Professional Documents
Culture Documents
Pmak Session3
Pmak Session3
Discussion Session 3
Bopjun Gwak
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 1 / 25
Today’s Agenda
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 2 / 25
The steady state of the general Solow model
α
1−α α
1−α
ŷ¯ = s
δ+n+g ; ȳ = At s
δ+n+g
1 1
¯
1−α 1−α
s s
k̂ = δ+n+g ; k̄ = At δ+n+g
ĉ¯ = (1 − s)ŷ¯
Return on capital:
ŷ¯
ȳ n+g+δ
r̄ = α ¯ = α = α
k̂ k̄ s
Note: The solutions for the per worker variables (ȳ, k̄ ) boil down to the
one of the basic Solow model if g = n = 0 for B = A1−α .
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 3 / 25
Growth rates in the steady state
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 4 / 25
Growth rates in the steady state
What is the growth rate of ȳt (output per worker) in the steady state?
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 4 / 25
Growth rates in the steady state
What is the growth rate of ȳt (output per worker) in the steady state?
Solution:
yt+1 At+1 yt
yt+1 − yt At+1 At − At ŷ¯ (1 + g) − ŷ¯
= yt = =g
yt At ŷ¯
Alternatively,
Yt
ŷt = AYt Lt t = ALtt = Aytt . Since ŷt is constant in SS, yt must grow at the
same rate as At , i.e. at g.
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 4 / 25
Growth rates in the steady state
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 5 / 25
Growth rates in the steady state
Alternatively,
ŷt = AYt Lt t . Since ŷt is constant in SS, then Yt must grow at the same
rate as At Lt , i.e. at n + g.
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 5 / 25
Growth rates in the steady state
Alternatively,
ŷt = AYt Lt t . Since ŷt is constant in SS, then Yt must grow at the same
rate as At Lt , i.e. at n + g.
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 5 / 25
Long-run growth in the general Solow model
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 6 / 25
Long-run growth in the general Solow model
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 6 / 25
Long-run growth in the general Solow model
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 6 / 25
Long-run growth in the general Solow model
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 6 / 25
Balanced Growth path
Balanced Growth
Output per worker yt , capital per worker kt , consumption per worker ct , and
the real wage wt grow at the same constant rate g. Total output Yt = yt Lt ,
total consumption Ct = ct Lt , total investment It = St = syt Lt , and the total
capital stock Kt = kt Lt grow at the same constant rate g + n + gn ≈ g + n.
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 7 / 25
Kaldor facts
Yt
1 GDP per worker Lt grows roughly at a constant rate over time X
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 8 / 25
Kaldor facts
Yt
1 GDP per worker Lt grows roughly at a constant rate over time X
Kt
2 Capital per worker Lt grows roughly at a constant rate over time X
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 8 / 25
Kaldor facts
Yt
1 GDP per worker Lt grows roughly at a constant rate over time X
Kt
2 Capital per worker Lt grows roughly at a constant rate over time X
3 The share of capital income rtYKt t and the share of labor income wt Lt
Yt in
GDP are roughly constant over time X
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 8 / 25
Kaldor facts
Yt
1 GDP per worker Lt grows roughly at a constant rate over time X
Kt
2 Capital per worker Lt grows roughly at a constant rate over time X
3 The share of capital income rtYKt t and the share of labor income wt Lt
Yt in
GDP are roughly constant over time X
Kt
4 The capital-output ratio Yt is roughly constant over time X
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 8 / 25
Kaldor facts
Yt
1 GDP per worker Lt grows roughly at a constant rate over time X
Kt
2 Capital per worker Lt grows roughly at a constant rate over time X
3 The share of capital income rtYKt t and the share of labor income wt Lt
Yt in
GDP are roughly constant over time X
Kt
4 The capital-output ratio Yt is roughly constant over time X
5 The return on capital rt is roughly constant over time X
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 8 / 25
Kaldor facts
Yt
1 GDP per worker Lt grows roughly at a constant rate over time X
Kt
2 Capital per worker Lt grows roughly at a constant rate over time X
3 The share of capital income rtYKt t and the share of labor income wt Lt
Yt in
GDP are roughly constant over time X
Kt
4 The capital-output ratio Yt is roughly constant over time X
5 The return on capital rt is roughly constant over time X
6 The real wage wt grows over time X
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 8 / 25
Labor-augmenting technological progress
Yt = F (Kt , Lt , At )
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 9 / 25
Why must techn. progress be labor-augmenting?
Consider
!
Kt 1
1=F , , At
Yt YL t
t
We know that KYtt is roughly constant over time, YLtt grows with a
constant rate g and At grows with a constant rate g.
Since the left-hand side of the equation above is constant, the inputs
on the right-hand side have to be constant too. This is only possible if
technological progress can be written as labor-augmenting in the
production function Yt = F (Kt , Lt , At ) = F (Kt , At Lt ).
