Income Effects Of Human Capital Investment

Human Capital Theory has been rapidly developed since Shultz and Becker proposed the concept of human capital. At the same time, a large number of empirical evidence shows that human capital is very important in economics. The economic returns to human capital investment are the key problem on which is focused by researchers. In recent years, with rapid development of the capability of collecting and dealing with individual data, the effect research of human capital investment become the hot point focused by society, that is, individual obtains the benefits from the investment in the aspect of education, train, labor mobility and medical care so on.Human Capital Theory considers that human capital is the ability of physical strength, health, experience, knowledge and technique embodied in human body, which is formed by formal education, train, migration, and medical care so on. For different investment, the effect is different on income, consume, quantity and returns. However, all these investments enhance people's technique, increase their knowledge and improve their health, so that workers have more flexibility in their employment, finally have more lifetime income.Theory needs support of evidence, especially in the field of human capital research. Becker thought, the main reason of its flourishing was"the tight combine between theory and demonstration, and theory research is usually inspired by the empirical results". Therefore, empirical study is important in the process of human capital theory development.Given a broad overview of the present reference, Mincer's income function and its extension are widely applied in the study of effects of human capital investment, but the unobservable individual heterogeneity is the problem disturbing the econometrician all along. So how to precisely estimate the contribution of human capital investment on individual income is very important. The other problem disturbing the econometrician is the validity of micro data. Since the micro data set is the basis of empirical study of human capital investment, the estimators maybe include the character of data to produce bias.During the past twenty years, the capability of analyzing complicated data is rapidly developed, especially McFadden and Heckman overcome the dependence of linear models in discrete choice analysis, censored and truncated models, consequently discrete choice and/or tobit model is used to estimate returns to human capital investment, such as expenditure in schooling, train, labor mobility and medical care so on.In our country, related empirical study began later because of lacking of micro data. There were some researchers'analysis based on micro data sets in recent years, but they paid more attention to the research content which is different from foreign scholars. Most of references are about the compare analysis of individual returns to human capital between different sexes, between different departments, between different areas and between different periods; the variation of returns during the process of economic system reform; and the rural situation of human capital investment. In the light of research methods, more researches apply to linear model regression so as to ignore the bias of estimators, so the results may be not precise. And due to the time of data earlier, corresponding empirical results maybe not reflect the real status, although some researchers tried to apply new econometric methods to estimate.Therefore, the paper applies to econometric methods to explore the income effect of human capital investment in China's labor market in the theory frame of economics based on micro survey data in 2006. The following problems are discussed:Firstly, the paper evaluates the effect of education on labors'income. The paper discusses ordinary least square and instrument variable methods respectively in homogeneous model and heterogeneous model. By analyzing, we find that there are three sources of bias in the estimate results: ability bias, return bias and measurement error bias. Increasing control variables is help for reducing ability, and finding another measurement way is help for reducing the measurement error bias. The IV estimate results can be the unbiased estimators when a series of hypothesis are satisfied. But it needs to be noted that the IV estimators are different with different instruments in heterogeneous model. Furthermore, the paper applies Heckman two-step process to eliminate sample selection bias and combines it to instrument variables methods.The estimate results show that the effect of education is substantial positive on individual income. By correcting ability bias and sample selection bias, the average returns to schooling of civil male and female are 7.7 percent and 7.3 percent respectively; those of rural male and female are 3.2 percent and 4.0 percent respectively. It presents that the average returns to schooling of citizen are more than villager. However, it does not provide direct evidence that the average returns to schooling of female are more than that of male. The results still show that, individual income is not linearly increased with the increasing of schooling level. The highest average return to schooling is college, the next is high school, middle school, and the lowest is primary school, in which the values between male and female of accepting middle or high school are very close, but the return to college of female is obvious higher than that of male which shows that income difference is reducing in the cohort. And for the estimation of local average treatment effect of college, whether people studied in key high school is a stronger instrument variable, and the average return to college who studied in key high school is higher than that of who did not.Secondly, the paper evaluates the effect of train on labors'income. Because of the validity problem of observable data, such as ignoring important variables or disturbed by macro economic trend, the paper applies a way of quasi-social experiment - propensity score-matching method to acquire more precise estimator. The key point of this method is finding the treated group similar to control group by matching, in which the aim of propensity score is reducing the dimensionality of matching.The estimate results show that the effect of train is substantial positive on labors'income. The income of male and female who accept train is increased by 157 and 133 yuan respectively than that of who do not. And the income effect of male is higher than that of female. Although the outcome of ordinary least square and cross section do not support the conclusion, the estimate results of propensity score-matching method are more closer to that of random experiment based on the comparing statistical characters between treated and control group. So it is more precise. Furthermore, we find there is obvious positive correlation relationship between the decision-making of accepting train and the signing of labor contracts, which means train is help for steadying relationship between employers and employees.Thirdly, the paper evaluates the effects of mobility on rural labors'income. With the deep development of market economy, mobility has been an important form of promoting labor employment, especially for rural labor. We define the main factors affecting labor mobility by statistical analyzing micro data, and test it using Logit model. The paper applies Heckman two-step process to estimate the economic benefits of mobility concerning sample selection bias based on extension of Mincer's income function.The estimate results show that the migrating action certainly increases the price of labors'knowledge and technique in labor market. In all labor, the returns to schooling and experience of real migrator are higher than non-migrant, in which the younger and educated people can obtain relative higher income after migration. Otherwise, the returns to schooling and experience of latent migrant are not only lower than that of real migrant, but also lower than latent non-migrant, which further supports the hypothesis of labor mobility theory, that is the motive of mobility is to obtain higher expectant benefits. Nevertheless, the results still show that some factors maybe affect decision-making of mobility, such as educational level, age, family background, and geographical factor so on. For example, it is not only reducing the probability of labor emigration, but also make them back home if there is a kid in family.Finally, the paper analyses the effect of educational policies on acquirement of human capital. We apply micro econometric method and gross effect analysis respectively to explore the effect of compulsory schooling law and educational expansion on individual schooling level and benefits.The implement of compulsory schooling law certainly increases the basic educational level, and indirectly raises the rate of coverage of high school. However, the implement of the law do not make residents'income substantial variation, which means that the income of people who attend school longer because of compulsory schooling laws is not obvious different from who do not attend, that is to say, their income in 2006 are not different.The analysis on educational expansion presents that the sign of total relative supply index and the special age group relative supply index between college graduates and high school graduates are both negative, and very remarkable. It means that whether changing structure of education or structure of age of labor cohorts can affect negative income effect for college graduates, that is, large increasing relative supply of college graduates will arouse large reducing college wage premium. All in all, the effect of human capital on individual income is very remarkable, and the effect of implement of related educational policies on attendance of education is also remarkable, but the effect of policies on income is still not confirmed.The paper pays attention to analyze the economic effects of education and other human capital investment, and do not discuss their social benefits. The reason is that it is difficult to measure the latter, but it does not mean other effects are not important, so a lot of work need to be done in the future. In a word, the research methods in this paper have important reference significance, and the research outcomes have important value to constitute and improve related policies.