Do T-Bills Beat Inflation? Barely Before Tax, Usually Not After

Before tax, yes, by a thin margin. From 1928 through 2025, rolling 3-month Treasury bills returned 3.37% a year while consumer prices rose 3.04% a year, a real return of 0.33% a year.¹ The longer Dimson, Marsh and Staunton series puts U.S. bills at 0.5% a year above inflation for 1900-2025.²

After federal income tax, the answer mostly reverses. Taxing each year’s interest at a flat 10% would have erased the entire 1928-2025 real return. At a flat 24%, a dollar of purchasing power held in bills for those 98 years would have shrunk to 64 cents.

There is also a third question that the historical series does not answer: whether your own cash earned the T-bill rate. The average bank deposit did not. In the FDIC’s end-of-August 2026 survey, the national average savings rate was 0.37% while the comparable Treasury yield was 3.63%.³

T-bills still do one job very well: holding nominal dollars you will need soon. Whether they suit money you will need in 20 years depends on the record, on your tax rate, and on what you are trying to protect.

98 years of T-bills against inflation

Damodaran’s annual dataset is the standard free source for this comparison. His T-bill return for each year is the average 3-month bill rate during that year, taken from the Federal Reserve’s series from 1934 onward, and his inflation figure is the December-to-December change in CPI-U.¹ Rolling bills this way is what a Treasury money market fund or a 0-3 month Treasury ETF approximates today, before fees.

The UBS Global Investment Returns Yearbook, built on the Dimson-Marsh-Staunton database, starts in 1900 and reaches a slightly higher figure for bills.² Its comparison with stocks and long bonds over the same 126 years shows the size of the gap between preserving purchasing power and growing it:

The two sources differ by under 0.2 points a year because they start in different years and build the bill and inflation series differently. Both put the long-run real return on U.S. bills below 1%. The U.S. is also a favorable case: the Yearbook notes it had the third-lowest inflation of the 35 markets in the database.² If you want the full argument for why a century of returns is a base rate rather than a forecast, Do 200 Years of Stock Returns Still Matter? covers it.

The average hides long losing stretches

A 0.33% average built from years that ranged from -15% to +13% says little about any one investor’s experience. Grouped by decade, bills lost to inflation in four of the nine full decades, and in the partial 2020s so far:

The 1940s were policy

The worst stretch followed from policy. In April 1942 the Federal Reserve, at the Treasury’s request, committed to hold 3-month bill yields at 3/8 of a percent to cheapen war financing, and kept the peg in place after the war while CPI inflation ran 17.6% from June 1946 to June 1947.⁴ In Damodaran’s data, bills returned 0.38% in 1946 against 18.1% inflation, a real loss of 15.0%. From 1941 through 1951 bills lost 4.9% a year in real terms, cutting purchasing power by 43%. The Treasury-Fed Accord of March 1951 ended the Fed’s commitment to fix Treasury rates.

Reinhart and Sbrancia call this financial repression: capped interest rates plus steady inflation that eroded the real value of government debt. They estimate that negative real rates liquidated 3% to 4% of GDP of U.S. and U.K. government debt per year, and that advanced economies had negative real interest rates roughly half the time from 1945 to 1980.⁵ Bill holders were the lenders on the other side of that liquidation. The 2010s and 2021-2022 rhymed on a smaller scale: with bills near zero, 2021 inflation of 7.0% produced a real loss of 6.5%.

The 1980s were also policy

The best decade came from the opposite stance. With the Fed holding short rates well above inflation to bring it down, bills returned 3.7% a year above inflation in the 1980s. The best 10-year window in the sample, 1981-1990, earned 4.0% a year real. More recently, bills beat inflation in each of 2023, 2024 and 2025 (by 1.9, 2.2 and 1.4 points). In both directions, real bill returns followed the Fed.

How often bills kept up with inflation

Decades are arbitrary. A cleaner test takes every overlapping holding period of a given length in the 1928-2025 data and asks whether bills beat inflation over it:

Before tax, about one 10-year holding period in two beat inflation, and about one 30-year period in three lost to it. Longer horizons did not make bills reliable; they only narrowed the range, from -5.0% to +4.0% a year for 10-year windows to -1.65% to +1.97% for 30-year windows. The most recent 30-year window, 1996-2025, lost 0.19% a year to inflation before any tax. After tax at 24%, three out of four 30-year periods lost purchasing power.

