近几十年来，世界范围内经济的飞速发展也带来了很多社会问题，收入不平等的程度不断加剧对社会和人类的发展都有着极大危害，随着社会关注度的逐渐升高，对收入不平等问题的剖析和研究也越来越多。在探寻收入不平等现象的原因方面，目前已有的研究多集中在人类社会和经济制度、体制、政策等的缺陷和不合理，这些研究可以帮助我们理解收入不平等的形成原因，但是现有的研究对非人为影响的机制性因素研究不够，即使在毫无人为因素的影响下，收入不平等也有其自身发展的规律，而这时影响它的就是非制度性因素。另外，多因素交互作用，收入不平衡的形成原因必然是多方面相互作用、共同促成的，很难看出其中某一因素对最终结果的影响程度。因此，我提出了针对单一因素对收入不平等影响的理论模型，这与对实证数据进行分析和处理而得出结论的研究方向也有所不同。本文结合数学推导以及数据模拟，以基尼系数和前1%人口收入份额作为衡量收入不平等程度的指标，分析了两大非制度性因素，即城市化和职业多样化，对收入不平等的影响作用，研究了亿万富翁人数与收入不平等的关系，最后探讨了组内收入不平等对整体不平等的影响。在模型相应的假设条件下可以发现，随着城市化趋势的发展，基尼系数先升高后降低，前1%人口收入份额随之下降，先急后缓；随着职业多样化的发展，在最高收入和最低收入保持不变时，基尼系数逐渐降低，前1%人口收入份额随之上升，先急后缓；而随着职业种类越来越多，不同职业的收入呈等比或等差分布时，基尼系数逐渐升高；亿万富翁人数与基尼系数、前1%人口收入份额均呈正相关关系；当组内基尼系数被纳入考虑时，公式计算的整体基尼系数和模型模拟的基尼系数都与实际基尼系数更为相符。研究过程首先考察最特殊、最简单的情况，然后逐渐贴近实际情况，加大复杂性，再进一步通过极端情况和数值模拟验证模型的合理性，最后查找相应的实证数据，通过现实情况考察模型的有效性和应用价值。这种由简到繁的技术路线与一般研究的从实证入手、由繁到简有很大不同。不得不说，这些模型并不完美，假设条件过于简单和理想化，并不能很好地解释真实世界里越来越严重的收入不平等现象。但是作为该类研究的基础模型，直击研究的核心问题而刚好可以忽略其他因素的干扰作用，使研究思路简明清晰。同时相关研究者可以通过借鉴研究思路而调整假设条件，来推导更加贴近实证情况的研究结论。

In the past a few decades, the rapid development of the world economy also brought many social problems. The continuous increase of income inequality hurts social community and human development. With more attention has been drawn on this issue, there are more analysis and research on income inequality.In terms of the causes of income inequality, many studies have focused on economic system, social policy, and governmental regulation. They could help us gain a better understanding that how income inequality comes into reality. However, few people notice how the semi-mechanical factors affect income inequality. Moreover, income inequality comes from the interaction among multiple factors. It is difficult to recognize how much effect is from each one of the factors. Therefore, I brought several theoretical models to discuss the effect of a single factor on income inequality. The method is quite different from other research which generates conclusions by conducting research on real data.In this paper, I use mathematical derivation and simulation to analyze two semi-mechanical effects, i.e., urbanization and vocational diversification, on income inequality. Gini coefficient and top 1% income share are used as the measurement of income inequality’s degree. I also study the relationship between the number of billionaires and income inequality. Finally, I discuss the effect of Gini coefficient within subpopulations on income inequality of the whole group. Under certain assumptions, Gini coefficient first increases and then decreases with urbanization. Top 1% income share decreases with urbanization. Gini coefficient drops while top 1% income share rises with vacational diversification, if the highest income and lowest income are constant. However, Gini coefficient increases with vocational diversification, if the income of each kind of occupation increases proportionately or arithmetically. The number of billionaires is positively correlated with Gini coefficient and top 1% income share. When taking Gini inequality within subpopulations into account, the overall Gini coefficients from calculation and simulation are more close to the real data. The models start from the simplest condition, and then come to the more practical situations. Numerical simulation is used to test formula. Real data is used to compare the results generated from the models. Needless to say, the models are imperfect and the assumptions are too simple and ideal. They cannot fully explain the rise of income inequality in the real world. However, the simple models do good job on focusing on the single factor and neglecting the interference of other factors, which makes the conclusions very clear and direct. Researches may use our methods but adjust the assumptions according to different cases.