant closures; exposure at ages 0-16; compulsory-school test taking
Evidence role: boundary
Attributed studies: Bingley, Cappellari, Ovidi (2023)
Citation check: Source estimate · IZA DP 16367 Table 2 Panel A col.2: probability difference .0054 (SE .0027), expressed as .54 percentage points. Regression constant .0785 is not an admitted control mean; no hazard ratio, relative-risk conversion or displaced-worker total. Published-version reconciliation remains open. IZA DP 16367 August 2023, Table 2 Panel A col.2, printed p.45 / PDF p.48; working-paper contrast, published-version reconciliation remains open
Range: IZA DP 16367 Table 2 Panel A col.2: probability difference .0054 (SE .0027), expressed as .54 percentage points. Regression constant .0785 is not an admitted control mean; no hazard ratio, relative-risk conversion or displaced-worker total. Published-version reconciliation remains open.

childhood parental violence exposure → adult partner victimization odds2.68

Calculation: Published Table 3, violence between parents, net of other risks: OR 2.68, reported 95% CI 1.49-4.82. Sex-adjusted pooled association, not a daughters-only causal estimate, not adult perpetration or injury, and not a household-incidence multiplier. No count conversion or daughter-chain composition.
Parameter support: 1.49 – 4.82
Study population: US community cohort, both sexes pooled; Table 3 receiving any partner violence
Evidence role: boundary
Attributed studies: Ehrensaft, Cohen, Brown, Smailes, Chen, Johnson (2003)
Citation check: Source estimate · Published Table 3, violence between parents, net of other risks: OR 2.68, reported 95% CI 1.49-4.82. Sex-adjusted pooled association, not a daughters-only causal estimate, not adult perpetration or injury, and not a household-incidence multiplier. No count conversion or daughter-chain composition. Published JCCP 71(4), 2003, Table 3, printed p.747 / PDF p.7
Range: Published Table 3, violence between parents, net of other risks: OR 2.68, reported 95% CI 1.49-4.82. Sex-adjusted pooled association, not a daughters-only causal estimate, not adult perpetration or injury, and not a household-incidence multiplier. No count conversion or daughter-chain composition.

Shared assumptions & later generations

Assumptions and limits

The control changes only the stated worker count. Totals scale with it; the study effects and their uncertainty do not change.

Fixed while you move the control: children per worker (2); follow-up window (20 years); every study effect size and its uncertainty; mortality and earnings baselines.

Children assumed: 2 children per displaced worker
Follow-up period: 20 years after displacement
Mortality timing: Five study phases; source offset +6 starts in follow-up year 7
Results shown: 3 illustrative totals and 1 later-generation earnings projection. The great-grandchild estimate remains unavailable.
Pinned model bundle: v1.54 · engine 0.4.1

Why local service jobs are not counted here

Moretti estimates how total tradable employment in a metro affects local services that sell locally, such as restaurants, retail, and business services. It does not estimate the effect of replacing one worker with another in an existing job.

For that reason, the model does not apply the local multiplier to this displacement scenario. It would need evidence of a net local loss of tradable jobs, with an industry-specific assumption, before it could report local service employment.

Explore the earnings gap across generations

Worker

80%of comparison earnings

Study estimate

Child

90.76%of comparison earnings

Study estimate

Grandchild

94.92%of comparison earnings

Model projection

Great-grandchild

Estimate withdrawnModel projection

Each figure compares expected adult earnings with the earnings expected without displacement. Dollar figures apply the model's $2,591,418 SSA lifetime-earnings benchmark to one descendant, not to an assumed number of grandchildren. The later-generation figures carry forward only the modeled child earnings gap. They do not add new harms from later displacement, health, neighborhood conditions, policy, or other events.

Displaced worker earnings: 80.0¢ per $1 expected without displacement · cross-study model assumption
Grandchild adult earnings: 94.9¢ per $1 expected without displacement · model-based projection
Grandchild lifetime earnings, per person: $2,459,722 · $131,696 below the $2,591,418 SSA benchmark
Great-grandchild adult earnings: Estimate withdrawn
Great-grandchild lifetime earnings, per person: Estimate withdrawn

How the model works

Downstream applies published study effects to a stated worker count. Each result keeps its own population, baseline, units, and assumptions.

Research & methods

Model graph

91 modeled quantities74 parameter linkspinned bundle v1.54 · engine 0.4.1each link carries a unit change and an evidence tier

Full ledger · EXACT-abstractdisplacement event → worker earningspersons → gap_multiplierdisplacement->worker_earningsConditional estimate; population or setting assumptions applySource parameter: 0.8 (support: 0.75 to 0.85)Study: jacobson1993;oreopoulos2008;davis2011Study population: US high-tenure displaced menOutcome level = 0.8 × comparison level; relative loss = 1 − 0.8

Range: Cross-study model band, not a confidence interval.

Original model note

This is the preserved bundle note. It may contain extraction or interpretation errors. The citation check above states the current source assessment.

long-run earnings multiplier; band = cross-study spread: JLS early-80s PA losses vs D&vW 2011 full-text synthesis (still 15-20% below controls after 20y)

All 74 links as a keyboard list · same details as the diagram

Study findings

Sources and assumptions

Study inputs and model assumptions

These values include source estimates, model conversions and assumptions. The labels identify unresolved source support. They are separate inputs; a worker-replacement scenario does not establish that every result applies.

Open findings: 6 reviewed study inputs
Totals in the calculator: Child earnings · excess deaths
Later generations: Structural persistence projection

Separate evidence families

These study families answer different questions. The model does not start them from a worker displacement.

Net metro tradable jobs → Local service employment: Needs a measured metro employment change and a named job mix.
Import exposure → Wages · political outcomes: A region-level trade-shock input, not a worker count.
Eviction order → Shelter use · earnings: A court order, studied by comparing cases assigned to strict and lenient judges.
Unconditional income → Schooling · youth crime: A recurring transfer program, not a displacement effect.
Neighborhood exposure → Adult mobility outcomes: Studied among families that moved between counties; combining it with displacement stays an assumption.

Null results stay in the ledger

The evidence corpus keeps null and negative findings. The receiving-community crime stream records a null aggregate effect on total crime beside the identified negative effect on violent crime. Nulls are not silent drops; they decide what the model refuses to claim.

