Here is the short version. A certificate of deposit pays you for a promise: you agree not to touch the money for a fixed term, and the bank agrees the rate cannot change on you. A savings account makes the opposite deal: the money is yours any morning you want it, and the rate is the bank's to change any afternoon it wants to. In 2026 the gap between the two is narrower than the advertising suggests, and the early withdrawal penalty can erase a CD's edge entirely if there is any real chance you break the term. This piece works the actual numbers, from the published national averages down to the month-by-month arithmetic of breaking a CD, so you can see what the lock buys before you agree to it.
Two ground rules. No bank names appear here, because the right answer is whatever your own institutions quote you this week, not what anyone was paying when this was written. And no rate predictions appear either, because nobody honest has any.
What the published averages say, and what they hide
The FDIC posts national average deposit rates, and as of August 17, 2026 they read: savings accounts 0.38 percent, 12-month CDs 1.71 percent, 24-month 1.57, 36-month 1.34, and 60-month 1.36. Two things in that list deserve a second look. First, the averages are weighted, in the FDIC's words, "by each institution's share of domestic deposits," so the giant banks that pay almost nothing dominate the figure, which is why your own bank's savings offer may sit far above 0.38 or, if it is one of those giants, right on it. Second, the curve is upside down: the average 12-month CD pays more than the average 60-month. In August 2026, locking money up for longer did not, on average, pay more. That single fact should change how you read every CD pitch this year.
A second source, counted a different way, tells the same story from another angle. The National Credit Union Administration publishes a quarterly comparison of bank and credit union rates; its table dated December 26, 2025 showed 1-year CDs averaging 2.95 percent at credit unions against 2.29 at banks, and 5-year CDs at 2.83 against 2.11. Those figures are older and computed differently, unweighted by deposit share, which is exactly why they run higher than the FDIC's. Between the two sources the honest reading is this: averages depend on who is counting, big banks pay little, credit unions and smaller institutions pay meaningfully more, and shorter terms have recently paid as well as long ones or better. Your job is not to memorize any of these numbers. It is to collect three real quotes of your own before moving a dollar.
The penalty math, worked all the way through
The Consumer Financial Protection Bureau's plain-language page on CDs says to compare three things: the term, the rate, and "the amount of the penalty for withdrawing money before the end of the term." Most people compare the first two and skip the third. The third is where the math turns.
Penalties are usually quoted as a number of months of interest, and six months is a common figure on a 1-year CD, though yours may be three or twelve, which is why the disclosure is the first thing to read. Now work it. Suppose $10,000 in a 12-month CD at 4.00 percent, with a six-month penalty, against a savings account paying 3.60 percent. The CD earns about $33 a month, the savings about $30. If you break the CD at month eight, you have accrued roughly $267, the penalty takes $200 of it, and you keep $67. The savings account over the same eight months paid $240. The lock cost you $173 for the privilege.
Here is the part worth writing on the disclosure itself. The CD's advantage over that savings account is about $3 a month. The penalty is $200. At $3 a month, recovering $200 takes over five years, and the CD only runs twelve months. So with a six-month penalty and a rate edge that thin, a broken CD loses to savings no matter when in the term you break it. Even with a full percentage point of edge, the break-even sits around two years, still past the maturity date. The conclusion is not that CDs are bad. It is that a CD is only worth holding with money you are genuinely certain to leave alone, because the arithmetic gives no partial credit for almost making it.
What a ladder actually fixes
A ladder is the standard answer to that certainty problem, and it is a good one when used for what it does. Take $60,000 you will not need for daily life. Instead of one 5-year CD, buy five $12,000 CDs maturing in one, two, three, four, and five years. Every year one rung matures. If you need the cash, it is there without a penalty. If you do not, you reinvest it at whatever rates then exist, again without a prediction. You are never more than twelve months from a fifth of the money, and you never bet the whole sum on today's rates being the good ones.
What a ladder does not do is manufacture yield. On the FDIC's August 2026 averages, with long CDs paying less than short ones, the longer rungs earned less, not more. A ladder is a liquidity schedule and a rate-averaging machine. Judge each rung against a savings account the way the last section did: rate edge per month against penalty, times your honest odds of needing that rung early. Money with real odds of being needed, the roof fund, the car fund, the account that absorbs a bad month, belongs in savings even at a lower sticker rate. Both are insured to the same $250,000, at banks through the FDIC and at credit unions through the NCUA, so safety is not the variable here. Timing is.
The honest trade, stated plainly
A CD's real product in 2026 is not extra yield, which the averages show is small or absent. It is certainty. The savings account paying 3.60 today can pay less next quarter without asking you; the CD cannot. If you are living on a fixed budget and want a known number of interest dollars landing in a known month, that certainty has genuine value, the same way a larger locked-in check has value in the Social Security claiming decision. Just price it as certainty, not as profit.
A fair way to decide, then. Split the money by its job. Anything you could plausibly touch inside a year goes to the best savings rate you can find at an insured institution. Anything with a firm date, the property tax bill, next year's trip, gets a CD maturing just before that date. Anything beyond that can ladder. Then read the penalty clause of every CD before signing, do the three-dollar-a-month arithmetic above with your own quotes, and keep the disclosures in the same folder as the maturity dates. The banks are counting on nobody doing that math. It takes ten minutes, and it is the only part of this decision that is entirely under your control.
