Transcript Document

Total Factor Productivity Elasticity of R&D Investment
Empirical Research on Korean Manufacturing Industries
Objectives
This work focuses primarily on the analysis of Korean
industries’ R&D effect to TFP(Total Factor Productivity),
an important indicator of innovation and economic
growth. We estimate TFP elasticity of R&D stock by
different sectors of Korea manufacturing industries.
Also, two kinds of R&D stock obsolescence rates were
used in this study. First approach is using the previous
research result of Suh(2005), which estimates different
obsolescence rate of the 7 industries(0.1160-0.4949)
using 695 Korean firms’ data. Second approach is using
0.125 suggested by many literatures.
Results
• 4 tech sector according to OECD classification
: high-tech is most high, but low-tech is not most low
Methodology
• R&D Stock (perpetual inventory method by Jeorgenson)
Sector
RD(stock)t = RD(flow)t + (1-γ) RD(stock)t-1
RD(stock)t0 = RD(flow)t0 {(1+g)/(g+ γ)}
Elasticity 1
Elasticity 2
(different Obsolescence rate) (Obsolescence rate=0.125)
Observation
High-tech
0.210682
0.218392
104
Mid-high tech
0.128771
0.089635
389
Mid-low tech
0.079498
0.078182
274
Low tech
0.115224
0.105929
448
* γ:obsolescence rate of R&D stock, g: average growth rate of R&D stock
• 23 industry classification according to KSIC
• TFP(Value Added Account Growth using Cobb-Douglas function)
Yi,t = A Ki,tα Li,tβ
lnAi,t Y = lnTFPi,t = lnYi,t – α lnKi,t – β lnLi,t
* Y:Value Added, A:TFP, K:Capital stock, L:Labor input, α:Capital stock ratio to value
added, β:labor ratio to value added, i:industy, t:year
• Elasticity(Co-integration test & Vector Error Correction Model)
H0(null) : lnTFPi,t – β lnRNDi,t-1 = 0
∆lnTFPi,t = λ ∆lnTFPi,t-1 + α ∆lnRNDi,t-2 + η ∆lnTFPi,t-1
+ β ∆lnRNDVi,t-2 + εi,t
* RND:R&D Stock, β:long-term elasticity to TFP, i: industry, t:year, ε:error
Data base
• Unbalanced panel of Korean manufacturing industry
covering the 12 year period 1995-2006
• Data sources: the Survey of R&D activities in Science
and Technology of Korea, the Survey of Mining and
Manufacturing of Korea, and Economic Statistics System
of the Bank of Korea
• Main variables : Value Added, investment in R&D of
each industrial sector , Capital·Labor input and ratio to
value added
Main Variable
Mean
Std. Dev.
Observation
ln (R&D Stock)
10.60903
2.123930
2262 (1995-2006)
ln TFP
1.226280
0.320276
2484 (1992-2006)
Approach
Research is conducting on 23 groups according to 2digit category of KSIC(Korea Standard Industry
Classification) and four sectors according to OECD
Contact
technology sector and product classification
Manufacturing Industry
Tech.
sector
Beverage and Food products
Low
Tobacco products
Low
Textile, except Apparel and Fur
Elasticity 1
(different
Obsolescence rate)
Elasticity 1
(Obsolescence
rate=0.125)
Observation
0.081833
0.092267
123
Low
0.086495
0.099189
60
Apparel and Fur product
Low
0.231579
0.242051
36
Leather, Luggage and Footwear
Low
0.056473
0.163516
25
Wood, Wood and Cork product, ex. Fur.
Low
0.123846
0.12916
32
Pulp, Paper, Paper product
Low
Publishing, Printing, Recorded media
reproduction
Low
Coke, Refined Petroleum, Nuclear fuel
Mid-low
16
Chemical material and product
Mid-high
106
Rubber and Plastic product
Mid-low
0.094769
0.105874
54
Non-metallic minerals
Mid-low
0.087547
0.065396
76
Basic metal product
Mid-low
0.107858
0.102285
67
Fabricated metal product
Mid-low
0.204147
0.20711
68
Other machinery and equipment
Mid-high
0.1398
0.150781
154
Office, Accounting machinery
High
Other electrical equip. and Transformers
Mid-high
0.093368
0.102369
64
Electric video and audio equip., telecom.
High
0.225062
0.244724
32
Medical, Precision and Optical inst.,
Watches and clocks
High
0.183596
0.188115
56
Motor Vehicles, Trailers, Semitrailers
Mid-high
0.09066
0.095663
24
Other Transport equip.
Mid-high
0.111528
0.120058
41
Furniture and other Manufacturing
Low
0.139033
0.147753
88
Recovery material reproducing
Low
Total
24
0.148981
0.172519
44
16
16
0.09901
0.115184
1215
Policy implications
‘TFP elasticity of R&D difference between each
sector/industry can be utilized for innovation-policy
decision making in national R&D strategy. Especially,
these estimates as basic data showing the difference
between technology sector / industries could be used in
Benefit-Effectiveness Analysis(BEA) for feasibility study
of new large scale R&D program.
Next steps / The way forward
• Estimation of obsolescence rate of each industry
• Estimation of R&D contribution ratio in economic
growth
• Study to explain low-tech sector’s high elasticity
SeungKyu Yi, Eunjung Ma
Korea Institute of Science and Technology Evaluation
and Planning(KISTEP)
Seoul, Korea
Tel. 82 2 589 2983 / 2811
E-mail: [email protected], [email protected],