Methodology & Data

How the Deflation Index is calculated, plus complete data downloads and alternative weighting methodologies.

Overview

The Deflation Index tracks technological cost reductions across four fundamental sectors: computing, communications, energy, and transportation. Each sector is measured using cost-per-performance metrics that capture the true force of technological progress.

The DI is a weighted geometric mean of these four sector indices, showing how technological deflation compounds over time.

DI = Computing^0.294 × Communications^0.235 × Energy^0.294 × Transport^0.176

Why geometric mean? Technological improvements multiply (compound), they don't add. A geometric mean properly captures this multiplicative relationship.

Where each sector represents fundamental technological progress:

Current Scope: These weights represent v3.0.2 composition with four sectors. Future versions (v4.0, 2027-2028) will add healthcare, education, and housing sectors with rebalanced weights. See limitations section for full methodology discussion.

Note: These weights represent the current v3.0.2 composition with four sectors. Future versions will add healthcare, education, and housing sectors with rebalanced weights.

The Four Sectors

Computing (29.4%)

Metric: $/GFLOPS, $/GB storage, $/GB RAM
Deflation: 99.91% (1990-2024)
Annual Rate: 35-40%
Coverage: 35 years
Sources: AI Impacts, Nordhaus, Backblaze

Communications (23.5%)

Metric: $/GB data, $/minute voice, $/Mbps bandwidth
Deflation: 99.8% (1990-2024)
Annual Rate: 30-35%
Coverage: 35 years
Sources: FCC, BLS, industry reports

Energy (29.4%)

Metric: $/kWh solar LCOE, $/kWh batteries, $/kilolumen LEDs
Deflation: 90% (2010-2024)
Annual Rate: 12% pre-2010, 27% post-2010
Coverage: 35 years
Sources: IRENA, BloombergNEF, DOE

Transportation (17.6%)

Metric: $/kWh EV batteries, autonomous tech costs
Deflation: 88% (2010-2024)
Annual Rate: 18-22%
Coverage: 15 years
Sources: BloombergNEF, DOE, Waymo

Note on Future Expansion: v4.0 will add healthcare, education, and housing sectors. Current v3.0.2 composition uses four sectors with normalized weights totaling 100%.

Calculation Method

Step 1: Measure Cost-Per-Performance

For each component within each sector, we track the cost per unit of performance over time:

Step 2: Calculate Component Indices

Each component is indexed to 100 in the base year (1990 for most, 2010 for energy/transport). As costs fall, the index falls proportionally.

Component Index = (Cost in Current Year / Cost in Base Year) × 100

Step 3: Weight Components Within Sectors

Multiple components within each sector are combined using geometric means and their respective weights. For example, Computing = (GFLOPS^0.60 × Storage^0.30 × RAM^0.10).

Step 4: Combine Sectors Into DI

The four sector indices are combined using the DI formula shown above, producing a single comprehensive measure of technological deflation.

Sector Weight Justification

Why these specific weights? The current weights (29.4%, 23.5%, 29.4%, 17.6%) reflect relative economic importance across three dimensions:

  1. GDP Contribution: Share of US GDP from these sectors
  2. Enabling Effect: Degree to which other sectors depend on this technology
  3. Deflationary Force: Magnitude of cost reduction in sector
Weight Rationale by Sector:

Computing (29.4%): High enabling effect (powers nearly everything), moderate GDP share, extreme deflationary force (99.9%+ reduction)

Energy (29.4%): High GDP share (universal input), extreme enabling effect (literally powers economy), high deflationary force (90% solar reduction)

Communications (23.5%): Moderate GDP share, high enabling effect (essential infrastructure), extreme deflationary force (99.8% data cost reduction)

Transportation (17.6%): Moderate GDP share, moderate-high enabling effect (critical for goods movement), high deflationary force (88% battery reduction, shorter time series)

Sensitivity Analysis: Are Results Robust?

