Each study here answers one question that comes up across the guides, such as how often Treasury bills beat inflation or how deep bitcoin's crashes have run, from public data. This page explains how the studies are built, dated and licensed, then sums up each one by topic; the full list follows at the end.
What is in each study?
Every study follows one layout, so they can all be read and checked the same way:
- A summary with the dated headline number, then a table of key numbers.
- A method table defining each measure, with the data sources and date range.
- A results section for each question, built around a table or chart.
- The limitations that apply, such as daily closes missing intraday moves or price returns leaving out dividends.
- Reuse terms and a suggested credit line.
The main table of each study downloads as a CSV from a fixed address, for example /data/t-bill-yields-vs-inflation.csv. The page and the download read the same file, so they always match.
Where does the data come from?
The studies are built from public sources, mostly official statistics and exchange price feeds. Each study's own page names the exact series it uses and links to them.
| Source | What it supplies | Used in |
|---|---|---|
| FRED, the St. Louis Fed's database | 3-month Treasury yield, CPI-U, Treasury yield spreads, the Fed's target rate, recession dates, the S&P 500 price index and the broad dollar index | T-bills vs inflation, yield curve inversions, Fed rate cuts, bitcoin vs the S&P 500, the US dollar |
| Robert Shiller's stock market data | Monthly S&P 500 prices, averaged from daily closes, and dividends | Fed rate cuts |
| World Bank Remittance Prices Worldwide | Sender fees and exchange-rate margins by corridor | Cost of sending $200 |
| World Bank World Development Indicators | Personal remittances received, in dollars and as a share of GDP | Remittances by country |
| European Central Bank, via Frankfurter | Daily euro reference rates, converted to dollar rates | US dollar vs world currencies |
| Bitstamp and Coinbase Exchange | Daily closing prices for bitcoin and five other coins | The four bitcoin and crypto return studies |
| DefiLlama | One daily price per stablecoin | Stablecoin depegs |
| Twelve Data, stored by the site | Split-adjusted daily closes for five stocks and two ETFs | Large-cap stock volatility |
How are the snapshots dated and refreshed?
Eleven studies are dated snapshots: a build script pulls each source, calculates the results and stamps them with the last date the data covers. The figures change only when the script is re-run and the study's text is rechecked against the new output.
Each source updates on its own schedule, so the studies end on different dates:
- Daily market series end between October 28 and November 1, 2026, and Shiller's monthly S&P 500 data runs through October 2026.
- CPI is monthly, so the T-bill study ends with September 2026; the next CPI release is due November 10, 2026. BLS published no CPI for October 2025, so that month and the 12-month periods that start or end on it are left out.
- The World Bank's latest corridor survey is for Q3 2025, with prices collected from July 22 to September 10, 2025.
- Remittance inflows are annual, and most countries' latest figures are for 2025.
The twelfth, on large-cap stock volatility, has no snapshot: it is recalculated from the daily closes the site stores, and its tables show an “unavailable” note until that price history is complete.
Worked example: Rebuilding September 2026's real T-bill yield from the CSV
The T-bill study's last row, for September 2026, shows a 3-month Treasury yield of 4.09% (the month's average) and CPI inflation over the previous 12 months of 3.74%. The method defines the real yield as (1 + yield) ÷ (1 + inflation) − 1:
- 1.0409 ÷ 1.0374 = 1.00337
- 1.00337 − 1 = 0.00337, or 0.337%, which the table rounds to 0.34%
Subtracting instead gives 4.09 − 3.74 = 0.35 points. That shortcut is close when inflation is low, but the gap between the two grows when inflation runs high.
How can you reuse and cite a study?
The study tables and CSVs are published under the Creative Commons Attribution 4.0 International license, CC BY 4.0. Anyone may copy, redistribute, adapt and build on them for any purpose, commercial or not. The license deed sets three conditions: give appropriate credit, link to the license and say what was changed, such as trimmed rows or a recalculated column.
A credit line names the analysis, its sources and the data's end date, placed next to links to the study and the license. For the yield curve study it reads: “Source: Sonic Wallet analysis of FRED daily Treasury spreads (T10Y2Y and T10Y3M) and NBER business-cycle dates, data through October 30, 2026.”
The license covers the analysis, not the raw inputs. The underlying series remain under their providers' terms, and some restrict reuse: S&P Dow Jones Indices, for one, does not allow its index values to be reproduced without permission. The editorial policy explains how studies are built, checked and corrected.
