Personal Saving Rate: What US Households Have Left

US personal income rises, yet the personal saving rate falls. The two readings can easily coexist. Households collectively can take home more after taxes and spend enough of the increment that a smaller fraction remains. A chart of bank balances describes something else again.

Did saving change in dollars or only as a share of income? Start by checking after-tax income, outlays and annual-rate units in the source documents. The calculations below distinguish the published record from an illustrative customer transaction and show how to check the result in dollars.

Key insight. BEA measures personal saving as disposable personal income less personal outlays. Dividing saving by disposable personal income gives the personal saving rate. It is a flow for the personal sector under national accounts definitions and is distinct from the stock of wealth in savings accounts.

Build the national accounts identity

Take an illustrative $100 of disposable income, $92 of consumption, $2 of personal interest payments and $1 of personal current transfer payments. Outlays total $95 and saving is $5, so the rate is 5%. Looking at consumption alone would imply $8 left over and a misleading 8% rate. The expenditure side is broader than a card-spending series.

Now lift disposable income to $104 and outlays to $100. Saving falls from $5 to $4 and the rate becomes 4 / 104, about 3.85%. The level of income rose. The fall in the saving rate can therefore reflect a change in spending and other outlays relative to income, rather than an outright drop in after-tax income.

Read units and revisions before attaching a market story

The monthly US nominal income and saving levels used here are generally published at seasonally adjusted annual rates in billions of dollars. Use matching units for both numerator and denominator. The common annual-rate factor cancels in the ratio. Adding twelve annualized monthly flows would overstate a calendar-year amount unless the frequency conversion is handled correctly.

Government transfers, taxes and income revisions can move the rate materially. National totals say little about the distribution of emergency cash across households. For the dollar, read the wider personal income and outlays release, including inflation and spending surprises. A historical saving-rate change and an exchange-rate move during the same reference month do not prove that markets reacted to the later publication.

Why a lower saving rate can accompany strong spending

A household or sector can finance stronger outlays from higher current income, lower saving or borrowing. The national accounting identity identifies the saving residual but does not identify how every individual financed a purchase. A one-time tax payment can lower disposable income and the saving rate despite unchanged underlying consumption. A transfer payment can produce the reverse. Examine disposable income, saving and outlays separately before attributing a change to confidence or financial distress.

If disposable income is reported at $20 trillion annualized in a month and saving at $1 trillion annualized, the ratio is 5%. Dividing only the income by twelve would manufacture a 60% rate. Divide both by twelve if you need approximate monthly flows in dollars before comparing them. For an annual saving rate, sum or average the matching flows first and divide. An unweighted average of monthly saving rates is generally a different statistic when income varies materially across the year.

A saving flow and a savings account tell different stories

Suppose a household sells an existing stock holding and deposits the proceeds into a savings account. Its bank balance rises. The sale converts one asset into another and does not by itself create new personal income and saving in the national accounts. Conversely, a household can save from current income by paying down a liability without adding to a savings account. These examples explain why a chart of deposits and the BEA personal saving rate can diverge without either series being wrong.

Imagine two months with disposable income of $100 and $200, and saving of $10 and $10. Monthly saving rates are 10% and 5%. Their simple average is 7.5%, while the combined rate is $20 / $300, or about 6.67%. The combined calculation gives the higher-income month more weight. This matters when income fluctuates sharply because of taxes or transfers. For a national annual comparison, use the correctly aggregated flows and the matching vintage instead of taking an unweighted average of published monthly rates.

The BEA personal sector is an aggregate. Even a rise in national saving could coexist with stress in households facing large rent, food or debt-service bills. The saving rate does not report median checking balances or the fraction of families with enough emergency cash. A question about distribution needs household-level evidence; a question about the macro contribution of consumption needs national-account detail. State which question you are answering before interpreting a one-month change as either strength or distress.

Revisions can alter each part of the identity. New income estimates may revise disposable income, and taxes or expenditure updates may revise outlays. Looking at an old saving-rate print and current income data can leave a discrepancy that disappears with a consistent vintage. If the interest is the dollar reaction to the original release, retain both the first published figures and later revisions in separate columns. A spreadsheet that overwrites the first estimate with history from a later download cannot reconstruct the market surprise on its own.

