Portfolios are systems
PFAtlas reads a portfolio as a structure of weights, risks, assumptions and beliefs rather than as a list of isolated assets.
METHODOLOGY
How PFAtlas reads portfolios
PFAtlas reads a portfolio as an architecture of capital: a set of weights, assumptions, exposures and historical analogies that can be compared across regimes.
Core note
Historical portfolios in PFAtlas are not literal reconstructions of past returns. They are conceptual proxies: modern, investable approximations designed to translate the economic logic of a wealth system into comparable asset-class behavior.
Methodology principles
Historical portfolios are proxies
Historical wealth systems are translated into modern asset-class approximations, not reconstructed literally from archival balance sheets or exact return ledgers.
Results are comparative, not predictive
Performance history helps compare behavior across regimes. It does not forecast the future or convert a proxy into a recommendation.
Chapter
Scope of analysis
Unit of analysis
Portfolio architecture
PFAtlas' unit of analysis is the portfolio architecture: the allocation logic that determines how capital is split across broad economic functions. It does not attempt to rank individual funds, managers or securities.
PFAtlas is interested in what a portfolio expresses: the future it assumes, the risks it accepts, the institutions it trusts and the regimes in which it may fail.
Interpretation
What PFAtlas measures
PFAtlas measures portfolio architecture rather than individual security selection. A portfolio is treated as a system of exposures: how much it allocates to productive assets, nominal assets, real assets, cash, monetary hedges or speculative scarcity.
The goal is not to say whether a portfolio is good in isolation, but to compare what kind of world it is built for.
- asset weights
- asset-class exposures
- concentration and diversification
- macro-regime behavior
- historical relatives
- structural fragilities
Boundaries
What PFAtlas does not measure
PFAtlas does not provide personal investment advice, tax advice, product recommendations or guarantees of future performance. It does not know the user's full financial situation, liabilities, tax status, pension rights, liquidity needs or behavioral constraints.
PFAtlas is a research and education atlas. It is designed to help users think more clearly about portfolio structure, not to prescribe a specific investment action.
Chapter
Portfolio taxonomy
Portfolio types
How different portfolio classes should be read
PFAtlas includes several kinds of portfolios, and they should not all be interpreted in the same way.
Modern model portfolios
Examples include 60/40, Three-Fund, Permanent Portfolio and All Weather. These are closer to investable model allocations.
Historical wealth-system portfolios
Examples include Roman Elite, Venetian Merchant, Waqf Endowment or East India Company style portfolios. These are conceptual translations of historical systems into modern proxies.
Famous-person or dynasty portfolios
These are interpretive approximations of wealth engines associated with a person, family or institution. They are not audited personal balance sheets.
Custom user portfolios
User-defined allocations entered through PFAtlas tools are mapped into canonical asset classes so they can be compared with the Atlas library.
Historical translation
Historical portfolios as conceptual proxies
A Roman elite portfolio, a Venetian merchant portfolio or a Waqf endowment cannot be backtested literally. PFAtlas maps their economic logic into modern asset-class proxies such as real estate, cash, bonds, commodities, equities, gold or cryptoassets.
The question is not "What exact annual return did this historical person earn?" The question is "What economic system did this portfolio represent, and how might a modern proxy for that system behave?"
Chapter
Translation and proxy logic
Mapping logic
Proxy mapping logic
PFAtlas maps historical assets by economic function rather than by surface similarity alone.
Proxy mapping is interpretive. It should be transparent, debatable and improvable.
| Historical function | Modern proxy family | Rationale |
|---|---|---|
| Land / estates | Real estate, productive real assets | Stores wealth through ownership, rents and scarcity. |
| Gold / precious metals | Gold, monetary metals | Monetary hedge and reserve asset. |
| Merchant capital | Equities, commodities, cash | Exposure to trade, inventory, liquidity and business risk. |
| Sovereign credibility | Bonds, cash | Claims on state capacity and monetary institutions. |
| Industrial ownership | Equities | Claims on productive corporate capital. |
Historical function
Land / estates
Modern proxy family
Real estate, productive real assets
Rationale
Stores wealth through ownership, rents and scarcity.
Historical function
Gold / precious metals
Modern proxy family
Gold, monetary metals
Rationale
Monetary hedge and reserve asset.
Historical function
Merchant capital
Modern proxy family
Equities, commodities, cash
Rationale
Exposure to trade, inventory, liquidity and business risk.
Historical function
Sovereign credibility
Modern proxy family
Bonds, cash
Rationale
Claims on state capacity and monetary institutions.
Historical function
Industrial ownership
Modern proxy family
Equities
Rationale
Claims on productive corporate capital.
Chapter
Data and comparison engine
Evidence layer
Asset series and data sources
PFAtlas uses long-run return series and proxy datasets to represent major asset classes such as equities, bonds, cash, gold, commodities, real estate, cryptoassets and inflation.
