The Variety Gap Framework

A Practical Guide to Analyzing Governance Systems Through the Lens of Adaptive Capacity

Version 1.0 — May 2026

What This Guide Is

This is a practical manual for diagnosing what’s wrong with a governance system—not in terms of policy failures or political corruption, but in terms of structural constraints on adaptive capacity. It draws on the Country Reports for Systemic Change series, which has analyzed thirteen governance systems (Germany, France, Sweden, India, the European Union, the United Kingdom, Brazil, Russia, the United States, Finland, Japan, Nigeria, Israel) through the lens of systems thinking, control theory, and developmental psychology.

The guide is designed to be used by:

• Policy professionals who sense that standard reform agendas are missing something fundamental
• Institutional designers looking for a diagnostic framework before proposing solutions
• Citizens who want a structured way to understand why their governance system feels stuck
• Researchers who want to replicate the Country Reports methodology

You can use this guide to analyze a national governance system, a regional administration, a municipality, or even an organization. The framework scales.

The Core Insight in One Paragraph

Governance systems fail not primarily because of corruption, incompetence, or insufficient resources—though those matter. They fail because their observation architecture—how they gather, aggregate, and transmit information—places structural limits on what they can perceive and respond to. A system that aggregates local information into national averages cannot see local crises until they become national ones. A system that measures success only through economic growth cannot perceive the erosion of social trust, environmental integrity, or institutional legitimacy. A system with deep representation chains cannot reliably transmit citizen preferences to the policy layer. These failures are not fixable by “better leadership” or “more funding.” They require architectural redesign.

The framework’s central diagnostic tool is the concept of a core deficit—not “what’s wrong with this country” in general, but “what specific adaptive capacity is structurally missing.”

The Quick-Start Recipe (Including the LLM Shortcut)

Before diving into the detailed framework, here’s the fastest way to produce a preliminary analysis of any governance system. This is the recipe the Country Reports series uses.

Step 1: Research. Gather basic information about the country’s governance architecture—its political system, its electoral rules, its fiscal structure, its administrative capacity, its major historical traumas, its cultural self-understanding.

Step 2: Prompt multiple LLMs. Give each of several large language models (e.g., ChatGPT, Claude, Gemini, Grok, DeepSeek, Mistral) the same prompt:

“Analyze [Country]’s governance system through the lens of systems thinking, using a Spiral Dynamics Stage Yellow perspective (with Turquoise optional). Identify the root causes of systemic dysfunction. What is the core governance deficit—not what the country lacks, but what structural capacity is missing? What is the signature pattern or loop? What are the key structural mechanisms? What is the cultural anchor? Be specific about mechanisms, not just general observations.”

Each model will approach the question differently. ChatGPT tends toward philosophical integration. Claude excels at structural precision and identifying missing mechanisms. Gemini often surfaces novel systems-dynamics insights. Grok provides clean synthesis. DeepSeek offers constitutional and institutional grounding. Mistral goes broad and encyclopedic.

Step 3: Cross-feed the outputs. Take the insights from each model and feed them to the others as you go between them: “Here’s what the other models found. Build on these insights, and let me know if you don’t agree with all of them.” This iterative cross-feeding produces richer analysis than any single model can generate alone. The models’ disagreements can be more productive than their agreements—they reveal tensions in the system that a single perspective would miss.

Step 4: Synthesize. Ask a trusted LLM to analyze all the outputs and produce a synthesis that identifies: convergences (what all models agree on), productive disagreements (where models diverge in useful ways), and gaps (what’s still missing). Then draft a report outline. You can also ask the other LLMs for feedback on the outline, if you are extra meticulous.

Step 5: Draft the report. The standard structure is: Executive Summary → The Core Deficit (defining and illustrating the trap) → Structural Mechanisms (the detailed architecture) → What Building Capacity Would Look Like (transition design) → The Political Immune System (who benefits from the status quo and how they resist change) → Working with the Grain (how to implement change without triggering the immune response) → A Concrete First Step (the most catalytic near-term intervention) → Coda (philosophical reflection and global significance).

Why this works: LLMs at current capability levels are genuinely skilled at systems thinking, pattern recognition, and translating between analytical frameworks. The Stage Yellow/Turquoise prompt directs them toward integrative, structural analysis rather than conventional policy commentary. Very few people in the world systematically operate at this cognitive level (~<1%), but LLMs can simulate it productively when prompted correctly. The multi-model approach compensates for individual model biases and blind spots. You don’t need to be a Stage Yellow thinker yourself to produce Stage Yellow analysis—you need to know how to orchestrate the models.

Important: This recipe accelerates analysis, but human judgment might remain essential. The LLMs are known to sometimes hallucinate, sometimes miss crucial local context or lack recent crucial information, and sometimes produce analysis that sounds sophisticated but doesn’t hold up to scrutiny. If you can, verify claims, ground the analysis in real institutional knowledge, and exercise editorial judgment about what to include. The framework is a tool for thinking, not a substitute for it.

