What a Century of T-Bill Returns Teaches About Cash
Investors who keep large sums in cash or short-term Treasury bills believe they are avoiding risk entirely. The long historical record shows that "safe" asset has barely stayed ahead of inflation, which means holding too much of it for too long is its own quiet form of loss.
The core principle: nominal safety, real erosion
U.S. Treasury bills are about as close as investing gets to a risk-free asset. They are backed by the full faith and credit of the federal government, mature in a year or less, and their price barely moves before maturity, which is why they anchor the short end of every yield curve and why economists use their yield as the baseline "risk-free rate" in nearly every model in finance. What bills do not protect you from is inflation. Across long historical stretches reaching back nearly a century, the average nominal return on Treasury bills has landed somewhere in the 3% to 4% per year range, while average inflation over those same long stretches has run close to 3% per year. Subtract one from the other using the basic relationship real return = nominal return − inflation rate, and the long-run real return on cash equivalents comes out close to zero, sometimes fractionally positive, sometimes fractionally negative depending on the exact window measured.
This is not a flaw in Treasury bills; it is close to their design intent. A security this liquid and this free of default and price risk should not be expected to also deliver a rich real return, because in an efficiently priced market, return and risk move together. Bills carry almost no risk, so investors do not need to be paid much beyond compensation for the erosion of purchasing power to hold them. Their job in a portfolio is capital preservation and liquidity for near-term needs, not long-term wealth building; that job has historically fallen to riskier assets willing to pay a premium for the privilege of being held through uncertainty.
The math: two worked examples of real return
Example 1: the erosion of a lump sum held in bills for 30 years. Suppose you keep $10,000 in Treasury bills earning an average nominal 3.5% per year for 30 years, while inflation averages 3.0% per year over the same span. The nominal ending balance compounds as 10,000 × (1.035)^30. Since 1.035 raised to the 30th power is approximately 2.807, the account grows to roughly $28,070 in nominal dollars, which looks like healthy, steady growth on a statement. But prices have also been rising the whole time: 1.03^30 is approximately 2.427, meaning something that cost $10,000 at the start now costs roughly $24,270. Divide the nominal ending balance by that price multiplier, 28,070 / 2.427, and the result is approximately $11,565 in today's purchasing power. Over three full decades, the real value of the pile grew only about 15.6%, or roughly 0.5% per year compounded, despite the account statement showing 3.5% nominal growth the entire time.
Example 2: comparing two five year windows with different inflation regimes. In a low inflation window, bills yield an average 2.0% while inflation runs 1.5%, giving a real return of roughly 0.5% per year; over five years, (1.005)^5 − 1 is about 2.5% cumulative real growth. In a high inflation window, bills yield an average 5.0% while inflation runs 6.5%, giving a real return of roughly negative 1.5% per year; over five years, (0.985)^5 − 1 is about negative 7.2% cumulative, meaning the saver's real purchasing power actually shrank by about 7% over five years despite collecting a 5% nominal yield the entire time. The lesson is that the nominal yield on bills tells you almost nothing about your real outcome without knowing the inflation regime it was earned in.
What close to a century of data shows
Data on U.S. Treasury bill returns and consumer price inflation extending back to the mid-1920s shows this near-zero real return pattern holding up remarkably consistently as a long-run average, even though it masks enormous variation window to window. There have been extended multi-year stretches, notably periods of unexpectedly high inflation, where bill holders suffered meaningfully negative real returns for years running, and other stretches, notably periods when central banks pushed short rates up faster than prices rose, where bill holders earned a modest but genuine positive real return. Averaged across the full near-century span, the two effects roughly cancel, landing close to zero. This is precisely why cash is described in finance as preserving nominal capital while offering no reliable long-run real growth: the long-run data backs that characterization almost exactly.
Over the same long span, asset classes willing to bear more risk delivered a materially higher average real return. Long-term government bonds have historically cleared inflation by a modest but positive margin over multi-decade periods, and diversified stock portfolios have historically cleared inflation by a substantially larger margin, commonly several times the real return of bonds. The pattern across all three asset classes is consistent with the basic risk and return relationship that runs through the whole discipline: bills are safest and grow purchasing power slowest, bonds sit in the middle, and stocks carry the most volatility and have historically compensated for it with the most real growth.
