Content uploaded by Ricardo Molero-Simarro
Author content
All content in this area was uploaded by Ricardo Molero-Simarro on Sep 18, 2016
Content may be subject to copyright.
Functional distribution of income, aggregate demand, and economic growth
in the Chinese economy, 1978-2007
1
Ricardo Molero-Simarro
2
Department of Economics, Loyola University Andalusia, Calle Energía Solar, 1. Edificios G,
2ª planta. 41014, Seville, Spain; and Department of Applied Economics I, Complutense
University of Madrid, Madrid, Spain.
(Received 15 September 2013; accepted 21 January 2015)
This paper seeks to analyse the relationship between the functional distribution of income,
aggregate demand and growth in the Chinese reform economy. For this purpose, the Bhaduri-
Marglin Model is used to indicate the theoretical possibility of both profit-led and wage-led
growth regimes. Previously, the principal literature on the evolution of factor shares in China was
reviewed. The statistical series for the period 1978-2007 are reconstructed to carry out our
analysis of the relations between capital share and investment, on one hand, and labour share and
consumption, on the other. Supported by empirical analysis and the model estimations, it is argued
that Chinese growth has been profit-driven. Finally, the implications are presented concerning
Chinese economic prospects.
Key Words: China; Economic Reform; Factor Shares; Aggregate Demand; Profit-Led Growth.
JEL Classifications: 011; E25; C13.
1. Introduction:
The spectacular expansion of the Chinese economy since the beginning of the reform process
has been accompanied by a significant increase in distributional inequality. The economic
reforms implemented since 1978 have allowed the Chinese economy to reach an average
growth rate of 9.9% between that year and 2007
1
. Investment and exports together have driven
growth (Zhu and Kotz 2010). On the one hand, exports have increased as a consequence of
declining labour costs. On the other hand, capital accumulation has been financed through
1
The Version of Record of this manuscript has been published and is available in the International Review of
Applied Economics, 29 (4), pp. 435-454, 25 February 2015:
http://www.tandfonline.com/doi/full/10.1080/02692171.2015.1016404.
Article DOI: 10.1080/02692171.2015.1016404.
2
E-mails: rmolero@uloyola.es.
internal savings, primarily from business profits, rather than households (He and Cao 2007).
Thus, as we shall see, the growth of the Chinese economy relies on the evolution of wage and
profits shares in GDP.
This paper links the evolution of labour and capital shares with China’s growth rates through
aggregate demand components by applying Bhaduri and Marglin’s theoretical framework
(1990) to the Chinese reform economy (1978-2007). The model has been estimated during the
last few years for several economies and groups of economies (Stockhammer and Onaran 2012).
Wang (2009) estimated the model for China during the 1993-2007 period, and Molero-Simarro
(2011) estimated the model for the period 1978-2007, as did Onaran and Galanis (2012), and
Jetin, Kurt and Su (2012) for the period 1980-2007. Wang’s results indicate a growth pattern
driven by profits. Molero-Simarro corroborates that pattern, taking into consideration only
domestic demand. Onaran and Galanis also find that China is profit-led, but this conclusion is
derived from exports rather than private investment. Meanwhile, Jetin, Kurt and Su focus on
the role of investment. Our goal is to confirm if the Chinese growth path between the years
1978 and 2007, as well as both internal and external demand, can be properly characterised as
“profit-led”.
The paper is organised in five main sections. After the introduction, there is a presentation of
the above-mentioned theoretical background to clarify the relationship between factor shares,
aggregate demand and growth. Second, there is a review of the current literature on the
evolution of the functional distribution of income in the Chinese economy and its effect on
growth. Third, this paper attempts to recalculate the available statistical data to obtain
comparable and consistent series of the variables to be analysed. Fourth, the topic of the
evolution of factorial distribution in China is revisited, explaining the factors that determined
its evolution and focussing, later, on its own role in pushing the different aggregate demand
components, which consequently determined the pace of economic growth. Fifth, in support of
the argument, there is an estimation of different equations that mathematically specify the
Bhaduri-Marglin Model. The conclusion refers to the implications of our results vis-à-vis the
Chinese government’s efforts to rebalance growth towards private consumption in the context
of the current global crisis.
2. Theoretical background:
In their seminal paper, Amit Bhaduri and Stephen Marglin (1990) consider the dual role of
wages as the most important element of production cost and the most important determinant of
aggregate demand. Bhaduri and Marglin state that, at a given labour productivity, there is a
distributional conflict between both profit margin and share, as well as the real wage. Real wage
has a negative effect on investment, via lower profit margin/share, and a positive effect on
consumption.
To determine the stronger of those two effects, they also define an investment function as
positively dependent on profit share and capacity utilisation through the effect the latter has on
the need for new production capacity, the effect the former has on future profits expectations,
and the role it also has as a source of finance for investment. Thus, the final effect of a change
in profit share on capacity utilisation and, therefore, on output depends on whether investment
is more or less responsive than savings to the positive effect that capital share on GDP has on
both. Then, the main question is to determine (empirically) if the positive effect of higher profit
share on investment is larger in relation to the negative effect it has on consumption.
On one hand, if the response of investment to a change in profit share is stronger than that of
savings, investment is the dominant component of aggregate demand, so growth is profit-driven.
Given the assumed higher propensity to save out of profits than out of wages, in this
“exhilarationist regime” the positive effect of an increase capital share on investment is also
higher than the negative effect of real wage restraint on private consumption. In this
exhilarationist regime, a lower real-wage growth rate generates higher capacity utilisation.
If the response of investment is weaker, however, consumption leads demand; thus, growth is
wage-driven. In this “stagnationist regime”, a lower profit share and a higher wage share allow
aggregate demand and capacity utilisation to increase as long as the positive effect on total
consumption is greater than the negative effect on capital formation. Then, in a wage-led pattern
of growth, a higher real wage generates higher employment.
If the economy is open, the results of the model in a context of wage-led growth may be altered.
The greater the share of exports and imports in the national income and the larger the elasticities
of both exports and imports vis-à-vis international price competitiveness, the less relevant are
the relationships described by the stagnationist regime. The negative effect of a lower real wage
on aggregate demand can be outweighed by the positive effect of lowered costs on external
sales. This can cause a shift from a wage-led to a profit-led wage regime.
To determine which effect prevails, it is necessary to estimate the effect a change in wage or
profit share has on aggregate demand components, including consumption, investment and net
exports, considering also additional control variables. In this work, the form of the consumption
function is:
C = f(W, R)
where C = final consumption; c = constant; W = compensation of employees; and R = gross
operating surplus. The expected signs of the parameters for wage incomes and profit incomes
are both positive. However, it is expected that the former will be higher than the latter due to
different propensities to consume out of wages and profits.