Paths the model refuses

A path without a compatible input stays blank. Local service employment needs measured net job loss, not a worker count. The work-banned asylum property-crime branch is recorded as findings and never merged into displacement effects.

Measured outcome · populationModel input

Worker earnings

Jacobson · Davis & von Wachter · High-tenure US displacement; range across three long-run studies

Model assumption

80.0¢

per $1 expected

Parent income

Hilger (2016) · US fathers laid off; $59,832 study baseline

Source unverified · estimate withdrawn

−$8,138

in layoff year

Adult child earnings

Oreopoulos · Page · Stevens · Canadian father–son firm closures

Model conversion

−9.24%

annual earnings

College enrollment

Hilger (2016) · US children of laid-off fathers, ages 18–22

Source unverified · estimate withdrawn

−0.432 pp

from 40.66%

Grade retention

Stevens · Schaller · US children, 1+ year after involuntary household job loss

Model conversion

+0.81 pp

from 5.5%

Birth weight

Lindo (2011) · US PSID; husband displacement; sibling comparison

Model conversion

−4.59%

births after loss

Marriage dissolution

Rege · Telle · Votruba · Norwegian married couples after plant closure

Model assumption

+11%

relative likelihood

Mortality odds

Sullivan · von Wachter · High-seniority male Pennsylvania mass layoffs

Model conversion

2.67×

displacement phase

Inspect the mortality timing profile · five study phases

Displacement2.670×Study range: 2.114× to 3.371×

+1 year1.984×Study range: 1.370× to 2.873×

+2–3 years1.732×Study range: 1.324× to 2.265×

+4–5 years1.271×Study range: 0.970× to 1.666×

+6 years onward1.135×Study range: 1.034× to 1.247×

The five phases come from the study's own year-by-year table. Each range is the statistical band around that study estimate; it is not any one person's range of outcomes. The calculator converts each phase with a cited mortality baseline and compares survival over the follow-up.

Methods & equations

The technical record for the pinned bundle: what each study result may be used for, how uncertainty is carried, and where every number can be inspected.

Each study result has a permitted use

The model first states what a study result measures. It then applies only the matching calculation: an earnings ratio changes an earnings level, a mortality ratio needs a mortality baseline, and a local-jobs result stays a jobs result. This prevents combining numbers that describe different things.

\begin{aligned} \text{level:}\;&v' = vp & \text{direct:}\;&v'=p \\ \text{remaining loss:}\;&v'=1-t(1-v) & \text{rate:}\;&q'=q_{0}\,p \\ \text{count ratio:}\;&n'=np & \text{elasticity:}\;&\Delta\%y=p\,\Delta\%x \end{aligned}

A rate becomes a count only after the model applies a cited baseline. A local-jobs multiplier cannot be added to an earnings result.

Later generations carry only the remaining earnings gap

The child estimate is the starting point. The model applies an intergenerational transmission assumption t to the part of the gap that remains. It does not multiply the original shock into every generation or add other harms later in life.

\begin{aligned} g_1 &= p_{\mathrm{displacement}\to\mathrm{child}} \\ g_2 &= 1-t_{12}(1-g_1) \\ g_3 &= 1-t_{23}(1-g_2) \\ \neq\;& g_1t_{12}\quad\text{or}\quad g_1g_2 \end{aligned}

If the child estimate has no gap (g₁ = 1), the later-generation estimates also have no gap. Grandchild and great-grandchild figures are structural projections, not direct study estimates.

Why the model does not add every paper together

The child-dollar result uses the direct displacement study. Divorce and family-size studies can describe the same children, so adding them would count one loss twice. Local-service-job research answers a separate question: what happens after a metro's net tradable employment changes.

Displacement→Child earningscounted onceDivorce / family size→Child earningsvisible, not addedNet tradable jobs lost→Local servicesrequires net-job evidence

Mortality adds risks across time

The model converts five study odds ratios with an annual mortality baseline and compares survival over 20 years. Source offset +6 starts in follow-up year 7.

\begin{aligned} q(r) &= \frac{rm}{1-m+rm} \\ S_Y &= \prod_{y=1}^{Y}(1-q_y) \\ D_{\mathrm{excess}} &= N(S_{Y,\mathrm{baseline}}-S_{Y,\mathrm{exposed}}) \end{aligned}

How the model checks complex uncertainty

For independent paths, the model calculates the first three moments of the combined distribution and compares its approximation with a joint simulation. Nonlinear or correlated paths use the simulation directly. Repeated transmission relationships share one uncertainty draw.

\begin{aligned} \mu &= \mathbb{E}[V] \\ \sigma^2 &= \mathbb{E}[V^2]-\mu^2 \\ \gamma_1 &= \frac{\mathbb{E}[V^3]-3\mu\sigma^2-\mu^3}{\sigma^3} \end{aligned}

Open the equation key and model rules

The equations use relative levels unless they explicitly produce a count or dollar amount. A level of 1 means the no-displacement comparison; a level below 1 means a loss.

v: earnings level before the next link; 1 means no earnings gap
v′: earnings level after the next link
p: the cited study estimate or multiplier
t: share of the remaining earnings gap passed to the next generation

g₁, g₂, g₃: child, grandchild, and great-grandchild earnings levels
n: a count before a count multiplier is applied
x, y: an input and its outcome in an elasticity calculation
Δ%: a percentage change

N: number of workers in the stated scenario
m: annual mortality risk in the comparison population
r: study odds ratio for mortality
q(r), Sᵧ, Dₑₓcₑss: exposed annual risk, survival through year Y, and excess deaths

q₀: annual rate before a study multiplier is applied
Y: number of follow-up years
V: a combined model result treated as a random variable in simulation
μ, σ², γ₁: the mean, variance, and skewness of V

Each rule limits how one type of evidence can enter a calculation.
Type	Operation	What it prevents
remaining loss	1 − t × (1 − g)	Treating a transmission rate as a new earnings level.
relative risk	Apply to a cited baseline	Reporting an absolute count without a baseline.
count multiplier	Multiply one count or dollar level	Chaining a jobs result into an earnings result.
elasticity	%Δ outcome = elasticity × %Δ input	Dropping the input unit or changing the sign.

Source studies

Confidence intervals stay with their studies

A confidence interval describes the sampling uncertainty of one study estimate. The model retains those source details. It does not call a combined model result a confidence interval.

