Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

An Examination of the CCP’s Strategies to Alleviate Discontent After the Great Recession of 2008

191 views

Published on

The following research project examines the various strategies of the Chinese Communist Party in the aftermath of the Great Recession of 2008, to reduce the discontent of the population.

Published in: Government & Nonprofit
  • Be the first to comment

  • Be the first to like this

An Examination of the CCP’s Strategies to Alleviate Discontent After the Great Recession of 2008

  1. 1. 1 An Examination of the CCP’s Strategies to Alleviate Discontent After the Great Recession of 2008. Luigi Caloi, Sahaj Sood and Alon Mor Dr. Christina Jenq Political Economy of East Asia December 12, 2016 Abstract In 2007, per the World Bank, China’s heavily export-oriented economy posted a growth rate of 14.195%.1 The onset of the Great Recession in the fall of 2008, following the collapse of the Lehman Brothers investment bank, significantly reduced world demand for Chinese exports. This contributed to a significant rise in unemployment in China which, in turn, led to an increase in discontent among the population. The political ramifications of this could have been catastrophic for the Communist Party (CP) as strong economic performance is one of the regime’s main sources of legitimacy. However, the CP successfully used its Hukou-based system of migration and instituted a RMB 4 trillion stimulus package to minimize the probability of political instability. We define political instability as the probability of the CP’s collapse2 . While literature exists that examines the ways in which the government used these two instruments to alleviate the rising levels of discontent after the crisis, no other paper appears to provide an empirical explanation behind the CP’s actions to endure the fallout from the recession. We employed a theoretical framework (another point of differentiation from the existing literature) and conducted a series of correlations and regressions to test many of the existing literature’s findings. Our tests show that urban provinces in China are more unstable than their rural counterparts and that the CP made rural provinces the primary beneficiaries of its generous stimulus package. Our theoretical framework explains these findings by considering investments in rural provinces as investments in political control and investments in urban provinces as investments in economic growth. At first glance it may appear that since good economic performance is an important source of legitimacy for the CP, the goals of political control and economic performance are mutually reinforcing. However, during periods of severe external shock such as the recession in 2008, our theoretical framework assumes this to certainly not be the case, underlining the importance for the CP to find the optimal allocation of its investment budget between economic performance/urban provinces and political control/rural provinces. 1 http://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG?end=2009&locations=CN&start=1978 2 Alesina, Alberto, Sule Ozler, Nouriel Roubini, and Phillip Swagel. “Political instability and economic growth.” Journal of Economic Growth 1.2 (1996): 189-211.
  2. 2. 2 1. Introduction China embarked on a period of market-oriented reform in 1978 under Deng Xiaoping, but did so using a “gradualist” approach as opposed to the “Big-Bang” approach of rapid reform adopted by many former countries of the Soviet Union.3 The gradualist approach aligned itself with the CP’s interests as it allowed the regime to initiate selective economic reform while creating new patronage resources4 —necessary to ensure the allegiance of newly created autonomous interest groups to the regime.5 In 1993, the change in the structure of power caused by the death of the eight elders saw a shift towards a more decisive approach towards reform.6 Economic performance has since become an important source of legitimacy for the CP. However, it is important to remember that the goals of good economic performance and maintenance of political control are neither mutually exclusive, nor completely mutually reinforcing. This means that there may be times when the CP must prioritize the maximization of political control over economic performance to remain in power. The most recent example of such a time was during the Great Recession of 2008. 1.1 The Great Recession China’s route to economic growth was heavily export-oriented, helped by the regime’s artificial devaluation of the renminbi.7 At the time of the crisis, the export sector employed an 3 Naughton, Barry. “Four: A Political Economy of China’s Economic Transformation. 1st ed. N.p.: Cambridge, n.d. 91-135. Print. 4 ibid 5 ibid 6 ibid 7 Wallace, Jeremy L. Cities and Stability: Urbanization, Redistribution, and Regime Survival in China. New York, NY: Oxford UP, 2014. Print.
  3. 3. 3 estimated 100 million people.8 These jobs were largely concentrated in China’s more affluent coastal areas, described by Wallace as “the economic vanguard of the country.”9 Between September 2008 and January 2009, the number of Chinese exports fell from approximately $140 billion to $90 billion.10 The average number of goods shipped out of China fell from approximately 4.5 billion in September 2008 to less than 1.5 billion in January 2009.11 Of interest to us was the difference between the severity of the impact of the recession on the coastal regions and the rural interior. Since most of the export activity was centered in the coastal regions, the negative impact of the recession would be greater felt in those regions. This is clearly illustrated in the following maps, which show the relative fall in 2009 exports relative to 2008 GDP by province (left) and industrial employment losses by province in 2009 (right).12 8 Tong 2012 9 Wallace, Jeremy L. Cities and Stability: Urbanization, Redistribution, and Regime Survival in China. New York, NY: Oxford UP, 2014. Print. 10 ibid 11 ibid 12 ibid
  4. 4. 4 Furthermore, Tong found that China’s top five most export-oriented provinces for 2008 of Guangdong, Fujian, Shanghai, Zhejiang, and Jiangsu were among the following year’s worst performers in industrial activities, writing that “while industrial employment declined by 0.1% nationwide, that in Guangdong and the three Yangtze River delta provincial units declined significantly more (by 3.8 percent in Guangdong, 6.5% in Shanghai, 3.3% in Zhejiang, and 7.1% in Jiangsu).”13 The less-export oriented provinces of Anhui, Hubei, Inner Mongolia and Sichuan showed comparatively better industrial growth in 2009.14 Given the importance of exports to the Chinese economy, unemployment soared during the recession. About two-fifths of the newly unemployed in the world economy following the crisis were Chinese.15 Kim Wang Chan estimated the number of unemployed Chinese to be greater than 20 million16 while Huang et al. place their estimate at around 50 million.17 The graph below shows the number of protests recorded per province in 2008, illustrating the rising discontent within the population. As shown, Guangdong, Fujian and Shanghai recorded among the highest levels of protests. These three, together with Jiangsu and Zhejiang, were among China’s most export-oriented sectors of 2008 per Tong.18 13 Tong 2012, 103. N.b. Industrial Employment data. 14 ibid 15 International Food Policy Research Institute (IFPRI) 2009; Xinhua Net 2009; Chan 2010d, 660. 16 Chan 2010c. 17 Huang et al. 2010. 18 Tong 2012, 103. N.b. Industrial Employment data.
