EN中文

GEOGRAPHICAL RESEARCH >

2018 , Vol. 37 >Issue 12: 2567 - 2575

DOI: https://doi.org/10.11821/dlyj201812016

Articles

Multi-dimensional analysis of housing segregation:A case study of Shenzhen, China

  • ZHANG Yu ,
  • TONG De ,
  • Ian MacLACHLAN
Expand
  • School of Urban Planning and Design, Peking University, Shenzhen 518055, Guangdong, China

Received date: 2018-05-31

  Request revised date: 2018-09-05

  Online published: 2018-12-20

Copyright

《地理研究》编辑部 所有
Fold

Abstract

Residential segregation has been a severe and widespread phenomenon in mega cities along with fast urbanization in China. Migrants from rural area flock into developed cities especially coastal regions for better job opportunities, which provide essential cheap labor for urban growth. However, their housing problems could not be resolved in formal housing either hindered by institutional barrier or unreachable housing price. The housing segregation gradually formed as locals reside in formal gated communities while migrants crowd in informal housing like urban villages, which is characterized with lower rent but substandard living conditions. The housing segregation in China derives from household registration system (hukou). The Index of Dissimilarity (ID) only emphasizes the unevenness of population distribution but could not fully manifest the segregation characteristics in density, location, proximity, etc. Inspired by the work of Massey Denton in multi-dimensional segregation, this article applies three measures of housing segregation (Clustering, Centralization, and Concentration) based on the ID to analyze the segregation between urban residents with and without hukou. It examines the multi-dimensional housing segregation based on hukou status using data from China’s 6th national census in 2010. The typical migrant city Shenzhen was chosen to conduct the case study, and the segregation index of three dimensions was calculated based on 55 sub-districts for comparison. The multi-dimensional segregation indexes showed that Shenzhen has high segregation problems at the city scale, but more homogeneous inside each district. The history, industrial structure and socioeconomic background of each district play a crucial role in the segregation. The outside-custom area provides more chances in labor-dense sectors and attracts more migrants to reside in a large scale, while the inside-custom regions are more advanced in informatics and financial sectors, which results in scattered spots of migrants housing. Cluster analysis reveals the three types of segregation, each of which has its unique processual mechanisms, and policy prescriptions. The study shows that the housing segregation has multiple dimensions and scales. Thus two sets of people could be featured by a single ID yet to be clustered or dispersed, central or peripheral, or concentrated or deconcentrated. Migrants may occupy continuous neighboring blocks in peripheral area, or densely reside in few scattered urban villages in inner city, or congregate in factory dorms alongside each industrial zone. Based on segregation patterns, locations and density, local governments should take different measures like redevelopment of targeted urban villages, large-scale public housing construction or cooperation with factories in worker dormitory improvement accordingly. This article contributes an innovative and comprehensive perspective to conceptualize housing segregation, and provides policy recommendations to deal with the social problems that arise from segregation in China. With the advancement of big data, more practical real-time housing management measures could be developed for practitioners to provide human-centric housing planning and avoid the housing polarization.

Key words： housing segregation; Clustering Index; Centralization Index; Concentration Index; Shenzhen

Cite this article

ZHANG Yu , TONG De , Ian MacLACHLAN . Multi-dimensional analysis of housing segregation:A case study of Shenzhen, China[J]. GEOGRAPHICAL RESEARCH, 2018 , 37(12) : 2567 -2575 . DOI: 10.11821/dlyj201812016

1 引言

中国快速城市化背景下,大量农村剩余劳动力涌入城市,特别是东部沿海经济较为发达的大城市,他们为发达城市输送了大量廉价劳动力,提供了推动城市迅速发展的重要生产要素,但同时也在一定程度上引发了一系列“城市病”,如城市居住空间日益分化、空间异质性增强,“分异”和“碎化”成为普遍趋势[1]。
以深圳为例,截至2015年末,深圳市非户籍人口达到782.88万人,占常住人口的68.8%（图1）。虽然非户籍人群内部具有异质性,既包含外来务工者,也包括收入和知识水平较高的人,但群体统计数据显示,非户籍人口具有明显的“三低”特征,即低教育水平、低收入、低稳定性[2]。他们中大部分人的居住需求难以通过正规住房市场解决,棚户区、城中村等非正规住房则以价格、区位等优势成为外来人口聚居区[3]。逐渐地,非户籍与户籍人口形成了日益明显的居住空间分异格局[4]。
Fig. 1 Demographic composition of Shenzhen based on residency status (1979-2015)