Taxes fall on the inflation part of the yield

T-bill interest is subject to federal income tax and exempt from state and local income tax.⁶ The federal tax applies to the whole nominal yield, including the part that only compensates you for inflation. That makes the after-tax real return shrink as inflation rises, even when the pre-tax real return stays the same:

Martin Feldstein worked out the theory in 1976: if lenders are to keep the same after-tax real return, the nominal rate has to rise by more than one point for each point of inflation, by 1/(1 - t).⁷ The data did not cooperate. Feldstein and Summers found that U.S. interest rates rose roughly one-for-one with inflation from 1954 to 1976, which meant, in their words, that “the real interest rate net of tax available to investors is reduced dramatically by inflation.” Their example: a lender in a 50% bracket earning 3% real with no inflation would see that net return “reduced to zero by a 6 percent inflation.”⁸

What that did to the historical record

Applying a flat federal rate to every year’s interest in the 1928-2025 data:

Treat this as a thought experiment. Nobody paid one rate for 98 years: brackets changed many times, top rates exceeded 90% in the 1940s and 1950s, and the 3.8% net investment income tax only began in 2013. A study using the rates in force at the time reaches the same conclusion. Jeremy Siegel’s Stocks for the Long Run (2nd edition) computed after-federal-tax real T-bill returns of -0.7% a year for 1913-1997 for an investor in the $50,000 bracket, and -1.6% a year for one in the top bracket.⁹

The 1981-2025 period, the best stretch for cash in the sample, shows how small the margin is. Bills returned 0.93% a year above inflation before tax. At 24%, that fell to 0.02%.

Try your own bracket

Two details change the result for a specific reader. The state exemption is worth real money if you live in a high-tax state and would otherwise hold a bank account, and SGOV vs. HYSA works through that comparison. And inside an IRA or 401(k), the federal tax drag is deferred or, in a Roth, removed, so the before-tax column is the relevant one there. The 3.8% NIIT applies to interest once modified AGI exceeds $200,000 (single) or $250,000 (married filing jointly), and those thresholds are not indexed for inflation.¹⁰

For today’s rates rather than history, the after-tax real yield calculator compares T-bills with TIPS, I bonds and savings accounts at your tax rates.

Why bills track expected inflation and miss surprises

The Fisher relation explains why bills hold up over the long run at all. A nominal short-term rate is approximately the real rate lenders require plus the inflation they expect. If investors expect 3% inflation and want 1% real, a 3-month bill should yield around 4%. Ang, Bekaert and Wei estimate that the unconditional U.S. real rate curve “is fairly flat around 1.3%.”¹¹

Fama and Schwert tested this directly. Over 1953-1971, U.S. government bills and bonds were a complete hedge against expected inflation, while stock returns were negatively related to it.¹² That result is tied to its sample and covers expected inflation only. A bill bought at 4% when the market expects 3% inflation pays 4% whatever inflation turns out to be over the next three months. If inflation jumps to 7%, you lose about 3% annualized in real terms on that bill, and only the next bill can reprice.

So rolling bills adapt to inflation with a lag, and only as fast as the market and the Fed move short rates. TIPS work differently: their principal is indexed to CPI.¹³ When the Fed keeps short rates pinned while inflation runs, as in the 1940s or 2021, the adjustment does not happen at all.

“Risk-free” depends on the horizon

Finance textbooks call the T-bill the risk-free asset. For a three-month horizon and a liability in nominal dollars, that is close to true: no meaningful credit risk, almost no price risk, and you know exactly what you will receive.

Over 30 years, an investor rolling bills knows neither the yield on any bill after the first one nor the inflation each one will face. Campbell and Viceira showed that conservative long-horizon investors should hedge the risk of falling real rates, and that long inflation-indexed bonds are the most suitable asset for it.¹⁴ Campbell, Shiller and Viceira put the conclusion plainly: “because short-term real interest rates vary over time, Treasury bills are not safe assets for long-term investors.”¹⁵ A rolled bill swaps price risk for reinvestment risk. The 1941-1951 and 2009-2021 periods are what that reinvestment risk looks like.