  5. 5. 5 1.2 The CP’s Response The CP used the Hukou-based system of migration and instituted a RMB 4 trillion stimulus package to alleviate the rising discontent.19 After the recession eradicated much of the economic incentive for migrants to work in the coastal areas, the Hukou-system created incentives for the so-called “temporary population” to return to their homes in the rural interior where, even without the relatively higher wages in the coast, the newly unemployed would have shelter and some land to cultivate for food-growing.20 Land, therefore, was the rural man’s social security. Huang et al. estimated that between 20 million and 40 million people migrated back to their Hukou-registered homes.21 19 Wallace, Jeremy L. Cities and Stability: Urbanization, Redistribution, and Regime Survival in China. New York, NY: Oxford UP, 2014. Print. 20 ibid 21 Huang et al. 2010 0 50 100 150 200 250 Protests (2008) Protests (2008)
  6. 6. 6 Meanwhile, the CP’s stimulus package was designed to benefit export-oriented firms on the coast in the short-term but boost the long-term employment prospects of the rural interior.22 One could perhaps be forgiven for assuming, given how the coastal cities were more adversely affected by the recession, that they would be the main beneficiaries of the stimulus package. The following maps show that this was not the case:23 The map on the right shows that the provinces of the rural interior were the main recipients of the center-approved debt in 2008.24 The map on the left, shown previously, illustrates the difference in severity of the recession between the coast and the rural interior. It shows that export losses in 2009 were greater in the coastal regions than in the rural interior.25 The decision to direct the stimulus package primarily towards rural provinces reflects a reality that will be developed further in the next section: the CP’s marginal benefit of investment in maintaining political control was greater than the CP’s marginal benefit of investing in economic performance. 22 Wallace, Jeremy L. Cities and Stability: Urbanization, Redistribution, and Regime Survival in China. New York, NY: Oxford UP, 2014. Print. 23 ibid 24 ibid 25 ibid
  7. 7. 7 2. Theoretical Framework We assume that at any given time the CP’s main aim is to remain in power. The payoff to the CP (probability of political stability) can be denoted by the following function, πg = f (e, c) where π, g, e and s stand for payoff to the government, government, economic performance, and political control respectively. As mentioned earlier, we define political instability as the probability of a CP collapse and we define political stability as the CP’s maintenance of political control. It is important to remember the important distinction between political stability and political control. The CP’s objective is to maximize the probability of political stability by finding the optimal allocation of its budget, the stimulus package, between economic performance and political control. This optimization can be denoted by the following function, MBe = MCe MBc = MCc where MBe and MCe refer to the CP’s marginal benefit and marginal cost of investment in economic performance, respectively and MBc and MCc refer to the CP’s marginal benefit and marginal cost of investment in political control, respectively. We assume that: i. Prior to the recession, the CP was optimally allocating its investments between its goals of economic performance and political control. We call this time t0 and is represented by the above optimization function.
  8. 8. 8 ii. The instability caused by the recession increased both the CP’s marginal benefit of investment in political control and the CP’s marginal benefit of investment in economic performance, but that the former rose by more than the latter. Simultaneously, the CP’s marginal cost of investment in economic performance also increases to reflect the rising opportunity cost of NOT investing in political control. We call this time t1, shown below: MBe < MCe MBc > MCc iii. The budget constraint faced by the CP that prevents it from dedicating unlimited resources to economic performance and political control is represented by the following function: Y = Ie + Ic where: Y = amount of the government’s budget that can be allocated to investment in economic performance and political control. Ie = investment in economic performance Ic = investment in political control iv. To maximize payoff/probability of political control during the recession, the CP must reallocate its budget/stimulus package to match the changes in the CP’s marginal benefits and marginal costs described earlier. This would mean increased investment into political control relative to economic performance. As investment into political control increases, the marginal benefit of doing so naturally decreases while the marginal cost of doing so
  9. 9. 9 increases. Simultaneously, the marginal cost of investing in economic performance decreases to reflect the falling opportunity cost of NOT investing in political control. At some optimal allocation of the budget, the CP will again maximize its payoff/probability of political stability. This occurs at t2: MBe = MCe MBc = MCc Graphical Representations of t0, t1, and t2: Change from t0 to t1
  10. 10. 10 So far, we have restricted our discussion of the sources of discontent to only include the unemployment caused by the recession. In the next section, we explore alternative sources of discontent. 3. Related Literature/Alternative Sources of Discontent From our research, we have gathered the following possible alternative sources of discontent, including unemployment: i. GDP per capita per province ii. Percentage change in GDP per capita per province iii. Social security as a percentage of GDP iv. Unemployment v. Unmarried men as a percentage of total population of men province Anticipations of future income are direct factors of current happiness. This helps explain China’s apparent social stability despite the “immense socioeconomic transition that previous rapid growth had brought.” The same simulations used to come to this conclusion also imply that Change from t1 to t2
