Title: Industrial and Aggregate Measurements of Productivity Growth in China, 19822000
1Industrial and Aggregate Measurements of
Productivity Growth in China, 1982-2000
- Jing Cao
- Tsinghua University
- Mun S. Ho, Dale W. Jorgensen
- Harvard University
- Ruoen Ren and Linlin Sun
- BeiHang University
- Ximing Yue
- Renmin University of China
2Main Research Question
- What are the sources of Chinas successful
economic growth? - Accumulation of capital?
- Total Factor Productivity Growth?
- Restructuring of the economy from agriculture to
manufacturing and services?
3Research Goals
- Estimate productivity growth by sectors
- Derive aggregate productivity growth estimates
using three different methods of aggregation - Aggregate Production Function
- Aggregate Production Possibility Frontier (PPF)
- Direct Domar-weighted Aggregation
- Use newly developed data based on time-series
Input-Output Tables and micro data from the
Survey of Labor Force - Construct detailed time series data at sector
level including appropriate deflators
4Time Periods
- Covers the period 1982-2000
- For purposes of analysis, four sub-periods are
considered - 1982-84 Characterised by growth in agriculture
and efficiency gains due to policies including
the introduction of household registration system - 1984-88 Growth due to industrial sector reforms
and open door policy launched - 1988-94 Social market economy and development
zones - 1994-2000 Declining role of SOEs and increasing
private ownership, tariffs were reduced to be
ready for WTO entry
5Past studies
- Chow (1993) not including services sector no
technical change during 1952-80 - Chow and Li (2002) 3.03 TFP growth during
1978-1998 - Borensztein and Ostry (1996) TFP growth -0.7
during 1953-78 3.8 during 1979-1994 - Fan, Zhang and Robinson (1999) 4.2 TFP growth
during 1978-1994 - Woo (1998) 1.1 TFP growth during 1979-1993
- Young (2003) 1.4 in non-agriculture
- Ren and Sun (2005) TFP growth at 2-digit level
Using Domar weights found 3.83 for the whole
sector during 1988-94
6KLEMS approach to TFP Growth measurement
- Economy divided into primary, industrial and
tertiary sectors, in which industrial sectors are
divided into 26 sectors - Capital input structures equipment and
automotive vehicles - Labor input number of workers classified by
sex, age and educational attainment - Intermediate inputs commodities produced
domestically plus imports put in the production
process
7Classification of Industries
8Capital and Labor Inputs
Labor Input
Capital Input
9Growth accounting at the industrial sector level
- The gross output of sector j in period t is
assumed to be produced with a Hicks-neutral
production function using various types of
capital, labor and intermediate commodities
- Productivity growth in each sector j from the
year t-1 to t is given by
- d ln Ajt provides a measure of TFP growth at the
industrial sectorl level.
10Value-added Functions
- We define a concept of industry value-added. The
real value added of sector j, , is defined by
writing output as a weighted share of value added
and intermediate input
- The following identity is applied
11Time-series I-O table
- An important technical innovation in the growth
accounting is to construct a internally
consistent time-series I-O table, labor input,
capital input data and price deflators. - Our data set is constructed based on 4 benchmark
official I-O tables (1981,1987, 1992, 1997)
including A tables and U tables, consistent to
the coverage and definitions of 1997 table - Then a time-series current value U tables are
constructed for 1981-2000, using Kuroda
minimization of errors approach. - Based on current value U tables and sectoral
price indices, a time series real value U tables
are constructed. - Finally, a time series KLEM data are extracted
based on the time-series real value U tables.
12Industry Output Growth ()1982-2000
13Industry Output Growth ()1994-2000
14Industrial TFP Growth ()1982-2000
15Industrial TFP Growth ()1994-2000
16Domar-Weighted Contribution ()1982-2000
17Domar-Weighted Contribution ()1994-2000
18Growth accounting at the aggregate level
- Three Alternative Aggregation Approaches
- Direct aggregation across industries A simple
Domar-weighted aggregate of industry-specific
productivity growth estimates - Aggregate Production Function Approach Aggregate
value added is considered as a function of
aggregate value of capital and aggregate labor
requires the assumption that there is perfect
substitution across sectors, same prices for
labor and capital in all sectors
19Growth accounting at the aggregate level
- Aggregate Production Possibility Frontier
Relaxes the assumption of the existence of an
aggregate production function and assumes value
added is not perfectly substitutable across
industries but retains assumptions on input side - We also discusses inter-relationships between
measures of productivity aggregate level using
the three different approaches. - TFP growth under production function approach is
equal to Domar-weighted sectoral growth plus
reallocation of value added across sectors plus
reallocation of labor and capital across sectors.
20Growth in Aggregate Value-added and its Sources
using production possibility frontier
21Aggregate Reallocation Effects ( per year)
22GDP Growth Decomposition
23Main Findings Sectoral productivity change
- TFP growth rates are quite varied across sectors
and across sub-periods - Preferred approach is to use the Production
Possibility Frontier approach allows TFP growth
estimates to be decomposed into Domar-weighted
sectoral growth and reallocation in labor and
capital - Our results shows that the aggregate TFP is 2.5
for 1982-2000, and a steady decline in TFP growth
over the four sub-periods - Effects due to reallocation of labor and capital
appear to be negligible when the whole study
period is considered.
24Some Caveats
- Data problems in the service sector,
- Some data issue in highly regulated sector, such
as electricity, oil and gas extraction, petroleum
and refining sectors, sectoral output might be
underestimated. - Deverticalization Problem There has been an
increase in sub-contracting and/or splitting up
of previously highly vertically integrated
enterprises, then gross value of output might
increase. - We do not have estimates of land input for some
sectors, but treat it as capital input, this
would lead to some bias as well.
25THANK YOU!