【作者】 苏鹏；

【导师】 孙巍；

【作者基本信息】 吉林大学 ， 数量经济学， 2014， 博士

【摘要】 自1978年改革开放以来，我国进入一个有史以来最好的发展阶段，经济高速增长带来居民收入水平的快速提高及其分配结构的剧烈演变，而且快速发展过程中出现问题及在问题解决中呈现出的反复和震荡，必然比发达国家来的明显和剧烈。中国这一特定的经济环境决定了其在整个发展过程中的理论创新和政策实践必将遇到许多不同于西方传统理论的新挑战和新课题。于是，从收入分布变迁出发，重新审视我国收入和消费间的作用机制，对寻求激发和释放居民需求的突破口来说，具有重要的理论和现实意义。本文以“经济增长→居民收入分布变迁→消费行为改变→消费市场效应”为思想主线，以居民收入分布变迁为切入点，且把其作为近年来我国居民消费需求演变的主要动力，提出了收入分布变迁消费市场效应的理论预期，该预期的核心内容是在市场机制主导下由居民收入分布变迁引致而来的居民消费需求偏好和档次的转变和局部多个市场轮番出现供不应求和价格虚高等非均衡现象的结果。为分析其合理性，本文从理论和计量两个展开讨论。首先通过对西方经典消费理论的简单回顾和评析，在储蓄动机不同的异质群体假设下分析出了收入和储蓄之间的“马鞍形”关系，此为西方收入分布和消费作用的主要理论观点。但通过对我国城镇居民收入分布变迁和需求的演变规律发现，此规律并不符合我国居民特有的间歇性周期波动消费行为及其“整体需求不足、局部需求旺盛”的需求现状，很难实现西方式的由微观到宏观的过度。因此本文转换思路，从收入分布视角，并结合我国居民消费行为规律，提出收入分布变迁消费市场效应的理论预期是很有理论必要的。要计量检验收入分布变迁消费市场效应的存在性和合理性，前提需完成收入分布变迁的测度及相关计量变量的设计。首先运用非参数核密度估计方法和多种函数形式参数方法相结合的方式，选取8次CHNS数据的大样本调查数据和统计年鉴中2000-2009年城镇居民收入分组数据对收入分布进行了拟合，结果显示我国城镇居民收入水平由低到高向上转移的趋势明显，而这种提高的不同步性又导致了收入差距的扩大。其次采用反事实分析法，分解得到度量收入分布变迁的均值变化、方差变化和残差变化三个动态特征的计量变量，分解结果中均值变化占据主导，方差变化次之，说明在收入分布变迁中经济发展的推动作用为主导，由现阶段的分配制度等导致的不同步次之。在收入分布变迁计量完成基础上，依据消费的经济逻辑理论，在异质性偏好的前提下，证明了收入分布变迁会对会对总消费产生组群下的水平效应、组内规模效应和组间分配效应三方面影响。而后采用CHIP数据中2002年和2007年的城镇数据，借助内生性门限回归的组群划分，通过对上述结果的实证检验表明，水平效应对总消费变化的影响占主导；规模效应则对总消费有抑制作用，源于组群间由高到低的人口回流；组间分配效应在整体上作用很小，表明组间收入差距对总消费的抑制作用没有想象中那么大，但群体间作用差异大，尤其是低收入组群对总消费的抑制作用最大。在上一步研究基础上进一步深入，对消费支出分布分位点上居民的收入分布变迁消费效应进行研究。为了保证微观数据间的可比性，对CHIP数据进行了人工面板改造，借助其门限模型的分位数回归，完成了消费分布分位点上收入分布变迁消费效应的度量，结果表明居民收入方差带来的离散效应在各分位点的消费效应中占据主导，而均值变化带来的水平效应表现乏力。因此，水平效应对绝对支出增长贡献的乏力，直接导致了其在现实中对消费率拉低的主导作用。另外，我国收入分布和消费分布间存在矛盾，其根源由对CHIP数据的统计分析表明，在于现阶段我国多数居民，尤其是中等阶层面临的消费结构困境。至于收入分布变迁对消费结构的影响效应研究，首先构造“反事实收入变量”，对AIDS模型进行动态性扩展，而后基于我国“十五”和“十一五”两期的城镇数据进行了实证检验，结果显示“十五”时期我国城镇居民消费结构整体上由生存型向发展型演变升级，整体收入的提高带来的水平效应是我国消费结构不断升级的主要原因；而“十一五”期间的水平效应不再显著，离散和异质效应占据主导，整体社会需求下滑；但收入差距对个别市场消费有一定积极作用，尤其对耐用品和文化娱乐服务等消费；异质效应作用下高收入阶层已有显著地的服务类消费需求，消费结构存在进一步升级的可能。上述分析说明水平效应决定整体需求，而离散和异质效应影响单个市场需求。最后进一步以家用车市场为例考察了收入分布变迁对单一市场的影响。基于按收入分组的城镇居民家用车拥有量等数据，采用基于消费函数回归的分解方法，度量了收入分布变迁对家用车市场的影响效应。结果显示收入分布变迁的水平效应几乎均起到了决定性的正向拉动作用，但其效果并未得到充分释放，原因在于离散效应对低收入者的抑制作用，但离散效应对中等以上的高收入人群的需求具有很强的正向促进作用，其现阶段在全国整体来看仍有积极作用；异质效应则由于数据原因略显微弱。更多还原