  11. 11. 11 a reversal in income expectations would lead to a quick sharp fall in levels of happiness, which could cause social instability (Frijters). Historical evidence support these findings; “social protest becomes more likely when people’s aspirations, influenced by recent achievements, exceed their subsequent achievements.”26 Implication 1: Lack of economic growth can lead dissatisfaction and to political instability Between 1990 and 2010, per capita consumption has increased fourfold in China. While incomes have increased for all income groups, income inequality has increased, causing a disparity in life satisfaction. Increasing levels of economic growth and urbanization have in fact changed China from one of the most egalitarian countries to one of the least. China’s shift towards economic liberalization and privatization has led to the removal social safety nets, in turn, increasing levels of income inequality and insecurity.27 In general, life satisfaction has become more dependent on income and perceived income. Several surveys have been conducted by Pew Research Center, the Chinese Academy of Social Sciences, Gallup, and the World Values Survey (WVS), measuring subjective well-being (SWB) of the Chinese population. Figure 1 displays a U-shaped pattern where life satisfaction in the Chinese population declined from 1990 to around 2000-2005 and then turned upward.28 This is consistent with Mancur Olson’s proposition that rapid economic growth in recently underdeveloped countries is a destabilizing force. As rural people move to urban centers they become aware of the high level of inequality and social injustice. Those in poverty become 26 Knight, John. “The Economic Causes and Consequences of Social Instability in China.” China Economic Review 25 (2013): 17-26. Web. 11 Dec. 2016. 27 Easterlin, Richard, Morgan, Robson, Switek, Holgozata, and Wang, Fei. “China’s Life Satisfaction, 1990-2010.” PNAS. 109.25 (6 Apr. 2012): 9775-9780. Web. 11 Dec. 2016. 28 ibid
  12. 12. 12 aware of not only their inequities but also of their prospects for change and begin to aspire higher, staying in cities and continuing growth. The downside is an intermediary rise in dissatisfaction.29 Implication 2: Rapid economic growth could lead to higher dissatisfaction Rather than steadily increasing along with GDP, the U-shaped pattern mirrors an inverted U-shape in the urban employment rate. Comparing these trends to the European transition countries, the fact that life satisfaction did not increase as markedly as income and output, indicates the fundamental importance of employment and the social safety net in determining social well-being. According to Pew surveys, in 2002, when unemployment was near its high point, nearly 48% of respondents indicated that their economic situation was “somewhat bad” or “very bad.” In 2010 when unemployment dropped, life satisfaction rose from 5.27 in 2002 to 5.85 in 2010. This indicated that not only the unemployed experience decreased in life satisfaction during troubling economic times.30 Implication 3: Unemployment can be a source of dissatisfaction and instability. Amongst the dissatisfied, migrants show to be the most affected by economic circumstances. As migrants increasingly move to urban areas, the life satisfaction of the urban population as a hole tends to decrease. 31Surveys show that long staying migrants have a higher 29 Olson, Mancur. “Rapid Growth as a Destabilizing Force.” The Journal of Economic History 23.04 (1963): 529-52. Web. 11 Dec. 2016. 30 Easterlin, Richard, Morgan, Robson, Switek, Holgozata, and Wang, Fei. “China’s Life Satisfaction, 1990-2010.” PNAS. 109.25 (6 Apr. 2012): 9775-9780. Web. 11 Dec. 2016. 31 Easterlin, Richard, Morgan, Robson, Switek, Holgozata, and Wang, Fei. “China’s Life Satisfaction, 1990-2010.” PNAS. 109.25 (6 Apr. 2012): 9775-9780. Web. 11 Dec. 2016.
  13. 13. 13 coefficient on the income variable than average.32 They are also more sensitive to average urban income per capita in their destination provinces. The China Household Income Project (CHIP) national survey of 2002 found that the income coefficient is highest in urban areas; the effect of income on satisfaction is greatest in urban cities.33 When migrants move from rural to urban areas, they become more dependent on income as a source of satisfaction. Implication 4: GDP per capita and rapid economic growth impacts the decision making of a migrant. Implication 5: Employment opportunity can impact the decision making of a migrant. According to two studies on the role of “perceived change in income over the previous five years” current satisfaction is sensitive on perceived change in income. A past rise in income increases satisfaction and a past fall in income decreases current satisfaction.34 Respondents in the categories of “unhappy” and “not at all happy” stated income as their primary reason for unhappiness. The next most important reason, reported by over 11% is “uncertainty about the future.” This suggests that insecurity is a problem.35 Migrant households settled in the city having rural hukou apparently have the lowest mean satisfaction among city dwellers. There are two reasons for this: (1) Rural people have narrow reference groups, with 68% of rural dwellers reporting that they primarily compare themselves to their “fellow villagers and neighbors.” (2) Migrants experience economic and 32 Knight, John, and Ramani Gunatilaka. "Aspirations, Adaptation and Subjective Well-Being of Rural–Urban Migrants in China." Adaptation, Poverty and Development (2012): 91-110. Web. 11 Dec. 2016. 33 Knight, John. “The Economic Causes and Consequences of Social Instability in China.” China Economic Review 25 (2013): 17-26. Web. 11 Dec. 2016. 34 Knight, John. “The Economic Causes and Consequences of Social Instability in China.” China Economic Review 25 (2013): 17-26. Web. 11 Dec. 2016. 35 Knight, John, and Ramani Gunatilaka. "Aspirations, Adaptation and Subjective Well-Being of Rural–Urban Migrants in China." Adaptation, Poverty and Development (2012): 91-110. Web. 11 Dec. 2016.