图1 1979-2015年深圳市按户籍状况人口构成

针对中国大城市居住空间分异问题的研究已取得不少有价值成果。与国外以种族问题为核心的研究[5,6,7,8,9]不同,国内相关研究重点关注户籍因素作用下的居住空间分异特征、格局、空间效应及形成机制[10,11]。诸多学者运用多元回归等方法,从政策、历史、经济要素等方面入手[12,13,14],证明了中国的居住空间分异与中国住房制度历史及改革进程、城乡二元户籍及土地制度、市场化及全球化背景下经济形态演化分异等因素有关[15,16,17,18,19]。而在不同城市,由于经济发展水平与产业布局模式不同,居住空间分异格局存在显著差异[20];同时,在同一城市内部,社区、街道和区等不同尺度上空间分异状况也存在差异[21,22]。
在居住空间分异程度方面,国内外学者从不同角度给出度量公式,如相异指数（Index of Dissimilarity）、隔离指数（Isolation Index）、交互指数（Interaction Index）等[23],这些指数普遍借鉴基尼系数的算法,重在揭示特定人群空间分布不均衡的态势。一般来说,相异指数<0.3表示人口分布比较均匀,0.3~0.6表示具有一定的居住分异态势,>0.6则表示存在比较严重的居住分异问题,政府需干预调节[24]。然而,现有各类居住分异指数的算法重在强调人口分布比例的不均匀程度,对在同样的人口分布比例下,由于人口聚居形态、居住区位和居住质量等方面差异导致的更进一步的分异状况及其空间效应,并未得到足够重视,在一定程度上降低了相异指数对空间管治政策制定的指导性,因此,有必要进一步丰富和改善居住分异指数的算法,以挖掘居住分异更深层次、更多元的内涵。本文以典型移民城市深圳为例,基于2010年全国第六次人口普查数据（简称六普数据）,以居住空间相异指数为基础,进一步构建集聚—分散度、中心—边缘度和极化—均质度指数,深度解析户籍和非户籍人口居住分异的多维特征、成因及管治方向,为研究快速城市化背景下大城市普遍出现的居住空间分异提供新鲜视角,同时为解决其带来的社会及空间治理困境提供更有针对性的决策依据。

2 居住空间分异的多维度内涵及其测算方法

2.1 相异指数

目前,国内外学术界在衡量居住空间分异时,普遍采用相异指数（Index of Dissimilarity）[25],公式如下：
ID = 1 2 × ∑ i = 1 n x i X - y i Y （1）
式中：xi代表第i个空间单元的某群体人口;X代表全域该群体人口;yi代表第i个空间单元的另一类群体人口;Y代表全域另一类群体人口。相异指数的阈值范围为0~1,代表了分异程度由最均匀到最分异。图2表示了全域非户籍人口比例相同的情况下,图2a的相异指数小于图2b（网格代表街道单元;网格颜色代表了本街道的非户籍人口比例,越深表示非户籍人口比例越大）。
Fig. 2 Diagramatic illustration of multi-dimensional housing segregation

图2 多维度居住空间分异内涵示意图

这一指数被用于研究基于户籍的居住空间分异时,仅考虑了各居住单元非户籍人口比例与全域的对比。然而,在相同比例下,非户籍人口与户籍人口居住地规模、区位及居住密度等方面的差异将会导致更进一步的分异格局（如图2c~图2e）及空间影响,产生不同的社会空间治理问题和政策需求[26]。基于此,本文进一步建立集聚—分散度、中心—边缘度和极化—均质度指数,对居住空间分异进行多维度解析,更深入、细致剖析分异特征及形成机制。

2.2 集聚—分散度

集聚—分散度指数用来测量一个区域中非户籍人口集聚（图2b）或分散布局（图2c）的程度,其公式为：
SP = ( X P xx + Y P yy ) T P tt （2）
P mm = ∑ i = 1 n ∑ j = 1 n m i m j c ij M 2 （3）
式中：cij代表空间单元i与j几何中心距离的负指数函数,即 c ij = e - d ij ;X、Y、T分别代表全域非户籍、户籍和总人口数;Pxx、Pyy与Ptt分别代表非户籍人口之间、户籍人口之间和全域总人口之间的空间邻近程度,计算公式分别使式（3）中的 m = x , y , t 。该指数以1为界,小于1的程度越大表明非户籍人口分布越分散,而大于1的程度越大表明非户籍人口分布越集聚[27]。集聚—分散度指数越高,意味着城市中成片的、甚至跨统计单元的贫民窟、流动人口聚居区等规模越大,此时,降低人口居住空间分异的政策手段除了针对外来人口的社会管理政策外,更应加强对重点集聚区域,如老旧城区和城乡结合部连片粗放开发区域的空间改造和管理。

2.3 中心—边缘度

中心—边缘度指数用于测量一个区域中非户籍人口相对分布在城市中心（图2b）或外围的程度（图2d）。其公式为：
ACE = ∑ i = 1 n X i - 1 A i - ∑ i = 1 n X i A i - 1 （5）
式中：Xi代表在全部的非户籍人口中居住在第i个空间单元的比例;Ai代表前i个空间单元的累计面积占全域面积的比例;i从1至n依次表示在城市中的区位由中心到边缘,该指数阈值范围是[-1, 1]。在相同的居住空间相异指数下,正的中心—边缘度指数一般发生在高度城市化或后工业化区域,值越大意味着非户籍人口居住更加趋向城市中心,此时的空间治理政策一方面应通过在郊区兴建保障性住房和增加低收入就业岗位,疏导外来人口向郊区布局;另一方面,应加强城市中心区域旧城改造及环境、交通等品质提升。而负的中心-边缘度指数一般发生在工业化或城市扩张初期的城乡结合部,空间治理重点则应是加强郊区基础设施和公共服务设施建设,提升城乡结合部城市化质量。

