M2 Money Supply: Reconcile the Measure Before Reading Markets

Learn the current U.S. M2 definition, May 2020 and July 2026 methodology changes, seasonal adjustment, revisions, nominal and real growth, component contributions, velocity, and limits of stock-market inference.

For educational purposes only; not investment advice. Investing may result in loss.

Direct answer

U.S. M2 is a Federal Reserve statistical money-stock aggregate, not all wealth, bank reserves, central-bank assets, available credit, or market liquidity. Under the July 2026 H.6 presentation, M2 equals M1 plus small-denomination time deposits and retail money-market-fund balances, less aggregate IRA and Keogh balances held at depository institutions and money funds.

M1 itself includes currency outside the Treasury, Federal Reserve Banks, and depository-institution vaults; specified demand deposits net of stated exclusions and collection items; and other liquid deposits, including checkable and savings deposits. Exact footnotes, exclusions, seasonal adjustment, units, frequency, vintage, and current H.6 methodology are part of the definition.

M2 is not a direct stock signal. Its level or growth can change because of fiscal flows, bank deposit creation or repayment, asset transactions, portfolio shifts, deposit pricing, seasonal patterns, statistical revisions, and definition changes. Equity effects depend on expectations, credit, spending, inflation, policy rates, term and risk premia, earnings, valuation, global dollar conditions, and what markets already priced.

How it works

Audit the series in this order:

Freeze the official definition and vintage. For the current construction, M2 = M1 + small time deposits + retail MMF balances - aggregate IRA and Keogh balances. Read every H.6 footnote and record release date, observation date, vintage, units, seasonal status, frequency, and whether a later benchmark, seasonal-factor, source-data, or netting revision changed history.
Respect classification breaks. Beginning May 2020, a Regulation D amendment removed the six-transfer limit on savings deposits; H.6 combined savings and other checkable deposits into other liquid deposits and included that category in M1 retroactively from May 2020. This raised M1 through reclassification but did not itself change M2. Beginning with the July 28, 2026 release, IRA and Keogh balances are netted once at aggregate M2 rather than from two components; non-seasonally adjusted M2 was unchanged, while seasonally adjusted M2 could receive minor revisions.
Reconcile components before interpreting totals. Separate currency, demand deposits, other liquid deposits, small time deposits, retail money funds, and retirement-account deductions. A transfer between two included components can change composition without changing M2; bank loan creation can create a deposit, while loan repayment can extinguish one, but government, securities, and cross-border transactions require full balance-sheet tracing.
Choose the statistic deliberately. Compare levels, dollar changes, component contributions, month-over-month or year-over-year growth, seasonally adjusted or unadjusted data, and annualized or nonannualized rates on consistent dates. For a simple nonannualized change, growth = (M2_t / M2_t-h) - 1; do not mix a one-month rate with a twelve-month rate.
Deflate matched nominal balances carefully. A diagnostic real-money-balance index is real M2 index = nominal M2 / price index × base. Match dates, frequency, seasonal treatment, price measure, revisions, and base. This quotient is not household purchasing power, investable cash, excess liquidity, or a causal stock valuation factor.
Calculate velocity with the published convention. FRED’s quarterly M2 velocity is M2V = quarterly nominal GDP at a seasonally adjusted annual rate / quarterly average M2. It combines an annualized flow with an average stock, so interpret the ratio under that convention; it is not a direct count of transactions or proof that a dollar literally changed hands that many times in the quarter.
Test transmission rather than chart resemblance. Map the change to fiscal balances, lending, deposits, money funds, rates, credit, velocity, inflation, nominal activity, earnings, risk premia, currencies, and lags. Use timestamped vintages, leads and lags, subperiods, alternative specifications, out-of-sample tests, and economic mechanisms before making any equity claim.

M2 is a domestic statistical aggregate. Cross-country labels are not automatically comparable because definitions, institutions, reserve frameworks, money funds, deposit insurance, capital controls, currencies, seasonal methods, reporting frequency, and revisions differ.