The form of the investment function is:
I = f(Y, R)
where I = gross fixed capital formation; C = constant; Y = gross domestic product (a proxy of
capacity utilisation, as is discussed below); and R = gross operating surplus. Both parameters
are expected to be positive.
Finally, the form of the net exports function is:
NX = f(Y, Y
f
, W, E)
where NX = share of net exports on GDP; C = constant; Y = gross domestic product;
Y
f
= main foreign partners weighted GDP; W = labour share; and E = nominal exchange rate.
The parameters for partners’ GDP, labour share (as a proxy of labour costs) and nominal
exchange rate are expected to be positive, but that for the domestic product is expected to be
negative due to its positive effect on imports.
As Stockhammer and Onaran (2012, 6) state, the total effect of, for example, a change in the
profit share is given by:
∆Y/∆∏ = (∂C/∂∏ + ∂I/∂∏ + ∂NX/∂∏) / [1 – (∂C/∂Y + ∂I/∂Y + ∂NX/∂Y + ∂G/∂Y)] (1)
where ∏ = profit share.
Although the value of the denominator, the standard multiplier, will determine the magnitude
of the effect, its sign is expected to be positive. Thus, the direction of the effect will be
determined by the sign of the numerator, “which is the sum of the partial effects of a change in
the profit share on the components of aggregate demand” (Stockhammer and Onaran 2012, 7).
Given that the partial effects of an increasing profit share (falling wage share) on both
investment and net exports is expected to be positive while its effect on consumption is expected
to be negative, the overall effect will depend on the relative magnitude of those partial effects.
As long as the absolute effect depends on the relative share of aggregate demand components
on GDP and consumption tends to be the main of those components, it could be expected that
domestic demand is wage-led, i.e., the sum of the partial effects of increasing profits on both
consumption and investment would be negative. Only when the positive effect of falling labour
costs on net exports is taken into consideration could it be expected that the sum of the three
partial effects is positive, i.e., the economy becomes profit-driven. In any event, only empirical
analysis could determine to what extent these expectations can be corroborated. Indeed,
particular cases regarding, for example, strong positive effects of increasing profit share on
investment in comparison with weak negative effects on consumption or high shares of
investment on GDP can make domestic demand become profit-led.
As we shall see, in China, economic liberalisation made possible a profit-led pattern of growth,
not only because of external competitiveness gains but also owing to high investment rates.
Increasing exports generated employment, but wages grew below productivity, creating a large
surplus to be reinvested. This investment fuelled growth. Before explaining the mechanism that
allowed all this to happen, in the next section, a review is presented in the next section of
previous explanations on the evolution of China’s factor shares are recalculated in Section 4.
3. Literature review:
Research on the functional distribution of income in China is very recent. Li, Liu, and Wang
(2009) characterise the general evolution of labour share in the economic development process
by a U-shaped curve, as with the Kuznets Curve. While they find that the wage share is lower
in the Chinese economy than in developed economies, they argue that China’s functional
distribution evolution between 1990 and 2006 is consistent with that pattern. Moreover, they
maintain that labour share will enter an upward trajectory after the traditional transfer of
workers among agricultural and industrial sectors is complete.
In accordance with this first explanation, Luo and Zhang (2009) claim that labour income share
is mainly correlated with the economy’s sectoral structural change. They notice three main
trends in national income distribution between 1993 and 2004: an upward one until 1995 due
to an increase of the importance of primary and tertiary sectors in GDP; second, a downward
one between 1996 and 2003 caused by the predominance of industrialisation and the decline of
wage share inside the three sectors; third, an abnormal drop in 2004, attributable to a change in
statistical accounting method, implying that the incomes of the self-employed came to be
counted in the profit share.
Wang (2009) does not discuss the reasons behind the decrease in labour share but was the first
to estimate the Bhaduri-Marglin’s Model for China. This author estimates the model for the
period 1993-2007, both for coastal area provinces and the whole country, relying on China’s
province functional distribution panel data. Wang’s main finding is the profit-led nature of the
Chinese economy, which “is caused by the larger impact of investment” on growth “relative to
that of consumption” (Wang 2009: 133), although the coastal area province internal demand
would be wage-led. Nevertheless, the pattern of growth “could not be sustained if external
demand contracts”, so “growth would be at risk, unless it becomes more dependent on private
consumption” (Wang 2009, 134).
Bai and Qian (2010) extend the period of study from 1978 to 2007. They maintain that the two
most important factors explaining evolution of functional distribution are structural
transformation and variations in the labour share within each sector of the economy. They argue,
however, that the drop of wage share is overstated due to the accounting methodological change
carried out by NBSCh in 2004. They adjust that year’s data, so the 5.25% point decline in labour
share between 2003 and 2004 becomes a 1.04% point recovery. Nevertheless, the inability to
apply that adjustment to subsequent years prevents the presentation of a completely consistent
series for the whole period.
Zhou, Xiao and Yao (2010) recalculate a consistent series for the three decade period by using
two alternative methods and calculating an average for the data. Then, they investigate the
evolution of raw labour, human capital, and physical capital. They divide labour compensation
between returns to human capital, which depend on years of schooling and work experience,
and raw labour, “the return from the physical expenditure of the labour” (Zhou, Xiao, and Yao
2010, 9). Meanwhile, they obtain physical capital by calculating the economy’s stock of capital
by the perpetual inventory method. These authors find that the decline in labour share from
1998 to 2006 would have been caused by the negative growth of raw labour, the deceleration
in the growth of human capital, and the high growth of physical capital. They argue that the
main cause of the fall in raw labour would have been the stagnant income of rural surplus
workers migrating to urban areas, who would have been paid below their contributions.
Onaran and Galanis (2012) analyse the Chinese economy’s functional distribution time series
as part of their research on the effects of the wage share widespread decline on global growth
rates. They find that China is “very strongly profit-led”, although “not due to investment” but
to “very strong export and import effects” (Onaran and Galanis 2012, 33). Onaran and Galanis
do not include SOE and collective enterprise investment in their estimates because, according
to them, “it makes no sense to treat these units as part of the same behavioural function as
private investment” (Onaran and Galanis 2012, 17). Thus, when these authors analyse the effect
of pro-profit redistribution on private investment, they find that it is insignificant; thus,
domestic demand is wage-led.
Finally, Jetin, Kurt and Su (2012) extend the period of study until 2010, although taking 1980
as the starting point. They correct the statistical series to avoid inconsistencies. They find, on
the one hand, that the evolution of the compensation of employees has a limited effect on
consumption. On the other hand, they argue that increasing profit share has pushed investment,
generating a profit-led domestic demand. According to them, this positive effect on growth
would have been reinforced by the increasing exports made possible by falling labour costs.