Crypto: returns, crashes, timing and stablecoins
Five studies use daily prices to test claims readers meet in the crypto guides.
| Study | Question it answers | Headline number | Data through |
|---|---|---|---|
| Bitcoin vs the S&P 500 | How have bitcoin's returns and risk compared with the S&P 500's over 1, 3, 5 and 10 years? | 61.0% a year for bitcoin and 13.8% for the S&P 500 price index over 10 years, with volatility of 68.3% and 18.1% | October 28, 2026 |
| Bitcoin drawdowns history | How deep have bitcoin's falls of 50% or more been, and how long did recoveries take? | 5 complete declines averaging 73.8%, and a median 469 days from the low to a new record | November 1, 2026 |
| Bitcoin DCA vs lump sum | Did investing all at once beat spreading purchases over 12 months? | The lump sum beat monthly buying in 68.3% of 167 start months, measured a year later | November 1, 2026 |
| Crypto returns by day of the week | Does any weekday reliably earn more for six major coins? | 2 of 42 coin-weekday averages had a t-statistic of 2 or more either way, against 2.1 expected by chance | October 31, 2026 |
| Stablecoin depeg history | How often have major stablecoins traded away from $1? | USDT was more than 0.5% off on 48 of 2,127 days, and USDC on 53 | November 1, 2026 |
Bitcoin's latest close in the drawdown study, $78,233 on November 1, 2026, was 37.3% below its $124,728 record of October 6, 2025. Two caveats matter here: the S&P 500 side of the bitcoin comparison leaves out dividends, and one price a day misses intraday extremes, such as USDC's slide on March 11, 2023. Live prices sit on the bitcoin price page.
Stocks and the Fed
Two studies look at the stock market: one at how the S&P 500 moved after Fed rate cuts, the other at how much the largest companies swing.
| Study | Question it answers | Headline number | Data through |
|---|---|---|---|
| S&P 500 after Fed rate cuts | How did the S&P 500 move after the first rate cut of each easing cycle since 1984? | Up 8.7% on average 12 months after the first cut, and higher in 7 of 9 cycles | October 2026 |
| How volatile are the biggest stocks? | How much have Apple, Microsoft, Nvidia, Alphabet and Tesla swung next to the SPY and QQQ funds? | None fixed: recalculated from daily closes | The latest stored close |
The 8.7% average hides a split in the same data: 16.7% across the five cycles with no recession within two years, against −1.2% across the four where one began. Returns run between monthly averages of daily closes, so they approximate close-to-close moves rather than match them. The stocks hub covers investing through index funds, and the SPY page charts the S&P 500 fund's price and returns.
Treasuries and inflation
Two studies cover what Treasury bills have earned after inflation and what the yield curve has signaled.
| Study | Question it answers | Headline number | Data through |
|---|---|---|---|
| T-bill yields vs inflation | How often have 3-month Treasury bills beaten CPI inflation since 2000? | 38.2% of 12-month periods; the real yield was 0.34% in September 2026 | September 2026 |
| Yield curve inversions history | How long and deep was each inversion since 1980, and did a recession follow? | 10 inversions of the 10-year minus 2-year spread, 7 followed by a recession within 2 years; 9 of the 10-year minus 3-month spread since 1982, 6 followed | October 30, 2026 |
The real yield was negative in 194 of the 320 months, including a 62-month run from November 2009 to December 2014. The longest inversion, 1,095 days of the 10-year minus 3-month spread from October 18, 2022, was followed by no recession within two years, and on October 30, 2026 both spreads were positive, at 0.38 and 0.78 percentage points. The Treasuries hub covers how bills work, and the T-bill calculator turns a discount rate into a purchase price and yield.
Money transfer and currencies
Three studies cover what it costs to send money abroad, where remittances go and how far the dollar has moved.
| Study | Question it answers | Headline number | Data through |
|---|---|---|---|
| Cost of sending $200 by corridor | What does it cost to send $200, or a fixed amount in the sender's currency, in 48 corridors? | 5.28% on average; 2 of 48 corridors at or below the UN's 3% target | Q3 2025 survey |
| Remittances by country | Which countries receive the most remittances, in dollars and as a share of GDP? | India received $150.7 billion; inflows equaled 57.7% of Tajikistan's GDP | Mostly 2025 |
| US dollar vs world currencies | How far has the dollar moved against 29 currencies over 1, 5 and 10 years? | Stronger against 17 of 28 currencies over 10 years; the Fed's broad dollar index rose 6.5% | October 30, 2026 |
The corridor figure is a simple average across the 48 corridors, not the World Bank's weighted global average, and corridors from the UAE, Saudi Arabia, Canada and Australia are priced in local currency, so their percentages compare but their dollar costs are not shown. The remittance table adds the 187 countries with data to $986 billion, 50.5% of it going to the top 10; Vietnam is absent because the World Bank has no figure for it after 2004, not because it receives little. The money transfer hub shows how to price a transfer yourself, and the currency converter shows today's exchange rates.
What can't these studies tell you?
They describe particular windows and samples, and past performance does not predict future results. Samples are often small, such as 9 easing cycles or 19 inversions across two spreads; price histories start when an exchange's records do; and testing 42 coin-weekday pairs turns up about two “significant” averages by chance alone. None of the studies is advice: they set out what the data recorded so readers can weigh it. The investing glossary defines the measures behind them, from drawdown to the Sharpe ratio.