Lebanon: Daily Risk Brief
Lebanon faces imminent state collapse amid active Israeli military operations and mass displacement. With 1.2 million internally displaced, a $9.2 billion fiscal deficit, and sustained airstrikes across southern territory, the country's stability trajectory is sharply downward. High confidence: escalation will continue absent
August 30, 2026 · Score 28.1
GeoMemo briefs are generated under an ICD 203 prompt: bottom line up-front, every assertion cited to the dispatch it came from, and probability framing that survives analyst scrutiny.
Start your morning with the GeoMemo Brief.
Never miss what’s moving. Ten minutes. Multi-capital lens. In your inbox by 07:30 EDT.
Global recession
Escalating trade tensions, tariffs, and geopolitical uncertainties, which can exacerbate global economic instability and increase the risk of a recession.
±1σ (typical), inside the normal range. Baseline noise.
±1σ to ±2σ (elevated), outside the usual range. Worth a second look.
Beyond ±2σ (critical), statistically unusual. Something changed.
Technical: σ (lowercase sigma, the standard-deviation symbol, not to be confused with Σ, the summation symbol) = (current probability − 60-day baseline mean) / baseline standard deviation. Baseline excludes the last 7 days so recent movement isn’t self-referenced.
Why it matters: a threat where the coverage concentrates in one country reads differently from one spread across a region. “KP 80% / US 15% / JP 5%” tells you North Korea’s provocation story is being told mostly by Korean press. “US 35% / RU 20% / CN 18%” tells you it’s a great-power story. Same probability, different interpretive weight.152 countries
Global food crisis
India's potential reduction in domestic farm output due to a weak monsoon and geopolitical disruptions, which could contribute to a global food crisis, albeit with a moderate and indirect impact.
±1σ (typical), inside the normal range. Baseline noise.
±1σ to ±2σ (elevated), outside the usual range. Worth a second look.
Beyond ±2σ (critical), statistically unusual. Something changed.
Technical: σ (lowercase sigma, the standard-deviation symbol, not to be confused with Σ, the summation symbol) = (current probability − 60-day baseline mean) / baseline standard deviation. Baseline excludes the last 7 days so recent movement isn’t self-referenced.
Why it matters: a threat where the coverage concentrates in one country reads differently from one spread across a region. “KP 80% / US 15% / JP 5%” tells you North Korea’s provocation story is being told mostly by Korean press. “US 35% / RU 20% / CN 18%” tells you it’s a great-power story. Same probability, different interpretive weight.195 countries
Russia-Ukraine ceasefire
A significant escalation of violence with a Russian strike killing 38 people, indicating a continuation of hostilities and making a ceasefire less likely in the near term.
±1σ (typical), inside the normal range. Baseline noise.
±1σ to ±2σ (elevated), outside the usual range. Worth a second look.
Beyond ±2σ (critical), statistically unusual. Something changed.
Technical: σ (lowercase sigma, the standard-deviation symbol, not to be confused with Σ, the summation symbol) = (current probability − 60-day baseline mean) / baseline standard deviation. Baseline excludes the last 7 days so recent movement isn’t self-referenced.
Why it matters: a threat where the coverage concentrates in one country reads differently from one spread across a region. “KP 80% / US 15% / JP 5%” tells you North Korea’s provocation story is being told mostly by Korean press. “US 35% / RU 20% / CN 18%” tells you it’s a great-power story. Same probability, different interpretive weight.183 countries
The sparkline reconstructs this daily average by forward-filling each constituent threat’s last known probability and averaging across the category. Categories with only 1 tracked threat (e.g. currently AI governance) are excluded, “average of one” isn’t an index.category-level composite · 30 days trajectory
The board shows what our model currently believes, backed by the evidence count (how many dispatches contributed) and confidence intervals. Not forecasting, continuous re-weighting of a claim against all available reporting.