Technical materials

Ranges and simulations are documented separately

The white paper and source repository show what changes when declared parameters vary, and the simulation results for eligible paths. Those are sensitivity checks, not odds that a real-world outcome falls inside a stated range.

Neither procedure resolves uncertainty about the documented exposure, baseline measurement, use in a new population, model choices, or effects that the evidence does not identify.

Reproduce the reference scenario

The pinned bundle (v1.54, engine 0.4.1) produces the numbers in the reference scenario:

downstream scenario --workers 1000 --children 2 --tradable-share 1 --exposure-years 20

Inspect the bibliography · 105 entries

The identified bibliography contains 108 sources and 214 named author records (people and organizations), 1979-2026.

The parameters cite 49 sources with 117 named author records; these are not counts of independent research teams.

The bibliography includes research, methods and context sources; inclusion does not establish peer review or independent validation. Modeled numbers carry their cited inputs and declared structural assumptions.

The bundle contains 110 reference records. Duplicate identifiers share one document entry. These labels describe the bundle’s extraction records; they do not certify that each source supports every modeled outcome.

Earnings Losses of Displaced Workers
Jacobson, Louis S. and LaLonde, Robert J. and Sullivan, Daniel G. · 1992 · jacobson1993
Source edition: Working-paper edition registered by this DOI; 1992. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: published abstract
DOI →
The Intergenerational Effects of Worker Displacement
Oreopoulos, Philip and Page, Marianne E. and Stevens, Ann Huff · 2008 · oreopoulos2008
Source edition: UC Davis author manuscript, Table 2, column 3 (PDF page 31); manuscript has no printed revision date. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →DOI →
Mortality, Mass-Layoffs, and Career Outcomes: An Analysis using Administrative Data
Sullivan, Daniel and von Wachter, Till · 2007 · sullivan2009
Source edition: NBER Working Paper 13626, Table 5, column 3 (PDF page 54). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →DOI →
Plant Closure and Marital Dissolution
Rege, Mari and Telle, Kjetil and Votruba, Mark · 2007 · rege2007
Source edition: Statistics Norway Discussion Paper 514 (2007), Tables 1–2 (PDF pages 12–13). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: published abstract
Source →
Job Displacement, Disability, and Divorce
Charles, Kerwin Kofi and Stephens, Melvin, Jr. · 2001 · charles2004
Source edition: Working-paper edition registered by this DOI; 2001. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
Is Making Divorce Easier Bad for Children? The Long-Run Implications of Unilateral Divorce
Gruber, Jonathan · 2000 · gruber2004
Source edition: NBER WP 7968, November 2000; unilateral-divorce-law exposure ITT, not individual parental divorce. The original model key is retained when the cited edition’s year differs.
primary PDF; SHA-256 c60742be3ee2802c5cb547475795f043e7b042258df664e792399005f35a6c80
Source mismatch · estimate withdrawn · Unilateral-divorce-law ITT does not identify the earnings effect of an individual parental divorce. Its sex/cohort estimates cannot justify a common .95 multiplier.
Range: Model band; the cited source does not establish this parameter.
Bundle extraction record: listed reference; extraction not verified here
Source →
The More the Merrier? The Effect of Family Composition on Children’s Education
Black, Sandra E. and Devereux, Paul J. and Salvanes, Kjell G. · 2004 · black2005
Source edition: NBER WP 10720 September 2004, abstract and study outcome definition; published QJE 120(2), 2005. The original model key is retained when the cited edition’s year differs.
primary working-paper PDF; SHA-256 d1914a1e21db77046631d3153ab46a9b72856325e34eceee06124b39fcaefc30
Source mismatch · estimate withdrawn · Outcome is education, with family-size effects becoming negligible after birth-order controls and twins IV. No .98 child-earnings multiplier follows; a null schooling finding cannot be converted to zero earnings harm.
Range: Model band; the cited source does not establish this parameter.
Bundle extraction record: listed reference; extraction not verified here
Source →
Intergenerational Income Mobility in the United States
Solon, Gary · 1992 · solon1992
Source edition: American Economic Review 82(3), 393–408 (1992). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
Income Inequality, Equality of Opportunity, and Intergenerational Mobility
Corak, Miles · 2013 · corak2013
Source edition: IZA Discussion Paper7520 (2013), literature synthesis. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
Where is the Land of Opportunity? The Geography of Intergenerational Mobility in the United States
Chetty, Raj and Hendren, Nathaniel and Kline, Patrick and Saez, Emmanuel · 2014 · chetty2014
Bundle extraction record: listed reference; extraction not verified here
DOI →
The Intergenerational Persistence of Human Capital: An Empirical Analysis of Four Generations
Lindahl, Mikael and Palme, M{\aa}rten and Massih, Sofía Sandgren and Sj{\"o}gren, Anna · 2012 · lindahl2015
Source edition: IZA DP 6463 April 2012, abstract and earnings-generation scope; earnings for three generations, education for four. The original model key is retained when the cited edition’s year differs.
primary PDF; SHA-256 9c3610b1a6e4a0640ff2710dee4a5c503a23d4a27bf2d1c723f3c51568d8135b
Source unverified · estimate withdrawn · The source measures earnings for three generations and education for four. It does not estimate .55 causal gap transmission to a fourth earnings generation. Retaining a projection would require an explicit uncalibrated assumption, not a source-identified default.
Range: Assumed persistence band; source support unresolved.
Bundle extraction record: listed reference; extraction not verified here
Source →
The Cycle of Violence
Widom, Cathy Spatz · 1989 · widom1989
Source unverified · estimate withdrawn · Retired identifier; use childhood_parental_violence_exposure->adult_partner_victimization_odds. Published Table 3, violence between parents, net of other risks: OR 2.68, reported 95% CI 1.49-4.82. Sex-adjusted pooled association, not a daughters-only causal estimate, not adult perpetration or injury, and not a household-incidence multiplier. No count conversion or daughter-chain composition.
Range: Unverified model band.
Bundle extraction record: listed reference; extraction not verified here
DOI →
Intergenerational Transmission of Partner Violence: A 20-Year Prospective Study
Ehrensaft, Miriam K. and Cohen, Patricia and Brown, Jocelyn and Smailes, Elizabeth and Chen, Henian and Johnson, Jeffrey G. · 2003 · ehrensaft2003
Source edition: Published JCCP 71(4), 2003, Table 3, printed p.747 / PDF p.7. The original model key is retained when the cited edition’s year differs.
numeric table extraction from publisher PDF viewed through browser; not original PDF byte hash; SHA-256 3901f2e76a1eee05a952909bf2dfe3bbecefad87b1e06b8afd5b1042e358500f
