【作者】 李松；

【导师】 苏春红；

【作者基本信息】 山东大学 ， 财政学， 2017， 硕士

【摘要】 随着医疗科技的发展与医疗卫生水平的提高,我国人口的预期寿命不断延长,人口老龄化加剧。为了缓解人口老龄化给我国带来的养老金支付压力,实现养老保险制度的可持续发展,延迟退休政策呼之欲出,其通过增加劳动年龄人口、减少退休人口而成为实现较低制度抚养比的一项可行性措施。然而,延迟退休的前提是将要进入或已经进入老年阶段的人口拥有可供工作的健康资本。那么,延迟退休究竟会对城镇从业人员的健康产生怎样的影响?该种影响又是通过什么渠道产生的?本文基于中国健康与养老追踪调查(CHRALA)数据库2011、2013年城镇职工"健康状况和功能"模块数据,分性别研究了我国退休对城镇职工的包含自评健康(sah)、躯体生活自理量表(psms)、工具性日常生活功能量表(iadl)、慢性病(chronic disease)、抑郁症(CES-D)五种健康指标的影响,并进一步探索了产生相应影响的机制,即是否因为退休后锻炼、社交、睡眠、吸烟、饮酒、时间使用等与健康相关的生活行为发生了变化而促使了该种影响的发生。本文在梳理退休对健康影响文献基础上,运用两阶段最小二乘法,并借助工具变量进行模糊断点回归(FRDD),实证分析了退休对健康的影响及影响机制。在本研究中存在遗漏变量、双向因果等内生性问题,加入控制变量可缓解遗漏变量问题,模糊断点回归方法则能很好解决双向因果这一问题,同时我国城镇职工退休制度恰好满足断点回归的假设条件,因而,本文采用模糊断点回归方法进行了相应回归。研究发现退休促进了我国城镇男性职工的自评健康水平,在10%显著水平上使该健康指标提高了 25%;对女性健康水平的影响则表现在慢性病的得病概率方面,女性退休使女性慢性病得病概率在5%显著水平上降低了 32%。然后,本文进一步探究了退休影响健康的机制。根据Grossman健康需求模型,本文提出"退休后与健康相关的生活方式及时间使用发生改变是退休对健康产生影响的原因之一"这一假设,运用与前文相同的回归方法,将包含锻炼、社交、睡眠、吸烟、饮酒等的健康行为及时间使用作为因变量,进行了进一步回归。研究发现:退休后男性使诸如交友、慈善、公园娱乐等的社交概率显著提高了 38%,且该种生活方式能使男性群体自评健康水平显著提高6个百分点,退休后女性社交概率有所增加,但该生活方式对女性慢性病得病概率却产生负向影响,因此得出退休后社交概率的变化是退休对男性健康产生显著作用的原因。最后,为检验结果的稳健性,依据FRDD方法有效的假设条件,本文在"控制变量在断点处的连续性"、"年龄密度函数的连续性"、"将其他值作为临界值来检验结果的显著性"三方面进行了稳健性检验,结果发现,控制变量及年龄密度函数在我国退休年龄处均是连续的,同时采用其他假设年龄作为断点时相应回归结果也不再显著,因而证明本研究所使用的回归方法是有效的。基于上述研究,本文认为为提高延迟退休的政策效果,提高职工的健康水平,我们需要灵活增加男性职工群体的社交概率,适当鼓励男性提高与朋友交往、去公园等场所进行跳舞等娱乐活动及参加社团的概率;同时,单位可不定期组织慈善活动,多多鼓励男性职工群体进行志愿者服务;并且,单位可主持修建诸如棋牌室、健身室等娱乐性场所,进而提高男性群体的社交娱乐几率;另外,可对男性群体进行课程培训及普及炒股、上网等知识,进而促进男性群体的健康水平。更多还原

【Abstract】 With the development of science and public health,Chinese life expectancy has been prolonged.The aging of population is becoming even worse.In order to pay for the pension in the future and make the old-age insurance system realize the sustainable development,delaying retirement policy will be carried out.It is a feasible measure which can realize the low elder support rate by increasing the working-age population and deceasing the retired population.However,the precondition to delay retirement is the health capital using for work for the persons who are going to be older people or have been older people.So what will be the effect of postponed retirement on urban employed persons’ health?What’s the reasons which drive such effects occurring?In this research,we study the effect of normal retirement on urban employed persons’ health which contain self-reported health(sah).physical self-maintenance scale(psms)、the instrumental activities of daily living(iadl)、chronic disease and CES-D and we also explore whether the change of the health behavior containing smoking、drinking、social contact、exercising、time using is the reason why such effect can be happen by using the data from China Health and Retirement Longitudinal Study(CHRALS)on 2011 and 2013.According to the researches,we do the FRDD by using the instrumental variable and 2SLS to study the effect of retirement on health and the mechanisms of such effect.In the research,.there are two endogenity problems which contain omitted variable bias and reverse causality.Adding control variable can help solve the omitted variable bias.Fuzzy Regression Discontinuity Design can circumvent the reverse causality.So we do the FRDD after teasing the essays which are related to the subject.We find that retirement improve the sah for males by about 25%at the 10%significant level;it reduce the probability of chronic disease for females by 32%at the 5%significant level.Then we study the mechanisms which promote such results.Basing on the healthy demand theory which proposed by Grossman,we hypothesize that the effects of retirement on health are resulted from the change which contain life style and time use what related to health.By using the same method,we set the exercise、social、sleep、smoke、drink et al as the dependent variable to do the same regress.We find that man increase the odds to social including making friends、charity、park entertainment by 38%which promote sah by 6%.Women add the social contact as well,however,it is negative for females’ chronic disease.At the last,we do the robust test.We investigate the robustness of the results with respect to the specification of the control variable、the specification of the age trend and the specification of the assignment variable.The results show that the control variable are continuous following the age trend、age density function curve are successive at the normal retirement age and the regression coefficient are not significant any more which reveal that our method is effective.On account of the results,we think that in order to enhance the effect of delaying retirement and improve the health level of employ,we need to encourage man to do much more recreational activities such as contacting with friends and going to the park.At the same time,organization in which one works can organize charity event.on any time and encourage person to do the volunteer activities.It can also build the entertainment places such as the chess and card room and so on to increase the chance to recreation.In addition to measures which mentioned in the above,we can also train the males on the stock market and net work.更多还原