Household wealth
Whose balance sheet has strengthened, and through which assets?
MacroRoom·Measure by MacroRoom, from Federal Reserve DFA and BEA
Rising assets and falling liabilities can strengthen a household balance sheet. This measure shows how net worth changes across income and wealth groups, and which components account for the change.
2017 dollars / household · Q2 2026
Change in the group mean’s purchasing power.
DataPrimary measure through Q2 2026
Household net worth
Assets less liabilities per household, in constant purchasing-power dollars.
Separate Fed rankings by income and by net worth; groups overlap across the two views.
Quarterly distributional estimates, not quarterly household surveys. Component changes combine valuation, transactions and changing group membership; they are not causal stock-price effects.
Loading history…
| Group | Through | Level | Change |
|---|---|---|---|
| Bottom 20% | Q2 2026 | $144,900 | +1.5% |
| 20th–40th percentiles | Q2 2026 | $207,900 | +0.8% |
| Middle 20% | Q2 2026 | $393,100 | +2.1% |
| 60th–80th percentiles | Q2 2026 | $685,300 | +4.7% |
| 80th–99th percentiles | Q2 2026 | $2,607,700 | +7.3% |
| Top 1% | Q2 2026 | $26,373,500 | +13.2% |
Component data
Accessible chart data
Historical values load with the chart.
Methodology
Wealth per household
Net worth is assets less liabilities from the detailed Distributional Financial Accounts, divided by that source’s household count. Amounts are deflated with the complete quarter’s average PCE price index, in 2017 dollars. The Fed constructs quarterly estimates using financial accounts and survey information; these are not new household wealth surveys each quarter.
Income and wealth are separate ranking systems. Selecting the top 0.1% ranks by net worth; it does not define a fixed-dollar ultra-high-net-worth class. Across time, group membership can change. The mean amount should not be read as the median household’s wealth.
The combined top 1% is selectable on its own. Its totals and household counts combine the 99th–99.9th and top 0.1% groups; subgroup means are weighted by household counts. It is excluded from the default comparison so it is not presented alongside its own parts as an additional nonoverlapping group.
What changed on the balance sheet
The waterfall compares the selected group’s real per-household components at the breakdown date and the chosen interval earlier. Deposits, equities and mutual funds, retirement and annuities, real estate, and other assets are mutually exclusive. Liabilities enter with the opposite sign. Components reconcile to the change in net worth within source rounding.
These are accounting changes. They combine asset valuations, transactions, revisions and changing group composition; they do not isolate a causal stock-market or housing-price effect. Wealth gains are not automatically available for spending.
Dates, changes and limits
Every group retains its actual latest observation period. Comparisons use calendar dates; missing reference dates remain unavailable. Dollar and index changes are percentages, while rates and survey balances change in percentage points. Percent changes from a nonpositive base are unavailable. Annual observations stay annual.
The indexed view sets all selected groups to 100 at their first common positive observation within the visible window. Changing the group selection or history can change that baseline. The table always shows the underlying levels. A rising index describes improvement relative to that baseline, not an absolute standard of financial health.
Source releases can revise earlier observations. Retrieval dates, source hashes and cached-source flags are retained in the data manifest. A failed request retains the last available source with a visible note; a refreshed page does not imply that every underlying source was newly published.
Inputs and sources
Retrieved 2026-09-23
Retrieved 2026-09-23
Each chart names its measure’s source. Group definitions and dates can differ across the supporting panels.
Download
Download MacroRoom’s constructed measure (CSV)
Institutional supporting series are available from the original publishers above. The CSV preserves the period, population, units, estimation status and source links.