How Long a Top Savings Rate Lasts

Which of these are you?

• You just saw a bank ranked first and wonder whether to move: the rank will probably change within weeks, and the rate behind it probably will not. Decide on the gap between your APY and the field, using the break-even math in Chapter 1.
• You moved to a top-10 bank and its position slipped: your rate most likely did not. Check your own APY, not the ranking, before you do anything.
• You think top rates are teasers that fade within months: in our data only 7 of 33 classified exits were the bank's own cut. Read the sections on rank churn and on how often banks change a rate.
• You want a rule for how often to look: the last section gives one, and Chapter 8 turns it into a written plan.

What did we measure, and how?

We ranked institutions by their best high-yield savings APY every day for 167 days and recorded how long each one stayed in the top 10. This is the Rate Half-Life study, and it is preliminary: one rate cycle, scraped rates, and only the institutions SwitchWize tracks.

The definitions, in the study's own wording. The daily panel holds, for each institution and day, the highest high-yield savings APY we observed. An institution's last observed rate is carried forward for up to 7 days, after which it drops out of that day's ranking. The top 10 is the 10 highest current values that day. A stay is an unbroken run of days in the top 10. A rate change is a move of at least 0.05 percentage points that is still in place at the next observation, so a one-day scrape error is not counted. Where an institution has several savings products, only its best rate that day is used.

The window runs from 2026-04-20 to 2026-10-03 and covers 126 institutions. It produced 63 stays. Of those, 43 have ended and 10 are still running. The other 10 were already in progress on the first day, so we cannot know when they began and we leave them out of the survival curve. That leaves 53 stays, and they belong to just 27 institutions: the same names re-enter the top 10 repeatedly.

How long does a top-10 spot last?

A top-10 spot lasted a median of 15 days. An estimated 69 percent of stays reached 7 days, 50 percent reached 14, 32 percent reached 30, and 17 percent reached 60. Of the 43 stays that ended, 16 ended within a week.

The median is the first day on which the survival estimate falls to 50 percent or lower. It is 15 days here because the estimate is 50 percent at 14 days (to the nearest whole percent) and has dropped below it by day 15.

What is a Kaplan-Meier estimate, and why not just average the stays?

A Kaplan-Meier estimate is a way to measure how long something lasts when some of the observations have not finished yet. Averaging only the stays that ended would understate how long stays last, because it throws away the ones still running. The method comes from Kaplan and Meier (1958).

The mechanics are short enough to do by hand. Take five hypothetical stays: four ended at 2, 4, 4 and 9 days, and one was still running at day 6 when tracking stopped. Walk through the days on which a stay ended. The share that survives each such day is the share that survived the last one, multiplied by the fraction of stays still at risk that did not end on that day.

On day 2, all 5 stays are at risk and 1 ends: 1 - 1/5 = 0.80. On day 4, 4 stays remain at risk and 2 end: 0.80 x (1 - 2/4) = 0.40. The median is therefore 4 days, the first day the estimate reaches 50 percent or lower. The running stay still counts toward the denominator through day 6, which is how it is used without pretending it ended. If you simply dropped it and asked what share of the 4 finished stays lasted past 4 days, you would get 1 of 4, or 25 percent, below the 40 percent the method gives.

The study applies exactly this calculation to its 53 stays, 10 of which were still running when the data ended.

Do banks lose the top spot because they cut, or because others pass them?

Mostly because others pass them. Of the 43 completed stays, 26 ended without the bank cutting its own rate by 0.05 points, 7 ended when the bank cut its own rate by 0.05 points or more, and 10 could not be classified because the bank had no fresh rate that day. Among the 33 classified exits, 79 percent were passes and 21 percent were cuts.

This is the headline finding, and it contradicts a common claim: that a top rate is a teaser that fades. If it were, we would see the leaders cut soon after arriving. We checked directly. Among 40 banks that entered the top 10 and had a rate on both the day they entered and 30 days later, the median change in their own APY was zero, and 4 of 40, or 10 percent, were lower. The ranking changed because the field moved around a mostly steady rate, not because the rate melted.

How often does a bank change a savings rate?

Rarely, for most institutions we track. Of 91 institutions with at least 60 days of data, 47 (52 percent) made no confirmed change of 0.05 points or more in the 167-day window, and the median was zero changes per 90 days. Across all 91 there were 73 confirmed changes.

Of the 73 changes, 41 were raises and 32 were cuts, so 44 percent were cuts. Rate changes happen in both directions, and the typical institution does neither often. We did not test why banks leave a rate alone, so treat any explanation as a hypothesis; the observation is that most do.