2.4 极化—均质度

极化—均质度指数用于测量一个区域中非户籍人口居住空间质量内部分化的程度（图2b、图2e）,各子区域间人口密度差异大则意味着居住空间质量分化程度大,公式为：
DEL = 1 2 ∑ i = 1 n x i X - a i A （6）
式中： a i 代表第i个空间单元的面积;A代表整个区域的总面积,其他参数与前文定义相同。该指数范围为0~1,越接近1则表明非户籍人口间居住密度差异越大,该类人群内部可能已出现进一步分异趋势;而指数趋向于0则代表该区非户籍人口居住密度相对均质,居住空间质量差异不大。在极化—均质度指数高的地区,空间治理应进一步剖析导致非户籍人口内部二次分异的原因,分层制定适用于不同区域、不同层次外来人口的空间治理政策,尤其是识别出高度拥挤的非户籍人口集聚区,通过定向提供廉租房或社会福利保障、棚户区改造等措施尽快疏散人口、降低安全隐患。

3 深圳居住空间分异实证分析

3.1 研究区概况及数据来源

本文采用六普数据作为计算多维居住空间分异指数的数据源,当时,深圳市下辖宝安、福田、光明、龙岗、罗湖、坪山、南山、盐田共八个行政区（新区）,包含55个街道（图3）。首先分别基于各区及街道人口数据,计算全市及分区相异指数,再进一步从集聚—分散、中心—边缘、极化—均质度三个维度计算并比较各区居住分异特征及成因的差异。
Fig. 3 Map of administrative districts and sub-districts in Shenzhen

图3 研究区行政区划及街道分布示意

由于分维度指数需用到街道几何中心、街道间距离、街道面积等空间数据,以深圳市行政区划矢量数据作为空间数据基础,利用ArcGIS平台开展相关计算。在计算中心—边缘度时,由于现实中的城市空间布局难以实现理想的单中心模式,城市空间区位优劣难以体现出严格的由中心向外围衰减趋势,故采用深圳市政府2013年发布的城市基准地价作为区位优劣的评价标准。

3.2 多维度居住空间分异结果

如表1所示,本文分别以区和街道数据为基础,以全市和区为统计单元,分别计算了全市及各区的相异指数、集聚—分散度、中心—边缘度和极化—均质度指数,发现深圳的居住空间分异程度在不同维度表现出不同特征。
Tab. 1 Statistics of multi-dimensional housing segregation indexes of Shenzhen

表1 深圳市分维度居住空间分异统计指标

排序 相异指数 多维度指数
集聚—分散度 中心—边缘度 极化—均质度
深圳市 0.684 深圳市 1.058 深圳市 0.693 深圳市 0.250
1 光明新区 0.352 盐田区 1.204 宝安区 0.797 龙岗区 0.406
2 宝安区 0.317 光明新区 1.187 罗湖区 0.755 罗湖区 0.386
3 南山区 0.264 宝安区 1.064 盐田区 0.750 盐田区 0.329
4 福田区 0.257 南山区 1.046 龙岗区 0.699 光明新区 0.245
5 龙岗区 0.231 福田区 1.020 坪山新区 0.693 福田区 0.183
6 盐田区 0.203 龙岗区 1.017 福田区 0.472 南山区 0.154
7 罗湖区 0.187 罗湖区 1.007 南山区 0.472 宝安区 0.135
8 坪山新区 0.040 坪山新区 1.000 光明新区 0.357 坪山新区 0.089

注：坪山新区与光明新区分别仅由两个街道构成,各维度分异指数计算结果可比性相对较差。

3.2.1 相异指数及其空间尺度差异 根据表1,深圳全市相异指数为0.684,按照国际经验,深圳已出现了较严重的非户籍人口与户籍人口空间分异现象。但从以区为单元的计算结果来看,相异指数基本在0.3以下,分异程度不大。市区两级计算结果的明显差异说明,深圳区与区之间非户籍人口相对规模差异较大,而各区内部差异较小。市级尺度的分异主要来自于深圳市原特区内外经济发展水平与产业结构的差异。宝安、龙岗等“关外”地区产业构成以劳动密集型制造业为主,制造业在岗职工人数较多（图4）,其主要构成为非户籍人口;而“关内”各区金融、信息产业等服务业较发达,劳动力素质高,户籍人口所占比例大。
Fig. 4 Number of in-service staff of each district by industry sector (2010)