Source unverified · estimate withdrawn · Retired identifier; use childhood_parental_violence_exposure->adult_partner_victimization_odds. Published Table 3, violence between parents, net of other risks: OR 2.68, reported 95% CI 1.49-4.82. Sex-adjusted pooled association, not a daughters-only causal estimate, not adult perpetration or injury, and not a household-incidence multiplier. No count conversion or daughter-chain composition.
Range: Unverified model band.
Bundle extraction record: full-text table
Source →
Local Multipliers
Moretti, Enrico · 2010 · moretti2010
Source edition: Author-hosted 2010 prepress manuscript, Section III and Tables 1–2 (PDF pages 4–5). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: published abstract
Source →DOI →
The Effects of School Spending on Educational and Economic Outcomes: Evidence from School Finance Reforms
Jackson, C. Kirabo and Johnson, Rucker C. and Persico, Claudia · 2015 · jackson2016
Source edition: NBER Working Paper 20847 (2015), abstract (PDF page 2); reported wage effect, not a reported 3%–10% interval. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →DOI →
Crime Rates and Local Labor Market Opportunities in the United States: 1979--1997
Gould, Eric D. and Weinberg, Bruce A. and Mustard, David B. · 2002 · gould2002
Source edition: Original Review of Economics and Statistics article on Stanford’s research archive. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →DOI →
Market Wages and Youth Crime
Grogger, Jeff · 1997 · grogger1998
Source edition: Working-paper edition registered by this DOI; 1997. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
Long-Term Earnings Losses due to Mass Layoffs During the 1982 Recession: An Analysis Using U.S. Administrative Data from 1974 to 2004
von Wachter, Till and Song, Jae and Manchester, Joyce · 2009 · vonwachter2009
Source edition: April 2009 conference draft. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
Does Job Loss Shorten Life?
Eliason, Marcus and Storrie, Donald · 2009 · eliason2009; eliasonstorrie2009
Bundle extraction record: listed reference; extraction not verified here
DOI →
When Work Disappears: Manufacturing Decline and the Falling Marriage-Market Value of Young Men
Autor, David and Dorn, David and Hanson, Gordon · 2019 · autor2019
Bundle extraction record: listed reference; extraction not verified here
DOI →
The Gender Wage Gap and Domestic Violence
Aizer, Anna · 2010 · aizer2010
Source edition: Author manuscript in PMC4123456, Table 2 Panel A col.3; machine-readable primary full-text XML. The original model key is retained when the cited edition’s year differs.
primary full text XML; SHA-256 c9498f4af8b33f072127eb3de99f9afefd13af120941c361eac8dd12428b3410
Source mismatch · estimate withdrawn · Retired identifier; use female_male_wage_ratio_change->female_assault_log_change. Table 2 Panel A col.3: -.813 (SE .317). Change in log female assault hospitalizations per absolute unit change in potential female/male wage ratio. Input .01 means one ratio percentage point; not a relative wage multiplier or log-log elasticity. Hospitalization outcome is not all household IPV incidence.
Range: The coefficient’s SE is 0.317; its interval does not resolve the unit mismatch.
Bundle extraction record: listed reference; extraction not verified here
Source →
Male Earnings, Marriageable Men, and Nonmarital Fertility: Evidence from the Fracking Boom
Kearney, Melissa S. and Wilson, Riley · 2017 · kearney2020; kearney2018fracking
Source edition: NBER Working Paper 23408 (2017), Table 8, Panel B, column 2 (PDF page 45). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →DOI →
The Trade Origins of Economic Nationalism: Import Competition and Voting Behavior in Western Europe
Colantone, Italo and Stanig, Piero · 2017 · colantone2018
Source edition: Accepted AJPS manuscript, Table 1 col.10 / PDF p.9; exposure definition PDF pp.5-6; footnote 5: all monetary figures deflated to 2006 euros. The original model key is retained when the cited edition’s year differs.
primary PDF; SHA-256 93398be18ac49dc887fdefb900668e3edc627d39ca2407f7bb968df7230d96b3
Source mismatch · estimate withdrawn · Retired identifier; use import_exposure_eur1000_per_worker_2y->radical_right_vote_share_change_pp. Accepted manuscript Table 1 col.10: .132 (SE .051) vote-share fraction per EUR1000/worker import increase over the two years before an election. Multiply output by 100 to obtain percentage points. No USD conversion, decadal-dose substitution or persistence assumption.
Range: Rounded coefficient ± 1.96 standard errors; the exposure-unit mismatch remains.
Bundle extraction record: full-text table
DOI →
E Pluribus Unum: Diversity and Community in the Twenty-first Century—The 2006 Johan Skytte Prize Lecture
Putnam, Robert D. · 2007 · putnam2007
Bundle extraction record: listed reference; extraction not verified here
DOI →
The Association Between Income and Life Expectancy in the United States, 2001--2014
Chetty, Raj and Stepner, Michael and Abraham, Sarah and Lin, Shelby and Scuderi, Benjamin and Turner, Nicholas and Schrag, Augustin and Cutler, David · 2016 · chetty2016jama; chetty2016
Source edition: Results: income-dollar concavity example ($14,000 to $20,000). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
Parental Job Loss and Infant Health
Lindo, Jason M. · 2010 · lindo2011
Source edition: IZA Discussion Paper 5213, September 2010, Table 2, column 3 (PDF page 26). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
Short-Run Effects of Parental Job Loss on Children's Academic Achievement
Stevens, Ann Huff and Schaller, Jessamyn · 2009 · stevens2011
Source edition: NBER Working Paper 15480 (2009), Table 4, column 1 (PDF page 38); sample mean in results prose. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →DOI →
Linking Job Loss, Inequality, Mental Health, and Education
Ananat, Elizabeth O. and Gassman-Pines, Anna and Francis, David V. and Gibson-Davis, Christina M. · 2017 · ananat2017
Bundle extraction record: published abstract
DOI →
Unemployment Impairs Mental Health: Meta-Analyses
Paul, Karsten I. and Moser, Klaus · 2009 · paul2009
Source edition: Published JVB 74, 2009, Table 1 Overall (out. ex), printed p.271 / PDF p.8. The original model key is retained when the cited edition’s year differs.
primary PDF; SHA-256 96897f014c89125dc65236d22b0f2a9e69849b7fb7e49ae7cf61a27d855a9e11
Bundle extraction record: full-text table
DOI →
The Impact of Single-Family Mortgage Foreclosures on Neighborhood Crime
Immergluck, Dan and Smith, Geoff · 2006 · immergluck2006
Bundle extraction record: listed reference; extraction not verified here
DOI →
Forced Sales and House Prices
Campbell, John Y. and Giglio, Stefano and Pathak, Parag · 2009 · campbell2011
Source edition: NBER Working Paper 14866 (2009), abstract and neighborhood externality estimates. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →DOI →
The Effects of High-Skilled Immigration Policy on Firms: Evidence from Visa Lotteries
Doran, Kirk and Gelber, Alexander and Isen, Adam · 2014 · doran2022
Source edition: Working-paper edition registered by this DOI; 2014. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
DOI →
Immigration and Crime: Assessing a Contentious Issue
Ousey, Graham C. and Kubrin, Charis E. · 2018 · ousey2018
Bundle extraction record: published abstract
DOI →
Recessions and the Costs of Job Loss
Davis, Steven J. and von Wachter, Till · 2011 · davis2011
