Female labour mobility, skill acquisition and choice of labour markets : theory and evidence from the Philippines

 0RD 1HSCUtiSlON .PAPER Report No. DRD284 ..F.Ed:tt\.LE LABOUR HOBiLITY,, SKILL ACQUlSLTION AND CHOICE OF LABOUR MARKETS: TH.EORY AND EVIDENCE FROM Tl"IE PHiLIPPINES by Oded Stark and Jennifer L. Lauby Development Research Department Economics and Research Staff WorLd Bank 'i'he World .Bank does not accept responsibility for the views expressed herein \vhich are those of the autllor(s) and should not be attributed to the World J.)ani< or to its affiliated organizations. The findings, interpretations, and conclusions are the results of research supported by the Bank; they do not ne'I.!BSSar ily represent official policy vf the Hank. The designations employed' tilt! pr~.::i>f.:!Utation of material~ and any maps u~ed in this document are solely for the eon.venience of the reader and do not imply the expression of any" = \.Jh<J.tsoever on th.t;; part of the tvorld Bank or its affiliates concerning Slat:U$ of r.lUY country, territory, cityll area, or Of its authorities, of· thl! delimitations of its boundaries, or national affiliation.  FEMALE LABOUR MOBILITY, SKILL ACQUISITION AND CHOICE OF LABOUR MARKETS: THEORY AND EVIDENCE FROM THE PHILIPPINES by Oded Stark and Jennifer L. Lauby * Paper submitted to the Development Research Department, The World Bank. Mailing Address: Oded Stark Harvard University 9 Bow Street Cambridge, Massachusetts 02138 *HarvardUniversity, Bar-Ilan University and The World Bank; Harvard University, respectively.  ABSTRACT Migration behavior by individuals, migration decisions and migration outcomes .are not neutral to the needs and constraints facing the migrants• families who stay put. In this paper we present and analyze evidence from the Philippines suggesting that the choice of migrant members and migration destination are largely determined by familial characteristics. We obtain several ·interesting insights into the migration process. The standard human capital approach explains the inverse relationship between the age of migrants and the propensity to migrate through the longer payoff period facing the .young. However, we find that the young age of migrants is explained by their greater amenability to familial income needs and familial manipulation. This amenability also seems to explain the preference for daughters over sons as migrants. Likewise, the initial labor market performance of migrants is accounted for not, as in human capital theory, by migrants' low skill levels but rather by familial needs which mandate participation in labor market activities that secure certain if low short run returns.  1 r. Introduction A major feature of economic change nowadays in Third World countries is the rapid growth of their cities caused, to a large extent, by the in-migration of men and women from rural communities. In many countries migration streams to the cities have been dominated by men. Yet in many Latin American countries and in the Philippines women's migration to to the cities is at least as prevalent as men's; in several other Asian countries the number of female migrants is growing more rapidly than the number of male migrants. However, the distinct migration patterns of women have not been examined in detail until quite recently. Most migration studies focus on the movement of men, with the assumptions either that men are t~e decision makers in the migration process and women are tied movers, or, if women migrate alone, that they follow the same routes, are motivated by similar considerations, and experience the same consequences as do male migrants. Recent work on women migrants has suggested that these assumptions may not be valid. In many developing societies ~ women increasingly migrate alone. (See, for example, Fawcett et. al. (1984); Trager (1984).) Since women perform different roles than men in society, the economy and the family, the reasons for their migration may also be different. In many cultures daughters are under the control of their parents to a greater degree than sons are, and, correspondingly, the migratory  2 behavior of women may be influenced more than that of men by familial considerations. Social scientists have tended to explain the migration of an individual by personal traits such as age, education, familial status, occupation, aspirations, attitude toward risk, etc. The human capital approach views migration as an action that an individual undertakes in order to increase the chances of getting a higher future return on investments in human capital assets such as education and training. Yet to a certain extent, all migrants are influenced by the constraints, opportunities and objectives of people around them and in particular by the needs and resources of their families (Stark (1984a), Stark and Lucas (1987)). Under fairly general conditions, a desirable strategy of a rural family in coping with unstable agricultural income is to send a family member to work in the city. A large proportion of rural-to-urban migrants in developing countries are unmarried and remit a significant part of their earnings home to their parents, thereby reducing the income variance associated with work in agriculture. Because of their immediate need for more certain or additional income, families may be more interested in present earnings than in future prospects. Reasons for migration may be thought of as forming a continuum from purely individual to family-dominated. Certain characteristics of migrants may allow us to estimate where an  3 the continuum the reasons for their migration tend to fall. In this paper, our main hypothesis is that in the Philippines women migrate for family reasons to a greater extent than do men. Furthermore, young women migrants are closer to the family end of the continuum than are older women migrants. The family needs which lead to migration may also help us understand why more migrants than nonrnigrants are in wage-earning occupations and why women migrants accept jobs which, while affording a steady short-term income, do not confer much long-term stability. It is qulte difficult to determine, let alone measure, the reasons for migration. In some studies migrants are interviewed and asked about factors impinging upon their decision to migrate. The notorious problems with this approach are that migrants may not remember all factors relevant at a previous point in time, may have changed their perceptions of the decision because of post migration experience, or may not be able to articulate the decision process. In this paper, family f' influence on the migration decision will be inferred from the relationship between family characteristics and migration by a daughter. For reasons to be specified below, we expect that family variables such as parents' education and occupation and the number of children in the family will be better predictors of th~ migration of daughters than of sons, and of the migration  4 of young daughters (aged 15 to 24) than of qaughters 25 years and older. II. Women in the Philippines In the Philippines ·a majority of rural-to-urban migrant~ are women, and a substantial p·roportion· of the labor force is female. The labor force participation ~ate of women 25 to 64 years old in 1976 was 45% (NEDA (1981)), and is probably higher today. Women make up over 50% of professionals, sales workers and service workers. In urban areas there are more women than men; the urban sex ratio (men per 100 women) is 95.5 and is 85.4 for the population 15 years old and over. There are indications that this imbalance is caused by the migration of very young women. In 1980, females made up only 47.9% of the urban population 10 to 14 years old, but 52.6% of those aged 15 to 19 (United Nations (1984)). Our contention is that a.significant part of the explanation for the considerable rural-to-urban migration by young women in the Philippines cannot merely be attributed to structural, macro-economic factors such as the nature of the light, labor-intensive industrialization and the rapid development of urban services. To a large degree, women's opportunities and roles in the economic sphere have been intricately shaped by the interplay of cultural and historical forces. Philippine culture has long recognized·the independent  5 economic role of women. In pre-Spanish times women not only worked in the fields and raised livestock but they also handled most of the trade (Infante (1975)). This tradition of women traders continues in the present day with women in control of retail sales in market stalls and small shops. A woman's migration from home in search of work opportunities is thus perfectly consistent with the acceptance in Philippine culture of such economic roles. The early colonization of the Philippines by Spain brought Christianity and with it a patriarchal structure of authority. The Spanish Code of Laws confined women to the home. They were forbidden to transact business or to dispose of property (Rojas- Aleta et. al. (1977)). The Christian teachings brought by the Roman Catholic priests inpluded admonishments to women to be obedient, passive and pious. As will be pointed out below, this cultural shift has an important bearing on the extent to which implicit contractual arrangements between parents and daughters are likely to be honored. Social norms affect economic behavior in general and, in particular, the distribution of gains between a daughter and her parents back at home which accrue from the daughter's labor market activity. They also bear upon adherence to distributing the migration gains according to a pre-migration agreed-upon rule. The American occupation from 1898 to 1946 introduced public education for both males and females. Over time, this has been  6 translated into educational attainment for the two sexes which by now is approximately equal (Mangahas (1976)). On the whole women are thus likely to possess the minimal human capital traits necessary to fac~litate entry into the urban labor market. Women have also gained political and other rights, but they have not been translated into similar pay for the performance of similar work {Szanton (1982)). It seems that the heritage of the Spanish colonial rule - a period during which women were excluded from most educational institutions and from political and economic roles in the urban areas (Szanton (1982)), has not been replaced completely by subsequent educational and political changes. Women can and do have access to urban labor markets, but the economic environment they face and the political factors underlying this environment appear insufficient to support or sustain severance of social and economic ties with their rural families. It thus seems that in comparison with men, women's urban jobs are likely to pay reasonably well and be moderately secure. It should be noted that if the urban job pays too well, the mig=ant may be able to save some income in the capital market of the urban economy in addition to, or instead of, remitting part of the income and thus escape her or his dependence upon the family earlier than otherwise. Similarly, the migrant may not need to depend on the family if the urban job is too secure. (Clearly, though, if the job pays very badly or is very insecure  7 the rural family may be required to continue supporting the migrant - an outcome the family will surely regard as unfavorable). Daughters may thus depend on their families more than sons do because their jobs do not pay as well and are less stable. Once we add to these considerations women's strong devotion to their families and their considerable willingness to abide by the decisions of their parents, we can see the rationale of Filipino families who so often choose to send a daughter rather than a son to work in the city as a strategy for advancing familial well-being. III. The Data Sources Two national surveys constitute the data sources for this paper. The National Demographic Survey (NDS), conducted in 1973 by the Population Institute of the University of the Philippines 1 covered a nation-wide random sample of households selected with a multi-stage procedure that insures proportional representation of each region of the country. The subfile used by us contains information about each married woman who resided in the .sample households, a total of 9331 women. Using personal interviews, this survey collected information ori place of residence at the time of the survey, eight years earlier, and at birth, as well as data on educational attainment, occupation, marriage and  8 childbearing experience, and other personal and family characteristics. The Status of women Survey (STW), conducted in 1976, collected more detailed data from a smaller sample of women and some of their husbands. The STW, in a two-hour interview, collected information on job histories, marriage and childbearing, and migration histories, along with education and family background variables. The nationwide sample consisted of 1,997 adults, 1,598 women and 399 men. The women were between the ages of 18 and 59 and the men were all husbands of married female respondents. Although only about a third of the population was living in urban areas at the time of the survey, half the sample was selected from rural areas and half from urban areas as defined by the National Census and Statistics Office~ This was a deliberate attempt to adequately study women living in urban areas of various sizes. The sampling for rural areas was carried out in three stages. First, one province was chosen to represent each of 10 regions~ (Western Mindanao was excluded from the study because of security problems.) Then the ru~al sample of 40 barrios was allocated among the sample provinces according to the size of the rural population in the region it represented. Specific barrios were chosen randomly. From each barrio, a sample of 20 households was drawn, and one woman aged 18 to 59 in each household was randomly chosen to be interviewed. Five husbands in each barrio were also  9 interviewed. The urban sample was divided into two parts: large urban areas, (Greater Manila, Cebu, Iloilo, Bacolod and Davao) and smaller areas with populations of 10,000 to 50,000. The 50 sample barangays were distributed among the urban areas according to their population size. A different sampling fraction was used for the large and small areas so that an adequate sample could be drawn from the small areas. The barangays were chosen randomly, 16 households from each barangay were sampled, and 16 women and four men were interviewed. IV. Migration Patterns of Filipinas Fi1ipinas are among the most geographically mobile of Asian iA women. Their high rates of migration and their predominance over men in rural-to-urban migration are confirmed by our data. In addition, the life history information shows that women are more likely than men to migrate as teenagers and that a good proportion of womenrs migration occurs well before marriage. In the NDS migration is measured by comparing place of residence at three points in time. Settled migrants are those who have moved between birth and 1965, but did not move between 1965 and 1973. Recent migrants are those who moved only in the eight years prior to the survey while frequent migrants are those who changed residence in both time periods. Respondents were divided into four groups based on size of place of residence at the time of the survey: Metropolitan Manila  10 residents; those in large regional centers (the cities of Cebu, Bacolod, Iloilo and Davao, which have populations of over 200,000 each); people living in other urban areas, which will be referred to as "towns"; and those in rural communities. Over two-fifths of the women in the NDS were migrants {42.6%). The majority of these were settled migrants (29.3% of all women), while 13% had migrated in the eight years·before the survey (Table 1). Thus migration appears to feature in the lives of a substantial proportion of Filipinas. Table 1: Recent and Settled Migrants by Current Place of Residence (1973 NDS) CURRENT RESIDENCE RURAL SMALL REGNL MANILA TOTAL TOWN CENTER NON-MIGRANTS 64.66% 56.39% 33.49% 35.75% 57.44% SETTLED MIGRANTS rural origin 19.39 23.97 33.72 41.06 24.11 urban origin 4 .. 