| As an important mode of human capital investment, firms' training is of positive significance to individual employee, individual firm and the country as a whole. However, firms' training is not popular in China. The article uses the theory model, which discusses firms' investment in training on condition of information asymmetry, to explain the relationship between firms' training investment and quit rate and the phenomenon of low investment in training in Chinese firms. The article is organized as follows:The third chapter analyses the status of Chinese firms' training investment in recent years and compare it with American firms' training investment. First, this chapter discusses the status of firm training in various industries, firms of various ownership and various regions in China, and then introduces the capital input, the contents and structure and the design and practice of American firm training. Finally, the status of firm training of two countries is compared in the respects mentioned above. And the conclusion is that training investment of Chinese firms is very low as a whole.The fourth chapter dwells on the theory model of firms' training investment on condition of information asymmetry. Firms' training investment and employees' quit rate are the result of the game among the training firms, their employees and the potential firms. There exits information asymmetry about employees' ability between training firms and their employees in training period. And the training firm can master the information regarding employees' ability after training. The potential firms do not provide training and so they can not hold the information about employees' ability and wages offered by the training firm as well. Given the wages offered by the training and potential firms, employees make the decision of whether to leave the training firm or not. Given the wage offered by firms in second-hand market and employees' quit rate, the training firm decides its wage to trained employees and investment in training. The potential firms figure out the productivity of employees in second-hand market according to the Bayes' Rule and decide its wage offered to them. There exits multiple equilibria in this game. And in the equilibrium of high training investment, the quit rate is low and vise versus. That is, the training investment and quit rate are negatively related and are necessary and sufficient conditions for each other.The fifth chapter analyses the status of employee' turnover in various industries and firms of various ownership and induces the model of employees' turnover. Factors in four respects affect employees' turnover, including social and economic factors, organization, job and personal characteristics. Social and economic factors include labor demand and supply and policies and laws regarding employees' turnover. The better the labor market operates and the stronger the labor demand is, the higher the quit rate is. The more is the labor supply than the labor demand and so the higher the unemployment rate is, the lower the quit rate is. The more perfect the relative policies and laws are, the more normative the employees' turnover is and so the quit rate can be maintained around a reasonable level. The organizational and job characteristics include job-involvement, organization-support and compensation, which are all negatively related with quit rate. The personal characteristics include age, education and traits such as the character. The older the age is, the weaker the quit intention is. Education affects turnover through specificity of human capital. The stronger the specificity of human capital is, the weaker the quit intention is.Finally, conclusions of individual chapter are summarized and policy implication and possible directions of future studies are discussed. Relative policies and regulations about employees' turnover should be carried out to improve firms' training in China. Meanwhile, firms should make efforts to improve organization and job characteristics to actively manage and control employees' turnover. |