The Yield Curve and Choosing a Term

Which of these are you?

• You know the date the money is needed. Buy the term that ends on or just before that date. The curve only decides between two CDs that both fit your date.
• You have money you will not touch for years and the sheet is steep. A longer term can pay. Compute the implied forward rate below and compare it with what you believe a shorter CD could plausibly earn later.
• The sheet is flat or inverted and your date is uncertain. The longer term charges you liquidity for nothing. Take the shorter term, or split the money across terms in a ladder.
• You are reading headlines about an inverted curve. Ignore the headline. The curve in the news is a Treasury curve. Your decision runs on your bank's rate sheet.

What the yield curve is, from Treasury's own method

The Treasury yield curve is a line of yields against maturity, built from the most recently auctioned Treasury securities. Treasury's methodology page says it uses indicative, bid-side price quotations, not actual transactions, gathered by the Federal Reserve Bank of New York at or near 3:30 PM each trading day. The inputs are bills (4-, 6-, 8-, 13-, 17-, 26- and 52-week), notes (2-, 3-, 5-, 7- and 10-year) and bonds (20- and 30-year).

Treasury fits a smooth curve through those inputs with a monotone convex spline applied to forward rates, which replaced an earlier quasi-cubic Hermite spline in December 2021. The result is a par yield curve: the yield at which a security of each maturity would be priced at par, meaning 100% of face value. Treasury then reads fixed maturities off the fitted curve. The daily table on home.treasury.gov publishes 1, 1.5, 2, 3, 4 and 6 months and 1, 2, 3, 5, 7, 10, 20 and 30 years, and calls these the constant maturity Treasury rates. Treasury says the rates are usually available by 6:00 PM Eastern on each trading day.

Two facts follow for a CD saver. First, the curve is a market price for government borrowing, so it contains no credit risk and no bank funding needs. Second, it is a fitted curve, so a published point such as the 3-year yield is read off a smooth line, not quoted from one security. This chapter states no current yields. The data page is the place to read today's curve, and the numbers below are hypothetical.

Three shapes, and what each means for a CD term

A normal curve slopes up: longer maturities pay more. A flat curve pays about the same at every maturity. An inverted curve slopes down: short maturities pay more than long ones. The same words apply to a CD rate sheet, listing the best APY at each term from 3 months to 5 years, even though a bank sheet does not follow Treasury's curve one for one.

The Federal Reserve staff note that explains this defines inversion as short-term yields rising above longer-term ones. It adds a useful decomposition from a companion note: the forward rate at a given maturity can be thought of as the market's expected short rate at that horizon plus a term premium. A steep curve therefore mixes two things, an expectation that short rates will rise and a payment for tying money up. You cannot separate them from the sheet alone, which is why the sections below test the trade in dollars instead of guessing the cause.

The premium per extra year is the spread divided by the extra years. On the normal sheet, 0.50 points across 4 extra years is 0.125 points, or $12.50 a year per $10,000. That is the price the sheet is offering you to give up access for a year, and it is the number to hold against the penalty and against the chance you need the money.

The implied forward rate: the number that decides

The implied forward rate is the average annual rate a 1-year CD must earn, rolled over each year for the next four years, to leave you exactly even with the 5-year CD. It is the break-even for the shorter term.

Formula: f = ((1 + y5)^5 / (1 + y1))^(1/4) - 1

Variables: y5 is the 5-year APY, y1 is the 1-year APY, and f is the average APY on the four later 1-year CDs. Annual compounding, no taxes, no penalties, hypothetical rates.

Worked example, normal sheet (hypothetical). You have $25,000, y1 = 4.00% and y5 = 4.50%.

1. Value at year 5 in the 5-year CD: 25,000 x 1.045^5 = $31,154.55.
2. Value at year 5 if the 1-year CD is rolled at a flat 4.00%: 25,000 x 1.04^5 = $30,416.32.
3. Difference: 31,154.55 - 30,416.32 = $738.23 in favor of the 5-year CD.
4. Implied forward rate: (1.045^5 / 1.04)^(1/4) - 1 = 4.63%.

The reading: if you believe the 1-year CD will average more than 4.63% over years 2 through 5, the short CD wins. If you believe it will average less, the long one wins. You do not need a forecast to use this, only a view on whether 4.63% is a high or low bar for your own money. If your date is a hard date inside 5 years, none of this applies, because you cannot use the 5-year CD at all.

Same test on the other shapes (hypothetical).

