Gender, remittances and household entrepreneurship: new perspectives from Malawi

Entrepreneurship is often seen as a core area of economic growth and development. This is apparent in the context of small and medium-sized enterprises (SMEs) which make up the majority of employment and income sources in both developing and developed markets (Muriithi, 2017). This has been particularly noted in the case of Africa, where links between economic growth and entrepreneurship have been found, suggesting that African countries are entrepreneurially driven (Ojong et al., 2021; Adusei, 2016). Despite this importance, much of Africa's entrepreneurial activities remain informal with finance gaps that limit their growth and survival (Kansheba, 2020; Choudhury and Goswami, 2019).

Malawi is one of the many developing countries whose economy consists of a large base of SMEs and informal entrepreneurial activities. Formal employment in the country is only at 11%, with much of the unrecognised employment coming from the informal service and agriculture sector, which is the largest employer in the country (Kilic et al., 2023; African Development Bank, 2018). While entrepreneurship is viewed as an important driver of economic growth (Muriithi, 2017), access to finance and credit are often cited as major barriers to SME growth within Malawi (Ndala and Pelser, 2019; Fanta et al., 2017), despite policy interventions by government agencies and NGOs offering support services such as business coaching.

The role of remittances as investments into societies and families have been shown to promote entrepreneurship, alleviate poverty, as well as improve education and housing (Yavuz and Bahadir, 2021; Chrysostome and Nkongolo-Bakenda, 2019; Adams, 2011; Vaaler, 2011). Remittance inflows to the developing world have grown drastically over the past decades from US$56 billion in 1995 to over US$334 billion as of 2010 (UNDP, 2015) and amounted to US$ 529 billion in 2018 (World Bank, 2019). While remittances to Malawi are relatively smaller compared to other African countries, inflows have increased substantially in recent years. In 2010, Malawi's remittance inflows stood at US$22 million, and this figure has now grown to US$264 million as of 2020 (World Bank, 2022). This brings into question their potential role as a source of financing for entrepreneurial activities in Malawi.

Moreso, gender dynamics are of relevance in the context of entrepreneurship as women play a sizeable role in economic growth, with over a third of businesses being female owned (Sajjad et al., 2020; Bardasi et al., 2007). However, the relationship between gender dynamics and entrepreneurship is complex as the many of the barriers that female entrepreneurs face differs in scale and scope. This is often the case with access to financing, ownership structures and business size. For instance, issues of asset ownership may play a role on access to finance with social norms dictating that asset ownership should be male-owned, resulting in female entrepreneurs having limited assets to borrow against. Access to financing is already established as a strong part of why businesses fail, and women entrepreneurs are less likely to receive financing than their male counterparts (Wang et al., 2022; Asiedu et al., 2013). Women-owned businesses are also often smaller in size, with many of these firms being in low-growth sectors that may already saturated (Chaudhuri et al., 2020; Carter and Marlow, 2006). In Malawi, women-owned businesses also tend to be informal microenterprises in nature (Finscope Malawi, 2019; Bardasi et al., 2007) and this raises concerns about their likelihood to grow and succeed beyond a start-up phase. Given the significant inflows of remittances in Malawi, and their ability to promote entrepreneurship, there is a need to understand how remittances can close the financing gap for female entrepreneurs and become a context-specific solution. As the role of remittances has often varied and offered mixed results (Nanyiti and Sseruyange, 2022; Yavuz and Bahadir, 2021; Vasco, 2013; Amuedo-Dorantes and Pozo, 2006), a contextual analysis is therefore required (Naudé et al., 2017) to understand how remittances can be used a source of financing for entrepreneurial activities in Malawi, particularly in case of women-owned business. With a clearer picture of how remittances may be operating within the African household context, this paper aims to advance both academic discourse and policies on how to make optimal use of remittance flows and inform the policy measures that need to be taken to ease the constraints female entrepreneurs face.

While the role of remittances on development has been studied in the context of developing economies ^[1], this has not extensively been studied in the case of entrepreneurial activities in Malawi, particularly at a household level. This has to do with a multitude of reasons, including the lack of availability of data to study these dynamics (Adams, 2011). The existing studies on Africa have also been more inclined to focus on outcomes such as poverty alleviation, health and education in relation to remittances (Adams and Cuecuecha, 2013) while others that have taken into account entrepreneurial activities often do so in the context of return migration (Thomas and Inkpen, 2013; Marchetta, 2012; Wahba and Zenou, 2012), which do not account for entrepreneurial activity necessitated by remittances. Household studies can be particularly important in this regard, as they may take into account the informal economy (Elgin and Erturk, 2019), which constitutes a large share of Africa's economic activity yet is often overlooked in other data sources. With limited research at an African level, there has also been less evidence looking at the gender dynamics of remittances and entrepreneurship. Studies on gender dynamics and entrepreneurship in Africa, have focused on the role of access to finance as formal credit (Ojong et al., 2021), and while these are often related to issues of size and access to collateral (Abor and Biekpe, 2006; Isaga, 2019), there are limited examples that explore the role of alternative sources of finance, including remittances.