The estimation results of Wang, Onaran and Galanis, and Jetin, Kurt and Su will be discussed
in Section 6. Meanwhile, an original account of the evolution of the functional distribution of
income will be presented in Section 5. Beforehand, the next section will seek to obtain new
series of functional distribution of income including net taxes on production, which are
excluded from Bai and Qian’s calculations and those of Zhou, Xiao and Yao, and which use
official China Statistical Yearbook data on self-employed workers instead of the data used by
these authors.
4. Statistical series comparison and recalculation:
Sources for Chinese functional income distribution are limited. Prior to 1992, the only existing
data are those of Hsueh and Li (1999), who collected “Gross Domestic Product by Primary
Distribution of Income” for the 30 provinces of China (excluding Chongqing, whose
municipality was not created until 1997) during the 1978-1995 period. From 1992 to 2007, the
National Bureau of Statistics of China provides three different factor share series in its
Statistical Yearbook: the “Flow of Funds Accounts (Physical Transactions)”, the “Intermediate
Use Part of Input-Output Table”, and the “Income Approach Components of Gross Regional
Product” (NBSCh, various years). The first runs through the whole subperiod; the second
presents only 1992, 1995, 1997, 2000, 2002 and 2005 data; finally, the cited regional statistics
must be completed with the data supplied by NBSCh (2007) to obtain a third series that includes
data through to the year 1993.
There are marked differences between the values given by each series, even among Input-
Output and Regional ones (at least, until 2002), although, contrary to those of the Flow of Funds,
both take the depreciation of fixed assets into account. Although there are no apparent criteria
to choose among, the latter series has been retained for two reasons: first, because it is the only
way to obtain a homogeneous series for the entire 30-year period under study because the
regional product by income approach is calculated following the same criteria as Hsueh and Li
for their 1978-1992 data; second, it allows this paper’s outcomes to be compared with those of
the earlier studies of Bai and Qian (2010) and Zhou, Xiao and Yao (2010), who have also
chosen this series.
However, there are two obstacles to that comparison. On the one hand, these authors exclude
net taxes on production from the factor distribution series. It is impossible, however, to assign
an a priori pro-labour or pro-capital effect to taxes. Thus, their exclusion could distort
understanding of labour share progression insofar as it would neglect significant changes
experienced by net taxes, for which the share of GDP went from 11.6% in 1983 to 15.8 in 2003.
Accordingly, to avoid that neglect, in this work, they are retained within the functional
distribution of income series.
On the other hand, all those authors implicitly exclude depreciation of fixed assets. Including it
in the operating surplus, as we do in this paper, would actually make it comparable in
international terms. Moreover, after adding depreciation of fixed assets, the gross gains of
capital share in the whole period are significant (nearly six percentage points) and account for
more than two-thirds of the more than four-point decrease in labour share that occurred during
the whole 1978-2007 period.
Finally, all of this forces us to recalculate the 2004-to-2007 data of the series with taxes
incorporated and depreciation of fixed assets summed up in operating surplus to avoid the
effects of the NBSCh 2004 accounting change. The method developed by Zhou, Xiao and Yao
is used. First, total compensation of both employees and self-employed workers has been
calculated by obtaining the average labour compensation of employees and multiplying it for
the total workforce:
CE
a
= (CE/E) · (E+SE)
where CE
a
= total adjusted compensation of employees; CE = total original compensation of
employees; E = number of employees; and SE = number of self-employed workers.
Alternatively, average operating surplus for employees has been obtained, as well as the
equivalent sum calculated for total self-employed workers, to sum it with the NBSCh original
data for compensation of employees:
CE
b
= CE + (SE · (OS/E))
where CE
b
= total adjusted compensation of employees; CE = total original compensation of
employees; SE = number of self-employed workers; OS = total original operating surplus; and
E = number of employees.
Next, wage share has been weighed by the two methods, and the average value chosen:
LS
av
= ((CE
a
/GDP) + (CE
b
/GDP))/2) · 100
where LS
av
= average adjusted wage share; CE
a
= total adjusted compensation of employees
calculated by the first method; CE
b
= total adjusted compensation of employees calculated by
the second method; and GDP = gross domestic product.
However, to employ the same source for all the series, the China Statistical Yearbook data on
self-employed workers have been utilised instead of the China Economic Census Yearbook
data that Zhou, Xiao and Yao employ. Unregistered workers calculated by Zhou, Xiao and Yao
have not been taken into account; providing their incomes are very likely not part of the NBSCh
calculation of the compensation of employees. Figure 1 compares this paper series with those
calculated on the basis of the NBSCh original data.
Figure 1. Recalculation of labour share series, 2004-2007. Note: percentage of GDP
Source: Own calculations based on data from Hsueh and Li (1999) ("Gross Domestic Product by
Primary Distribution of Income (by the distribution approach)") for the 1978-1992 period; NBSCh
(2007) (“Components of GDP by Income Approach by Region”) for the 1993-2004 period; and
NBSCh (various years) ("Income Approach Components of Gross Regional Product") for the 2005-
2007 period. Since 2004 our series adjusted following the method presented by Zhou, Xiao, and Yao
(2010), but employing China’s Statistical Yearbook data for self-employed individuals.
As we shall see in the next section, which will be corroborated in Section 6, the pattern evolution
of functional distribution of income matches, to a great extent, that followed by aggregate
demand components and growth
.
At least since the second period of the reform, Chinese
economic growth has been dependent on both exports and investment, thus following a profit-
led path. The mechanism that allowed profit share to increase will be explained next.
5. Functional distribution, aggregate demand and growth during China’s economic
reform, 1978-2007:
During the early years of Chinese economic reform, productivity improvements in rural areas
were accompanied by increases in the purchasing prices of agricultural products, improvements
in farmer living standards and reductions in rural-urban inequality. Nevertheless, gradually, the
terms of trade between agricultural and industrial products turned again to favour the latter,
increasing the urban-rural gap. The Chinese government intervened in grain markets during the
35,0
40,0
45,0
50,0
55,0
60,0
1978
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
GDP by Region
Table
GDP by Region
Table Adjusted
whole reform period, first, by establishing agricultural product purchasing prices and
subsequently by setting minimum prices for buying in the market exchange.
Agricultural prices were cut at the end of the 1980s to contain inflation (Sicular 1988). In the
1990s, state agencies did not have sufficient funding to cover costs, so they were unable to offer
peasants substantially higher minimum prices than the market prices (Colby, Tiao, and Tuan
2001), thereby causing financial losses to them when good harvests pushed market prices down.
Moreover, since the entry of China into the World Trade Organization in 2001, China has used
imports to increase farm product supply (Yang and Li 2008), further pushing agricultural prices
down. These price policies worsened rural household average income, substantially increasing
the urban-rural income gap, which reached 3.33 points in 2007
2
.