图4 2010年深圳市各区分行业在岗职工人数

3.2.2 多维度居住空间分异格局及成因 与相异指数不同,集聚—分散度、中心—边缘度、极化—均质度三维指数并未出现深圳市区间的尺度差异。各区集聚—分散度指数近似为1、中心—边缘度指数均为正的结果表明,各区虽未出现大规模连片的非户籍人口集聚区,但其分布占据较优区位;而极化—均质度指数在0~0.5间分散分布,表明各区内非户籍人口内部居住质量存在不同程度的差异（表1）。
城中村是深圳非户籍人口分布的主要区域,各区居住空间分异特征与城中村的空间分布具有高度关联。根据2007年深圳市城中村人口调查数据,全市320个城中村内居住着637万人口,占全市人口一半以上,其中非户籍人口达到595万,是户籍人口的14.2倍,城中村已经成为流动人口聚居区的代名词[4]。而城中村形成于城市建设用地向郊区扩张过程,在中国大多数传统城市中,城中村通常位于城市中心区外围或城乡结合部[28,29]。然而,改革开放以前,深圳是以传统农业、渔业为主的边陲小镇,建成区范围十分狭小,整个辖区呈现出农村居民点分散布局的态势。正是由于深圳缺乏旧城基底,其城市空间扩张途径并非如其他城市般由老城区向郊区蔓延,而是以罗湖、蛇口、沙头角等毗邻香港的口岸为中心,选择临近农村居民点周围连片、廉价的农田开展城市建设,并不断发展壮大。因此,深圳城市发展自始几乎就是城市包围农村的态势,由原农村居民点发展成的城中村自然很少位于城市边缘,而是分散布局于城市建成区内。而由于城中村周围几乎被高度建成区包围,难以出现其他城市城乡结合部外来人口聚居区大面积无序蔓延的情况,因此深圳并未出现成片、甚至跨街道的非户籍人口聚居区。
在极化—均质度指数方面,全市和大部分区指标偏低,说明基于户籍的人口居住分异主要体现在非户籍与户籍人口之间,非户籍人口内部基于人口密度差异的进一步分异并不明显。但是,龙岗、罗湖等区该指标已经达到0.4左右,表明该地区内部各街道间居住质量存在较大差异,非户籍人口内部的分层治理需求较大。龙岗地处深圳郊区,是2012年世界大学生运动会场馆所在地,“大运新城”等局部地区的高质量城市建设导致区内建成环境、居住质量差异较大;而罗湖区是深圳的老城区,局部地区的城市更新是导致区内非户籍人口内部居住质量差异明显的主要原因。由此可见,城市空间治理应避免局部建设带来的两极分化,特别需防止高质量居住空间营造过程迫使原本居住在本地的低收入外来人口迁居至附近的外来人口聚居区,寻求居住替代,加剧高质量新区周围外来人口聚居区居住品质的恶化。

3.3 基于聚类分析的居住空间分异治理建议

如前所述,居住空间分异的多维内涵可揭示出不同的社会经济空间现象,其成因和空间治理手段也不尽相同。为了更深入理解深圳各辖区非户籍与户籍人口居住空间分异的特点,有的放矢地制定差异化治理政策,本文综合各区在三个维度上的指标差异,进一步开展聚类分析,将深圳非户籍与户籍人口居住空间分异类型划分为三类（图5）。
Fig. 5 Tree diagram of cluster analysis results

图5 聚类分析树状图

首先,龙岗、罗湖与盐田区为第一类,非户籍人口居住空间具有中心分布和内部差异较大的特征,其主要成因是局部地区的城市更新导致区内建成环境差异较大,低收入外来人口被迫向附近未被更新的区域迁移,尤其是一些区位条件好、交通可达性强的城中村。该地区的空间治理应加强对城市更新项目社会效应的评估,通过集中兴建保障性住房等方式解决城市更新后原有租客的居住问题,或通过局部地区产业结构调整降低区内外来人口就业比例。同时,对外来人口密度较大、居住质量较差的定点区域,政府应加强财政投入,提升基础设施和配套服务供给水平。
其次,福田、南山与光明新区为第二类,非户籍人口呈现出集聚和边缘分布的特征,即主要集中在区内成片的工业区附近居住,包括大量企业员工宿舍。因此,该类外来人口居住空间的治理重点是通过与企业或工业园区合作,统筹外来人口居住管理,兴建或提升宿舍区居住质量。同时,在通过城市更新等方式调整产业和人口结构时,建议以企业为单元统筹考虑员工迁居事宜,加强外来人口的组织化管理。
第三,宝安与坪山区的各项指标都比较平均,可划为第三类。但值得注意的是,这两个区是深圳非户籍人口比例最大的区,非户籍人口占总人口的比例均超过88%。该类地区对居住空间分异的治理更宜采取全区普适性政策,如借鉴国内外经验在普通商品房小区配建保障性住房、加强城中村综合整治提升外来人口居住质量等。同时,也要吸取罗湖等区经验,防止城市更新等局部建设加剧区内空间极化。

4 结论与讨论

居住空间分异是一个具有多尺度、多维度特征的复杂现象,单一的相异指数、隔离指数等无法准确刻画在相同的人口分布比例下,由于人口聚居形态、居住区位和居住质量等方面差异导致的更进一步的分异状况及其空间效应。因此,本文构建了集聚—分散度、中心—边缘度和极化—均质度三个分维指数,进一步挖掘居住空间分异的多维内涵,及其所揭示出的不同的社会经济空间现象、成因及空间治理重点。进而运用六普数据对深圳开展实证研究,在计算了全市及各区分维指数的基础上,分析了深圳人口居住空间相异指数特征及空间尺度差异,多维居住空间分异格局特征及成因,并通过聚类分析将深圳非户籍与户籍人口居住空间分异类型划分为三类,分类提出空间治理政策建议。
本文为更加深入细致地理解居住空间分异内涵及其成因、背后所反映的社会问题和治理手段提供了更加丰富的视角,可作为进一步开展相关研究的有效切入点。特别是在大数据时代,更小尺度及基于个体空间行为的多源数据应逐渐补充或替代基于行政区域的空间统计数据,这将更有利于揭示更深层次的分异特征和成因,从而提出操作性、时效性更强的空间治理决策建议。

The authors have declared that no competing interests exist.