Source edition: NBER Working Paper17638 (2011, revised March2017), earnings-loss discussion. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →DOI →
Importing Political Polarization? The Electoral Consequences of Rising Trade Exposure
Autor, David and Dorn, David and Hanson, Gordon and Majlesi, Kaveh · 2016 · autor2020
Source edition: Working-paper edition registered by this DOI; 2016. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: published abstract
DOI →
Did Austerity Cause Brexit?
Fetzer, Thiemo · 2018 · fetzer2019
Source edition: Edition registered by this DOI; 2018. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
Growing Up with a Single Parent: What Hurts, What Helps
McLanahan, Sara and Sandefur, Gary · 1994 · mclanahan1994
Bundle extraction record: listed reference; extraction not verified here
Source →DOI →
Does Growing Up in a High Crime Neighborhood Affect Youth Criminal Behavior?
Damm, Anna Piil and Dustmann, Christian · 2013 · damm2014
Source edition: Aarhus Economics Working Paper 2013-17, Tables 3 and 5 (PDF pages 40 and 42). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
International Migration, Remittances, and Household Investment: Evidence from Philippine Migrants' Exchange Rate Shocks
Yang, Dean · 2006 · yang2008
Source edition: Working-paper edition registered by this DOI; 2006. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
Liquidity versus Wealth in Household Debt Obligations: Evidence from Housing Policy in the Great Recession
Ganong, Peter and Noel, Pascal · 2020 · ganong2022
Source edition: American Economic Review 110(10), 3100–3138 (October 2020). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
Intimate Partner Violence in the Great Recession
Schneider, Daniel and Harknett, Kristen and McLanahan, Sara · 2016 · schneider2016
Source edition: Tables 1 and 3 and results prose on predicted probabilities. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
Exposure to Domestic Violence: A Meta-Analysis of Child and Adolescent Outcomes
Evans, Sarah E. and Davies, Corrie and DiLillo, David · 2008 · evans2008
Source edition: Published Aggression and Violent Behavior 13, 2008, printed p.135 / PDF p.5, Table 1 and sections 3.2.1-3.2.2. The original model key is retained when the cited edition’s year differs.
primary PDF; SHA-256 a876bf1205d8b7a3a80cb493e42c810f0c5b98783bd44626565d5236f7a46f3f
Bundle extraction record: full-text table
DOI →
Relationships Between Poverty and Psychopathology: A Natural Experiment
Costello, E. Jane and Compton, Stephen N. and Keeler, Gordon and Angold, Adrian · 2003 · costello2003
Bundle extraction record: listed reference; extraction not verified here
DOI →
Parents' Incomes and Children's Outcomes: A Quasi-Experiment Using Transfer Payments from Casino Profits
Akee, Randall K. Q. and Copeland, William E. and Keeler, Gordon and Angold, Adrian and Costello, E. Jane · 2010 · akee2010
Source edition: Author manuscript of the 2010 AEJ Applied article; Tables 4 and 6. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
Early-Childhood Poverty and Adult Attainment, Behavior, and Health
Duncan, Greg J. and Ziol-Guest, Kathleen M. and Kalil, Ariel · 2010 · duncan2010
Bundle extraction record: full-text table
DOI →
Mother's Education and the Intergenerational Transmission of Human Capital: Evidence from College Openings
Currie, Janet and Moretti, Enrico · 2003 · currie2003
Source edition: Published Quarterly Journal of Economics article, November 2003, 1495–1532. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
Criminal Careers up to Age 50 and Life Success up to Age 48: New Findings from the Cambridge Study in Delinquent Development
Farrington, David P. and Coid, Jeremy W. and Harnett, Louise M. and Jolliffe, Darrick and Soteriou, Nadine and Turner, Richard E. and West, Donald J. · 2006 · farrington2006
Source edition: NCJ 216170; September 2006 report, 94 pages. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
Recent Immigrants: Unexpected Implications for Crime and Incarceration
Butcher, Kristin F. and Piehl, Anne Morrison · 1997 · butcher1998
Source edition: Working-paper edition registered by this DOI; 1997. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
Austerity and the Rise of the Nazi party
Galofr{\'e}-Vil{\`a}, Gregori and Meissner, Christopher M. and McKee, Martin and Stuckler, David · 2017 · galofrevila2021
Source edition: Working-paper edition registered by this DOI; 2017. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
The Labor Demand Curve Is Downward Sloping: Reexamining the Impact of Immigration on the Labor Market
Borjas, George J. · 2003 · borjas2003
Bundle extraction record: listed reference; extraction not verified here
DOI →
Rethinking the Effect of Immigration on Wages
Ottaviano, Gianmarco I. P. and Peri, Giovanni · 2012 · ottaviano2012
Bundle extraction record: listed reference; extraction not verified here
DOI →
The Impact of the Mariel Boatlift on the Miami Labor Market
Card, David · 1989 · card1990
Source edition: Working-paper edition registered by this DOI; 1989. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
Group Loyalty and the Taste for Redistribution
Luttmer, Erzo F. P. · 2001 · luttmer2001
Bundle extraction record: listed reference; extraction not verified here
DOI →
A Meta-Analytic Test of Intergroup Contact Theory
Pettigrew, Thomas F. and Tropp, Linda R. · 2006 · pettigrew2006
Bundle extraction record: listed reference; extraction not verified here
DOI →
The Economic and Fiscal Consequences of Immigration
{National Academies of Sciences, Engineering, and Medicine} · 2017 · naidich2020
Bundle extraction record: listed reference; extraction not verified here
DOI →
Employment Hysteresis from the Great Recession
Yagan, Danny · 2017 · yagan2019
Source edition: Working-paper edition registered by this DOI; 2017. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
Where Have the Middle-Wage Workers Gone? A Study of Polarization Using Panel Data
Cortes, Guido Matias · 2016 · cortes2016
Bundle extraction record: listed reference; extraction not verified here
DOI →
The Short- and Long-Term Career Effects of Graduating in a Recession
Oreopoulos, Philip and von Wachter, Till and Heisz, Andrew · 2012 · oreopoulos2012
Bundle extraction record: listed reference; extraction not verified here
DOI →
Job displacement and stress-related health outcomes
Browning, Martin and Danø, Anne Møller and Heinesen, Eskil · 2006 · browning2006
Bundle extraction record: listed reference; extraction not verified here
DOI →
Rising Morbidity and Mortality in Midlife among White Non-Hispanic Americans in the 21st Century
Case, Anne and Deaton, Angus · 2015 · case2015
Bundle extraction record: listed reference; extraction not verified here
Source →DOI →
Unemployment and Suicide: A Cohort Analysis of Social Factors Predicting Suicide in the US National Longitudinal Mortality Study
Kposowa, Augustine J. · 2001 · kposowa2001
Bundle extraction record: listed reference; extraction not verified here
DOI →
Macroeconomic Conditions and Opioid Abuse
Hollingsworth, Alex and Ruhm, Christopher J. and Simon, Kosali · 2017 · holliingsworth2017
Bundle extraction record: listed reference; extraction not verified here
DOI →
The Rise in the Disability Rolls and the Decline in Unemployment
Autor, David H. and Duggan, Mark G. · 2003 · autor2003
