Return migrants’ perceptions of living conditions in Ethiopia: A gendered analysis

Abstract

It is widely accepted that people migrate to seek better living opportunities, and migration experiences affect life conditions considerably upon return. Research focused on gendered perspectives has illustrated that men and women experience migration differently and that this difference clearly affects their living conditions upon return. However, few studies have sought to further examine the factors that determine the differences in men and women’s perceptions upon return. This paper investigates the case of Ethiopian returnees by exploring three sets of independent variables: (1) the migration experience, (2) the return experience, and (3) the post-return conditions in Ethiopia. For this analysis, we utilize a subsample of returnees based on a household survey conducted among 1,284 households in five regions of Ethiopia in 2011. The subsample consists of 146 returnees and provides valuable information on the diverse experiences of returnees. Both the descriptive and logistic regression analyses demonstrate that: first, women have significantly worse perceptions of their living conditions upon return to Ethiopia than men do; second, migration experiences are the most significant variables that influence perceptions of living conditions upon return. Considering this latter result, we conclude by highlighting the importance of situating return experiences within the wider context of the entire migration cycle and recognizing the long-term effects of migration experiences that persist after return.

1. Introduction

According to neoclassical migration theories, return migration is a failed attempt at the migration project (Cassarino 2004; De Haas 2010). Neoclassical theory assumes that migrants return to their origin country when they do not achieve an increase in their income abroad. Conversely, the New Economics of Labour Migration (NELM) associates return migration with success and positive outcomes of the migration experience. According to the NELM, migrants may choose to return when their migration project is accomplished; that is, when they have maximized their income and saved enough money to bring home (Cassarino 2004). From a NELM perspective, living conditions should improve upon return, which we take as the central premise in this paper. However, research is increasingly demonstrating that migrants may not always be able to improve their living conditions upon return (Kuschminder 2014; Van Houte 2014). There are increasing costs and risks of migration with increased border controls, tighter employment markets in destination countries, and more arduous journeys that are required to reach destination countries. These challenges make it even more difficult for migrants to achieve their migration aspirations of returning home and the ultimate goal of improving their living conditions.

In this paper, we examine how return migrants perceive their living conditions upon return and whether they perceive that their living conditions have improved or worsened in comparison with their conditions prior to migration. Although there is substantial research on objective changes upon return, such as in the area of employment, less is known about how returnees perceive the effects of migration upon their return (Rogge 1994; Fransen and Kuschminder 2012; Cassarino 2014). Furthermore, returnees’ perceptions upon return have mainly been examined through qualitative ethnographic studies of smaller groups (Ammassari 2009; Wong 2014) and, to a lesser degree, in quantitative studies (Gmelch and Gmelch 1995; Ruiz, Siegel and Vargas-Silva 2015). In this paper, we contribute to the latter strand of research. Specifically, we use a unique dataset of return migrants in Ethiopia to examine their perceptions of living conditions upon return. The dataset features a balanced sample of male and female returnees, which enables us to construct a gendered analysis of returnees’ perceptions. Ample research has demonstrated that men and women have different experiences in migration. The next section of this paper will highlight the importance of gender in assessing perceptions of living conditions upon return.

Our second objective is to further understand the role of different stages in the migration cycle in influencing returnees’ perceptions of their living conditions upon return. This objective is achieved by examining three sets of independent variables: (1) the migration experience, (2) the return experience, and (3) the post-return conditions in Ethiopia. This analysis enables further understanding of how migrants’ experiences shape return perceptions. We attribute persistent differences to the large gender inequalities in the country and the gendered migration flows, as will be discussed in the concluding section.

The remainder of the paper is divided into six sections. Section 2 examines the role of gender in migrants’ experiences and perceptions upon return. Section 3 presents a case study with a brief overview of migration and return dynamics in Ethiopia. Section 4 discusses the role of the migration cycle in perceptions upon return and presents the hypotheses formulated. Next, in section 5, we discuss the data and methods used, followed by the results in section 6. The conclusion in section 7 examines the relationships among gender, migration experiences and return perceptions.

2. Gender, migration experiences and living conditions upon return

Over the past two decades, a growing body of research has demonstrated the importance of understanding gender considerations in migration experiences (such as Mahler and Pessar 2001; Martin 2004; Levitt 2011). Gender roles in the public and private spheres are constructed through cultures and societies, and these roles play out differently through migration opportunities and experiences. A small but growing body of literature examines the role of gender in migrants’ perceptions upon return (Gmelch and Gmelch 1995; Ammassari 2009; Guarneri 2014; Wong 2014; Ruiz, Siegel and Vargas-Silva 2015). The most important finding from this literature is that women have different perceptions and adjustment processes upon return than men do. For example, Gmelch and Gmelch (1995) found that women in Ireland were 58 per cent less content with their life upon return compared with their level of contentedness in the country of migration, whereas only 23 per cent of men reported being less content in Ireland. A study on return to Burundi showed that women were less likely to think that migration had improved their social status, increased their ability to contribute to their community or given them more decision-making power within their household (Ruiz, Siegel and Vargas-Silva 2015). These findings are attributed to the different impacts of forced migration on men and women and their positions within their communities and households upon return.

