1

The multiplier describes the response of output (GDP) to an autonomous

change in spending.
An important question about a national economy the size of the multiplier. The multiplier is the ratio between an autonomous change in
spending that is a change in spending unrelated to income and
the resulting change in GDP. The autonomous change might be an
increase or decrease in government spending; a change in investment
due to a change in expected profits, financial conditions, or some
other factor independent of current output; or a change in exports.
When we are discussing business cycles, we are often interested in
how much GDP responds to changes in investment; when we are
discussing the government budget, we are interested in the change
in GDP in response to a change in government spending. But the basic question is the same: how much does GDP rise in response to an
increase in spending, and how much does GDP fall in response to a
decline in spending?
For example, if we think that an increase in government spending of $10 billion will result in GDP rising by $15 billion, then the
multiplier would be 1.5. (In this case, there would be $5 billion of additional private spending, on top of the government spending.) The
multiplier will depend on the country, on economic conditions, and
on the specific budget changes involved. But there are some things
we can say about it in general.
Here is one way to think about the multiplier. Suppose the city
of New York begins some major new expenditure on expanding
the subway, lets say. The subway expansion itself counts in G, government final expenditure, and adds to GDP. But each person who
receives an income from the project individual employees on the
project, obviously, but also the owners of businesses that contract
with the city will spend some of that income. (The employee might
buy a new winter coat; the contractor might buy a new summer
house.) Some of the income will be saved, or spent on imported
goods, or paid in taxes but some will be spent on local goods and
services. The sellers of those goods and services will then receive
income, some of which will be spent locally in turn. And the people
who receive that income will spend some of it, and so on. Because
there are leakages at each stage, the increase in spending will come
to an end eventually. But it will be greater than original spending.
The multiplier tells us, if we add up all the additional spending how
does it compare to the original government spending that started the
process? It is clear that people spend a lot of their income on locally
produced goods and services, the multiplier will be high; if large
parts of income are saved, taxed, and spent on imported goods, the

Multiplier. The relationship between

a change in investment, government
spending, or other autonomous expenditure, and the change in output
that results from it. Mathematically,
Y
the multiplier can be expressed as A
where Y is output, A is autonomous
expenditure, and means change.
Autonomous. Describes a change in
spending that is independent of current
income.

Leakage. Uses of income that do

not contribute to aggregate demand,
and do not directly create income for
other units in the economy. The most
important leakages are saving, imports
and tax payments.

multiplier will be low.

The key assumption behind the multiplier is that peoples consumption depends mostly on their current income. There are two
reasons for this link. First, many households (and businesses) are
liquidity-constrained they would like to be spending more than
they currently are, but they cannot because they dont have savings
available and they cannot (or do not want to) borrow more. For anyone in this position, higher income will increase spending simply
because it allows them spend more. Second, even people who have
savings or could borrow more, still have to decide what is a reasonable level of spending. Since we cant predict the future, most people
or businesses use current income as a rough guide to what income is
likely to be in the future. So when income increases, people are likely
to want to spend more, unless they have strong reason to think that
the current increase is only temporary.
The most common use of the multiplier is to estimate the effects of
fiscal policy on output. The multiplier used in this way is referred to
as the fiscal multiplier. But the same analysis applies to any change
in autonomous spending spending that changes for reasons unrelated to current income. The multiplier describes how much GDP
changes in response to a change in investment, or a change in exports, just as much as it describes how much GDP changes in response to a change in government spending.
Its important to remember that the multiplier is describing the
relationship between changes in GDP and changes in autonomous
spending, not the level or growth rate of GDP itself. A large multiplier means that GDP is very responsive to changes in spending. A
large multiplier does mean that the level of GDP is high, or that GDP
is growing more rapidly. A large multiplier means that when some
spending in the economy rises, GDP will rise by a lot, and when
some spending falls, GDP will fall by a lot. A small multiplier, conversely, means that GDP does not change very much when spending
changes. When governments are considering the use of fiscal policy, a high multiplier is a good thing, since it will make policy more
effective. But when we are thinking about business cycles, a large
multiplier may be a bad thing, since it means that GDP will change a
lot whenever the level of business investment changes, creating more
instability.

The multiplier is the ratio of the final change in GDP to the initial change
in spending.
Mathematically, the multiplier is simply the ratio of the resulting
change in output to the initial change in spending. Using Y for out-

Liquidity-constrained. Describes a
household, business or other economic
unit that is spending less than it would
otherwise choose to because of a lack
of current access to cash or credit. This
is distinct from spending that is low
because of low income or wealth.