Example

Use a stylized reconciliation to keep level, growth, real balances, velocity, and composition separate:

Current-definition total: suppose M1 is $18,500.0000 billion, small time deposits are $1,100.0000 billion, retail money-market funds are $2,200.0000 billion, and aggregate IRA and Keogh balances are $300.0000 billion. Then M2 = 18,500 + 1,100 + 2,200 - 300 = $21,500.0000 billion. Do not subtract retirement balances again from the two gross components.
Nominal and real growth: if prior M2 was $20,800.0000 billion, nominal growth is (21,500 / 20,800) - 1 = 3.3654%. If a matched price index rises from 300.0000 to 309.0000, the real-balance index moves from 20,800 / 300 × 100 = 6,933.3333 to 21,500 / 309 × 100 = 6,957.9288, only 0.3547% growth. The chosen price index and revisions matter.
Velocity convention: if quarterly nominal GDP at a seasonally adjusted annual rate is $30,000.0000 billion and quarterly average M2 is $21,500.0000 billion, M2V = 30,000 / 21,500 = 1.3953. Do not divide one quarter’s nonannualized spending by a month-end M2 observation and call the result the same series.
Composition without new aggregate money: moving $400.0000 billion from an included savings deposit to an included retail money fund is Δ savings = -$400.0000 billion and Δ retail MMF = +$400.0000 billion, so the direct transfer’s ΔM2 = $0.0000 before other transactions, valuation, timing, classification, or measurement effects. A change in one component is not automatically a change in total M2.

Risks

Use the official H.6 definition, footnotes, exclusions, and current netting method.
Record observation date, release date, retrieval date, vintage, units, and frequency.
Distinguish seasonally adjusted from not seasonally adjusted series.
Expect annual seasonal-factor, benchmark, source-data, and historical revisions.
Treat May 2020 M1 reclassification separately from a change in M2.
Treat the July 2026 retirement-balance netting change separately from economic flows.
Reconcile currency, deposits, small time deposits, retail money funds, and deductions.
Do not add gross components and then deduct IRA and Keogh balances twice.
Separate component migration from creation or destruction of aggregate deposits.
Match month-end, monthly average, quarterly average, and quarterly-flow dates.
Label month-over-month, year-over-year, annualized, and nonannualized growth correctly.
Match nominal M2 with a stated price index before calculating a real diagnostic.
Use the published annualized-GDP over quarterly-average-M2 convention for M2 velocity.
Do not interpret velocity as a directly observed transaction count or stable constant.
Separate M2, the monetary base, reserves, central-bank assets, credit, wealth, and market liquidity.
Trace fiscal, bank, money-fund, securities, foreign, and portfolio-shift balance sheets.
Test rates, credit, inflation, earnings, risk premia, currency, velocity, and policy expectations.
Control leads, lags, overlapping crises, structural breaks, trend, and common causes.
Do not compare countries without reconciling definitions, institutions, currency, and methods.
Require robust out-of-sample evidence before using M2 in an investment decision.

Common misconceptions

“Every increase in M2 is new money entering stocks.” Deposits may be held, spent, repaid, transferred among included components, or associated with transactions that never purchase equities.
“M2 is the Fed’s balance sheet or bank reserves.” These are distinct aggregates with different holders, components, and transmission mechanisms.
“The 2020 jump in M1 proves an equal economic liquidity shock.” Much of the discontinuity reflects the reclassification of savings deposits after the Regulation D change; M2 was not raised by that reclassification itself.
“M2 velocity measures how often each dollar transacts in a quarter.” The published ratio divides annualized quarterly nominal GDP by quarterly average M2 and is an aggregate accounting ratio.
“A close M2-stock chart proves causality.” Trend, policy, fiscal actions, crises, inflation, earnings, discount rates, lags, revisions, and sample choice can generate misleading correlation.

Sources

Federal Reserve Board: Money Stock Measures - H.6.
Federal Reserve Board: Money Stock Measures - H.6 Technical Q&As.
Federal Reserve Bank of St. Louis FRED: M2 Money Stock.
Federal Reserve Bank of St. Louis FRED: M2 Money Stock, Not Seasonally Adjusted.
Federal Reserve Bank of St. Louis FRED: Velocity of M2 Money Stock.
U.S. Bureau of Economic Analysis: Gross Domestic Product.