Other qualitative studies have also shown that women experience the readjustment process upon return differently than men do, particularly when returning to patriarchal societies (Ammassari 2009; Wong 2014). In Ghana, Wong (2014) found that highly skilled women who had instigated return with their husbands were more likely to experience divorce upon return due to different gender role expectations for the couple’s life in Ghana compared with those in America. Ammassari (2009) also found that different expectations of gender roles in communities upon return placed additional pressures on both husbands and wives in their return experiences and negatively influenced their perceptions. Finally, Van Houte (2014) showed that in Afghanistan, men were generally more strongly reintegrated than women were and that for women, return meant a move back to a more traditional role within the household.

The limited existing literature indicates that return perceptions differ between men and women. The reasons for these differences, however, are attributed quite differently across studies. This paper contributes to this growing body of research by first examining the case of Ethiopia, a country in which migration flows have been steadily feminized (Kuschminder, Andersson, and Siegel 2012), and then applying a holistic approach that helps to explore the distinct effects of each stage of the migration cycle on perceptions upon return, with a focus on living conditions. We focus on living conditions because returnees perceived changes in living conditions reflect the perceived economic outcomes that are expected to result from migration.

3. An overview of migration and return dynamics in Ethiopia

Since the early 2000s, Ethiopia has been experiencing a dynamic period of economic growth as one of the world’s fastest growing economies (African Development Bank Group 2017). At the same time, Ethiopia remains one of the poorest countries in the world, ranked 173 of 187 on the UNDP Human Development Index (UNDP 2016). Considering this context, this section provides a gender-based comparison with regard to background characteristics (e.g. education, age of first marriage), motivations and patterns of migration and return migration experiences.

In Ethiopia, women are disadvantaged relative to men. Ethiopia ranks 129 out of 154 countries on the gender inequality index (GII) (UNDP 2016) and ranks 89 of 102 countries on the OECD social institutions and gender equality index (OECD 2014). Lower secondary school completion rates are slightly slower for girls (28 per cent) than boys (30.8 per cent), with fewer girls also progressing to secondary school from primary school (96.6 per cent compared with 100 per cent of boys) (World Bank 2017). This has been attributed to higher domestic work responsibilities for girls than boys, leaving girls few opportunities to study and perform in school (Presler Marshall et al. 2016). Poor school results have been found to be a key driver of girls’ migration (Presler Marshall et al. 2016). In Ethiopia, there is also a high prevalence of female child marriage in certain regions (ODI 2014). It has been reported that parents force their daughters to marry young and swiftly divorce to meet cultural norms and to then migrate to the Middle East so that the remittances return to the family and not to the husband’s family (Jones et al. 2014). Emigration from Ethiopia is thus entwined with other social and gender norms.

Some evidence suggests that the drivers of migration from Ethiopia are different for men and women. Although unemployment is a driver for both men and women, large discrepancies in unemployment rates continue to exist between men and women (Kibret 2014). Figure 1 shows the urban unemployment rate by gender in Ethiopia from 2009 to 2012. The fieldwork for the data used in this paper was conducted in 2011, when the urban unemployment rate of women was more than double that of men, at 25 per cent versus 11 per cent. Furthermore, women are more likely to participate in informal labour, with only 41 per cent of women working in the formal sector in 2011 (Broussard and Tekleselassie 2012). It is evident that women experience more employment challenges in Ethiopia than men do. For both men and women, emigration from Ethiopia has increased substantially over the past decade. The clear drivers of these flows are unemployment and poverty.

Figure 1.

Urban unemployment rate by gender, 2009–12. Source: Central Statistics Agency, Urban Employment Unemployment Survey, 2012.

Three central irregular migration routes currently exist from Ethiopia (IOM 2010). The first is to other locations in Africa (IOM 2010). Most commonly, young men migrate to South Africa in search of work (Horwood 2009). The second is migration through Sudan to Libya, frequently with the aim of travelling to Europe (IOM 2010). The third and most common migration flow is from Ethiopia to Middle Eastern countries (IOM 2010). The route by land is across the strait of Aden to Yemen, with the objective of migrating onwards to the Kingdom of Saudi Arabia. Despite the conflict in Yemen and the large-scale deportation of more than 160,000 Ethiopians in 2013–14 from the Kingdom of Saudi Arabia back to Ethiopia, more than 80,000 Ethiopians arrived in Yemen in 2015 (RMMS 2016). The journey to Saudi Arabia by foot is highly dangerous.

The majority of women migrating from Ethiopia go to the Middle East (Kuschminder, Andersson and Siegel 2012). They are able to fly directly and do not have to cross via Yemen (Kuschminder 2014). Emigration from Ethiopia has shown strong feminization. The IS Academy, Migration and Development: A World in Motion project survey found that 60 per cent of all current migrants were women (Kuschminder and Siegel 2014). Migrants’ destinations are also highly gendered; 68 per cent of migrants in the Middle East are female. According to the Ethiopian Ministry of Labour and Social Affairs (MOLSA), 175,427 women migrated legally via private employment agencies from July 2012 to 2013 (cited from Atnafu and Adamek 2015). Research has demonstrated, however, that the majority of this flow occurs through unregistered agencies and irregular networks, indicating that the actual flows are likely much higher (ILO 2011; Fernandez 2013; Kuschminder 2014). A central difference from male migration to the Middle East is that female migration frequently involves far less arduous journeys to the desired destinations. However, upon arrival, women often encounter much more difficult situations then men do in the Middle East.