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 10 / 25
Why must techn. progress be labor-augmenting?
Example 1
(the function cannot be written with labor-augmenting At ):
1
Multiply Yt by Yt
α !1−α
Kt At
1 = Yt
Aαt
|{z} Yt |{z}
const. | {z } | Lt {z } growing
const. const.
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 11 / 25
Why must techn. progress be labor-augmenting?
Example 2
(the function can be written with labor-augmenting At ):
Yt = F (Kt , Lt , At ) = A1−α
t Ktα L1−α
t
= Ktα (At Lt )1−α = F (Kt , At Lt )
1
Multiply Yt by Yt
α !1−α
Kt At
1 = Yt
|{z} Yt
const. | {z } | Lt {z }
const. const.
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 12 / 25
2. Growth accounting vs. levels accounting
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 13 / 25
Growth Accounting
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 14 / 25
Growth Accounting- some figures
Data: OECD
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 15 / 25
Levels Accounting
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 16 / 25
Levels Accounting
Hi = eφ(Ei ) Li
where
I φ(.): a not yet specified function
I Ei : average years of schooling in country i
I Li : labor input in country i
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 17 / 25
Levels Accounting
How do we specify φ(E)?
Production function of country i: Yi = Kiα (Ai eφ(Ei ) Li )1−α
Wages:
∂Yi
wi = = (1 − α)Kiα (Ai eφ(Ei ) Li )−α Ai eφ(Ei )
∂Li
= (1 − α)Kiα A1−α
i e(1−α)φ(Ei ) L−αi
Consider the wages of two workers (1,2) in the same country i with
different levels of education:
w1,i = (1 − α)Kiα Ai1−α e(1−α)φ(Ei,1 ) L−α
i ; w2,i = (1 − α)Kiα A1−α
i e(1−α)φ(Ei,2 ) L−α
i
w1,i
= e(1−α)[φ(E1,i )−φ(E2,i )] ⇔ ln(w1,i )−ln(w2,i ) = (1 − α)[φ(E1,i )−φ(E2,i )]
w2,i
⇒ We can compare wages of workers in one country to detect a
shape for φ(E).
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 18 / 25
Level Accounting
Hall and Jones (1998) describe φ(Ei ) as piecewise linear: For the first
4 years of education the rate of return is 13.4 %, for the next 4 years
10.1% and for education beyond 8 years 6.8 %.
0.134 Ei
if 0 ≤ Ei ≤ 4
φ(Ei ) = φ(4) + 0.101(Ei − 4) if 4 < Ei ≤ 8
φ(8) + 0.068(Ei − 8) if 8 < Ei
1.6
1.4
1.2
1
0.8
0.6
0.4
0.2
0
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Hall and Jones (1998) assume the same function φ(Ei ) for all
countries. Do you think this is plausible?
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 19 / 25
Levels Accounting
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 20 / 25
Levels Accounting
Hall and Jones (1998) find that productivity differences are the most
important factor to explain the differences in the levels of output per
worker between countries.
Output per worker in the five rich countries is about 30 times higher
than in the five poorest countries (figures for 1988).
I Differences in capital intensity and educational attainment across
countries are only about 2 times higher in the rich countries compared to
the poor.
I Productivity over 8 times higher in the rich countries.
However, differences in education matter as much as capital intensity!
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 21 / 25
Does educationTHE
matter for economic growth?
EMPIRICS OF ECONOMIC GROWTH 425
(1). Test for unconditional convergence:
TABLE III
TESTS FOR UNCONDITIONALCONVERGENCE
Note. Standard errors are in parentheses. Y60 is GDP per working-age person in 1960.
Table III from Mankiw, Romer and Weil (1992): ”A contribution to the empirics of economic
essentially zero. growth“,
There Quarterly
is no tendency
Journal of for poor countries to grow
Economics.
faster on average than rich countries.
Table III does show,
Bopjun Gwak (Goethe University Frankfurt)
however, that there is a significant
PMAK - Discussion Session 3 May 16, 2018 22 / 25
426 QUARTERLYJOURNAL OF ECONOMICS
Does education matter for economic growth?
TABLE IV
(2). Test for conditional convergence:
TESTS FOR CONDITIONALCONVERGENCE
Note. Standard errors are in parentheses. Y60 is GDP per working-age person in 1960. The investment and
population growth rates are averages for the period 1960-1985. (g + 8) is assumed to be 0.05.
Note. Standard errors are in parentheses. Y60 is GDP per working-age person in 1960. The investment and
population growth rates are averages for the period 1960-1985. (g + 8) is assumed to be 0.05. SCHOOL is the
average percentage of the working-age population in secondary school for the period 1960-1985.
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 3 May 16, 2018 25 / 25