The safe asset is the one that matches the liability:

A matched TIPS has its own costs. Its market price moves with real yields, which matters if you have to sell early. In a taxable account the annual inflation adjustment to principal is reported as income each year even though you are not paid it until maturity.¹⁶ At maturity you receive the greater of the inflation-adjusted or original principal.¹³ On September 30, 2026, a 20-year TIPS yielded 3.18% above inflation.¹⁷ Building a TIPS ladder covers the mechanics. I bonds are another route, with federal tax deferred until redemption but a $10,000 annual purchase limit per person and a 12-month lockup.¹⁸

Most household cash earns less than T-bills

Every number above describes T-bills. Much of the money Americans call cash sits in bank accounts, and banks pay less. The FDIC’s deposit-weighted national rates, posted September 21, 2026 from end-of-August data:³

A national average weights each bank by its deposits, so it is pulled toward the large banks that hold most of them; many online banks pay several times as much. But the gap is structural. Drechsler, Savov and Schnabl show that banks use market power over depositors to widen deposit spreads when the Fed raises rates: the average spread beta in their data is 0.54, so deposit rates capture roughly half of a rate increase.¹⁹ In the second quarter of 2026, the industry’s whole cost of funding (deposits plus other borrowing) was 2.03% of earning assets, while 13-week bills averaged about 3.7%.²⁰

Two Fed surveys describe where households keep it. In the Fed’s 2022 Survey of Consumer Finances, 98.6% of families had a transaction account (a category that includes money market funds and brokerage cash, not only bank accounts), with a median balance of $8,000. Only 1.1% held bonds directly, the category that includes T-bills, and 6.4% held savings bonds.²¹ In the Fed’s 2025 household survey, 59% of adults had a savings account, money market account or CD.²²

Money market funds now move a large share of cash close to bill rates. Households and nonprofits held $5.33 trillion in money market funds at mid-2026, against $14.86 trillion in deposits and currency, so about a quarter of their cash earned something near the bill rate.²³ Total money market fund assets were $7.94 trillion in late September 2026, $3.11 trillion of it in retail funds.²⁴ For a saver whose money sits at the national average rate, though, the T-bill record is an upper bound. At 0.37%, and with the TIPS market expecting about 2.36% inflation, that saver loses roughly 2 points of purchasing power a year before tax.

When holding T-bills long term makes sense

A cash allocation can stay in a portfolio for decades while each dollar in it is earmarked for the next year or two. That distinction answers the usual objection that a 30-year horizon rules out cash. Reasonable long-term uses:

• Emergency reserves and operating cash. Their job is to be there at par on a bad day. See how to size one.
• Known spending in the next few years. A house down payment, tuition, or a car is a nominal liability with a date on it. Stocks fell more than 35% in 1931, 1937 and 2008.¹
• A retiree’s spending reserve. One or two years of withdrawals in bills, refilled from the portfolio, so a crash does not force selling. See sequence-of-returns risk.
• Low risk capacity. Someone who cannot absorb a drawdown, or already has enough, can rationally accept a lower expected real return.
• Behavior. If holding some cash keeps you from selling stocks in a panic, the cash is worth more to you than its yield.
• Periods of high real yields. On September 30, 2026, 13-week bills yielded 4.13% and the inflation rate implied by 10-year TIPS was about 2.36%.¹⁷ If inflation runs at that rate, bills would earn about 1.7% real before tax and about 0.8% at 24%, better than most of the historical record. That holds only until the Fed cuts.

When it doesn’t

• Long-run growth money. Bills turned $1 of purchasing power into $1.80 over 126 years; stocks turned it into $3,296.² If the money is for spending 20 years out, bills are the low-volatility choice, but historically they have also been the low-return one.
• A fixed real liability far in the future. A matched TIPS hedges it; rolled bills leave you exposed to whatever real rates the next two decades bring.
• “Dry powder” held indefinitely. Cash held to buy after a crash that has not come is a market-timing position. Its cost is the equity premium forgone while you wait. Lump sum vs. DCA covers the evidence on waiting to invest.
• Large balances in a taxable account in a high bracket. At 32% and above, the historical after-tax real return was below -0.7% a year. If you hold bills for the long term, hold them where the tax does not apply when you can.

What we recommend

Use T-bills, or a Treasury money market fund or 0-3 month Treasury ETF, for money you will spend in nominal dollars within roughly three years, and for a standing reserve of that size. They are among the best instruments available for that job: state-tax-exempt, backed by the Treasury, and repriced every few weeks.

Do not count on them to grow purchasing power. Before tax, their long-run real return has been close to zero, and after tax in most brackets it has been negative. For inflation-adjusted spending more than a few years out, a TIPS maturing that year matches the need. For long-run growth, the evidence favors a diversified stock portfolio sized to what you can hold through a crash.