Conditions in the Middle East for Ethiopian migrants are challenging for both men and women. Men tend to work as guards or daily labourers or on farms (De Regt and Tafesse 2016). The majority of women perform domestic work. A growing body of literature has documented the abuse that female domestic workers face in the Middle East (e.g. Silvey 2006; Fernandez 2010, 2011; ILO 2011; Mahdavi 2011, 2013). At the same time, there is growing evidence of the hardships that Ethiopian men face in the Middle East (de Regt and Taffesse 2016). For both men and women, migration to the Middle East is a high-risk endeavour.

It is therefore unsurprising that return and reintegration can be challenging for Ethiopian migrants, particularly those who may have suffered in the Middle East. In a study of 44 female returnees from Ethiopia, Kuschminder (2014) identified 27 per cent of the sample as vulnerable upon return, meaning that women met one of the following criteria: living in extreme poverty, experiencing social exclusion, or self-identifying as having poor well-being. Although this study is not representative, it is a troubling finding that one in four female returnees from the Middle East is highly vulnerable upon return. Other studies have also identified challenges in women’s return to Ethiopia, including continued unemployment, financial instability, lack of social acceptance, and severe mental health issues (ILO 2011; Atnafu and Adamek 2015). As female emigration and return to the Middle East has become the central migration corridor from Ethiopia, far less research has been conducted on the experiences of male returnees. This paper is the first known study to explicitly compare the differences in perceptions of return outcomes between male and female returnees in Ethiopia.

4. Examining differences in return perceptions

To examine differences in perceptions of living conditions upon return, we examine three sets of independent variables: the migration experience, the return experience, and post-return conditions. These independent variables are rooted in the assumption that the migration cycle and aspirations are important in determining return perceptions. In this paper, we examine whether the different stages have different relationships with the returnees’ perceptions of their living conditions upon return.

4.1 Migration experience

The migration experience is operationalized in this paper through the following five variables: the duration of the migration episode, employment in the destination country, feelings of integration while abroad, the frequency of contact with social networks in Ethiopia while abroad, and the sending of remittances while abroad.

Competing arguments exist regarding the effect of integration on experiences in the origin country after return, which is reflected in the theoretical debate discussed above between neoclassical approaches to return versus the NELM approach. NELM implies that migrants’ socio-economic position in the destination country can positively affect their return experiences and perceptions. Cassarino (2004) viewed integration at the destination as an essential component of migrants’ readiness to return that allows them to save money, make new acquaintances and enhance knowledge, skills and expertise for return. In this respect, perceived social integration and economic integration in the destination country may support a successful return as the final stage of the emigration cycle (Boyd and Grieco 2003). Thus, we hypothesize that returnees who were economically and socially integrated in the destination country will have significantly better perceived living conditions upon return (H1 and H2). We operationalize economic and social integration through the variables of being employed in the destination country (H1) and feeling integrated in the destination country (H2).

In the immigrant integration literature, the most commonly used indicator of integration has been duration of migration (Van Tubergen 2006). When integration is perceived as a process of settling in the destination country, the time component of the concept becomes central to the debate. All else being equal, the integrative aspect of time is strongly accepted in the literature. Although it cannot explain all the variations in different integration patterns, time is positively associated with better social and economic integration in the destination country (Van Tubergen 2006). Regarding return migration, Cassarino (2004) argued that migration should include a sufficient duration abroad to enable returnees to mobilize resources to bring home. To illustrate, Ruiz, Siegel and Vargas-Silva (2015) found that migrants who remained abroad longer had more positive perceptions of their migration experience. Hence, we hypothesize that returnees who have stayed longer in the destination country are significantly more likely to have better perceived living conditions upon return (H3).

Limited research exists on the impact of transnationalism on perceptions of return. However, based on research on integration and transnationalism, it can be argued that the ability to send remittances and maintain connections with social networks in Ethiopia can impact return living conditions. Transnational social contacts cannot be directly associated with an indication of positive migration experiences. Proponents of assimilation theory argue that it is mainly migrants who are marginalized in the destination country who seek social relations with their contacts in the origin country as they are struggling abroad (Bilgili 2015). In this regard, increased transnational social engagement can be associated with negative experiences abroad. However, more recent empirical research in the field has shown that economically and socially better-integrated migrants also maintain strong social contacts with their family and friends in the origin country (Bilgili 2015). Transnational social contacts maintained through telephone calls, e-mails, or Skype help to sustain the close relationship between migrants and their social network in the origin country and assist in the social reintegration process. Sending remittances, whether for altruistic or contractual reasons, also has a strong social component and helps to maintain a positive and closer rapport with the social network in the origin country. Consequently, we argue that returnees who maintain social contact with their origin country while abroad and who are able to send remittances are significantly more likely to have better living conditions upon return (H4 and H5).