43 4.61 9.24 7.87 5.22 RECENT MIGRANTS rural origin 4.62 5. 80. 12.93 10.01 6.02 urban origin 1.23 1.54 1.15 1.45 1.31 FREQUENT MIGRANTS rural origin 5 .. 00 4.69 8.08 2.35 4.68 urban origin 0.70 3.01 1.38 1.52 1.21 '.rOTAL lOOoOO% 100.00% 100.00% 100.00% 100.00% (6001) (1431) (433) (1449) (9314)  11 There are significant differences in the proportion of migrants living in each of the four size-of-residence groups described above. While 64.7% of rural residents have never migrated, only 35.8% of Metro Manila women and 33.5% of women in regional centers have not migrated. The proportion of non- migrants in small towns, 56.4%, falls between these two extremes. Looking in more detbil at the 64.2% of Metro Manila residents who are migrants, we see that 76.2% are settled migrants, most (84% of settled migrants) from rural birthplaces. Of all migrants, 17.8% have moved to Manila in the past eight years; again a large majority (87%) come from a rural area. Fifty-six of the Manila women in the survey (6.0% of all migrants) changed residence both between birth and 1965 and between 1965 and 1973. The STW survey, with its migration history data, allows the measurement of migration at various ages. Most migration at very young ages is undertaken with parents. In order to examine mainly migration undertaken by women alone, or, after marriage, with their husbands and children, any change of residence before the age of 15 will be disregarded in this section.  12 Table 2: Migration Before and After Age 25 by Current Place of Residence (1976 STW) FEMALES CURRENT RESIDENCE AGE AT MIGRATION RURAL SMALL REGNL MANILA TOTAL TOWN CENTER 15 to 24 26.6% 36.5% 35.6% 45.9% 32.4% 25 and over 12.7 14.0 13.8 12.0 13.1 Both age groups a.o 9.8 5.6 5.0 8.0 No migration 52.8 39.8 45.0 37.1 46.5 after age 14 TOTAL 100.00% 100.00% 100.00% 100.00% 100.00% (798) {480) (160) (159) (1597) MALES 15 to 24 20.9% 25.0% 35.9% 42.5% 25.8% 25 and over 18.9 19.2 18.0 30.0 20.0 Both age groups 6.5 18.3 10.3 2.5 10.0 No migration 53.7 37.5 35.9 25.0 44.2 after age 14 TOTAL 100.00% 100.00% 100.00% 1,.00.00% 100.00% (201) {120) (39) (4 0) (400) Fifty-four percent of the women in the STW survey had migrated after age 14 (Table 2). Of these, over half, or 32% of the total sample, migrated between the ages 15 to 24, 13% migrated only after age 24 and 8% moved in both age brackets. There is slightly more migration from ages 15 to 19 {25% of  13 women migrate) than from 20 to 24 (22%). The predominance of migration at young ages is especially true of female residents of Manila, half of whom migrated between the ages of 15 and 24. The STW contains data on men which allows some comparisons of migration patterns. The 400 men in this sample are all husbands of women respondents and so are only representative of men living with their wives. Men are less likely to migrate at ages 15 to 24 and more likely to migrate after age 24 than are women. Overall migration rates using this measure are slightly higher for men, so this sample does not reflect the national trend of higher migration rates for women. There may be job-related reasons why women mov~ to the cities at very young ages, while men who migrate to cities do so when they are somewhat older. Jobs open to women, as maids or launderers, can be performed by girls who are unskilled and not yet full-grown, whereas jobs performed by male migrants, like construction work or dock labor, often require physical strength and job skills which teen-aged boys may not have atfained. Support for this suggestion is found in census data which show that urban women of ages 15 to 19 are more likely to be employed than are young men of the same age. This is the only age group in which more- females than males are employed. The demand for marital partners may be another reason that men in their twenties move to the city. If many of the young women from rural communities migrate to urban areas in their  14 teens, rural men of marriageable age who stay in the rural areas may face low probabilities of marital matches. Hence, female selective migration motivated, at least initially, by labor market considerations may entail male selective migration which is motivated by marital market considerations. v. Effects of Family Characteristics on Migration Many Filipinas migrate at very young ages and the destination for a large majority of them is Metro Manila. It is our contention that the decision to migrate and migration patterns are often not purely individual but rath~r are determined by family resources and needs. There are many aspects of family life and economic conditions that might affect whether or not a daughter migrates. These include family income, parents' occupations, and other measures of the family's financial situation, which would indicate both the family's need for additional independent income from a migrant daughter, and the resources available to finance her move to the city. Clearly, the benefits conferred by a daughter earning away from home need to be compared with the opportunity costs associated with her absence from home. The extent to which a daughter is needed for ~ork in the family's home may be determined by the occupations of her parents as well as by the number of children in the family. In order to send a daughter to the city the family also requires some information about employment  15 opportunities available there; often the dominant sources of information and help are relatives and village-mates who have already moved to the city. While all these factors may influence migration, the analysis in this paper is constrained by the variables included in our data sets. The NDS contains information on father's education and occupation, but none on mother's work or education. The STW includes data on mother's occupation and education as well as on the number of siblings in the family. The woman's own education, available in both surveys, indicates h~r own skills and training. It also reflects her family's resources. Father's occupation can be used both as a rough indicator of the family income and as a measure of the type of family enterprise which may determine whether or not a daughter's work. is needed at home. In order to test the effect of type of father's occupation on a daughter's migration, CATFIT, 4' a type of path analysis which was developed for categorical variables (Nie et. al. (1980), Davis (1976)), was used to estimate path coefficients. Father's occupation was divided into three categories: nonmanual, farm and manual occupations (Figure 1). The coefficients in a CATFIT path diagram are easy to interpret since they are expressed as proportions. The .099 next to Nt)nmanual indicates that 9. 9% of fathers are ~n this categoryJ while 74~4% are in agriculture. The manual category  16 FATHER'S OCCUP .498 .099 .108 ,,, ,,, -.282 ,,,,, ,,, ,,,rt' .089 ,,, ,,, , - 042 tlilfJIIIIIIIJIL-----,ro---' 1 ,' f . • ,......•••••••'''" -- ,,,, f' I ............''''''""'••••••• . 