References

Publishing order | Descend order by publishing year | Descend order by cited within
[1]
李志刚, 吴缚龙, 肖扬. 基于全国第六次人口普查数据的广州新移民居住分异研究. 地理研究, 2014, 33(11): 2056-2068.运用第六次全国人口普查数据,分析当前广州新移民的居住空间分异问题,探讨中国大城市新移民居住空间的整体格局、分异程度及其分异机制,并以此实证中国城市社会空间理论。研究表明:广州新移民的空间分布总体上表现出近郊集中和远郊分散的特征,其中省内新移民较多集中在中心区外围,省外新移民集中分散在远郊。广州新移民与本地常住人口的差异指数为0.48,分异指数为0.46,隔离指数为0.56,超过美国亚裔移民分异的平均水平。此外,各区的分异度水平存在较大差异。机制分析表明,历经30多年的市场化进程,制度因素对新移民聚居区的影响减弱,单位因素对新移民聚居无明显影响,市场因素的作用增强。与西方情况类似,人口和家庭因素对新移民聚居有一定影响,其中年龄因素和婚姻状况是影响新移民聚居的重要因素。总体上,市场化下中国城市新移民的社会隔离正不断转化为明显的空间分异问题。为推进社会融合,应尽快采取社会空间重构措施(如社会规划和社区规划)予以应对。

DOI

[Li Zhigang, Wu Fulong, Xiao Yang.Residential segregation of new migrants in Guangzhou, China: A study of the 6th census. Geographical Research, 2014, 33(11): 2056-2068.]

[2]
林李月, 朱宇, 梁鹏飞, 等. 基于六普数据的中国流动人口住房状况的空间格局. 地理研究, 2014, 33(5): 888-898.住房是流动人口融入城市、实现市民化过程中必须解决的关键问题。基于2010年第六次人口普查数据,采用住房拥有率、租住房率、住房面积指数、住房不受干扰指数、住房质量指数和住房费用指数6个指标考察流动人口的住房状况,并综合运用数理统计、空间自相关和系统聚类法揭示流动人口住房状况的属性特征、空间分布与集聚类型。研究发现,与城镇常住人口相比,流动人口的住房状况较差。从空间分布看,流动人口住房状况的各项指标具有显著的空间正相关,在空间分布上不仅存在集聚现象,而且有明显的集聚中心。研究结果还表明,流动人口住房条件综合状况可划分为较好、中等、中等偏下、较差4级类型区,在全国尺度上的空间分布除个别类型外具有团块聚合的结构特征。在考虑社会公平的前提下,应分类解决不同类型区域流动人口的住房问题。

DOI

[Lin Liyue, Zhu Yu, Liang Pengfei, et al.The spatial patterns of housing conditions of the floating population in China based on the sixth census data. Geographical Research, 2014, 33(5): 888-898.]

[3]
冯长春, 李天娇, 曹广忠, 等. 家庭式迁移的流动人口住房状况. 地理研究, 2017, 36(4): 633-646.近年来"家庭式迁移"日益成为流动人口迁移的主要趋势,并对城市居住的独立性、权属和质量提出了现实需求。运用2009年环渤海、长三角、珠三角、成渝四区域12市的2394份抽样调查问卷,采用Logistic回归分析等计量方法,探究家庭式迁移的流动人口住房特征及影响因素。研究发现,"独住型""夫妻同住型""两代同住型"和"三代同住型"的流动家庭住房特征存在显著差异,其购房比例和住房质量依次提高。流动家庭的住房权属和质量受到家庭社会经济特征、家庭类型、地理因素以及流动家庭与老家联系和在流入地融入程度的影响。研究发现,如果纳入城市归属感、留城意愿及与老家的联系等变量,将会显著弱化户口对住房的作用。由于不同类型家庭所处的社会经济状况和应对策略不同,住房特征产生了家庭分异。因此,政府应当制定梯度化的住房管理政策,以此推动流动家庭逐步实现"固化"到城市。

DOI

[Feng Changchun, Li Tianjiao, Cao Guangzhong, et al.Housing outcomes of family migrants at the place of destination. Geographical Research, 2017, 36(4): 633-646.]

[4]
仝德, 冯长春, 邓金杰. 城中村空间形态的演化特征及原因: 以深圳特区为例. 地理研究, 2011, 30(3): 437-446.

[Tong De, Feng Changchun, Deng Jinjie.Spatial evolution and cause analysis of urban villages: A case study of Shenzhen Special Economic Zone. Geographical Research, 2011, 30(3): 437-446.]

[5]
Rugh J S, Massey D S.Racial segregation and the American Foreclosure Crisis. American Sociological Review, 2010, 75(5): 629-651.The rise in subprime lending and the ensuing wave of foreclosures was partly a result of market forces that have been well-identified in the literature, but it was also a highly racialized process. We argue that residential segregation created a unique niche of minority clients who were differentially marketed risky subprime loans that were in great demand for use in mortgage-backed securities that could be sold on secondary markets. We test this argument by regressing foreclosure actions in the top 100 U.S. metropolitan areas on measures of black, Hispanic, and Asian segregation while controlling for a variety of housing market conditions, including average creditworthiness, the extent of coverage under the Community Reinvestment Act, the degree of zoning regulation, and the overall rate of subprime lending. We find that black residential dissimilarity and spatial isolation are powerful predictors of foreclosures across U.S. metropolitan areas. To isolate subprime lending as the causal mechanism through which segregation influences foreclosures, we estimate a two-stage least squares model that confirms the causal effect of black segregation on the number and rate of foreclosures across metropolitan areas. We thus conclude that segregation was an important contributing cause of the foreclosure crisis, along with overbuilding, risky lending practices, lax regulation, and the bursting of the housing price bubble.

