کیفیت شبکه های اجتماعی و تصمیمات سرمایه گذاری های آموزشی

All individuals belong to a social network with certain quality level. This paper analyzes the role of the quality of the social network in the educational decision making process. I propose a measure for quality of network based on the schooling level and the labor position of the members of the net. The analysis compares individuals who are similar in at least two characteristics: socioeconomic level and intellectual ability. Although they belong to the same type of community (poor), they differ in the composition of their social network. The higher the quality of the network, the higher the probability of investing in education. Hence, socially disadvantaged and equally intelligent individuals may end up acquiring different schooling levels.

It is well known that the level of family income play an important role in determining the amount of human capital investment that an individual is willing to undertake. Although public education is free in many countries at basic and medium level of schooling – there is no fee – and relatively cheap at the superior level, there are other costs like transport, food and clothes, among others, that poor families cannot afford. Besides, credit markets for education are incomplete and exclude most of the low income potential applicants. What we can observe in cities like Cali (Colombia) and many others in Latin America,1 is that young poor people tend to leave school much earlier than wealthier individuals. In Cali, the lack of attendance rate among youngsters from 18 to 26 years old is around 82% in the poorest neighborhoods (eastern and mountainside areas), while for the wealthier zone the rate is 50%. In the case of 11–17 years old youngsters, the rates are 19% and 10% for the poorer and the wealthier areas respectively (Zuluaga and Benitez, 2009). Perhaps what is driving the poorer to skip school is a belief that, either way, investing or not in schooling, good jobs will be given to the wealthier. This belief may discourage the decision maker to attain higher levels of education, as he believes the instrumental value of education is low.2 This paper aims at exploring the existence of an additional factor influencing the educational investment decisions, i.e. the social networks3 that individuals belong to. Social network theory goes beyond traditional researches in which the individual's social and economic decisions are only determined by individual traits. Instead, network theory states that both personal characteristics and links with the members of the social network are important determinants of individuals’ behavior and decisions. One example of this influence is the impact of social networks in altering the incentives that individuals have to acquire education. In particular, I state that the “quality” of the network influences the perceived returns to education, encouraging or discouraging individuals’ investment in additional years of schooling. For my purpose, I define the “quality” of a social network as determined by the schooling level and links with the labor market of the network's members or ties (e.g. family members, neighbors, colleagues, friends, classmates, teachers, among others). It is worth mentioning that the term “quality” is used here as an attribute (ascribable) of the network, which is not related to the value of the individuals belonging to the network. In this sense, the meaning of quality that I use here is close, in nature, to the meaning that Becker and Lewis (1973) and other authors used in their papers referring to children's quality, which is different from referring to the value of a son or a daughter. In their context, a child's quality is a combination of endowment (inheritance) and household expenditure on the child. Having clarified this important point, the quotation marks for the word quality will not be used any more in the text. There are no previous works that focus on defining quality of networks. It is more usual to find definitions of the quality of a tie referring to the strength of the relationship (Granovetter, 1973). This paper's definition of network quality is associated with the idea that certain characteristics of the members of the network may (positively or negatively) influence the individual's behavior and decisions. Perhaps other members’ attributes besides schooling and occupational position are also important to capture this impact, however, this paper will only focus on these two characteristics. Specifying a method to estimate network quality will allow us to empirically verify the influence of networks on the schooling investment, or any other socioeconomic achievement. In this analysis, the impact of parents’ characteristics is separated from the effect of the rest of members of the network. The idea is to check the influence of networks, after controlling for parental background. It is possible to find individuals belonging to the same community or neighborhood, who share certain attributes like family income and ability, ending up at different schooling levels, expected future income and expected social mobility. The analysis of social networks may offer us an attractive hypothesis to explain this phenomenon and to explore why policies of educational expansion favor only a small portion of low income individuals. This paper proceeds as follows. The second section corresponds to the literature review, where previous contributions about the influence of social interactions on schooling investment are briefly presented. In the third section of the paper, it is made explicit how the social network quality affects educational investments of individuals. The network's quality has a potential relevant effect on the individuals’ perception of the returns to education, which in turn influences their educational decisions. I propose a specific measure for the quality of social networks, whose information requirements are: (i) quality of each member of the network, based on educational level and labor position, and (ii) the weight of each member. An important definition is the “key tie”. This is a concept characterizing a non-relative member of the network who plays a decisive role in determining the overall quality of the social network. A key tie is an initially weak tie (under the kinship criterion) who turns up to be a strong tie if we adjust his weight by factors like closeness, intimacy, economic support and admiration. Although the concepts of weak and strong ties are commonly used in the literature of social networks, there is no consensus on their precise definition. I do not pretend to be more accurate in defining the concepts here, instead, I adopt an ordering for strength of ties originally based on kinship: family (stronger), friends and acquaintances (weaker), and subsequently modified by the mentioned adjustment factors (closeness, intimacy, economic support and admiration). These adjustment factors may lower the weight of originally strong ties and could make a weak tie become a key tie. The fourth section corresponds to the empirical calculations. Existing databases do not allow us to determine the members of an individual's network, nor their characteristics. In order to obtain the required information for measuring the network quality of a group of individuals, a survey was carried out. This survey was applied to a target group and a comparison group. Individuals in both groups are similar in their intellectual ability and socioeconomic conditions – they live in the same type of poor neighborhood. Those in the first group have continued studying after secondary school whereas those in the comparison group have not. Through the survey, I find out the schooling level and the labor position of each member of the individuals’ network in order to estimate the quality of the network. I then specify a Logit model to test the influence of the network quality on the decision of individuals to continue studying, controlling for parental background and network size. The information captured through the survey is valuable because it helps us to determine the appropriate reference group likely to influence our individuals’ decisions. The last section gives some conclusions and recommendations.

