Professional Documents
Culture Documents
Pmak Session4
Pmak Session4
Discussion Session 4
Bopjun Gwak
June 4, 2018
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 1 / 32
Today’s Agenda
1 Levels Accounting
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 2 / 32
1 Levels Accounting
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 3 / 32
Levels Accounting
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 4 / 32
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 1 June 4, 2018 5 / 32
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 1 June 4, 2018 6 / 32
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 1 June 4, 2018 7 / 32
Levels Accounting
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 8 / 32
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 1 June 4, 2018 9 / 32
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 1 June 4, 2018 10 / 32
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 1 June 4, 2018 13 / 32
2 R&D-based endogenous growth model
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 14 / 32
R&D-based endogenous growth model
Capital accumulation
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 15 / 32
The production of new technology - φ
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 16 / 32
The production of new technology - λ
λ < 1: ”stepping on toes”, the more research there is, the more
probable that firms are working on the same ideas.
λ = 1: no ”stepping on toes” effect.
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 17 / 32
Technological Dynamics
At+1 − At
gt ≡ = ρAφ−1
t LλAt (1)
At
can be written as a first-order difference equation
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 18 / 32
Semi-endogenous Growth Model
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 19 / 32
Endogenous Growth Model
φ = 1 and n = 0
Constant growth rate of technology:
g = ρLλAt (6)
Scale effect: If sR is higher, then the growth rate of GDP (per worker)
in the long run is higher.
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 20 / 32
3 Business Cycles: The short run
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 21 / 32
Defining Business Cycles
Business Cycles
Fluctuations in the aggregate economic activity
A cycle consists of periods of expansion, followed by periods of
contractions in aggregate economic activity (trough - peak - trough)
A full business cycle lasts for more than one year
Business cycles are recurrent but not periodic
Recesssions
Negative growth of real GDP for at least two quarters
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 22 / 32
Measuring Business Cycles
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 23 / 32
The Hodrick-Prescott (HP) Filter
T
X T
X −1
min (yt − gt )2 +λ [(gt+1 − gt ) − (gt − gt−1 )]2
g
t=1 t=2
| {z } | {z }
A B
g
A: minimize the deviation of trend yt to actual output yt .
g
B: minimize the variation in the slope of yt (the slope of the long-run
growth component should not vary much).
The higher λ the higher the penalty for slope differences:
λ = 0: trend equals the original series; λ → ∞: linear trend (for
quarterly data: λ = 1600).
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 24 / 32
US Real GDP
Trend Component
16.8
16.4
log
16
15.6
15.2
1970 1975 1980 1985 1990 1995 2000 2005 2010
Cyclical Component
5
percentage dev. from trend
−1
−3
−5
1970 1975 1980 1985 1990 1995 2000 2005 2010
Data: St.Louis Fed FRED and SPF survey; US. Bureau of Labor Statistics.
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 25 / 32
Studying Business Cycles
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 26 / 32
US Private Consumption and Investment
Consumption and Output
5
Yc
3 Cc
1
percent
−1
−3
−5
1970 1975 1980 1985 1990 1995 2000 2005 2010
−10
−20
Data: St.Louis Fed FRED and SPF survey; US. Bureau of Labor Statistics.
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 27 / 32
Some Business Cycle facts:
Bopjun Gwak (Goethe University Frankfurt) PMAK - Discussion Session 1 June 4, 2018 28 / 32