The same pattern shows up in our earlier, shorter study, We Tracked Bank Rates for 90 Days. Its window, threshold and counting method differ from this chapter's, so do not add the two sets of counts together. Both say the same thing about direction: when banks move, they move up as well as down.

Banks are allowed to change a variable rate without telling you in advance. Regulation DD requires 30 days of advance notice before a change that may reduce your APY, but it exempts changes in the interest rate and APY of a variable-rate account (12 CFR 1030.5(a)(2)(i)). The new rate may only show up on your dashboard or your next statement.

What does this mean for you?

Your own rate is the number to watch, and the ranking is not. A bank can lose its top spot with no change in what you earn, and a bank can quietly cut your rate while it stays ranked. Compare your APY to what comparable accounts pay on a schedule, and move only when the gap repays the cost.

Here is a hypothetical to make that concrete. You hold $25,000 at 3.50 percent. A competing account pays 4.20 percent. Moving takes 1.5 hours that you value at $25 an hour, so the switching cost is 1.5 x $25 = $37.50, with no fees and no bonus. We do not know how long the 4.20 percent will last, so we compare three assumptions. These scenarios are stated by us, not taken from the study.

• A: the new rate holds for all 12 months.
• B: the new rate holds for 2 months, then loses 0.10 percentage points each month until it reaches your current rate (month 9).
• C: the new rate starts at 4.20 percent and loses 0.10 points every month from month 1, so it pays 4.10 percent in month 1.

Even the fast-decay scenario stays positive, because the floor is your own rate and the cost is small. That is the useful result: when a move is cheap, a rate that fades still repays it. Scenario C leaves only $4.81 of margin, though, and if the cost were higher the sign would flip. The calculator below uses scenario B's shape with your inputs. Put in the decay you consider realistic; the study does not give you one.

The Bank Switch ROI calculator adds a sign-up bonus and ongoing friction if you want the fuller picture.

Why shouldn't you chase the top spot?

Because position turns over every few weeks, and each move costs time. The median stay is 15 days. If you moved to every new leader, you would move about 24 times a year (365 days divided by 15, rounded down). At $37.50 of time each that is 24 x $37.50 = $900, against $169.04 a year of extra interest if the 0.70-point gap in scenario A held for the full year.

The comparison is rough on purpose. A top-10 spot is a rank, not a promise about your APY, and you would not actually move at each change. But it shows the shape of the trade: the rate gap pays you a few hundred dollars a year on a mid-sized balance, and chasing position spends more than that.

The gap you need is smaller than most people assume when the balance is large and larger when it is small. With a cost of $37.50 and a 12-month payback, the smallest rate gap that repays the move is 0.15 percentage points on $25,000 and 0.75 points on $5,000 (cost / balance: $37.50 / $25,000 = 0.15 percent). Chapter 1 builds this table out for more balances and time costs.

What are the limits of this data?

Read the findings as a first look. The window is 167 days, a single rate environment, and the institutions are the 126 SwitchWize tracks, not the whole market.

• Our set is not the market. A bank we do not track could outrank everyone in it.
• Rates are scraped from public pages. A change at the bank may show up on our side a day or more later.
• Promotional and balance-tiered rates are not separated out. Use each bank's own terms for what you would actually earn.
• One rate cycle. Nothing here predicts what happens in a falling or rising cycle. Whether banks follow the Fed is a separate question with its own one-event snapshot, in Chapter 4.
• A stay is a measure of rank. It does not tell you how long any particular rate lasts for a particular customer.
• Ten stays could not be classified. Where a bank had no fresh rate on its exit day, we could not say whether it passed or cut.

The full data are on the study page, with both per-institution and per-stay downloads.

How often should you check your rate?

Check on a schedule and after events, not daily. A quarterly comparison of your own APY with the field catches the changes that matter, because at most banks we track the best rate changes rarely, and a daily check would mostly watch the ranking, which is noise for your purposes.

A workable rule, stated as an assumption and not a finding: compute the dollar gap between your APY and the best comparable insured account you would actually open. If that gap, held for 12 months, would not repay your switching cost, do nothing. If it would, run the break-even math with a rate decay you consider plausible and decide. Add checks when your bank emails a rate notice, when your balance changes a lot, or after a Fed move.

For today's market, the best savings APY we track is 4.27% and the national average is 0.38%. Read your own APY off your account page and compare it with both. The best savings accounts guide lists the accounts behind the top number, and Chapter 2 puts a dollar figure on the gap.