Skip to content

Velocity of money

Nominal GDP per unit of a chosen money stock: an income-velocity ratio whose money definition, time units and vintage matter.

Velocity of money is nominal income generated per unit of a chosen money stock over a stated period. The usual income velocity measure divides nominal GDP by a compatible average money supply. For a quant, it is a useful way to describe how nominal activity and money balances move together, provided the denominator and time convention stay explicit.

It is an aggregate accounting ratio. A reading of two per year does not establish that every physical dollar changed hands twice, or that households all spend at the same speed.

The flow divided by the stock

Write nominal GDP as PYPY, where PP is a compatible output price level and YY is real output. With MM denoting the average money stock for the period,

V=PYM,MV=PY.V = \frac{PY}{M}, \qquad MV = PY.
The numerator is a flow of production valued at current prices; the denominator is a stock of money balances. Using an annual GDP flow gives velocity in units of per year. GDP measures domestic production, so this income-velocity ratio does not count all payments: purchases of existing financial assets, for example, are not purchases of newly produced output (BEA on financial transactions). See the FRED M2 velocity notes for an official series construction.

Consider two fictional annual observations, with all monetary amounts in the same currency and in billions:

ObservationAnnual nominal GDPAverage money stockIncome velocity
A5002502 per year
B5502502.2 per year
The calculations are 500/250=2500/250=2 and 550/250=2.2550/250=2.2. Nominal output rises while the measured stock stays unchanged, so the ratio increases. This does not reveal how much of the output increase came from higher prices rather than real economic growth.
Fictional annual observations: nominal GDP rises from 500 to 550 billion, average money stock stays at 250 billion, and income velocity rises from 2 to 2.2 per year.
Fictional annual observations: nominal GDP rises from 500 to 550 billion, average money stock stays at 250 billion, and income velocity rises from 2 to 2.2 per year.
Fictional annual observations, same currency and money definition. A larger nominal GDP flow divided by an unchanged average money stock produces a higher income-velocity ratio.

Which money, and which period?

There is no definition-free velocity number. A ratio using M1 answers a different measurement question from a ratio using M2 or the monetary base. The Federal Reserve's H.6 definitions distinguish currency, deposit aggregates and reserve balances; substituting one for another changes the denominator.
Time units matter just as much. Suppose a fictional quarter contains nominal output of 125 billion and its average money stock is 250 billion. The ratio is 125/250=0.5125/250=0.5 for that quarter. Expressing the same output at an annual rate gives 4×125=5004\times125=500 billion per year and velocity of 500/250=2500/250=2 per year. These describe the same quarter under different rate conventions; the annual-rate version is not a forecast of the year's output.
The US FRED M2V series is quarterly and seasonally adjusted, using nominal GDP divided by quarterly average M2. Its GDP input is reported at a seasonally adjusted annual rate. Thus the observation frequency is quarterly while the numerator carries an annual-rate convention. When reproducing or comparing velocity, record the jurisdiction, money definition, stock averaging rule, seasonal adjustment and annualisation. Preserve the release vintage and availability date when using it in a historical signal: revised inputs were not necessarily known at the original decision date.

An identity needs a model before it becomes a prediction

Rearranging MV=PYMV=PY does not make VV fixed or establish which variable caused another to change. A claim that faster money growth must produce a particular inflation rate needs additional assumptions about velocity, real output and their responses. McCallum and Nelson make this distinction explicitly in Section 2 of Money and Inflation: Some Critical Issues.

For empirical work, treat an observed velocity change as something to explain. A fitted relationship can be useful, but stability is a hypothesis to test across money definitions, sample periods and vintages. The ratio alone supplies no universal inflation trigger or trading rule.

Sources

Last updated
  • money
  • macroeconomic measurement
  • nominal gdp
← Back to the glossaryBack to top ↑