Average Bank Account Balance in America (2026): By Income Bracket

Type "average bank account balance" into a search engine and you'll get one number: $62,410, sourced to the Federal Reserve. It's accurate, and it's also close to useless, because it's a mean -- an average pulled far above what most people actually hold by a relatively small number of very large balances. The Fed's own median, the amount where half of households hold more and half hold less, is $8,000. That's a much better single number. But it's still one number for the entire country, and the real answer depends heavily on where your income actually falls. This report breaks it out by income bracket, using a direct recomputation of the Fed's own public microdata rather than the headline figure alone, and shows what the gap between "average" and "your actual rate" is worth today. This page is reviewed by the SwitchWize Editorial Team; the figures are sourced below with dates.

Mean vs. median: why the number you've heard is misleading

Every site citing "the average bank account balance" is citing the same source: the Fed's triennial Survey of Consumer Finances (SCF), most recently fielded in 2022 (the 2025 wave is expected around Q4 2026). The Fed reports two very different numbers from that same survey: a mean of $62,410 and a median of $8,000.

The mean is what you get by adding up every household's balance and dividing by the number of households. The problem is that a mean is extremely sensitive to outliers, and bank balances have a lot of them: a relatively small share of households hold balances in the hundreds of thousands or millions, and those balances pull the average far above what a typical household holds. The median -- the balance where exactly half of households hold more and half hold less -- isn't affected the same way, which is why it's the more honest single number for "what does a typical American have in the bank."

SwitchWize didn't just repeat the Fed's $8,000 figure. We downloaded the Fed's own public 2022 SCF microdata (the same underlying dataset the Bulletin is built from) and recomputed the weighted median directly, using the Fed's own analysis weights across all five multiple-imputation implicates. Our result: $7,850 -- close enough to the Fed's published $8,000 to confirm the number rather than just cite it, with the small difference explained by normal rounding and weighting-method variation.

The number that actually matters: by income bracket

Neither $62,410 nor $8,000 describes a specific household, because both are single national figures folding together everyone from minimum-wage earners to the top 1%. The Fed's own survey already sorts households into six income-percentile brackets, and recomputing the median balance within each one tells a much more useful story:

That's a 139x gap between the bottom and top brackets, all collapsed into one national "average" that describes almost nobody exactly. Notice that the 40th-60th percentile bracket -- literally the middle fifth of the country by income -- lands almost exactly on the overall median ($7,200 vs. $7,850 nationally), which is a useful sanity check that the middle of the income distribution and the middle of the balance distribution roughly line up, even though the tails pull hard in opposite directions.

What the gap is actually worth, right now

Knowing where your balance falls relative to your income bracket is interesting. What's actionable is what that balance could be earning. As of August 2026, the best widely-available savings APY is 4.10%, against a FDIC national-average savings rate of 0.38% -- a live gap of 3.72 percentage points. (This piece is dated; the live dataset and Bank Gap Index always show today's real figures.) Applied to each bracket's real median balance instead of a single flat example:

Every bracket is leaving real money on the table -- proportionally similar (the gap is the same 3.72 points everywhere), but very different in dollar terms depending on the balance behind it. See the full breakdown of what that means for you:

The honest counterargument

A few real limits are worth stating plainly. The 2022 SCF is the most recent wave available; a 2025 wave is expected around Q4 2026, and these figures will need refreshing once it publishes. "Transaction accounts" in the Fed's methodology combines checking, savings, money market, and call accounts -- it is not a single savings-account figure, so it will read higher than "how much is in my savings account" for someone who keeps most of their liquidity in checking. And a median by income bracket still isn't a target: age, region, homeownership, and debt load all shift what's typical for otherwise-similar households, and none of this is personalized financial advice.

None of that changes the core point. Whatever your income and whatever your actual balance, comparing it to "the national average" tells you almost nothing, because that average blends brackets 139x apart. Comparing your account's rate to what's actually available today is the comparison that's worth acting on.

Methodology

Overall and by-bracket balances are computed directly from the Federal Reserve's 2022 Survey of Consumer Finances Summary Extract public microdata (rscfp2022.dta), using the Fed's own analysis weights (WGT/5 across the 5 multiple-imputation implicates) and the Fed's own six-category income-percentile classification (INCCAT) -- not transcribed from the Bulletin PDF. The computed overall median ($7,850) is validated against the Fed's own published figure ($8,000) as a sanity check on the weighting and variable handling. Live rate-gap figures (4.10% best available, 0.38% national average) come from SwitchWize's Bank Gap Index, the same live inputs used across our rate-spread indices, and will differ from the exact figures above whenever rates move.

How we source this. Balance-by-income figures are our own direct computation over the Fed's public microdata, not a secondary citation. Live rate figures come from SwitchWize's own rate database. See our methodology and editorial team. We take no payment for organic rankings.