And check what your cash earns. Moving a balance from a 0.37% savings account to a Treasury money market fund or ETF paying close to the bill rate adds about 3 points of yield for very little added risk.

Key takeaways

• Before tax, T-bills beat inflation by 0.33% a year over 1928-2025 (0.5% over 1900-2025), turning $1 of purchasing power into $1.38.
• The record is uneven. Bills lost to inflation in 40 of 98 years and in about half of all 10-year holding periods. The worst losses came when the Fed held short rates down during high inflation.
• Federal tax erases the margin. A flat 10% rate on each year’s interest wiped out the 1928-2025 real return; at 24%, $1 became $0.64.
• Bills adapt to expected inflation with a lag and do not protect against surprises. TIPS index to CPI directly.
• The average bank deposit earns well below the bill rate: 0.37% on the average savings account against 3.63% on Treasuries in August 2026.
• Match the asset to the job. Bills for near-term nominal spending and reserves, TIPS for dated real spending, stocks for growth.

How Summitward helps

Frequently asked questions

Do T-bills keep up with inflation?

Over the very long run, slightly. From 1928 to 2025 they returned 0.33% a year above CPI inflation before tax. Over shorter periods they often did not: only about half of all 10-year periods beat inflation, and bills lost purchasing power in the 1940s, 1950s, 1970s, 2010s and 2020-2025.

Are T-bills a good inflation hedge?

Partly. Because they mature within a year, their yields catch up with expected inflation as you roll them. They do not protect against inflation that arrives before rates rise, or against a central bank that keeps rates below inflation. TIPS and I bonds are indexed to CPI and hedge inflation directly.

Do T-bills beat inflation after taxes?

Historically, not for most taxpayers. In the 1928-2025 data, a flat federal rate above about 10% on each year’s interest produced a negative real return. Siegel’s calculation with actual historical tax rates found -0.7% a year for 1913-1997 for an investor in the $50,000 bracket. In an IRA or 401(k), the before-tax figures apply.

Is T-bill interest taxed by states?

No. Interest on Treasury bills is subject to federal income tax but exempt from state and local income tax. Treasury ETFs and money market funds pass the exemption through only for the share of income that comes from Treasury securities, which varies by fund, and some states set minimum shares.

What is the real return on T-bills today?

On September 30, 2026, 13-week bills yielded 4.13%. If inflation matches the 2.36% a year implied by 10-year TIPS, that is about 1.7% real before tax and about 0.8% after a 24% federal tax. Both figures depend on that inflation assumption, and bill yields change with every Fed decision.

Are T-bills better than a high-yield savings account?

Often, after tax, for savers in states with an income tax, because bill interest is state-exempt. The best online savings accounts can pay close to bill yields; the national average account does not. SGOV vs. HYSA calculates the break-even APY for your state.

Related guides

• Where to Park Your Cash: SGOV, T-bills, HYSAs and munis by tax bracket and reserve layer
• The Best Inflation Hedges Are Boring: TIPS and I bonds, with an after-tax real yield calculator
• TIPS Ladder: matching inflation-adjusted spending year by year
• Most Stocks Lose to T-Bills. The Market Still Wins.: T-bills as the hurdle for individual stocks
• What Real Return Should You Assume for Stocks?
• Can You Live Off Treasury Interest Forever?