We tested four alternative weighting schemes to ensure our findings aren't artifacts of weight choices:

Robustness Check: Alternative Weightings

To test sensitivity, we calculated results under alternative weighting schemes:

Weighting Approach Weights (C/Co/E/T) Annual DI Cumulative DI
Current (Multi-Factor) 29/24/29/18 -9.21% -96.25%
Equal Weights 25/25/25/25 -8.45% -95.04%
Expenditure-Weighted 25/20/30/25 -8.43% -94.99%
GDP-Only Weights 20/15/40/25 -8.37% -94.88%

Result: The core finding—approximately 94-96% cumulative deflation at 8.3-9.2% annually—is robust across reasonable weighting approaches. The current weighting produces the highest deflation rate because it gives more weight to the fastest-deflating sectors (Computing, Communications).

Detailed analysis: See Weight Justification document for complete methodology, derivations, and comparisons to alternative approaches.

v4.0 Plan: When healthcare, education, and housing sectors are added (2027-2028), we'll switch to expenditure-weighting as primary method while keeping current multi-factor approach as alternative for comparison.

Step 5: Calculate Annual Rates

Year-over-year percentage changes show the annual deflation rate. The average annual rate (1990-2024) is -9.21%.

Data Sources

All data comes from authoritative, publicly available sources. Every data point includes source citations in the Excel files.

Source What We Use Reliability
Federal Reserve (FRED) M2 money supply, CPI data 100/100 (A+)
Bureau of Labor Statistics CPI components, historical pricing 100/100 (A+)
IRENA Solar LCOE, renewable energy costs 95/100 (A)
BloombergNEF Battery costs, EV economics 95/100 (A)
AI Impacts Computing power costs (FLOPS) 90/100 (A-)
DOE LED efficiency, transportation data 95/100 (A)
FCC Broadband pricing, network costs 90/100 (A-)
Backblaze Hard drive cost tracking 85/100 (B+)

Average source reliability: 92/100 (A-grade). We prioritize government agencies, international organizations, and industry leaders with transparent methodologies.

Download the Data

All data is open and free. Download the Excel files, verify every formula, and build your own analysis.

Complete Dataset

DI v3.0.2 with all sectors, M2, and gap analysis

Download Excel

GitHub Repository

All source data, scripts, and documentation

View on GitHub

Advanced: Alternative Weighting Methodologies

Why provide variants? To demonstrate the robustness of our findings across different methodological approaches. All variants use the same source data and sector indices—only the weights differ.

1. Multi-Factor Weighting Primary

Methodology: Balances three factors: GDP contribution, enabling effect (how much other sectors depend on it), and deflationary force (magnitude of cost reductions).

Weights: Computing 29.4% • Communications 23.5% • Energy 29.4% • Transportation 17.6%

Result: 96.26% cumulative deflation | 2024 DI: 3.74 | Annual avg: -9.21%

Use case: Recommended default. Most comprehensive approach accounting for both direct and indirect economic impact.

Download Primary

2. Equal-Weighted

Methodology: Simple equal weighting across all sectors with no assumptions about relative importance.

Weights: Computing 25% • Communications 25% • Energy 25% • Transportation 25%

Result: ~95.8% cumulative deflation | 2024 DI: ~4.2 | Annual avg: ~-9.0%

Use case: Simple baseline, conservative estimate, easy to explain and defend.

Download Equal-Weighted

3. Expenditure-Weighted

Methodology: Weighted by household and business expenditure patterns—where money is actually spent.

Weights: Computing 25% • Communications 20% • Energy 30% • Transportation 25%

Result: ~95.5% cumulative deflation | 2024 DI: ~4.5 | Annual avg: ~-8.8%

Use case: Consumer-centric perspective, "where did my money go?" analysis, policy discussions.

Download Expenditure-Weighted

4. GDP-Weighted

Methodology: Weighted by direct GDP contribution—traditional economic weighting.

Weights: Computing 20% • Communications 15% • Energy 40% • Transportation 25%

Result: ~94.2% cumulative deflation | 2024 DI: ~5.8 | Annual avg: ~-8.3%

Use case: Macro-economic analysis, GDP impact studies, conservative estimate.

Download GDP-Weighted

Sensitivity Analysis Summary

Variant 2024 DI Cumulative Annual Avg
Multi-Factor (Primary) 3.74 -96.26% -9.21%
Equal-Weighted ~4.2 ~-95.8% ~-9.0%
Expenditure-Weighted ~4.5 ~-95.5% ~-8.8%
GDP-Weighted ~5.8 ~-94.2% ~-8.3%

Key Finding: All variants show substantial technological deflation (94-96% cumulative). The core finding is robust across all methodologies. Choice of weights affects magnitude but not direction.

