Testing H5: Intensity-Impact Testing H6: Duration Matters

The Intuition We Wanted to Prove

It seems obvious: deadlier wars should hurt markets more. If an intervention involves more troops, more casualties and more escalation, the economic damage should be proportionally larger. We also expected that longer conflicts would compound these effects over time, creating a clear relationship between war intensity and market pain.

The reality is more nuanced. Our analysis reveals that intensity alone doesn't predict market outcomes. We observe a correlation between duration and negative returns, but here's the catch: 20-year conflicts like Afghanistan inevitably overlap with other major economic events like the 2008 financial crisis, Iraq War and the dot-com bubble. Correlation doesn't prove causation and we cannot isolate duration's true causal effect from these confounding factors.

Not all military interventions are equal in scale or severity. Some are brief and symbolic, while others are prolonged, escalatory and carry substantial human and geopolitical costs.

To analyze how markets respond to these differences, we construct a single intensity index that captures multiple dimensions of intervention severity.

But how do you combine lethality, escalation and duration into a single number without making arbitrary choices? This is where PCA comes in.

Understanding the Intensity Index

How we quantify severity without guessing

The Problem: Avoiding Arbitrary Choices

We could have simply decided that fatalities matter twice as much as duration or that troop deployment is the most critical factor. But that would be subjective, we would be imposing our human bias on the historical data.

The Solution: Letting the Data Speak

Instead, we used Principal Component Analysis (PCA). This technique allows the data itself to determine the mathematical "weight" of each factor based on how they correlate naturally.

How it works: If high fatalities and long durations tend to happen together, PCA detects this pattern and combines them into a single signal. The result is a purely data-driven Intensity Score from 0 to 1 that captures the true "magnitude" of an intervention.

What Goes Into the Intensity Score?

Lethality (Weight: 0.65)

The human cost. Based on fatality counts (log-transformed to handle the huge range from minor skirmishes to major wars). This is the strongest contributor and deadlier conflicts score higher on intensity.

Escalation (Weight: 0.58)

How aggressive is the posture? Combines 9 variables including hostility levels, whether it's declared a "war," and military activity indicators. Higher escalation means more aggressive military engagement.

Duration (Weight: 0.49)

How long does it last? Measured in days (log-transformed). Longer interventions indicate sustained military commitment. Interestingly, this has the lowest weight, suggesting lethality and escalation matter more for "intensity."

Why Does This Matter?

The first principal component explains 55% of the total variance across these three dimensions. That's a moderate correlation, meaning lethality, escalation and duration tend to move together, but not perfectly. Some interventions are deadly but brief like Desert Storm, while others are prolonged but low-intensity like peacekeeping missions. This nuance is exactly what we need to test whether intensity predicts market reactions.

Now comes the crucial test of our Intensity-Impact hypothesis (H5): do high-intensity interventions actually cause bigger market crashes?

Each point in the visualization below represents one of our 29 interventions, positioned by its intensity score and the market's cumulative response.

If our hypothesis is correct, we should see a clear downward slope where more intensity equals worse returns. But look closely at the outliers. They tell a different story.

Intensity vs Market Reaction

Each point represents an intervention. Toggle between return and volatility metrics.

The Tale of Two Wars: Afghanistan vs. Desert Storm

Afghanistan: Persistent Negative Returns

The War in Afghanistan stands out with a -28% cumulative abnormal return, the sum of daily market underperformance over the entire conflict. At intensity 0.86 and lasting 7,632 days, it's both severe and prolonged. This cumulative drag represents the sustained economic cost of open-ended military commitments compounded over 20 years.

Desert Storm: High Intensity, Positive Returns

Operation Desert Storm (intensity: 0.59) achieved +7.46% CAR despite being a major military operation. The difference? It lasted just 348 days with a clear objective and decisive outcome. Markets don't fear intensity; they fear uncertainty. A swift, successful operation actually boosted confidence by demonstrating U.S. military capability and removing the uncertainty of "what happens next."

The Duration Confound (H6)

This is where our Duration Matters hypothesis (H6) becomes complex. Vietnam (3,092 days, intensity 0.91) and Afghanistan (7,632 days, intensity 0.86) are the two longest conflicts and they're also the two worst performers. However, correlation doesn't prove causation. Multi-decade conflicts inevitably overlap with other major economic events like the 2008 financial crisis, concurrent Iraq War, and the dot-com bubble. We cannot isolate duration's true causal effect from these confounding factors.

Key Insight

Intensity alone doesn't predict market outcomes. While longer conflicts correlate with worse returns, we must be cautious about causation. Multi-decade wars overlap with other major economic events, making it impossible to isolate duration's true effect. Quick, high-intensity operations with clear objectives can actually boost markets by reducing uncertainty. Markets price in geopolitical risk gradually and don't panic proportionally to how many troops are deployed.

The scatter plot tells a story, but can we quantify it? Let's look at the statistical relationship between intensity and market reactions across all 29 interventions.

We group interventions into four intensity quartiles to see if the pattern holds when aggregating the data.

Correlation by Intensity Quartile

Average market reaction grouped by intervention intensity level.

