Accelerator
Accelerator
Initial points
1. The model is a synthesis of the Kahn-Keynes multiplier and the accelerator theory of investment1. 2. The accelerator model is based on the truism that, if technology (and thus the capital/output ratio) is held constant, an increase in output can only be achieved though an increase in the capital stock.
P. Samuelson. Interaction Between the Multiplier Analysis and the Principle of Acceleration, Review of Economic Statistics (May 1939).
The accelerator
Firms need a given quantity of capital to produce the current level of output. If the level of output changes, they will need more capital. How much more? Change in capital = accelerator change in output (10.1)
But firms can only increase their capital stock by (positive) net investment. How much?
Net investment = accelerator change in output (10.2)
Capital/Output ratio
If we do not allow for productivity boosting technical change, then the capital output ratio is held constant.
If fact, this is what we are assumingno technical change.
We assume that = 3a . That is, it takes 3 dollars worth of capital to manufacture $1 worth of shoes.
Hence if the demand for shoes increased by say, $10, there would be a need for $30 in additional capitalor equivalently, $30 in net investment.
aSherman & Kolk claim this is a reasonable figure since estimates show that GDP
is typically equal to 1/3 the value of the capital stock.
Time period
1 1 to 2 2 2 to 3
Shoe Machinery
$300
$330 $60
3
3 to 4
$130
$5
$390
$15
4
4 to 5 5 5 to 6 6
$135
$0 $135 -5 $130
$405
$0 $405 -$15 $390
Yt = Ct + It
(1)
Ct C cYt 1
(2)
We assume that investment in the current period (It) is equal to some fraction () of change in output in the previous period (or lagged output):
It (Yt 1 Yt 2)
1
(3)
We assume that C depends on lagged, rather than current, income. Also note that for our simplified economy, Y = YD.
Yt C (c v)Yt 1 Yt 2
(4)
To get a homogenous equation, we ignore the constant C To get a standardized form, let A = c + . Also, Let B = . Thus we can write:
Yt At 1 Bt 2 0
(5)
Note for the mathematically inclined: equation (5) is a 2nd order (homogenous) difference equation.
It can be shown that: 1. There will be cyclical fluctuations in the time path of national income (Yt) if A2 < 4B. 2. If B = 1 (and presuming that A2 < 4B), then cycles are constant in amplitude.
3. If B < 1 (and presuming that A2 < 4B), then cycles are dampedthat is, amplitude is a decreasing function of time.
4. If B > 1 (and presuming that A2 < 4B), then cycles are explosivethat is, amplitude is a increasing function of time. 5. There will be no cyclical fluctuations if A2 > 4B.
$996 996 992 985 975 964 953 942 933 927 928 928 936 945 957 969 978 987 992
Assumptions: (1) Y is $996 in period 1 and $1000 in period 2; (2) C = 96 + .9Yt - 1; and (3) = 1
Multiplier-Accelerator Model Assumptions: (1) Y is $996 in period 1 and $1000 in period 2; (2) C = 996 + .9Yt -- 1; and (3) = B = 1 Data in Billions
1020
1000
980
960
940
920 900
11
13
15
17
19
21
Time Period
Damped oscillations
B < 1 and A2 > 4B
Time period
Explosive oscillations
Time period
Qualifications/limitations
This model is based on a crude theory of investment. There is no role for expected profits or animal spirits. The time lag between a change in output and a change in (net) investment can be significantthe investment process (planning, finance, procurement, manufacturing, installation, training) is often lengthy. J. Hicks pointed out that, for the economy as a whole, there is a limit to disinvestment (negative net investment). At the aggregate level, the limit to capital reduction in a given period is the wear and tear due to depreciation.