Microsoft Word EcRefChina Oct06. doc

party authorities (outside intervention) relative to the managers of firms. 
52
See, for instance, Choe and Yin (2000). Similar explanations, in terms of principal-agent relations, for 
the low productivity of SOEs (as well as their profitability before its recovery from about year 2000) 
have been used, for instance, by Mi and Wang (2000). 
It remains to be seen whether the new government agency designed to manage physical government 
assets, the state-owned Asset Supervision and Administration Committee, will create some arm’s-length 
distance between politicians and managers of state enterprises. 
53
Chow (2002, Chapt. 19) reports subjective impressions, by himself and others, of poor service in 
government-operated service institutions in China.

---

32
So far, I have only referred to 
indirect 
evidence – “smoking guns” – of 
inefficiencies in the Chinese growth path. In principle, these inefficiencies should 
also be reflected in estimates of total factor productivity (TFP) growth in the context 
of aggregate production functions – although calculations of this variable are 
probably even more hazardous in China than in developed countries.
54
Most such 
calculations indicate yearly rates of TFP growth in the interval 3-4 percent during 
the reform period. However, we may want to confine the calculation to 
contributions to output growth through better technology and organization in 
individual firms, sometimes called “multifactor productivity” (MFP) growth. In this 
case, it is reasonable to exclude both reallocation gains across sectors (basically 
shifts of labor from agriculture to other sectors) and improvement in human capital 
through education and training. Each of these two factors is often estimated to have 
contributed nearly one percentage point (per year) to the output growth rate in China 
during the reform period.
55
When these factors are excluded from calculations of 
TFP growth, we obtain figures (of MFP growth) approximately in the interval of 1.5 
to 2.5 percent per year during the reform period.
56
 
Compared with the pre-reform decades, for which MFP growth seems to have been 
close to zero (OECD, 2005a), these figures appear to be fairly good. However, the 
picture requires modification in important respects. First, part of MFP growth 
during the reform period reflects once-and-for-all efficiency improvements in 
connection with the agricultural reforms in the period 1979-1984, as well as 
temporary productivity spurts in other sectors in the mid-1990s when prices were 
deregulated. Second, it would seem that much of the improvement in technology 
and organization in the manufacture sector, so far, is the result of the entry and 
expansion of non-mainland firms rather than improvements in a broad spectrum of 
domestic firms (Tseng and Zebregs, 2002; OECD, 2002, pp. 195-230).
57
Some 
54
Indeed, Holz (2005b, section 4a) argues that there is no stable aggregate production function for China 
for the reform period. 
55 
For estimates of reallocation gains arising from the contraction of agriculture see, for instance, 
Borensztein and Ostry (1996); Woo (1998); and OECD (2005a, Table 1:4). Kuijs and Wang (2006b) 
conclude that these reallocation gains have gradually fallen in recent years; they argue that the 
contribution was 0.8 percentage points per year in the period 1993-2003.
56
For surveys of studies of TFP and/or MFP growth, see, for instance, Heytens and Zebregs (2003);
Blanchard and Giavazzi (2005); OECD (2005a); and Wu (2006). 
57
Whalley and Xin (2006) estimate that such firms contributed no less than 40 percent of China’s GDP 
growth in 2003 and 2004. 

---

33
studies also suggest that the rate of TFP and MFP growth has gone down during the 
last decade, although it is impossible to know if this new trend will continue and, if 
so, how strong it will be.
58
While we would expect falling returns on real 
investment, other factors are likely to have positive effects on GDP growth, such as 
the faster restructuring and privatization of the SOEs, the gradually higher R&D 
spending (starting from a low level), increased competition as a result of joining the 
WTO, and better economic integration of various geographical areas within China 
itself.
On the basis of existing evidence, it is reasonable to conclude that (except mainly 
for the sector of foreign firms) economic growth in China has largely been 
resource
extensive
in the sense of heavily relying on capital accumulation, inputs of raw 
materials and energy, and the depletion of environmental resources.
Naturally, both the public discussion and the scholarly literature have emphasized the 
huge regional differences in per capita income growth – with the eastern (coastal) 
provinces as the leaders, and the mountainous provinces in the west, along with the 
“rustbelt” areas in the north as the laggards; see, for instance, Démurger et al. (2002). 
Official statistics suggest that while GDP has grown by about 11.5 percent per year during 
the reform period in the most successful provinces (Fujian, Guangdong, Jiangsu and 
Zhejiang), the growth rate has been about half of that in the least successful province 
(Quinghai); see, for instance, Wei and Liu (2004).
However, such extreme comparisons 
give an exaggerated picture of the 
overall
dispersion of the distribution of growth rates 
across provinces. The standard deviation of the per capita growth rate across provinces 
during the period 1980-2002 seems to have been 1.5 percentage point, which, however, is 
enough to generate large differences in levels of per capita income across provinces.
59
The 
level 
of per capita GDP is currently reported to be about seven times higher in the 
most developed than in the least developed province, even if we exclude the very poorest 
province (Guizhou).
58
For instance, OECD (2005a, Table 1.4) reports a fall in multifactor productivity growth from 3.4 
percent per year during 1993-1998 to 1.3 percent per year during 1998-2003. Angang and Zheng (2004) 
report a fall from 3.7 percent during 1991-95 to 0.6 percent during 1996-2001.
59
According to Blanchard and Giavazzi (2005, p. 6), the standard deviation of the 
level
of income per capita 
across provinces increased by 72 percent between 1998 and 2003. 

---

34
Indeed, the regional differences among China’s provinces (and large cities) are so large 
that China looks like a continent with a mixture of emerging industrial countries and some 
of the poorest “countries” in the world.
It is also well known that per capita income differs 
drastically between urban and rural areas within provinces. For instance, official statistics 
suggest that the ratio of average income in urban areas is about three times (3.3) as high as 
in rural areas. However, this is a highly uncertain figure because it may not fully reflect 
differences in costs of living in rural and urban areas – a measurement defect that is usual 
in other countries as well. Although average income and consumption have increased 
rapidly in nearly all parts of the country during the reform period (Chow, 2006a; 
Ravallion and Chen, 2006), these geographical differences are an important background 
for widespread dissatisfaction with the 
social 
situation in the country – an issue to which I 
now turn. 

---

35