4.2 Return experience

Turning to the final stage of the migration cycle, research has demonstrated that the conditions of return form a central element in the reintegration experiences and perceptions of returnees (Gmelch and Gmelch 1995; Cassarino 2014; Kuschminder 2014; Van Houte 2014). We operationalize the return experience through the following two variables: the voluntary or forced nature of the return and whether savings were brought back in the return.

The return episode itself is important in understanding the return conditions. The dichotomy most commonly used to depict the return episode is whether the return is forced or voluntary. Forced return refers to migrants who do not decide on their own volition to return, such as deportees and repatriates. However, it can also be used to include groups with limited agency in their return decisions, such as assisted voluntary returnees who do not have the right to stay in the destination country. There is a significant debate within the literature on the boundaries of voluntary and forced returns regarding degrees of voluntariness and compelled decision making when people have limited options available (Blitz, Sales and Marzano 2005; Webber 2011; Cassarino 2014; van Houte, Siegel and Davids 2014). Therefore, voluntary return includes degrees of voluntariness, from deciding to return entirely upon one’s own free will to agreeing to return to avoid negative consequences, such as forced removal. The degree of voluntariness is not the focus of this article; however, the issue is important in considering these dichotomies. Furthermore, forced returnees are often associated with irregular and vulnerable migrants, whereas voluntary migration is associated with highly skilled migrants or successful migration episodes.

Forced and voluntary returnees have varied return perceptions and experiences. Cassarino (2014) argued that forced return implies an interrupted migration cycle in which migrants are not able to achieve their aspirations. Experiences include the inability to achieve their financial goals and improve their living conditions. Therefore, we hypothesize that forced returnees will have significantly worse perceptions of their living conditions upon return compared with voluntary returnees (H6).

Second, some migrants do not engage in sending remittances while abroad but instead choose to save their earnings and bring them back upon their return. This practice is more common among migrants who have shorter durations abroad. However, migrants who have an unexpected return (often forced) may not be able to bring anything with them on their return, which can be a source of shame for the returnee. Therefore, we hypothesize that migrants who are able to bring financial savings with them on their return will have better perceptions of their living conditions upon return (H7).

4.3 Post-return conditions

In this paper, we use the term ‘post-return conditions’ to conceptualize the experiences of the structural environment in Ethiopia upon return. We examine the following three variables to operationalize post-return conditions: duration since return, employment since return, and transnational social contacts since return.

The indicator most commonly used to assess objective return conditions is employment (Kilic et al. 2009; Kirdar 2009; Alquezar Sabadie et al. 2010). In agrarian societies, the issue of land acquisition, which clearly relates to livelihood, has been found to be one of the most central factors in return (Fransen and Kuschminder 2012). We hypothesize that returnees who were employed or engaged in agriculture after returning have significantly better perceived living conditions upon return (H8).

Second, the duration since return can affect returnees’ perceptions of their living conditions upon return. The initial experiences of return can be quite challenging, and with time, returnees experience an adjustment phase. In a survey of 135 return migrants to a Barbadian village, 53 per cent were so dissatisfied after their first year at home that they believed they would have been happier abroad. However, after three years in Barbados, the level of dissatisfaction dropped to 17 per cent (Gmelch 2004). Individuals’ memories and perceptions are post facto to their migration, meaning that people’s perceptions can be changed based on experiences since the migration episode (Skeldon 1994). Differences in time since return should therefore be accounted for in assessing return perceptions. We hypothesize that returnees who have been in the origin country longer have better perceived living conditions upon return as they are able to rebuild after their migration (H9).

Finally, following from the preceding section, transnational ties can also influence perceptions upon return. Post-return experiences are reinforced by transnational practices that help to enhance the ability to mobilize resources (Kuschminder 2014; Van Meeteren et al. 2014). De Bree et al. (2010) emphasized that transnational practices performed after return also shape post-return experiences. For some, continuous transnational practices, whether social or economic, may be fundamental to establishing post-return belonging. Namely, sustained transnational networks through mobile communication and the Internet can help migrants remain engaged in the globalized world. Kuschminder (2014) found that returnees who maintained transnational ties upon return were more reintegrated upon return. Transnational tie maintenance upon return allows for the continued circulation of ideas, knowledge, and identities formed from potential integration in the country of migration. Thus, we hypothesize that returnees who have maintained transnational ties upon return have better perceived living conditions upon return (H10).

5. Methods and data

This paper is based on the Migration and Development: A World in Motion project data collected in Ethiopia. A household survey was conducted among 1,284 households across five different regions in Ethiopia from March to May 2011. Surveys were completed for three types of households: households that had a member living abroad at the time (migrant household); households that had a member who had lived abroad and returned (return household); and households that had no experience of international migration (non-migrant household). The surveys were conducted in the following five regions of Ethiopia: Amhara, Oromia, Southern Nations Nationalities and People’s Region (SNNPR), Tigray, and Addis Ababa, which together account for 96 per cent of the Ethiopian population. In each region, three different woredas (districts) were selected for sampling, producing 15 data collection sites. The sampling strategy was based on a two-stage approach. First, a listing was conducted at each site to identify different categories of households (migrant, return migrant or non-migrant households). Based on this identification, households were randomly selected for enumeration in each site with an equal proportion of migrant or return migrant households to non-migrant households. The data are not nationally representative and cannot be generalized to represent all Ethiopian migration but provide valuable information on the diversity of experiences in different types of households.