Source edition: Quarterly Journal of Economics, February 2003, 157–206. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
Does the availability of high-wage jobs for low-skilled men affect welfare expenditures? Evidence from shocks to the steel and coal industries
Black, Dan A. and McKinnish, Terra G. and Sanders, Seth G. · 2003 · black2003
Source edition: Edition registered by this DOI; 2003. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
Public Goods and Ethnic Divisions
Alesina, Alberto and Baqir, Reza and Easterly, William · 1997 · alesina1999
Source edition: Working-paper edition registered by this DOI; 1997. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
Migration and Development Brief 39
{World Bank} · 2023 · worldbank2023
Source edition: Migration and Development Brief 39, December 2023. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
The Fragile Middle Class: Americans in Debt
Sullivan, Teresa A. and Warren, Elizabeth and Westbrook, Jay Lawrence · 2000 · sullivan2000
Source edition: Original Yale University Press book, 2000. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
Public service reductions associated with placement of homeless persons with severe mental illness in supportive housing
Culhane, Dennis P. and Metraux, Stephen and Hadley, Thomas · 2002 · culhane2007
Source edition: Edition registered by this DOI; 2002. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
DOI →
A Comparison of Three Methods for Selecting Values of Input Variables in the Analysis of Output from a Computer Code
McKay, Michael D. and Beckman, Richard J. and Conover, William J. · 1979 · mckay1979
Bundle extraction record: listed reference; extraction not verified here
DOI →
A Distribution-Free Approach to Inducing Rank Correlation Among Input Variables
Iman, Ronald L. and Conover, W. J. · 1982 · iman1982
Bundle extraction record: listed reference; extraction not verified here
DOI →
Making Best Use of Model Evaluations to Compute Sensitivity Indices
Saltelli, Andrea · 2002 · saltelli2002
Bundle extraction record: listed reference; extraction not verified here
DOI →
Global Sensitivity Indices for Nonlinear Mathematical Models and Their Monte Carlo Estimates
Sobol', Ilya M. · 2001 · sobol2001
Bundle extraction record: listed reference; extraction not verified here
DOI →
Decomposition of the Continuous Ranked Probability Score for Ensemble Prediction Systems
Hersbach, Hans · 2000 · hersbach2000
Bundle extraction record: listed reference; extraction not verified here
DOI →
Calculated Risks: How to Know When Numbers Deceive You
Gigerenzer, Gerd · 2002 · gigerenzer2002
Source edition: Original Simon & Schuster book (2002), author publication list. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
The {China Syndrome}: Local Labor Market Effects of Import Competition in the {United States}
Autor, David H. and Dorn, David and Hanson, Gordon H. · 2012 · adh2013
Source edition: NBER Working Paper 18054 (2012), Table 7, Panel B, column 6 (PDF page 29). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →DOI →
The Impact of Family Income on Child Achievement: Evidence from the {Earned Income Tax Credit}
Dahl, Gordon B. and Lochner, Lance · 2008 · dahl2012
Source edition: NBER Working Paper 14599, December 2008, revised November 2010; Table 3, column i (PDF page 38). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
Parental Job Loss and Children's Long-Term Outcomes: Evidence from 7 Million Fathers' Layoffs
Hilger, Nathaniel G. · 2016 · hilger2016
Source edition: Published AEJ: Applied Economics 8(3), 2016, Table 3, printed p.265 / PDF p.19. The original model key is retained when the cited edition’s year differs.
primary PDF; SHA-256 5f8f90917584d1b6d211dd46fb003de99c65fd79d51bb2d5b07ffc9a9b7105f8
Bundle extraction record: full-text table
DOI →
Identifying the Effect of Unemployment on Crime
Raphael, Steven and Winter-Ebmer, Rudolf · 2001 · raphael2001
Source edition: Published JLE 44(1), 2001, Table 5 col.6, printed p.275 / PDF p.18; unemployment is a fraction. The original model key is retained when the cited edition’s year differs.
primary PDF; SHA-256 a594607279054a5e18fd651a7e1ccc5a6d9353e886b6397a37711767d23b0f1d
Source unverified · estimate withdrawn · Retired identifier; use unemployment_rate_change_pp->property_crime_log_change. Table 5 col.6: b=5.018 (SE 1.134) per unit unemployment fraction. Per 1pp change, log crime changes by .05018; percent response is 100*expm1(.05018*dose). The old 5.018 percent reading was a first-order approximation, not an exact percent effect.
Range: Source-table extraction unverified.
Bundle extraction record: full-text table
DOI →
The Long-Term Impact of the {Earned Income Tax Credit} on Children's Education and Employment Outcomes
Bastian, Jacob and Michelmore, Katherine · 2018 · bastian2018
Source edition: Publisher 2018 accepted preprint, Table 2 col.6, PDF p.32; 2013 dollars; ages 13-18 exposure, ages 22-27 earnings. The original model key is retained when the cited edition’s year differs.
primary PDF; SHA-256 e62e763adbaff2dbb1da66dc6431f3ce7d0faa156a613e3b68ee709cada62fbb
Bundle extraction record: full-text table
DOI →
Importing Political Polarization? The Electoral Consequences of Rising Trade Exposure
Autor, David and Dorn, David and Hanson, Gordon and Majlesi, Kaveh · 2020 · autor2020polarization
Source edition: Published AER 110(10), 2020, Table 4 col.6, printed p.3167 / PDF p.29; employment-weighted import penetration pp. The original model key is retained when the cited edition’s year differs.
primary PDF; SHA-256 e85aaaad15cc2f6220fb45e17803d2f5f30b24eb271dda3c211b418012756e6c
Source mismatch · estimate withdrawn · Retired identifier; use import_penetration_change_pp->gop_win_probability_change_pp. Table 4 col.6: 24.08 (SE 12.07) Republican win-probability percentage points per 1pp employment-weighted CZ Chinese import penetration change. Not per USD1000/worker and not a national election probability. Other vote-share columns and pre-2010 effects are not replaced by this result.
Range: The published coefficient interval cannot repair the exposure-unit mismatch.
Bundle extraction record: full-text table
Source →
Caution! Men Not at Work: Gender-Specific Labor Market Conditions and Child Maltreatment
Lindo, Jason M. and Schaller, Jessamyn and Hansen, Benjamin · 2013 · lindo2018maltreatment
Source edition: NBER Working Paper 18994 (2013), Table 3, Panel A, column 3 (PDF page 41). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →DOI →
Housing and Employment Insecurity among the Working Poor
Desmond, Matthew and Gershenson, Carl · 2016 · desmond2016
Bundle extraction record: full-text table
DOI →
Eviction and Poverty in American Cities
Collinson, Robert and Humphries, John Eric and Mader, Nicholas S. and Reed, Davin K. and Tannenbaum, Daniel I. and van Dijk, Winnie · 2022 · collinson2024
Source edition: OUP accepted QJE manuscript, cover date September 30, 2022; download stamp September 26, 2023. Table VI cols.4,6, printed p.45 / PDF p.46; quarters 5-8. The original model key is retained when the cited edition’s year differs.
primary PDF; SHA-256 80c99a680a0096b16f05d04adbec0dac518dfc157acc8bebf4b7e112e393c180
