The maths, in daylight.

02 / Look-through

Country weights aggregate across funds.

For ETF i, let c_i,k be the reported weight of country k inside that fund, as a decimal. The portfolio country exposure is:

A 60% MSCI World allocation and a 40% emerging-markets allocation therefore contribute 0.60 × the World country weight plus 0.40 × the EM country weight. We do not re-normalise the result to hide cash, derivatives or non-equity positions.

03 / Company look-through

Effective company weight is additive.

For security j inside ETF i, let h_i,j be the security's reported fund weight as a decimal. Its effective portfolio weight is:

The same issuer can appear in several funds. We aggregate by a normalized issuer name (removing punctuation and common corporate suffixes such as Inc., Ltd., PLC, NV and Corp.), so Apple held by four ETFs is one row with several contributions. This is an issuer-name approximation, not a legal-entity identifier. It may merge or split entities with unusual naming.

The “fund contribution” columns show both the weight inside the ETF and the effective weight after your allocation. The reported country comes from the SEC filing. If the same normalized name carries conflicting countries, the row is flagged.

04 / Benchmarks

Benchmark shares use one reference year.

For a selected metric, the benchmark share of country k is its value divided by the sum of positive values across all countries with a reported value in the snapshot:

The difference is E_k − B_k, expressed in percentage points (pp). It is not a relative percentage change. The benchmark is normalised across the countries that actually reported the metric, so its coverage is shown beside the chart.

GDP is an economic output measure, not a target for stock ownership. Population is a demographic measure. Listed-market capitalisation measures listed domestic companies and is not total private wealth, bond value or real estate. The comparison is a lens, not a model of the “correct” portfolio.

05 / Limits

What the model refuses to pretend.

• Issuer country ≠ revenue exposure. A company organised in the US can earn most of its revenue elsewhere. We do not infer revenue geography.
• Holdings are snapshots. Funds change. The filing date, retrieval date and source are shown; stale data is flagged.
• Coverage differs by metric. The market-cap indicator covers fewer countries than population or GDP. The UI shows the count and missing values.
• Company aggregation is approximate. Names are normalized; share classes and cross-listings can complicate the mapping.
• No UCITS-wide coverage yet. The current catalogue uses US-listed reference funds with public SEC filings. A commercial UCITS holdings feed is a separate licensing decision.
• No natural-resource metric. A reliable, commercially usable country-level series was not selected; inventing or proxying one would weaken the comparison.
• Not advice. The tool does not assess suitability, tax, currency risk, costs or your objectives.