7 44 --·~ FARMf•''''''''''l'''''lt .356 HIGH SCH is educational· attainment of at least high school MIGB65 is migration between birth and 1965 MIG6573 i~ migration between 1965 and 1973 All displayed coefficients are significant at a= .05 Figure 1: Effects of Type of Father's Occupation on Migration using CATFIT coefficients {1973 NOS)  17 does not appear in the diagram, but forms the base category, against which the effects of the other categories are measured. The next variable is High School, which indicates whether or not the respondent has at least one year of high school education. The .498 above High School indicates that 49.8% of the base category, daughters of men with manual occupations, have some high school education. The positive path from Nonmanual father's occupation to High School indicates that these daughters are more likely than the base category to have gone to high school and the coefficient says that this difference is 22.7 percentage points. In other words, 49.8 + 22.7, or 72.5% of daughters of nonmanual workers have been to high school. The negative arrow from farming says that only 21.6% of farmers' daughters have been to high school. The migration variables in this diagram measure migration between birth and 1965 and more recent mig~ation between 1965 and 1973. Earlier migration is positively related to attending high school and to having a father with a white-collar job. Farmer•s daughters are the least likely to migrate. Even when we control for the low levels of education of farm children, they still have the lowest rates of migration. Daughters of carpenters, mechanics and other craftsmen are more likely to migrate, perhaps because they are not expected to participate in such occupations, whereas farmers' daughters may be of help on  18 the farm. More ~ecent migration is also related to education and is positively related to previous migration. There are other aspects of a family's situation, aside from father 1 s occupation, which may affect the likelihood of a daughter's migration. These include mother's education and occupation and the number of siblings in the family. While studies of socio-economic status often focus solely on the l occupation and income of the men in the household, it is increasingly recognized that the women are important contributors to the income and prestige of the family. Especially when we look at the experience of a daughter, her mother's education and occupation are very important predictors of the type of work she will perform (Stevens and Boyd (1980)). The number of siblings may also play a role in the decision whether or not a daughter will migrate both because more children mean more mouths to feed and a need for more income, and because a large number of children insures that someone will be left to help out with work at home, even if one or two siblings move to the city. For both of these reasons it is expected that migrants would tend to come disproportionately from large families. The STW is the source of information on these variables. This survey also contains information on the age of women at migration, and thus allows us to check whether the effects of family characteristics are different for younger migrants than  19 older ones. In general, we expect that women who migrate between the ages of 15 and 24 will be affected most by family traits, indicating that their migration is more a manifestation of family needs than individual aspirations. Mother's work outside the horne might be expected to affect the migration pattern of her daughter in several different ways. As she helps to earn the family's income, she and her husband may expect a daughter to also work outside the home and so may be more likely to send her to the city. The daughter may want to follow the model set by her mother and so may be more willing to accept a role as income earner for the family. However, when a mother supplements the family income, there may be less need for extra earnings from a migrant daughter. In this analysis, father's occupation is measured with a prestige scale developed for the Philippines (Lauby (1976)). The prestige score for each occupation was estimated using the average education, income, age and number of persons in the occupation, assigning each occupation a score from 1 to 100. A path diagram using data from the 1976 STW (Figure 2) shows that the two family variables which help predict migration between the ages of 15 to 24 are the prestige of father's occupation which has a negative effect on migration and. number of siblings which is positively related to migration. It is interesting that education, which is usually positively related to migration does not have an effect on migration at young ages,  20 .089 MIG25+ .706 · · · · · · · ·~ ......"......... -.080 ,,,, 1111 111111,,,___ _ _...., MIG1524 SIBLINGS PAED Is fatherts educational attainment in years of schooling MAED is mother's educational attainment in years of schooling PAPRES is the prestige score of father's education EDUC Is respondent's educational attainment in years of schooling MAWORK is whether or not mother was employed outside the home SIBLINGS is number of brothers and sisters MIG1524 is migration between the ages of 15 and 24 MIG25+ is migration after the age of 24 All displayed coefficents are significant at a= .05 Figure 2; Effects of Family Characteristics and Education on the Migration of Women (1976 STW)  21 although in the next figure we shall see that education does positively affect migration after age 24. That daughters of fathers with lower prestige occupations, who probably also earn low incomes, are the ones most likely to migrate helps support the claim that families in need of extra income send young daughters to the city. The number of siblings may also indicate a need for more income, as well as the labor supply available for substitution in farming or other tasks. When this path analysis is conducted separately for rural-born and urban-born women, an important difference is that the number of siblings has a strong positive relationship with migration for women from small towns, but an insignificant relationship for rural-born women. A large family may be able to find work for all its members on a farm, but in a to~>?n a large number of children may be a liability resulting in daughters being sent to the city. A mother working outside the home has no effect on her daughter migrating. It may be that the positive effect of the attitude toward women's work is offset by the negative effect of increased family income. Perhaps mother's work does influence her daughter's choice of occupation, but not her migration pattern. Mother•s education does not have a direct effect on migration, but does affect her daughter's education, and this effect is stronger than that of father's education. The relationships in the path diagram may be affected by the age of the women. Not only would age be related to  22 increased migration after age 24, but it should also have a negative correlation with education, as the schooling Qf younger cohorts has been increasing. The next diagram .