Individual Sector Data

Download sector-specific indices with detailed calculations and source data.

Computing

Performance-adjusted price index

Download

Communications

Data transmission cost index

Download

Energy

Solar + LED cost index

Download

Transportation

Battery cost index

Download

Data Quality Standards

Every data point in the Deflation Index is graded, sourced, and verified. We maintain rigorous quality standards:

The Numbers

Quality Grading System

Grade Score Criteria
A (Excellent) 85-100 Government agencies, peer-reviewed research, industry gold standards
B (Good) 70-84 Reputable industry reports, academic estimates, established analysts
C (Fair) 50-69 Secondary sources, interpolations, reasonable estimates
D (Poor) <50 Weak sources, speculative estimates (not used in index)

Conservative Assumptions

Where uncertainty exists, we choose the more conservative estimate. This means the true technological deflation is likely higher than we measure. The index understates the gap, not overstates it.

Transparency Commitment

Every Excel file includes:

Anyone can verify our work. That's the standard we hold ourselves to.

Institutional Review

The Deflation Index methodology is open for formal scrutiny. We believe rigorous measurement improves through expert review, institutional feedback, and collaborative refinement.

Seeking Expert Review

We welcome formal review from:

Current Status:

Partnership Opportunities

We're looking for:

Data Integration

Bloomberg, Refinitiv, FactSet, and other platforms—let's discuss data feeds and API access

Research Partnerships

Co-author research, validate methodology, collaborate on sector expansion

Sector Expansion

Domain experts for v4.0: healthcare economists, education researchers, housing analysts

Critical Feedback

Challenge assumptions, propose alternatives, identify improvements

How to Engage

Institutional inquiries: For data licensing, API access, or partnership discussions, contact us directly.

Technical feedback: If you spot errors or have methodology suggestions, open an issue on GitHub with details.

Research use: Cite the Deflation Index in research, analysis, and publications. We encourage independent verification.

Contact: [email protected]

Our commitment: Precision, transparency, and intellectual honesty. If you see something wrong, we want to know. If you have ideas to improve the work, we're listening.

The DI-M2 Gap

The core measurement reveals a systematic divergence between two opposing forces:

Annual Gap: 14.9 Percentage Points

DI-M2 Gap = |Tech Deflation Rate| + M2 Expansion Rate
= 9.21% + 5.7% = 14.9pp annually

Technology delivered 9.21% annual deflation across the four sectors. Meanwhile, the money supply expanded 5.7% annually. These opposing forces create a 14.9 percentage point gap every year.

Cumulative Divergence (1990-2024)

DI: 100 → 3.74 (96.26% deflation)
M2: 100 → 650.2 (550% expansion)

Over 35 years, this annual gap compounds into massive divergence. Technology made these sectors 96% cheaper while the money supply expanded 550%. The question: where does the productivity go?

The answer: Into asset prices, intermediaries, financial complexity, and system overhead—not into lower prices for consumers.

Known Limitations & Future Improvements

We're transparent about where the Deflation Index is currently limited. Better to acknowledge constraints openly than to overstate capabilities.

Current Scope Limitations

Sectors Covered (v3.0.2):

Coverage: The current four sectors represent approximately 40% of measurable technological deflation, not 100%. We prioritize defensible measurement over comprehensive coverage—better to measure four sectors perfectly than twenty sectors poorly.

Expansion Timeline: v4.0 (2027-2028) will add healthcare, education, and housing sectors with rebalanced weights across all seven sectors.

Methodological Limitations

Weighting System:

Index Construction:

Conservative Assumptions:

What We Don't Measure

Why These Limitations: We prioritize verifiable, reproducible data over comprehensive coverage. Each limitation represents an opportunity for future improvement and is documented for transparency.

Data Frequency Limitations

Current:

Planned:

Our Commitment to Transparency

These limitations don't invalidate the core finding: massive technological deflation occurred (-96.26% cumulative) while consumer prices rose (+155%). This divergence exists regardless of methodological choices.

What changes with better methodology: The precise magnitude of the gap. Not whether it exists.

We welcome critique: Challenge our assumptions, propose alternatives, suggest improvements. The goal is the best measurement possible, not defending any particular method.