  14. 14. 14 social discrimination in the city.36 With economic well-being and perceived income being determinants of social satisfaction, it is natural that upon moving to urban cities, rural migrants experience decreases in satisfaction. Implication 6: Lack of social security can cause dissatisfaction and instability Yet even though migrants in urban areas are less satisfied on average than rural residents, 56% of migrants feel that urban living gives greater happiness. There is a notion, originating from sociological literature and furthered in companion papers on subjective well-being, that aspirations depend on income relative to reference groups. Migrants compare themselves to their urban counterparts, drawing inference about their future income. Since 1988 the national Gini coefficient of household income per capita rose from 0.38 to 0.49 in 2007 (then the highest in Asia). The apparent inequality in cities and attachment to relative income help explain why happiness scores fail to show a strong positive correlation between increase in happiness and rapid growth of incomes in China.37 Insecurity is another factor of unhappiness. Before the reform of state-owned enterprises in the mid-1990s, urban residents enjoyed lifetime employment, coined the “iron rice bowl.” The collapse of social welfare systems lead to unemployment, coinciding with U-shape analysis in figure 1. Figure 3 shows that since 1990 the disparity in life satisfaction between the upper third income group and the lower third has increased. In 1990 the proportion of respondents reporting a high level of life satisfaction was 68% vs 65% respectively. In 2007 this shifted to 71% vs 42% (Easterlin). China’s economic restructuring undermined the health care system Health care has 36 Knight, John. “The Economic Causes and Consequences of Social Instability in China.” China Economic Review 25 (2013): 17-26. Web. 11 Dec. 2016. 37 Knight, John, and Ramani Gunatilaka. "Aspirations, Adaptation and Subjective Well-Being of Rural–Urban Migrants in China." Adaptation, Poverty and Development (2012): 91-110. Web. 11 Dec. 2016.
  15. 15. 15 become increasingly privately financed, though still mostly public. However health care costs have risen, pricing out the lower tiers of the population and in some cases causing people to fall into poverty.38 China’s market-oriented economic reforms has made the employment-based social security system inadequate and inefficient. Previously state-owned enterprises (SOEs) provided workers with health care, job security, and pension funds. But SOEs’ disappointing performance after China’s liberalization has led to cuts in SOE support, fundamentally changing China’s industrial structure. Specifically, the elderly are under-cared for. In 2000 71.4% of older people relied on their children for support while only 16.9% received pension.39 In response, the government has since focused on getting more people employed and revamping the pension system. The goal is to fund the elderly with money from the payroll tax (9% of wages) and to create a defined contribution system for current employees’ pension.40 Our data analysis finds that the government has invested more heavily into regions with decreasing pension funds, indicating that employment and economic growth may be a priority in ameliorating dissatisfaction from decreasing pensions. Another source of insecurity is the uneven sex ratio among young people. In 2008 the ratio of males to females was 1.21, 1.15, and 1.14 among young people aged 5-9, 10-14, and 15- 19 respectively. The result is intense marriage competition not only among young men, but also 38 Organization for Economic Development and Cooperation (2010) China in the 2010s: Rebalancing Growth and Strengthening Social Safety Nets (Organization for Economic Development and Cooperation, Beijing 39 Leung, Joe. “Social Security Reforms in China: Issues and Prospects.” International Journal of Social Welfare 12.2 (2003): 73-85. Web. 11 Dec. 2016. 40 Feldstein, Martin. “Social Security Pension Reform in China.” China Economic Review 10.2 (1999): 99-107. Web. 11 Dec. 2016.
  16. 16. 16 among their parents. Wei and Zhang (2011) find that households with a son “increase their savings in a competitive manner to improve their son’s relative attractiveness in marriage.41 Furthermore, rural households with a son have higher conditional income, attributed to provide for their son.42 Implication 7: Unmarried men as a percentage of total population of men can be a source of dissatisfaction and instability. In the next sections, we provide empirical support for the theory that investments in rural provinces represent investments towards political control while investments in urban provinces represent investments towards economic performance. 4. Data Description Protest data is for obvious reasons a very sensitive data in China. For that reason, one can hardly have access to it. In order to solve this issue, we collected protest data from Google Database of Events, Language, and Tone (GDELT). GDELT has a computer algorithm that scans the web for all print, web, and radio news articles mentioning a given event for any regions. We used GDELT to gather protest data for China. Official Chinese GDP figures have faced criticism for inaccuracy43 and so must be used with caution. However, we trust that for our purposes—to compare the GDP among provinces— the data should be valid. We collected GDP per capita per province data and unmarried men data 41 Wei, Shang-Jin, and Xiaobo Zhang. "The Competitive Saving Motive: Evidence from Rising Sex Ratios and Savings Rates in China." Journal of Political Economy 119.3 (2011): 511-64. Web. 11 Dec. 2016. 42 Knight, John, Li Shi, and Deng Quheng. "Son Preference and Household Income in Rural China." Journal of Development Studies 46.10 (2010): 1786-805. Web. 11 Dec. 2016. 43 Cary, Eve. "The Curious Case of China's GDP Figures." The Diplomat. The Diplomat, 2013. Web. 13 Dec. 2016.