5.1 Sample

The survey was answered by one main respondent per household, who identified at the beginning of the survey whether there were any return migrants in the household. A return migrant was defined as any household member who lived abroad for a minimum of three months and had since returned for a minimum of three months. This definition was chosen to include seasonal migration, which occurs annually for a shorter period, usually three to eight months. Once return migrants in the household were identified, the enumerator arranged an interview with one or more of these individuals to complete the return module of the survey. This module was devoted to the experiences of return migrants during and after their migration episode. The data used in this study are primarily derived from the return module, which was answered by roughly one returnee per household. The final dataset contained 152 return migrants. Due to missing values, this paper provides information on 146 returnees.

5.2 Variables

Table 1 presents the operationalization of dependent, independent and control variables used in the paper. The dependent variable of this study concerns perceived personal living conditions upon return. The respondents were asked the following question: ‘Compared to prior to your migration, would you say your own living conditions have improved or become worse?’ Those who indicated that their living conditions (very much) improved received a score of 1, whereas those who said that their living conditions remained the same or became (much) worse received a score of 0.

Table 1.

Dependent, independent and control variables

Dependent variable
Perceptions of living conditions	Living conditions (very much) improved=1.
	Living conditions stayed the same or became (much) worse=0.
Independent variables
Migration duration	Short-term migration refers to 3 months to 12 months; medium-term migration refers to 13 months to 60 months; and long-term migration refers to 61 months and over.
Employed while abroad	Employed=1. Unemployed=0.
Feeling integrated abroad	This variable has three levels:
	1=not at all; 2=a little bit; 3=very much.
Contact with social network while abroad	This is a continuous eight-level variable ranging from:‘no contact at all’ (1) to ‘having contact every day’ (8).
Remittances sent while abroad	Sent remittances during migration=1.
	Have not sent remittances=0.
Savings brought back	Savings brought back upon return =1.
	Have not brought back savings=0.
Forced return (versus voluntary return)	A proxy for forced return is based on two questions: ‘Who was involved in your decision to return to the origin country?’ and ‘Have you received any support from people or organizations?’
	Those returnees who indicated that the destination country government was involved in the decision and/or supported the migrant were considered forced returnees.
Years since return	Less than 1 year=1.
	Between 1 and 2 years=1.
	Between 3 and 10 years=3.
	More than 10 years=4.
Employment since return	Unemployed=0.
	Inactive (doing housework, permanently disabled, or retired)=1.
	In paid work or self-employed in business=2.
	Self-employed in agriculture=3.
	For this variable, returnees who were currently not employed or inactive were coded as 2 if they had a job at least once since their return
Transnational social contacts since return	No contact=0.
	Transnational contacts maintained=1.
Control variables	Age; marital status; highest level of education; region of return

5.3 Analysis

Given the interest in gendered differences, in the first part of the analysis, we examined descriptive-level differences between male and female returnees with respect to the independent and dependent variables. For the main empirical analysis, we first performed collinearity checks. We only retained variables for which the condition number would not pass the unacceptable level (rule of thumb: CN < 35). We also checked correlations between variables before making final decisions about the models. There was no problematic correlation between the independent variables (rule of thumb: correlation∼ < 0.6). Considering that the dependent variable of the study was binary, we conducted logistic regression analyses. We present the odds ratios. We first ran the model with the control variables, gender and duration of migration. In a second step, we included groups of independent variables separately to observe their unique role in explaining returnees’ perceptions of their living conditions. In a final model, we included all independent variables and control variables to make an overall assessment of the ways in which different factors relate to returnees’ perceptions. Moreover, this approach allowed us to test whether the effect of migration duration is mediated by other migration-related factors.

6. Results

6.1 Descriptive analysis

Looking at the individual level characteristics, we observe that the average age of the respondents was 36 years old, and the men in the sample were younger than the women (see Table 2). The majority was married (56 per cent), but male returnees (73.81 per cent) were more likely to be married than female returnees (34.38 per cent). The highest level of education attained was captured by the following values: 0 = no formal education (23.65 per cent); 1 = primary education (38.51 per cent); and 2 = secondary and tertiary education (37.84 per cent). There was a fairly equal distribution across the levels of education. However, a large share of female returnees (50.00 per cent) had at least secondary education, whereas most male returnees completed primary education (45.24 per cent). Whereas female returnees resided mainly in Addis Ababa (34.38 per cent) and Tigray (25.00 per cent), male returnees were concentrated in the Oromiya (46.43 per cent) and Amhara (22.62 per cent) regions.

Table 2.