Sources

1. Damodaran, A. “Historical Returns on Stocks, Bonds and Bills: 1928-Current”. NYU Stern. The source for every 1928-2025 figure in this guide, including the decade, rolling-window and flat-tax calculations, which we computed from his annual series. The T-bill return is the average 3-month bill rate over each year (Federal Reserve series TB3MS from 1934; the 1928-1933 rates have no stated source). Inflation is the December-to-December change in CPI-U (for example, 7.0% for 2021 and 18.1% for 1946).
2. Dimson, E., Marsh, P., & Staunton, M. UBS Global Investment Returns Yearbook 2026, Summary Edition. The source for the 1900-2025 U.S. nominal and real returns and terminal real wealth (Figure 12), and for the U.S. having the third-lowest inflation of the 35 markets covered.
3. FDIC. National Rates and Rate Caps, monthly update of September 21, 2026, based on data as of the last business day of August. Rates are averages weighted by each institution’s share of domestic deposits.
4. Federal Reserve History. “The Treasury-Fed Accord”. The source for the 3/8% bill-rate peg from April 1942, the 17.6% CPI inflation from June 1946 to June 1947, and the March 4, 1951 accord.
5. Reinhart, C.M., & Sbrancia, M.B. (2015). “The Liquidation of Government Debt”. Economic Policy, 30(82), 291-333; NBER Working Paper 16893. The source for negative real rates roughly half the time in advanced economies during 1945-1980 and the 3% to 4% of GDP annual liquidation estimate for the U.S. and U.K.
6. TreasuryDirect. Treasury Bills. “Federal tax due on interest earned. No state or local taxes.”
7. Feldstein, M. (1976). “Inflation, Income Taxes, and the Rate of Interest: A Theoretical Analysis.” American Economic Review, 66(5), 809-820.
8. Feldstein, M., & Summers, L. (1978). “Inflation, Tax Rules, and the Long-Term Interest Rate”. Brookings Papers on Economic Activity, 1978(1), 61-109. The source for the roughly one-for-one response of interest rates to inflation over 1954-1976 and both quotations. The paper’s 1/(1 - t) result is stated for the corporate tax rate as a ceiling on what borrowers can pay; the lender-side version is in Feldstein (1976).
9. Siegel, J.J. (1998). Stocks for the Long Run, 2nd ed. McGraw-Hill, Table 8-1. Real T-bill returns after federal income tax at historical rates. Later editions update the figures.
10. IRS. Net Investment Income Tax, and Questions and Answers on the NIIT. The source for the 3.8% rate, the inclusion of interest, the thresholds ($125,000 for married filing separately), and their not being indexed for inflation.
11. Ang, A., Bekaert, G., & Wei, M. (2008). “The Term Structure of Real Rates and Expected Inflation”. The Journal of Finance, 63(2), 797-849.
12. Fama, E.F., & Schwert, G.W. (1977). “Asset Returns and Inflation”. Journal of Financial Economics, 5(2), 115-146. Sample period 1953-1971.
13. TreasuryDirect. Treasury Inflation-Protected Securities (TIPS). The source for CPI indexation of principal and the greater-of-adjusted-or-original-principal rule at maturity.
14. Campbell, J.Y., & Viceira, L.M. (2001). “Who Should Buy Long-Term Bonds?”. American Economic Review, 91(1), 99-127.
15. Campbell, J.Y., Shiller, R.J., & Viceira, L.M. (2009). “Understanding Inflation-Indexed Bond Markets”. Brookings Papers on Economic Activity, Spring 2009. The source for the quoted sentence on Treasury bills and long-term investors.
16. TreasuryDirect. Tax Forms and Tax Withholding. The source for annual reporting of TIPS inflation adjustments on Form 1099-OID before they are paid.
17. U.S. Department of the Treasury. Daily Treasury Bill Rates, Par Yield Curve and Par Real Yield Curve, September 30, 2026: 13-week bill 4.13% (coupon equivalent), 10-year nominal 5.29%, 10-year real 2.93%, 20-year real 3.18%. The 2.36% inflation rate is the 10-year nominal minus the 10-year real yield.
18. TreasuryDirect. I Bonds. The source for the $10,000 electronic purchase limit, the 12-month minimum holding period, the three-month interest penalty before five years, and federal tax deferral with state exemption.
19. Drechsler, I., Savov, A., & Schnabl, P. (2017). “The Deposits Channel of Monetary Policy”. The Quarterly Journal of Economics, 132(4), 1819-1876. The average deposit spread beta of 0.54 (0.61 for the largest banks).
20. FDIC. Quarterly Banking Profile, Second Quarter 2026, Table III-A: cost of funding earning assets, all insured institutions. The 3.7% bill comparison is our average of Treasury’s daily 13-week coupon-equivalent rate for April-June 2026.
21. Board of Governors of the Federal Reserve System (2023). “Changes in U.S. Family Finances from 2019 to 2022: Evidence from the Survey of Consumer Finances”, Table 3 and Appendix B (definitions of transaction accounts and bonds).
22. Board of Governors of the Federal Reserve System (2026). Economic Well-Being of U.S. Households in 2025: Savings and Investments, Table 28.
23. Board of Governors of the Federal Reserve System. Financial Accounts of the United States (Z.1), households and nonprofit organizations, second quarter 2026: money market fund shares (FL153034005), checkable deposits and currency (FL153020005), and time and savings deposits (FL153030005).
24. Investment Company Institute. Money Market Fund Assets, week ended September 23, 2026.