教育、户籍转换与城乡教育收益率差异

Education,Migration and the Urban-Rural Difference in Returns to Education

为何农村教育收益率低于城市,但农村居民的教育需求却很高?这种看似矛盾的现象可以从教育帮助农村居民突破户籍限制的角度进行解释。利用中国居民收入调查数据(CHIP2013),通过将极度正向选择性的永久移民分离出来,并放回到农村样本中去以修正样本选择偏差,以1986年义务教育法实施的自然实验作为工具变量,我们发现农村教育收益率有显著提高,由原始样本的3.2%提高到5.7%。对农村居民而言,往往是高等教育更有利于他们突破户籍限制,因而,我们还估计了大学教育收益率,发现城市大学教育收益率为50.4%,农村大学教育收益率也高达50.6%,考虑到城市户籍附着的潜在福利收益,农村大学教育收益率可能会更高。这意味着,政府应该进一步加强对农村的教育投资,以促进更多农村居民进入城市体系,从而有利于我国经济的结构性转型。

Summary： Returns to education are closely related to a country＇s economic system. Before 1978, the economy in China was governed by a planning system, and education was not rewarded. Earlier studies usually find low returns to education in urban China and even negative returns in rural China. Since open reform, returns to education have risen, and returns to education in urban China are now approaching the world level. Many studies find that returns to education for rural people are much lower than those of urban people. However, their demand for education, especially higher education, is greater than that of urban people. It is puzzling that returns to education are lower while demand for education is higher for rural people. In this paper, we try to solve this puzzle. There is a special arrangement in the Chinese household registration system （the hukou system）. The hukou system divides China into two sectors, urban and rural. Rural hukou holders could not move to urban China in the past. Now, restrictions are lighter： rural hukou holders can work in cities, but they cannot enjoy the same welfare benefits as local urban hukou holders. However, there are some ways to change hukou status. Education is one of the most important channels to by which to obtain urban hukou status. Usually, when a rural child enters college, his or her hukou status will change to urban hukou. Joining the army is another way to change hukou status. With the recent rapid urbanization, land expropriation has become another important way for rural people to obtain urban hukou status. We call such people ＂permanent migrants＂. Permanent migrants are usually not included in investigations of rural people; the remaining rural people will be the selected sample. Their education levels are lower, as is their potential income. This is why earlier studies obtained low estimates of returns to education for rural people. In this paper, we use a unique dataset, the China Household Income Project 2013 （CHIP2013） , which consists of three samples： the rural sample, the urban sample, and the migrant sample. CHIP2013 provides hukou transfer information, including whether, when, and why rural people transfer their hukou status to urban. This information can help us to identify permanent migrants so that we can include them in the rural sample to address the sample selection problems. After adjusting the rural sample, we obtain much higher returns to education for rural people than earlier estimates in the literature. In particular, we get similar magnitudes of returns to higher education for rural and urban people. This implies that education plays an equivalent role for rural people. We first estimate the classical Mincer equation with the original sample and the adjusted sample. Rural and urban returns to education are 3.8% and 8.5% , respectively, for the original sample： urban returns are much larger than rural returns. After adding permanent migrants to the rural sample, the estimates are 5.6% and 8.6% , respectively. Because higher education enables rural people to transfer their hukou status, we estimate the returns to higher education with the same procedure. The rural and urban returns are 23.6% and 36.8% , respectively; after adjustment, the estimates are 33.7% and 36.1% , respectively： the gap is decreased. The omitted variable problem is a concern when identifying the causal effect of education on income. We take advantage of the fact that the data contain admission scores （of college entrance examinations） for higher education and re-estimate the Mincer equation by controlling them. The estimates are smaller, which is consistent with the ability bias story in econometrics. As only controlling scores may not be sufficient, we also use the implementation of compulsory law in 1986 as an IV to identify the causal effect of returns to education. The IV estimates provide similar results, and the rural and urban returns to higher education are 50.6% and 50.4% , respectively. The empirical results suggest that the returns to education for rural people are not low and that the government should invest more in education in rural China, which will help smooth China＇s ongoing process of structural transformation.