In the inverted row, the 5-year CD pays 0.50 points less, and the short CD wins unless later rates average below 3.88%. Locking an inverted curve for years means you pay to lose flexibility. Rolling the 4.50% CD at a flat 4.50% ends at $31,154.55 against $30,416.32 for the 5-year, a $738.23 advantage at a flat 4.50%, shrinking to zero if later years average 3.88%.

Your CD sheet is not Treasury's curve

Banks price CDs from their own funding needs, competition and balance-sheet appetite, so a bank's 5-year CD rate can sit above or below the 5-year Treasury yield, and an inverted Treasury curve does not force an inverted CD sheet. Treasury's curve is the public benchmark for what safe money earns at each maturity, and it is worth watching as a cross-check when you are deciding whether a bank's long CD is generous or stingy. Your CD sheet is the actual menu.

To build your own curve, list the best APY for 3, 6, 12, 24, 36 and 60 months from one rate table on one date, then compute the premium per extra year between the terms you are choosing between. The CD rates we track are one place to gather that list. Treat any sheet as a snapshot: it changes when banks reprice, and the shape you saw last month may not be today's.

What a longer term costs when plans change

The premium you earn from a longer term is small compared with what you can lose by needing the money early. Early withdrawal penalties are set by each bank and stated in the account agreement, so the 180 days used here is a hypothetical, not a range from any bank.

Worked example, premium against penalty (hypothetical). A $25,000, 5-year CD at 4.50%, with a penalty of 180 days of simple interest at the contract rate.

1. Penalty = principal x rate x days / 365 = 25,000 x 0.045 x 180 / 365 = $554.79.
2. Premium over the 1-year CD in year one = 25,000 x (0.045 - 0.040) = $125.00.
3. Years of premium to repay one penalty = 554.79 / 125.00 = 4.4 years.

One break wipes out about 4.4 years of the premium. This is the mechanism behind a sensible rule: when the sheet pays a small premium for a long term, the term is only worth taking on money whose date you are sure of. The penalty mechanics, and the break-even replacement APY for breaking a CD, are in Chapter 3.

Reading the curve: a decision procedure

Run these steps in order. Each one narrows the choice before the next.

1. Write the date. For each block of money, write the month you need it. Any term that ends after that month is disqualified, unless you accept the penalty.
2. List the best APY at each qualifying term from one rate table on one day.
3. Compute the premium per extra year between your two candidate terms: (longer APY minus shorter APY) divided by the extra years.
4. Compute the implied forward rate for the longer term with the formula above.
5. Ask one question about the forward rate: is that a high or a low bar for a one-year CD over the years that follow? If it is a high bar, the longer term tends to be the better trade. If it is a low bar, the short term tends to be.
6. If you cannot answer step 5, split the money. Put part in the shorter term and part in the longer one. That is a ladder or a barbell, compared in Chapter 12.

• When the sheet is steep, the premium can be worth taking on money you will not need. Check the forward rate: a steep sheet has a high forward rate, and that high bar is the market's price for tying money up.
• When the sheet is flat, the premium is near zero, so the reason to go long is only to remove the risk that the short CD reprices lower when it rolls. Decide whether that protection is worth losing access.
• When the sheet is inverted, the shorter term pays more and stays flexible, so the case for long terms is thin unless your date forces one.

A flat or inverted sheet is not a reason to sit in cash forever, and a steep sheet is not a reason to reach for the longest term. The steps produce a term that fits your dates and a fair price for stretching it.

The recession-signal caveat

Federal Reserve staff have documented that the ten-year minus three-month Treasury spread narrowed before each of the six most recent recessions dated by the National Bureau of Economic Research, as of that 2018 note, and that the curve was inverted on that measure before five of them. A separate Fed note on the near-term forward spread says long-term spreads are likely to be affected by other factors unimportant for forecasting recessions, and suggests the near-term spread may predict recessions only because it impounds expectations market participants have already formed.

That literature is about recessions. It does not tell you what your bank's CD rates will do, and it does not tell you whether to lock. Treat an inverted Treasury curve as a fact about relative prices today. It does not change the procedure above: the date sets the term, the sheet prices the premium, and the forward rate is the break-even.

Where the maturity date lands matters too

A term is only useful if the maturity date lands when you need the cash. A 24-month CD bought today ends 24 months from today, and the grace period after maturity is short and set by each bank. Check the calendar, not just the label. For staggering maturities so a piece of your money comes due on a regular schedule, see the ladder chapters.

Use the CD ladder calculator to see how a spread of terms blends into one average maturity and one blended APY. The average maturity of a ladder of N equal rungs spaced s months apart is s x (N + 1) / 2 months, so five annual rungs average 36 months.