(Figure 3) includes age as an independent variable and compares the path coefficients for women and men. In the previous diagra~, migration aft~r age 24 seemed to be positively related to father's prestige. However, when age is controlled this relationship disappears. As we might suspect, migration after age 24 is positively related to age, simply because older women have had more years in which to migrate. The small. but positive effect of education ofi migration after age 24 is also evident in this diagram. The various path diagrams have shown that family characteristics, especially father's occupation and the number of siblings, are related to the migration of daughters. It is our hypothesis that family characteristics are more important in explaining the migration of daughters than that of sons. If the hypothesis is verified, it would lend support to the contention that in the Philippines family needs determine the migration of daughters while migration by sons is determined more by their own individual aspirations and needs. The path diagram for the males in the STW sample supports the hypothesis, since none of the family variables are related to the migration of a son, either before or after age 25o While it would be preferable to test this hypothesis using a larger and more representative male  23 FEMALES MALES AGE -.209\ .744 I MIGJ52~ ~IBLlNGS I ~ Figure 3: Effects of Age and Family Characteristics on the Migration of Women and Men {1976 STW)  24 sample, the results support the tentative notion that a daughter's migration is controlled by her family to a larger extent than a sot!s migration. VI. Effects of Migration on Occupation and Income Families may not only determine whether or not a daughter migrates but also exert influence upon the type of work she obtains when she gets to the city. This may contribute to the explanation of why jobs held by migrants differ from those of urban-born women, as has been found in a study of Asian women (Shah and Smith (1984)). Jobs can be divided into those that generate wages; unpaid work in a family enterprise and self employment. While working for wages often indicates a job in the modern sector, there are also low-wage jobs in traditional service occupations, such as maid, waitress or shop assistant. women who are self employed often own their own small shop or are market vendors or dressmakers. These women may not earn a steady monthly income, but their jobs can relatively easily be combined with child care and housework and so are preferred by many married women. Unpaid family workers are women with the lowest personal income and the poorest chance for mobility • ..  25 Table 3: Type of Occupation by Migrant Type (1973 NDS) MIGRANT TYPE TYPE OF NON- SETTLD RECENT FREQNT TOTAL OCCUPATION MIGRNT MIGRNT MIGRNT MIGRNT 28.9% 41.2% 49.1% 43.7% 34.6% WAGE EARNER SELF EHPLOYED 46.5 45.3 35 .. 8 43.7 45.3 FAMILY WORKER 24.6 13.5 15.0 12.6 20.0 TOTAL 100.0% 100.0% 100.0% 100.0% 100.0% (2131) (1106) (226) (190) (3654) x2 = 111.775 df = 6 prob = 0.000 No matter what their destination, migrant women are more likely than nonrnigrants to be wage earners (Table 3). This is true of migrants to rural communities as well as of those who move to large cities. we might expect that fewer migrant women would be workers in a family enterprise since many women migrate alone. Yet women migrants are also less likely to be self employed. The differences are greatest in rural communities where recent migrants are twice as likely as nonmigrants to earn wages (41% for migrants versus 21% far nonmigrants). Half of nonmigrants and 38% of recent migrants are self employed. In Metro Manila, 70% of recent migrants and 74% of frequent migrants are wage earners, versus 67% of nonmigrants and 64% of settled migrants. Settled migrants are more likely than recent  26 migrants to be self employed and their jobs are more similar to those of nonmigrants. The effect of migration on the type of occupation is made clearer in the diagram with CATFIT coefficients (Figure 4). As expected, women who live in cities are more likely to be in wage-earning jobs, while women living in rural communities tend to be self-employed or to work in a family enterprise (the base category in this model). However, when birthplace is controlled, as it is in this diagram, or when place of current residence is controlled (diagram not shown), migrants are more likely to be wage earners than are non-migrants. When birthplace and educational attainment are controlled, there are 10 percentage points more wage earners among recent migrants than among nonmigrants. For women with less than a high school education, 23% of nonmigrants born in small towns (the base category for birthplace) and 15% of nonmigrants in rural areas are wage earners. This increases to 25% for rural-born migrants and to 60% for migrants with a high school education. There is also a negative effect of recent migration on being s~lf employed. Fifty-eight percent of less educated rural- and town-born women are self employed. Of similar women who have recently migrated, only 50% are self employed. There are several possible explanations for why relatively more migrants are found in wage occupations, even after controlling for education and birthplac~. One is the young age  27 BIRTHPLACE .232 .400 .130 .352 .117 - ~188 . -- ---~--- .333 .585 All displayed coefficients are significant at a==.05 Figure 4: Effects of Migration on Type of Work using CATFIT coefficients (1973 NOS)  28 of recent migrantsa It may be that young women are more likely to have modern-sector jobs that eatn wages while older women are more likely to be self employed in more traditional sales and craft jobs. Independent of education, age is a factor in the type of job a woman performs because of the rapidly changing mix of occupations in the labor market. Whe~ older women started to work, there were fewer wage-eatning jobs avsilable and .so these women tended to be self employed in craft or sales occupations. With the rapidly growing urban centers and increasing industrialization have come more wage-earning jobs and employers who readily hire young women with no labor force ex~erience. The data show that young women are more likely than older women to be wage earners (Table 4). Forty-three percent of working women aged 15 to 29 earn wages, as against 38% of women 30 to 44 and 26% of those 45 and over. However, within each age group, migrants are still more likely than·nonmigrants to earn wages. Thus age may explain the type of work only partially. Another explanation, and one that is not so easy to test, is that self employment may not be as accessible to migrants as it is to urban-born women. ~o be successful in one•s own business may depend not only on owning some capital, but also on information and network capital, viz. having connections with prospective customers, suppliers and persons in authority, etc. This might explain why settled migrants are more likely to be  29 self. employed than are women who have recently moved to the city. Table 4: Type of Occupation by Migrant Type and Age (1973 NDS) TYPE OF OCCUPATION MIGRANT WAGE SELF FAMILY TOTAL TYPE EARNER EMPLD WORKER NONMIGRANT 28.9% 46.5% 24.6% 100.0% 15 to 29 years 35.9 35.5 28.5 100.0 30 to 44 31.4 45.1 23.4 100.0 45 and over 21.7 54.6 23.7 100.0 SETTLED MIGRANT 41.2 45.3 13.5 100.0 15 to 29 years 50.6 34.9 14.5 100.0 30 to 44 45.7 41.7 12.6 100.0 45 and over 31.0 54.8 14.2 100.0 RECENT MIGRANT 49.1 35.8 151t0 100.0 15 to 29 years 56.8 28.4 14.8 100.0 30 to 44 46.6 37.5 15.9 100.0 45 and over 40.0 46.0 14.0 100.0 FREQUENT MIGRANT 43.7 43.7 12.6 100.0 15 to 29 years 55.4 28.6 16.1 100.0 30 to 44 41.1 48.9 10.0 100.0 45 and over 34.1 52.3 13.6 100.0 N (1266) ( 165 6) (732) (3654) Yet it may also be that migrants willingly choose to take wage jobs and the reasons that they do so may be the same as those that spurred their very migration: their family's short-run need for a stable source of income. If it is true that the family is instrumental in deciding that a daughter • _., . •·, ~ • f >• ~ '" I. · ol 0 c<. :). , • "< I., 30 should migrate, it is probable that the family would also have a say in the pattern of earnings and the kind of job the daughter undertakes in the city. The family often has ties to relatives and friends in the city through whom it finds work and a place to stay for the daughter. If the family wants a quick and steady source of income, it may choose a wage-earning job, and since it may expect remittances from a daughter (as well as willingness to remit) to decrease after a few years when she may marry and have other obligations, the family may not be too concerned with the long run stability of the job or in the long term chances for upward mobility. In the early stages of industrialization and modernization, family ties tend to be strong, and parents maintain control over unmarried children, particularly daughters. One way that rural families can partake of some benefits of urbanization and industrialization while, at the same time, avoiding some of the costs associated with these processes, is to send a daughter to be a wage earner in the city while they themselves stay put. While both educated urban-born women and rural-to-urban migrants are more likely to hold wage-earning jobs than to be self-employed, the types of jobs held by each group are often very different. The next table (Table 5) shows in more detail that th~ types of occupations held by women are contingent upon their migration status.  