Transcript R&D and Development

R&D, the National Innovation
System, and Growth
William F. Maloney
Office of the Chief Economist, LAC
Joint work with M. Bosch and D. Lederman
See 2003 flagship: Closing the Gap in Education and Technology
5 Questions
1.
How strong is LAC’s innovation effort in
comparative terms?

2.
3.
4.
5.
Lederman and Saenz data set (2003)
Is this optimal given potential impact on
growth?
What determines levels of R&D investment?
How efficiently does LAC use this invesment?
Why are we not Finland?
R&D Superstars
R&D
GDP
 GDP 
 1
 2 
GDP
CAP
 CAP 
2
5.0%
Predicted & Observed R&D/GDP
4.5%
4.0%
Israel
Finland
3.5%
3.0%
2.5%
Korea
2.0%
1.5%
China
1.0%
India
0.5%
Argentina
Mexico
0.0%
4
5
6
7
8
Log GDP per Capita
9
10
11
-2
-1
0
1
2
Residuals from Benchmarking
Regression: Asia
1970
1980
1990
2000
Year
China
Korea
India
T aiwan
Indonesia
T hailand
2
Latin America
-2
-1
0
1
Do NR offer fewer
opportunities for
innovation?
1970
1980
1990
2000
Year
Argentina
Costa Rica
Brazil
Mexico
Chile
Venezuela
-2
-1
0
1
2
Advanced Countries: X- NR
1970
1980
1990
2000
Year
Australia
Ireland
Canada
Netherlands
Finland
Sweden
Finland: Rise in R&D Generalized
Percent Value Added
Finland: Rise in R&D/VA
23%
9
8
7
6
5
4
3
2
1
0
od
o
F
1976
1990
Pr
.
d
o
.
.
r
.
r
ls
ls
ls
ts
h
d.
h
e
e
q
a
a
a
c
c
o
c
r
E
r
th
a
u
ic
a
ap
et
e
x
P
d
t
O
n
M
P
m
i
M
l M por
ro
e
T e od
tc .
M
c
a
h
i
P
o
s
s
l
ic
t.
C
e
l
r
n
a
a
e
t
t
W
e
a
B
e
r
M
nec
l
T
M
n
o
E
o
N
N
s
it l e
Do Returns to R&D Justify
Efforts of Superstars?
US firm level/industry data- social returns