Historical workbook: Personal saving as percent of disposable income

The workbook follows the us personal sector’s income left after outlays from a US dollar perspective. Its question is specific: Did saving change in dollars or only as a share of income? Keep three columns beside the dates. Column one is monthly us disposable personal income in billions at annual rate. Column two is monthly us personal saving in billions at annual rate. The calculated column shows personal saving as percent of disposable income. Using the same definitions in every row makes it possible to explain an outlier by returning to the inputs instead of changing the story after the price has moved.

Divide monthly personal saving by monthly disposable personal income and multiply by 100. Both nominal levels are seasonally adjusted annual rates in billions of dollars. Compare the calculated value with the rounded BEA saving rate. A rise or fall in the rate alone does not tell you whether saving rose in dollars.

There are 120 monthly observations from 2015-01 through 2024-12. The historical FX observations come from DEXUSEU. For the independent dollar comparison below, one unit of EUR is always expressed in US dollars. Where the published quote gives foreign currency per dollar, the comparison takes its reciprocal. This keeps a rise in the comparison series synonymous with appreciation of the named foreign currency against USD. The calculation of the main worksheet stays in its original units.

Selected records from the workbook. The table displays rounded values. Statistics and reconstructions use the unrounded inputs.
PeriodMonthly US disposable personal income in billions at annual rateMonthly US personal saving in billions at annual ratePersonal saving as percent of disposable income
2015-0113,335.7000835.60006.2659
2016-0513,834.7000732.00005.2910
2019-1216,358.10001,020.10006.2361
2020-0418,703.30005,955.200031.8404
2020-0717,987.10003,213.000017.8628
2022-0618,735.8000444.00002.3698
2024-1222,450.50001,135.30005.0569

Read one row from the source to the answer

Start with 2019-12, near the chronological middle of the sample. The two inputs were 16,358.1000 for monthly us disposable personal income in billions at annual rate and 1,020.1000 for monthly us personal saving in billions at annual rate. Applying the stated rule gives 6.2361 for the saving share. A reader reconstructing the published number should carry more precision than the table displays. The market comparison used an FX mark on 2019-12-31, quoted at 1.1227 in the ordinary EUR/USD convention. It is not necessarily the observation time of the other inputs.

Disposable income of 16,358.1 billion and saving of 1,020.1 billion, both at annual rates, yield a saving share of 6.236%. The BEA rounded published rate for this reference month is 6.2%. The implied annualized personal outlays figure is 15,338.0 billion dollars. This identity includes more than personal consumption expenditures. Using a narrower consumption series as a direct replacement can leave an apparently unexplained gap.

Hold disposable income fixed and increase annual-rate personal outlays by $100 billion. Saving would fall by the same $100 billion, taking the modeled saving share to 5.625%. Raise both income and outlays by $100 billion, and dollar saving stays constant while its share becomes 6.198%. This distinction between a dollar amount and a share is precisely why a falling saving rate can accompany several different household-sector narratives.

Which input explains the largest change?

Across adjacent months the largest absolute change in the saving share occurred between 2020-03 and 2020-04. The calculated value went from 12.4023 to 31.8404, a signed difference of 19.4381. A headline describing the two end values is accurate but incomplete. To identify what generated the change, first update only the first column. Then update the second while holding the new first column fixed. This procedure accounts for the whole change without assuming which input moved first in real time.

The first input went from 16,370.4000 to 18,703.3000. With the second held at its starting value of 2,030.3000, that gives a hypothetical intermediate result of 10.8553. The first step therefore contributes -1.5470 to the measured change. Replacing the remaining old input with 5,955.2000 contributes another 20.9851. Add the two signed contributions to recover 19.4381 before display rounding. A negative contribution offsets a positive one even if both raw inputs rose.

