Professional Documents
Culture Documents
Benjamin Wache
28 August 2019
Overview
Key aspects:
I Very stylized
I Capital accumulation and technology are the main drivers for
macroeconomic growth
I Constant savings rate
Structure of the model
where,
2 other factors (e.g., land, human capital) are ignored (for now)
Production function
Y (t)
y (t) =
A(t)L(t)
F (K (t), A(t)L(t))
y (t) =
A(t)L(t)
K (t)
y (t) = F ,1
A(t)L(t)
y (t) = F (k(t), 1)
y (t) = f (k(t)) (1)
Production function
CRS assumption...
F (K (t), A(t)L(t))
= K (t)α (A(t)L(t))−α
A(t)L(t)
!α
K (t)
F (k(t), 1) =
A(t)L(t)
f (k(t)) = k(t)α
Example
I f (0) = 0α = 0
I f 0 (k(t)) = αk(t)α−1 > 0
I f 00 (k(t)) = α(α − 1)k(t)α−2 < 0
Example
I MPK = αk(t)α−1
I MPL = (1 − α)Ak(t)α
MPL can be rewritten:
A[k(t)α − αk(t)α ]
X (t) − X (t − ∆)
∆
I if ∆ approaches 0, the change of X over time becomes:
∂X (t)
Ẋ (t) =
∂t
I We can now write a continuous time version of the growth
rate at time t as:
Ẋ (t)
g=
X (t)
Dynamics of the model
I Note that taking the time derivative of ln X(t) will also give
the growth rate:
I Given the initial values of A(0), L(0) and K (0) (all > 0),
assume that labour and technology will exhibit constant
growth rates (implying exponential growth)
I A(t) = A(0)e gt
I L(t) = L(0)e nt
I If we take logs and then differentiate with respect to time we
will obtain:
Ȧ(t)
=g
A(t)
L̇(t)
=n
L(t)
Savings and Investment
We can derive the path of k(t) applying the chain rule with respect
K (t)
time to k(t) = A(t)L(t)
K̇ (t) K (t)
k̇(t) = − [Ȧ(t)L(t) + A(t)L̇(t)]
A(t)L(t) [A(t)L(t)]2
Three possibilities:
I k(t) grows because there are more savings than investment
needed to keep k constant if
We will try to see what is the growth rate of the main variables
when k(t) is constant or when economy is on balanced growth
path.
I Given that K (t) = k(t)A(t)L(t) and k(t) is constant, capital
grows in steady state at the rate n + g
K (t)
L(t) which is capital per worker grows at rate g
I
Y (t)
L(t) which is output per worker grows at rate g
I
K (t)
I We can clearly see that k ∗ and y ∗ increase. Therefore, L(t)
Y (t)
and L(t) increase too.
I We cannot say anything about changes of consumption per
worker:
c ∗ = f (k ∗ ) − (n + g + δ)k ∗
∂c ∗ ∂k ∗
= [f 0 (k ∗ ) − (n + g + δ)]
∂s ∂s
∗
Remember that k is an increasing function of s. The result
depends on the relation between f 0 (k ∗ ) and n + g + δ
Transition
Figure 9: Jump
Transition
Researchers have tried to test if the capital per worker values (with
both a time series approach and across countries analysis)
predicted by the model match with the income distribution
I Over time: model implies that the capital-to-worker ratio is
approximately 10 times larger today than 100 years ago for
industrialized countries. Capital-to-output ratios are
approximately constant
I Across countries: model implies that the the capital-to-worker
ratio is approximately 20 to 30 times larger in industrial
countries, not 100. Capital-to-output ratios are approximately
2 to 3 times larger in industrial countries than in developing
countries
I observed K/L ratio differences are too small to account for
observed differences in income
Returns to capital
Obtaining consistent measures of variables, such as capital, over
time and across countries can be tricky. An alternative approach is
to look at its price (or return), which in theory should be less prone
to measurement error. Large differences in capital per worker
should then also imply huge difference in capital’s rate of return.
I If markets are competitive, the rate of return to capital should
equal its marginal product, f0 (k), less the depreciation rate, δ
α−1
I f0 (k)=αy α