settled-reach/docs/workshops/commodity-catalog/burnelli-sheldon-round1.md

Burnelli-Sheldon: Commodity Catalog Analysis for #801

I've reviewed D-171-D-183, my Round 2 model, the Round 1 cross-discussion, and the wiki corporation corpus. Here's my full economic analysis.

---

1. COMMODITY ENUMERATION FROM FIRST PRINCIPLES

I start from the question: what sectors does an interstellar civilization require? Then I enumerate what's produced in each sector, constrained by the wiki lore.

The answer is 32 commodity types: 24 physical goods in a 3-tier production hierarchy + 8 services.

Primary Sector -- Raw Materials (8 types)

#	Commodity	Description	Production geography
R1	Metallic ore	Iron, copper, structural metals. Bulk extraction from planetary/asteroid mining	Common -- most rocky bodies have some deposits
R2	Rare minerals	Exotic crystals, rare earths, gemstones. Low volume, high value	Concentrated -- 5-10 systems with significant deposits
R3	Timber/biomass	Harvested wood, structural/decorative grade. Forest worlds only	Regional -- requires established biosphere (Nordmark Skog's territory)
R4	Agricultural produce	Food crops, livestock, grain. Perishable, bulk	Ubiquitous -- any terraformed world produces some
R5	Water/volatiles	Water ice, atmospheric gases. Life support critical	Ubiquitous on ocean/ice worlds; life-or-death expensive on bare-rock stations
R6	Fusion fuel	Deuterium, He-3. Continuous consumption -- not a one-time purchase	Common -- gas giants and ice moons are primary sources
R7	Organic compounds	Biochemicals, pharmaceutical precursors, biological feedstock. Includes brach fiber (Braemar, D-177)	Regional -- requires complex biospheres
R8	Stone/aggregate	Construction stone, ceramics feedstock. Kvitfjell Blue marble is the premium variant	Common, but premium variants are monopolistic

Why 8 instead of D-173's "7": Stone and metallic ore follow fundamentally different production chains. Ore is smelted into metals; stone is cut/shaped for construction and decoration. Bifrost Marmor doesn't process their marble through a smelter. Collapsing them would force marble through a metallurgical chain that doesn't represent reality. D-173 said "approximately 30" -- 32 is within that margin.

Secondary Sector -- Intermediate Goods (8 types)

#	Commodity	Primary inputs	Description
I1	Refined metals	Metallic ore	Structural steel, copper wire, aluminum. Backbone of construction
I2	Advanced alloys	Metallic ore + rare minerals	High-performance materials for hulls, gate parts, precision machinery
I3	Structural panels	Timber + refined metals	Pre-fabricated building components (Nordmark Skog produces these)
I4	Processed food	Agricultural produce + water	Shelf-stable, transportable food products (Mercado distributes these)
I5	Electronics/components	Rare minerals + refined metals	Circuits, processors, control systems
I6	Chemicals/pharmaceuticals	Organic compounds + water	Medicines, industrial chemicals, life support consumables
I7	Drive cores	Rare minerals + advanced alloys	Propulsion components. High-value, specialized
I8	Textiles/composites	Organic compounds + chemicals	Clothing, technical fabrics, composite materials (brach fiber products)

Changes from Round 2: Replaced "hull plates" with "advanced alloys" (hull plates are a specific product from advanced alloys, not a separate intermediate category). Replaced "habitat modules" with "textiles/composites" -- habitat modules are a finished assembly, not an intermediate input. Textiles/composites IS an intermediate: it feeds into consumer goods, luxury goods, and habitat construction.