Grilliches and Lichtenberg 1984
Terleckyj 1980, Scherer 1982
71%
>100%
X country

-Coe and Helpman 1995 G7
123%
Lederman and Maloney: extend to LDCs



Five year period averages, 1960-2000
43-53 countries
GMM system estimator (Blundell and Bond 1998,
Arellano and Bover 1995)
Estimating Rate of Return to I,
R&D
I
R&D
 ln Y  rk ( )  rs (
)  (1   ) ln L
Y
Y
• Derived from production function
• Ratio of rs/rk = R&D*/R&D
•
US: 28%/7% = rise by 4 X. ( Jones and
Williams 1998)
• Interact with GDP/capita
Rate of return to R&D
Dependent Variable: Growth of GDP (Constant PPP), five year averages 1975-2000.
Methodology: GMM System Estimator
(1)
(2)
(3)
(4)
(5)
Initial level of gdp per capita
Investment/GDP
Labor growth
R&D/GDP
Tertiary Enrollment ratio
NR-Leamer
R&D*(gdp per capita)
R&D*(NR-Leamer)
Investment/GDP*(GFP per capita)
Countries
Observations
0.035
1.299
0.509
3.193
(6)
*** 0.001
-0.009 ** 0.000
0.001
0.088 ***
*** 0.189 *** 0.328 *** 0.237 *** 0.271 *** 0.883 ***
*** 0.600 *** 0.495 *** 0.754 *** 0.484 *** 0.771 ***
*** 1.382 *** 0.518 *** 1.022 *** 9.622 *** 9.290 ***
0.056 *** 0.028 * 0.053 ** 0.023 **
0.001 ** -0.006 *** -0.006 ***
-0.300 ***
-1.029 *** -0.992 ***
0.371 *** 0.328 ***
-0.131 ***
-0.079 ***
53
43
43
43
43
43
162
107
107
107
107
107
ROR R&D ~78%
Predicted Rates of Return
120%
5
4.5
LAC should invest 2.5 to
10 times more in R&D !!
100%
4
Rate of Return
3
60%
2.5
2
40%
NIC
JAM
BRA
CRI
MEX
ARG
CHL
KOR
USA
FIN
1.5
1
20%
0.5
0%
2000
0
7000
12000
17000
22000
Constant 1995 PPP US$
Returns to R&D
Returns to Investment
R&D/Inv
Optimal R&D/Investmtent
3.5
80%
Why do Rich Countries Invest
More in R&D?
Dependent Variable: R&D/GDP
Estimation Method: GMM System Estimator
(1)
(2)
R&D/GDP at t-1
0.90 *** 0.86
Log (GDPpc)
0.22 *** 0.21
GDP growth
-1.22
-0.15
Real Interest Rate
-0.45
-0.23
Private Credit/GDP
0.07
Sd Growth
Log (IP Index)
Gov.Cons./GDP
Quality of Res. Inst.
Collaboration
Leamer Index
Sargan Test (p-value) 0.36
0.45
0.32
0.35
2nd Order Serial
Observations
73
73
Countries
30
30
***
***
*
***
(3)
(4)
(5)
0.89 *** 0.87 *** 0.90
0.22 *** 0.21 *** 0.15
-0.03
0.00
1.12
-0.40 *** -0.07
-0.86
(6)
(7)
*** 0.80 *** 0.87 ***
*** 0.00
0.07 *
*** -0.53
-0
*** -0.04
-0.11
(8)
(9)
0.96 ***
0.18 ***
0.09
0.52 *
0.68 *** 0.69 ***
-0.02
0.02
2.68 **
1.11
-0.15
-0.49 *
0.17
(10)
-0.44 ***
0.16 ***
0.99 ***
1.22 ***
0.39 ***
1.11 **
1.12 **
1.15 *
0.98 ***
0.96
0.42
73
30
0.96
0.37
73
30
0.53 ***
0.52
0.35
73
30
0.47
0.33
73
30
0.58
0.33
73
30
0.42
0.40
73
30
0.26
0.35
73
30
-1.38 ***
0.50
0.47
73
30
LAC’s efficiency of converting
R & D into patents,TFP is also low
Patents = B1R&D + Bp country*R&D
20.00%
15.00%
10.00%
-5.00%
-10.00%
NOR
FIN
ISR
KOR
TWN
VEN
URY
PER
MEX
CRI
COL
CHL
BRA
0.00%
ARG
5.00%
National Innovation System
Global Knowledge Economy
Human Capital
Think Tanks/
Antenna
University
Innovation
Clusters
Firms
Innovation & TFP Growth
Other Public Policies:
Rules of the Game
Infrastructure (ICT)
Subsidies/Tax incentives
Coordination Initiatives
Global Knowledge Economy
Most R & D $ and scientists in universities
but collaboration with private sector is low
(interviews with entrepreneurs: score 1-7)
7.0
6.5
6.0
5.5
5.0
Finland: 40% have formal
arrangements with U;
Chile: 25% and not very
fruitful
Public think tanks poor and
captured
60% R&D funds dedicated to
basic science; US 15%
4.5
4.0
3.5
3.0
2.5
2.0
Arg
Bra
Chi
Col
CR
Mex
Chn
Quality of Scientific Inst.
Esp
Cor
Ind
Irl
Aus
Sw e
Isr
Univ/Private Sector Collaboration
Fin
EUA
Why are we not Finland?
Not addressing serious mkt failures in innovation


Few incentives to R & D - tax breaks, subsidies
Weak efforts to help firms learn

Incubators, research parks, consortia
Disfunctional NIS:

Explains more dynamic NR sectors in Scandinavia.



see From Natural Resources to the Knowledge Economy (2001)
LAC’s NIS-closer to the Holy Roman Empire than Finland.
No mkt forces assure elements of NIS work together
FIN