At the same time, peasants suffered a loss of access to basic services due to the disappearance
of communes. In addition, industrial development moved increasingly to the coastal cities. Thus,
a combination of push and pull factors generated a growing amount of migrants flowing from
rural to urban areas following the evolution of agricultural product prices. Between 1991 and
1996, prices increased and then significantly decreased until the year 2005. Indeed, although
they gained momentum during the last three years of the period under study, they recovered
their 1996 level only in 2007. The push effect of low prices made rural-urban migration increase
to above 5% annually,
3
making rural migrant workers the labour force that enabled the
expansion of urban industry (Wen 2008).
The working and living conditions of the so-called nongmingong, however, have been poorer
than those enjoyed by urban residents. The Maoist era had established a system for controlling
internal population movements. The hukou, or household registration system, made it
practically impossible for a rural dweller to obtain an urban residence permit. The last three
decades transformed the economy operating logic, but the hukou has still been in effect.
“Undocumented” rural migrants to the cities are deprived of public services and legal protection.
The tide of undocumented migrants in the coastal cities has caused a difference in informality
rates, the number of working hours and salaries paid between migrants and the rest of urban
workers (Cai, Du, and Wang n.d.; ILO 2008). However, at the same time, it has also pressed
down the general wage level, holding the annual real rate of growth of industrial wages well
below the increase in worker productivity. Meanwhile, productivity in the secondary and
tertiary sectors increased by an average of 14.3% between 1991 and 2007, and wages grew by
an average of just 5.5%.
4
Thus, contrary to previous analyses, the mechanism that ultimately explains the evolution of
functional distribution is linked to the evolution of peasant income. As a consequence of
government intervention in the agricultural product market, the migration from the countryside
to the coastal cities has increased the floating population to at least 137 million in 2007,
according to the data of Chan (n.d.). This limited the evolution of real wages, labour costs and
wage share. Although employment in secondary and tertiary sectors grew by 6.9% annually in
the whole 1978-2007 period,
5
the labour share reached a level of just 42.5%. In this way, since
the early 1990s, the relationship between the countryside and the city has become the key to
passing world market competitive pressure on wages (Figure 2).
According to Gong and Yang (2010, 69), “the unlimited supply of labour under the dual
economic structure is the main reason for the steady decline in the share of wage income in
national income. In this structure, the unlimited supply of labour not only prevents wages from
responding to supply and demand situation in the labour market but also makes them insensitive
to labour productivity and price changes. This suggests that when there is a rise in labour
productivity or economic growth and price rises, the rise in wages may not be sufficiently
marked. The benefits derived from economic growth or labour productivity increases have
been, to a very large extent, converted into profits and not wages. Thus, the share of wage
income in national income will fall as the economy develops”. The main difference with our
argument is that they do not take into consideration the role of agricultural prices controlled by
the government in explaining the supply of labour from the countryside.
The productivity-wages gap created by the control of agricultural prices allowed China’s
competitiveness gains. In this sense, the importance of that gap has been doubled. On the one
hand, relatively low wages in comparison with productivity made possible the increasing
attraction of foreign direct investment (FDI) flows into the Chinese economy. According to
UNCTAD data, China’s share of world FDI inflows grew from 0.1% in 1980 to 8.3% in 2003.
Chinese government FDI policies (Shirk 1994) ensured the technology transfer process from
transnational corporations. Because of this, “by 2006, China’s export basket was highly
sophisticated and one of the most diversified in the world” (Felipe et al. 2013, 813).
Figure 2. Urban-rural gap and labour share, 1978-2007. Note: Income gap and percentage of GDP.
Sources: Own calculations based on NBSCh (various years) and same as in Figure 1.
On the other hand, the productivity-wages gap made possible the marked reduction of nominal
unit labour costs. Technological transfer and export basket diversification and sophistication
further increase industrial productivity. The continuous devaluation of the renminbi until 2005
helped to reduce China’s exports prices as well. However, only falling labour costs, which were
cut by nearly half between 1990 and 2007
6
, can explain the large increase of China’s share in
world total merchandise exports, which according to UNCTAD data rose from 0.8% in 1978 to
8.7% in 2007 (Figure 3).
Decreasing low and medium income household earnings due to falling labour share and
precautionary savings due to loss of access to basic services have caused private consumption
to grow slowly. Nevertheless, this slow growth of consumption has been compensated by rising
exports, which have clearly gained relevance in aggregate demand (Figure 4).
Owing to falling labour costs and rising external sales, the profitability of industrial firms has
increased markedly, at least during the reform’s second period (Lu et al. 2008). Increasing
surpluses have been the main source of investment financing. On the one hand, FDI has had a
minor role in capital accumulation. According to the UNCTAD data, it accounted for just an
average of 6.6% of gross fixed capital formation in the period 2006-2007. On the other hand,
according to the NBSCh (various years) data, domestic loans accounted for just 15.3% of the
funds for investment in fixed assets in 2007, falling from 20.5% in 1995. Thus, it has been the
reinvestment of Chinese firm profits, which explains the high rates of investment maintained.
Self-raising funds for investment shares rose to 60.6% in 2007.
At the same, together with increasing exports, capital accumulation has driven GDP growth
during much of the reform period. Indeed, capital-deepening since 1993 also explains labour
productivity increases (Lo and Zhu 2010). Thus, growth has followed the path of business
profits (Figure 5), depending on both exports and investment (Zhu and Kotz 2010).In the end,
the growth pattern seems to have been profit-led.
Figure 3. Labour costs and competitive gains, 1978-2007. Note: Index number and percentage of total.
Sources: Own calculations based on NBSCh (various years) and UNCTAD.
Figure 4. Labour share, household consumption and exports, 1978-2007. Note: Percentage of GDP.
Source: Same as in Figure 1 and NBSCh (various years).
Figure 5. Profits, investment and growth, 1978-2007. Note: Percentage of GDP and annual variation.
Source: Same as in Figure 1 and NBSCh (various years).
On the whole, this growth path allowed China to become the world’s second largest export and
the third largest economy in 2007. Nonetheless, profit-led growth has created conditions that
fuel inequality. The rising share of profits in national income has resulted in a higher top 10%
income share in household distribution of disposable income, which increased from 16.5% in
1985 to 25.5% in 2007 (Molero Simarro 2012).
In this paper, however, we focus on the analysis of the relationship between functional
distribution and growth through the evolution of aggregate demand components. As we have
observed, the growth of the Chinese economy has been dependent on exports and investment,
at least since 1992. As long as the evolution of both of them in China relies on the evolution of
labour costs and profitability, the growth pattern seems to have been profit-led. In the next
section, we will try to corroborate all that by estimating the model’s relations to the Chinese
reform economy.
