Conflicts
Linking a death to an industry is the step this project will be criticised for, so it is scored in the open: nine criteria, a published guide, a source for every score, and five scenarios that filter the same scores in five different ways.
Twelve wars have been scored so far, chosen to cover the range of the argument rather than to be easy — a war paid for by small-scale mining, one paid for by diamonds on one side and offshore oil on the other, a genocide with almost no economic side, and an oil state whose war has nothing to do with its oil. The records hold over a thousand wars and conflicts. That gap is shown here because a small sample presented as the whole picture is the easiest way to mislead with true numbers.
Scored wars
In order of the number of deaths the publisher recorded, not of anything this model calculated — ordering by a model result would suggest a ranking the model has not earned. A war with two periods has been scored twice, because its economy changed.
| War | Type | Deaths recorded | Scored periods | Industries | Scores |
|---|---|---|---|---|---|
| Government of Rwanda - Civilians
UCDP 472 |
against civilians | 794,375 | 1 | 3 | 19 |
| Russia - Ukraine
UCDP 13243 |
involving a government | 372,593 | 1 | 8 | 45 |
| Ethiopia: Government
UCDP 267 |
involving a government | 349,545 | 1 | 6 | 20 |
| Syria: Government
UCDP 299 |
involving a government | 287,894 | 2 | 8 | 50 |
| Afghanistan: Government
UCDP 333 |
involving a government | 286,253 | 1 | 6 | 33 |
| Sudan: Government
UCDP 309 |
involving a government | 78,223 | 1 | 7 | 36 |
| Iraq: Government
UCDP 259 |
involving a government | 69,447 | 1 | 7 | 29 |
| Sri Lanka (Ceylon): Eelam
UCDP 352 |
involving a government | 60,816 | 1 | 5 | 20 |
| Jalisco Cartel New Generation - Sinaloa Cartel
UCDP 13737 |
between armed groups | 30,574 | 1 | 5 | 25 |
| Angola: Government
UCDP 327 |
involving a government | 30,245 | 1 | 4 | 29 |
| DR Congo (Zaire): Government
UCDP 283 |
involving a government | 28,659 | 1 | 7 | 37 |
| Colombia: Government
UCDP 289 |
involving a government | 26,144 | 1 | 6 | 36 |
| Nigeria: Government
UCDP 297 |
involving a government | 14,618 | 1 | 6 | 31 |
| Burkina Faso: Government
UCDP 360 |
involving a government | 12,300 | 1 | 7 | 27 |
Death and event counts: Uppsala Conflict Data Program, Georeferenced Event Dataset v26.1, CC BY 4.0.
How much of the records this covers
14 conflicts are scored. The ledger holds 1,581.
The twelve were chosen to cover the range of the argument, not all the records: a war over resources, a rebellion paid for by small-scale mining, a collapsing state, a war fought through local allies, a war between countries, two wars with very little economic side, and one whose economy is a criminal market no classification includes. Between them they test every part of the scoring, including the parts that are meant to refuse.
Scoring nine criteria across several industries against evidence is research, not data entry, and it is the slowest part of this project. Until many more are scored, no total across wars is published — a total over twelve wars presented as a world figure would be wrong by about as much as it left out.
The five scenarios, and the three settings that set them apart
One set of scores, five filters. Every difference between the scenarios is in this table; no scenario has scores of its own.
| Scenario | Evidence required | Lowest confidence allowed | Industries included | Other passed on to products |
|---|---|---|---|---|
| Economic-embedded S-MAX |
no | LOW | every named industry | yes |
| Evidence-weighted S-EVID |
yes | LOW | every named industry | no |
| Resource-only S-RES |
yes | LOW | ENERGY, MINING, AGRICULTURE, LAND, WATER, FORESTRY | no |
| Arms and resources S-ARMS |
yes | LOW | ENERGY, MINING, AGRICULTURE, LAND, WATER, FORESTRY, ARMS, MILITARY_PROCUREMENT | no |
| Conservative S-CONS |
yes | HIGH | every named industry | no |
No scenario stretches the remaining shares. A share a scenario declines to pass on moves to Other and stays in the total, so a share backed by evidence keeps the same size in all five, and the gap between them is a published number rather than a hidden adjustment.
Questions
Why is “Other” so large for some conflicts?
Because for some wars it should be. Other holds the economic activity that no named industry accounts for, and it is scored like any industry rather than being whatever is left over. A separatist war over citizenship, or a genocide organised by officials and carried out where people lived, has little to do with where things are produced — and a model that could not say so would have to invent an economic story for exactly the cases that most obviously do not have one. Every war whose Other is above 30% has a written review saying whether it needs more scoring or is truly outside the named industries.
Does a high share mean the sector caused the war?
No. It means that, when harm is followed along supply chains, that much of the war’s human cost travels with that industry — a claim about accounting, not about motives. The main scenario says plainly that it measures the economic activity linked to a war, not the reasons anyone fights.
Who did the scoring?
Every score records the source it was read from, when it was entered and by whom, and any later change is written to a log that can only be added to — never edited or deleted. If you think a score is wrong, the scoring guide and the source are both on the war’s page; that is why they are published.
Why are only conflicts scored, and not workplaces or disease?
Harm at work is already reported by industry, so it only needs matching up, not judgement. Harm from poverty and pollution needs the same nine criteria applied to a different kind of subject, and that comes later. The scoring shown here is what all four kinds of harm will use.