DOI PMID

[6]
Schnare A B.The Persistence of Racial Segregation in Housing. Washington D. C.: The Urban Institute, 1978: 1-30.

[7]
Johnston R J.哲学与人文地理. 北京: 商务印书馆, 2000: 11-51.

[Johnston R J.Philosophy and Human Geography. Beijing: The Commercial Press, 2000: 11-51.]

[8]
Hutchinson P M.The Effects of Accessibility and Segregation on the Employment of the Urban Poor. Lexington: Lexington Books, 1974: 77-96.

[9]
Streitweiser M L, Goodman J L.A survey of recent research on race and residential location. Population Research and Policy Review, 1983, 2(3): 253-283.This article reviews the post-1975 theoretical and empirical research on race and residential location in metropolitan areas of the United States. We interrelate the main themes of recent research, focusing on the causes and consequences of racial residential segregation. Racial prejudice and discrimination, black suburbanization, school segregation, labor market discrimination, and city/surburban environmental differentials are among the issues examined.

DOI

[10]
黄友琴, 易成栋. 户口、迁移与居住分异: 以武汉为例的实证研究. 城市发展研究, 2009, 16(6): 36-46.计划经济时期的中国城市居住分异程度很低。转型期间快速的城市化和住房的市场化导致了大量人口的迁移迁居,使原来相对均质的单位社区转变为多元的、异质的城市空间,各种新型社区如破旧的移民社区和富有的门禁社区也都应运而生。城市的居住分异日益明显,居住区位也日渐成为社会经济地位的标志。根据武汉市2000年人口普查0.1%按户抽样数据,将市区人口分为市内未迁居居民、市内迁居居民、市外永久移民和市外暂时移民,从户口、迁移迁居和居住的关系分析不同人群在城市中的居住区位及分异状况。与西方城市相比,武汉市区目前的居住分异和隔离程度不高,但各分区的居住隔离程度存在明显差异,特定人群在城市某些区位的分异隔离程度已达到相当高的程度,各类人群也存在一定程度的孤立性。政府相关部门应对此予以重视。

DOI

[Huang Youqin, Yi Chengdong.The urban contact of Changsha-Zhuzhou-Xiangtan urban agglomeration: Based on the urban flow. Urban Development Studies, 2009, 16(6): 36-46.]

[11]
钟奕纯, 冯健. 城市迁移人口居住空间分异: 对深圳市的实证研究. 地理科学进展, 2017, 36(1): 125-135.基于深圳市第六次人口普查数据,将迁移人口按照户籍地划分为市内迁移、省内迁移和省际迁移3种类型.通过计算区位熵分析街道迁移人口比重在全市中的水平,并用空间自相关来识别其空间集聚状况,以揭示迁移人口的空间分异格局,进而用分异指数刻画迁移人口的分异程度.然后以街道迁移人口比重为因变量,住房因素和就业因素为自变量,采用OLS模型、空间滞后模型和空间误差模型来分析和解释迁移人口居住空间分异的影响因素.结果表明:①迁移人口区位熵呈现出较为明显的圈层结构分布特征,以南山区和福田区为中心向外依次为市内、省内和省外迁移人口;②迁移人口的居住分布存在空间集聚,市内迁移人口集聚分布在行政中心周围,省内迁移人口集中分布在商业中心周围,省外迁移人口集中分布在工业园区较多的关外街道;③省外迁移人口与本地人口之间的居住分异程度最高;④省外迁移人口的居住空间分布显著地受住房因素的影响,省内迁移人口的居住空间分布则受就业因素的影响,住房和就业因素对市内迁移人口的居住空间分布的影响不显著.

DOI

[Zhong Yichun, Feng Jian.Residential spatial differentiation of migrant population within the city: A case study of Shenzhen. Progress in Geography, 2017, 36(1): 125-135.]

[12]
Bond Huie S A, Frisbie W P. The components of density and the dimensions of residential segregation. Population Research and Policy Review, 2000, 19(6): 505-524.The purposes of this research are to examine the relationships between density and residential segregation and to propose a technique for the more precise measurement of social density. Using data from the 1990 US Census for the fifty eight largest metropolitan areas in the United States, we explore the applicability of measuring social density by examining how the dimensions of segregation are related to the components of race-specific and non-racespecific density. Findings suggest that density is an important part of our understanding of the processes involved in the segregation of race/ethnic groups and further that the measurement of social density can make a significant contribution to research on the concentration of poverty, joblessness, and violence.

DOI

[13]
Chih Hoong Sin.Segregation and marginalisation within public housing: The disadvantaged in Bedok New Town, Singapore. Housing Studies, 2002, 17(2): 267-288.This paper examines segregation within public housing in Bedok New Town, Singapore. The highly structured and regulated public housing sector, accommodating 86 per cent of the total Singapore population, provides an interesting look at the issue of 'choice' and 'constraint', and their implications for segregation. Using the index of dissimilarity to measure evenness of distribution and the P* index to measure social interaction and isolation, the data show that lower-income members of Indian ethnic background had become more segregated between 1980 and 1990. The eligibility criteria and allocation procedures pertaining to public housing help channel certain groups of residents into a narrow array of housing types in strictly defined locations. Particular socio-demographic features of lower-income Indians, coupled with their numerical inferiority, leads to a weak position within the housing market. The issue of constrained choice is especially relevant for this group of public housing residents.

