Demo: TCA Implementation

This page outlines a demonstration implementation of True Cost Accounting using real-world food products to test the framework’s ability to:

1. Distinguish between products with different externality profiles
2. Show how the weighted norm reveals “sustainability traps” that simple addition misses
3. Demonstrate the unknown penalty incentive mechanism
4. Provide visual comparisons and consumer-facing signals

Experiment Overview

Objective: Compare Multi-Dimensional TCA (with weighted norm) against simple additive “true price” methods to demonstrate:

• How the norm penalizes concentrated harms
• How the unknown penalty enforces transparency
• How consumer health ($C_4$) directs away from ultra-processed foods

Product Selection: Choose 5-6 products with distinct externality profiles across social, environmental, and health dimensions.

Phase A: Product Selection

Select products that test different aspects of the framework:

Product 1: Brazilian Beef (1 kg, conventional feedlot)

Profile: High environmental ($C_3$), moderate social ($C_2$), moderate health ($C_4$)

• Why: Tests environmental intensity (carbon, water, land)
• Expected: High $C_3$ dominates in norm

Product 2: Dutch Organic Carrots (1 kg)

Profile: Low across all dimensions

• Why: Baseline for low-externality fresh produce
• Expected: Low $C_1$, low externality norm

Product 3: Ultra-Processed Soda (1 liter)

Profile: Low social ($C_2$), low-moderate environmental ($C_3$), high health ($C_4$)

• Why: Tests consumer health penalty (Nutri-Score E, NOVA 4)
• Expected: High $C_4$ creates large externality signal despite low other components

Product 4: Fair Trade Dark Chocolate (100g bar)

Profile: Low social ($C_2$ due to certification), moderate environmental ($C_3$), moderate-good health ($C_4$)

• Why: Shows how verification reduces unknown penalty
• Expected: Transparent supply chain → lower penalty

Product 5: Generic Milk Chocolate (100g bar, unknown origin)

Profile: Unknown social ($C_2$), unknown environmental ($C_3$), moderate health ($C_4$)

• Why: Tests unknown penalty mechanism
• Expected: 95th percentile priors create large penalty vs. Fair Trade version

Product 6: Ultra-Processed Breakfast Cereal (500g box)

Profile: Low-moderate social ($C_2$), moderate environmental ($C_3$), high health ($C_4$)

• Why: Tests NOVA 4 + poor Nutri-Score combination
• Expected: Health component dominates

Phase B: Data Collection

Primary Data Sources

Environmental ($C_3$) Indicators

• EXIOBASE v3.8 for sector-country intensities
• ecoinvent for specific products where available
• Agri-footprint for agricultural products

Indicators to collect:

• kg CO₂e per kg product
• m³ water (scarcity-adjusted) per kg
• ha·year land use per kg
• Eutrophication, acidification potentials

Social ($C_2$) Indicators

• Eora MRIO + ILOSTAT for sector-country labor intensities
• Fair Trade certifications for verified products
• TPF case studies for chocolate, coffee supply chains

Indicators to collect:

• Child labor hours per kg (for cocoa, coffee sectors)
• Living income/wage gap per kg
• OHS DALY intensity per € output → convert to per kg

Health ($C_4$) Indicators

• Open Food Facts for Nutri-Score, NOVA, ingredients
• GBD Study for DALY calibration

Indicators to collect:

• Nutri-Score (A/B/C/D/E)
• NOVA (1/2/3/4)
• Added sugar, salt content

Applying the Unknown Penalty

For Product 5 (Generic Milk Chocolate):

1. Origin unknown → assume “Cocoa, Global” sector
2. Query EXIOBASE for cocoa production:
   • 95th percentile carbon intensity: 8 kg CO₂e/€
   • 95th percentile water intensity: 5 m³/€
3. Query Eora + ILOSTAT for cocoa social intensity:
   • 95th percentile child labor: 0.5 hours/€
   • 95th percentile living income gap: €0.40/€
4. Price assumption: €2/100g
5. Apply 95th percentile values × monetization factors

Result: Generic chocolate receives conservative high externality estimate until supply chain is verified.

Phase C: Calculation Battle

Calculate Simple “True Price” (Addition)

$P_{\text{simple}}(p)=C_1+C_2+C_3+C_4$

No weighting, no norm—just sum all externalities.

Calculate TCA with Weighted Norm

$N_{ext}(C)=\sqrt{w_2{C}_2^2+w_3{C}_3^2+w_4{C}_4^2}$

$P_{\text{TCA}}(p)=C_1+N_{ext}(C)$

Baseline test weights: $w_2=w_3=w_4=1/3$ (equal weighting)

Alternative test weights: $w_2=0.4,w_3=0.35,w_4=0.25$ (social emphasized)

Phase D: Results Visualization

Table: Component Costs

Product	$C_1$ (€)	$C_2$ (€)	$C_3$ (€)	$C_4$ (€)	$P_{\text{simple}}$ (€)	$N_{ext}$ (€)	$P_{\text{TCA}}$ (€)	Difference
Brazilian Beef (1kg)	8.00	1.50	106.92	12.50	128.92	107.19	115.19	-10.6%
Dutch Carrots (1kg)	1.50	0.10	0.80	0.50	2.90	0.96	2.46	-15.2%
Soda (1L)	1.50	0.20	2.50	58.00	62.20	58.05	59.55	-4.3%
Fair Trade Chocolate (100g)	3.50	0.80	4.20	8.50	17.00	9.66	13.16	-22.6%
Generic Chocolate (100g)	2.00	35.00	22.00	8.50	67.50	43.00	45.00	-33.3%
Ultra Cereal (500g)	3.00	1.20	5.50	72.00	81.70	72.18	75.18	-8.0%