  17. 17. 17 from China Data Online.44 Percentage or urban population was taken from the annual data per province from the National Bureau of Statistics.45 Data on the registered rate of unemployment in urban areas was taken from the the Reporting Form System on Training and Employment Statistics, which provided by the Ministry of Human Resources and Social Security. We used the CEIC China Database to gain access to this data.46 Finally, data on pension fund per province, which we used as a proxy to social security, was taken from the Department of Population and Employment Statistics of the National Bureau of Statistics, again using the CEIC China Database.47 5. Methodology In this section, we outline our methodologies to test the following two hypotheses: I. Hypothesis: In 2008, the more urbanized a province, the more unstable the province The stimulus package was primarily intended to stabilize China’s political economy by creating new employment opportunities in the rural interior to counter the brimming discontent.48 These measures were based on the premise that dispersing discontent among the larger area of the rural interior represented a more manageable proposition for the CP than if the discontent were to be clustered/concentrated within a few coastal provinces and cities.49 Our empirical analysis to test 44 "China Yearly Macro-Economics Statistics(Provincial)--All China Data." China Yearly Macro-Economics Statistics(Provincial)--All China Data. N.p., n.d. Web. 13 Dec. 2016. 45 "NBS Statistical Data." National Bureau of Statistics of China. N.p., n.d. Web. 13 Dec. 2016. 46 "体验最完整的一套超过128个国家的经济数据." Compare Economic Data for over 120 Countries CEIC. N.p., n.d. Web. 13 Dec. 2016. 47 ibid 48 Wallace, Jeremy L. Cities and Stability: Urbanization, Redistribution, and Regime Survival in China. New York, NY: Oxford UP, 2014. Print. 49 ibid
  18. 18. 18 the notion that clustered/concentrated urbanization shows a positive relationship with the level of discontent (using number of protests as a proxy) is divided into two steps: Step 1: We performed a correlation between number of protests per province in 2008 and urban population as a percentage of total population per province in 2008. We also conducted correlations between the number of protests per province in 2008 and the other following sources of discontent to uncover any possible alternative relationships: i. GDP per capita per province (2008) ii. GDP per capita growth per province (2008) iii. Unmarried men as a percentage of total population of men per province (2008) iv. Registered unemployment in urban areas per province (2008) v. Pension fund as a percentage of GDP per province (proxy for social security fund) (2008) Step 2: We substantiated our correlations with the following OLS regression model: Protests in 2008 = β0 + β1 percentage of urban population (2008) + β2 xi We then added the other variables into the model to test the possible alternative relationships. For that, we ran the following regressions: 1. Protests in 2008 = β0 + β1 percentage of urban population (2008) + β2 GDP growth (2008) 2. Protests in 2008 = β0 + β1 percentage of urban population (2008) + β2 percentage of unmarried men (2008)
  19. 19. 19 3. Protests in 2008 = β0 + β1 percentage of urban population (2008) + β2 registered urban unemployment rate (2008) 4. Protests in 2008 = β0 + β1 percentage of urban population (2008) + β2 social security/people>65 (2008) 5. Protests in 2008 = β0 + β1 percentage of urban population (2008) + β2 GDP per capita (2008) II. Hypothesis: In 2008, the government directed its stimulus package primarily towards the rural interior to mitigate discontent After testing the first hypothesis, we attempt to show how the CP’s stimulus package was strategically directed to further incentivize the reverse-migration of the so-called “temporary population” back to their Hukou-registered jurisdictions.50 According to our theoretical framework, after the recession, the CP’s marginal benefit of investment in political control increased relative to that of economic performance. Since our findings support the argument that more urbanized provinces are more unstable than their rural counterparts, it follows that any CP investment towards rural regions represents an investment in increased political control. Step 1: Our first step was to correlate percentage change in fixed asset investment (same proxy used by Wallace for the stimulus package)51 from 2008 to 2009 with the following factors influencing migration: i. GDP per capita per province (2008) 50 Wallace, Jeremy L. Cities and Stability: Urbanization, Redistribution, and Regime Survival in China. New York, NY: Oxford UP, 2014. Print. 51 Wallace, Jeremy L. Cities and Stability: Urbanization, Redistribution, and Regime Survival in China. New York, NY: Oxford UP, 2014. Print.
  20. 20. 20 ii. Percentage change in GDP per capita per province (2007 to 2008) iii. Rural population as a percentage of total population per province (2008) iv. Social security as a percentage of GDP per province (2008) v. Registered urban unemployed per province (2008) vi. Total unemployment rate per province (2008) vii. Unmarried men as a percentage of total population of men per province (2008) Step 2: We substantiated our correlation analyses with an OLS regression model. Our hypothesis is that the two main factors that explain the direction of the stimulus package are (i) low urbanization (rural areas) and (ii) low GDP per capita. We face the obvious challenge that GDP per capita and urbanization rate are correlated. Thus, of the two variables, we picked one—GDP per capita—to construct the following model: Percentage change in fixed asset investment = β0 + β1 GDP per capita + β2 xi We then ran the following regressions to test for alternative explanations and variables: 1. Percentage change in fixed asset investment (2008 to 2009) = β0 + β1 GDP per capita (2008) + β2 social security/people>65 (2008) 2. Percentage change in fixed asset investment (2008 to 2009) = β0 + β1 GDP per capita (2008) + β2 registered urban unemployment rate (2008) 3. Percentage change in fixed asset investment (2008 to 2009) = β0 + β1 GDP per capita (2008) + β2 GDP per capita percentage change (2007 to 2008) 4. Percentage change in fixed asset investment (2008 to 2009) = β0 + β1 GDP per capita (2008) + β2 percentage of urban population (2008)
  21. 21. 21 5. Percentage change in fixed asset investment (2008 to 2009) = β0 + β1 GDP per capita (2008) + β2 percentage of unmarried men (2008) 6. Percentage change in fixed asset investment (2008 to 2009) = β0 + β1 GDP per capita (2008) + β2 percentage on non-employed (2008) 6. Empirical Analysis Testing Hypothesis I: We conducted a series of correlations and regressions between number of protests per province in 2008 and an index of sources of instability in 2008 to test the plausibility of the notion that urban areas are more unstable than rural areas. A. Correlations Finding 1: The figure below clearly indicates that the number of protests per province recorded in 2008 was greater than the same recorded in 2006 and 2007 for all twenty-nine provinces shown except for the following seven: Tianjin, Hebei, Liaoning, Jiangxi, Henan, Yunnan, and Ningxia. Beijing and Tibet, outliers, were excluded from the graph altogether. Beijing regularly sees a disproportionately larger number of protests per year than other provinces, while we believe it is reasonable to assume that most protests in Tibet are in some way related to the Central Tibetan Authority’s (CTA) protracted struggle for independence from the People’s Republic of China (PRC).