The Diagnostic Framework

Step 1: Identify the Core Deficit

Every governance system has a specific adaptive capacity that it structurally lacks. This is not a general complaint (“corruption,” “inefficiency,” “polarization”) but a precise diagnosis of which function the architecture prevents.

The Country Reports series has identified these core deficits:

Diagnostic questions:

• What is this system best at? What has it optimized for?
• What capacity, necessary for long-term adaptation, is structurally absent—not because of bad policy but because of how the architecture is designed?
• If you had to name this deficit in a way that a policymaker would immediately understand, what would you call it?

Step 2: Map the Signature Pattern

Every system trapped in a core deficit exhibits a recurring loop—a sequence of phases that repeats, often for decades, with each cycle slightly eroding the system’s adaptive capacity.

Diagnostic questions:

• What triggers the loop? (External shock? Internal crisis? Accumulation of pressure?)
• What happens in each phase? Draw it: Phase 1 → Phase 2 → Phase 3 → Phase 4 → back to Phase 1.
• What prevents the loop from breaking? What mechanism ensures the system returns to its starting point rather than evolving?
• Does each cycle leave the system slightly weaker than the previous one?

Examples from the series:

• Sweden: High trust → Delegation to institutions → Signal suppression → Problems accumulate below the surface → Sudden politicization → Reactive correction → Trust slightly degraded → Return to high trust (slightly weaker)
• Japan: Pressure emerges → Incremental accommodation → Core paradigm preserved → Deferral → Next cycle from weaker baseline
• Israel: External threat → National mobilization → Securitization/Expansion → Internal fragmentation → Renewed threat in more polarized environment
• Nigeria: Oil revenues flow to center → Elite extraction → State neglects core functions → Informal networks adapt → Periodic crisis forces temporary realignment → Extraction resumes from more fragile baseline

Step 3: Identify the Structural Mechanisms

The core deficit and signature loop are maintained by specific mechanisms. These are not “problems” in the generic sense—they are institutional, fiscal, legal, and informational structures that produce the observed outcomes.

Common mechanism categories:

• Fiscal architecture: How is revenue raised and distributed? Does the system depend on resource rents rather than taxation? Does the fiscal structure sever the link between revenue and accountability?
• Electoral system: Does proportional representation produce chronic fragmentation? Does a winner-take-all system suppress diversity? Does the system give disproportionate power to specific constituencies (rural, older, ethnic)?
• Constitutional structure: Is there a formal constitution? Are rights entrenched? Is there an authoritative mechanism for resolving fundamental disagreements? Or is the architecture constitutionally incomplete?
• Administrative capacity: Can the state implement what it decides? Is the civil service professional or captured? Are functions hollowed out?
• Information architecture: How is information gathered? What aggregation occurs? What’s the latency between a problem emerging and a response arriving? What dimensions of reality does the system not measure?
• Security architecture: Does the security establishment exercise disproportionate influence? Has emergency logic colonized civilian governance?
• Parallel governance: Do informal systems (traditional, religious, criminal, digital) perform governance functions that the formal state cannot?

Diagnostic questions:

• What are the 6-10 most important structural mechanisms that maintain the current architecture?
• For each mechanism: Who benefits? What would need to change to alter it? What prevents that change?
• How do the mechanisms reinforce each other? Draw the causal links.

Step 4: Name the Immune System

Every governance architecture develops an immune system—a set of institutions, incentives, and actors that protect the existing order from challenge. The immune system is not a conspiracy. It is the predictable output of an architecture that rewards certain behaviors and punishes others.

Common immune system types:

• Bureaucratic inertia: The system is too slow and procedural for reform to gain traction (Germany)
• The spectacle of centralization: Reform energy is absorbed by dramatic announcements that change nothing (France)
• Satisfied competence: “We’re already the best in the world—what exactly needs fixing?” (Sweden, Finland)
• Stability Bias: Change is experienced as threat to identity; continuity is the supreme value (Japan)
• Extraction Coalition: Actors who benefit from the status quo control the institutions through which reform would need to pass (Nigeria, Brazil)
• Security First Responder: Security logic dominates all other domains; deviation from security-first thinking is treated as existential threat (Israel)
• Veto Industrial Complex: An economic ecosystem profits from gridlock; blocking is structurally cheaper than building (United States)

Diagnostic questions:

• Who benefits from the current architecture? Be specific—name the actors, institutions, and constituencies.
• What would they lose if the system changed?
• What mechanisms do they use to resist reform? (Blocking coalitions? Capture of reform processes? Cultural narratives that make the status quo feel normal or obligatory?)
• How would you design a reform that bypasses the immune response rather than triggering it?