Secondary Sector -- Final Goods (8 types)

#	Commodity	Primary inputs	Description
F1	Heavy equipment	Refined metals + electronics + drive cores	Mining rigs, construction machinery (Stalownia's core product)
F2	Vehicles/craft	Advanced alloys + electronics + drive cores	Ships, ground vehicles (MVG territory)
F3	Consumer goods	Processed food + textiles + electronics	Daily necessities, household items, personal tech
F4	Luxury goods	Agri produce + textiles + chemicals	Premium products with terroir/brand identity: Calloway whisky, VGV wine, Bifrost marble, fine spirits
F5	Implant hardware	Electronics + chemicals + advanced alloys	Neural interfaces, augmentation, lattice components (Prometheus Labs)
F6	Gate components	Advanced alloys + drive cores + electronics	Gate maintenance and construction parts (Gate Corporation monopoly)
F7	Habitat modules	Structural panels + electronics + chemicals	Complete living/working units for stations and settlements (Cygni B station sections, 2-5y lead time per D-177)
F8	Freight haulers	Refined metals + drive cores + electronics	Commercial cargo vessels. Demand driven by trade volume, not population

Why "freight haulers" separate from "vehicles": Different demand functions. Freight hauler demand = f(trade_volume, fleet_age). Vehicle demand = f(population, wealth). They respond to different economic signals. A trade boom increases freight hauler demand without affecting personal vehicle demand. This distinction matters for the simulation's responsiveness to trade-flow changes.

Tertiary Sector -- Professional Services (5 types)

#	Service	Location-bound?	Notes
S1	Legal services	Yes	Contract enforcement, dispute resolution, Commission certification. Demand partly compliance-driven
S2	Financial services	Yes	Banking, credit, clearing, currency exchange. Infrastructure for all trade
S3	Medical/re-embodiment	Yes	Healthcare, neural backup, re-embodiment tech. Regulation varies by zone
S4	Insurance/risk	Yes	Cargo, liability, corporate risk management
S5	Information brokerage	Yes	Commercial intelligence, market data, confidential analysis (Adams & Ford, Ferreira Monteiro)

Quaternary Sector -- Luxury/Experiential Services (3 types)

#	Service	Location-bound?	Notes
X1	Tourism	Yes	Travel experiences, cultural tourism. Demand = f(cultural_significance, accessibility)
X2	Entertainment media	Yes	Holo-content, performances, cultural production
X3	Fine dining/hospitality	Yes	Restaurants, hotels, gastronomy (Talbreu territory)

Services are non-transportable through gates (D-173 confirmed). This is the key distinction. You cannot arbitrage legal services across systems. A system with excellent financial services has a structural competitive advantage that can't be shipped away. This creates geographic stickiness for economic activity -- corporations cluster near good service providers.

Total: 8 raw + 8 intermediate + 8 final + 5 professional services + 3 luxury services = 32 commodities.

---

2. BASE PRICING IN TRACTUS

Base prices represent equilibrium price at a typical producing node. I set these to create three properties:

• Raw materials cheap, finals expensive (accumulated cost principle)
• Processing margins realistic (2-4x for intermediates, 1.5-3x for finals)
• Absolute price spread wide enough that transport costs create meaningful differentials

Raw Materials

Commodity	Base (T/unit)	Rationale
Water/volatiles	2	Near-free on ocean worlds. The most abundant commodity in the universe
Agricultural produce	5	Common output from any terraformed world
Stone/aggregate	8	Heavy, cheap. Premium variants (marble) command 10-50x at source
Metallic ore	10	Bulk mining baseline. The reference commodity
Timber/biomass	15	Less common -- requires established forests
Fusion fuel	20	Critical utility. Price reflects continuous demand, not scarcity
Organic compounds	35	Complex biospheres only. Includes pharmaceutical-grade biologicals
Rare minerals	80	Scarce, location-specific. The premium raw material