  22. 22. 22 The figure below illustrates the same phenomenon, including the disproportionate number of protests recorded in Beijing and Tibet: 0 50 100 150 200 250 Tianjin Hebei Shanxi InnerMongolia Liaoning Jilin Heilongjiang Shanghai Jiangsu Zhejiang Anhui Fujian Jiangxi Shandong Henan Hubei Hunan Guangdong Guangxi Hainan Chongqing Sichuan Guizhou Yunnan Shaanxi Gansu Qinghai Ningxia Xinjiang Number of Protests per Province (2006, 2007, 2008) excl. Beijing and Tibet 2006 2007 2008 0 500 1000 1500 2000 2500 3000 Beijing Tianjin Hebei Shanxi InnerMongolia Liaoning Jilin Heilongjiang Shanghai Jiangsu Zhejiang Anhui Fujian Jiangxi Shandong Henan Hubei Hunan Guangdong Guangxi Hainan Chongqing Sichuan Guizhou Yunnan Shaanxi Gansu Qinghai Ningxia Xinjiang Tibet Number of Protests per Province (2006, 2007, 2008) incl. Beijing and Tibet 2006 2007 2008
  23. 23. 23 Finding 2: The scatterplot below is a graphical representation of the statistically significant positive correlation between number of protests per province in 2008 and the percentage of urban population per province in 2008. This finding suggests that the coastal and urban regions saw more protests in 2008 than did rural regions. Correlation: 0.585766819; p- value: 0.001326126 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00 100.00 0 50 100 150 200 250 %ofurbanpopulation(2008) Protests per province (2008) Number of protests per province (2008) and % of urban population (2008) excl. Beijing, Tibet, Tianjin, and Gansu
  24. 24. 24 Finding 3: We observe a statistically significant negative correlation between GDP growth per province from 2007 to 2008 and the number of protests recorded per province in 2008, suggesting that rural provinces, which are more likely to have higher GDP growth rates, are less conducive to protests than urban provinces, which are more likely to have lower GDP growth rates. We exclude outliers Beijing and Tibet. Correlation: -0.54089; p-value: 0.00244867. Finding 4: The correlation between GDP per capita per province in 2008 and number of protests per province in 2008 is positive and statistically significant, implying that urban provinces, whose residents are likelier to have higher incomes, tend to record a greater number of protests than rural provinces, whose residents are likelier to have lower incomes. We exclude outliers Beijing, Tibet, Tianjin, and Gansu. Correlation: 0.58037796; p-value: 0.00150527. Finding 5: The correlation between the number of unmarried men as a percentage of the number of total men per province in 2008 and the total number of protests per province in 2008 was found to be positive and statistically insignificant. The correlation implies that the higher the number of unmarried per province, the greater the level of frustration among the men in society regarding their inability to find a spouse, leading to greater protest. We exclude outliers Beijing and Tibet. Correlation: 0.15779432; p-value: 0.41363673. Finding 6: We found a small, positive correlation between the percentage of registered urban unemployed residents per province in 2008 and the number of protests per province in 2008, suggesting that the higher the registered unemployment rate per province, the greater the
  25. 25. 25 number of protests in the year 2008. We exclude outliers Beijing and Tibet. Correlation: 0.11765383; p-value: 0.1176 Finding 7: Social security per population size older than 65 was found to be positively correlated to the number of protests per province for 2008. The statistically significant positive correlation implies that the greater the proportion of people over the age of 65 benefiting from social security, the greater the number of protests. While this correlation is counter-intuitive, it could simply be reflecting the correlation between social security and GDP per capita per province. We exclude outliers Beijing, Tibet, Gansu, and Fujian. Correlation: 0.509351926; p- value: 0.00665499. B. Regressions The empirical evidence supports the premise that percentage of urban population is the biggest factor for the growth of protests in 2008. First, when we added registered urban unemployment rate, social security or GDP per capita to the model (regressions 3, 4 and 5 respectively), all showed unreliable results. For the three, the variables’ p-values were extremely high (e.g. in regression 5, 2008 urban percentage of population p-value was 0.409, and 2008 GDP per capita p-value was 0.499). This could be explained by the fact that these measures are correlated with each other. Regression 1, on the other hand, was our most reliable result. We found that: Percentage change in protest (2008 -2009) = 36.33742 + 1.9961 urban percentage of total population – 3.8384 GDP growth.
