AQALY Framework Application: Carpenter & Kanepajs, 2026
This page applies the AQALY-4PD-9L framework (the Animal Quality-Adjusted Life Year with 4 Physical Domains assessed on a 9-Level welfare scale; Espinosa, Browning & Treich, 2025) to quantify and monetise the sow welfare difference between farrowing crates and free-farrowing systems.
Marie conducted two complementary assessments of sow welfare under farrowing crate and free-farrowing (Swedish model, 6.5 m²) systems, using the Five Domains framework (Mellor, 2016) and EURCAW/Baxter et al. (2011) welfare indicators.
Welfare Compromise: each indicator scored on Severity, Intensity, Duration, and Confidence (0–4), combined into a composite grade: A (none, 0–0.20), B (low, 0.21–0.40), C (mild-moderate, 0.41–0.60), D (marked-severe, 0.61–0.80), E (very severe, 0.81–1.00). Welfare Enhancements: opportunities for positive experiences scored 0 (none) to +++ (high-level).
| Indicator | Crate | Free farrowing | ||
|---|---|---|---|---|
| Welfare Compromise |
Welfare Enhancements |
Welfare Compromise |
Welfare Enhancements |
|
| Environment | ||||
| Space and freedom to move | E | 0 | A | +++ |
| Comfortable climatic conditions | D | 0 | A | ++ |
| Separate pen into functional areas / elimination behaviour | E | 0 | B | ++ |
| Health | ||||
| Injuries / Skin lesions | D | 0 | A | ++ |
| Behaviour | ||||
| Nest-building and explorative behaviour | D | + | C | ++ |
| Interaction between sow and piglets | D | 0 | A | +++ |
| Freedom from sow at weaning | D | 0 | A | +++ |
Marie's Impact Assessment scores (Table 1) use a Severity/Intensity/Duration scale (0–4 each) that produces composite grades A–E, plus Enhancement ratings (0 to +++). To translate these into the AQALY 1–9 welfare scale, each subdomain grade is computed as: grade = 5 + compromise level − enhancement level, where compromise maps A=0, B=1, C=2, D=3, E=4 and enhancement maps 0=0, +=1, ++=2, +++=3. A grade of 1 (best) means no compromise and high enhancement; 9 (worst) means very severe compromise and no enhancement. Where no Impact Assessment score exists, Marie's Five Domains numeric scores (−2 to +2) are converted as grade = 5 − 2 × score, or qualitative descriptions are translated to the 1–9 scale by expert judgment.
Sources: Impact Assessment workbook, Domains_Sows_Piglets workbook
Evidence tiers: Tier 1 = directly scored in the Impact Assessment; Tier 2 = scored by Marie outside the IA; Tier 3 = inferred from system assumptions (description shown in italics).
| Subdomain | Crate | Free farrowing | Tier | Rationale |
|---|---|---|---|---|
| Water/food intake | 1 | 1 | 2 | Marie +2: easy access, no competition |
| Foraging material | 7 | 7 | 2 | Marie −1: minimal provision |
| Fibre diet | 9 | 9 | 2 | Marie −2: liquid feed, lack of fibre |
| Average |
| Subdomain | Crate | Free farrowing | Tier | Rationale |
|---|---|---|---|---|
| Thermal conditions | 8 | 3 | 1 | IA Climate: C(3/3/3) D/0 to A/++ |
| Substrate | 7 | straw bedding, 3/4 solid floor → 3 | 3 | Crate: Marie −1. Free: Assumptions sheet |
| Confinement | 9 | 2 | 1 | IA Space: C(4/4/4) E/0 to A/+++ |
| Air/noise/light | 7 | can regulate exposure via movement → 6 | 3 | Crate: Marie −1. Free: marginal improvement |
| Environmental monotony | 8 | 4 | 1 | IA Functional areas: C(3/3/4) to B/++ |
| Average |
| Subdomain | Crate | Free farrowing | Tier | Rationale |
|---|---|---|---|---|
| Disease/injury | 8 | 3 | 1 | IA Injuries: D/0 to A/++ |
| Body condition/fitness | 7 | movement maintains muscle, bone strength → 4 | 3 | Crate: Marie −1. Free: inferred from movement |
| Average |
| Subdomain | Crate | Free farrowing | Tier | Rationale |
|---|---|---|---|---|
| Locomotory | 9 | freer movement, pen-limited (6.5 m²) → 3 | 1 | IA Space + description |
| Foraging | 7 | manipulate enrichment/foraging material → 4 | 3 | Crate: Marie −1. Free: description |
| Nest-building | 8 | 4 | 1 | IA Nest: D/+ to C/++ |
| Maternal behaviour | 8 | 2 | 1 | IA Interaction: D/0 to A/+++ |
| Social behaviour | 7 | 7 | 2 | Both: individual pens. Marie −1 |
| Escape behaviour | 8 | avoidance can be shown, distance limited → 3 | 1 | IA Separation + description |
| Resting behaviour | 7 | posture changes + straw bedding → 3 | 3 | Crate: Marie −1. Free: description |
| Comfort behaviour | 7 | space + substrate enable grooming → 4 | 3 | Crate: Marie −1. Free: description |
| Average |
The AQALY framework assesses animal welfare across four physical domains on a 1–9 scale. Lower grades indicate better welfare. Domain grades are continuous averages of their subdomain scores.
| Grade | Welfare level | Interpretation |
|---|---|---|
| 1 | Best possible | Reference state; no welfare decrement in this domain |
| 2 | Significant positive | Very good conditions; minor sub-optimality only |
| 3 | Mild positive | Good conditions |
| 4 | Minor positive | Above-neutral; basic positive provisions present |
| 5 | Neutral | Neither positive nor negative welfare in this domain |
| 6 | Minor negative | Below-neutral; some compromise of needs |
| 7 | Mild negative | Clear welfare compromise; chronic mild suffering |
| 8 | Severe negative | Significant suffering; multiple needs unmet |
| 9 | Very severe negative | Worst state; extreme and sustained suffering |
Grades 3, 5, 7, 9 measured directly via DCE-TTO (Espinosa et al. 2025). Grades 2, 4, 6, 8 interpolated via PCHIP.