DOI

[14]
Houston D.Changing ethnic segregation and housing disadvantage in Dundee. Scottish Geographical Journal, 2010, 126(4): 285-298.Dundee has a small black and minority ethnic (BME) population, which has been neglected by previous research, as have BME populations in small towns and cities generally. As in other British cities, the residential locations of the main BME groups are distinct from that of the white population. After briefly reviewing the history of settlement in Dundee, this paper examines the extent to which patterns of ethnic segregation have changed between 1991 and 2001. Some moves towards dispersal and suburbanisation are identified but there are important contrasts between different BME groups. The implications of segregation for housing availability are assessed through Census of Population data. The hypothesis is posed that the consequences of segregation for housing disadvantage are greater in small cities such as Dundee.

DOI

[15]
李志刚, 吴缚龙. 转型期上海社会空间分异研究. 地理学报, 2006, 61(2): 199-211.

[Li Zhigang, Wu Fulong.Sociospatial differentiation in transitional Shanghai. Acta Geographica Sinica, 2006, 61(2): 199-211.]

[16]
刘小平, 黎夏, 陈逸敏, 等. 基于多智能体的居住区位空间选择模型. 地理学报, 2010, 65(6): 695-707.

[Liu Xiaoping, Li Xia, Chen Yimin, et al.Agent-based model of residential location. Acta Geographica Sinica, 2010, 65(6): 695-707.]

[17]
兰宗敏, 冯健. 城中村流动人口的时间利用及生活活动时空间结构. 地理研究, 2010, 29(6): 1092-1104.

[Lan Zongmin, Feng Jian.The time allocation and spatio-temporal structure of the activities of migrants in 'village in city': Surveys in five 'villages in city' in Beijing. Geographical Research, 2010, 29(6): 1092-1104.]

[18]
廖邦固, 徐建刚, 梅安新. 1947-2007年上海中心城区居住空间分异变化: 基于居住用地类型的视角. 地理研究, 2012, 31(6): 1089-1102.Based on long-term residential land-use data,this paper makes a calculation on the dissimilarity of diverse residential lands,which might be used as reference in the perspective of physical changes of residential spaces.(1)This study classifies residential land use of downtown Shanghai into 6 types:garden house and villa(coded as R1),high-rise apartment before 1949and workers'community after 1949(R2),commercial residential building(R2N),li-nong residential building(R3),shanty town(R4)and rural house(E6).Then,calculations are made on the spatial differentiation,i.e.the index of dissimilarity(D),spatial-modified dissimilarity index(D(s)),multi-group dissimilarity index(D(m))and spatial-modified multi-group dissimilarity index(SD(m))of various land-use types on the spatial scale of blocks and towns.(2)The result shows that the changing of residential spatial differentiation in different time series is not affected by scale effects or whether the dissimilarity index is spatial-modified or not.(3)From 1947 to 2007,in the type of garden house and villa,the dissimilarity maintains high,while the dissimilarity of commercial residential building keeps decreasing.In other types,however,the dissimilarity has a wave change.(4)D(m)of residential land-use shows that residential segregation might be notable in 1947,and decreases obviously from 1947 to 1979,while D(m)of residential space decreases obviously,and increases significantly from 1979 to 2007.(5)The relation between the hierarchy and the dissimilarity of residential land differs in various periods.Before 1949,the dissimilarity is high within high-rank residential land,whereas the index is quite low in medium and low rank residential land.During the socialist period,the rank and the dissimilarity have a positive correlation.In the transitional period,a"Vshaped"pattern can be found,which means that the dissimilarity of high rank and low rank residential land is high,and low dissimilarity can be seen in medium rank residential land.This indicates that the residential space of Shanghai has been polarized in terms of physical environment.

DOI

[Liao Banggu, Xu Jiangang, Mei Anxin.Evolution of residential differentiation in Central Shanghai City (1947-2007): A view of residential land use type. Geographical Research, 2012, 31(6): 1089-1102.]

[19]
蒋亮, 冯长春. 基于社会-空间视角的长沙市居住空间分异研究. 经济地理, 2015, 35(6): 78-86.随着住房商品化政策的实施,城市居住空间在房价的“过滤”和社会经济差异的“分选”机制作用之下,不同职业背景、收入状况、价值取向的居民在住房选择上趋向于同类相聚,形成一种居住分化甚至相互隔离的状况。从社会—空间视角出发,利用住房数据与问卷调查相结合的方式,对长沙市的居住分异情况进行研究。通过聚类分析将长沙住宅划分成五类,归纳总结其分布和居民特性。以街道为单位从宏观和中观尺度,对各住宅阶层的分异指数进行定量研究,从而得到长沙居住空间分异的特征。最后,通过构建多元排序logistic住宅阶层选择模型,对分异的微观动因——居民的社会属性进行分析。

DOI

[Jiang Liang, Feng Changchun.The study of residential differentiation in Changsha based on the social-spatial perspective. Economic Geography, 2015, 35(6): 78-86.]