In the next section, we will try to corroborate that assertion by estimating the model’s relations
to the Chinese reform economy.
6. Estimation results:
In support of the stated pattern, the results are presented for the Bhaduri-Marglin Model
estimation for China during the 1978-2007 period using regional panel data. The model has
been applied to several national economies and economic areas taken as a whole (Stockhammer
and Onaran 2012). According to Stockhammer and Onaran (2012, 8), “most studies conclude
that domestic demand is wage-led […]. Thus, demand is profit-led only when the effect of
distribution on net exports is sufficiently high to offset the effects on domestic demand, and this
is likely only in small open economies”. This contradicts to some extent the results found by
Wang (2009), Molero Simarro (2011) and Jetin, Kurt and Su (2012), although it fits with those
of Onaran and Galanis (2012), who, as was already noted in Section 2, found that China’s
economy is profit-led but only due to strong export effects.
This paper estimates single equations for consumption, investment and net export functions.
7
.
Panel data estimation techniques have been used instead of time series, as is usual in the
literature that estimates the Bhaduri-Marglin Model. Baltagi (2005, 4-7) lists several benefits
from using panel data; among others, it allows one to control for individual heterogeneity; it
gives less collinearity among the variables, more degrees of freedom and more efficiency. In
our case, this is especially important due to the relatively low number of observations of our
sample, which would most likely not be sufficient to carry out a time series analysis.
Given the nature of our data, the presence of fixed effects in the panel to be estimated is
expected. However, regression f-values have been compared with critical values to determine
if fixed effects estimates are preferable to pooled ones. At the same time, Hausman tests have
been carried out to determine if they are also preferable to random effects estimates. On the
other hand, explanatory variables have been included in both contemporary and one-year-lag
values to capture the whole short-run effects on the explained variables.
As with Wang, we mainly rely on the specifications of Stockhammer, Onaran and Ederer (2009).
Regional compensation of employees and operating surplus are used as independent variables,
assuming that they are exogenous. Gross, instead of net, operating surplus is taken, not
summing up net taxes to factor shares, as explained in Section 4. Statistical sources for factor
shares are those specified above. GDP, as a proxy of capacity utilisation, foreign partner income
and exchange rates are also used as explanatory variables. Interest rates have also been taken
into consideration, but as we noted below, they have been found not to be significant.
Regional household consumption, gross fixed capital formation and net exports are the
dependent variables. Public investment is included in total investment. Although it would be
preferable to estimate separately the behaviour of private investment, as Onaran and Galanis
(2012) did, separate public and private investment series are not available at the regional level,
so panel data estimation is not possible. Moreover, given the public nature of rural land’s
property rights and the state’s role in the economy, China cannot be considered a fully capitalist
economy. Indeed, although its share in total investment in fixed assets decreased, state-owned
and collective-owned enterprise investment still accounted for 31.8% of investment in 2007.
8
Thus, it is difficult to wholly understand China’s pattern of growth without taking into account
public investment as well. In accordance with this relevance of public investment, in the
Chinese case, operating surplus does not have to be considered a proxy for business expectation
of future returns. On the contrary, its inclusion as an explanatory variable is justified by its
central role as the main financial source for investment.
We analyse a cross-section panel data for China’s 31 provinces (including its four
municipalities and its five autonomous regions but not its two special administrative regions,
as long as the data are not available for them). Unlike Wang, however, the estimates are
calculated for the period 1978-2007, the whole 30-year period that includes the beginning of
the Chinese economic reform to the world financial turmoil of 2008 onwards. Due to Chinese
administrative reforms, there are no available data for the Hainan and Tibet provinces for
several years in the 1978-1992 period, in addition to the Chongqing municipality for that whole
period and the period 1993-1996. There are also no Jiangxi province data for 1978 and 1979.
Nevertheless, the data panel is strongly balanced.
All regional-level data are deflated to constant prices (1978=100) by GDP deflator own
calculation. The logarithm of the variables is calculated to avoid heteroscedasticity problems
derived from their exponential growth tendency. The only exception is net exports, for which
the ratio on GDP is used instead of its logarithmic to prevent negative data loss. The ECM
specification for the logarithms of the variables does not give statistically significant
cointegration results in either of the equations (consumption, investment and net exports). In
addition, the results of the Levin-Lin-Chu test suggest that all variables, except exchange rates,
contain a unit root, i.e., variables are integrated of order one. Thus, specifications in first
differences are used to estimate stationary series.
First, estimation results are presented (Table 1) for a consumption function defining the
relations between compensation of employees and operating surplus as independent variables
and final consumption as a dependent variable. The F-value for the significance of the fixed
effects indicates that pooled data estimates are preferable, as originally expected, given the
nature of our data. In addition, the Hausman test confirms that fixed-effects estimates are also
preferable to random effects estimates. Breusch-Pagan and Wald tests indicate
heteroscedasticity. To address that problem, the model is estimated by obtaining White robust
standard errors. Wooldridge tests give evidence of autocorrelation in the panel data in both of
the specifications. Thus, an AR(1) autoregressive term is also added to them.
Table 1. Consumption function estimation
Equation 1a Equation 1b
DV dlci dlci
Coefficient t-statistic Coefficient t-statistic
cons 0.000 0.17 0.000 0.19
dlwi 0.732 16.35 0.730 36.49
dlri 0.160 4.79 0.162 9.24
lagdlwi 0.014 0.94 0.015 0.78
lagdlri -0.027 -2.18 -0.027 -1.59
AR(1) -0.098
R
2
0.960 0.960
Note: DV= Dependent variable; ci = final consumption; c = constant; wi = compensation of employees;
ri = gross operating surplus. “l” denotes the Neperian logarithm of the constant renminbi value of the
variable. “d” the calculus of their differences. “lag” is the one year lag of the variables.
In both of the specifications, the coefficients for wages and profits are significant for their first
differences but not for their lags. The values of the coefficients for wages are higher than those
for profits. These values should be interpreted as elasticities. Their differences confirm,
however, that the propensity to consume out of profits is lower than that out of wages, as
indicated by the Bhaduri-Marglin Model.
Second, an investment function has been estimated, relating the behaviour of gross fixed capital
formation to that of gross domestic product and operating surplus (Table 2). GDP is taken as
the standard proxy of capacity utilisation. This is needed because there are no official data on
capacity utilisation later than 2002 in China’s statistical sources. Several specifications have
been tested, including interest rates in the equations. However, the specifications finally chosen
do not take into account those rates because of the lack of significance (as Wang also found),
even at 10%, and the weak robustness of the equations. Again, fixed effects estimates are
preferable to both pooled data and random effects estimates. Heteroscedasticity problems are
also addressed by estimating the equations with robust standard errors, and the AR(1) term is
also added to avoid autocorrelation.
