Hundreds of topics. One connected model. No hidden errors.
Most social value methods value one topic at a time - a fixed figure per outcome, borrowed from population averages, then added together.
What everyone else does
| Approach | How It Works | What It Misses |
|---|---|---|
| WELLBY methods | Survey data compared to population baselines | One topic at a time; no person-level paths |
| SROI | Borrowed substitute values from value banks | No comparison with what would have happened anyway |
| HACT Social Value Bank | Multiply counts by fixed per-outcome values | Single-topic, population-average, transfer-based |
| Hedonic pricing | Infer value from property prices | No person-level welfare model; confounds many effects |
| Multi-model analysis | Different teams model topics separately | Results stapled together; conservation not enforced |