Transcript Human Capital, Industrial Growth and Resource Curse

Human Capital, Industrial Growth
and Resource Curse
by Elena Suslova (NES’06)
and
Natalya Volchkova
(CEFIR at NES)
Motivation: cracking the “black box”
of “resource curse”
• High oil prices → “resource curse”
• Is there a charm against the curse?
• Literature – provide us with a number of
speculations on the origin and propagation
of the curse. Little of empirical backup.
• Test various channels of “resource curse”
transmission
• Policy application – case of Russia
Literature: are natural resources a
blessing or a curse?
 Three channels of negative effects transmission
 Dutch Disease
 Sachs&Warner’95, 97, 99: cross country studies revealed
negative relation between natural resource abundance and
growth rates
 Spatafora&Warner’95, Hutchison’94: time series analysis
does not confirm diagnosis
 Political Economy
 Auty’01, Paldam&Svendse’00: huge rents provoke rentseeking, corruption, postponement of reforms, competitive
industrialization
 Human Capital Development
 Gylfason’01, Leamer et al.’99: resource intense sectors
absorb national savings while creating only a few eminently
qualified jobs, thus preventing the development of innovative
industries
Research question:
• Is there a human capital
underdevelopment channel within the
“black box” of “resource curse”?
How to crack the box?
• Cross-country growth studies:
– omitted variables problem
– endogeneity issues
– failure to distinguished among possible mechanisms
of transmission
• Difference in differences - cross-country and
cross-industry growth - study a la
Rajan&Zingales’98: is there a growth
disadvantage of industries that are more human
capital intensive in the resource abundant
economies?
– fixed country and industry effects
– mostly exogenous explanatory variables
– model the transmission mechanism
Model: Leamer et al’99
• Assumptions
– Growth mechanism: capital accumulation–
both physical and human
– Open economy: production pattern is
determined by comparative advantage a la
Hecksher-Ohlin model
• Compare the implied dynamics of human
capital accumulation between two
countries: rich in natural resources vs.
poor in natural resources
Resource abundant economy
Higher (more skilled) human capital is required
Capital substitution makes labor cheaper
Physical capital accumulation
Capital accumulation makes labor more expensive
Higher (more skilled) human capital is required
Resource poor economy
Model prediction:
• Weaker private incentives to invest in
development of higher human capital in
resource rich economies compared to resource
poor economies
• Resource rich economy needs to overcome
coordination problem with respect to
development of higher human capital in order to
switch to next product mix
• Warning: the story is not about the lower volume
of human capital but about the deficit of marginal
skilled human capital in resource rich economies
Hypotheses:
• the higher is demand of industrial sector for high
skilled labor the slowly it grows relative to
industries less dependent on high skilled labor in
resource rich economies compared to resource
poor economies
• we could not expect that there is the
differentiated effect of resource abundance on
industries based on their demand for average or
lower skilled labor.
Data: industrial sectors’ demand for
human capital
• Abowd et al. ( 2003) estimate the human capital index
for each of 68 millions of U.S. workers (which covers
45% of U.S. labor force) that were surveyed in 1992
within Longitudinal Employer - Household Dynamics
(LEHD Program’s individual, employer, and employment
history databases).
• Then each individual human capital index was placed
into the industry where the firm she employed in belongs
to.
• This allows constructing the comparable distribution of
the level of human capital within U.S. industries.
Distribution of human capital within
U.S. industries
Metallurgy
12
10.6
10.6
11.2
11.4
11.1
10.3
9.2
10
8.4
%
8
9.2
7.9
6
4
2
0
1
2
3
4
5
6
7
8
9
10
Machinery (excl. electrical)
16
13.5
14
12
%
10
9.5
7.9
8.2
8.8
2
3
4
9.5
10.4
10.8
11.3
10.1
6
7
8
9
8
6
4
2
0
1
5
deciles of human capital
10
Measures of various levels of
human capital demand
Manufacturing sector
Petroleum and coal
products
Machinery, except
electrical
Iron and steel+
Nonferrous metals
Transport equipment
Paper and products
Printing and publishing
Wood products, except
furniture
Electric machinery
Textiles
Food products +
Beverages
Other manufacturing
products
Share of employees with human capital whose level is in
deciles from …-to….