The outcomes of the regression are also close to those anticipated: the sign of the contemporary
variables indicates a contemporary relationship of investment with both capacity utilisation
(represented by GDP as proxy) and gross operating surplus. The lags of the variables are not
significant. However, as a contemporary variable, operating surplus is strongly significant,
contrary to the calculations of Onaran and Galanis, which do not include the investment of
public enterprises. The values of the coefficients for GDP and profits are closer than other
studies indicate for several countries, confirming the importance of profits financing investment
in the Chinese economy.
Third, as a net exports function, the share of net exports on GDP is estimated in relation with
regional GDP, the weighted GDP of China’s five most important trade partners, labour share
as a proxy of labour costs, and nominal exchange rate. The dependent variable is in
differentiated form, as the logarithms of the independent ones. In this case, random effects
estimates are preferable to fixed effects estimates. Robust standard errors are obtained to correct
heteroscedasticity, and again, an AR(1) term has been added to avoid autocorrelation.
Table 2. Investment function estimation
Equation 2a Equation 2b
DV dlii dlii
Coefficient t-statistic Coefficient t-statistic
cons -0.003 -0.27 -0.006 -0.42
dlyi 0.629 3.15 0.592 7.99
dlri 0.533 3.00 0.565 8.19
lagdlyi 0.106 0.89 0.089 1.23
lagdlri -0.076 -0.71 -0.053 -0.80
AR(1) -0.244
R
2
0.814 0.814
Note: DV= Dependent variable; ii = gross fixed capital formation; c = constant; yi = gross domestic
product; ri = gross operating surplus. “ln” denotes the Neperian logarithm of the constant renminbi value
of the variable. “d” is the difference of the respective logarithms. “lag” is the one year lag of the variables.
Table 3. Net exports function estimation
Equation 3a Equation 3b
DV dnxy dnxy
Coefficient t-statistic Coefficient t-statistic
cons -0.027 -3.84 -0.028 -5.46
dlyi 0.374 7.26 0.379 12.45
dlyf 0.029 0.38 0.041 0.65
dlwy -0.168 -3.19 -0.166 -5.52
dlei -0.018 -0.20 -0.031 -0.43
lagdlyi -0.152 -4.31 -0.148 -4.90
lagdlyf 0.035 0.75 0.023 0.42
lagdlwy -0.088 -2.15 -0.086 -2.89
lagdlei -0.048 -0.93 -0.036 -0.61
AR(1) 0.051
R
2
0.627 0.627
Note: DV= Dependent variable; nxy = net exports on GDP; c = constant; yi = regional GDP; yf = main
foreign partners weighted GDP; wy = labour share; ei = renminbi/dollar nominal exchange rate. “ln”
denotes the Neperian logarithm of the constant renminbi value of the variable. “d” is the difference of
the respective logarithms. “lag” is the one year lag of the variables.
Surprisingly, the coefficients for foreign income are not significant, and neither are those for
the exchange rate. In addition, contrary to expectations, the sign of the regional income
coefficient is positive for the contemporary variables, although it is negative when presented in
lag form. Nevertheless, labour cost coefficients are significant in both contemporary and lagged
values. Moreover, their negative signs are expected, although the values of the coefficients are
lower than those found in previous studies, as we shall see.
Overall, high R
2
values (0.960 for the consumption function; 0.814 for the investment function;
and 0.627 for the net exports function) indicate a good performance for the model. Wald tests
confirm that parameters are jointly significant in all the equations. Thus, results seem to support
the relationship between the different components of functional distribution of income and
aggregate demand derived from the Bhaduri-Marglin Model.
To determine the distribution aggregate component that has driven Chinese economic growth,
it is still necessary to calculate the effect of a change in primary distribution on GDP growth. It
is usual in the literature to calculate partial effects and total effects at the mean and at the end
of the sample to assert the distinct impact of distributive changes once the distribution has been
altered. Here, we calculate them at the mean of the variables for the whole estimation period,
at the 2007 values and also at the values at the beginning of the sample in 1979; thus, an attempt
is made to distinguish between the different periods of Chinese reform. With that aim, we use
expression (1) discussed in Section 3. Nonetheless, it is important to bear in mind that only the
coefficients obtained from the specifications that are corrected by adding an AR(1) term are
used. The results are listed in Table 4.
In the calculation made at the 1979 level, the overall effect of an increase of profit share on
GDP is negative. In the other two levels, the effect is positive, i.e., growth would be driven by
profits. However, if only the effects on internal demand are considered, the negative effect on
an increase of profit share on consumption would be higher than its positive effect on
investment both at the 1979 and mean levels. Only at the 2007 level could internal demand be
labelled as profit-led. In any event, as stated above, if the positive effect of decreasing labour
costs on net exports is taken into account, China’s growth model is nearly unambiguously
profit-led.
Table 4. Distributional effects on GDP calculations.
∂C/∂∏ ∂I/∂∏ Internal
Effects
∂NX/∂∏ Total Effects
Equation
1b
Equation
2b
Equations
1b and 2b
Equation
3
Equations
1b, 2b and 3b
Mean level
-0.285
0.217
-0.068
0.079
0.011
1979 level
-0.324
0.211
-0.113
0.088
-0.026
2007 level
-0.238
0.245
0.007
0.070
0.077
Compared with previous studies, the negative partial effects of increasing profit share on
consumption (which range from -0.238 to -0.324) are lower than those at the mean level of the
sample calculated by Onaran and Galanis, who found a 0.412 negative effect. However, our
calculations are higher than those calculated by Wang at the mean and 2007 levels (-0.18 and -
0.20, respectively), and they are also substantially larger than the positive effects of wage share
on consumption calculated by Jetin, Kurt and Su for the 1980-2007 (0.02), 1980-1998 (0.01)
and 1999-2007 (0.03) periods.
Onaran and Galanis do not find significant effects of profit share on investment due to their
lack of consideration of state and collective enterprise investment. After including this
investment, we found a positive effect, but in comparison with the Wang calculations (0.43 and
0.48) and especially with respect to the Jetin, Kurt and Su calculations of the negative effect of
wage share on investment (-0.81, -0.70 and -0.88), our calculations (which range from 0.211 to
0.245) are substantially lower. Thus, they are our calculations of the external demand effect
(which range from 0.070 to 0.088) compared with those of both Wang (0.25 and 0.28) and Jetin,
Kurt and Su (who obtained -0.90, -0.58 and -1.15 effects of wage share on net exports) as well
as those of Onaran and Galanis (1.986). This is very likely due to the different periods and
specifications analysed.
