Materials

Harm reaches a product through the materials it is made of. This is the list those materials come from, and an honest account of how far down it the model currently reaches.

The model shares an industry’s harm among its materials according to what that industry actually produced, country by country and year by year: crops and livestock (from the UN food agency, FAOSTAT), metals and minerals (from the US Geological Survey, USGS), oil, gas and coal (from the US Energy Information Administration, EIA), and water (from the UN’s AQUASTAT). An industry that produces no raw material — transport, trade, manufacturing, services — is never split into materials it did not produce. Under the main scenario its harm is shared out according to everything the country where it was counted produces, and that assumption is stated beside the figure; the other scenarios keep it in an unplaced share. Both shares are shown on this page, because a list that looks complete while much of it rests on an assumption would be the most misleading thing here.

How far the model reaches

99.8% of the attributed harm reaches one of 38 named materials. 0.2% is left unplaced.

72.4% of the harm linked to the economy reaches its material only through the production basket of the country where it was counted — that is, shared out by what that country produces: an assumption, graded LOW, not a measurement.

Under the main scenario, every death the model cannot place follows the country’s production basket — what the country produces, by value: a war nobody has scored, the part of a scored war the evidence could not place, harm in industries that make or serve rather than dig or grow (building, transport, manufacturing, trade), and weapons whose maker is unknown. Only a country with no recorded production in any year keeps its share unplaced. Water supply has production figures of its own (FAO AQUASTAT), so deaths from unsafe water reach water; but water is never part of this sharing-out: by weight it would swamp everything else, and it has no price to weigh by value.

The other scenarios make no such assumption. In them, three things stay unplaced: harm that could not be linked to an industry; harm in industries that make or serve rather than dig or grow, which this method does not follow into what they buy; and harm in industries that have a material but no production figures. Work injuries reported only as a national total are shared out by how many people work in each industry and how dangerous that work is. Metal ores have mine production for every year from 2010 to 2025 (USGS, each year from the newest edition that covers it); outside those years a country takes its own nearest year, marked low confidence.

Industries linked to a material: 16
Industries with a worked-out mix: 9
Materials that harm reaches: 38

Energy

Energy
Material	Unit	Producing countries	State
Crude petroleum	t	103	production loaded
Natural gas	m3	103	production loaded
Coal and lignite	t	82	production loaded
Uranium	t	—	no production data
Electricity	MWh	—	no production data

Metals and minerals

Metals and minerals
Material	Unit	Producing countries	State
Cobalt	t	17	production loaded
Copper	t	14	production loaded
Lithium	t	11	production loaded
Nickel	t	17	production loaded
Gold Conflict mineral (tin, tantalum, tungsten, gold)	t	21	production loaded
Tin Conflict mineral (tin, tantalum, tungsten, gold)	t	16	production loaded
Tantalum Conflict mineral (tin, tantalum, tungsten, gold)	t	13	production loaded
Tungsten Conflict mineral (tin, tantalum, tungsten, gold)	t	15	production loaded
Iron ore	t	18	production loaded
Bauxite	t	18	production loaded
Rare earth elements	t	13	production loaded
Silver	t	14	production loaded
Zinc	t	13	production loaded
Phosphate rock	t	55	production loaded
Manganese	t	15	production loaded
Natural graphite	t	23	production loaded
Salt	t	27	production loaded
Silica sand	t	36	production loaded
Natural diamonds (rough)	t	19	production loaded

Farming

Farming
Material	Unit	Producing countries	State
Wheat	t	124	production loaded
Milk	t	185	production loaded
Natural rubber	t	36	production loaded
Maize	t	171	production loaded
Rice	t	139	production loaded
Soybeans	t	117	production loaded
Sugar	t	156	production loaded
Coffee	t	80	production loaded
Cocoa	t	60	production loaded
Palm oil	t	50	production loaded
Cotton	t	98	production loaded
Livestock	t	193	production loaded
Timber	m3	170	production loaded
Fish and seafood	t	—	no production data
Tobacco (unmanufactured)	t	143	production loaded
Wood charcoal	t	191	production loaded

Land and water

Land and water
Material	Unit	Producing countries	State
Agricultural land	ha	210	production loaded
Forest	ha	210	production loaded
Mining concession	ha	—	no production data
Fresh water	m3	183	production loaded
Hydroelectric capacity	MW	—	no production data

Infrastructure

Infrastructure
Material	Unit	Producing countries	State
Port	unit	—	no production data
Pipeline	km	—	no production data
Railway	km	—	no production data
Shipping lane	km	—	no production data
Electricity grid	km	—	no production data
Telecommunications network	unit	—	no production data

Weapons

Weapons
Material	Unit	Producing countries	State
Small arms and light weapons	unit	—	no production data
Small-arms ammunition	t	—	no production data
Armoured fighting vehicles	unit	—	no production data
Military aircraft	unit	—	no production data
Naval vessels	unit	—	no production data
Bombs, missiles and warheads	t	—	no production data

Money and trade levers

Money and trade levers
Material	Unit	Producing countries	State
Sanctions	USD	—	no production data
Resource rents	USD	—	no production data
Taxation	USD	—	no production data
Trade control	USD	—	no production data
Smuggling	USD	—	no production data
State revenue	USD	—	no production data
Foreign direct investment	USD	—	no production data
Sovereign debt	USD	—	no production data
Market access	USD	—	no production data

Chemicals

Chemicals
Material	Unit	Producing countries	State
Nitrogen fertiliser (N)	t	—	no production data
Phosphate fertiliser (P2O5)	t	—	no production data
Potash fertiliser (K2O)	t	—	no production data

Questions

Why is the list longer than what has data?

Because the list is a choice about what the model should cover, and the data is a schedule of what has been loaded so far. Listing every material and showing which of them have data makes the gap visible; listing only those would make the model look complete, and every later addition look like a change of mind.

Why share an industry’s harm by what it produced, rather than by money?

Because the question is which physical thing the harm travels with. Money is the right measure for an economic footprint and the wrong one for following a death to a tonne of something. The model works out both a split by weight and a split by value, from the same production figures, and never adds them together. The figures use the split by value (tonnes times the price): by weight, gold mined beside iron ore would get almost nothing, and sugar cane, which is mostly water, would outweigh cocoa, coffee and meat. Compared year by year, the two splits are equally steady; by value a country’s harm lands on what its economy is actually made of.

Does a big share mean the material is dangerous to produce?

No. It means that much of a war’s harm travels with that material when harm is followed along a supply chain. The danger to the people who produce it is a separate kind of harm, with its own data, counted separately.

Where is the number of deaths per tonne?

It is calculated, but not published yet. It exists, tied to a particular calculation and version of the method, but it has no range yet and has not passed the test on sixteen past years. A project only gets to publish a wrong figure like that once.

Methodology → · Conflicts → · Sources →