【Abstract】 Since the reform and opening up in1978, China has entered a period of rapid development,and the rapid economic growth has brought a rapid increase in the level of income and itsdistribution also evolution drastically in a short time, so problems encountered and solved laterappeared again and again must come more obvious and severe than in developed countries. Chinathis particular economic environment determines the theoretical innovation and policy practicesthroughout the development process is bound to encounter many new challenges and newtheoretical issues which is different from the Western. As a result, from the perspective of theincome distribution evolution, studying the mechanism of income and consumption and seeking abreakthrough of releasing residents’ demand has an important theoretical and practical significance.In view of the above considerations, this paper follows the research mentality of “EconomicGrowth→Income Distribution Evolution→Changes of Consumer Behavior→Evolution ofthe Level and Structure of Consumer Demand”, puts changes of the income distribution as thestarting point and as the driving force of consumer demand evolution, and then we propose thetheoretical expectation that consumer market effects of income distribution evolution, on which wecarry out the micro-econometric study deeply.Firstly give a simple review and assessment of the Western classical theory, and then under theassumption of heterogeneous groups that have different motives in savings, we get the "saddle"relationship between income and savings, which can be used as the Western theoretical basis ofeffect analysis between income distribution and consumption. But through statistical analysis ofstatus that our urban residents’ income distribution evolution and demand change, we find that thisrule does not meet our residents unique intermittent cyclical fluctuations of consumer behavior andour status quo demand of "insufficient overall demand, strong local demand", so it is difficult torealize the transition from the micro to the macro by the way as Western. As a result, we convertideas to the income distribution, and propose the theoretical expectation of consumer market effectsof income distribution evolution finally.In order to characterize income distribution evolution and design income quantitativevariables that follow-up study required. Firstly, base on the large sample survey data of eight year’sCHNS data and2000-2009statistical yearbook urban group data, through a method combination of non-parametric kernel density estimation and several forms of parameter-function, we get theconclusions that the trend of China’s urban residents income transfer from low to high upsignificantly, but this increase is not synchronized, which has led to a widening income gap.Secondly, we obtain the mean change, variance change and residual change of income distributionevolution by counterfactual analysis, where the mean change is dominant, followed by the variancechange, last is residual change, by which we know that the promoting role of the economicdevelopment is the most important and the second is the distribution system.Then from the perspective of income distribution changes, according to economic logic ofconsumption theory, and under the premise of heterogeneity, we establishe a mathematical model tostudy how the evolution of income distribution impacts on the total consumption, whose resultshow that income distribution changes would generate the level effect, scale effect in the group anddistributional effect between the groups. Adopted in2002and2007CHIP urban data, based onthreshold regression, empirical results show that level effect dominates on the impact of changes intotal consumption, scale also reflects a certain inhibitory effect for the high to low populationreflux, and the overall effect of distributional effects between the groups is very small, so impactionof the income gap on total consumption is not so big as imagined, but effects between groups havea big difference, especially low-income groups take the greatest inhibitory effect on totalconsumption.Based on the research in the previous chapter, takes a further study on the demand effects ofincome evolution on the expenditure quantiles. In order to ensure comparability of data, weconstruct a quasi-panel of CHIP data, and by quantile regression of its threshold model we finshiedthe estimate. The main conclusion is that scatter effect caused by income variance bring dominanteffect on expenditure quantiles, while the level effect play poor performance. Showed that leveleffect contributed weakly to spending, a direct result is that its leading role of consumption ratereduce; while there are significant inconsistencies between the distribution of the income andexpenditure, its roots are likely to be the consumption dilemma of durable goods that majority ofresidents faced.As for the study about effects of income distribution evolution on the consumption structure.By constructing two "counterfactual income variable", we construct a new model that dynamicexpansion of AIDS. Then based on China urban data of "Tenth Five-Year" and "EleventhFive-Year", we finish the empirical test and results show that in the Tenth Five-Year consumptionstructure of urban resident overally upgrade from subsistence to developmental level, and the mainreason is the level effect brought by the improvement of residents’ average revenue. While during the Eleventh Five-Year period, the horizontal effect is not significant any more, scatter andheterogeneity effects take the dominant role, which cause decline of the whole society demand. Butincome inequality still has some positive effects on individual consumer market, especially for themarket of durable goods and consumer services, culture and entertainment. The high-income grouphas a significant class of service consumer demand, showing that there will be a further escalationof our consumption structure. So level effect major impacts the overall demand, while variance andresidual effects are more inclined to affect the specific market demand.In the end, we take further study on the effect of income distribution change on the singlemarket with family car market as an example. Income distribution evolution generate three effects,including level effect, scatter effect and heterogeneity effect on consumer demand for family carsand other durable goods. Empirical analysis found that, level effect almost always play a major rolein promoting car’s consumption; scatter effect has a greater inhibition on the families ofmiddle-class and less, but has a positive impact on the higher income groups, especially for thehighest income group, whose scatter effect is greater than the level effect, so it has a smallerpositive effect on the whole; as for heterogeneous effect, it’s very weak. Based on revenue trends,there is a possibility the automotive market shrinking, so income distribution reform and industrialplanning process needs to be more cautious.更多还原