In summary, the Chinese economy can be characterised as driven by profits if external demand
effects are taken into consideration. Relatively high consumption effects and relatively low
investment effects lead us to consider that China’s domestic demand has been likely wage-led,
only becoming profit-led when investment reached its peak values in GDP.
7. Conclusions:
Unlike many accounts of the Chinese reform process, which explain functional distribution of
income solely on the basis of the Kuznets Curve or account for the evolution of economic
growth as unconnected with distribution, this study develops an alternative approach in an effort
to highlight the existing link between agricultural prices, rural-urban migration and labour share,
on one side; and functional distribution, aggregate demand and growth on the other. The meagre
improvement of rural incomes pushed rural-urban migration, restraining the growth in
industrial wages well below productivity. Lower labour costs allowed China’s external sector
to gain competitiveness, increasing the economy’s share in world exports. Meanwhile, the
surplus created thereby financed China’s internal investment, fostering capital accumulation
and further improving productivity and pushing growth.
The results generally support the hypothesis concerning the positive relationship between profit
share and growth for the whole 1978-2007 reform period. This was already stated by Wang
(2009) but only for the 1993-2007 years. The main difference with the results obtained by
Onaran and Galanis (2012) lies in this paper’s inclusion of public investment in gross fixed
capital formation figures. However, this only made it possible to assert a profit-led internal
demand, as Wang also found for the 1993-2007 years, in one out of three of our calculations of
the effects of increasing profit share. In any event, the inclusion of SOE investment, however,
should be understood in terms of the role profits have financing investment in China, not as
confirming a private business behavioural function. Indeed, it would clarify the lack of
significance of interest rates in explaining investment behaviour.
Although this paper focused on the relations between factor shares, aggregate demand and
growth, it can be claimed that in China, the evolution of the Gini index would have been linked
to that of the primary distribution of income. The increase in profit share would have generated
a consequent in-depth enrichment of top income households, so the relationship between
growth and inequality, mediated by functional distribution, would be positive. Nevertheless, as
stated in the Bhaduri-Marglin Model, widespread class conflict would have been avoided
thanks to increasing employment and real wages growth, which allowed absolute wage bill
gains.
Overall, it is possible to posit some implications of this research for Chinese economic
prospects. The decrease in exports to external markets forced the Chinese government to launch
a 4 trillion renminbi stimulus package (equivalent to nearly US$600 billion) in November 2008
to maintain growth rates. In addition, the Chinese government had already started to develop a
wide range of distributive and redistributive policies: agricultural tax abolition in 2006; new
labour law in 2008; health care reform in 2009; an appeal to increase provincial minimum
wages to 40% of the average during the XII Five-Year Plan period, 2011-2015; a 5% increase
in the SOE percentage of profits to be paid to the state beginning in 2015 to finance social
security; and so on. In this way, the government tries, at the same time, to prevent social
instability, by constructing a “harmonious society”.
For the time being, those measures, along with a steady increase in the agricultural product
producer prices every year since 2007, with the exception of 2009, have allowed rural incomes
to increase faster than urban incomes. The urban-rural gap dropped from 3.31 to 3.13 points,
and wages rose by 12.3% annually.
9
Nevertheless, it seems to have not been sufficient to alter
China’s distributive and growth patterns. Rural-urban migration increased further to at least 145
million in 2009, whereas the labour share still fell, and the capital share increased between 2009
and 2011.
10
Consequently, consumption decreased by an extra half point as a proportion of
GDP (from 49.6% in 2007 to 49.1% in 2011), while investment augmented it even further (from
41.6% in 2007 to 48.3% in 2011),
11
so net exports reduced its share of GDP because of the
world crisis.
Therefore, according to this paper’s analysis, if the Chinese growth model is to be reoriented,
it would be necessary to deepen the measures already taken by pushing agricultural prices
upward and by using additional policy tools to increase the labour share. Although we found a
lower sensitivity of exports to labour costs in comparison with the Onaran and Galanis
calculations, it would also be necessary for a change ‘in the composition of exports via a shift
towards products with a lower price elasticity of demand’ (Onaran and Galanis 2012, 44) to
ensure that the external sector still contributes positively to growth.
Acknowledgements:
The author acknowledges the generous help of Nacho Álvarez Peralta and Ángel Vilariño; the
useful comments on a previous version of this paper made by Javier Cuenca Esteban, Yolanda
Fernández Lommen, Dic Lo, Jo Michell, Özlem Onaran, Enrique Palazuelos Manso and
Engelbert Stockhammer; and the insightful suggestions of all the participants in both the
Universidad Complutense de Madrid’s ‘Economía Internacional y Desarrollo’ postgraduate
programme’s seminar, and in the V Workshop on Economic Modelling organized by the
Departament of Economics of Loyola University Andalusia, especially Melania Salazar-
Ordóñez who participated as a discussant. I would also like to thank the referees for providing
useful comments which served to substantially improve the paper. Needless to say, all the
remaining mistakes are my own.
Funding information:
This work was supported by Complutense University of Madrid’s.
Notes:
1. Own calculations based on NBSCh (various years) data.
2. Own calculations based on NBSCh (various years) data.
3. Own calculations based on Chan (n.d.) data.
4. Own calculations based on NBSCh (various years) data.
5. Own calculations based on NBSCh (various years) data.
6. Own calculations based on NBSCh (various years) data. Nominal unit labour costs have been
calculated as the ratio of nominal average wage (wage bill divided by the number of workers) in the
secondary and tertiary sectors and the real labour productivity (GDP at constant prices divided by the
number of workers) in the same sectors.
7. According to Onaran and Galanis (2012, 7), “[t]he single equation approach fails to utilize the
fact that consumption, investment and net exports (and state expenditures) add up to GDP. To address
this aspect as long as the endogeneity of the wage share, a systems approach, like the VAR approach
used by […] Onaran and Stockhammer (2005), may be a solution. However, this comes with its own
problems, because results are more difficult to interpret. It is not possible to detect the precise economic
relationships that lead to changes in demand in response to distribution when using the systems
approach. Nevertheless, it is important to note that the convenience of interpretation of the results of the
single equation approach come at the price of some bias because the system-dimension is ignored”.
8. Own calculations based on NBSCh (various years) data.
9. Own calculations based on NBSCh (various years) data.
10. NBSCh does not provide the income approach components of gross regional product data for
2008. Moreover, a jump in compensation of employees and operating surplus data for 2009-2011 can
be observed, so the calculation criteria seems to have been altered again. Labour share grew from 39.7%
in 2007 to 46.6% in 2009, falling to 44.9% in 2011. Meanwhile, capital share decreased from 46.1% in
2007 to 37.9% in 2009, rising to 39.9% in 2011. If this paper’s recalculation methodology is applied,
labour share increases from 42.5% in 2007 to 48.6% in 2009, decreasing to 47.2% in 2011, and capital
share decreases from 43.3% in 2007 to 37.9% in 2009, increasing to 39.9% in 2011. However, even
these calculations are tentative because NBSCh also altered total employment figures for 2006 and 2007
in its 2011 Statistical Yearbook. Unfortunately, as long as the Bureau does not update the 2004 and 2005
data, it is not possible to use these new figures to recalculate factor shares during this research period.