2-10 3-10 4-10 5-10 6-10 7-10 8-10 9-10
10
91.7
84.7
77
68.1
58.1
47.4
36.4
25.6
15
90.5
82.6
74.4
65.6
56.1
46
35.6
24.8
13.5
89.3
88.2
88
87.7
78.7
77.4
77
78.8
67.5
66.7
66.3
69.8
56.1
56
55.8
60.3
45
45.4
45.5
50.5
34.7
35.2
35.9
40.7
25.5
25.7
27.1
31.1
17.1
17
19
21.6
9.2
8.9
11
12.1
87.3
86.1
85.3
76.8
74
74.4
66.5
63.2
65.1
56.4
53.3
56.5
46.6
44.1
48.2
37.2
35.6
40
28.2
27.7
31.7
19.5
19.9
22.9
10.6
11.4
12.9
83.7
71.9
61.5
51.6
42
33
24.8
17.3
9.8
81.7
68.4
57.2
47
37.8
29.6
22.2
15.4
8.9
Data: other industrial
characteristics
• Average annual real growth rates of
manufacturing sector in 1980-1990
– Nominal value added data from UNIDO
(United Nation Industrial Development
Organization) database for 3-digit ISIC codes
(Rev.2)
– GDP deflator obtained from WDI (World
Development Indicators) database.
• Share of sector in total manufacturing
value added in 1980 from UNIDO
database.
Data: country level
• raw hydrocarbon production of the economy as a
share of country’s GDP
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Japan
Singapore
Korea
Spain
Turkey
Austria
France
Israel
Portugal
Sweden
Greece
Finland
Jordan
Morocco
Philippines
South Africa
Belgium
Costa Rica
Kenya
Zimbabwe
Chile
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
Sri Lanka
Denmark
Germany
New Zealand
Italy
Bangladesh
India
Brazil
Pakistan
Australia
United Kingdom
Netherlands
Colombia
Peru
Canada
Malaysia
Mexico
Norway
Egypt
Nigeria
Venezuela
0.000
0.001
0.004
0.007
0.008
0.013
0.015
0.017
0.037
0.046
0.050
0.068
0.079
0.123
0.125
0.141
0.161
0.165
0.361
0.415
0.469
Example
• Norway: Machinery grew at a 4 percent
lower annual real rate than Metallurgy
• Belgium: Machinery grew at 2 percent
higher rate than Metallurgy
Estimated equation
Growth i,k=Const +Σkβk*Dummy for countryk+
+ Σiδi*Dummy for industryi+
+γ*Share of industryi in IVA1980+
+λh*(Demand of industryi for HCh*
*Resource abund.of countryk)
+εi,k
Estimation results
Dependent variable: Industry's Real Annual Growth, 1980-1990
Initial share
-0.935
-0.934
-0.933
-0.934
-0.936
(2.93)***
(2.92)***
(2.92)***
(2.92)***
(2.92)***
-0.937
(2.92)***
-0.939
(2.92)***
-0.940
(2.92)***
-0.938
(2.91)***
Interaction term: (Share of hydrocarbon production to GDP, 1980) * (Sum of upper deciles from …)
2 to 10
-0.822
Higher level of HC
(0.82)
3 to 10
-0.869
(1.38)
4 to 10
-0.888
(1.66)*
5 to 10
-0.949
(1.86)*
6 to 10
-1.080
(2.07)**
7 to 10
-1.305
(2.26)**
8 to 10
-1.667
(2.42)**
9 to 10
Higher level of HC
10th decile
Real effect
Observations
R-squared
-2.335
(2.47)**
-0.002 -0.005 -0.006
417
0.43
417
0.44
417
0.44
-3.936
(2.22)**
-0.006
-0.008
-0.01
-0.01
-0.01
-0.01
417
0.44
417
0.44
417
0.44
417
0.44
417
0.44
417
0.44
Robust t-statistics in parentheses
* significant at 10%; ** significant at 5%; *** significant at 1%
Results
• If we rank industries based on the demand for
very sophisticated labor (top deciles of human
capital distribution) then we observe the
significant systematical losses in growth rates of
industries with higher demand relative to those
with lower demand in countries rich in natural
resources compared to resource poor countries.
• The estimated losses become smaller and
insignificant as we use ranking of industries
based on the demand for less sophisticated
labor.
Interpretation of results
• Natural resource abundance which is an
exogenous characteristic of the country serves
as an impediment for manufacturing sectors that
depend on sophisticated human capital.
• There is no systematic effect of natural resource
abundance on the growth of industrial sectors
when we differentiate industries based on their
average human capital level demands.
• This implies that one of the links between natural
resource abundance and industrial growth might
be through an underdevelopment of very skilled
human capital.
Policy application
• One of the possible charms against the resource
curse: investment in education
• Leamer at al: “If the model is somehow backed
up with hard evidence, the policy advice is very
clear: Governments in countries that are in a
stage of “old product mix” but close to the stage
of “new product mix” should be making major
improvements in their educational systems, in
particular eliminating the dumbbell educational
systems that were economically efficient in old
product mix but inappropriate in new one.”
Relevant to Russia?
• Managerial survey:
– Tough deficit of middle level managers
– Most of top managers are foreigners
• Business development evaluations: lack of
skilled labor always mentioned as an important
impediment for business
• At the same time:
– 90’s: most of “Institutes” changed the titles to
“Universities”.
– General negligence to professional schools –
technical schools.
– May be it is time to revert some “Universities” to
“Professional schools”