Much Ado about Nothing, Almost: Factor Income Distribution in China

1
Much Ado about Nothing, Almost Factor Income
Distribution in China

• Chong-En Bai
• Zhenjie Qian
• Tsinghua University

2
Trend since 1978
3
Observations

• Capital share increased from 37.34 in 1978 to
  45.23 in year 2006.
• Labor share declined from 49.8 in 1978 to 40.61
  in 2006.
• Share of taxes on production increased from 12.85
  in 1978 to 14.16 in 2006.
• There is an abrupt change between 2003-2004
  capital share increased from 39.93 to 44.35, and
  labor share declined from 46.16 to 41.55.

4
Trend since 1978 (net of production taxes)
5
Questions

• What is behind the recent changes in factor
  income distribution in China?
• What explains the abrupt changes between 2003 and
  2004?

6
Why Do We Care?

• Factor income distribution may affect size income
  distribution.
• What Is China Doing to Its Workers? (by Arvind
  Subramanian, Peterson Institute, in Business
  Standard, New Delhi, February 8, 2008)
• This might well be the mother of all
  redistributions.
• Is the dramatic decline in labor's share of the
  economic pie ominous?
• Will the decline in labor's share of the economic
  pie be reversed through political change? That
  may be China's big question.

7
Why Do We Care?

• Factor income distribution may affect size income
  distribution.
• A workers' manifesto for China (Economics focus,
  Economist, Oct 11th 2007) Many countries have
  seen a fall in the share of labour income in
  recent years, but nowhere has the drop been as
  huge as in China. This partly reflects China's
  large pool of surplus labour, which has depressed
  wages relative to the economy's large
  productivity gains.

8
Why Do We Care?

• Factor income distribution may affect size income
  distribution.
• Many economists in China have also talked about
  the rapid decline in labor share and proposed
  policies in response to the decline.
• The government has adopted or is considering
  adopting policies to deal with the issue.

9
Why Do We Care?

• Return to capital
• Investment rate has been increasing. (figure)
• Capital output ratio has also been increasing.
  (figure)
• How has the rate of return to capital changed?
  The answer depends on capital share.
• Kaldors stylized facts about the growth of
  advanced industrial economy
• Fact one Real output per capita grows at more or
  less constant rate over fairly long periods of
  time
• Fact two The stock of real capital, crudely
  measured, grows at a more or less constant rate
  exceeding the rate of growth of labor input
• Fact three The rates of growth of real output
  and the stock of capital goods tends to be about
  the same, so that the ratio of capital to output
  shows no systematic trend
• Fact four The rate of profit on capital has a
  horizontal trend
• Fact three and four imply that factor income
  share in output should be constant

10
Preview of Results
N1 Aggregate Labor share start to decline in 1995
11
Preview of Results

• The elasticity of substitution between capital
  and labor in industry sector is not significantly
  different than 1.
• The change in the relative price of capital and
  labor is not a significant factor behind the
  change in labor share in industry sector.

12
Outline

• What Explains the Abrupt Changes between 2003 and
  2004?
• Structural Transformation
• Change of Labor Share in Industry

13
What Explains the Abrupt Changes between 2003
and 2004?

• Before 2004, all the income of the self-employed
  was counted as labor income.
• Since 2004, income of the owners of the
  individual businesses, a major form of
  self-employment other than rural households, has
  been ascribed as capital income.
• In 2004 census, its stipulated that the
  operating surplus of state-owned and
  collective-owned farms should be counted as labor
  compensation and almost half of the provinces
  follow this method in year 2004, which explains
  why there is an abrupt increase in the labor
  share in primary industry.

14
What Explains the Abrupt Changes between 2003
and 2004?

• Adjustments made using the 2004 census data
• In table 1-26 of China Economic Census Yearbook
  (NBS, 2007), there are items such as operating
  revenue, operating expenditure including employee
  compensation and payable taxes, book value of
  fixed assets for individual businesses by
  industry, which can be employed to calculate
  value-added, operating surplus for individual
  businesses by industry according to the method
  given in DNA (2007).
• Depreciation book value of fixed assets 5
• Operating surplus operating revenue-operating
  expenditure-depreciation
• Net production taxes and labor compensation
  reported in NBS (2007)
• Value added Depreciation Operating surplus
  Net production taxes Employee compensation
• We subtract calculated operating surplus of
  individual businesses from the reported total
  operating surplus by industry and add it to
  employee compensation. With this adjustment, we
  recalculate labor share.

15
What Explains the Abrupt Changes between 2003
and 2004?

• We also adjust for the inconsistency between
  national and provincial data on individual
  businesses, by assuming that labor productivity
  of individual businesses at the provincial level
  is the same as that at the national level.
• As we do not have enough information to judge how
  much operating surplus of state-owned and
  collective-owned farms are counted as labor
  compensation in 2004, we could not make
  adjustment to eliminate the effect of the
  associated change. We recalculate aggregate labor
  share in 2004 using the actual labor share in
  primary industry in year 2003.

