The remarkable differences in the incidence of poverty among provinces are largely explained by the interprovincial income inequality. For example, per capita income in Guangdong is 11,975 yuan, which is 3.5 times as large as the per capita income of the lowest income province, Qinghai.
Urban poverty is sensitive to two main determinants: per capita mean income and inequality. If inequality is measured by the Gini index, we can run the following simple regression to derive the poverty elasticities.
There are 31 observations from 31 provinces in 1998. Because the poverty incidence for each province is measured by both low and high poverty lines, the regression model can be run in two different versions. The first uses the poverty incidence measured by the low poverty lines. The second uses the poverty incidence measured by the high poverty lines. The regression results are reported in Table 2.
All the estimated coefficients are statistically significant below the 5% significance level. The incidences of poverty, either measured by the low or high poverty line, are highly sensitive to the mean incomes and income inequality. Since the model is specified in a double-log form, the estimated coefficients are the corresponding elasticity of poverty incidence with respect to the independent variables. For example, the elasticity of poverty measured by the low poverty line with respect to mean income is _1.572. This means that if mean income rises by 10%, poverty incidence will decline by about 16%. The results show that poverty incidence is slightly more sensitive to mean income if it is measured by the high poverty line than if it is measured by the low poverty line.
Poverty incidence is also sensitive to income distribution. For example, if the Gini coefficient increases by 10%, the poverty incidence measured by the high poverty line will rise by almost 15% if per capita mean income is unchanged. The results also show that poverty incidence is far more sensitive to income distribution if it is measured by the low poverty line than if it is measured by the high poverty line. If the Gini coefficient rises by 10%, the poverty incidence measured by the low poverty line will increase by over 30%.Table 1
Poverty incidence by province in urban China (1000 people and percentage)Table 2
Effects of income and income distribution on poverty incidence
3. The incidence of rural poverty and its response to income and inequality
There are no data on food and nonfood consumption at the household level to derive rural poverty lines in the same way as for the urban sector. As a result, we have to rely on estimates provided by the statistical authority, the World Bank, and some other independent researchers. The starting point is to use the estimates provided by the Chinese government. The official poverty line for 1990 was set at 318 yuan per capita per year.
The official poverty line has been criticised for being too low. Hence, it tends to underestimate the extent of poverty, although it is a useful benchmark for poverty analysis. The World Bank (1997) used a higher line at 454 yuan per capita per year in 1990. It was based on $1 per day per capita at PPP. The imputed PPP exchange rate was 1.24 yuan per dollar, but the official exchange rate was 4.83 yuan per dollar in 1990. In empirical studies, different authors have used different poverty lines. We do not repeat their methodologies of estimations, but various studies can be found in Gustafsson and Li (1998), Gustafsson and Wei (2000), Jalan and Ravallion (1998, 2000), Khan and Riskin (2001), Riskin (1994), and Riskin and Li (2001).2 In this paper, however, we follow World Bank (1997) and Yao (2000) using $1 per day per capita at PPP. Poverty lines in different years are adjusted by the consumer price index for the rural population or the rural CPI. The respective poverty lines are 561 and 801 yuan per capita in 1995 and 615 and 877 yuan per capita in 1998 at current prices.
It is important to note that the rural poverty lines are substantially lower than the urban ones. In 1998, the low and high urban poverty lines are 1869 and 2316 yuan per capita, respectively. The corresponding rural poverty lines are only 615 and 877 yuan per capita. The rural/urban poverty lines ratios are 33% and 38%, respectively. In the same year, per capita urban income was 5633 but per capita rural income was 2447 yuan, with a per capita rural/urban income ratio of 43%.
It is well recognised that the rural and urban populations are highly heterogeneous,not only because of the large urban/rural income disparity, but also because of the large differences in the cost of livings between the cities and the countryside. However,the significant differences of poverty lines are unlikely to be explained by the differences in the cost of livings. It is likely that in comparison with the rural poverty lines, the urban poverty lines are overestimated. Alternatively, one may argue that the rural poverty lines are underestimated in comparison with the urban poverty lines.Therefore, we should be cautious when making poverty comparisons for the urban and rural sectors.
We are unable to obtain household survey data from the (National Bureau of Statistics)NBS. However, we have obtained data from two alternative sources: the Chinese Academy of Social Science (CASS) for 1995 and the Ministry of Agriculture (MOA)for 1998.
The data set for 1995 contains about 8000 households from 19 provinces. The data set for 1998 contains about 15,000 households from the same provinces. Although the data sets come from different sources, they have similar characters and contain household level income and population information. To make the two data sets comparable, we use provincial level rural CPI to convert the values in 1995 into comparable values in 1998. In other words, the income levels are all measured in 1998 prices.
The sample provinces had a total rural population of 777 million people in 1998. They represent about 83% of the national rural population. From the sample data, per capita income was 2377 yuan. The national average rural income was 2163 yuan per capita (NBS, 2000), which was 25% higher than in 1995 (1727 yuan/head).
The Gini coefficient for the whole sample is .4606, which is 22% higher than the Gini coefficient derived from the 1995 data (.3773). The Gini coefficient varies widely across provinces, ranging from .30 in Hubei to .52 in Yunnan. Except Zhejiang, Guangdong, and Yunnan, all the provinces have a Gini coefficient smaller than the national Gini coefficient. This is an indirect indication of large interprovincial income inequality. Per capita income ranges from 1226 yuan in Gansu to 4378 yuan in Beijing. Most provinces in the East have a per capita income much higher than the national average. These include Beijing, Liaoning (2930), Jiangsu (2789), Zhejiang (3815), and Guangdong (4283). The lowest income provinces are concentrated in the west, including Shaanxi (1619) and Gansu.
The poverty headcount rates by province are presented in Table 3. At the national level, the incidence of poverty increased by 2.36 percentage points from 1995 to 1998 if it is measured by the low poverty lines, but declined by 1.37 percentage point if it is measured by the high poverty lines. This reversed change in poverty rates using different poverty lines suggests that poor people are highly concentrated around the low poverty lines. As a result, the incidence of poverty is highly sensitive to alternative measurements of poverty.
The increase in poverty measured by the low poverty lines is particularly worrying, not because of the rise in poverty incidence, but because of the implied difficulty in poverty reduction in rural China. National average per capita income actually rose by about 25% from 1995 to 1998, which is an impressive growth, but national income growth did not benefit the poor due to rising inequality.
The changes in poverty headcount rates at the provincial level are far more complicated than at the national level. If we focus on the change in poverty rates measured by the high poverty lines, although most provinces experienced a significant reduction in poverty, some provinces experienced an increase. The largest reduction in poverty was found in the following provinces: Liaoning, Jilin, Zhejiang, and Hubei. Most provinces experiencing a rise of poverty were highly concentrated in the central regions, including Anhui, Jiangxi, and Henan. One obvious explanation was that there was a prolonged and severe flood in central China in 1998, which must have had a detrimental impact on agricultural production and farm incomes. The estimated results suggest that the poor households are highly susceptible to natural disasters.
上一页 [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] ... 下一页 >>