Fiscal multiplier. The multiplier

applied to changes in government
spending.
Autonomous spending. Spending that
does not depend on current income.

put, A for autonomous spending, and to mean change, we can

write:
multiplier =

Y
A

(1)

Or equivalently,
Y = multiplier A
This second form is usually more relevant we normally use the
multiplier to predict or estimate the change in GDP resulting from a
given change in spending..
For example, suppose you think the multiplier is 2. Then if there
is an autonomous increase in spending of $ 100 billion (for example,
an increase in government spending, an increase in exports, or an
increase in investment), then you would expect GDP to increase by
$100 billion 2 = $200 billion.
Similarly, if we can observe or estimate the change in output that
results from a change in autonomous spending, we can calculate the
multiplier. For example, one study of the effects of 2009 stimulus
bill looked at how much federal health care spending had increased
in various states. They found that for each $2 million of additional
government spending, output in that state was about $3 million
higher. This implies a multiplier of 32 , or 1.5.
Equation 1 is the definition of the multiplier it is just what the
multiplier means.

We can estimate the multiplier if we know what fraction of income is saved,

what fraction is spent on imports, and what fraction is taken in taxes.
We can make a quantitative estimate of how large the multiplier
might be, and what factors will make it larger or smaller. To do this,
we construct a simple model a set of equations that describe the
aspects of the economy that are relevant to the problem. By solving
the equaitons, we can come up with a number that we hope will
apply to the real world situation that the equations describe.
Our model starts with the national income identity:

Quantitative. Measured or estimated

numerically. The opposite is qualitative,
which refers to something that can be
described but not given as a number.
Model. A set of equations used by
economists and others to describe or
predict the behavior of some system in
the real world.

Y = C + I + G + (X M)
Next we make some behavioral assumptions. First, we will take
government spending, exports and (for the moment) investment as
exogenous that is, fixed or determined outside the model. (To say
something is exogenous means that we are taking it as given our
model does not try to explain it. The variables a model does explain
are called endogenous. In this case, Y, C and M are endogenous.)

Behavioral. Describing the choices of

the actors within a model. A behavioral
equation is one with parameters that
have to be estimated on the basis of
data. Behavioral equations may be more
or less accurate approximations of the
phenomena they describe, but they will
never be exactly correct.
Exogenous. Determined outside the
model. Variables that a model does
not try to explain, but simply takes as
given.
Endogenous. Determined within the
model. Variables that a model explains
that is, whose values depend on other
other variables in the model.

Another way of saying that a variable is exogenous is to say that

our model is treating it as fixed. Economists often convey the idea
that a variable is fixed by writing a bar over it. So if x is fixed, we
Using this notation, we write:
write it as x.
I = I
G = G
X = X
Consumption, on the other hand, is endogenous, because it depends
but
on current income. There is also an exogenous component, C,
most consumption spending, we think, is determined by current
income.
We write this relationship as
C = C + mpc YD
YD is disposable income, that is, income after taxes and transfers.
The parameter mpc is the marginal propensity to consume; it is
sometimes written as c. The equation tells us a story, or theory, about
consumption behaves
For example, if mpc = 0.75 then each dollar of additional disposable income would cause 75 cents of additional consumption
spending. So if disposable income in the economy as a whole rises
by $200 billion, then we would expect total consumption to rise by
$200 0.75 = $150 billion.
The equation includes YD , or disposable income, not Y. That is because we think consumption depends on the income actually available to households to spend. Disposable income does not include tax
payments, which are not available to households; but it does include
transfers like Social Security, unemployment insurance or interest
payments, which are treated by households as current income just as
wages and profits are. In other words:
YD = Y T + Tr
where T is total tax collections by the government and Tr is total
transfer payments to households.
We also think that taxes and transfer payments are closely linked
to current income. Income taxes are, of course, collected on income,
and sales taxes take a percentage of most purchases of goods and
services. Meanwhile, many transfer payments increase when income falls. For example, unemployment normally rises when output

Disposable income. Income available

to households after transfers and tax
payments.
Parameters. Numbers in an equation
that describe the relationships between
the variables.
Marginal propensity to consume. The
fraction of each additional dollar of
income that is spent on consumption.
c and mpc can be used interchangeably
as abbreviations for the marginal
propensity to consume.

falls, meaning higher unemployment insurance payments. So we can

write:
T Tr = tY
Combining this with the previous equation gives:
YD = (1 t)Y
We can think of t as the marginal tax rate on income.
Finally, we think a fixed share of spending goes to imports:

Marginal tax rate. The fraction of each

additional dollar of income that is taken
in taxes.

