256
Capitalism with Chinese Characteristics
does or does not improve welfare or to what degree it improves welfare. One piece of evidence apparent in Table 5.1 is that since 2002, GDP growth and personal income growth again began to converge. I discuss the details later but again, this development took place under the leadership of Hu Jintao and Wen Jiabao, who, as is widely known, began to put more emphasis on improving the living standards of the Chinese people rather than on GDP growth per se.
That there can be a divergence between GDP growth and welfare gains is a cautionary tale to those economic observers who formulate their grandiose views of the Chinese economy on the basis of GDP data alone. It is for this reason that this book treats GDP performance as the beginning of the analysis rather than as the end of the analysis. In Chapter 1,1 cite research that contrives analytical devices to comport manifestly inefficient institutions and policies on the one hand with the excellent GDP performance on the other. Maybe a simpler and a more analytically productive approach ought to have been to probe more deeply into the complications and implications of GDP performance.
1.3 Equity
It is well established that China today is among the most unequal societies in the world. According to Khan and Riskin (2005), China had a Gini coefficient (which measures inequality of income distribution) of 45 in 2002, compared with 31.6 in Korea (1998), 32.5 in India (late 1990s), and 34.3 in Indonesia (2002). By 2006, China's Gini coefficient reached 49.6, according to a report by the Chinese Academy of Social Sciences based on a survey of 7,140 households (Dyer 2006). China has now surpassed or is in the process of surpassing the level in Latin America, the region widely known as having the worst income inequality in the world (and troubling economic performance). According to the data provided by Khan and Riskin (2005), the Gini coefficient was 46.5 in Costa Rica, 52.2 in Argentina, 57.1 in Chile, and 58.5 in Brazil.
A part of this increase in the dispersion of income distribution is not surprising. To be sure, the income distribution during the socialist period was compressed, but the price for this was an extraordinarily high level of poverty. One mechanism to achieve absolute egalitarianism under central planning was the complete suppression of incentives in resource allocation. Moving away from the absolute egalitarianism of the central planning era and toward a greater reliance on economic incentives is likely to lead to an expansion of income gaps among individuals endowed with different
Capitalism with Chinese Characteristics
257
dispositions and capabilities. From an efficiency perspective, income disparities rise for "good" reasons.
There can also be an entirely different reason for rising income disparities. In this case, certain groups or individuals are privileged - by the political process, financial system, and regulations - over others to grab a larger share of the economic gains. For example, the privileged groups or individuals can enter into businesses that are off-limits to others; under this scenario, the income disparities rise because the economic processes are anti-competitive. A more subtle form of anti-competitive economic processes is the restrictions not on the economic opportunities per se but on the social opportunities - the acquisition of psychological inclinations and educational and physical capabilities to participate in economic opportunities. In this case, economic opportunities may be distributed equally, but the social opportunities to participate in them are not. Income disparities can rise as a result and for "bad" reasons.
But here is the analytical challenge: The "good" and "bad" mechanisms for income disparities lead to an observationally identical result - a rising Gini. We do not know which set of factors is behind the rising income disparities or which set of factors is more important. Both for analysis and for drawing the right policy implications, it is critical to identify the true mechanisms - economic incentives or blockage of economic opportunities -behind China's rising income disparities during the reform era. Let me suggest a way to think about this issue.
A simple, although crude, way to distinguish between the economic incentive story and the anti-competitive story is to assume that 100 percent of the income disparity in the United States is the result of market incentives. Under this assumption, the Gini coefficient of the United States can serve as an upper threshold between the "good" and the "bad" mechanisms for income disparity. It is an upper threshold because in the United States, racial discrimination and the political power of big business can also be construed as obstructions to economic opportunities. For our purposes, it is better to err on the side of caution against drawing a false positive and, therefore, it makes sense to set a high threshold on the basis of the US level rather than on the basis of the more egalitarian East Asian level.
Estimates of Gini coefficients often vary by sources and by analysts. For the sake of simplicity, I adopt the estimates provided in Khan and Riskin (2005) who directly compared the Gini coefficient for China with that of other countries, including the United States. According to them, the recent Gini coefficient for the United States is 40.8. If this is the threshold, China crossed it sometime in the early 1990s. China's Gini was 38.2 in 1988 and
258
Capitalism with Chinese Characteristics
45.2 in 1995 (Khan and Riskin 1998). In 1980, China's Gini was 28 (Khan and Riskin 2001). Based on these numbers, we can argue that in the 1980s, the rise in the Gini was due to the workings of the economic incentives, whereas in the 1990s, it was due to the blockage of economic opportunities.