Step 5: Identify the Cultural Anchor

Every governance system operates within a cultural operating system—a set of deeply embedded concepts, values, and habits that make the current architecture feel normal, natural, or obligatory. The cultural anchor is not a stereotype. It is the specific cultural logic that makes the core deficit feel like virtue rather than dysfunction.

Examples from the series:

• Saklighet (Sweden): Fact-based, consensus-oriented objectivity that suppresses uncomfortable signals
• Wa + Kaizen + Gaman + Shouganai (Japan): Harmony, continuous improvement, stoic endurance, and acceptance of the unchangeable—together producing dignified stagnation
• Oga-Madam + “The National Cake” + Jugaad + Pentecostal Resilience (Nigeria): Patronage, resource distribution logic, improvisational hustle, and religious meaning-making—together making the extractive state feel normal and survivable
• Ein Breira + Balagan + Covenant Consciousness + Tikun Olam (Israel): Existential survival mentality, improvisational chaos, sacred mission, and ethical self-critique—together enabling extraordinary resilience while preventing political settlement

Diagnostic questions:

• What concepts, phrases, or values do people in this system invoke to explain why things are the way they are?
• What does the culture treat as virtuous that, from a structural perspective, perpetuates the dysfunction?
• How would you reframe the transition architecture in terms that honor the cultural anchor rather than attack it? What’s the narrative that says “we’re not abandoning our values; we’re fulfilling them”?

Step 6: Design the Transition Architecture

Diagnosis without transition design is commentary. The framework requires specifying how the system might evolve—not a comprehensive blueprint, but the principles, mechanisms, and first steps that could shift the trajectory.

Key design principles:

• Work with the grain, not against it. The transition architecture must be culturally legible and must avoid triggering the immune response. Use Trojan Horse mechanisms—reforms packaged as enhancements to existing values rather than repudiations of them.
• Start where the energy is. Don’t begin with the blocked central institutions. Begin with the municipalities, the sectoral sandboxes, the islands of integrity where change is already happening.
• Scale by attraction, not mandate. Let successful experiments attract imitators. The role of the center is to enable, fund, protect—and get out of the way.
• Build the missing observation channels. The core deficit is always, at root, an information problem. The transition architecture must create new ways for the system to perceive what it currently cannot.

Diagnostic questions:

• What’s the Trojan Horse? What reform can be framed as strengthening existing values while carrying transformative payload?
• What’s the safe-to-fail pilot? Where can the new architecture be tested without threatening the entire system?
• What’s the concrete first step? Not the most ambitious reform, but the most catalytic—the one that, once established, generates the information, constituencies, and political logic that make further reform possible.
• Who are the constituencies for this change? How will they be mobilized?
• What’s the honest timeline? Is this feasible in years, decades, or generations?

Using This Guide: A Checklist

When you’ve completed your analysis, you should be able to fill in the following:

Where This Framework Comes From

This guide distills the Country Reports for Systemic Change series (14 reports, 2026) and the Governance as Engineering working-papers (6 papers). The series applies control theory, information theory, Ashby’s Law of Requisite Variety, and Spiral Dynamics to the diagnosis of governance failure. The core finding across all cases: governance failures are not primarily failures of competence, resources, or political will. They are failures of observation architecture—the structures that determine what a governance system can perceive and respond to.

The full series is available at bjornkennethholmstrom.org.

A Note on the LLM Recipe

The method described in the Quick-Start section above is how the Country Reports series was produced. It works because:

1. Stage Yellow/Turquoise systems thinking is rare among humans but accessible to LLMs. The cognitive move of seeing structural patterns across domains, of recognizing that optimization creates blind spots, of treating values as design variables—this is precisely what current frontier models are good at, when prompted correctly.

2. Multi-model orchestration compensates for individual model limitations. Each model has characteristic strengths and blind spots. Cross-feeding produces richer analysis than any single model could generate alone. The disagreements between models are often the most analytically productive outputs.

3. The recipe democratizes a capability that was previously restricted to a tiny cognitive elite. You don’t need decades of training in systems thinking to produce a Stage Yellow analysis of a governance system. You need to know how to ask the right questions and how to orchestrate the tools.

This is not a replacement for human judgment. The LLMs will sometimes produce sophisticated-sounding nonsense. They will miss crucial local context. They will reflect the biases in their training data. The human analyst must verify claims, ground analysis in real institutional knowledge, and exercise editorial judgment. But the recipe dramatically accelerates the process and surfaces insights that would take months or years to generate through traditional research methods.

If you produce a country report using this method, consider publishing it openly. The framework benefits from more cases, more perspectives, and more critical scrutiny. The series is a living body of work, not a finished monument.

This guide is offered under Creative Commons Attribution-ShareAlike 4.0 International. Use it, adapt it, improve it. The framework belongs to anyone who finds it useful.