The table below shows representative series used inside PFAtlas. It separates the direct coverage that is visible from the treatment used to preserve comparability across the backtest window.
| Series | PFAtlas role | Source / proxy label | Direct coverage | Backtest treatment / note |
|---|---|---|---|---|
| US Stocks | US equity growth proxy | Damodaran | 1970-2024 in the current annual dataset | Used directly in the current comparison window. |
| MSCI World | Global developed-equity proxy | MSCI | 1970-2024 in the current annual dataset | Used as the developed-world equity anchor where applicable. |
| ACWI | Global all-country equity proxy | MSCI | 2001-2024 direct ACWI-linked coverage | Pre-2001 is extended using MSCI World plus emerging-market weights, as noted in the series metadata. |
| US 10Y Treasury Bonds | Nominal sovereign-bond proxy | Damodaran | 1970-2024 in the current annual dataset | Used directly as a long-duration nominal bond sleeve where relevant. |
| US Cash / T-Bills | Cash / bill reserve proxy | FRED | 1970-2024 in the current annual dataset | Used directly as the short-duration liquidity anchor. |
| Gold | Monetary metal reserve proxy | Damodaran | 1970-2024 in the current annual dataset | Used directly as the core monetary-metal series in modern comparisons. |
| Broad Commodities (futures TR) | Broad commodity futures total-return proxy | Modern futures history with earlier reconstructed extension | Pre-1991 values are backfilled or reconstructed index history, as stated in the source note. | |
| BTC | Digital scarcity proxy | Coin Metrics (PriceUSD) | Bitcoin-era price history only | Used where crypto exposure is present; there is no pre-Bitcoin direct history. |
| Spain Stocks | Spanish equity lens | Jordà-Schularick-Taylor Macrohistory Database | 1970-2024 in local historical currency | Used directly for Spanish equity comparisons in the current window. |
| Spain Housing / Property | Spain housing / property proxy | Jordà-Schularick-Taylor Macrohistory Database | 1970-2024 in local historical currency | Treated as a domestic housing/property total-return proxy, not as a like-for-like listed REIT series. |
| US REITs / Listed Real Estate | US listed real-estate / REIT proxy | FTSE NAREIT All Equity REITs TR USD | 1972-2024 in USD listed real-estate total-return form | Used for US listed real-estate / REIT exposure. In PFAtlas this is the market-traded property lens, not a direct housing or land series, and not a global REIT history. |
| US Housing / Property | US housing / property proxy | Jordà-Schularick-Taylor Macrohistory Database | 1970-2024 in USD total-return proxy form | Used for housing, land, estate and direct-property sleeves when a US-base housing/property lens is more appropriate than listed REITs. |
| CPI | US inflation lens | Jordà-Schularick-Taylor Macrohistory Database | 1970-2024 in the current annual dataset | Used when real returns are shown from a US investor perspective. |
| CPI | Spain inflation lens | Jordà-Schularick-Taylor Macrohistory Database | 1970-2024 in the current annual dataset | Used when real returns are shown from a Spanish investor perspective. |
| FX | EUR / USD currency lens | EUR-era data plus synthetic ECU extension | 1971-1998 uses ECU/USD as a synthetic euro proxy; 1999 onward uses EUR/USD. | |
| Spain FX to USD | Spain local-currency to USD lens | FRED / Penn World Table | 1970-2000 peseta history, then EUR-linked continuation | 1970-2000 uses peseta-per-USD history converted to annual returns; 2001 onward links to EUR/USD. |
Series
US Stocks
PFAtlas role
US equity growth proxy
Source / proxy label
Damodaran
Direct coverage
1970-2024 in the current annual dataset
Backtest treatment / note
Used directly in the current comparison window.
Series
MSCI World
PFAtlas role
Global developed-equity proxy
Source / proxy label
MSCI
Direct coverage
1970-2024 in the current annual dataset
Backtest treatment / note
Used as the developed-world equity anchor where applicable.
Series
ACWI
PFAtlas role
Global all-country equity proxy
Source / proxy label
MSCI
Direct coverage
2001-2024 direct ACWI-linked coverage
Backtest treatment / note
Pre-2001 is extended using MSCI World plus emerging-market weights, as noted in the series metadata.
Series
US 10Y Treasury Bonds
PFAtlas role
Nominal sovereign-bond proxy
Source / proxy label
Damodaran
Direct coverage
1970-2024 in the current annual dataset
Backtest treatment / note
Used directly as a long-duration nominal bond sleeve where relevant.
Series
US Cash / T-Bills
PFAtlas role
Cash / bill reserve proxy
Source / proxy label
FRED
Direct coverage
1970-2024 in the current annual dataset
Backtest treatment / note
Used directly as the short-duration liquidity anchor.