Intermediate Goods

Commodity	Base (T/unit)	Input cost	Margin	Notes
Processed food	15	~8T	88%	Low-margin, high-volume. Mercado territory
Refined metals	30	~20T	50%	Bulk processing. Thin margin, huge volume
Textiles/composites	50	~32T	56%	Value-add from raw organics
Structural panels	45	~30T	50%	Timber + metals assembly. Nordmark Skog
Chemicals/pharma	60	~36T	67%	Higher value-add, specialized processing
Advanced alloys	100	~39T	156%	High-tech processing, significant value-add
Electronics/components	120	~49T	145%	Fabrication is extremely high value-add
Drive cores	200	~112T	79%	Precision assembly from expensive inputs

Final Goods

Commodity	Base (T/unit)	Input cost	Margin	Notes
Consumer goods	80	~41T	95%	Mass market, moderate margin
Habitat modules	150	~104T	44%	Heavy assembly, high input cost
Heavy equipment	250	~106T	136%	Complex, specialized (Stalownia)
Freight haulers	300	~135T	122%	Capital goods, durable
Vehicles/craft	350	~156T	124%	High-spec, precision (MVG)
Implant hardware	400	~86T	365%	Enormous value-add. Neural tech premium
Luxury goods	500	~46T	987%	Price is demand-driven, not cost-driven. Brand/terroir premium IS the product
Gate components	800	~248T	223%	Monopoly pricing (Gate Corporation). Precision + political leverage

Key observation on luxury goods: The 987% margin isn't a flaw -- it's the point. Calloway whisky isn't expensive because barley costs a lot. It's expensive because of 400 years of heritage, 12-20 year aging, and terroir that cannot be replicated. The input recipe captures physical requirements; the price captures brand economics. The corporate behavioral agent layer (Monopolist/Specialist archetypes) is what maintains these margins against competitive pressure.

Services

Service	Base (T/unit)	Notes
Entertainment media	30	High volume, low per-unit
Information brokerage	50	Data analysis and intelligence
Fine dining/hospitality	60	Cultural premium
Insurance/risk	75	Financial product
Tourism	100	Premium experience
Legal services	100	Compliance-driven demand
Financial services	120	Infrastructure premium
Medical/re-embodiment	200	Highest-value service. Re-embodiment is the premium tier

Service pricing varies dramatically by location. These are at-node base prices for a node with average service quality. Core systems with deep labor markets might price legal services at 80T; a frontier station with one overworked lawyer might price them at 300T. Service price variation is driven by labor supply, not transport costs.

---

3. ELASTICITY CLASSIFICATION

Five steady-state categories plus a panic mechanism for supply shocks.

Category 1: Perfectly Inelastic (epsilon approx 0.0-0.1)

"We need this to survive. Price doesn't matter."

• Water/volatiles (on stations without local source)
• Fusion fuel (continuous consumption -- stations cannot operate without power)
• Medical/re-embodiment (health services are non-discretionary)

A 20% supply drop: demand stays flat. Price spikes absorb the entire adjustment. These commodities create the most dramatic price events during shortages.

Category 2: Highly Inelastic (epsilon approx 0.2-0.4)

"We need this, but we can reduce waste or ration."

• Agricultural produce / processed food (people eat less meat, stretch supplies)
• Chemicals/pharmaceuticals (defer non-critical treatments, reduce industrial use)
• Legal services (compliance-driven -- Commission still requires certification)
• Financial services (banking is infrastructure -- can't opt out)
• Refined metals (basic construction continues, just slower)

A 20% supply drop: demand drops ~5-8%. Significant price increase, modest demand adjustment.

Category 3: Unit Elastic (epsilon approx 0.8-1.2)

"We adjust our plans proportionally."

• Metallic ore (mining expansion/contraction tracks price)
• Structural panels (construction delays when expensive)
• Electronics/components (upgrade cycles stretch)
• Consumer goods (people buy less, repair more)
• Habitat modules (expansion deferred, not cancelled)
• Insurance/risk (coverage levels adjusted)
• Information brokerage (discretionary intelligence purchases scale with budget)
• Textiles/composites (fashion cycles and replacement rates adjust)

A 20% supply drop: demand drops ~15-25%. Price and quantity adjust roughly proportionally.

