Food Security

Causal loop diagrams live or die on where you draw the boundary. Too wide and everything connects to everything; too narrow and you miss the loop that matters. We set the boundary at the global staple-crop production-and-trade system: soil, inputs (fertilizer, water, seed), yields, prices, farmer livelihoods, trade flows, and the policy reflexes (subsidies, export bans, reserves) that react to them.

We exclude (for this pass): end-consumer nutrition behaviour, food waste at retail/household, and the deep ecology of specific agroecosystems. Each deserves its own diagram.

We also name the tension we expect to find: the things that raise yields this season often lower the productive base over decades. That intuition will become a loop.

Try it yourself

Before drawing a CLD, fix the system boundary in one sentence: what's
in, what's out. Then write the one tension you already suspect is at
the system's core — it usually becomes the central loop.

---

H1 — The dominant system (and its strain)

Input-intensive, yield-maximising, globally traded monoculture. What it's built on:

• Synthetic fertilizer (nitrogen via Haber-Bosch, phosphate, potash) as the productivity engine.
• Long, efficient, low-redundancy global supply chains (grain from a few breadbaskets; fertilizer through a few chokepoints).
• Yield-per-hectare as the master metric.
• Subsidy regimes oriented to production volume.

Signs it's fading:

• Soil degradation eroding the base: ~30% of agricultural land degraded; 1.7B people in declining-yield areas. (2)
• Fertilizer-supply fragility: ~30% of traded fertilizer transits the Strait of Hormuz; 2022 (Ukraine) and 2026 (Iran-related) shocks both spiked prices. (3)
• Yield-growth plateauing for major cereals in some regions.
• Price volatility transmitting shocks to import-dependent (esp. African) economies.

H3 — The emerging system (fringe signals)

Diversified, soil-regenerating, shorter-and-redundant, less input-dependent:

• Regenerative / agroecological practices — cover cropping, reduced tillage, compost/biochar, integrated livestock — shown to maintain yields while cutting input costs and rebuilding soil organic matter. (4)
• Diversified and regionalised supply chains with deliberate redundancy.
• Soil-as-capital metrics (soil organic carbon) entering farm accounting and carbon markets.
• Dietary shifts and alternative proteins reducing pressure on land and feed-grain.
• Bio-based inputs partially substituting synthetic fertilizer.

H2 — The contested middle

• Precision agriculture (variable-rate inputs, sensing, AI agronomy). H2− mostly — it optimises the industrial system, extending H1's life, though it can reduce input intensity.
• Vertical / controlled-environment agriculture. H2−/H2+ — useful for some high-value crops; energy economics keep it niche for staples.
• Carbon-farming markets. H2+ if they genuinely fund soil rebuilding; H2− if they become offset-greenwashing that leaves practice unchanged.
• Gene-edited climate-resilient cultivars. H2−/H2+ — can entrench input dependence or enable lower-input resilience, depending on deployment.
• Alternative proteins. H2+ structurally (reduces feed-grain land) but contested on cost, taste, and ultra-processing concerns.

The Three Horizons discipline: label each H2 item H2− (captured by H1) or H2+ (opens H3), and notice that the same technology can be either depending on the system it's embedded in. Precision ag in a yield-max paradigm is H2−; the same tools in a soil-as-capital paradigm are H2+.

Try it yourself

Map the food system (or your sub-system) in Three Horizons.
- H1: 4–5 pillars of the dominant system + signs of strain
- H3: 4–5 fringe practices pointing at a different system
- H2: 4–5 contested innovations, each labelled H2− or H2+
For each H2 item, note what would flip it from − to + (usually it's
the surrounding paradigm, not the tech).

---

Now we open the hood. A CLD names variables, connects them with signed arrows (+ means they move together; − means they move oppositely), and identifies closed loops as reinforcing (R) — vicious or virtuous spirals — or balancing (B) — self-correcting. (5)

Variables

1. Synthetic fertilizer use
2. Short-term crop yield
3. Soil health (organic matter, structure, biology)
4. Food availability
5. Food price
6. Farmer profitability
7. Investment in soil-building practices
8. Land degradation
9. Land abandonment
10. Export restrictions (policy reflex)
11. Global tradable supply
12. Supply-chain concentration / fragility

Signed edges (selected)

• Synthetic fertilizer use +→ short-term yield
• Synthetic fertilizer use +→ soil degradation (over time; masks declining fertility)
• Soil degradation −→ soil health
• Soil health +→ short-term yield (and yield resilience)
• Soil degradation +→ fertilizer demand (need more to get the same yield)
• Short-term yield +→ food availability
• Food availability −→ food price
• Food price +→ farmer profitability
• Farmer profitability +→ investment in soil-building
• Investment in soil-building +→ soil health
• Food price +→ export restrictions (spikes trigger bans)
• Export restrictions −→ global tradable supply
• Global tradable supply −→ food price (less supply, higher price)
• Supply-chain concentration +→ fragility (shock transmission)

Named loops

R1 — The Fertilizer Treadmill (reinforcing, vicious). Fertilizer use → short-term yield (good) — but fertilizer use → soil degradation → lower baseline fertility → more fertilizer needed to hit the same yield → more degradation. The yield signal masks the erosion of the base. This is the system's core trap: it looks like it's working right up until it isn't.