[20]
刘望保, 翁计传. 住房制度改革对中国城市居住分异的影响. 人文地理, 2007, 22(1): 49-52.住房制度改革是中国城市居住分异的重要影响因素。住房制度改革后，城市居民可根据自身社会经济特征和消费偏好，选择居住区位和住房与邻里质量。住房选择行为相对自由化。商品房的开发在住房制度改革后得到鼓励，房产商在政府宏观调控指引下，自主选择投资区位，建设不同层次的住房和邻里，以满足不同层次居民的居住需求。不同类型企业、不同工龄的职工在享受房改房的机会明显不同，造成享受者和未享受者住房条件的差异。从而形成居住分异。公共部门的干预使得住房类型构成多样化，出现了经济适用房、房改房和廉租房等资助房，与商品房并存；种种约束条件使得社会群体在不同类型住房中分布并不均衡。形成居住分异。

DOI

[Liu Wangbao, Weng Jichuan.The impact of housing reform on residential differentiation in urban China. Human Geography, 2007, 22(1): 49-52.]

[21]
Pu Hao.The effects of residential patterns and Chengzhongcun housing on segregation in Shenzhen. Eurasian Geography and Economics, 2015, 56(3): 308-330.As cities in China undergo growth and transformation, they continue to absorb migrants from both ends of the economic spectrum, giving rise to socially mixed cities. As this occurs, the cities experience an elevated level of residential segregation due to the emergence of new forms of enclave urbanism, such as gated communities andchengzhongcun(villages-in-the-city). Factors including historical legacy, land institutions, and property-led development have contributed to this divided residential pattern at the neighborhood level. However, at larger geographical scales, the degree of segregation depends on whether the provision of different housing types is systematically segregated among urban districts. This paper, using Shenzhen as a case study, examines the spatial logic of the divided pattern of the population by analyzing the distribution of both urban residents and housing provisions. The analysis explores segregation between the privilegedhukouholders and underprivileged non-hukoumigrants as well as the spatial separation of formal urban housing andchengzhongcun. As expected, non-hukoumigrants are largely segregated fromhukouholders due to their much-constrained choice of housing and the widespread availability ofchengzhongcun. A rather low degree of segregation is manifest at the sub-district level. The pattern is somewhat more desirable, as it maintains a more spatially equitable setting that enables disadvantaged groups to reside within short distances of jobs and amenities. Nevertheless, urban renewal programs targeted atchengzhongcunare most likely to jeopardize such a pattern of housing, which may aggravate segregation at the larger geographical levels.

DOI

[22]
Clark D, Davies W K D, Johnston R J. The application of factor analysis in human geography. Journal of the Royal Statistical Society, 1974, 23(3/4): 259-281.

[23]
Denton M N A. The dimensions of residential segregation. Social Forces, 1988, 67(2): 281-315.

DOI

[24]
陈颂, 汪鑫, 那鲲鹏, 等. 转型新时期上海房权空间分异格局和机制研究. 城市发展研究, 2016, 23(7): 18-23.

[Chen Song, Wang Xin, Na Kunpeng, et al.Study on tenure-based housing segregation in transitional Shanghai. Urban Development Studies, 2016, 23(7): 18-23.]

[25]
James M S.A generalized index of dissimilarity. Demography, 1981, 18(2): 245-250.The index of dissimilarity can be interpreted as the ratio of the number that must be moved from cells of excess to cells of deficit to achieve even distribution. This interpretation is used to generalize the index in two directions. First, the index is made applicable to more than two groups at a time. Second, an index and a test of significance are made available for explorations of cells of a two-way contingency table. DISSIM is the name of a computer program which provides these calculations for contingency tables.

DOI PMID

[26]
Jakubs J F.A distance-based segregation index. Journal of Socio-Economic Planning Science, 1981, 15(6): 129-131.A tool for measuring segregation in settlement patterns is introduced. This in an extension of the well-known Index of Dissimilarity. By incorporating locations of areal units into the measurement process directly, the distance-based approach substantially reduces the dependence upon size and number of data observations characteristic of the Index of Dissimilarity and other approaches. Experimental tests are reported. These suggest that the locational index constitutes an improvement and is applicable in comparative studies, either cross-sectional or longitudinal.

DOI

[27]
White M J.The measurement of spatial segregation. American Journal of Sociology, 1983, 88(5): 1008-1018.The index of dissimilarity has come to be the principal statistic for measuring segregation, particularly urban residential segregation by race. Recently, though, a literature has arisen which criticizes the dissimilarity index and proposes revisions or alternative statistics. Here a statistic is derived that explicitly incorporates the spatial relationships among the geographic parcels into the tabulation, a feature absent from the dissimilarity index and its competitors. This proximity statistic is compared with other indices and is found to be somewhat successful in distinguishing between single-cluster and multiple-cluster residential settlement patterns.

DOI

[28]
王婷, 余丹丹. 边缘社区更新的协作式规划路径: 中国“城中村”改造与法国“ZUS”复兴比较研究. 规划师, 2012, 28(2): 81-85.

[Wang Ting, Yu Dandan.Cooperative approach for marginal community renewal: Comparison between china's urban village and France's ZUS. Planners, 2012, 28(2): 81-85.]

[29]
班茂盛, 方创琳. 国内城市边缘区研究进展与未来研究方向. 城市规划学刊, 2007, (3): 49-54.我国城市边缘区研究可划分为1980年代至1990年代中期和1990年代至今两个阶段,对不同时期城市边缘区研究的内容进行了回顾与总结,以及研究特点和研究中存在的问题进行了评述,探讨了未来研究中需要注意的问题.

DOI

[Ban Maosheng, Fang Chuanglin.Progress in research on urban fringe and basic frame of research in the future. Urban Planning Forum, 2007, (3): 49-54.]

Options
Outlines

/

〈 〉