Control variables—Background characteristics

	Female		Male		Total
	No.	%	No.	%	No.	%
Age (mean)***	39.01		32.79		36.32
Married*** (versus single)
No	42	66	22	26	64	43
Yes	22	34	62	74	84	57
Highest education level
None	13	20	22	26	35	24
Any primary	19	30	38	45	57	38
Any secondary	32	50.00	24	29	56	38
Region of return
Addis Ababa	22	34	12	14	34	23
Amhara	5	8	19	23	24	16
Oromiya	13	20	39	46	52	35
SNNPR	8	13	1	1	9	6
Tigray	16	25	13	16	29	20
Destination countries
Middle East	50	78	42	50	92	62
Africa	11	17	38	45	49	33
Other	3	5	4	5	7	5

Note:

***

indicates statistically significant mean difference across groups at the 1% level.

Finally, more female returnees (78.13 per cent) migrated to the Middle East than male returnees (50.00 per cent). African countries also represented a major destination for male returnees (45.24 per cent), whereas less than one-fifth of female returnees had migrated to other African countries. The majority of women migrating within Africa were headed to Sudan for domestic work.

With regards to the dependent variable of the study, only slightly more than half of the sample indicated that their personal living conditions had improved compared with prior to their migration (see Table 3). However, there were significant differences between men and women. Whereas only 44.44 per cent of female returnees had a positive perception of their living conditions, more than 62.00 per cent of male returnees believed that their living conditions had improved after returning. In comparison with men, we can perceive a clear pattern of women finding their situation upon return and the effect of their migration experience more negative. The larger proportion of women who find their experiences negative is quite striking and concerning and encourages us to explore the reasons behind this difference between women and men .

Table 3.

Dependent variable

Living conditions improved***	Female		Male		Total
	No.	%	No.	%	No.	%
No	35	56	31	37	66	45
Yes	28	44	52	62	80	55

Note:

***

indicates statistically significant mean difference across groups at the 1% level.

Focusing on the migration experiences of men and women in the sample, we observe that men tend to have shorter migration episodes (see Table 4). Compared with 37 per cent of men who were away for less than one year, only 30 per cent of women resided abroad for between 3 and 12 months. Approximately 46 per cent of women had a migration duration of up to 5 years. Female and male returnees had similar employment status while abroad. Approximately two of every three returnees were employed when abroad. However, a relatively larger share of female returnees ‘very much’ felt integrated when abroad (20.31 per cent). Female and male returnees also had significantly different experiences with respect to their contact with their social network in Ethiopia. Whereas almost 40 per cent of male returnees did not have any contact with family and friends in their origin country, more than 80 per cent of female returnees had contact at least every three months. Similarly, a much larger proportion of women (70.31 per cent) sent remittances while abroad compared with men (45.24 per cent).

Table 4.

Independent variables—Migration experience

	Female		Male		Total
	No.	%	No.	%	No.	%
Migration duration
Short	19	30	31	37	50	34
Medium	30	47.88	33	39	63	42
Long	15	23	20	24	35	24
Employed in destination
No	22	34	29	35	51	34
Yes	42	66	55	65	97	66
Feeling integrated abroad
Not at all	35	55	56	67	91	61.49
A little bit	16	25.00	19	23	35	23.65
Very much	13	20	9	10	22	14.86
Contact with social network while abroad***
No contact	7	11	33	39	40	27
Every few years to every three months	4	7	11	13	15	10
Every three months	20	31	18	22	38	26
Every month or more	33	52	22	26	55	37
Remittances sent while abroad***
No	19	30	46	55	65	44
Yes	45	70	38	45	83	56

Note:

***

indicates statistically significant mean difference across groups at the 1% level.

As shown in Table 5, slightly more than half of women and men brought savings back with them to Ethiopia. A much larger share of men (28.57 per cent) than women (4.69 per cent) were forced returnees.

Table 5.

Independent variables—Return experience

	Female		Male		Total
	No.	%	No.	%	No.	%
Savings brought back
No	32	50	35	42	67	45
Yes	32	50	49	58	81	55
Forced return***
Voluntary return	61	95	60	71	121	82
Forced return	3	5	24	29	27	18

Note:

***

indicates statistically significant mean difference across groups at the 1% level.

Table 6 shows the post-return conditions. Because the feminization of migration is a fairly recent phenomenon in Ethiopia, it is not surprising that a larger share of female returnees returned to Ethiopia within the last year, compared with 40 per cent of male returnees who had been back in Ethiopia for more than 10 years.

Table 6.

Independent variables—Post-return conditions

	Female		Male		Total
	No.	%	No.	%	No.	%
Years since return***
Less than 1 year	20	31	7	8	27	18
Between 1 and 2 years	13	20	16	19	29	20
Between 3 and 10 years	26	41	27	32	53	36
More than 10 years	5	8	34	41	39	26
Employment since return
Unemployed	13	20	4	5	17	12
Inactive	32	50	10	12	42	28
Employed/Own business	19	30	42	50	61	41
Employed in agriculture	0	00.00	28	33	28	19
Transnational social contacts since return
No, no contact	41	64	64	76	105	71
Yes, transnational contacts maintained	23	36	20	24	43	29

Note:

***

indicates statistically significant mean difference across groups at the 1% level.