M = mY
The parameter m is the marginal propensity to import, or import
share. It represents the fraction of total spending that goes to imports.
For example, if m = 0.2, then 20 cents out of each new dollar spent
in the economy goes to imports, leaving the circular flow. So in this
case, if GDP increases by $50 billion, we would expect imports to
increase by $50 billion 0.2 = $10 billion. Notice that this equation
includes Y, not YD . Thats because imported goods may be used
for investment, government projects, or for exports, as well as for
consumption. So it makes sense to think of imports as a fraction of
total spending rather than a fraction of disposable income.
Next we substitute the equations for consumption, investment,
government spending, exports and imports back into the national
income identity:
Y = C + mpc YD + I + G + X mY = mpc(1 t)Y + I + G + X mY
Now we solve for Y:
Y mpc(1 t)Y + mY = C + I + G + X
Y=

1
(C + I + G + X )
1 mpc(1 t) + m

Now, lets combine the exogenous terms into a single variable,

autonomous spending, or A for short:
A = C + I + G + X
And lets write the equation in terms of changes (given as usual by
the symbol) rather than as levels.
Y =

1
A
1 mpc(1 t) + m

(2)

Import share. The fraction of total

spending on goods and services that
goes to imports.

Comparing Equation 2 to Equation 1, you can see that we have just

found the multiplier.
In some cases, we may want to ignore the effects of taxes and
trade, to keep things simpler. That is the same as setting t and m to
zero. So in that case, Equation 2 simplifies to:
Y =

1
A
1 mpc

Then the multiplier is just 11mpc .

You should use this form of the multiplier only when you are sure
that you dont need to worry about taxes or trade.
Otherwise, there is one further step to make Equation 2 a little
easier to work with. Instead of using mpc, the marginal propensity
to consume, lets use mpsthe marginal propensity to save. Since all
disposable income is either consumed or saved, we know that:
mpc + mps = 1
Or equivalently,
mps = 1 mpc
We sometimes write s rather than mps for the marginal propensity
to save.
Rewriting Equation 2 with mps instead of mpc in gives us:
Y=
So

1
mps+tmpst+m

1
A
mps + t mps t + m

(3)

is the general form of the multiplier.

For the US, a reasonable first guess for the multiplier might be around
1.5.
One reason to write the multiplier in terms of saving rather than
in terms of the marginal propensity to consume is that it makes it
clearer how the multiplier is connected to the circular flow of income.
Each of the terms in the denominator of Equation 3 corresponds to
a leakage out of the circular flow. Saving, paying taxes, and buying
imports are all alternatives to buying domestically produced goods;
they are all ways that income can be used that do not create income
for other households in the same country. So the greater the fraction
of income that is used in these ways, the fewer times a new dollar of
spending will circulate through the economy an the less additional
income it will create.
What we can see in Equation 3 is that the larger are the various
leakages, the smaller will be the multiplier. If savings, taxes and

The multiplier is given by

1
s + t st + m
where s is the marginal propensity to
save, t is the marginal tax rate, and m is
the marginal propensity to import.

imports are low, the multiplier will be large; if savings, taxes and
imports are high, the multiplier will be small. Its possible for the
multiplier to be less than one, but, in this simple model, it will always be positive. In other words, no matter how high are tax rates,
imports and saving, an increase in spending will always lead to some
increase in total output.
Lets fill in some plausible numbers for the US. Imports in the US
are about 15% of national income, but we know that the import share
reliably rises in booms and falls in recessions. In other words, the
marginal import propensity is higher than the average import propensity. Statistical evidence suggests that in the US, a 1% rise in income
typically leads to a 2% rise in imports. (In other words, the income
elasticity of imports seems to be around 2 in the US.) So a reasonable
value for m is around double the import share, or 0.3. Savings are
quite low in the US, and consumption responds strongly to current
income. In fact, some people will respond to an increase in income
by increasing their consumption by even more than the change in
income, implying a negative savings propensity. (This may happen
because a higher income makes it easier to borrow money, or because
some purchases, especially durable goods, are lumpy you have to
buy them all in one piece.) But overall, mps is certainly positive, especially since an important component of savings is retained earnings
corporate profits that are not paid out to shareholders. Overall,
mps = 0.2 is a reasonable first guess. Finally, federal taxes are around
15% of GDP, and state and local taxes are another 10%. Income taxes
of course vary with income. Some other taxes, like corporate profit
taxes, vary more than proportionately with income, while others, like
inheritance taxes, dont vary much with current income at all. Transfers also include payments that vary with income, like Medicaid and
unemployment benefits, and payments that dont vary with income,
including the two largest transfers, Medicare and Social Security.
Overall, a value of 0.2 seems reasonable for t as well.
Put these estimates together and we have:
Y = A(

Retained earnings. Profits that are kept

by the business that earned them, rather
than paid out to shareholders. Retained
earnings are an important form of
saving in the economy. Historically,
corporations have paid out about half
their profits and retained about half.

1
1
) = A(
) A 1.5
0.3 + 0.2 + 0.2 0.04
0.66

And in fact, 1.5 is quite close to many recent econometric estimates of the multiplier based on historical data.

Econometric. Statistical analysis of

economic data.