Series
Gold
PFAtlas role
Monetary metal reserve proxy
Source / proxy label
Damodaran
Direct coverage
1970-2024 in the current annual dataset
Backtest treatment / note
Used directly as the core monetary-metal series in modern comparisons.
Series
Broad Commodities (futures TR)
PFAtlas role
Broad commodity futures total-return proxy
Source / proxy label
Direct coverage
Modern futures history with earlier reconstructed extension
Backtest treatment / note
Pre-1991 values are backfilled or reconstructed index history, as stated in the source note.
Series
BTC
PFAtlas role
Digital scarcity proxy
Source / proxy label
Coin Metrics (PriceUSD)
Direct coverage
Bitcoin-era price history only
Backtest treatment / note
Used where crypto exposure is present; there is no pre-Bitcoin direct history.
Series
Spain Stocks
PFAtlas role
Spanish equity lens
Source / proxy label
Jordà-Schularick-Taylor Macrohistory Database
Direct coverage
1970-2024 in local historical currency
Backtest treatment / note
Used directly for Spanish equity comparisons in the current window.
Series
Spain Housing / Property
PFAtlas role
Spain housing / property proxy
Source / proxy label
Jordà-Schularick-Taylor Macrohistory Database
Direct coverage
1970-2024 in local historical currency
Backtest treatment / note
Treated as a domestic housing/property total-return proxy, not as a like-for-like listed REIT series.
Series
US REITs / Listed Real Estate
PFAtlas role
US listed real-estate / REIT proxy
Source / proxy label
FTSE NAREIT All Equity REITs TR USD
Direct coverage
1972-2024 in USD listed real-estate total-return form
Backtest treatment / note
Used for US listed real-estate / REIT exposure. In PFAtlas this is the market-traded property lens, not a direct housing or land series, and not a global REIT history.
Series
US Housing / Property
PFAtlas role
US housing / property proxy
Source / proxy label
Jordà-Schularick-Taylor Macrohistory Database
Direct coverage
1970-2024 in USD total-return proxy form
Backtest treatment / note
Used for housing, land, estate and direct-property sleeves when a US-base housing/property lens is more appropriate than listed REITs.
Series
CPI
PFAtlas role
US inflation lens
Source / proxy label
Jordà-Schularick-Taylor Macrohistory Database
Direct coverage
1970-2024 in the current annual dataset
Backtest treatment / note
Used when real returns are shown from a US investor perspective.
Series
CPI
PFAtlas role
Spain inflation lens
Source / proxy label
Jordà-Schularick-Taylor Macrohistory Database
Direct coverage
1970-2024 in the current annual dataset
Backtest treatment / note
Used when real returns are shown from a Spanish investor perspective.
Series
FX
PFAtlas role
EUR / USD currency lens
Source / proxy label
Direct coverage
EUR-era data plus synthetic ECU extension
Backtest treatment / note
1971-1998 uses ECU/USD as a synthetic euro proxy; 1999 onward uses EUR/USD.
Series
Spain FX to USD
PFAtlas role
Spain local-currency to USD lens
Source / proxy label
FRED / Penn World Table
Direct coverage
1970-2000 peseta history, then EUR-linked continuation
Backtest treatment / note
1970-2000 uses peseta-per-USD history converted to annual returns; 2001 onward links to EUR/USD.
Performance and assumptions
Comparison window, direct coverage and backtest treatment
Portfolio performance depends on assumptions: rebalancing frequency, start date, asset availability, inflation adjustment, currency treatment and missing-data handling.
PFAtlas distinguishes between the comparison window and direct source coverage. Some series are available directly across the full period. Others are extended, linked or proxied to preserve comparability across portfolios.
- backtest window
- direct source coverage
- rebalancing assumptions
- nominal vs real returns
- currency basis
- inflation basis
- unavailable asset series
- proxy substitutions
Investor lens
Country, currency and inflation
Nominal returns show the growth of money. Real returns show the growth of purchasing power. Currency treatment matters because the same portfolio can produce different outcomes for a US, Spanish or global investor.
Where the data allows it, PFAtlas distinguishes investor perspective by country, currency and inflation treatment rather than assuming a single universal reporting lens.
Chapter
Limits of the model
Limits
Known limitations
The atlas is a model of portfolios, not a perfect reconstruction of financial history. Proxy choices simplify reality. Historical returns do not predict future returns. Some portfolios are better understood as intellectual systems than as directly investable strategies.
PFAtlas cannot fully model taxes, transaction costs, leverage, illiquidity, confiscation risk, political privilege, private-market access or behavioral constraints unless those assumptions are explicitly included.
Responsible reading
How PFAtlas should be read
Use PFAtlas to compare structures, ask better questions and identify hidden assumptions. Do not treat a high historical return as a recommendation, or a historical proxy as proof that a system will work in the future.
The most useful question is not "Which portfolio won?" but "What future does this portfolio require in order to work?"
Related routes