Now reverse the order. Updating the second input first produces 36.3779. Its first-step contribution under this order is 23.9756, and the remaining contribution is -4.5375. Addition and subtraction formulas give the same attribution either way. Products and quotients can distribute a joint interaction differently. If the two allocations differ, report the order you used or average the two contributions for each input. Both reconciliations describe an identity. Neither establishes an economic cause. Using matching rates of flow preserves the intended denominator. A change in national saving may coexist with very different experiences across households.

Average of both replacement orders. The two allocations sum to the full change.
Measured changeAllocated to first inputAllocated to second input
19.4381-3.042222.4803

Use a typical period to test a dramatic headline

Ordering the calculated record by level gives a minimum of 2022-06 (2.3698) and a maximum of 2020-04 (31.8404). The distance between them is 29.4706 in the units of the worksheet. The mean is 7.3048 and the median is 6.1634. A mean gives every period equal weight and includes the influence of extreme observations. A median locates the middle of the ordering. Neither measures the experience of a specific US account with transactions concentrated on one or two dates.

A second question is how much the saving share generally changed between adjacent observations. Its median absolute change was 0.2448. The transition from 2016-04 to 2016-05 lies closest to that amount, going from 5.5359 to 5.2910. In that interval the first input moved by 12.1000 and the second by -33.2000 in their respective units. Compare those with the largest-change interval. The same final-sized change can come from modest moves in both inputs or large movements that largely offset.

For the high observation, ask what the same calculation would have returned if only the first input had taken its median sample value. Using 16,385.3000 in place of 18,703.3000 while keeping the second input at 5,955.2000 gives 36.3448. This is a deliberately constructed comparison. Those values need never have occurred together. If an input is contractually fixed or mechanically related to the other one, interpret the substitution as an arithmetic sensitivity exercise rather than a market scenario that could have been traded.

Does the dollar chart agree with the worksheet?

The largest absolute adjacent change in the dollar value of one EUR occurred from 2020-06 to 2020-07. The market comparison changed from 1.123700 to 1.182200 USD per unit, or 5.21%. By contrast, the worksheet moved by -0.4838 in its own units. A percentage currency return and a change in the saving share are different measurements. Comparing their signs is meaningful only after specifying the direction and recognizing when one input is itself a transformation of the exchange rate.

Of the 119 adjacent pairs, 119 had a measurable change in both series. In 60 they moved in the same signed direction, and in 59 they moved in opposite directions. The remainder had a flat result in at least one column at the stated numerical tolerance. These counts describe a sample, with no assumption of independence between observations. Where the translation formula already uses the exchange rate, part or all of the association can be mechanical. An agreement count alone is unsuitable as a trading win rate.

Compare a period where the signs explicitly disagreed. From 2020-06 to 2020-07, the dollar value of one EUR moved 5.21% and the worksheet changed by -0.4838. The first input ended at 17,987.1000 and the second at 3,213.0000. Record the observation even if it challenges your preferred narrative. Then return to bea personal income and outlays accounts to test whether the disagreement reflects definitions, different sampling dates or a genuinely different economic exposure. The outcome by itself does not choose between these explanations.

What was knowable before the last observation?

Keep the last month outside the record when setting a reference. The earlier 119 observations had a median worksheet value of 6.1730. The last result was 5.0569 in 2024-12. A simple baseline using the immediately preceding value, 5.6825, misses the last result by -0.6256 on an actual-minus-baseline convention. The earlier median misses by -1.1161. These baselines answer different questions about stability and representativeness.

Extend the last-value test across every adjacent pair and take the mean absolute difference. The result is 1.0199, compared with a median absolute difference of 0.2448. The mean gives unusually large misses their full weight. The median describes a typical miss. When you evaluate a real budget, multiply a per-unit figure only by its proper exposure, and account for any amount the formula already includes. Multiplying a dollar total by the entire invoice a second time would produce a fictitious level of risk.

These tests run on one documented set of historical observations. A revised series, an annual value compiled after year-end or a market quote taken after a release can include information unavailable at the hypothetical decision. To turn the worksheet into a real-time study, save the exact version that was available, the timestamp at which it became available and the quote at which a decision could be implemented. Keep a dated bea release and matching saving and income observations with that file. A comparison that relies on a later release remains a historical reconciliation even if it happens to fit the chart perfectly.