Source mismatch · estimate withdrawn · Retired identifier; use eviction_order->quarterly_earnings_gap_q5_8. Table VI col.6 IV earnings difference -613 (SE 248), divided by matching col.4 non-evicted sample mean 4254, not year-1 mean 4300. This scaled index holds the descriptive comparison mean fixed; it is not a separately identified complier-counterfactual earnings ratio. Never composed into displacement earnings.
Range: The source SE is 248; the model’s denominator is from the wrong period.
Bundle extraction record: full-text table
Source →
The Impacts of Neighborhoods on Intergenerational Mobility {I}: Childhood Exposure Effects
Chetty, Raj and Hendren, Nathaniel · 2016 · chettyhendren2018
Source edition: NBER Working Paper 23001 (2016, revised May 2017), Appendix Table V, column 1 (PDF page 57). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →DOI →
Early Childhood Poverty, Immune-Mediated Disease Processes, and Adult Productivity
Ziol-Guest, Kathleen M. and Duncan, Greg J. and Kalil, Ariel and Boyce, W. Thomas · 2012 · zg2012
Bundle extraction record: full-text table
DOI →
Parental Job Loss and Children's Careers
Huttunen, Kristiina and Riukula, Krista · 2019 · huttunen2019
Source edition: IZA Discussion Paper 12788, November 2019. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
Job Displacement Among Single Mothers: Effects on Children's Outcomes in Young Adulthood
Brand, Jennie E. and Simon Thomas, Juli · 2014 · brand2014
Source edition: Author manuscript of the 2014 AJS article; Table 3, propensity-score kernel matching TT column. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →DOI →
Effect of Job Loss Due to Plant Closure on Mortality and Hospitalisation
Browning, Martin and Heinesen, Eskil · 2011 · browningheinesen2012
Source edition: AKF working paper, April 2011; Tables 6.1–6.2 (PDF pages 23 and 25). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
The Effect of Unemployment on the Mental Health of Spouses: Evidence from Plant Closures in Germany
Marcus, Jan · 2012 · marcus2013
Source edition: SOEPpapers 488 (2012), Table 3, specification 4 (PDF page 17). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
When It Hurts the Most: Timing of Parental Job Loss and a Child's Education
Bingley, Paul and Cappellari, Lorenzo and Ovidi, Marco · 2023 · bingley2026
Source edition: IZA DP 16367 August 2023, Table 2 Panel A col.2, printed p.45 / PDF p.48; working-paper contrast, published-version reconciliation remains open. The original model key is retained when the cited edition’s year differs.
primary PDF; SHA-256 f8acbb13956b08b654876e9ee7b0958e74c91ee9c6d5e804a6eb5ff40e731c0d
Source mismatch · estimate withdrawn · Retired identifier; use displacement_event->exam_noncompletion_change_pp. IZA DP 16367 Table 2 Panel A col.2: probability difference .0054 (SE .0027), expressed as .54 percentage points. Regression constant .0785 is not an admitted control mean; no hazard ratio, relative-risk conversion or displaced-worker total. Published-version reconciliation remains open.
Range: Coefficient uncertainty scaled by an unsupported denominator; not an estimated hazard ratio.
Bundle extraction record: full-text table
Source →
Short-Run Effects of Job Loss on Health Conditions, Health Insurance, and Health Care Utilization
Schaller, Jessamyn and Stevens, Ann Huff · 2015 · schallerstevens2015
Bundle extraction record: listed reference; extraction not verified here
DOI →
Job Loss, Firm-Level Heterogeneity and Mortality: Evidence from Administrative Data
Bloemen, Hans and Hochguertel, Stefan and Zweerink, Jochem · 2018 · bloemen2018
Bundle extraction record: full-text table
DOI →
Job Displacement, Family Dynamics, and Spousal Labor Supply
Halla, Martin and Schmieder, Julia and Weber, Andrea · 2020 · halla2020
Source edition: AEA-hosted manuscript, Table 2 (PDF page 39), subgroup tables (PDF pages 41–43). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →DOI →
The Effect of Job Displacement on Couples' Fertility Decisions
Huttunen, Kristiina and Kellokumpu, Jenni · 2012 · huttunenkellokumpi2016
Source edition: IZA Discussion Paper 6707, July 2012; Table 3, plant-closure sample (PDF page 39). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
Short-Run Effects of Parental Job Loss on Child Health
Schaller, Jessamyn and Zerpa, Mariana · 2015 · schallerzerpa2019
Source edition: NBER Working Paper 21745, November 2015; Table 2, father-displaced tenured column (PDF page 31). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
Consequences of Parental Job Loss on the Family Environment and on Human Capital Formation: Evidence from Plant Closures
Mörk, Eva and Sjögren, Anna and Svaleryd, Helena · 2019 · mork2019
Source edition: IZA Discussion Paper 12559, August 2019. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
Like Father, Like Son? A Note on the Intergenerational Transmission of IQ Scores
Black, Sandra E. and Devereux, Paul J. and Salvanes, Kjell G. · 2008 · blackdevereuxsalvanes2009
Source edition: IZA Discussion Paper 3651 (2008), Results (PDF page 8), Table 3. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
From the Cradle to the Labor Market? The Effect of Birth Weight on Adult Outcomes
Black, Sandra E. and Devereux, Paul J. and Salvanes, Kjell G. · 2005 · black2007
Source edition: NBER Working Paper 11796 (2005), Table 5, All Twins / Ln(Earnings) FT (PDF page 50). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →DOI →
Birth Weight and Risk of Type 2 Diabetes Mellitus, Cardiovascular Disease, and Hypertension in Adults: A Meta-Analysis of 7~646~267 Participants From 135 Studies
Knop, Michael R. and Geng, Tianti T. and Gorny, Anna W. and Ding, Rui and Li, Changwei and Ley, Sylvia H. and Huang, Tao · 2018 · knop2018
Source edition: Results: dose-response analyses of T2DM and CVD per 1 kg birth weight. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
Separated at Girth: US Twin Estimates of the Effects of Birth Weight
Royer, Heather · 2009 · royer2009
Bundle extraction record: published abstract
DOI →
Intergenerational Consequences of Family Disruption
McLanahan, Sara and Bumpass, Larry · 1986 · mclanahanbumpass1988
Source edition: IRP Discussion Paper 805-86 (May 1986), Appendix Table A1, white women, first background-adjusted block (PDF page 26). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: listed reference; extraction not verified here
Source →
Does Undocumented Immigration Increase Violent Crime?
Light, Michael T. and Miller, Ty · 2018 · lightmiller2018
Source edition: Original 2018 article manuscript, Table 2, Model 5; Table 1 units. The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →
Comparing Crime Rates between Undocumented Immigrants, Legal Immigrants, and Native-Born {US} Citizens in Texas
Light, Michael T. and He, Jingying and Robey, Jason P. · 2020 · lightheobey2020
Bundle extraction record: full-text table
DOI →
Crime and Immigration: Evidence from Large Immigrant Waves
Bell, Brian and Fasani, Francesco and Machin, Stephen · 2013 · bellfasanimachin2013
Bundle extraction record: full-text table
DOI →
Do Immigrants Cause Crime?
Bianchi, Milo and Buonanno, Paolo and Pinotti, Paolo · 2010 · bianchibuonannopinotti2012
Source edition: February 2010 working-paper manuscript on a public mirror; Equation 3 and Table 5, column 3 (PDF page 33). The original model key is retained when the cited edition’s year differs.
Bundle extraction record: full-text table
Source →