One supporting piece of evidence for this hypothesis comes from Khan and Riskin (2005). According to their analysis, both urban and rural entrepreneurship - defined as self-employment businesses - improved income distribution, although each in different ways. Rural entrepreneur-ship was found to increase within-rural income inequality but it decreased rural-urban income inequality. Their analysis shows that within-rural inequality contributed very little to China's overall income inequality, whereas rural-urban income differences had a huge effect. Thus, the net effect of rural entrepreneurship was an equalization of income. Urban entrepreneurship, on the other hand, improved within-urban income distribution because small-scale self-employers in the urban areas tend to be poor. An extension of this finding to the theme of this book leads to the conclusion that in the 1990s, blockage of business opportunities for small-scale entrepreneurs exacerbated income inequalities. This dynamic also explains the income distribution pattern in Shanghai. Shanghai is the extreme version of the policy model that restricted small entrepreneurs, with grave implications for income distribution.
There are other telltale signs that in the 1990s the rising Gini was a result of factors other than economic incentives. The economic incentive story is most compatible with rising income disparities at the individual level -such as educated individuals earning more than uneducated individuals -rather than at the group level, especially groups of individuals with incidental characteristics (e.g., place of birth or race). The following is one of the most unusual findings by scholars who have looked at income distribution in China in some detail: In the 1990s, there was a sharp decline in individual income inequality and there was a sharp rise in group income inequality, primarily in the rural vis-á-vis the urban groups. This is from Khan and Riskin (2005), who report that both the rural Gini and the urban Gini declined between 1995 and 2002, whereas rural-urban income disparities rose sharply. This is true whether or not migration is taken into account.
Another intriguing finding is that factors such as place of residence, in fact, increased in importance in explaining income distribution. This is reported by Sicular et al. (2007). They show that the variables at the individual or household levels, such as the size of household, education, and age, explain about 54 percent of urban-rural income differences in 1995 but only about 20 percent in 2002. That is, in 2002, 80 percent of the income
Capitalism with Chinese Characteristics
259
differences are a result of incidental factors such as the geographic residence of the households, compared with only 46 percent in 1995. This finding is rather remarkable considering the following. One is that their paper implicitly already incorporates the story of economic incentives by including education and other variables. So, another reading of their finding is that the importance of economic incentives declined between 1995 and 2002. The other factor is that between 1995 and 2002, China was supposed to have experienced an increase in the regional mobility of labor as rural migrants moved more freely between rural and urban areas. The rising importance of geographic factors in their econometric exercise is quite unexpected given this increase in labor mobility.25 Although the scholars who have generated these findings do not explicitly make this argument, I would argue that these findings are largely consistent with the story of blocked opportunities rather than with the story of rising economic incentives.
Our third clue comes from the rising social tensions in Chinese society. Although few China economists take the sharp increase in social unrest in China as a data point in their perspectives on China (in sharp contrast to other China social scientists), these incidences of social unrest help to put the rising Gini in perspective. In general, economists are less concerned about the relative deprivation - rising incomes of all groups in the society but at different rates - than about the absolute deprivation (some groups losing income relative to the levels of their past incomes). If the rising Gini is a result of relative deprivation, the implications are relatively benign. If, however, the rising Gini is a result of absolute deprivation, it is altogether a different story.
The rising levels and the degree of severity associated with social unrest in China suggest the looming possibility of absolute deprivation. In the post-Tiananmen political environment in China, social unrest - demonstrations, protests, riots, strikes, and so forth - are highly risky undertakings. The individuals involved face realistic and swift prospects of arrest and severe punishment. It is unlikely that millions of Chinese participated in these highly risky activities simply because of relative income differences. A hypothesis centered on absolute deprivation seems to be more appropriate to the phenomenon in question.
Protests in China increased at a stunning rate. Between 1993 and 1997, the total number of demonstrations rose from 8,700 to 32,000.26 According to official figures released by the Ministry of Public Security, there were 58,000 large-scale incidents of unrest in 2003, 74,000 in 2004, and 87,000 in 2005. In an ominous development, in September 2007, more than 2,000 demobilized soldiers rioted simultaneously in two cities 770 miles
260
Capitalism with Chinese Characteristics
apart from each other, indicating a high level of coordination. The ex-soldiers were protesting their poor living conditions ("Thousands of Ex-Soldiers Riot in China" 2007). Another group of statistics is even more startling. According to Professor Li Shuguang, a vice dean at the graduate school at Chinese University of Politics and Law, in 2005 the central government received 30 million petitions from Chinese citizens addressing various grievances. Professor Li also reveals that between 1979 and 1982, the Chinese government received only 20,000 similar petitions annually (Chen 2006). Professor Li is in a position to know. His university is closely associated with the Ministry of Public Security, which handles petitions.