A further wrinkle worth understanding is that the government also issues a security purpose-built to solve this exact problem: Treasury Inflation-Protected Securities, commonly called TIPS, whose principal value adjusts with the official consumer price index, so the real return is set directly at purchase rather than depending on guessing future inflation correctly. A TIPS bond bought to yield a real 1.5% will deliver approximately 1.5% real return regardless of whether inflation over its life turns out to be 2% or 8%, because the principal itself rises with prices before the fixed real coupon is applied on top. Ordinary Treasury bills carry no such adjustment; their nominal yield is fixed (or reset frequently at auction) without any built-in inflation protection, which is precisely why their real return has fluctuated with whatever inflation happened to do, rather than sitting reliably near a known target the way TIPS real yields do.
How this applies in real portfolios
The practical implication is straightforward but frequently ignored: cash and cash equivalents should be sized to match near-term, known needs, not treated as a long-term savings vehicle. An emergency fund covering three to six months of expenses, a house down payment you plan to use within a year or two, or funds earmarked for a known near-term tax bill all belong in bills, high yield savings, or money market funds, because the priority for that money is that it be there and intact when needed, not that it grow. Money with a multi-decade horizon, by contrast, sitting in cash "to be safe" is quietly losing ground to inflation year after year even while the account balance climbs, and the opportunity cost compounds the longer it sits there.
This shows up most visibly after a market downturn, when investors who sold risk assets in a panic often leave the proceeds sitting in cash for years afterward, waiting for a clearer signal to reinvest that never quite arrives. That waiting period is exactly when the near-zero real return on cash does its quiet damage, and it is also, historically, when risk assets have delivered a disproportionate share of their long-run real return, since recoveries tend to be front-loaded relative to the following expansion. The investor who moved to cash to feel safe and stayed there too long typically pays for that safety twice: once in foregone recovery gains, and again in ongoing inflation erosion on the cash itself.
Actionable breakdown
- Match cash holdings to genuinely near-term needs.
- Three to six months of expenses for emergencies.
- Known expenses due within one to two years.
- Expect roughly zero long-run real return from cash.
- Nominal balance growth is not the same as wealth growth.
- Subtract expected inflation before judging any cash yield.
- Put multi-decade money into assets with positive expected real return.
- Bonds for a modest, more stable real return.
- Stocks for the largest historical real return, with more volatility.
- Reassess large idle cash balances on a fixed schedule.
- Revisit at least once a year, not just after a scare.
- Set a written reinvestment plan before a downturn happens.
There is also a tax dimension that makes the real return on ordinary bills even less generous than the pretax figures above suggest for a taxable account. Interest from Treasury bills is subject to federal income tax (though exempt from state and local tax), so an investor in a 32% federal bracket earning a 3.5% nominal yield keeps only about 2.38% after tax, which, against 3.0% inflation, produces a real after-tax return of roughly negative 0.6% per year, a genuine, quiet erosion of purchasing power even before accounting for any state tax exposure on other cash holdings like ordinary savings accounts. This tax drag is one more reason large cash balances held for reasons other than genuine near-term need deserve regular scrutiny rather than being left on autopilot.
Common pitfalls
The first and most common pitfall is mistaking nominal safety for financial safety. A bill portfolio that never shows a negative number on the statement can still lose most of its purchasing power over a working career, a real and frequently underestimated risk precisely because it never produces a scary, headline-grabbing monthly loss the way a stock decline does.
The second is holding excess cash after a downturn out of fear and letting that turn into a multi-year default, missing the recovery years when risk assets have historically delivered a disproportionate share of their long-term advantage.
The third is anchoring to a recent high nominal yield on cash, such as a period when short-term rates rose sharply, without checking what inflation was doing at the same time; a headline 5% yield during a stretch of 6% inflation is a real loss, not a windfall, even though it sounds impressive next to the near-zero rates of prior years.
The fourth is treating money market funds and savings accounts as fundamentally different animals when they are functionally close cousins of Treasury bills, both economically and in their long-run real return profile, so shopping among them for the best nominal rate matters far less than deciding how much belongs in cash at all.
None of this is an argument against holding bills at all; it is an argument for holding the right amount, sized to a genuine purpose. A well-run household or practice keeps enough in bills, money market funds, or high yield savings to cover true near-term needs with certainty, and treats every dollar beyond that as long-horizon capital that belongs somewhere with a positive expected real return after tax and after inflation, reviewed and rebalanced on a regular schedule rather than left to drift upward simply because it feels comfortable to have a large cash cushion.
The bottom line
Treasury bills and cash equivalents protect nominal capital and offer real safety for near-term needs, but they have historically delivered close to zero real return over long stretches, so they belong in the safety bucket of a plan, never the long-term growth bucket.
Related reading: cash and emergency funds guide, market history guide, what sets interest rates, what stocks and bonds have actually delivered, inflation.