Regarding employment status since return, although a majority of women were employed while abroad, they were more likely to perform housework once they returned. Only one in every three women was employed since returning. In contrast, half of the men were employed or had their own business. Moreover, one-third of men worked in agriculture. Finally, most returnees appeared to cut off contact with their social network abroad once they returned. Only one-third of the returnees continue to stay in contact with their friends, family and colleagues abroad. However, more female returnees (35.94 per cent) than male returnees (23.81 per cent) appeared to maintain their transnational social ties.

6.2 Main results

This section presents the results of the regression analysis. Table 7 presents the results of the logistic regressions as odds ratios. The first model focuses only on background characteristics, the second model includes variables regarding migration experiences, the third model focuses on return experiences, and the fourth model examines the associations between conditions upon return and perceived living conditions. The fifth and the final model incorporates all independent and control variables to provide a general picture.

Table 7.

Main results of logistic regression results

Living conditions improved	Background characteristics	Migration experiences	Return experiences	Conditions upon return	Total model
	(Model 1)	(Model 2)	(Model 3)	(Model 4)	(Model 5)
Male	2.09 (0.95)	5.42*** (3.33)	2.78** (1.43)	2.33 (1.37)	8.74*** (7.41)
Migration duration
Short		Ref.			Ref.
Medium		3.34** (1.75)			3.14** (1.96)
Long		7.27** (5.84)			8.84** (8.18)
Employed abroad		1.49 (0.78)			1.25 (0.72)
Feeling integrated while abroad
Not at all		Ref.			Ref.
A little bit		2.72* (1.63)			2.55 (1.69)
Very much		7.44*** (5.42)			8.18*** (6.57)
Contact with social network in Ethiopia while abroad		1.03 (0.16)			1.00 (0.02)
Sent remittances while abroad		2.97* (1.83)			2.36 (1.62)
Age	1.00 (0.02)	1.00 (0.02)	0.99 (0.02)	0.99	0.99 (0.02)
Married (vs. single)	2.72** (1.30)	2.66* (1.49)	2.95** (1.47)	3.58**	3.17* (2.05)
Highest level of education
None	Ref.	Ref.	Ref.	Ref.	Ref.
Low	0.64 (0.33)	0.83 (0.50)	0.59 (0.33)	0.47 (0.28)	0.68 (0.46)
Medium/High	1.75 (1.02)	1.98 (1.39)	1.24 (0.77)	0.96 (0.67)	1.43 (1.19)
Region of return
Addis Ababa	Ref.	Ref.	Ref.	Ref.	Ref.
Amhara	0.22** (0.16)	0.10** (0.11)	0.23* (0.18)	0.14** (0.13)	0.08** (0.09)
Oromiya	0.29** (0.17)	0.34 (0.24)	0.33* (0.20)	0.44 (0.28)	0.50 (0.40)
SNNP	0.11** (0.11)	0.15 (0.18)	0.19* (0.19)	0.12** (0.13)	0.26 (0.32)
Tigray	0.13***(0.09)	0.09*** (0.08)	0.20** (0.15)	0.11***(0.09)	0.09** (0.09)
Forced return (vs. voluntary return)			0.31** (0.18)		0.46 (0.33)
Savings brought back			2.89*** (1.15)		2.05 (1.07)
Time since return
Less than 1 year				Ref.	Ref.
Between 1 and 2 years				1.21 (0.90)	2.16 (1.99)
Between 3 and 10 years				0.56 (0.41)	0.87 (0.80)
More than 10 years				1.59 (1.47)	1.23 (1.49)
Employment since return				Ref.	Ref.
Unemployed upon return
Inactive				1.62 (1.54)	0.83 (0.95)
Employed/ Own business				3.05 (2.24)	3.51 (3.04)
Employed in agriculture				3.16* (1.93)	1.55 (1.13)
Current contact with social network abroad				3.16** (1.56)	1.92 (1.08)
Constant	1.40 (1.32)	0.06** (0.09)	1.01 (1.02)	0.66	0.03* (0.06)
Observations	146	146	146	146	146
Pseudo R2	0.136	0.284	0.196	0.215	0.344

Note: Odds ratios are presented.

***

p < 0.01,

**

p < 0.05,

*

p < 0.1 (2-tailed).

First, we observe that across all models, male returnees have significantly more positive perceptions about their own living conditions after their return. In fact, the final model illustrates that men are approximately 8.6 times more likely to think that their living conditions improved compared with prior to migration. Second, model 2 illustrates that migration experiences have a significant effect on how returnees perceive the changes in their living conditions. Returnees who have lived abroad longer, who have felt integrated while abroad and who have managed to send monetary remittances back home are more likely to think that their living conditions have improved. However, being employed while abroad is not significant. This final result is surprising, but its effect may be due to the remittance-sending variable, which is perhaps illustrative of the returnees’ economic well-being while abroad.

As shown in model 3, return experiences are also contributors to a more positive perception upon return. In fact, voluntary returnees are approximately three times more likely than forced returnees to think that their conditions have improved. Moreover, migrants who were able to bring savings back with them are 2.8 times more likely to have positive perceptions. These results indicate that migration and return experiences are determining factors of post-return perceptions. Model 4 shows that having transnational contacts is positively associated with the perception of improvement in living conditions (OR: 3.16, p-2 sided < 0.05). Overall, however, models 4 and 5 indicate that conditions upon return are less influential than migration experiences as time since return, employment status and transnational social ties do not seem to be significantly related to perceptions of living conditions. The results indicate that we can accept hypothesis statements 1–4 and reject the remaining hypothesis statements.