We already have some preliminary evidence of the absolute deprivation in China. In Chapter 4, we saw that the poorest 10 percent of Shanghai's population lost income every year since 2001. It turns out that this was a development at the national level as well. During the period when GDP growth averaged more than 10 percent, a World Bank study reports that the income of China's poorest 10 percent of the population declined by 2.4 percent every year between 2001 and 2003 (McGregor 2006b). This is the first documented evidence that a large number of Chinese people -130 million people - have actually experienced an absolute reduction in their living standards. That this absolute deprivation occurred at a time of double-digit GDP growth is a worrisome sign that Chinese growth has taken on an inherently anti-poor bias.
2 The Other Path
The title of this section is taken from the seminal book by Hernando de Soto. Hernando de Soto (1989), a Peruvian economist, documents the barriers to indigenous, small-scale entrepreneurs in his native country of Peru. In a real social science experiment, de Soto assembled a research team to follow all the required bureaucratic procedures to set up a one-employee garment factory. The process took the team members 289 days and cost them a total of $1,231, equivalent to three years of an average Peruvian income. Hernando de Soto shows the massive, self-inflicted harm to a poor struggling economy because of policies repressing indigenous entrepreneurship.
The ideas of de Soto gained wide acceptance and acclaim among policy makers and academics around the world - until China came along. In his book, The End of Poverty, Jeffrey Sachs (2006), an influential development economist, questions the basic premise in de Soto's work. China's growth experience figures heavily in Sachs's critique. He argues that economic
Capitalism with Chinese Characteristics
261
growth is not a single-factor process and many factors other than the security of private property rights and the policy treatment of private-sector firms contribute to growth. China's growth, Sachs argues, did not depend first on "solving the deeds and titles."
Sachs does not recognize the distinction I make in Chapter 1 between the personal security of a proprietor and the security of her property. Sachs is right that China did not first solve "the deeds and titles" and then grow its economy. But China did solve the personal security of millions of those holding deeds and titles as a first order of policy business. China was moving directionally toward liberalism and, by the "nasty, brutal, and short" standard of the Cultural Revolution, did so by leaps and bounds. The incentive effect was massive.
An empirically accurate framing of the China story is entirely consistent with the essence of de Soto's claim - the most important contributory factor to broad-based economic growth is indigenous entrepreneurship. Among Chinese policy makers as well as among Western observers in the 1990s, there was an obsession with the supposed growth-boosting effects of foreign direct investment (FDI). (Incidentally, just at the time when de Soto's experiment encountered numerous difficulties in registering an indigenous firm in Peru, the country, under President Fernando Belaunde from 1980 to 1985, enthusiastically wooed FDI by tax and policy concessions.27) The leaders of Shanghai routinely highlighted the number of Fortune 500 MNCs making investments in the city, not the growth of household income, as their achievement. For many years, Western analysts habitually wrote off the economic prospects of India simply because that country was unable to attract FDI.
My argument that the decade of the 1980s was the true China miracle implicitly assigns zero weight to FDI in explaining China's economic takeoff. In the 1980s, very little FDI flowed into China. Let me make this view explicit here. I do it in two ways. First, I compare Zhejiang with Jiangsu. I have referenced Zhejiang several times in this book as the most successful entrepreneurial economy in China. There is another aspect to the Zhejiang story - it has attracted very little FDI. Jiangsu, a neighboring province, has exactly the opposite combination. Like Shanghai, Jiangsu has systematically repressed indigenous entrepreneurship with its left hand while lavishly courting foreign investors with its right hand. (The direction of the hand analogy is intentional.) Here is a contrast between the two provinces: Zhejiang has outperformed Jiangsu in every meaningful dimension of economic performance. I present more details on this in the following section.
262
Capitalism with Chinese Characteristics
I then present a stylized comparison between China and India. There are both positive and normative aspects to this comparison. On the positive side, the stellar performance of the Indian economy should debunk many myths about growth - the outsized role of FDI being one of them. On the normative side, the rise of India undermines, hopefully fatally, the intellectual underpinnings of an idea that, in fact, never had any empirical support in the first place - that democracy is anti-growth. As China ponders the question of whether to begin to reform its politics, a closer look at India's experience is highly relevant.
