Return Data Files

Birju Patel

This project consolidates useful stock, bond, and currency return data.

50+ years of coverage for interest rates, bond returns, currency exchange rates, and more.

Some returns are reported in excess of the risk-free rate. An excess return represents the returns to a zero-cost strategy that borrows at a short-term interest rate to buy the target asset.

For an unhedged series, the strategy borrows in U.S. dollars and converts to foreign currency. For a hedged series, the strategy borrows in the foreign currency. All returns are measured in U.S. dollars.

Equities

• developed_ex_us.csv

  • Frequency: Annual
  • ex_us_hedged_er: Currency hedged USD index excess return
  • ex_us_unhedged_er: Unhedged USD index excess return

Annual foreign equity index returns intended to approximate the MSCI EAFE Index.

Methodology

• 2015 to 2025
  • Uses returns from publicly traded index ETFs tracking the MSCI EAFE Index.
• 1975 to 2014
  • Uses a weighted sum of currency-adjusted excess returns for developed-country equity markets excluding the United States and Canada.
  • Country weights are based on each country’s share of total equity market capitalization within the investable universe.
• 1950 to 1974
  • Uses the average country weights from the 1975 to 2014 period.

Bonds

• sov_bonds.csv

  • Frequency: Monthly
  • erHedgedUSD: Currency hedged long-term bond excess return
  • erHedgedUSD: Unhedged long-term bond excess return

Returns on 10-year developed-country sovereign bonds.

Methodology

The series approximates monthly returns by using changes in bond yields.

Yields are assumed to be par yields. Return estimates are calculated using the method outlined in Swinkels (2019). Additional implementation details are available on the Portfolio Optimizer blog.

Computed returns are not suitable for testing momentum-based strategies. Monthly yields are measured as within-month averages, which mechanically smooths sharp yield changes across adjacent months. This creates spurious autocorrelation in monthly returns and materially overstates the returns to momentum-based trading strategies.

• yields3M.csv

  • Frequency: Monthly
  • yield3M: Short-term interest rate

Methodology

The short-term interest rate is either the central bank policy rate or 3 month government bill yield, whichever has the most available data.

• yields10Y.csv

  • Frequency: Monthly
  • yield10Y: Long-term interest rate

Methodology

Long-term yields are 10-year sovereign bond par yields. Monthly yields are reported as the average daily yield over that month.

Currencies

• fx_daily.csv

  • Frequency: Daily
  • USDPerForeign: Value in USD of single unit of foreign currency
  • priceReturn: USD price return of foreign currency
  • fxExcessReturn: Excess return of foreign currency

Daily currency exchange rates, price fluctuations, and excess returns.

Methodology

Measurements are point-in-time estimates, recording market prices observed in the NYC FX market at 12:00 PM EST. This series is suitable for testing momentum strategies.

Foreign currency excess return is the return to borrowing USD and lending foreign currency at the U.S. and foreign country risk-free rate.

Other

• US_CPI.csv

  • Frequency: Monthly
  • CPI: Unadjusted Consumer Price Index for All Urban Consumers

Utility Scripts

Install the Python dependencies with:

pip install -r requirements.txt

• transform.py

  • Interactive tool for creating derived return series in custom/.
  • Select a source return series, then choose Real total return or Nominal total return.
  • Nominal total return adds the U.S. 3-month risk-free return from bonds/yields3M.csv.
  • Real total return converts to nominal total return, then inflation-adjusts with US_CPI.csv.
  • Output file names and transformed column names reflect the selected transformation.

Run interactively:

python transform.py

List available source series or run directly:

python transform.py --list
python transform.py --input currencies/fx_daily.csv --column fxExcessReturn --transform nominal_total
python transform.py --input equities/developed_ex_us.csv --column ex_us_hedged_er --transform real_total

• plot.py

  • Interactive tool for plotting log cumulative nominal returns to screen.
  • Supports the built-in stock, bond, and currency series only; custom/ series are not included.
  • Prompts let you select multiple assets, then choose Plot.
  • Charts display with matplotlib; no PNG file is saved.

Run interactively:

python plot.py

Additional Resources

High-frequency, long-history equity return data may be found at the Ken French Data Library. For example, daily U.S. equity market excess returns are reported from 1926 to 2026. Returns on the Fama-French cross-sectional equity factor portfolios, regional portfolios, industry portfolios, and other useful sorts are also reported here.

These returns can be easily integrated into Python workflows using the Pandas DataReader library.

Sources

• World Bank Group
  • Market capitalization of listed domestic companies, current US dollars: CM.MKT.LCAP.CD
• Organisation for Economic Co-operation and Development (OECD)
  • Long-term interest rates: IRLT
  • Short-term interest rates: IR3TIB, IRSTCI
• Federal Reserve
  • Spot exchange rates (monthly): EXUSEU, EXUSUK, EXUSAL, EXCAUS, EXJPUS, EXSZUS, EXFRUS, EXGEUS
  • Spot exchange rates (daily): DEXUSEU, DEXJPUS, DEXUSUK, DEXCAUS, DEXUSAL, DEXSZUS, DEXNOUS, DEXSDUS, DEXUSNZ
• JST Macrohistory Database
  • Equity total return: eq_tr
  • USD exchange rate: xrusd
  • Short term interest rate: bill_rate
• Bureau of Labor Statistics
  • Unadjusted Consumer Price Index for All Urban Consumers: CPIAUCNS

Licensing

This dataset is made available under the Creative Commons Zero v1.0 Universal Public Domain Dedication (CC0-1.0).

The datasets are approximations of the returns of target assets. The data is provided as-is, with no guarantee of accuracy or completeness.

This dataset is provided for research and educational purposes only. It is not investment advice.