What level of dollar saving corresponds to half a percentage point?

Hold disposable income fixed at its observed annual-rate level and raise the saving rate by 0.5 percentage point. The required nominal change in saving equals income times 0.005. Divide the resulting annual-rate amount by twelve for an approximate one-month flow under the same convention.

The output column below is billions of dollars at annual rate equivalent to 0.5 point of disposable income. This is a second calculation on the same dated records. State the assumption before reading the result so it is clear which parts of the comparison come from observations and which come from the proposed decision. The smallest calculated result occurs in 2015-01, at 66.6785. The largest occurs in 2024-12, at 112.2525. A negative result, where the formula allows one, means the order of the comparison matters. It does not mean that the source supplied a negative market price.

Billions of dollars at annual rate equivalent to 0.5 point of disposable income. Repeated selections are shown once.
PeriodFirst inputSecond inputOriginal worksheet resultSecond calculation
2015-0113,335.7000835.60006.265966.6785
2019-1116,400.20001,127.60006.875582.0010
2024-1222,450.50001,135.30005.0569112.2525

For a central case use 2019-11, where the result is 82.0010, against a full-sample median of 81.9265. The associated FX observation was 1.10190 in the EUR/USD convention on 2019-11-29. Compare it with the largest case rather than simply picking the most recent observation. If an assumed quantity, policy or percentage in the calculation would have been different at the time, describe the result as a controlled comparison across historical conditions. Avoid presenting it as a transaction the source actually recorded.

The first 60 records, 2015-01 through 2019-12, average 73.6942 on this second calculation. The later 60 records, 2020-01 through 2024-12, average 97.9585. The signed difference is 24.2643. This gives equal weight to each retained observation and asks how the same rule behaved over two consecutive intervals. It does not adjust for changes in business activity or prove that the start of the later interval marks a regime change. The sign of this controlled adjustment is known by construction. The historical saving rate can still fall if income and outlays change together in another pattern.

Can a small discrepancy come from rounding alone?

Compute personal saving divided by disposable personal income from matching monthly nominal annual-rate series. Use unrounded values and compare with the published saving rate at its displayed precision. A few hundredths of a percentage point can disappear when rounded. A large discrepancy warrants checking units, frequency and vintage first.

Then calculate implied personal outlays from income less saving. This additional reconciliation helps distinguish an income story from an outlays story and avoids confusing household asset appreciation with current personal saving.

How to identify the price setup on a EUR/USD chart

A US saving-rate change is an accounting result in a monthly release. The five-minute EUR/USD exercise below specifies how a reader could rehearse a distinct directional idea on a chart. Higher saving does not establish that dollar offers disappeared at any given price.

Use a five-minute demo chart with a named time zone, such as America/New_York. The first decision is where an existing swing could attract orders. In the drawing the prior high sits at 1.11932. Mark it while the chart still ends before the next challenge. The subsequent wick reaches 1.12021, above the earlier reference, and the market returns through it. ICT describes that sequence as a liquidity sweep. Candles show traded or quoted prices on a feed, so the sequence does not establish any individual institution’s identity or the number of orders at the swing.

The next test is displacement through the earlier internal swing at 1.11710. In this schematic the outer wicks of bars five and seven leave a three-candle fair value gap between 1.11725 and 1.11779. Write down both boundaries before testing the retest. A wick across the initial level without displacement is a different observation. A displacement without a later return can leave a correctly identified setup unfilled. Preserve each outcome so a replay does not silently select only the attractive-looking retracements.

Chart 1. Identify the sweep and three-candle imbalance
EUR/USD | hypothetical spot pricesPrior high 1.11932Sweep 1.12021Prior swing 1.11710Gap 1.11779 / 1.11725Green: up Red: down Blue: reference Purple: sweep Gold: gap
Schematic EUR/USD bars. Compare the wick of bar five with the wick of bar seven. The middle bar provides displacement. The shaded interval records the gap between the outside wicks.