R2 — Panic Hoarding (reinforcing, vicious). A price spike → export restrictions (each country protects its own) → less global tradable supply → higher price → more restrictions. 2008, 2011, and 2022 all showed this loop firing. It converts a regional shock into a global crisis through rational individual behaviour.

B1 — Price–Demand Balance (balancing). Higher price → reduced consumption / substitution → lower price. The textbook stabiliser — but slow, and brutal for the poor, who "balance" by eating less.

B2 — Degradation–Abandonment (balancing, grim). Soil degradation → land abandonment (~3.6M ha/yr) → less production on degraded land. A balancing loop that "solves" degradation by removing land from the system — at catastrophic human cost.

R3 — The Regenerative Spiral (reinforcing, virtuous — currently weak). Investment in soil-building → soil health → yield resilience + lower input costs → farmer profitability → more investment in soil-building. This is the loop H3 needs to strengthen. Today it's dominated by R1 because input subsidies and yield-only metrics starve it.

Reading the diagram

The system is held in its fragile shape by R1 (fertilizer treadmill) and R2 (panic hoarding) — two vicious reinforcing loops — while the loop that could deliver resilience, R3 (regenerative spiral), is weak. The balancing loops (B1, B2) "stabilise" the system through human suffering (eating less; abandoning land). That's the finding: the system is stable but in a bad way, and its self-correction mechanisms are cruel.

Try it yourself

Build a CLD for your system.
1. List 8–12 variables.
2. Connect with signed edges (+ move together, − move oppositely).
3. Find closed loops; label each R (reinforcing) or B (balancing) and
   give it a one-sentence story and an evocative name.
4. Identify which loops dominate (hold the current shape) and which
   are weak (the ones a transition would need to strengthen).
Output as: variables / signed edges / named loops. Then write the
"reading" — what holds the shape in place?

---

Donella Meadows ranked places to intervene in a system from weakest to strongest. (5) We locate food-system interventions on her ladder.

Weak leverage (parameters / subsidies):

• Adjusting fertilizer subsidies, buffer-stock sizes, tariff levels. Politically popular, low system-change. (Most food policy lives here.)

Medium leverage (feedback & information flows):

• Strengthening the balancing feedback that doesn't run through suffering: strategic grain reserves, the G20 AMIS (Agricultural Market Information System) for transparent stock data to defuse R2 panic-hoarding. Information transparency is a classic Meadows lever — it weakens the panic loop by removing the uncertainty that drives hoarding.
• Anti-export-ban coordination agreements (rules that dampen R2).

High leverage (rules, goals, paradigm):

• Changing the master metric from yield-per-hectare to something like yield-resilience-per-hectare or soil-organic-carbon-adjusted productivity. This re-weights R1 vs R3 directly.
• Repricing soil as capital so that degrading it shows up on a balance sheet — turning R3's "investment in soil-building" into a financially rational act.
• The paradigm shift: from food-as-commodity (optimise price/volume) to food-as-security (optimise resilience/redundancy). This is the highest-leverage and slowest-moving intervention; it's where the Three Horizons H1→H3 transition actually happens.

The CLD shows why low-leverage interventions disappoint: tweaking subsidies inside R1 doesn't weaken R1. You have to change the metric or the paradigm to shift which loops dominate.

Try it yourself

Place 5–7 candidate interventions on Meadows's leverage ladder
(parameters → feedback/info → rules → goals → paradigm). For each,
state which loop in your CLD it acts on, and whether it weakens a
vicious loop or strengthens a virtuous one. Notice how many of the
"obvious" interventions sit at the weak end.

---

1. What did the CLD show that Three Horizons alone didn't? — That H3 (regenerative) isn't failing to spread because it's a worse idea; it's failing because R1's subsidies and metrics starve R3. The transition is a loop-dominance problem.
2. Where is the diagram too clean? — It treats "farmers" as one variable. Smallholder vs industrial, irrigated vs rainfed, Global North vs South behave completely differently. A real CLD would disaggregate.
3. What loop did you leave out? — The political-economy loop: input manufacturers and commodity traders benefit from R1 and lobby to keep subsidies and metrics where they are. That loop is why the high-leverage interventions are politically hard.
4. What 2026 action does this surface? — Strengthen AMIS-style transparency (defuses R2 cheaply); pilot soil-carbon metrics in farm accounting (strengthens R3); resist the reflex to "intensify" (deepens R1).
5. What does this refuse? — To predict famines or prices. To pretend the paradigm shift is easy. To substitute for agronomic and regional expertise.

Try it yourself

Reflect in <60 words each: what did the CLD show that the scan didn't;
where is the diagram too clean; what loop (esp. political-economy) did
I leave out; what 2026 action follows; what does this refuse to do?

---

To make this felt, the Design side would pick a leverage point and prototype it. For instance: a 2035 farm balance sheet that books soil organic carbon as a depreciating/appreciating asset — a Design Fiction artifact that embodies the "soil-as-capital" paradigm shift, and lets people argue about whether financialising soil helps or harms. That would seed a Hexagon 2 walkthrough (see Topic 5 and Topic 13 for the Design Fiction form).

---