South Asian countries show different patterns of internal migration according to gender. In India, females have long dominated internal migration flows; censuses between 1971 and 2011 show that around 70 per cent of lifetime migrants were women. By contrast, in Bangladesh, in the last three decades of the twentieth century, more males than females moved (Afsar, 2000). A detailed study on one rural sub-district with excellent longitudinal data also shows more males than females moving out (Alam & Khuda, 2011). But in both countries, the patterns of migration have long been highly gendered: movement from rural areas to the cities (especially inter-state movement in the case of India) has been male-predominant, whereas localised patterns of movement have been femaledominated. Why this female predominance in localised migration? It results from marriage customs in the sub-continent, characterised (especially in north India) by village exogamy and women moving to their husband’s village on marriage. These villages were normally not very far from the woman’s home area, but sometimes further afield. Thus marriage (near-universal in South Asia) requires migration by most females after they reach marriageable age. While there is some short distance male migration as well, some of it seasonal, mainly for the purpose of seeking work, this is greatly exceeded by the marriage migration of females (Skeldon, 1985, p. 45). Because of the rules adopted in censuses for the recording of migration, in many countries much of the localised movement is not captured. It all depends on what boundaries have to be crossed in order to be considered as a migrant. If a district boundary has to be crossed, then moves within a district will obviously not be recorded. In the Indian census, the lowest units for determining migratory movements are the villages and towns whose boundaries are administratively defined. Any change of residence beyond the village or town/city boundary qualifies the person to be classified as a migrant. Given the pattern of village exogamy, most women’s moves to their husband’s place of residence on marriage are captured in the migration data. This is not the case in Bangladesh, however, where a district border must be crossed in order to be recorded as a migrant. (There are 64 districts in Bangladesh, with average populations of about two and a half million). Village exogamy is commonly though not universally practiced in Bangladesh, and most moves following marriage are less than 20 km (Rahaman, Akter, & Rahaman, 2010, Figure 2); many of these would be within the same district. The Bangladesh census would therefore not capture all marriage moves. The gender differences in migration patterns in India are striking. According to the 2001 Census, whereas 70 per cent of recent female migrants moved to rural areas, 51

中文翻译：

---

南亚女性移民的新模式

南亚国家根据性别显示出不同的内部迁移方式。在印度，女性长期以来一直是国内移民的主要来源。1971年至2011年的人口普查显示，一生中大约70％的移民是女性。相比之下，在孟加拉国，二十世纪的最后三十年，男性迁徙的人数多于女性（Afsar，2000年）。对具有良好纵向数据的一个农村街道的详细研究还表明，外出迁徙的男性多于女性（Alam＆Khuda，2011）。但是，在这两个国家中，长期以来，移徙的模式已被高度性别化：从农村地区到城市的流动（尤其是印度的州际流动）主要是男性，而局部的流动方式则以女性为主。为什么女性在本地移民中占主导地位？它来自次大陆的婚姻习俗，其特征是（尤其是在印度北部）村庄外婚和妇女结婚后搬到丈夫的村庄。这些村庄通常离妇女的住所不远，但有时更远。因此，婚姻（在南亚几乎是普遍的）要求大多数女性在达到结婚年龄后才移居。虽然也有一些短距离的男性移徙，其中一些是季节性的，主要是为了找工作，但女性的婚姻移徙大大超过了这一点（Skeldon，1985，第45页）。由于人口普查中采用了记录移徙的规则，因此在许多国家中，许多本地化活动没有被捕获。这一切都取决于必须跨越哪些边界才能被视为移民。如果必须跨越区域边界，那么显然不会记录区域内的移动。在印度人口普查中，用于确定迁徙活动的最低单位是在行政上定义边界的村庄和城镇。超出村庄或城镇/城市边界的任何居住变更都可以将该人归为移民。考虑到乡村外婚的模式，大多数女性在结婚时搬到丈夫居住地的举动都记录在移民数据中。但是，在孟加拉国情况并非如此，在孟加拉国必须越过地区边界才能被记录为移民。（孟加拉国共有64个地区，平均人口约为两百万半）。尽管在孟加拉国普遍没有实行乡村外婚制，但结婚后的大多数举动距离不到20公里（Rahaman，Akter和Rahaman，2010年，图2）。其中许多将在同一地区内。因此，孟加拉国的人口普查不会涵盖所有的婚姻动向。印度移徙模式的性别差异十分明显。根据2001年的人口普查，尽管最近有70％的女性移民移居农村地区，但51