2.1 The Zhejiang Model
All our indicators are better than those of Ningbo [in Zhejiang province], except per capita income.
- Wang Mang, the mayor of Suzhou of Jiangsu province, 2004
In this quote, Mayor Wang of Suzhou city of Jiangsu province gets right to the essence of our tale of two provinces. Compared with Zhejiang, Jiangsu has everything on its side - FDI, high-tech industrial parks (with heavy support from another FDI-heavy economy, Singapore), bank loans, and massive investments - except for one thing that actually matters, economic performance. Mayor Wang cites the GDP per capita data. In doing so, he is understating the true differentials between the two provinces. The relative difference between Zhejiang and Jiangsu - the ratio of Zhejiang to Jiangsu -in terms of GDP per capita was 1.11 in 2006 (NBS 2007a, p. 71). But the gap was much larger in terms of per capita household income, a better measure of the material well-being of the average person. In terms of urban household income per capita, the relative difference is 1.30 and in terms of rural household income per capita, the relative difference is 1.26. The gap is even larger when measured by specific components of household income. Consistent with the idea that an entrepreneurial economy is better at wealth creation, an average urban resident in Zhejiang earned an asset income 3.4 times that of her counterpart in Jiangsu province.
These differences are not mere statistical abstractions. They have real welfare implications. In 1990, an average resident in these two provinces had a roughly identical life expectancy at birth: 71.37 years in Jiangsu and 71.78 years in Zhejiang. In 2000, the gap increased: 73.91 years in Jiangsu and 74.70 years in Zhejiang. There are other objective differences. An average rural resident in Zhejiang consumes and owns more telephones, computers,
Capitalism with Chinese Characteristics
263
color TV sets, and cameras than her counterpart in Jiangsu. She also lives in a bigger house.28
That Zhejiang's life expectancy surpassed that of Jiangsu is an extremely important data point: Zhejiang is a catch-up story. Today, Zhejiang is the richest province in terms of per capita GDP and per capita household income. (In this and the following comparisons, I exclude Beijing, Shanghai, and Tianjin because these three metropolitan areas do not have an agricultural sector.) To put it simply, Zhejiang is rich because it has grown faster. Jiangsu is also among the richest provinces in China today, but it is rich because it has always been rich. In 1980, Zhejiang was ranked No. 7 in the country in terms of per capita GDP, compared with Jiangsu's No. 3 position. Today, Zhejiang has the highest per capita GDP (minus Beijing, Shanghai, and Tianjin), whereas Jiangsu retains its No. 3 ranking. In 1980, Zhejiang and Jiangsu had the same level of rural household per capita income. By 1990, the Zhejiang/Jiangsu ratio was 1.15 and by 2006, it rose to 1.26 (NBS 2007a, p. 368). There are other performance differences as well. Jiangsu was more indebted, had much higher investment/GDP ratios, and a higher non-performing loan ratio. Thus, Jiangsu carries some of the same traits of China as a whole - it has grown very fast but it requires massive resources to power its growth.
A comparison of these two provinces is a near-perfect natural experiment. Their geographic conditions are almost identical. Located next to each other, both are coastal. Jiangsu is to the north and Zhejiang is to the south of Shanghai. The two provinces are also similar in terms of their history of entrepreneurial development. In 1952, private firms accounted for 57 percent of the sales value in the retail sector in Jiangsu and 60 percent in Zhejiang.29 At the beginning of the reforms, the size of the industrial non-state sector was quite similar as well. Historically speaking, these two provinces were among the most entrepreneurial and culturally developed in China. In the first half of the 20th century, both supplied industrialists and entrepreneurs to Shanghai, and throughout Chinese history, the two provinces produced some of the most prominent literary and political figures.30
The two provinces differed in two critical aspects. First, Zhejiang relied substantially less on FDI for its economic performance as compared with Jiangsu. In the second half of the 1980s, both provinces drew very little FDI, as measured by the proportion of FDI to total fixed-asset investments. In Jiangsu, the ratio was only 0.63 percent, about the same as the ratio in Zhejiang (0.65 percent). In the first half of the 1990s, as China became more open to FDI, this ratio rose in both provinces but much more rapidly in
264
Capitalism with Chinese Characteristics
Jiangsu. On average, FDI accounted for 13.6 percent of fixed-asset investments in Jiangsu, which was more than twice the level in Zhejiang during the same period (5.7 percent). Other measures, such as output and export shares by FIEs, indicate the same contrast.