Key observations:

1. Soda & Cereal: High $C_4$ dominates both methods, but norm slightly reduces signal vs. simple addition
2. Generic Chocolate: Unknown penalty creates massive $C_2$ due to child labor/forced labor 95th percentile priors
3. Brazilian Beef: Environmental component is so large it dominates both methods
4. Fair Trade: Verification dramatically reduces social cost vs. generic version

Visual: 4-Dimensional Vector Plot

Create radar/spider charts showing $(C_1,C_2,C_3,C_4)$ for each product:

• Brazilian Beef: Huge spike in $C_3$
• Soda: Huge spike in $C_4$
• Generic Chocolate: Large spike in $C_2$ (unknown penalty)
• Fair Trade Chocolate: Balanced, smaller spikes
• Carrots: Small, balanced profile

Visual: Price Comparison Bar Chart

Show side-by-side:

• Market price ($C_1$) in dark gray
• $P_{\text{simple}}$ (simple addition) in light blue
• $P_{\text{TCA}}$ (with norm) in green

Highlight products where norm and addition differ most → reveals “sustainability traps”

Phase E: Analysis of Results

Identifying “Sustainability Traps”

Definition: Products where simple averaging looks acceptable but one dimension has catastrophic harm.

Example from results:

Generic Chocolate:

• Simple average externality: $(35+22+8.5)/3=€21.83$
• Norm externality: $\sqrt{(35^2+22^2+8.5^2)/3}=€25.3$
• Interpretation: Norm is ~16% higher, but more importantly, it surfaces that high $C_2$ (child/forced labor) shouldn’t be offset by moderate other components

Ultra Cereal:

• If this had slightly lower $C_4$ but very high $C_3$, simple addition would average them; norm would penalize the outlier more heavily

Transparency Signal Impact

Compare Fair Trade Chocolate vs. Generic Chocolate:

Component	Fair Trade (verified)	Generic (unknown)	Penalty Effect
$C_2$	€0.80	€35.00	43.75×
$C_3$	€4.20	€22.00	5.24×
Total impact	Verified as low	Penalized for opacity	Huge incentive to verify

Conclusion: Unknown penalty creates massive financial incentive to:

1. Disclose supply chain origins
2. Obtain third-party audits
3. Achieve certifications (Fair Trade, Rainforest Alliance, etc.)

Health Directing

Products with Nutri-Score E and NOVA 4 (ultra-processed, nutrient-poor):

• Soda: $C_4=€58/L$ (assuming regular consumption pattern)
• Ultra Cereal: $C_4=€72/500g$

Effect: These costs are 20-40× the market price, creating a strong signal that:

• Reflects true public health burden (diabetes, cardiovascular disease, obesity costs)
• Directs consumers toward whole foods
• Incentivizes reformulation (reduce sugar, improve processing)

Phase F: Consumer-Facing Implementation

Retail Display Options

Option 1: 4-Color Traffic Light

• Each light sized proportional to $C_k^2$ (squared to match norm)
• Red: $C_2$ (social)
• Orange: $C_3$ (environmental)
• Yellow: $C_4$ (health)
• Green: $C_1$ (economic)

Visual: Immediately shows which dimension dominates

Option 2: Externality Norm Score (0-100)

• Normalize $N_{ext}(C)$ to 0-100 scale (0 = best, 100 = worst)
• Display as single number with color coding
• Brazilian Beef: 92/100 (high externality)
• Carrots: 8/100 (low externality)

Option 3: Receipt Summary

At checkout, show:

Market Total:        €45.50
Externality Total:   €287.30
True Cost Total:     €332.80

Top contributors:
- Beef (1kg):        €107.19 externality
- Generic Chocolate: €43.00 externality

Effect: Makes aggregate choices visible without overwhelming per-item.

Phase G: Link to GitHub Repository

Placeholder: Full implementation with code, data, and visualizations will be available at:

GitHub Repository: github.com/reinierkruisbrink/true-cost-accounting-demo

Contents:

• Python scripts for data extraction (EXIOBASE, Open Food Facts APIs)
• Calculation notebooks (Jupyter) showing step-by-step TCA computation
• Visualization scripts (matplotlib, plotly) for charts and comparisons
• Data files (product profiles, intensities, monetization factors)
• Documentation (reproducibility guide, data sources, assumptions)

Status: (To be created—this is a placeholder for the actual experiment)

Next Steps for Full Implementation

1. Data Pipeline:

   • Automate queries to EXIOBASE, Eora, Open Food Facts
   • Build concordance tables (product → sector)
   • Extract intensity distributions and quantiles

2. Calculation Engine:

   • Implement component calculation ($C_2,C_3,C_4$)
   • Apply TPF monetization factors
   • Compute norm with configurable weights
   • Handle missing data with unknown penalty

3. Visualization:

   • Radar charts for multi-dimensional profiles
   • Bar charts for price comparisons
   • Interactive web tool (Streamlit or Dash) for exploring products

4. Validation:

   • Compare results with existing TPF case studies
   • Sensitivity analysis on weights and quantile choices
   • Expert review of DALY mappings and monetization assumptions

5. Publication:

   • Academic paper with methodology and results
   • Policy brief for regulators
   • Open-source code and data for transparency and replication

---

Previous: Data Sources
Next: Critiques
Parent: TCA