  26. 26. 26 In other words, the regression analysis shows us that if we held the urbanization rate of two provinces constant, a 1% increase in GDP growth will result in a 3.84% decrease in the number of protests. This result is consistent with the hypothesis that GDP growth is correlated to stability and high rates of urbanization are linked to high numbers of protests. Moreover, this regression analysis recorded the highest adjusted R2 among all our regression analyses (0.3947). However, the p-value for GDP growth was 0.051, which fails our statistical significance test. We face some limitations due to the natural fact that most of the alternative variables that could explain a rise in protests—lower GDP growth, unemployment, social security—were correlated with urbanization rates in China in 2008. Yet, our empirical analysis gives support to the premise that higher rates of urbanization in a province increases the likelihood of protests in that province. Our results are summarized in the following table:
  27. 27. 27 Red: P-value > 0.05 Green: Adjusted R2 > 0.3168 (the Yellow: Lowest P-value, but > 0.05 Adjusted R2 when Urbanization rate is the only variable)
  28. 28. 28 Testing Hypothesis II: We conducted a series of correlations and regressions between percentage change in fixed asset investment per province between 2008 to 2009 and an index of sources of discontent in 2008 to test the plausibility of the notion that the CP strategically formed an incentive scheme for the newly unemployed population in coastal cities to migrate back to their Hukou-registered jurisdictions. We expect investment to be negatively correlated to urbanization rates and positive correlated to GDP per capita. A. Correlations Finding 1: The percentage change in investment from 2008 to 2009 per province was negatively correlated with number of protests in 2008 per province. At first glance, one could doubt our hypothesis that the government prioritized political control during the recession, because of the statistically significant negative relationship of the stimulus package with the protest data. Correlation: -0.62; p-value = 0.000557565. 0 50 100 150 200 250 0 5 10 15 20 25 30 35 40 45 Protests % of investment change % change of investment (2008 - 2009) and protests (2008)
  29. 29. 29 However, as we saw in part 1, protests are negatively correlated to rural areas. Since our first hypothesis is that the stimulus package should go to rural regions to mitigate discontent, this first correlation does not contradict our hypothesis. Finding 2: Percentage change in investment from 2008 to 2009 per province was negatively correlated with percentage of urban population in 2008 per province When analyzing the investment change data, Sichuan stands as an outlier (the percentage change on fixed asset investment for Sichuan was 59.54%). This already gave us an intuitive idea that the investment had indeed gone to rural regions. We then excluded Sichuan and Tianjin from our data set, and ran the correlation of percentage of urban population and the stimulus package. The correlation confirms our initial hypothesis. Correlation: -0.564403209; p-value = 0.00142623. 0 5 10 15 20 25 30 35 40 45 0.00 10.00 20.00 30.00 40.00 50.00 60.00 70.00 80.00 90.00 100.00 Investment%Change % of Urban population % Urban population (2008) v. Investment % change (2008-2009) (excl. Sichuan and Tianjin)
  30. 30. 30 Finding 3: We found that the correlation between the GDP per capita and the percentage investment change was -0.708771096. Again, we excluded Sichuan and Tianjin because both were outliers. The correlation was statically significant, with p value = 0.0000168 and it also had the highest R-squared of all the correlations we ran (R2 = 0.502356467) and the lowest standard error (8.82413E-05). This result supports our hypothesis that investment should be directed to poorest areas to create a compelling incentive scheme for migration. Correlation: -0.708771096; p-value: 0.0000168; R2 : 0.502356467; SE: 8.82413E-05 Finding 4: We found that the GDP per capita growth from the year 2007 to 2008 and the percentage investment change was positively correlated and statistically significant. This positive relationship was neither surprising, nor did it add a new element to our explanation. Due to the negative correlation of GDP per capita and GDP growth in we expected that the stimulus package would be directed towards areas that were growing (considering that they are usually poorer/more rural). We test this explanation in our forthcoming regression analysis. Correlation: 0.484540105; p-value: 0.005738978. Finding 5: We found that the correlation between social security and the investment change was 0.477737384. The p-value was 0.007586965. We used pension fund divided by total population above the age of 65 as our proxy for social security. This relationship supports our previous finding that China lacks a robust social safety net, and that safety net plays a role in decision making of an unemployed migrant of whether to stay in an
  31. 31. 31 urban area or go back. In other words, this indicates that the government might have opted to invest in regions that had a lower safety net, to maximize the incentive scheme for migration. On the other hand, this could also be reflecting the positive correlation of GDP per capita and social security. To further test these two possibilities, we run a regression analysis with both variables—GDP per capita and social security. Correlation: 0.477737384; p-value: 0.007586965 Finding 6: We found a statically significant correlation between registered urban unemployment rate and the percentage investment change from 2008 to 2009 of 0.44993708, and a p-value of 0.014325681. Most notably, the coefficient here was the highest among all of the findings, equal to 7.632114755. The standard error here, however, was 2.915358632, which made us cautious about the validity of this finding. Considering the role that the existence of job opportunities play in the decision-making of a recent unemployed migrant, it would not be surprising if the CP targeted provinces with high urban unemployment. The possibility of an unemployed migrant finding a job in his/her Hukou- based jurisdiction directly increases the probability that he/she will migrate back. Correlation: 0.44993708; p-value: 0.014325681 Finding 7: We found that the correlation between the percentage of unmarried men to total population of men per province in 2008 and the percentage investment change from 2008 to 2009 was -0.108551701. However, the p-value for this correlation was statistically insignificant. Correlation: -0.108551701; p-value: 0.561054883
  32. 32. 32 Finding 8: Finally, we found that the correlation between the percentage investment change from 2008 to 2009 per province and the percentage of non-employed per province was 0.03. Yet, the p-value here was remarkably high: 0.876615829. We calculated the percentage of non-employed of a province with the following formula: ((Population > 15 years old and < 64 years old) – total employed) ______________________________________________________________________________________________________________________________ (Population > 15 years old and < 64 years old) This formula takes into account those that are currently not working but also not seeking a job. In other words, one explanation for the correlation’s unreliability is because the formula accounts for the non-employed, for instance, women who opt not to work and stay home. Correlation: 0.03; p-value: 0.876615829. Our findings are summarized in the following table:
  33. 33. 33 B. Regressions Regression Analysis: Finding 9: In all six regressions, on the one hand, all variables became statistically insignificant when paired with GDP per capita. On the other hand, GDP per capita remained statistically significant in all six. This increases the robustness of the conclusion we reached from our correlation analyses: GDP per capita was the main variable driving the decision making of the CP with regards to the direction of the stimulus package. Moreover, in the regression that included GDP per capita as the sole variable, its coefficient was -0.000461. When we included the other six variables, its coefficient varied between -0.000844 and -0.000407, but remained mostly close to its initial value. This low variation increased our confidence in the robustness of the result. Next, percentage of urban population was the second most important variable. From all six regressions we ran, percentage of urban population was the only variable that increased the adjusted R2 of the regression (from 0.484 to 0.527) and it had the smallest p-value when paired with GDP/capita. However, its p-value was still above 0.05 (p-value = 0.0748) so we concluded that it was, nevertheless, statistically insignificant. Although we reached a single-variable regression formula, this result shouldn’t be downplayed. Existing literature discussing the CP’s stimulus package also show the CP’s prioritization of investing in rural areas over urban areas. Our results agree with the existing literature, but also empirically indicate that GDP per capita was the main variable driving the CP’s decision making regarding the direction of the stimulus. This result reflects the CP’s strategy to not only invest in the rural regions, as already discussed in the existing literature, but
  34. 34. 34 to ensure investment in the poorest of the rural regions, in order to maximize the efficiency of its incentive scheme for migration. Yet, we must be cautious when interpreting these results. Instead of indicating the relevance of GDP per capita to explain the direction of the stimulus package, the regressions could be misleading because some of the variables used are correlated, leading to possible alternative explanations for causality. For instance, GDP per capita is usually higher in provinces that have higher urbanization rates. See below for a summary of our analyses:
  35. 35. 35 Red: P-value > 0.05 Green: Adjusted R2 > 0.4839 Yellow: Lowest P-value, but > 0.05 Adjusted R2 when GDP per capita is the only variable) 7. Conclusion To summarize, the purpose of our paper is to test and support two hypotheses: I. In 2008, the more urbanized a province, the more unstable the province II. Therefore, in 2008, the government directed its stimulus package primarily towards the rural interior to alleviate discontent. Our theoretical framework shows that this was rational for the CP because of its aim of minimizing political instability, which is defined as the probability of the CP’s collapse. In our payoff optimization function, we define the payoff as the probability that the CP remains in power (political stability) and it can use economic performance and the maintenance of political control as its instruments to achieve this. In economic terms, during the recession, the CP’s marginal benefit of investment towards political control increases relative to the CP’s marginal benefit of investment towards economic performance because of the rising discontent among the population. Simultaneously, it follows that the CP’s marginal cost of investment towards economic performance rises relative to the CP’s marginal cost of investment towards political control to reflect the rising opportunity cost of not investing in political control. To achieve the optimal allocation of resources that maximizes government payoff/minimizes political instability, the CP must dedicate a larger portion of its resources towards political control than economic performance until such time that the following conditions are satisfied:
  36. 36. 36 MBe = MCe MBc = MCc where MBe and MCe refer to the CP’s marginal benefit and marginal cost of investment in economic performance, respectively and MBc and MCc refer to the CP’s marginal benefit and marginal cost of investment in political control, respectively. We develop this framework to eventually conclude that during the recession, the CP viewed investments in rural areas as contributing more towards political control and investments in urban areas as contributing more towards economic performance. We reach this conclusion empirically, by running a series of correlation and regressions that support it. We also find that, not only did the CP choose to invest more in rural areas during the recession, they chose to invest in the poorer of the rural regions in order to maximize the efficiency of their incentive scheme to further encourage the reverse- migration of the “temporary population” back to their Hukou-based jurisdictions.
  37. 37. 37 8. Works Cited Naughton, Barry. “Four: A Political Economy of China’s Economic Transformation. 1st ed. N.p.: Cambridge, n.d. 91-135. Print. http://data.worldbank.org/indicator/NY.GDP.MKTP.KD.ZG?end=2009&locations=CN&start=1 978 Alesina, Alberto, Sule Ozler, Nouriel Roubini, and Phillip Swagel. “Political instability and economic growth.” Journal of Economic Growth 1.2 (1996): 189-211. Wallace, Jeremy L. Cities and Stability: Urbanization, Redistribution, and Regime Survival in China. New York, NY: Oxford UP, 2014. Print. Tong 2012, 103. N.b. Industrial Employment data. International Food Policy Research Institute (IFPRI) 2009; Xinhua Net 2009; Chan 2010d, 660 Huang et al. 2010. Knight, John. “The Economic Causes and Consequences of Social Instability in China.” China Economic Review 25 (2013): 17-26. Web. 11 Dec. 2016. Easterlin, Richard, Morgan, Robson, Switek, Holgozata, and Wang, Fei. “China’s Life Satisfaction, 1990-2010.” PNAS. 109.25 (6 Apr. 2012): 9775-9780. Web. 11 Dec. 2016. Olson, Mancur. “Rapid Growth as a Destabilizing Force.” The Journal of Economic History 23.04 (1963): 529-52. Web. 11 Dec. 2016. Knight, John, and Ramani Gunatilaka. "Aspirations, Adaptation and Subjective Well-Being of Rural–Urban Migrants in China." Adaptation, Poverty and Development (2012): 91-110. Web. 11 Dec. 2016.
  38. 38. 38 Organization for Economic Development and Cooperation (2010) China in the 2010s: Rebalancing Growth and Strengthening Social Safety Nets (Organization for Economic Development and Cooperation, Beijing Leung, Joe. “Social Security Reforms in China: Issues and Prospects.” International Journal of Social Welfare 12.2 (2003): 73-85. Web. 11 Dec. 2016. Feldstein, Martin. “Social Security Pension Reform in China.” China Economic Review 10.2 (1999): 99-107. Web. 11 Dec. 2016. Wei, Shang-Jin, and Xiaobo Zhang. "The Competitive Saving Motive: Evidence from Rising Sex Ratios and Savings Rates in China." Journal of Political Economy 119.3 (2011): 511-64. Web. 11 Dec. 2016. Knight, John, Li Shi, and Deng Quheng. "Son Preference and Household Income in Rural China." Journal of Development Studies 46.10 (2010): 1786-805. Web. 11 Dec. 2016. Cary, Eve. "The Curious Case of China's GDP Figures." The Diplomat. The Diplomat, 2013. Web. 13 Dec. 2016. "China Yearly Macro-Economics Statistics(Provincial)--All China Data." China Yearly Macro- Economics Statistics(Provincial)--All China Data. N.p., n.d. Web. 13 Dec. 2016. "NBS Statistical Data." National Bureau of Statistics of China. N.p., n.d. Web. 13 Dec. 2016. 1 "体验最完整的一套超过128个国家的经济数据." Compare Economic Data for over 120 Countries CEIC. N.p., n.d. Web. 13 Dec. 2016.

×