1. Aim
The project aim was to analyse the prevalence and importance of family relations in Swedish workplaces. The aim was decomposed into the following questions: (1) How does the kinship level vary across individuals, firms, and in particular across lines of businesses and regions? (2) Is kinship bias observable among the personnel present at workplaces? (3) To what extent is an observed kinship bias rational or irrational from the firms’ point of view, in the sense that it (a) improves economic performance compared to a situation with no kinship bias in recruitments of personnel and/or (b) informants representing the firm/organization see their experience of kin related recruitment as successful or problematic?

2. Work flow
The quantitative part (based on longitudinal data on all individuals and firms in Sweden 1990-2012) started with a mapping exercise to find out the regional concentrations (co-occurrences) of family members in workplaces. The main purpose was to reveal the geographical and sectorial features of workplaces with high concentrations of family members (RQ 1). In parallel, we developed a proxy for kinship bias (RQ 2). In short, this is done by, for each individual and each potential kinship tie (spouse, child, parent, cousin etc.), first allocate all employees a new randomly selected workplace within commuting distance from the home. This new workplace should hire similar competences as the original one so that it actually could constitute a real employment opportunity. Based on this reallocation, we could then establish whether a bias is present or not by comparing the co-occurrences of kinship in cases when other employment opportunities are available. For certain workers in certain regions (small and peripheral) no other potential employers where present, meaning that the potential co-occurrence is not a real bias, but rather a result in being located in a thin labour market. This measure was then used in the next-coming quantitative analyses (RQ 3a) on the geography of kinship bias and the impact thereof on firm performance (productivity and/or survival). It should be noted that since the indicator for kinship in the database is based on the older generation (two siblings are identified because they are linked to the same mother and/or father), an undisclosed number of ties are omitted if a parent never has been registered in Sweden.

To obtain more in-depth information on RQ 1 and address RQ 3b, 40 semi-structured interviews were performed between May 2014 and February 2015. 30 of these 40 firms are in the private sector, 1 belongs to the public sector, 6 are municipal-runned and 3 governed by the state. The firms where located from Haparanda in the north to Malmö in the south and were of various sizes (4 had less than 10 employees and 5 firms had more than 1000 employees). We identified our informants based on internet-searches and social networks. In contrast to the original plan to also interview employees, we only focussed on employers or HR-personnel as it was difficult to find employees that wanted to express that they had been either decimated or received privileges due to kinship.

3. Main results
The average share of kinship occurrences at Swedish workplaces is 14%. About 10% of these are found in regional labour markets where no other alternative employment opportunity was present, while 90% could be attributed to kinship bias. That is, co-occurrences within workplaces that cannot be explained by being located in a thin labour market. Whether this is a small or large number is difficult to say, since none such similar study on an entire economy has been performed previously. We do however find that some degree of kinship bias is prevalent in almost all workplaces, but is more frequent in smaller regions and in sectors where there are less formal entry requirements (e.g., low-tech manufacturing and services). Moreover, our findings clearly show that the bias related to kinship (as compared to kinship co-occurrences) are far more common in larger regions. In fact, the 10% of all kinship co-occurrences that is due to a shortage of alternative employers all are found in small and peripheral regions.  

Not only does the concentration of kinship vary across regions, so does also the economic significance of kinship. Based on the interviews, the rationale for kinship recruitment is that it in certain regions can be difficult to find suitable skills matching the demands of the firm. Then children, or other relatives, are better than random candidates due to presence of mutual trust and some degrees of peer pressure. When regressing the effect of family concentrations on the performance of firms, we do however find that the economic benefits of kinship mainly is found in larger regions, not in small regions where the presence of kinship concentrations are higher (but the bias is lower). Hence, while kinship on the one hand mitigate labour marker deficiencies in smaller regions, it promotes performance in regions where several potential candidates (presumably) are present for each job. In relation to the conceptual discussion on the impact and interplay of different proximity dimensions in relation to firm performance and regional development, this leads to the conclusion that too much social proximity could harm performance, while in combination with other externalities in the region the social proximity induced via kinship actually constitute an advantage. We also find that the type of familial relationship within the firm matters for performance, not the presence itself. The most beneficial type of relation is the co-location of parent-child(ren), in particular if the parent is part of management and the child(ren) has/have similar training (education) as the parent.

Finally, based on the interviews we identify clear tensions in regard to hiring based on kinship. On the one hand, many respondents argue for strong principles of not giving easier access to employment based on kin, but that kinship on the other hand still can be an advantage as it often speeds up the recruitment process. HR-representatives thus identifies both risks and advantages with the increasing professionalization of recruitment practises. The pros and cons of kinship recruitment is also, according to the informants, highly dependent on the type of work relation within the family. If family relations are combined with asymmetric power relations at work, this could cause undesired tensions at the workplace. The worst-case scenario in this respect is that such tensions could harm the long-term development of the organization.   

4. New research questions
Based on the interviews, a clear notion that a recent (10-15 years) shift in hiring practises has been taking place, in particular in relation to the perceptions of recruitment based on kinship. It would therefore be interesting to further analyse the origins for this increasing professionalism and how such values are related to practices formed by continuity within organizations. While the findings in this project are focussed on the employer perspective, it would be interesting to gain further knowledge on the employee perspective. That is, how do various groups of workers perceive possibilities and constraints of working with relatives? Perhaps even more importantly, how do groups of workers not working with relatives perceive their career prospects in workplaces where the kinship bias is high? In relation to the vast literature on the role of weak ties (friends, colleagues etc) for finding a job, it would be interesting to further analyse distinctions between such weak ties and the stronger ties related to kinship. While cronyism often is regarded as something positive, nepotism often has a negative connotation although they in certain cases could be comparable. The role of weak ties in relation to both job-to-job mobility and firm performance is something that we only briefly explored in this project but a quantification of such processes definitely deserves more attention.

5. International dimensions
Apart from participation in international conferences (see below), a number of new international contacts has been established via this project. For example, the PI was visiting researcher at both UCLA (2016) and MIT (2017). Two academic milieus with outstanding competences in network studies. In particular, the collaboration with two post-docs at MIT (Balazs Lengyel and Riccardo di Clemente) has rendered in new methods to study economic networks based on micro-data. We have also invited Rodrigo Basco (American University of Sharja) to the department to act as discussant on a PhD-thesis in relation to this project.

6. Diffusion of results
The project has resulted in 7 peer-review articles (4 accepted thus far), 1 PhD-thesis, and 1 professional report (see further details in the publication list). One of the published papers (Adjei et al, 2016) was awarded the ”2016 Best paper award” in the journal Regional Science, Regional Studies. Apart from that, results from the project have been presented in a number of international conferences and workshops: IGU in Krakow (2014), NECTAR cluster 4 meeting in Toledo (2016), Association of American Geographers in Chicago (2015) and San Fransisco (2016), the Druid academy conference in Bordeaux (2016) and at a special session on family business and regional development at the Regional Studies Associations annual conference in Dublin (2017).

Results have also been disseminated to the broader public. For example, SR Västerbotten (2016-09-27), Västerbottens-Kuriren (2016-09-27), Ukeavisen Ledelse (2016-08-12), Arbetsliv (2016-08-05), Kollega (2016-07-01) and Arbetet (2016-07-01). The professional report which contextualize and summarize the interviews has been provided to all informants and is also downloadable at department web-page.