The second difference is that Zhejiang was a vibrant entrepreneurial economy, whereas Jiangsu was a more statist economy.31 Jiangsu and Zhejiang represent two contrasting development models in China, a phenomenon first noted in 1986 by Professor Fei Xiaotong, China's most prominent sociologist. The Wenzhou model is characterized by a heavy reliance on private initiatives, a noninterventionist government style in the management of firms, and a supportive credit policy stance toward private firms. (Wenzhou is a city in southern Zhejiang province, hence the name of the model.) The foundation of the Wenzhou model was established in the 1980s, as a 1990 World Bank TVE study notes. Byrd and Lin (1990, p. 34) characterize the Wenzhou model as follows:
The famous "Wenzhou" model is characterized by free development of private enterprises (mostly household undertakings), a thriving financial market based to a large extent on private financial institutions, and extensive commercial relationships with distant parts of China.
In sharp contrast, the "Sunan model," which prevailed in Jiangsu, was highly interventionist and discriminatory against indigenous private entrepreneurship. In Jiangsu, private enterprises "are tolerated, but their development has been constrained by limits on loans, restricted access to inputs, and environmental and other regulations" (Svejnar and Woo 1990, p. 80). The Sunan model also carried a strong industrial policy approach with a heavy emphasis on the role of the government, rather than the private sector, in economic development. The model emerged in the 1980s and persisted until the late 1990s. (Since the late 1990s, the province has partially moved away from this model of economic development by privatization of the collective TVEs and more financial support for the private sector.)
The level of micromanagement in the Sunan model was extensive. Wuxi, a city in Jiangsu, is widely regarded as a progenitor of the Sunan model. In 1985, the Wuxi government adopted the following measures: (1) penalties for skilled workers who left collective TVEs for other jobs, including barring their family members from jobs in TVEs; (2) thorough status checks on enterprise registration documents and procedures; and (3) limits on managers' pay to three times the average payroll (Luo 1990, p. 150).
All of these measures were designed to constrain indigenous private firms by denying them access to quality human capital, raw materials, and finance.
Capitalism with Chinese Characteristics
265
And they all succeeded. In 1985, collective TVEs constituted 36 percent of the total number of industrial non-state firms in Wuxi but contributed 96 percent of the gross value of industrial output. The private sector in the industrial arena was inconsequential (Svejnar and Woo 1990, pp. 67-69). In terms of the share of private TVEs in total TVE output, in 1987, Zhejiang had a higher share, at 16 percent, compared with 11 percent in Jiangsu. But the differences grew greater in the 1990s. By 1997, in terms of the share of the private TVEs, it was 52 percent in Zhejiang and 28 percent in Jiangsu (Ministry of Agriculture 2003). Although the two models were formulated in the 1980s, it was the policies of the 1990s that accentuated their economic differences.
After Professor Fei proposed these two models, Chinese economists rigorously debated their respective merits. Now, the debate has been settled overwhelmingly in favor of the Wenzhou model. As I show earlier, the economy of Zhejiang grew faster and its residents are much richer than those in Jiangsu. We can speculate about why a stronger indigenous entrepreneurial economy is more efficient than a combination of statism and FDI. One reason could be that stronger indigenous entrepreneurship is associated with a larger local supply network, which expands local incomes more directly. In contrast, an FDI model relies heavily on export-processing that has low domestic value-added. The profits of this business model accrue to the foreign investors rather than to the local entrepreneurs. Another piece of evidence is that although both provinces have a similar export/GDP ratio, Zhejiang has a far lower import/GDP ratio, as compared with Jiangsu. A stronger local production base also enables foreign firms to source locally. In 2006, the foreign trade balance of FIEs in Zhejiang was US$18.6 billion, compared with US$16.2 billion in the case of Jiangsu (NBS 2007a, p. 741). However, FIEs in Jiangsu exported 3.3 times as much as FIEs in Zhejiang, suggesting a far greater domestic content of export production for FIEs in Zhejiang.
In a statist economy, the primary contributions of FDI maybe ameliorative in nature rather than additive to economic growth. That is, efficiency-enhancing FDI contributes to growth by offsetting the inefficiencies of the state sector. So, its contributions to growth are stunted. A counterfactual scenario is that the Sunan model would have done even worse without the FDI. This reasoning may explain why Jiangsu seemed to court FDI more eagerly than Zhejiang: Because Jiangsu systematically suppressed indigenous entrepreneurship, it needed to expend precious resources on importing foreign entrepreneurship. In this scenario, FDI still contributed to growth, but not nearly as much if Jiangsu had had a more vibrant private sector.