7. Conclusion—Gender and the migration cycle in returnees’ perceptions of living conditions upon return

In this paper, we adopted a NELM perspective that claims that a key goal of migration is to improve the living conditions at home for those who are left behind as well as for the migrants themselves upon their return. Migrants who choose to return home often do so with the intention of having achieved their migration aspirations and being able to return to improved living conditions compared to prior to their migration. However, migrants are often unable to achieve their migration aspirations during their migration experience. Return without achieving migration aspirations may be due to an ‘interrupted’ migration cycle (Cassarino 2014) or a specific pattern of events that leads migrants to perceive a failed migration episode. Return after an unsuccessful migration experience can understandably result in not achieving the goal of improved living conditions.

We acknowledge that there are several limitations in this study. First, the data are not nationally representative and reflect a relatively small sample size. Second, there are causality issues as we assume that the returnees’ migration experience has an effect on their current perceptions. However, perceptions may be based on the omitted variables or relational issues. Further, we do not compare perceptions to objective indicators, which may provide a different story. Third, perceptions may change over time and are retroactive. Fourth, we are unable to account for variations in migrants’ status while abroad—that is, whether they were regularized or irregular, their income earned at the destination, and the role of the destination of the migration episode in influencing perceptions upon return. Finally, we are unable to account for any mental health issues that respondents may have experienced upon return and at the time of the interview. Despite these limitations, this paper contributes to understanding gendered migration perceptions upon return in the under-researched case of Ethiopia.

Returning to the focus of this paper, we examined whether Ethiopian return migrants perceived their living conditions as having improved after their return migration and the factors that influenced their perception of their living conditions upon return. Only 55 per cent of return migrants to Ethiopia felt that their living conditions improved from their migration. This finding was highly gendered, and it is quite concerning that only 45 per cent of female respondents felt that their living conditions improved upon their return. The regression results further demonstrated the significance of the differences between male and female returnees, with male respondents being 8.6 times more likely to perceive better living conditions upon return than their female counterparts.

In the Ethiopian context, this finding is not particularly surprising. Global indices, such as the gender inequality index in which Ethiopia ranks 129 of 145 counties, remind us that women are disadvantaged compared to men in Ethiopia. Research has demonstrated that Ethiopian female domestic workers in the Middle East frequently do not make enough money to improve conditions at home (ILO 2011; Kuschminder 2014). Their work is often vital to maintaining the livelihoods of their families, who are supported by their remittances at home, but the salaries only maintain their families’ livelihoods and do not necessarily improve the conditions of the family (Kuschminder 2014). However, it is quite striking that male respondents are able to perform so much better in their migration episodes that they have significantly higher perceptions of living conditions upon return.

In this study, we sought to understand the conditions that lead to an improved perception of living conditions upon return by further examining the migration experience, return experience, and post-return conditions. As discussed above, the results indicate that the migration experience is more significant than the return experience or post-return conditions in determining perceptions of living conditions upon return. Specifically, returnees who had a longer duration abroad and felt integrated abroad were more likely to perceive better living conditions upon return. These findings can be associated with Cassarino’s (2014) argument for the importance of preparedness and the life cycle in return migration, such as having a longer migration abroad, feeling integrated abroad and sending remittances. We can also infer that the returnees in this study had a more complete migration cycle and were able to achieve their migration aspirations prior to their return. Future studies should examine the extent to which returnees perceived that they accomplished their migration cycle successfully and how such self-evaluations relate to their living conditions upon return.

Other research has indicated that returnees to Ethiopia face several challenges in their reintegration upon return, in large part due to their return experience and the post-return conditions (Kuschminder 2014; De Regt and Tafesse 2016). The lack of significance between the return experience and post-return conditions in the final model is unclear and is contrary to our expectations. This discrepancy highlights the need for further research to understand how the different stages of the migration cycle determine perceptions of living conditions after return, how the stages are interrelated, and how migrants navigate the challenges they face across these stages. In particular, these issues must be explored in a comparative country context with different types of return migrants. Exploring differences between returnees who settle in areas with different levels of urbanization would also be interesting as our analysis illustrated that the returnees in the capital had the most positive perceptions of their living conditions.

Our findings provide evidence of the importance of considering return perceptions within the wider context of the entire migration cycle and that different stages can have different levels of significance. Somewhat surprisingly, the migration experience itself was found to be most significant in understanding returnees’ perceptions of their living conditions upon return. This finding stresses the need for policy and programming interventions to focus on ensuring protection and support for migrants while they are abroad rather than focusing only on reintegration assistance upon return. Policies that inform migrants, particularly women migrating to the Middle East, of their rights pre-departure and provide them with information on how to receive support and assistance (if that assistance is available) abroad, may be more effective in improving living conditions upon return by preventing negative experiences and trauma than reintegration support provided after the fact. Migration experiences have a substantial and enduring effect on returnees’ lives not only during the period when they reside abroad but also when they return.