Category 4: Elastic (epsilon approx 1.5-2.5)

"We can wait, substitute, or do without."

• Heavy equipment (capital expenditure -- defer the purchase, extend existing equipment life)
• Vehicles/craft (deferrable, repairable)
• Freight haulers (fleet expansion pauses; existing fleet continues)
• Advanced alloys (use standard metals at a quality penalty)
• Drive cores (overhaul existing cores instead of replacing)
• Entertainment media (pure discretionary)
• Fine dining/hospitality (eat at home)

A 20% supply drop: demand drops ~30-50%. Buyers actively defer, substitute, or exit the market.

Category 5: Highly Elastic / Luxury (epsilon approx 3.0+)

"Nice to have. If the price moves, we're out."

• Luxury goods (Calloway whisky, VGV Classe, Kvitfjell marble -- nobody needs these)
• Gate components (extremely long procurement cycles; orders deferrable by years)
• Implant hardware (upgrades, not life support -- existing implants keep working)
• Tourism (first thing cut in any downturn)
• Stone/aggregate (at non-source nodes -- local recycled materials substitute)

A 20% supply drop: demand drops 60%+. Market collapses to essential buyers only.

Panic Mechanism (not a category -- a conditional modifier)

Perverse elasticity isn't a steady-state property. It's a regime shift triggered by acute supply shocks. I model this as a stockpile-target mechanism:

• Each active node maintains a target stockpile (in weeks-of-consumption) for critical commodities
• When actual stockpile < panic threshold, the node generates excess demand proportional to the shortfall
• This creates a positive feedback loop: shortage -> hoarding demand -> deeper shortage -> price spike
• The loop is bounded by the stockpile target (demand normalizes once reserves rebuild)

Panic-flagged commodities and thresholds:

Commodity	Panic threshold	Behavior
Fusion fuel	< 2 weeks supply	Station managers build emergency reserves. Demand spikes 2-3x
Water/volatiles	< 1 week supply	Life-or-death hoarding. Demand spikes 3-5x
Processed food	< 2 weeks supply	Consumer panic buying. Demand spikes 1.5-2x
Chemicals/pharma	< 3 weeks supply	Hospital/industrial stockpiling. Demand spikes 1.5x
Rare minerals	N/A (price-triggered)	Speculative hoarding when price_trend > +15%/tick for 5+ ticks

Why this is better than a "perverse elasticity" category: It produces panic behavior from a sensible micro-foundation (rational stockpile management) rather than a magic coefficient. The panic is bounded, temporary, and self-resolving once supply normalizes. It also interacts correctly with the stockpile_weeks signal from D-181 -- the signal IS the mechanism.

---

4. PRODUCTION CHAIN RECIPES

Leontief fixed-coefficient recipes (D-173, D-178). Each row says: "to produce 1 unit of output, you need exactly these inputs." No substitution within a recipe. Substitution happens at the STRUCTURAL level -- multiple raw sources can feed the same intermediate through different recipes (D-173).

Raw -> Intermediate Recipes

Output (1 unit)	Input 1	Input 2	Notes
Refined metals	2.0 metallic ore	--	Simple smelting. High volume, low complexity
Advanced alloys	1.5 metallic ore	0.3 rare minerals	Rare minerals are the bottleneck input
Structural panels	1.0 timber	0.5 refined metals	Timber-dependent -- forest world proximity matters
Processed food	1.5 agri produce	0.2 water	Water is rarely the constraint (ubiquitous)
Electronics	0.5 rare minerals	0.3 refined metals	Rare minerals are the scarcity driver
Chemicals/pharma	1.0 organic compounds	0.3 water	Biosphere access is the real bottleneck
Drive cores	0.4 rare minerals	0.8 advanced alloys	Double dependency on rare minerals (direct + via alloys)
Textiles/composites	0.8 organic compounds	0.2 chemicals	Includes brach fiber path (organic compound variant)

