The numbers behind the assets are a personal collection compiled from many reputable public sources including annual fund reports, index providers, academic research, and a good number of my own calculations. The goal is to present accurate, consistent, and actionable data that dutifully tracks real-world performance.
All data tracks the total return that includes both the capital appreciation of the asset and any reinvested dividends or interest.
Data Prioritization
All of the returns data follows the same basic sourcing hierarchy:
1. Index data: These historical numbers come from the underlying ideal models that index funds are designed to track. You may also see them used as benchmarks in various websites and charting tools that compare a fund to its index. This is the cleanest data free of costs, fees, and methodologies unique to specific funds.
2. Index fund returns: These are total return numbers that come directly from the annual reports of the funds. One representative fund is chosen per asset, and I refund the expense ratio to study the return before fees.
3. Academic resources: These are numbers provided by investing researchers who study historical data and publish their findings. I also include my own stock and bond calculations in this category.
4. Replacement assets: When no direct or indirect data is available, I substitute similar-but-different assets to provide a bit of historical context. Only a few assets fall into this category, and are mostly small and value stocks outside of the US prior to 1975. So for example, when no ex-US small cap value data is available I substitute ex-US large cap blend. This only affects the very oldest data, and the returns should still be in the same ballpark in normal portfolio percentages.
Here’s how it all comes together. Once a year in early January, I assemble all of my various public data sources in one spot (all carefully selected to track the same index definitions) and start going down the priority list from 1 to 4. For each asset and year, I use the highest-ranked source that is available. The end result is a composite history representing the most accurate and comprehensive collection of free data that I can assemble.
Index Methodologies
The “index” that index funds track is a standardized bucket of securities which follows set rules defined by an index provider like MSCI, FTSE, or Bloomberg. Every index methodology is a little different, and there is no one ideal index construction. But all share similar goals, and standard indices using the same terms like “large cap blend” or “small cap value” are usually on the same page in terms of returns within a reasonable margin of error.
For consistency in the historical data, the Portfolio Charts prefers sources that follow the same standard methodology rules common in real-world index funds. That may sound like an obvious choice, but some academic sources quote numbers that work well for things like factor research but not for backtesting the behavior of funds you can buy.
These are the definitions I look for:
Stocks
When available, stock data follows capitalization-based breakpoints rather than fixed company counts.
| Stock Type | Capitalization Breakpoints |
|---|---|
| Large Cap | 0-85% |
| Small Cap | 85-98% |
Since mid caps are commonly defined as the capitalization band between 70-85%, the large cap data also includes mid caps. While this may sound strange, that’s how most large cap funds work in the real world. Companies like MSCI refer to this as the “standard” index, and it’s why most large cap index funds explicitly say they cover large and mid-cap stocks.
If you’re wondering why small caps stop at 98%, that’s also an industry standard. The very smallest companies from 98-100% are more commonly classified as micro caps.
For stock valuation, “value” means the half of the market with the lowest valuations at any given time and “growth” means the half of the market with the highest valuations. “Blend” is the entire market including both growth and value. Relative valuation is a little counter-intuitive for deep value stock pickers who think in absolute terms or researchers looking at fixed deciles in the Fama-French research, but it’s the more accurate way to think about value index investing.
Bonds
Portfolio Charts bonds are all issued by government treasuries and do not include corporate, agency, or municipal bonds. All numbers are unhedged. The bond data uses standard definitions based on ranges of bond maturities.
| Bond Type | Maturity Range |
|---|---|
| Bills | 0-1 year |
| Short Term | 1-3 years |
| Intermediate Term | 3-10 years |
| Long Term | 10-30 years |
More Info
To learn more about how index construction works, I recommend the following pages:
- Investing With Style: How Size and Value Actually Work: Deep explanations of stock index methodologies
- High Profits at Low Rates: The Benefits of Bond Convexity: Longer explanation of how bond maturities and interest rates affect performance
- Each Assets page includes details on how the asset is defined.
- Also see the Portfolio Charts Calculations section below to learn how many of the older numbers are calculated.
But if you take just one thing away from the index methodologies section, let it be this:
Portfolio Charts data is carefully curated to accurately model the vast majority of applicable real-world ETFs within a reasonable margin of error. So the numbers are actionable and not purely academic.
Representative Index Funds
Collecting data histories is a constant process of continuous improvement, and I do not publish a detailed record of every original source for every year. Just know that I work extremely hard to research and compile quality data histories.
If you’re looking for practical source examples, the Assets section contains a full collection of real-world ETFs that I personally screened for index methodologies reasonably similar to the Portfolio Charts data. Never fully take my word for it, as I do sometimes make mistakes and fund providers do change methodologies from time to time. But the assets collection should be a great place to start to better understand each asset class.
And for older years that modern fund histories do not cover, see the original sources below.
Original Sources
If you’re looking for historical data for your own personal collection, here’s a list of free public resources worth exploring.