11. All these data are from NBSCh: Statistical Yearbook 2012 (“Gross Domestic Product by
Expenditure Approach).
References:
Bai, Chong-En and Zhenjie Qian. 2010. “The Factor Income Distribution in China, 1978-2007.”
China Economic Review 21: 650-670.
Baltagi, Badi H. 2005. Econometric Analysis of Panel Data. West Sussex: John Wiley and Sons.
Third Edition.
Bhaduri, Amit, and Stephen A. Marglin. 1990. “Unemployment and the real wage: the
economic basis for contesting political ideologies.” Cambridge Journal of Economics 14, 375-
93.
Cai, Fang, Yang Du, and Meiyan Wang. n.d. “Employment and Inequality Outcomes in China.”
Mimeo. Institute of Population and Labour Economics, Chinese Academy of Social Sciences.
Chan, Kam Wing. n.d. “China, Internal Migration”. In Ness, Immanuel y Peter Bellwod (eds.).
n.d.: The Encyclopedia of Global Migration. Blackwell.
Colby, Hunter, Xinshen Diao, and Francis Tuan. 2001. “China’s WTO Accession: Conflicts
with Domestic Agricultural Policies and Institutions”. Trade and Macroeconomic Divisions
Working Paper 68. Washington: International Food Policy Research Institute.
Felipe, Jesús, Utsav Kumar, Norio Usui, and Arnelyn Abdon. 2013: “Why Has China
Succeeded and Why It Will Continue to Do So?” Cambridge Journal of Economics 37: 791-
818
Gong, Gang and Guang Yang. 2010. “Analysis of China's Unequal Income Distribution from
the Perspective of Functional Income Distribution.” Social Sciences in China 31 (4): 53-72.
He, Xinhua, and Yongfu Cao. 2007. “Understanding High Saving Rate in China.” China &
World Economy 15 (1): 1-13.
ILO (International Labour Organization). 2008. “Rural-Urban Migrations and Policy
Responses in China: Challenges and Options.” ILO Asian Regional Programme on Governance
and Labour Migration Working Paper 15. Geneva: ILO.
Jetin, Bruno, Ozan Ekin Kurt, and Anna Su. 2012. “Unbalanced distribution of income and
Unbalanced Chinese Growth: Single Equation Estimations Based on Bhaduri/Marglin Model.”
Paper presented at the Political Economy and the Outlook for Capitalism Conference, AHE,
IIPPE and AFEP, Paris, June 5-8 2012.
Li, David, Linlin Liu, and Hongling Wang. 2009. “Changes in the Labor Share of GDP: A U-
Shaped Curve.” Social Sciences in China 30 (4): 131-153.
Lo, Dic, and Zhang Zhu. 2010. “Making Sense of China’s Economic Transformation.” Review
of Radical Political Economics 43 (1): 33-55.
Lu, Feng, Guoqing Song, Jie Tang, Hongyan Zhao, and Liu Liu. 2008. “Profitability of China’s
industrial firms (1978–2006).” China Economic Journal 1 (1): 1-31.
Luo, Changyuan, and Jun Zhang. 2009. “Labor Income Shares (LIS) in Economic Development:
An Empirical Study Based on China’s Sectoral-Level Data.” Social Sciences in China 30 (4):
154-178.
Molero Simarro, Ricardo. 2011. “Functional Distribution of Income and Economic Growth in
the Chinese Economy, 1978-2007.” School of Oriental and African Studies Department of
Economics Working Papers 168. London: University of London.
Molero Simarro, Ricardo. 2012. “Primary Distribution, Top Incomes and Inequality in China,
1978-2007.” PERI Working Paper Series 301. Amherst: University of Massachusetts Amherst.
Onaran, Özlem and Giorgos Galanis. 2012. “Is Aggregate Demand Wage-Led or Profit-Led?
National and Global Effects.” ILO Conditions of Work and Employment Working Papers Series,
40. Geneva: International Labour Organization.
Onaran, Özlem, Engelbert Stockhammer. 2005. “Two Different Export-Oriented Growth
Strategies: Accumulation and Distribution a la Turca and a la South Korea.” Emerging Markets
Finance and Trade 41 (1): 65-89.
Shirk, Susan. 1994. How China Opened its Door: The Political Success of the PRC´s Foreign
Trade and Investment Reforms. Washington: Brooking Institution.
Sicular, Terry. 1988. “Agricultural Planning and Pricing in the Post-Mao Period.” The China
Quarterly 116: 671-705.
Stockhammer, Engelbert, and Özlem Onaran. 2012. “Wage-Led Growth: Theory, Evidence,
Policy.” PERI Working Paper Series 300. Massachusetts: University of Massachusetts Amherst.
Stockhammer, Engelbert, Özlem Onaran, and Stefan Ederer. 2009. “Functional Income
Distribution and Aggregate Demand in the Euro Area.” Cambridge Journal of Economics 33:
139-159.
Wang, Peng. 2009. “Three Essays on Monetary Policy and Economic Growth in China.” PhD
diss., University of Ottawa.
Wen, Tiejun. 2008. How China’s Migrant Labour is Becoming the New Proletariat. Mimeo:
Renmin University of China.
Yang, Dennis Tao, and Yuanfang Li. 2008. “Agricultural Price Reforms in China: Experience
from the Past Three Decades.” Agroalimentaria 14 (27): 13-23.
Zhou, Minghai, Wen Xiao, and Xianguo Yao. 2010. “Unbalanced Economic Growth and
Uneven National Income Distribution: Evidence from China.” Institute for Research on Labor
and Employment Working Paper 2010-11. Los Angeles: University of California.
Zhu, Andong and David M. Kotz. 2010. “The Dependence of China's Economic Growth on
Exports and Investment.” Review of Radical Political Economics, 43 (9): 9-32.
Statistical Sources:
Hsueh, Tien-tung & Li, Qian. 1999: China's National Income, 1952-1995. Boulder: Westview
Press.
NBSCh (National Bureau of Statistics of China). 2007.: Data of Gross Domestic Product of
China, 1952-2004. Beijing: China Statistics Press.
NBSCh (National Bureau of Statistics of China). various years: China Statistical Yearbook.
Beijing: China Statistics Press.