Intermediate -> Final Recipes

Output (1 unit)	Input 1	Input 2	Input 3	Notes
Consumer goods	0.3 processed food	0.3 textiles	0.2 electronics	Broad input mix -- diversified
Habitat modules	1.5 structural panels	0.4 electronics	0.3 chemicals	Panel-heavy -- timber scarcity cascades here
Heavy equipment	1.0 refined metals	0.3 electronics	0.2 drive cores	Stalownia's recipe
Freight haulers	1.5 refined metals	0.5 drive cores	0.3 electronics	Metal and drive-core intensive
Vehicles/craft	0.8 advanced alloys	0.3 electronics	0.3 drive cores	Premium materials -- MVG uses alloys, not base metal
Implant hardware	0.5 electronics	0.3 chemicals	0.2 advanced alloys	Electronics-dependent -- rare mineral cascade
Luxury goods	0.3 agri produce	0.2 textiles	0.2 chemicals	Beverage-heavy aggregate (whisky, wine). Brand premium is in the PRICE, not the recipe
Gate components	1.0 advanced alloys	0.5 drive cores	0.4 electronics	The most input-intensive final. Triple rare-mineral dependency

Scarcity Cascade Analysis

The interesting question per D-178 (Leontief model) is: what happens when a single raw material becomes scarce?

Rare minerals shortage -> cascade:

• Direct: electronics down, advanced alloys down, drive cores down
• Downstream: implant hardware down significantly, gate components down severely, vehicles down significantly, heavy equipment down, freight haulers down
• The ENTIRE high-tech sector collapses. This is realistic -- rare earths are the bottleneck of real-world advanced manufacturing too.

Metallic ore shortage -> cascade:

• Direct: refined metals down, advanced alloys down
• Downstream: structural panels down, heavy equipment down, freight haulers down significantly, habitat modules down
• Construction and heavy industry halt. High-tech (electronics, implants) is less affected because it uses rare minerals, not bulk ore.

Timber shortage -> cascade:

• Direct: structural panels down
• Downstream: habitat modules down
• Narrow cascade -- but habitat module shortage means NO NEW STATIONS. Politically explosive.

Organic compounds shortage -> cascade:

• Direct: chemicals down, textiles down
• Downstream: luxury goods down, implant hardware down, consumer goods down, habitat modules down
• Broad but moderate cascade. Affects quality of life across the board.

Agricultural produce shortage -> cascade:

• Direct: processed food down, luxury goods down (beverages)
• Downstream: consumer goods down
• The food chain is SHORT -- 2 tiers max. But processed food is Category 2 (highly inelastic), so the price effect is enormous even if the quantity effect is small.

The coefficient ratios are designed to produce these cascades. Rare minerals appear as an input to 3 intermediates (electronics, advanced alloys, drive cores), making them the critical bottleneck commodity. This is intentional -- it creates a single-point-of-failure that the storyteller can exploit for dramatic supply shocks.

---

5. PRICE FORMATION SANITY

Transport cost model

Transport cost per hop = base_rate x bulk_modifier, where:

Bulk class	Modifier	Commodities
Liquid/gas	1.3x	Water, fusion fuel, chemicals
Bulk solid	1.2x	Ore, stone, timber, agri produce
Standard	1.0x	Refined metals, structural panels, processed food, textiles
Compact	0.8x	Electronics, drive cores, advanced alloys, consumer goods
Precision	0.6x	Rare minerals, implant hardware, luxury goods, gate components

Base gate rate: 5-12% per hop (D-178). Using 8% as the average, effective per-hop costs:

Bulk class	Effective cost/hop	Price after 5 hops	Price after 10 hops
Liquid/gas	10.4%	1.64x	2.69x
Bulk solid	9.6%	1.58x	2.50x
Standard	8.0%	1.47x	2.16x
Compact	6.4%	1.36x	1.86x
Precision	4.8%	1.26x	1.60x

Geographic price differential examples

Metallic ore (base 10T, bulk solid):

• At source: 10T
• 5 hops: 15.8T (+58%)
• 10 hops: 25T (+150%)
• This is meaningful. A system 10 hops from mining creates strong incentive for local extraction or recycling.