Where I’d Start
Simba’s Backtesting Spreadsheet — Collected from various sources around the web and maintained by the Bogleheads community, this is the definitive source of high-quality free asset data for everyday investors. Start with the Data_Series sheet.
Backtest by Curvo — This helpful Europe-centric site maintains a terrific collection of ETF and index data for a wide variety of different assets. This includes things like European bonds and real estate that can be tricky to find elsewhere.
Fund Providers
Index fund issuers are required to publicly report returns for their funds. The simplest way to find this data is to search for a fund on a data aggregator like Morningstar or Google Finance. Another method is to visit the fund provider sites directly. Here are a few of the most prominent:
Note that the lists on each site may vary based on the home country you specify.
Index Providers
MSCI End of Day Index Data Search — The popular index provider that many funds track offers a great tool for searching historical data. Its coverage of international markets is especially thorough.
Morningstar Indexes — The index provider preferred by Vanguard offers annual returns for recent years.
FTSE Historic Index Values — The official source for the most recent 2 years of FTSE index returns.
Industry Associations
Nareit Historical Values & Returns — The organization that promotes REITs in the US supplies thorough historical data going back to 1972.
LBMA Precious Metals Prices — The London Bullion Market Association is the place to go for the historical spot prices of a variety of precious metals.
Government Agencies
OECD — An excellent free resource for all kinds of economic data around the world. I particularly like its data for inflation and short-term bills.
Federal Reserve Economic Data (FRED) — A great source for all kinds of economic data including bond returns and inflation rates.
Federal Reserve Exchange Rates — Exchange rates relative to the USD for any currency you can think of since 1971. When using this data, just pay attention to the format. Some countries report local/USD, while others are USD/local. It’s important to be consistent.
International Monetary Fund — A massive resource for all types of international data. Not everything is applicable to investing, but the bond numbers are particularly thorough.
US Department of the Treasury — My go-to source for detailed yield curves in the United States.
European Central Bank — Among tons of other EU-centric data, they have a really nice collection of Euro area yield curves.
Eurostat — More Europe data, including a few specialty things like old ECU currency calculations before the Euro.
Portfolio Charts Calculations
Stock Index Calculator — My own work reconstructing realistic size and value indices from Fama-French source data using common index fund methodologies.
Bond Index Calculator — My personal tool for reconstructing bond index returns from the underlying interest rates.
Academic Resources
Fama-French Data Library — A massive database of US and international stock data broken down by various contributing factors. It requires a bit of a learning curve to use, but is extremely thorough and well-sourced.
Independence International Associates — A remarkable collection of all types of worldwide stock data from 1975-1996.
Shiller Data — Robert Shiller’s aggregate US Stock, bond, and inflation dating back to 1871.
JST Macrohistory Database — An incredible collection of a wide variety of economic data for 17 different countries from 1870-2015. This is the source data from “The Rate of Return on Everything, 1870–2015.” in the Quarterly Journal of Economics.
Crestmont Research Stock Market Matrix — The original inspiration for the Portfolio Charts Heat Map, this stock market history is packed full of useful information.
Fair Use
Portfolio Charts is educational in nature and my goal is simply to help investors make informed choices. In order to build simulated asset histories to model historical portfolio performance, I make use of a variety of public data freely found on the internet. I respect copyright and take several proactive steps to honor the rights of data providers:
Light Touch
Each asset history is an aggregate collection of purely factual information from multiple independent sources. I use only annual total returns — a single number per asset per year — not the granular daily data that providers license commercially. The limited asset classes I study are also just a tiny fraction of larger data collections. My focus is on asset allocation, not individual funds or securities.
Creative Application
All outside data is modified through inflation adjustments, exchange rates, and other calculations before being further transformed into unique visualizations to help tell a story. Nothing is copied in its original form.
No Redistribution
Raw source data is not available anywhere on the site. The online tools only serve calculated results that have already been transformed, so there are no original figures to copy in the first place.
Public Attribution
I openly credit the people and organizations whose work makes this research possible. The Original Sources section above links directly to my favorite resources, and I am always happy to point curious researchers to reputable providers.
Non-Commercial Use
Portfolio Charts is not in the data business. I give research away for free as part of the larger education mission. It is supported by the generosity of patrons, but they do not receive additional portfolio data in return beyond what is already available for everyone.
If you have any questions, please contact me. I take this seriously and will do everything I can to address any concerns.
Disclaimer
Never assume the data here is completely accurate. Not only may there be a few mistakes in the numbers, but the historical data is also updated from time to time by the primary sources as more information becomes available.
All numbers have been modified from their original form, and the results of your own fund may vary from what you see here. The goal is simply to be reasonably accurate for general portfolio backtesting purposes.
All data is from 1970 to the present, which is the most freely available for such a wide variety of assets. Past performance is no guarantee of future returns.