How to trade the retest on demo

For the hypothetical short position, the order waits at 1.11754 after the gap has formed. The protective stop is 1.12071, beyond the swept extreme. The target is 1.11087, which requires the quoted pair to move lower. Cancel the order if price reaches invalidation before entry, the agreed observation window expires, or the executable spread makes the original risk budget infeasible. A chart mid touching a line does not prove the customer’s bid or offer filled. In a live record those two events need separate timestamps.

The second drawing shows one favorable branch solely to establish the accounting. A losing branch exits at the stop under an exact-fill assumption. A third branch never enters and incurs no modeled round-trip charge. If stop and target both lie within one observed bar, the wick alone does not reveal their order. Seek finer time-stamped evidence or retain an ambiguous result rather than labeling the trade a winner. For an event study, select the exit, expiry and treatment of ambiguous sequences before viewing the release-day chart.

Preserve the release-stage inflation and spending details as well as the saving rate. A hypothetical stop-out on EUR/USD does not tell you which component of the income and outlays report investors valued most.

Chart 2. Price the entry, protective stop and target
EUR/USD | hypothetical spot pricesEntry 1.11754Stop 1.12071Target 1.11087Green: up Red: down Blue: reference Purple: target
Illustrative continuation after a retest. Lines show the planned short order. The favorable path is one possible sequence and carries no probability estimate.

Worked trade in a dollar-denominated demo account

Illustrative position. The prices are separate from the historical observations and the outcome has no assigned probability.
FieldHypothetical execution
Pair and sideEUR/USD short
Size27,000 EUR
Entry1.11754
Protective stop1.12071
Target1.11087
Gross risk and potential reward$85.59 and $180.09
Assumed round-trip charges$3.93

First subtract the entry and stop in absolute price units. The distance is 0.00317 USD per EUR, compared with 0.00667 to the target. Multiply each by 27,000 base units. That yields 85.5900 USD of gross risk and 180.0900 USD of potential gross gain. The figures are already in the USD account currency. Keep the full numerical precision through this step.

Gross planned risk is $85.59 and potential gain is $180.09. Divide gain by risk: $180.09 / $85.59 = 2.10. On a risk-to-reward convention that is 1:2.10. Now assume $3.93 in total round-trip charges in either completed branch, with executable entry and exit prices in the example. Net risk is $85.59 plus $3.93, or $89.52. Net potential reward is $180.09 minus $3.93, or $176.16. The net R:R is 1:1.97 from $176.16 / $89.52. These ratios use full-precision amounts before display rounding.

If the stop slipped one further pip, the modeled exit would move to 1.12081 and the completed loss including the assumed charge would become $92.22. Compare this with the original $89.52. Slippage, funding and market access can all affect an actual instrument. The purpose of this branch is to make the size and account-currency conversion explicit. A promising reward-to-risk ratio says nothing about the frequency with which this particular entry, protective exit and target are actually executable.

To annotate the rehearsal, mark the existing swing first, then identify displacement and the interval between the outer candle wicks. For definitions and session context, see the school’s lessons on fair value gaps, order blocks, the London session, market structure shifts, and the Judas swing.

Sources and calculation notes

BEA personal saving rate. BEA personal income and outlays.

Historical series and original records. DEXUSEU series, units and frequency · DSPI series, units and frequency · PMSAVE series, units and frequency · PSAVERT series, units and frequency. The calculations use observations available when checked on November 15, 2026. Subsequent revisions can change retrospective comparisons.

Frequently asked questions

Is the personal saving rate the interest rate on a savings account?

No. It measures the share of disposable income that remains after personal outlays.

Are capital gains personal saving in this identity?

Changes in the market value of assets are distinct from the current-income and outlays flow.

Can income rise while the saving rate falls?

Yes, if outlays increase enough relative to disposable income.

Why use disposable rather than total personal income?

Disposable personal income is the after-personal-current-tax measure used in the BEA saving identity.

Does the rate describe a typical household’s available cash?

It is a national aggregate. Distribution, debt service and household balance sheets require additional evidence.

To take the release from the data table to a dated trading record, use US Retail Sales and Forex, Core PCE and EUR/USD, the news-release filter, and hindsight-free replay.