Sethu Nguna.
Independent project · ISS African Futures Conference · 2026

Sovereign Futures Lab

Presented & facilitated

Participants sit in a fictional country’s Debt Management Office six weeks before a $750 million maturity. The treasury reports $780 million in cash. How much of it can actually be used is the first thing they have to find out.

← Explore all projects
Audience
Foresight practitioners at the ISS African Futures Conference, 14 October 2026
Format
About 6 minutes of preparation, 20 minutes of timed casework in eight stages, 10–12 minutes of debrief
Research base
AidData’s How China Lends and How China Collateralizes
Built with
React, TypeScript and Vite; Supabase; a deterministic consequence engine; runs in the browser
The Sovereign participant workspace, with the eight-stage decision chain on the left and the case facts on the right
The participant workspace: eight stages the facilitator unlocks in turn, with the case facts that set the clock.
The design problem

The challenge

Research on how sovereign loans are secured explains the mechanisms: lender-controlled accounts, confidentiality clauses, linked facilities. Foresight practice needs a way to rehearse the decisions that depend on them, under a deadline, with partial evidence, in a role that can recommend but not commit.

The logic

Bound the exercise to one role and one decision chain. Make the evidence visible but uncertain, the consequences deterministic and every submission immutable, so the debrief can reconstruct exactly what each participant knew when they decided.

What participants produce

An internal negotiation-preparation brief that states what is known, unknown and conditional, and a written reflection on what they will carry into real decisions.

My role

I designed the method and the scenario, built the participant application and facilitator console, and presented and facilitated the session at ISS.

Facilitator console
Facilitator console design showing session health, institutional workflow position and a live activity log
Facilitator console design reference: session health, workflow position and live activity. Figures shown are illustrative.

The exception desk

Routine evidence requests get authored responses and delays. A deliberately small exception desk lets the facilitator answer unusual requests in the voice of Treasury, Legal, the IMF, the Official Creditor Committee or a creditor, without opening a general messaging system.

How a session runs

Prepare, decide, reconstruct.

About forty minutes, with the casework clock running only while participants are deciding.

01

Prepare · ~6 min

Orientation to the role and its authority boundary, and a Learning Bridge, both outside the casework clock.

02

Unlock by discussion · 20 min

The facilitator opens each stage when its discussion condition is met, not when time passes: mandate, liquidity, evidence, record, linkage, disclosure, then the brief.

03

Request and wait

Evidence requests return after authored delays. Anything unusual goes to the exception desk.

04

Submit the brief

Each participant submits a versioned brief, preserved exactly as submitted.

05

Debrief · 10–12 min

Reconstruct the evidence each participant had at submission, compare pathways without ranking them, discuss one fictional consequence and run one counterfactual with its fixed assumptions named.

06

Transfer

Each participant records what they will do differently in a real decision under uncertainty.

Design decisions

Every mechanic has a job.

The exercise trains how to decide with evidence that is incomplete and authority that is limited.

DecisionWhat it trainsIn the simulation
A role that can recommend but not commit Separating analysis and recommendation from sovereign commitment. The Debt Management Office holds the loan agreements and claims record, but not the cash position, the legal reading or the authority to release anything externally.
A balance that isn’t what it seems Distinguishing reported cash from usable liquidity. Reported liquidity is $780m. Once the evidence is in, $240m is restricted and $60m protected, leaving $480m usable.
Evidence that arrives late Prioritising requests by the decision they serve. Requests return after delays, so requested evidence is not yet available evidence.
Two claims, kept apart Not letting a legal label stand in for operational fact. Copper-export receipts flow into a restricted account and are swept for debt service. Whether two loan facilities share a dependency is a separate question.
Disclosure within limits Setting a boundary between transparency and confidentiality. Legal allows a redacted functional summary. Full contract text needs consent.
Unlock by discussion, not by clock Keeping the room on reasoning, not pace. Each stage card gives the facilitator a purpose, an opening question, cues to listen for and an unlock condition.
Immutable submissions Honest reconstruction instead of hindsight. The debrief works from the version each participant actually submitted.
Deterministic consequences Discussing causes without pretending to predict. Consequences are fictional and rule-based, and the debrief says plainly that they are neither a forecast nor a competence score.
Evidence

What’s documented

Presented and facilitated at the ISS African Futures Conference on 14 October 2026. Each release runs automated tests on the deterministic engine and the interface, and the participant and facilitator surfaces are live.

Limits

What the evidence can’t show

No participant outcome data has been published. The lab is deliberately a bounded slice of the larger Sovereign Room method, not a full negotiation simulation, and its consequences are fictional rather than predictive.

Supporting artifacts

Explore the work.

Open the lab ↗

The live application. Participant sign-in needs a workshop code.

Source ↗

Application source, scenario content, research cards and the facilitator runbook

Start a conversation

Build the capability.
Make it work in practice.

Learning strategy, institutional systems and decision practice.

sethnguna@gmail.com ↗