Error bars show standard deviation.

What the Statistics Actually Tell Us

Return Correlation

Loading correlation statistics...

Volatility Correlation

Loading correlation statistics...


Translating the Numbers to Words
What Does r = -0.36, p = 0.057 Mean?

The Pearson correlation of -0.36 suggests a moderate negative relationship: as intensity goes up, returns tend to go down. But here's the catch: the p-value of 0.057 is just above the traditional 0.05 threshold. In plain terms, there's about a 6% chance this pattern is just random noise. We can't confidently claim intensity predicts returns.

Why Pearson and Spearman Disagree

Pearson measures linear relationships while Spearman measures rank-based relationships. Pearson says -0.36 (moderate negative), but Spearman says +0.06 (essentially zero). This discrepancy is a red flag: a few extreme outliers like Afghanistan and Vietnam are pulling the Pearson correlation down, but when we just look at rankings, there's no trend.

The Q4 Collapse: Duration in Disguise

The quartile chart shows cumulative returns (summed over each intervention's duration). The highest intensity quartile (Q4: 0.49-0.91) averages -264.73% cumulative return compared to +39.43% for the other three quartiles combined. But this extreme gap is misleading because Q4 contains Vietnam (3,092 days) and Afghanistan (7,632 days), the two longest conflicts. When you sum daily returns over 20 years, even small daily losses become enormous. Duration, not intensity, drives these cumulative numbers.

The Statistical Verdict

The borderline p-value (~0.06) means we cannot confidently claim that intensity alone predicts market returns. The relationship exists but is confounded by duration. High-intensity interventions tend to last longer and it's the prolonged exposure to uncertainty that hurts markets, not the intensity per se. This is why our Intensity-Impact hypothesis (H5) is only partially confirmed: intensity matters, but only when combined with duration.

Aggregate relationships between intensity and returns reveal broad trends, but individual interventions often deviate from the average pattern.

This comparison tool allows interventions to be examined side by side, highlighting how differences in intensity translate into sector-level outcomes.

By directly contrasting two cases, the tool helps clarify when intensity amplifies risk and when other factors dominate market behavior.

Compare Interventions

Select two interventions to compare their sector impacts side by side

Comparison Insights

Across our 29 interventions, 58.6% achieved positive cumulative abnormal returns, challenging the assumption that military action always hurts markets. The key differentiator isn't intensity alone, but how intensity interacts with duration and strategic clarity.

The Intensity Paradox

High-intensity interventions (>0.5) average -3.33% CAR, but this masks huge variation. The Lebanese Civil War (intensity: 0.73) achieved +4.33% CAR while Afghanistan (intensity: 0.86) suffered -28.08% CAR. The difference? Lebanon lasted 581 days with clear objectives; Afghanistan dragged on for 7,632 days.

The Regime Change Puzzle

Desert Storm and Just Cause had nearly identical intensity (~0.58), yet Desert Storm returned +7.46% while Just Cause lost -1.72%. The sector breakdown reveals why: Desert Storm boosted Consumer Discretionary by +0.73% daily while Just Cause dragged it down -0.67%.

Sector Resilience

Financial Services and Consumer Defensive are the most stable sectors during interventions (std dev ~0.26%). Meanwhile, Telecommunications shows the highest volatility (std dev 1.33%), making it a high-risk, high-reward sector during military conflicts.

Energy Sector Surprise

Counter-intuitively, Energy doesn't always benefit from military interventions. The sector averaged -0.23% daily returns across all interventions. The best Energy performance came from Gulf of Sidra (+0.81% MAR), while Libya-Egypt tensions saw the worst (-1.84% MAR).

The Verdict: Testing Our Hypotheses

H5
Intensity-Impact
PARTIAL

We expected: High-intensity wars will cause bigger market crashes.

Reality: Correlation exists (r = -0.36) but isn't statistically significant (p = 0.057). Outliers like Afghanistan and Vietnam drive the pattern, not a consistent relationship.

Why: Intensity matters only when combined with duration. Swift, decisive operations (Desert Storm: +7.46% CAR) can boost markets.

H6
Duration Matters
PARTIAL

We expected: Longer wars will cause more cumulative damage to markets.

Reality: Afghanistan (7,632 days, -28% CAR) and Vietnam show the worst returns, but these 20+ year conflicts overlap with other major events (2008 crisis, dot-com bubble).

Why: Long wars inevitably coincide with other economic shocks. We observe correlation, but cannot claim duration causes the negative returns.

RQ3
How does intervention intensity influence market reactions?

The data tells a story that challenges conventional wisdom:

  • Moderate negative correlation: r = -0.36 (Pearson), but p = 0.057 means we can't be confident this isn't random noise.
  • Outliers dominate: Afghanistan (-28% CAR) and Vietnam pull the average down; Spearman correlation (+0.06) shows no rank-based trend.
  • Duration correlation (with caveats): Longer conflicts correlate with worse returns, but confounding events make causal claims difficult.
  • Swift operations can boost markets: Desert Storm (intensity 0.59, +7.46% CAR) shows that decisive action reduces uncertainty.