31 Data from other Asian nations suggest that the occupations of migrant women differ from those of nonrnigrants (Shah and Table 5: Occupations of Female Migrants and Nonmigrants ALL WOMEN MIGRANT TYPE NON- SETTLD RECENT FREQNT TOTAL MIGRNT MIGRNT MIGRNT MIGRNT PROFESSIONALS 13~0% 19.3% 11.7% 13.0% 14.9% MANAGERS 0~6 2.5 0.6 0.0 1.2 CLERICAL WORKERS 4.9 6.8 4.9 7.0 5.6 SALES WORKERS 27.0 29.7 25.8 33.0 28.1 FARM WORKERS 30.7 19.6 22.7 17.4 26.1 CRAFT WORKERS 18.5 12 .. 2 13.5 18.3 16.2 SERVICE WORKERS 2.2 3.4 8.0 5.2 3.1 DOMESTIC SERVICE 3.0 6.5 12.9 6.1 4.9 TOTAL 100.00% 100.00% 100.00% 100.00% 100.00% (1427) (797) (163) (115) (2502) x2 = 140.899 df = 21 prob = 0.000 MANILA RESIDENTS PROFESSIONALS 22.2% 21.2% 5.7% 21.4% 20.4% MANAGERS 1.8 5.3 0.0 o.o 3.4 CLERICAL WORKERS 20.5 15.9 8.6 14.3 17.0 SALES WORKERS 38.0 27.4 25.7 35.7 31.6 FARM WORKERS 0.6 0.4 o.o 0.0 0.4 CRAFT WORKERS 11.7 17.7 11.4 7.1 14.6 SERVICE WORKERS 1.8 1.8 20.0 14.3 3.6 DOMESTIC SERVICE 3.5 10.2 28.6 7.1 9.0 TOTAL 100.00% 100.00% 100.00% 100.00% 100.00% (171) (226) (35) (14) (446) x2 = 87.560 df = 21 prob = 0.000  32 Smith (1984}). This finding is consistent with this table. The occupation structure of women in the 1973 NDS is very similar to that of women in the 1970 Census. About 15% of working women are professionals. Over a quarter are farm workers and another quarter are sales workers, most of whom are self-employed vendors or shop proprietors. Craft work, done by 16% of working women, includes factory work, but the majority in this category work at home or in small shops as dressmakers, weavers or potters. Only 8% of all women are in service occupations, including domestic work as maids and launderers (lavanderas). Focusing on Manila residents points out the large differences in the occupations of recent migrants and nonmigrants. Twenty-two percent of nonmigrants but only 6% of recent migrants are professionals. Part of this difference is due to the young age of recent migrants. However, nonmigrants are also more likely to be clerical workers and sales workers~ Recent migrants are much more likely to be in service work, in particular domestic work as maids or lavanderas. Settled migrants, who have been in Manila for at least eight years, have an occupational distribution more similar to that of nonmigrants, although they are more likely than Manila-born women to be in domestic service and less likely to be sales workers. The proportion of settled migrants who are professionals is high at 21%.  33 In general, recent migrants are found in the jobs requiring the least skills, yielding low pay. These jobs do not offer the prospects of upward mobility or advancement. Of 156 women who worked in domestic service in 1965, only 47% were still employed in any occupation eight years later. Of those who were employed in 1973, the large majority, 60 women, were still in domestic service. Only 14 of the 156 had moved to other jobs in sales or crafts. Thus women who migrate to the city do not find many opportunities to secure a stable or prestigious job. However, if their migration is not motivated by individual aspirations, but by the short-run needs of their family, the basic concern would be to earn a quick and steady income. Thus, the next question to consider is whether migration indeed increases a woman•s income. The women in the STW earned an average of 231 pesos a. month (Table 6. At the time of the survey the official exchange rate was around 8 pesoa to the dollar.) While this seems like a small amount in dollars, it is an important contribution to the family income in a poor country where the average family income was under 600 pesos a month at the time of the survey. It should be mentioned, though, that women earn less than men. The mean monthly income for men in the survey was 418 pesos. As expected, rural women earn less than urban women. However, it is surprising that women in small towns earn more on tne average than those in large cities. One explanation for  34 this rather substantial difference (~15 per month in towns Table 6: Mean Monthly Income by Age at Migration and Currept Residence in pesos (1976 STW) AGE AT l-1IGRATION CURRENT NON- 15-24 25 + 15-24 TOTAI; RESIDENCE MIGRNT & 25. + RURAL 152. 65 167.04 155.35 138.50 155.66 SMALL TOWN 293.06 290.78 320.42 431.61 315.44 LARGE CITY 187.96 191.57 266.60 246.75 208.05 TOTAL 205.34 230.71 247.29 311.50 231.06 N ( 270) (202) (105) ( 66) ( 643) versus 208 in cities) might be the larger numbers of migrants in large cities who depress unskilled wages. Yet when we compare the incomes of nonmigrants only, the gap between small and large cities remains. In the towns more women are self employed while in the large cities more are wage earners. Self employed women, working in shops and as dressmakers, may not earn a steady income, but tend to average a higher income than do wage earners. If women seem to do better in small towns than in large cities, earning higher incomes and holding self-employed occupations that easily combine with childcare and household tasks, why do not more women migrate to the towns instead of flocking to the large city? The small town "bundle" of opportunities would seem to better fit the personal needs of the  35 migrants themselves. Yet the needs of the sending family may be better served by the opportunities for wage earning in the large city. That women who with equal ease can migrate to small towns end up in large cities in spite of the adverse effect this has on their own long-run employment and income prospects could constitute evidence in support of our family decision hypothesis. Families send young daughters to the city to add to the family income. our data indicate that migrants to the large city have a higher average income than rural nonmigrants. Yet not only migration, but family characteristics, education and age also have effects on income. A path diagram will help to disentangle these effects. Th~ largest direct effects on income for rural born women are their occupation and education {Figure 5). There is also n smaller positive e~fect of early migration on income. With educational attainment and occupation controlled, women who moved out of a rural community at a young age earn more than those who stayed behind. It is interesting to note that there is no direct effect from migration after age 25 on income, even though these women earn much more than do young migrants. The higher income of these women is explained by their higher educational attainment and by their father's occupation which have positive effects on their occupation and thereby increase their income.  