This paper explores the role of the network quality on the schooling decisions of individuals. I want to show that, beyond socioeconomic background, individuals with higher quality of network – compared to their peers living in the same type of neighborhood – will experience higher probability to continue studying. Specifically, the network quality influences the perception of individuals on the returns to education, which determines the probability to continue higher education. I compare two groups of people belonging to the same type of poor neighborhood and with similar intellectual ability. The difference between them is that individuals in the first group have continued studying after secondary school, while those in the comparison group have not continued. A survey was applied to individuals of the two groups in order to obtain information about the quality of their social networks. A measure for quality network is proposed, based on the weighted sum of the qualities of the network's members. The quality of role models is related to their schooling level and their position at the labor market. As for the weight of each member, it is considered that the type of relationship (relative, friends or acquaintances) is not sufficient criterion to determine it. Other factors such us intimacy, confidence, emotional and economic support are relevant and may convert an initially weak tie into a key tie. Social networks are important in shaping the motivation, aspirations and expectations of individuals. These, in turn, affect the behavior and decisions of people in crucial aspects such as investment in human capital. It suggests, in line with the existing literature, that educational achievements of members of poor communities may generate a multiplicative effect, through social networks, beyond the private return of the educated individual. Results of the Logit model suggest the relevance of the quality of the network in schooling decisions, controlling for parental background. This might be one of the reason why it is possible to find individuals belonging to the same type of community or neighborhood, who share certain attributes like family income and ability, ending up at different levels of schooling, expected future income and expected social mobility. If we accept that social networks matter, a relevant question could be then: how can the quality of networks for individuals in poor neighborhoods be strengthened? Let us mention (not develop) some strategies towards that direction. The two last strategies do not follow directly from the analysis of this paper. Still, they are mentioned for their relevance to improve the network's quality of impoverished individuals. - From the collected information it was seen that, in several cases, the positive influence of a teacher plays an important role in shaping individuals behavior. From the target group, 17% (22%) of the individuals had teachers whose weight was adjusted by 4 (3) factors. There should be some mechanisms to strengthen the relationship teacher-students to better exploit this channel of positive influence. - Cities must offer more spaces to be indistinctly used by people from different income groups. Cities in Colombia are designed to promote polarization, which decreases the opportunities for poor youth to expand their social networks, losing the potential advantages of higher information received through possible weak ties. At least from the government initiatives, this practice should be reversed in order to avoid segregation. - Social housing (viviendas de interés social) should not be constructed in isolated places, as it is usual in Colombia. It is important that they are integrated to the city so that inhabitants may benefit from the infrastructure of the city and may easily socialize. - Students from high quality private universities have more chances to engage in the labor market, because of the good networking and prestige of the academic programs. Education in Colombia is polarized and poor students normally are excluded (not in theory but in practice) from private universities. Thus, good students from public schools should be guaranteed a given number of undergraduate scholarships in high quality private universities. This should not reduce the opportunity for them to access the scarce current scholarship offers. - A more ambitious but highly beneficial proposal is that at least one member of each family should benefit from a scholarship to cover the complete educational cycle. This member will have a positive incidence in the whole family group. Although social networks are important, it should be stressed that they are not a substitute for well-planned public policy and well-targeted public spending. Instead, social networks complement the governmental action to avoid educational exclusion of the poor and the segmentation of the labor market.