Each domain grade maps to a coefficient from Table A2 (Espinosa et al. 2025, merged sample N=1,062, PCHIP interpolation). For fractional grades, coefficients are linearly interpolated between the two nearest integer-grade values.
Coefficients are welfare decrements estimated via discrete choice experiments and time-trade-off tasks. Grade 1 = 0 (reference); grade 9 = most severe. Higher absolute values mean worse welfare in that domain.
AQALY per year = 1 + βNutrition + βEnvironment + βHealth + βBehaviour
Welfare gains monetised via the human QALY threshold (EUR 147,000), scaled by a moral weight representing how much a pig-year matters relative to a human-year.
Consumer WTP (Clark et al. 2017): meta-analytic consumer willingness-to-pay for pig welfare improvements = EUR 0.54/kg pork, from 13 studies and 90 WTP measures. BCR = 0.54 ÷ 0.0071 = 76.1×, applied uniformly to the EUR 0.22/SCD economic cost.
Welfare benefit alongside economic cost to the producer, across three valuation approaches.
| Valuation method | Per sow-crate-day | Per kg of pork | BCR |
|---|---|---|---|
| Consumer WTP (Clark et al. 2017, EUR 0.54/kg) | 0.22 → 16.73 | 0.0071 → 0.54 | 54.0× |
| Direct: Neuron count (1/44) | 0.22 → | 0.0071 → | |
| Direct: Hedonic capacity (0.515) | 0.22 → | 0.0071 → |
All values in EUR. Cost = economic cost to producer. Benefit = monetised welfare gain. BCR = benefit-cost ratio (benefit ÷ cost).
The Clark et al. (2017) meta-analytic WTP of EUR 0.54/kg used in the first toggle can be cross-checked against other estimates and proposals:
The Animal Quality-Adjusted Life Year (AQALY) framework was developed by Espinosa, Browning, and Treich (2025). It adapts the human QALY concept for animal welfare valuation using a discrete choice experiment (DCE) with time trade-off (TTO) methodology.
The 4PD-9L variant used here assesses four physical domains (Nutrition, Environment, Health, Behaviour) on a nine-grade scale. Coefficients were estimated from a merged sample of N=1,062 respondents.
Subdomain layer. Domain grades in this assessment derive from 18 subdomains, built from Marie's Five Domains assessment of sow welfare in crate and free-farrowing (Swedish model, 6.5 m²) systems. Each subdomain is scored on the same 1–9 scale, and the domain grade is the arithmetic mean of its subdomains. This produces continuous (fractional) domain grades.
Evidence tiers. Subdomain scores draw on three tiers of evidence:
Coefficient interpolation. The AQALY coefficient array provides values at integer grades 1–9. For fractional domain grades, coefficients are computed via linear interpolation between the two nearest integer-grade coefficients. For example, a grade of 7.50 uses the midpoint between the grade-7 and grade-8 coefficients.
Formula: AQALY/year = 1 + βNutrition + βEnvironment + βHealth + βBehaviour
The resulting AQALY difference is pro-rated to a 28-day lactation block and monetised at the EU QALY threshold (EUR 147,000), scaled by a moral weight.
Consumer WTP (Clark et al. 2017, EUR 16.73/SCD): Meta-analytic consumer willingness-to-pay for pig welfare improvements. Clark et al. (2017) synthesised 54 studies (335 WTP estimates, 17 countries). The pig-specific weighted mean is EUR 0.54/kg (13 studies, 90 measures, I² = 98.33%; GRADE: LOW). BCR = EUR 0.54 ÷ EUR 0.0071 per-kg cost = 76.1×; applied to EUR 0.22/SCD cost gives EUR 16.73/SCD. Methodologically distinct from the AQALY framework: a top-down stated-preference approach rather than bottom-up welfare science.
Neuron count (1/44 = 0.0227): The ratio of cortical neurons between pig (~2.2 billion) and human (~86 billion, of which cortical ~16 billion) brains.
Hedonic capacity (0.515): The Rethink Priorities welfare range estimate, which aggregates across multiple theories of consciousness and hedonic capacity. Substantially higher, reflecting that many ethical frameworks assign pigs considerably more than neuron-proportional moral weight.
The economic cost of EUR 0.22 per sow-crate-day comes from the EUFF economic module, which estimates the net producer cost of transitioning from farrowing crates to free-farrowing systems. This figure integrates capital costs (pen conversion), operating cost changes (labour, feed, mortality), and productivity effects across the EU-27 herd.
EUR 18 per sow per year and EUR 0.0071 per kg of pork are derived from the same module using standard denominators (2.3 litters/year, 28.2 marketed pigs/sow at 90.4 kg HCW = ~2,550 kg/sow/year).