Electronics (base 120T, compact):

• At source: 120T
• 5 hops: 163T (+36%)
• 10 hops: 223T (+86%)
• Moderate gradient. Electronics are worth shipping long distances because they're compact and valuable.

Gate components (base 800T, precision):

• At source: 800T
• 5 hops: 1,008T (+26%)
• 10 hops: 1,280T (+60%)
• 15 hops: 1,625T (+103%)
• Even with precision bulk class (cheapest to ship), gate components still double over 15 hops. Frontier systems face enormous gate maintenance costs. This is a STRUCTURAL driver for frontier economic grievance and Compact formation.

Water (base 2T, liquid/gas):

• At source: 2T
• 10 hops: 5.4T (+170%)
• Low absolute differential, but stations consuming millions of units face massive aggregate costs. Water-poor stations far from ice sources have a permanent cost disadvantage.

Fusion fuel (base 20T, liquid/gas):

• At source: 20T
• 5 hops: 32.8T (+64%)
• 10 hops: 53.8T (+169%)
• Energy costs nearly TRIPLE over 10 hops. This means frontier systems far from gas giants have permanently elevated operating costs. Combined with Category 1 inelasticity (epsilon approx 0), fuel price is the single biggest structural cost for remote stations.

Sanity verdict

The price formation produces:

• 3:1 to 4:1 ratios for bulk commodities across the full diameter of the Reach (reasonable -- EVE-style but not extreme)
• 1.5:1 to 2:1 ratios for precision goods across the same distance (realistic -- high-value goods justify long-distance trade)
• Steep gradient for bulk, shallow for precision -- this correctly reproduces the real-world pattern where raw material transport costs dominate manufacturing location decisions
• Geographic specialization emerges naturally: Systems near raw sources process them. Systems near multiple intermediate sources do final assembly. Remote systems specialize in services. Nobody has to script this -- it falls out of the transport cost structure.

One concern: The 5-12% range in D-178 needs pinning. I recommend the range be a function of two variables:

• Route congestion (high-traffic routes are more efficient: 5-7%)
• Commodity bulk class (the modifier table above)

This means the same gate hop costs different amounts for different goods. Shipping ore through a busy core route costs ~6% x 1.2 = 7.2%. Shipping ore through a minor frontier gate costs ~11% x 1.2 = 13.2%. This is realistic -- logistics infrastructure concentrates on high-volume routes.

---

6. WHAT MY ROUND 2 MODEL WAS MISSING

A. Services (now included)

My Round 2 model had 22 physical commodities only. The lead correctly added professional and luxury services. Services are economically distinct because they're non-transportable -- you can't arbitrage a lawyer across gates. This creates geographic stickiness: corporations cluster near systems with good services. A system without them remains a resource extraction node. This is the "service economy" effect that drives real-world urbanization, now operating at interstellar scale.

Services also consume goods without producing them (D-173). A hospital consumes chemicals and electronics but doesn't produce a physical output. This makes service-heavy systems net importers, creating persistent trade deficits that must be balanced by service exports (people traveling TO the system for medical/legal/financial services). This is interesting trade dynamics.

B. Information as a commodity

Is raw data/intelligence a tradeable good? Mercado's procurement database, VGV's 350-year Comptoir archive, Adams & Ford's publications -- these are canonically valuable.

My recommendation: Model information as a SERVICE (information brokerage), not a physical commodity. Information doesn't have a bulk class or transport cost in the physical sense. It doesn't flow through production chains. It affects PRICES (better information -> better trading decisions) rather than being a link in a production chain. The corporate behavioral agent layer handles this -- an Intermediary archetype with intelligence assets commands better pricing than one without.

