Research papers and supporting documents for CHIP valuation
Will AI and automation raise or lower the real value of human labor? This paper argues that the answer depends not on the technology itself but on who has access to it and at what price. Under distributed access, each hour of human work becomes more productive, reservation wages rise, and new forms of "no-collar" work emerge. Under concentrated access — enforced by regulatory capture — workers are excluded from the gains.
The paper traces the argument from historical precedent through Baumol's cost disease to CHIP's evolution from a measured index to a market-discovered price, and makes the case that open AI is an economic prerequisite for the value of human time.
Should CHIP track currency inflation over time? This paper examines the original study's deflation methodology, reveals that deflation cancels algebraically in the CHIP formula, and develops the concept of a "deflationary baseline" — the expectation that in a free economy, technology should make goods cheaper over time. Proposes a trailing-window production methodology that publishes CHIP in nominal current-year dollars.
Key result: Real CHIP is approximately stable at $2.80–$3.15/hr (2005–2022). Nominal CHIP rises with inflation, automatically hedging against currency devaluation — exactly the behavior a labor-anchored unit should exhibit.
How should country-level CHIP values combine into a single global rate? This paper evaluates five weighting schemes — GDP, labor-force, unweighted, freedom-weighted, and HDI-weighted — through both philosophical analysis and empirical comparison. It addresses the circularity concern (does GDP-weighting just measure rich-country wages?) and publishes per-country multipliers from 0.04× to 2.39×.
Recommendation: GDP-weighting ($2.68/hr) is defensible: it accounts for population, aligns with the CHIP definition's emphasis on productive efficiency, and its circularity is broken by the distortion factor. HDI-weighting ($2.20/hr) brackets the value from below.
A focused evaluation of the original CHIP valuation methodology: its theoretical framework (Solow-Swan / Cobb-Douglas), data sources (ILOSTAT, PWT, FRED), the innovative distortion factor approach, and how well each design choice aligns with the canonical CHIP definition. Identifies strengths (theoretically grounded, reproducible, strong definition alignment) and limitations (informal economy bias, weighting sensitivity, approximate capital separation).
Verdict: The $2.53 estimate is a credible first approximation. Post-review updates summarize how workbench studies have since addressed several of the open questions identified.
This follow-up study evaluates the robustness of the original CHIP value computation, initially proposed to anchor the MyCHIPs currency to the global average of one hour of unskilled work. It incorporates ICT (Information and Communications Technology) capital data to test whether the inclusion of technology-specific capital affects the CHIP estimate.
Calculating a reasonable CHIP estimate requires analyzing mountains of data from markets across the world. This page extrapolates the CHIP valuation estimate over time by adjusting the base value for US Dollar inflation, providing a means of estimating the present CHIP value between more complex recalculations.
The foundational white paper: a conceptual framework to quantify the baseline value of the CHIP currency using econometric estimates from panel data across 89 countries over 1992–2019. Establishes the $2.53/hour result that anchors all subsequent work.
October 2023 update: Improved R code accessing public APIs (ILO, FRED, Penn World Tables) for reproducible, up-to-date CHIP valuations.
A full reproduction of the original R-based study in Python. Validates the original $2.53 result ($2.33/hr with fresh API data, within 1% of target) and provides the infrastructure for the workbench studies below.
The CHIP Workbench is a modular Python analysis environment built to systematically test the assumptions, sensitivity, and robustness of the CHIP valuation methodology. Each study below is independently reproducible — see the workbench README for setup and usage.
Develops a methodology for producing current-year CHIP estimates using PWT 11.0 data (2000–2023). Tests trailing-window smoothing, PWT bridge strategies for post-PWT years, and CPI-based extrapolation.
Key finding: 2022 nominal CHIP of $3.17/hr, closely matching the site's independent CPI-extrapolated value of $3.18. A 5-year trailing window reduces year-over-year volatility by 75%. CPI extrapolation is validated as a reliable bridge method between full recalculations.
Tests how sensitive the global CHIP estimate is to the choice of aggregation weights, comparing GDP-weighted, labor-weighted, unweighted, freedom-weighted, and HDI-weighted schemes across 85 countries.
Key finding: CHIP ranges from $1.67/hr (labor-weighted) to $2.85/hr (freedom-weighted) — a 51% spread, making weight choice a first-order methodological decision. GDP-weighting ($2.68/hr) is defensible: it produces the second-highest result and no single country dominates (USA = 24% weight). HDI-weighting is the most balanced alternative.
Tests how stable CHIP estimates are across data vintages by comparing PWT 10.0 vs PWT 11.0 results for 2000–2019. Examines whether previously published CHIP values shift when new PWT releases revise historical national accounts.
Key finding: Mean upward revision of +5.9% between PWT versions, with early years showing extreme shifts (up to +37.6%). For the mature period (2002–2019), revisions average 3.8%. Recommendation: always use the latest PWT release and anchor to recent years with a trailing window.
Examines how CHIP evolves year-by-year from 1992 to 2019 and tests whether nominal (un-deflated) CHIP produces different results from the deflated approach. Identifies an 11-country stable panel with consistent coverage.
Key finding: Deflation cancels algebraically in the CHIP formula — both elementary and average wages are scaled by the same deflator, so nominal and deflated CHIP are mathematically identical. Real CHIP is approximately stable at $2.80–$3.15/hr (2005–2019), and nominal CHIP tracks the GDP deflator closely.
Analyzes data availability across countries and time periods to assess CHIP estimate reliability. Examines coverage from ILOSTAT (employment, wages, hours), Penn World Tables, and the FRED GDP deflator.
Key finding: 123 countries have sufficient data to estimate CHIP, but wage data is the binding constraint (134 countries vs 208 for employment). Of these, 79 countries have excellent coverage (15+ years). Recommended common analysis range: 2000–2019.
All research code, data pipelines, studies, and supporting documents are available in the open-source CHIP GitHub repository. The workbench studies are fully reproducible from the workbench directory.
Areas where further investigation could strengthen the CHIP methodology
The current CHIP estimate relies on a single theoretical framework: the Cobb-Douglas production function with a distortion factor. How sensitive is the result to this choice? We have outlined — but not yet implemented — several alternative approaches:
If any of these alternatives yields a materially different CHIP value, that would be important to know. If they converge on the same range, it strengthens confidence in the current estimate. Either outcome advances the research.
Interested in contributing? The outline, evaluation criteria, and model comparison framework are ready. The workbench provides the data pipeline and infrastructure — a new study can be scaffolded in minutes. Funding may be available for bona fide research participants. Get in touch.