Index Basket Construction: Build Your Own Trade Indices

Learning Objectives

By the end of this module, you will:

What Are Trade Indices?

Definition: A trade index is a composite metric combining multiple underlying data points into a single normalized value.

Structure:

Index Value = Σ (Component_i × Weight_i)

Where:

Example: Trans-Pacific Supply Chain Index

Components:

If normalized values are LA=110, Panama=95, Shanghai=105, SCFI=120:

Index = (110 × 0.40) + (95 × 0.20) + (105 × 0.25) + (120 × 0.15) = 108.25

Key Principle: Indices compress complexity into signal. A single index value tells you "supply chain health" without tracking 4+ separate metrics.

Why Build Custom Indices?

Advantages Over Single-Component Markets:

  1. Diversification: One component's noise (LA Port labor strike) doesn't derail entire thesis if other 3 components perform as expected

  2. Holistic Exposure: Capture end-to-end supply chain (origin port + chokepoint + destination port + pricing)

  3. Correlation Trading: Exploit relationships between components (if Shanghai volume up but LA volume down, bottleneck is in between—trade the spread)

  4. Customization: Build indices matching your analytical edge (e.g., Asia-Europe chokepoint specialist creates Suez/Malacca/Bosphorus index)

When to Use Index vs Component Markets:

Use Index Markets When:

Use Component Markets When:

Quotable Framework: "Indices are for themes. Components are for tactical bets. If your edge is 'Trans-Pacific trade will boom,' trade the index. If your edge is 'LA will steal share from Oakland,' trade the components."

Index Construction: Step-by-Step Framework

Step 1: Define Index Objective

Question: What economic reality are you trying to measure?

Examples:

Trans-Pacific Flow Index: Measure volume of goods moving from Asia to U.S. West Coast

Southeast Asia Chokepoint Basket: Measure geopolitical risk to maritime routes connecting Asia to Europe/Middle East

U.S.-China Trade Intensity Index: Measure bilateral trade relationship strength (volume × efficiency)

European Port Congestion Index: Measure operational stress at major EU container ports

Clear objective determines component selection. Don't mix unrelated metrics (e.g., Suez transits + U.S. retail sales—different phenomena).

Step 2: Select Components

Criteria:

  1. Relevance: Component directly measures aspect of index objective
  2. Verifiability: Data is publicly available, objectively measurable
  3. Timeliness: Data updates frequently (weekly/monthly, not annual)
  4. Historical availability: Need 2+ years of data to backtest correlations

Example: Trans-Pacific Flow Index

Objective: Measure Asia→U.S. West Coast container trade flows

Component Candidates:

| Metric | Relevance | Data Source | Frequency | Historical Depth | |--------|-----------|-------------|-----------|------------------| | LA Port TEUs | ✓ Direct (destination) | Port Authority | Monthly | 20+ years | | Long Beach TEUs | ✓ Direct (destination) | Port Authority | Monthly | 20+ years | | Oakland TEUs | ✓ Direct (destination) | Port Authority | Monthly | 20+ years | | Shanghai outbound | ✓ Direct (origin) | Port Authority | Monthly | 15+ years | | SCFI Asia-USWC rate | ✓ Indirect (pricing proxy) | Shanghai Shipping Exchange | Weekly | 10+ years | | Panama Canal transits | ✗ Indirect (some cargo, but also non-Asia flows) | Canal Authority | Monthly | 20+ years | | U.S. retail sales | ✗ Demand driver, not flow | Census Bureau | Monthly | 30+ years |

Selected Components (most relevant + direct):

  1. LA Port TEUs
  2. Long Beach TEUs
  3. Shanghai outbound TEUs
  4. SCFI Asia-USWC freight rate index

Step 3: Normalize Components

Problem: Components have different units and scales:

Can't average raw values (3.8M + 900k + 1,200 = meaningless)

Solution: Normalize each component to common scale (typically 0-150 with baseline=100)

Normalization Formula:

Normalized Value = (Actual Value / Baseline Value) × 100

Baseline Options:

Option A: Historical Average (e.g., trailing 12-month average)

Option B: Fixed Reference Period (e.g., January 2020 = 100)

Option C: Seasonal Baseline (e.g., same month prior year = 100)

Best Practice: Use Option A (trailing 12-month average) for most indices—balances stability and growth adjustment.

