逆向选择是经济学理论研究的重要课题，也是国际上保险经济学研究的前沿课题。由于逆向选择，保险需求受到抑制，保险市场的资源不能够得到有效的配置。对于中国这样的新兴保险市场，这种现象可能更为突出，但相关的研究却非常有限。因此，本论文集中研究了保险市场的逆向选择问题，并特别针对国内汽车险市场进行了深入的实证研究，这对我国保险市场的发展具有重要的实践指导意义。本论文研究的创新性内容包括三个方面。首先，论文对国际前沿研究的保险市场逆向选择模型进行了理论拓展，具有明显的理论创新价值。论文严格证明了在投保人风险类别无穷多的情况下，Rothschild-Stiglitz模型的基本结论仍然成立，即高风险类别的投保人会选择较高的保险金额，较低的免赔额。论文还就道德风险、风险分类、异质偏好对投保人的逆向选择和最优保险需求影响进行了讨论，说明了模型结论的稳健性。其次，论文依据理论拓展的结果，在保险经济学领域创新和发展了精微模拟的研究方法，考察了一个动态的保险市场。模拟结果表明，存在逆向选择时，较高的投保人风险厌恶程度和损失额的不确定性是保险市场可以长期平稳运作的必要条件；而通过学习和对投保人出险历史的分析，保险公司能够准确识别投保人的风险类别，从而使逆向选择问题不再严重。最后，我们利用中国汽车险市场中商业汽车险和个人汽车险的实际数据，采用多种特征变量和计量经济学模型对理论模型进行了实证检验。结果显示，高风险类别的投保人在投保时确实趋向于购买更多的车损险保额，更低的免赔额，从而验证了中国汽车险市场保单持有人确实存在明显的逆向选择问题。即使部分剔除道德风险因素后，中国的汽车险市场仍然存在显著的逆向选择现象。而如果采用多期保单，则有助于减轻保单持有人的逆向选择问题。这些实证研究结果，以及与国际上对成熟市场研究成果的比较，不仅支持了理论模型、精微模拟的结论，而且对于市场发展和保险公司的业务创新也具有明显的指导意义，验证了研究对于中国保险市场发展的实际价值。

Adverse selection is not only an important issue in economic theory research, but also the frontier of the international insurance economic research. Because of adverse selection, insurance demand was inhibited and resources can not be efficiently allocated in the insurance market. This phenomenon may be more prominent in the emerging insurance markets such as in China. In China the related researches are very limited. Therefore, this paper focuses on the problem of adverse selection in the insurance market, and conducts an in-depth empirical study on the domestic automobile insurance market, which provides an important practice guide for the development of China's insurance market. The innovative content of this thesis includes the following three aspects.First, the thesis tries to extend the adverse selection model of the insurance market. It is rigorous proved that the basic conclusion of the Rothschild-Stiglitz model still holds, i.e., the higher risk category policyholders will opt for more insurance amount, less deductible. To exam the robustness of the model, the impacts of the moral hazard, risk classification, heterogeneous risk aversion on the insurance demand and adverse selection are also discussed.Second, the thesis develops the micro simulation research method in the field of insurance economics. Based on the results of previous model extensions, we investigate a dynamic insurance market by simulation. It is shown that the higher risk aversion and the higher loss severity uncertainty are necessary for the long term operation of the insurance market with the existence of adverse selection. And through study and analysis of the insurance history, the insurers can accurately identify the policyholders’ risk category, so that the adverse selection problems will not be serious.Finally, the thesis empirically exams the models’ prediction by the data of China’s commercial and personal automobile insurance markets. Results show that high-risk type insurance policyholders do tend to buy more insurance and lower deductible, which proves there are indeed obvious adverse selection problems in China's automobile insurance market. The effects of adverse selection are still very significant even the impacts of moral hazard are excluded. But the multi-period insurance policy may help to alleviate the problem of adverse selection. The empirical findings, as well as the comparison with the findings in the mature insurance markets, not only strongly support the conclusions of model extensions and simulation studies, but also provide some practical guides for the market development and the innovations in insurance companies’ business, which are valuable for the development of China's insurance market.