Unidentified records · excluded from citations

These original bibliography records could not be identified. They are retained for inspection and are not used as evidence.

[
  {
    "key": "adermon2018",
    "author": "Adermon, Adrian and Lindahl, Mikael and Waldenstr{\\\"o}m, Daniel",
    "title": "Intergenerational Mobility in Pension Income: An Adoptee Perspective",
    "year": "2018",
    "evidence": "canonical"
  }
]

Source identity unresolved. This retained model record is not verified evidence.

[
  {
    "key": "lindo2018",
    "author": "Lindo, Jason M. and Schaller, Jessamyn and Hansen, Benjamin",
    "title": "Causes and Consequences of Child Maltreatment: Evidence from the Supplemental Nutrition Assistance Program",
    "year": "2018",
    "evidence": "canonical"
  }
]

Source identity unresolved. This retained model record is not verified evidence.

The pinned release restores six original estimates and adds seven corrected source contrasts. Three original earnings relationships remain unsupported. It has 64 available rows and ten retired identifiers. The archives retain the original records. Source contrasts do not establish generic displacement effects, headline totals or independent causal validation.

Parameters v1.54 · engine 0.4.1

Workers displaced?

Parent income · first year

Child lifetime earnings lost

Excess deaths · 20 years

Place baseline comparison

The effects of displacement

Family income

Child earnings

Health

Family income

Child earnings

Health

How does a father’s job loss affect his child’s future marriage?

Inside the calculation

Family income

Child earnings

Health

Earnings 7

Education 8

Health 11

Family 3

Assumptions and limits

Explore the earnings gap across generations

How the model works

Study inputs and model assumptions

Each study result has a permitted use

Later generations carry only the remaining earnings gap

Why the model does not add every paper together

Mortality adds risks across time

How the model checks complex uncertainty

Confidence intervals stay with their studies

Ranges and simulations are documented separately

Reproduce the reference scenario

Workers displaced