Example Normalization:

LA Port, December 2024:

Shanghai, December 2024:

Now both components on same scale (100 = baseline performance).

Step 4: Determine Weights

Question: How much should each component contribute to final index value?

Weighting Approaches:

A. Equal Weighting (simplest):

B. Volume/Size Weighting:

C. Inverse Volatility Weighting:

D. Analytical Judgment:

Recommended Hybrid Approach:

Start with size-based weights, then adjust for:

  1. Diversification: Cap any single component at 40% (avoid over-concentration)
  2. Signal Quality: Upweight components with low measurement error
  3. Practicality: Round to 5% increments (easier to rebalance and explain)

Example: Trans-Pacific Flow Index

Initial Size Weights (based on 2024 annual TEU volume):

| Component | Annual Volume | Raw Weight | Adjusted Weight | |-----------|---------------|------------|-----------------| | LA Port | 9.5M TEUs | 18% | 35% (upweighted as key USWC hub) | | Long Beach | 8.8M TEUs | 17% | 30% (paired with LA) | | Shanghai | 47M TEUs | 48% | 25% (downweighted to avoid domination) | | SCFI Rate | N/A (pricing) | 17% | 10% (supplemental signal) |

Final Weights: LA 35%, Long Beach 30%, Shanghai 25%, SCFI 10% (sum=100%)

Rationale:

Step 5: Calculate Index Value

Formula:

Index = Σ (Normalized_Component_i × Weight_i)

Example: Trans-Pacific Flow Index, December 2024

| Component | Normalized Value | Weight | Contribution | |-----------|------------------|--------|--------------| | LA Port | 108.6 | 35% | 38.01 | | Long Beach | 102.5 | 30% | 30.75 | | Shanghai | 105.0 | 25% | 26.25 | | SCFI Rate | 115.0 | 10% | 11.50 | | Total | - | 100% | 106.51 |

Trans-Pacific Flow Index = 106.51 (6.51% above baseline)

Interpretation: Trans-Pacific container flows are 6.5% stronger than 12-month average, driven by elevated freight rates (+15%) and solid LA performance (+8.6%).

Try index calculation on Ballast → Trans-Pacific Flow Index

Correlation Analysis: When to Combine Components

Key Question: Should you combine highly correlated or uncorrelated components?

Answer: Depends on your goal—diversification vs concentration.

Strategy 1: Diversification (Uncorrelated/Low Correlation)

Objective: Reduce index volatility by combining independent signals

Target Correlation: 0.2 to 0.5 (some relationship but not lockstep)

Example: Global Maritime Chokepoint Index

Components:

Correlation Matrix (hypothetical, based on historical data):

| | Suez | Panama | Malacca | |--|-------|--------|---------| | Suez | 1.00 | 0.35 | 0.42 | | Panama | 0.35 | 1.00 | 0.28 | | Malacca | 0.42 | 0.28 | 1.00 |

Average correlation: ~0.35 (low to moderate)

Benefit: If Red Sea crisis tanks Suez transits (-30%), but Panama and Malacca are unaffected, index only falls -10% (weighted average).

Use Case: Hedging portfolio—you want exposure to "global chokepoint risk" without being devastated by single-region event.

Strategy 2: Concentration (High Correlation)

Objective: Amplify signal by combining reinforcing metrics

Target Correlation: 0.7 to 0.9 (strong co-movement)

Example: U.S. West Coast Port Capacity Index

Components:

Correlation Matrix (realistic estimates):

| | LA | Long Beach | Oakland | |--|-----|-----------|---------| | LA | 1.00 | 0.88 | 0.75 | | Long Beach | 0.88 | 1.00 | 0.72 | | Oakland | 0.75 | 0.72 | 1.00 |

Average correlation: ~0.79 (high)

Implication: All three ports move together (same supply chain, shared infrastructure, common labor). Combining them doesn't diversify—it CONCENTRATES the bet on "USWC port system."