36 PRES is the prestige score of the daughter's occupation INCOME is the daughter's monthly income in pesos All displayed coefficients are significant at a= .05 Figure 5: Effects of Migration on Personal Income for Rural-born Women (1976 STW)  37 A similar path diagram for males shows that migration does not have an effect on income once education and occupation are controlled. This suggests that women get some special advantage from migrating to the city that men do not enjoy. Perhaps women, more than men, are underemployed in rural communities. Underemployment can be measured in many ways; in this case it will be defined as not being able to work as many hours as desired. If women who move to the city are able to work longer hours than they would have if they remained in the rural community, this may help explain why they are able to earn more. Especially in the large cities, there are appreciable differences in the hours worked by migrants and nonmigrants. While urban-born women work 44 hours a week, all migrant groups work more than this, with recent migrants working 51 hours and settled migrants, an average of 47 hours a week. Rural nonmigrants work an average of 37 hours a week. Thusu on average, a recent migrant from a rural community to the large city can expect to work 14 more hours a week than she did at home. This longer work week may explain the increase~ income that migrants enjoy. The path diagram with hours added supports this explanation (Figure 6). When hours are controlled, the direct effect of migration on income disappears. In other words, migration both before and after age 25 increases the hours a woman works and this is what increases her income. Father's education also  38 INCOME HOURS is the number of hours worked per week All displayed coefficients are significant at a= .05 Figure 6: Effects of Hours of Work on Income for Rural-born Women (1976 STW)  39 increases hours worked, perhaps because educated fathers have connections that help them find good jobs for their daughters. While father's education increases his daughter's income through increasing the hours she works, the daughter's own education affects her income through the occupation she enters. To summarize our findings, there are three paths to a higher income. One is getting a good job, and for this a good education is important. Another path to a higher income is working longer hours, and this is the path taken by most migrants. A third path would be the combination of a good job with migration and longer hours, which is the pattern of some older migrants. Young women migrants to the city do not get prestigious jobs or jobs that have high hourly earnings. They do get jobs that allow them to work more hours a week than they would in rural areas, and so increase their income. VII. Concluding Comments Even when an individual migrates alone, he or she may be acting as an agent on behalf of a principal back at home. In many cultures, the family is an especially strong unit that exerts influence over a daughter or son even after they are adults and have left home. In such cultures, the family can be thought of as an economic unit as well as a social grouping that transcends well-defined labor markets and geographical loci.  40 The foregoing analysis raises and partly addresses three questions: first, what are the factors which make it more likely that a family will send a member to the city, and second, why, in some societies, does the family usually choose to send a daughter rather than a son. The third question is what are the consequences of the family decision on the type of occupation and work history of the woman migrant. A family will decide to send a migrant if there is a need for additional or more steady income, if the expectation that migration will confer such benefits is reasonably high, and if the opportunity cost associated with migration is low. In our analysis, father's occupational prestige is a rough indicator of the economic position of the family. Fathers with the poorest jobs are the ones who are the most likely to send a young daughter to the city. The number of siblings in the family also reflects the income·needs of the family and how easy it would be to substitute for the departure of a member. Larger families are more likely to send a member to the city. Another factor that a family must weigh in its decision is the type of work performed by the family. Both sons and daughters of farmers are expected to help in farm work and this may help explain why children of farmers are the least likely to migrate. On the other hand, daughters of nonmanual workers, most of whom work for wages and do not own their own enterprise, are the most likely to migrate. Daughters of manual workers  41 fall in between these two groups in their propensity to migrate. Girls may be expected to help in pottery-making and basket- weaving, but not in wood-working or carpentry. Putting these pieces together, a family most likely to send a member to the city would have a large number of children with the father in a low-paying occupatibn that does not allow for participation by daughters. While both the prestige and type of father's occupation affect a daughter's migration, her mother's work does not seem to bear a significant effect. This may be because mother's work has both income and substitution effects on the decision to have a daughter move to the city. If the mother works outside the home, she adds to the family income which makes it less necessary for the daughter to migrate. Yet if the mother takes part in a family enterprise, tbe daughter's work may not be needed at home. Our analysis has suggested some ways a family's position may affect its decision about migration by one of its members. But why is it that in many societies the family usually decides to send a son, while in the Philippines (and, for that matter, in Latin America, too) a daughter is the one most often sent? Part of the explanation may have to do with the work performed at home and the possible jobs available in the city. The work of sons on the farm and in craft manufacturing in the rural community may be more valued than that of daughters. In the  42 cities, many jobs are clearly sex segregated. Jobs dominated by women include domestic service and sales jobs, as well as clerical jobs and work in labor-intensive manufacturing such as garment production~ Employers often prefer to hire women because they are considered to be more docile and compliant and can also be paid less than men (Snow (1978)). There are, of course, jobs open to men, but many of these, such as construction work and labor on the docks, are seasonal and may depend on such factors as size of crop production and even the weather. Although women may earn less per day than men, their jobs are such that they may give more certain income throughout the year than jobs held by men. This job stability may be strongly preferred by risk-averse rural families seeking to minimize the variance associated with their income. The likelihood that once migrating, the migrant member will share his/her urban-earned income with the family back at home could be of equal importance. Families in the Philippines may be willing to rely on daughters to supplement their income because traditionally daughters maintain close ties with their families of origin even after marriage. Daughters and sons inherit equally from parents and family lineage is bilateral, with equal ties to both parents' families. In addition, daughters are taught to be responsible family members, taking care of younger siblings, for instance, while sons are given more freedom and are expected to be more independent. Thus a  43 daughter sent to the urban area can be relied upon to help support her parents and to aid in the education of brothers and sisters. Even though she might be expected to earn less, a higher proportion of her income is likely to be shared (remitted). If the daughter establishes