C. Genetic material / biological IP

Re-embodiment technology implies tradeable biological templates. Recommendation: Fold into medical/re-embodiment services. The physical substrate (if any) is a variant of organic compounds. The service IS the re-embodiment process, not the material.

D. Cultural IP / brand value

VGV's classification system, Calloway's 400-year heritage. These are economically real but intangible. Recommendation: Handle through the corporate behavioral agent layer. A Monopolist/Specialist archetype commands premium pricing for the same physical commodity. The brand premium shows up in the PRICE, not the commodity catalog. This is why luxury goods have a 987% margin over input cost -- the margin IS the brand.

E. Salvage/recycling

In a centuries-old civilization, recycled materials are a real input. Recommendation: Don't add a separate commodity. Instead, model salvage as a supply-side modifier on metallic ore and refined metals at established systems. Old stations and decommissioned ships feed local recycling, effectively reducing the cost of refined metals at systems with long settlement history. This is a productivity seed dimension -- old systems have cheaper access to metals, new ones must import.

F. Contraband

D-174 lists canonical contraband: unlicensed lattice components, Sol-denominated luxury goods, unregistered commercial intelligence, unlicensed re-embodiment. These are NOT separate commodities. They're shadow-economy variants of existing commodities with different pricing and legality. The shadow_economy_intensity parameter (D-174) handles the pricing differential. The commodity catalog doesn't need a "contraband" entry.

G. Production ubiquity flag (NEW)

One thing missing from both my Round 2 model and the current D-173 record: a production ubiquity classification per commodity. This determines the geographic price gradient:

Ubiquity	% of inhabited systems producing	Effect on price geography
Ubiquitous	50%+	Small price variation (water, agri produce)
Common	20-50%	Moderate variation (ore, stone, timber, fusion fuel)
Regional	5-20%	Significant variation (organic compounds, most intermediates)
Concentrated	1-5%	Large variation (rare minerals, advanced alloys, electronics)
Monopolistic	<1% (1-3 systems)	Extreme variation (gate components, terroir-locked luxury goods)

This flag is essential for the simulation to correctly distribute production across the map during initialization. Without it, the generate_corporations binary has no guidance on how to spread production. I recommend adding this to the commodity record schema.

H. Continuous consumption flag (NEW)

Fusion fuel and water aren't purchased once -- they're consumed continuously. Every station and settlement has a base demand that exists regardless of economic conditions. Recommendation: Add a demand_model field to the commodity record:

• market -- demand driven by economic activity and consumer preference (default)
• utility -- demand driven by population and infrastructure (fusion fuel, water)
• compliance -- demand driven by regulatory requirements (legal services, certification)

This affects how the tatonnement adjusts: utility commodities have demand floors that never drop to zero, even in a deep recession.

---

SUMMARY TABLE

Category	Count	Round 2	Delta
Raw materials	8	7	+1 (stone/aggregate)
Intermediate goods	8	8	0 (but 2 replaced)
Final goods	8	7	+1 (habitat modules promoted from intermediate)
Professional services	5	0	+5 (new)
Luxury services	3	0	+3 (new)
Total	32	22	+10

New record fields I recommend adding to D-173 / the TOML schema:

1. production_ubiquity: ubiquitous / common / regional / concentrated / monopolistic
2. bulk_class: liquid_gas / bulk_solid / standard / compact / precision
3. demand_model: market / utility / compliance
4. panic_flagged: bool + panic_threshold_weeks (for stockpile-target mechanism)
5. political_sensitivity: bool (flag for shadow economy / currency zone modifiers)

These 5 fields plus the existing D-173 fields (tier, elasticity_class, base_price, bulk_class, political_sensitivity) give us a complete commodity record.

Waiting for team feedback. Happy to defend any of these numbers or revise coefficients if the cascade analysis reveals unwanted dynamics.