Use Case: Conviction trade—you believe entire USWC region will outperform, and you want full exposure to that theme.

When High Correlation Is Optimal: If your analytical edge is thematic (macro trade view, policy forecast), you WANT concentrated exposure. Diversification dilutes your edge.

Quotable Framework: "Diversify when you're hedging uncertainty. Concentrate when you have conviction. Correlation is a tool, not a rule."

Worked Example: Building Southeast Asia Chokepoint Basket

Objective: Create an index measuring geopolitical risk to maritime routes connecting Southeast Asia, Middle East, and Europe.

Step 1: Component Selection

Chokepoints in Scope:

  1. Malacca Strait: Connects Indian Ocean to South China Sea (80k+ transits/year)
  2. Suez Canal: Connects Mediterranean to Red Sea (20k+ transits/year)
  3. Bab el-Mandeb Strait: Connects Red Sea to Gulf of Aden (30k+ transits/year)

Metrics:

Step 2: Normalization

Baseline: Trailing 12-month average transits for each chokepoint

December 2024 Data:

| Chokepoint | Actual Transits | 12-Mo Avg | Normalized | |------------|-----------------|-----------|------------| | Malacca | 6,800 | 6,500 | 104.6 | | Suez | 1,620 | 1,850 | 87.6 | | Bab el-Mandeb | 2,100 | 2,400 | 87.5 |

Step 3: Weighting

Approach: Volume-based with geopolitical adjustment

| Chokepoint | Annual Transits | Raw Weight | Adjusted Weight | Rationale | |------------|-----------------|------------|-----------------|-----------| | Malacca | 82,000 | 61% | 40% | High volume but historically stable (downweight) | | Suez | 19,500 | 15% | 30% | Critical Europe-Asia link (upweight) | | Bab el-Mandeb | 28,000 | 21% | 30% | High conflict risk (upweight) |

Total: 100%

Rationale: Malacca dominates by volume (61%) but is geographically distant from Middle East conflict zones. Suez and Bab el-Mandeb are more geopolitically sensitive (Red Sea tensions, Yemen conflict, Iranian navy presence) → upweight their importance in "risk index."

Step 4: Calculate Index

Index = (104.6 × 0.40) + (87.6 × 0.30) + (87.5 × 0.30)
      = 41.84 + 26.28 + 26.25
      = 94.37

Southeast Asia Chokepoint Index = 94.4 (5.6% below baseline)

Step 5: Interpretation

Index below 100 indicates reduced chokepoint traffic vs normal. Why?

Signal: European-bound cargo is rerouting around Africa (Cape of Good Hope), skipping Suez and Bab el-Mandeb. This is geopolitical risk materializing.

Trade Setup:

Binary Market: "Will Southeast Asia Chokepoint Index remain below 95 for Q1 2025?" (YES at $0.52)

Analysis: Red Sea tensions unlikely to resolve quickly (Houthi attacks ongoing). Shipping lines committed to Cape routing for 3+ months (insurance, route planning lead time).

Trade: Buy YES at $0.52 (bet index stays suppressed)

Outcome (April 2025): Index averaged 92.8 in Q1 2025 → YES pays $1 → 92% profit

Try chokepoint baskets on Ballast → Southeast Asia Chokepoint Index

Rebalancing Strategies

Question: How often do you update component weights and baselines?

Static Index (No Rebalancing)

Approach: Set weights and baseline at inception, never change.

Example: Trans-Pacific Index launched Jan 2024 with weights LA=35%, LB=30%, Shanghai=25%, SCFI=10%. These weights never adjust.

Pros:

Cons:

Best For: Short-lived markets (1-2 year horizon) or highly stable supply chains

Dynamic Index (Periodic Rebalancing)

Approach: Update weights quarterly or annually based on recent data.

