Russia–Ukraine alignment (Fan, annual)

## What this measures

Annual alignment between Russia and Ukraine, scaled from −1 (maximum conflict) to +1 (maximum cooperation). The score is a dynamic measure that averages the tone of geopolitical events over the year, adjusted for persistence and decay.

## Background

Annual Russia–Ukraine alignment, −1 to +1.

**The events behind this series were compiled by a large language model (Gemini 2.5 Pro with web search), not collected by hand.** The paper validates reliability across four model families (0.88); there is no human-coded ground truth. Alignment is the dynamic score (annual Goldstein average ÷10 with persistence, depreciation 0.3). A partner counts once it has at least five effective events.

This is a research release that ends at 2024 and updates only when the author publishes a new version; /api/cron/fan-geopolitics-sync checks monthly and re-ingests when the file changes. For a daily read on the same scale, see the GDELT alignment series.

Sources: Tianyu Fan (2025), "Measuring Geopolitical Alignment and Economic Growth"; Tianyu Fan, Mai Wo, and Wei Xiang (2025), "Geopolitical Barriers to Globalization"; Global Geopolitical Events Database v1.0, June 2026 (MIT).

## Why it matters

The Russia–Ukraine relationship shapes energy markets, grain supply, defense procurement and regional stability across Europe and Central Asia. Organizations with exposure to Eastern European operations, commodity flows or security risk need a quantitative read on whether the bilateral stance is improving or deteriorating. A sustained shift in alignment can precede sanctions, trade disruptions or investment climate changes that affect planning horizons.

## How to read it

A score near zero indicates neutral or mixed relations; values below zero indicate net hostility; values above zero indicate net cooperation. A fall toward −1 signals rising tension; a rise toward +1 signals warming ties. Compare this series to other bilateral alignment pairs or to the daily GDELT alignment series to distinguish long-term trends from short-term noise.

## What it does not say

This series does not forecast future events or capture sub-annual turning points; it is a lagging, annual summary. The events behind the score were compiled by a language model, not coded by hand, so the measure reflects model interpretation rather than human-verified incident logs.

## Source

Tianyu Fan (2025) via the Global Geopolitical Events Database v1.0 (MIT), compiled by a large language model with web search and validated across four model families. Palanor ingests the research release monthly and updates when the author publishes a new version.