16
What Explains the Abrupt Changes between 2003
and 2004?

• Adjustments results

Notes N1 labor share calculated using reported
data N2 labor share adjusted using Census data
N3 labor share adjusted using Census data and
provincial employment numbers in individual
businesses N4 labor share in primary industry
in year 2004 is replaced by that in year 2003

• Conclusion
• The jump in reported labor share between 2003 and
  2004 is the result of the change in the way we
  tally the income of the self-employed.

17
Structural Transformation

• Observation
• Labor share of the primary sector is much higher
  than other industries.
• Labor share of the primary sector is
  overestimated since all the income of rural
  households engaged in primary industry production
  is counted as labor income

18
Structural Transformation

• Observation
• The structural transformation resulted in the
  decline of the primary sector and rise of the
  tertiary sector.
• As labor share in the primary industry is much
  higher than that of the tertiary industry, labor
  share declined with the structural
  transformation.
• As labor share in the primary industry is
  overestimated, the effect of structural
  transformation would not have been as large if we
  obtain true estimate of labor share in primary
  industry

19
Structural Transformation

• If theres no structural transformation between
  1995 and 2004, then the labor share in 2004 would
  have been 57, rather than 53.6
• Structural transformation can explain 31
  percentage points in the change of labor share.

20
Change in Industry

• Change in labor share in industry explains 15
  percentage points in the change in aggregate
  labor share.
• What has happened?

21
Change in Industry Theory

• Utility function
• Production function
• Firm objective SOEs are interested in the size
  (output and/or employment) of the firm as well as
  profits.
• labor share formula for labor share
• The elasticity of substitution between capital
  and labor is important.

22
Change in Industry Predictions

• Ownership effect labor share is higher when the
  firm has a stronger size preference.
• Monopoly power labor share is lower when the
  firm has stronger monopoly power
• The change in the relative price between capital
  and labor is reflected in capital-output ratio in
  efficient term.
• Elasticity of substitution between factors
  determines the relationship between labor share
  and capital-output ratio

23
Change in Industry Data

• Description of the data
• Annual survey of industrial firms conducted by
  the National Bureau of Statistics of China from
  1998 to 2005. Industrial survey covers all SOEs
  and non state-owned enterprises with annual sales
  over 5 million Yuan.
• Capital share ratio of operating profit and
  accounting depreciation to value added at factor
  cost
• Three proxies for monopoly power price markup
  HHI CR10
• Two groups of proxies for ownership equity
  shares (req_x) control rights (D_x)
• Capital-output ratio ratio of fixed assets at
  book value to value added at factor cost
• Capital augmenting technical parameter
  controlled by year dummies
• Other factors are controlled by industry dummies

24
Change in Industry Methodology

• Description of the methodology
• System GMM estimation estimates level and
  difference equations simultaneously, each using
  lags of difference term and level term of
  endogenous variables as GMM instruments. This
  deals with the problem of the endogeneity of
  capital-output ratio.
• System GMM estimation estimate level equation so
  that explicit fixed effect such as region dummies
  and industry dummies can be estimated.
• System GMM estimation uses both between group and
  within group information when estimating level
  and difference equation and hence obtain precise
  estimation for the difference of capital share
  between enterprises with different ownership
  structure.

25
Change in Industry Empirical Results
Notes N1, base model N2, log model legend
plt.1 plt.05 plt.01
26
Change in Industry Robust Checks
legend plt.1 plt.05 plt.01
27
Change in Industry Robust Checks
Notes N1, observations in sample no less than 3
years N2 observations in sample since 1998
legend plt.1 plt.05 plt.01
28
Change in Industry Decomposition
-81 of actual increase in capital share can be
predicted by the model -1/2 of the predicted
capital increase is contributed by ownership
restructure -1/4 of the predicted capital
increase is contributed by monopoly power
change -contribution of the restructure across
region and industries is trivial
29
International Comparison

• The potential problems with international
  comparison
• Data compatibility the tally of income of the
  self-employed differs across countries
• Gollin (2002) estimated labor share by country
  assuming income of self-employment as capital
  income, comparable to that adopted by the NBS

30
Rate of Return to Capital Revised
31
Preview of Results
N1 Aggregate Labor share start to decline in 1995
32
Conclusion

• The elasticity of substitution between capital
  and labor is not significantly different than 1.
• The change in the relative price of capital and
  labor is not a significant factor behind the
  change in labor share.

33
Investment rate since 1978
Back
34
Capital-output ratio since 1978
Back
35
Structural Transformation
Back