Example: Recalculate size-based weights every January 1 using prior 12 months' data.

Rebalancing Rule:

2024 Weights (based on 2023 data):

2025 Weights (based on 2024 data):

Pros:

Cons:

Best For: Long-term indices (5+ year horizon) or rapidly evolving markets

Baseline Rolling Window

Approach: Baseline updates continuously (e.g., trailing 12-month average recalculated each month).

Example:

December 2024 Baseline: Average of Jan-Dec 2024 data January 2025 Baseline: Average of Feb 2024-Jan 2025 data (drops Jan 2024, adds Jan 2025)

Effect: Each component's normalized value is always relative to "recent normal," not ancient history.

Pros:

Cons:

Best For: Volatility/change detection (you care about deviations, not absolute levels)

Try rebalancing strategies on Ballast → Index Methodology Lab

Advanced: Spread Trades Between Indices

Concept: Trade relative performance of two indices to exploit structural themes.

Example: Trans-Pacific vs Trans-Atlantic Index Spread

Thesis: U.S. import demand shifting from European suppliers to Asian suppliers due to USMCA incentives.

Indices:

Trans-Pacific Index (Asia→U.S.):

Trans-Atlantic Index (Europe→U.S.):

Spread: Trans-Pacific 110 - Trans-Atlantic 98 = +12 points

Market: "Will Trans-Pacific vs Trans-Atlantic spread exceed +15 points in Q2 2025?"

Analysis:

Trade: Buy YES at $0.48

Outcome: Q2 average spread = +16.8 points → YES pays $1 → 108% profit

Quotable Framework: "Index spreads isolate structural shifts. If you trade single indices, you're exposed to global demand shocks. If you trade spreads, you isolate RELATIVE changes—less noise, purer signal."

Common Pitfalls

Pitfall 1: Over-Diversification

Problem: Building 10-component index with equal weights (10% each).

Why It Fails: Each component contributes only 10% to index. Need very large moves in multiple components to see meaningful index change. Signal drowns in noise.

Solution: Limit to 3-5 components. If you need more, create sub-indices (e.g., "Port Index" + "Chokepoint Index" + "Freight Rate Index" → combine into "Supply Chain Super-Index").

Pitfall 2: Mixing Units Without Normalization

Problem: Averaging raw values of different units (TEUs + transit counts + rate indices).

Why It Fails: TEUs are in millions, transits in thousands, rates in hundreds → large components (TEUs) dominate just by scale, not economic importance.

Solution: Always normalize first (all components to 0-150 scale), THEN apply weights.

Pitfall 3: Static Weights in Dynamic Markets

Problem: Using 2020 weights in 2025 index when market structure changed.

Why It Fails: If LA doubled capacity 2020-2025 but weight stayed 35%, index underrepresents LA's current importance.

Solution: Rebalance weights annually or use % of recent volume (automatically adjusts).

Pitfall 4: Correlation Blindness

Problem: Combining 4 highly correlated components (0.9+ correlation) thinking it diversifies risk.

Why It Fails: If all components move together, you haven't diversified—you've just averaged 4 versions of same signal.

Solution: Calculate correlation matrix. If average correlation over 0.7, you're not diversifying (which is fine if that's your goal—just be aware).

Pitfall 5: Ignoring Data Lags

Problem: Combining weekly data (AIS vessel counts) with monthly data (port TEUs) in index.

Why It Fails: Weekly component updates 4x faster, dominates index movements even if weighted equally.

Solution: Match data frequencies (all monthly, or aggregate weekly to monthly) or downweight higher-frequency component.

Frequently Asked Questions

1. How many components should an index have?

Ideal: 3-5 components. Fewer than 3 = not really an "index" (just a pair trade). More than 5 = over-diversification (diminishing returns on complexity).

2. Can weights be negative?

Rarely. Negative weight means index goes UP when that component goes DOWN (inverse relationship). Possible for hedging indices but confusing. Better to use positive weights and trade opposite side (sell index if you're bearish).

