How Long Do Interstate Conflicts Last? Evidence from UCDP/PRIO..
How long do wars between states typically last? Using the UCDP/PRIO Armed Conflict Dataset (ACD) v25.1, we analyzed the durations of interstate conflict episodes from 1946 to 2024. We also split episodes by intensity to see whether high-intensity wars tend to last longer.
This post summarizes the results:
Data and definitions.
Source: UCDP/PRIO Armed Conflict Dataset (ACD) v25.1.
Scope: Interstate conflicts only (type_of_conflict = 1), years 1946–2024.
Unit of analysis: Conflict episodes. We reconstruct episodes per conflict_id by grouping contiguous active years, starting a new episode when:
UCDP marks an episode end in the previous year (ep_end), or
there is a gap of more than one year in activity.
Duration: For each episode, Duration = end_year − start_year + 1 (in years).
Intensity classes:
High-intensity: the episode contains at least one year with UCDP intensity_level = 2 (≥1000 battle-related deaths in that year).
Low-intensity: all years in the episode have intensity_level = 1 (<1000 battle-related deaths per year).
Note: This is a standard and transparent way to turn annual conflict-year panels into episode durations, consistent with the event logic in UCDP.
Key results:
Overall (all interstate episodes, 1946–2024; N = 21 episodes)
Mean duration: about 5.6 years
Range: 1 to 14 years
Interquartile range (IQR): 7 years
By intensity
High-intensity (≥1000 battle deaths in at least one year of the episode)
N = 7 episodes
Mean duration ≈ 8.4 years
Median duration = 9 years
Low-intensity (<1000/year)
N = 14 episodes
Mean duration ≈ 4.1 years
Median duration = 4 years
Interpretation: High-intensity interstate episodes last notably longer on average than low-intensity ones. The median difference is striking: 9 years versus 4 years.
What the graphic shows:
We created a compact two-panel figure:
Left panel: bar chart of mean episode duration (overall and by intensity).
Right panel: boxplot showing the distribution of durations by intensity class.