a home in the city, it can be used as a means for younger siblings to avail themselves to urban education and employment opportunities, and a daughter nlay be seen as being better able to perform this function as a surrogate parent than a son. Girls are ta~9ht to be responsible in their service of the family, and are dependent on the decisions of their parents concerning their education, work and even their social activities. It is no wonder then that the close ties between the daughter and her parents give rise to a causal relationship between her migration pattern and the characteristics or her family, while by and large the migration of sons is not related to family traits. The choices that migrant women make concerning their work may be at least partially explained by the familial nature of the migration decision. Migrants to large cities are found in different types of jobs than are urban-born women. They are in wage-earning jobs such as domestic service in which they earn a steady income to help support their families. They are less likely to be self employed in occupations which might give them more stable employment throughout their lives, yet yield less or less certain income in the short run.  44 The urban occupations that migrants hold allow for little upward occupational mobility, and they more often lead to unemployment after a few years. What women achieve by migration is the chance to earn more than they would had they remained in the rural community. Migrants increase their income not by earning a higher hourly wage but by work~ng more hours per week in the city. Thus migration is a solution to underemployment in the rural area and allows the woman and her family to make better use of her time. Faced with declining income from farming and rural crafts families in rural communities often decide to send one of their members to the city to earn additional or more certain income. Whom they choose to send will depend on the type of work that remains to be performed at home, what opportunities there are for work in the city, and who can be counte~ on to send horne part of the urban wage. In the Philippines it is daughters who can be expected to better meet the family needs, hence, it is they, rather than sons, whom the family entrust with the migration task. And it is not that females are inherently more mobile than males; the clue to deciphering the patterns of individual migration by sex may lie not in individual attributes per se but in the specific interactions of these attributes with familial objectives and opportunities.  45 REFERENCES Boserup, Ester, 1970. Women's Role in Economic Development (London: George Allen & Unwin). Davis, James A., 1976. "Analyzing Contingency Tables with Linear Flow ~raphs," in D. R. Heise, ed. Sociological Methodology (San Francisco: Jessey-Bass). Eviota, Elizabeth and Peter C. Smith, 1984. "The Migration of Women in the Philippines," Women in the Cities of Asia James T. Fawcett, et. al.,(eds.), (Boulder, Colorado: Westview Press). Fcwcett, James T., s. E. Khoo, Peter C. Smith, (eds.), 1984. Women in the Cities of Asia (Boulder, Colorado: Westview Press). Hendershot, Gerry E., 1975. "Fertility, Social Class and Outmigration from Two Rural Communities," Population Development in Southeast Asia John Kantner and Lee McCaffrey, (eds.), {Lexington, MA.: D.C. Heath). Infante, Teresita R., 1975. The Woman in Early Philippines and among the Cultural Minorities (Manila: Unitas Publications). Jelin, Elizabeth, 1977. "Migration and Labor Force Participation of Latin American Women," Signs vol. 3, 129-141.  46 Koo, Hagen and Peter c. Smith, 1983. "Migration, the Ur:ban Informal Sectora and Earnings in the Philippines," Sociological Quarterly vol. 24, 219-232. Lauby, J~nnifer L, 1977. The Job Hierarchy: A Study of Occupational ·Prestige in the Philippines (M.A. Thesis, University of the Philippines). Lucas, Robert E. B. and Oded stark, 19~5. "Motivations to Remit," Journal of Political Economy val. 93, 901-918. Mangahas, Mahar, (ed.), 1976. Measuring Philippine Development {Manila: Development Academy of the Philippines). National Economic and Development Authority, 1981. Philippine Yearbook 1981 (Manila: NEDA). Nie, Norman H., et. a!., 1980. user's Guide to the scss Conversational System (New York: McGraw Hill). Rojas~Aleta, Isabel, Teresita Silva and Christine Eleazar, 1977. A Profile of Filipino Women: Their Status and Roles (Manila: Philippine Business for Social Progress}. Shah, Nasra M. and Peter c. Smith, 1984. "Women at Work in Asia," women in the Cities of Asia James T. Fawcett, et. al.,(eds.} (Boulder, Colorado: westview Press). Snow, Robert T, 1978. "Export-Oriented Industrialization and its Impact on Women Workers," Philippine Sociological Review vol. 26, 189-199.  47 Stark, Oded, 1984a. " Migration Decision Making: A Review Article," Journal of Development Economics vol. 14, 251-259. Stark, Oded, 1984b. "Rural-to-Urban Migration in LDCs: A Relative Deprivation Approach,~ Economic Development and Cultural Change vol. 32, 475-486. Stark, Oded and David Levhari, 1982. "On Migration and Risk in LDCs," Economic Development and Cultural Change vol. 31, 191-196. Stark, Oded and Robert E. B. Lucas, 1987u "Migration, Remittances and the Family," Harvard University Migration and Development Program,Discussion Paper Series, Paper No. 28. (Forthcoming in Economic Development and Cultural Change). Stevens, Gillian and Monica Boyd, 1980. "The Importance of Mother: Labor Force Participation and Mobility," Social Forces vol. 59, 189-199. Szanton, M. Christina Blanc, 1982. "Women and Men in Iloilo, Philippines: 1903-1970," Women of Southeast Asia, Penny Van Esterik, (ed.) (Northern Illinois University). Trager, Lilian, 1984. "Family Strategies and Migration of Women," International Migration Review vol. 18, 1264-1277. United Nations, 1984. Demographic Yearbook 1982 (New York: Dept. of International Economic and Social Affairs).  Some Recent DRD Discussion Papers 267. Protectionism and the Debt Crisis, by S. van Wijnbergen. 268. Inventories as an Information-Gathering Device, by S. Alpern and D. J. Snower. 269. Welfare Dominance: An Application to Commodity Taxation, by S. Yitzhaki and J. Slemrod. 270. The Causal Role of Minimum Wages in Six Latin American Labor Markets, by M. Paldam and L. Riveros. 271. Tax Evasion, Corrumption and Administration: Monitoring the People's Agents Under Symmetric Dishonesty, by A. Virmani. 272. Characteristics and Operation of Labor Markets in Argentina, by C.E. Sanchez. 273. Growth and Structural Change in East Africa: Domestic Policies, Agricultural Performance and World Bank Assistance, 1963-1986, Part I, by U. Lele and R. Meyers. 274. Growth and Structural Change in East Africa: Domestic Policies, Agricultural Performance and World Bank Assistance, 1963-196, Part II, by U. Lele and R. Meyers. 275. Abstracts of Development Research Department Publications: April 1986 - April 1987. 276. Korea's Macroeconomic Prospects and Major Policy Issues for the Next Decade, by V. Corbo and s.w. Nam. 277. The Pricing of Manufactured Goods During Trade Liberalization: Evidence from Chile, Israel, and Korea, by V. Corbo and P.D. McNelis. 278. Fiscal Policy and Development Strategy in Southern Asia, by G.F. Papanek. 279. Evolution of the Tunisian Labor Market, by C. Morrisson. 280. Labour Allocation Across Labour Markets Under Different Informational Schemes and the Costs and Benefits of Signalling, by 0. Stark and E. Katz. 281. Mobility, Skill and Information, by 0. Stark and E. Katz. 282. Labour Mobility and Intrafamilial Income Transfers: Theory 'and Evidence from Botswana, by 0. Stark and R. Lucas. 283. Labour Migration, Income Inequality and Remittances: A Case Study of Mexico, by 0. Stark, J.E. Taylor and S. Yitzhaki.