3. What if a component's data isn't published one month?

Option A: Use prior month's value (carry forward) Option B: Exclude component and re-weight others (e.g., if LA data missing, distribute its 35% across other 3 components) Option C: Void that month's index (don't calculate)

Ballast uses Option A (carry forward) with disclaimer if data is over 60 days stale.

4. Can I trade components separately after defining an index?

Yes. Index construction doesn't lock you into trading only the index. You can trade index for thematic exposure AND trade individual components for tactical views.

5. How do I backtest an index?

Download historical data for all components (past 5 years). Apply normalization and weights using current methodology. Calculate historical index values. Compare to outcomes you would've forecasted (or actual prediction market resolutions if available).

6. What if two components measure similar things (LA and Long Beach are adjacent ports)?

That's fine IF your objective is "USWC port system" (they're complements). But if you want "diversified global port index," they're redundant—choose one or combine them into single "USWC" component.

7. Should freight rates (pricing) be weighted equally with volume metrics?

Usually NO. Freight rates are more volatile and noisy. Typical approach: 5-15% weight to price signals, 85-95% to volume signals (volumes are what you're measuring; prices are validating signal).

8. Can indices include non-trade data (GDP, PMI, etc.)?

Technically yes, but it muddles interpretation. "Trade flow index" should measure FLOWS (TEUs, transits). If you want composite economic indicator, build separate "Trade Health Index" with macro variables—but that's harder to objectively resolve in prediction markets (GDP has revisions, PMI is survey-based).

9. How do I handle outliers (e.g., port strike causing -50% drop one month)?

Option A: Include outlier (index reflects reality) Option B: Smooth with 3-month moving average (reduces noise) Option C: Use median instead of mean for baseline (outliers don't skew baseline)

For prediction markets, Option A (include outliers) is standard—traders are forecasting actual outcomes, including strikes.

10. Can I create index of indices?

Yes (meta-index). Example: "Global Supply Chain Super-Index" = 50% Trans-Pacific Index + 30% Trans-Atlantic Index + 20% Intra-Asia Index. Be cautious—adds complexity, harder to interpret, and correlation across sub-indices might be high (reducing diversification benefit).

11. What's the difference between index and portfolio?

Index: Single composite metric calculated by fixed formula (everyone computes same value from same data)

Portfolio: Collection of positions with dynamic management (trader discretion on weights, entry/exit timing)

Indices are passive/systematic. Portfolios are active/discretionary.

12. How do I know if my index is "good"?

Backtesting Criteria:

Run 2-year backtest. If you can't predict index movements better than coin flip, redesign components/weights.

Ready to Apply What You've Learned?

Turn knowledge into action.

Start Trading on Ballast Markets →

Use prediction markets to apply the concepts from this learning module. Trade real contracts based on port volumes, shipping delays, chokepoint transits, and tariff impacts.

Next Steps

Practice Exercises:

  1. Build Custom Index: Choose 4 ports/chokepoints YOU follow. Gather 12 months of data. Normalize, weight, calculate index. Compare to your intuition (does index value match your qualitative sense of strong/weak trade?).

  2. Correlation Analysis: Calculate correlation matrix for 5 major ports (LA, Long Beach, NY/NJ, Savannah, Houston). Which pairs are over 0.7 correlated (redundant)? Which are less than 0.4 (diversifying)?

  3. Rebalancing Simulation: Create static index with Jan 2023 weights. Create dynamic index with annual rebalancing. Compare index values for 2023-2024. How much did they diverge?

Continue Learning:

Try on Ballast Markets:

Advanced Resources:

Disclaimer

This content is for educational purposes only and does not constitute financial advice. Index construction involves subjective choices (component selection, weights, rebalancing rules). Different methodologies yield different results. Backtest performance does not guarantee future accuracy. Prediction markets on indices involve risk. Start with small positions and only risk capital you can afford to lose.