This paper extends existing studies (Banda, 2020; Kangmennaang et al., 2018; Thomas and Inkpen, 2013) in several ways. Firstly, this paper looks to extend its focus beyond just rural communities, using a larger sample and a quantitative method, which differs from the approach of Banda (2020) who qualitatively focused on rural communities in one district of Malawi. Kangmennaang et al. (2018) assessed a study on 1,000 households in northern and central Malawi, emphasising the relationship between remittance receipt on food security and household asset accumulation but did not account for the role of remittances on household entrepreneurship. By focusing on a national sample, this study will give a more representative outlook on the role of remittances, with a particular focus on entrepreneurial activities, as opposed to the role of asset accumulation and poverty alleviation. Secondly, Thomas and Inkpen (2013) focused predominantly on return migration and its interaction with self-employment. While a relationship with remittances is present, this study places a greater emphasis on remittances as a whole and its interactions with gender. This research allows for an understanding of female entrepreneurship when provided with access to capital in the form of remittances. This is particularly important as female entrepreneurs often face a myriad of context specific barriers (Isaga, 2019), despite making up half of the entrepreneurs in Sub-Saharan Africa (De Vita et al., 2014) and being the fastest growing cohort of entrepreneurs globally (Woldesenbet Beta et al., 2024).

While macroeconomic research may show the effects of remittances in terms of financial development and its effect on pushing economic growth (Mohammed, 2022; Gupta et al., 2009), household data can allow researchers to investigate individual interactions with remittances using a more nuanced approach. The analysis of recent survey data may provide important insights into how remittances may be operating for entrepreneurial purposes and what more can be done to better leverage these inflows for sustainable development in emerging economies.

On this premise, the rest of the paper is organised as follows: Section 2 presents the theoretical development pertaining to this research and undertakes a review of the empirical literature, while Section 3 outlines the empirical strategy employed in addressing the research question. Sections 4 and 5 discuss the results of the empirical analysis and policy recommendations, respectively.

Applying a gendered view of remittances and entrepreneurship is not entirely new, as research from this area suggests that gender and social norms play a part in the decisions to send and receive remittances (Mahmud, 2020; Pickbourn, 2016), which could in turn influence the social preference of remittance use. Yet, research on how gender and social norms interact and the extent to which these social norms affect women's entrepreneurial activities has so far been “fragmented and scattered across multiple theoretical platforms” (Bullough et al., 2022). On one hand, neo-classical entrepreneurship theory draws on the contingency theory to explain how entrepreneurs identify and exploit entrepreneurial opportunities to create wealth, but decision-making, wealth creation and resource allocation is contingent on the internal and external situation (Childs et al., 2022; Hitt et al., 2001). Resource allocation needs to adopt a gendered perspective in examining the internal situation of entrepreneurs as assets are often in the male spouse's name, thus restricting the ability of female entrepreneurs to acquire capital. From an external point of view, gendered differences in access to the acquisition of start-up capital also exist because the pay gap contributes towards lower business loan acceptance rates for women compared to men (Greer and Greene, 2003). On the other hand, entrepreneurial feminism (Orser et al., 2013, p. 242) views women as “change agents who exemplify entrepreneurial acumen in the creation of equity-based outcomes that improve women's quality of life and well-being through innovative products, services and processes”.

While we agree that entrepreneurial activity is based on a needs analysis conducted by the entrepreneur and is driven by multiple factors, such as internal and external contexts, social norms and resource possession or depravation, we also appreciate that achieving entrepreneurial outcomes varies on the basis of gender as men and women have different social and professional networks, labour market participation rates, and access to business capital, which is crucial for fostering innovative products, services and processes. As such, we approach this theoretical knowledge gap by applying the liberal feminist theory, more specifically the sub-domain of equality feminism as a framework that engages the role of gender on the relationship between remittances and entrepreneurial activities. Moreso, we apply this sub-domain of equality feminism as it regards men and women as equals and states that the solution to women's limited social and economic achievements is to remove the barriers that restrict women's economic participation. Furthermore, equality feminism advocates that access to opportunities should neither be gendered, nor constrained on the basis of sex. In removing barriers to opportunities, substantive equality often means establishing opportunities in areas where they do not currently exist (Nanni, 2023).

Greer and Greene (2003) go on to argue that the entrepreneurial literature that considers gender often focuses on the extent to which women's experiences also shape their entrepreneurial behaviours. Equality feminism therefore provides an appropriate lens to view the role of remittances on entrepreneurship from a gendered perspective as it endorses the creation of opportunities in breaking down barriers to economic activities such as entrepreneurship. By applying the theory of equality feminism, we gain an avenue to acknowledge that all people, irrespective of gender, can contribute to the development of society. When social and economic opportunities are available and accessible to women without any discrimination or bias, the level of participation by women in entrepreneurship increases. As remittances act as a way to relax household budget constraints, a gendered view of remittance receipt can allow for the assessment of male and female decisions to engage in entrepreneurial activities. This is particularly relevant when considering how remittances may change the allocation of household expenditure according to the preferences of the remittance receiver and their corresponding bargaining power (Duflo and Udry, 2004; Göbel, 2013). As such, we hypothesis that:

As migration and remittances are closely linked as a subject area, Thomas and Inkpen (2013) assessed the likelihood of self-employment amongst migrants returning to Malawi using census data. The study highlighted that migrants from Western countries were less likely to engage in self-employment in comparison to those from African countries and within the Southern African Development Community (SADC). Despite that remittances were not central to their inquiry; they were highlighted, and their research identified that remittance receipt is associated with self-employment in non-agricultural sectors. Past studies on remittances within the African context have also examined the role of remittances in promoting development and attaining the Sustainable Development Goals (SDGs) in Africa (Akanle et al., 2022) or assessed the extent to which remittances can be source of economic growth across 36 African countries (Nyamongo et al., 2012). Azam and Gubert (2006) examined how remittances buttress consumption in times of adverse shocks or vulnerability for families in Mali and Senegal, while Anyanwu and Erhijakpor (2010) determined the impact of remittances on poverty reduction in 33 African countries.

Several studies on remittances are also located outside of Africa, as studies on the role of remittances in Central and South America establish the use of remittances for household investment or consumption but not for entrepreneurial investment. For instance, Amuedo-Dorantes and Pozo (2006) studied the impact of remittances on business ownership in the Dominican Republic and discovered that remittance recipients were less likely to own businesses compared to those that did not. The authors attributed this to the income effect of remittance receipt leading to the purchase of leisure goods or alternative investments in education and housing. Vasco (2013) also assessed the role of remittances and migration on business ownership in rural Ecuador which showed that neither migration nor remittances were associated with being self-employed. This is largely due to Ecuadorians preference for channelling remittances into less risky assets, such as land, and their reluctance to engage in entrepreneurship following recent economic turmoil.

In trying to understand contextual factors surrounding remittances and entrepreneurship, Yavuz and Bahadir (2021) found a positive link between migrant remittances and new business creation from the perspective of ethnic diversity. While insightful, their findings potentially underestimated the effects of remittances due to limited data on the informal economy. Remittances may also have varying effects in the context of gender. As noted by Acosta (2007) in El Salvador, remittance receipt was associated with more entrepreneurship involvement among females, compared to those that did not receive any remittances. However, Acosta's definition of entrepreneurship, which includes managing a business, creates ambiguity between those managing their own enterprises and those managing on behalf of others.

From the empirical literature, it is also clear that due to the mixed nature of research, contextual factors still need to be used to fully understand how remittances operate for entrepreneurial purposes and can be leveraged, if at all. Naudé et al. (2017) points out that forming a better understanding of remittance use requires an assessment of their origins, where they are being directed, and what their intended use is to better understand the interactions between remittances and entrepreneurship.

This paper therefore seeks to extend the body of knowledge by specifically assessing remittance and entrepreneurial activities using Malawian household data. This may help in giving more insight in how remittances may function within the SADC region and build on this sub-sector of research from an African perspective by using Malawi as a case in point. By focusing on the gender dynamics of remittance receipt and entrepreneurial activity, this paper also seeks to understand whether remittances act as a positive catalyst for entrepreneurship against the backdrop of existing gender gaps and limited financing opportunities for female entrepreneurs who often face higher barriers in accessing finance.

This research uses secondary cross-sectional data from the Malawi Fifth Integrated Household Survey, collected between 2019 and 2020 (National Statistical Office, 2020). This nationally representative survey is conducted by the Malawi Government in collaboration with the National Statistical Office, with technical assistance from the World Bank and Food Policy Research Institute. The survey is conducted every 3–5 years and typically covers 12,288 households across Malawi. However, due to Covid-19, the most recent survey only covers 11,434 households, representing the opinions of 50,476 respondents, which is still considered nationally representative. This study focuses on respondents who completed the remittance section of the survey on Children Living Elsewhere, resulting in a sample of 20,547 respondents. The survey employs a stratified two-stage sample design which identifies respondents listed in the 2018 national census and covers all national districts. Other datasets were considered, including Malawi Census Data and the FinScope Survey, which consists of financial inclusion data. However, the Household Survey was chosen due to the variety of variables studied, including the gender of the household head as a consideration when assessing the role of remittances on female entrepreneurship. Given that equality feminism is applied as the theoretical lens through which this research is applied, this consideration aligns with the theory's assertions that the solution to women's limited social and economic achievements is the promotion of opportunities in areas where they do not currently exist (Nanni, 2023), like the use of remittances in driving female entrepreneurship. Furthermore, the Household Survey was selected due to being the most up-to-date publicly available dataset for the research aims of this study.

This section discusses the variables employed in the empirical analysis. Table 1 provides a summary of the variables, including definitions, expected signs, and basic descriptive statistics. These statistics were calculated using household-level data and are consistent with the definitions described in this section.

1. Dependent variable: Household Entrepreneurship

In a number of studies, entrepreneurship has been measured as self-employment (Finkelstein Shapiro and Mandelman, 2016; Yang, 2008). The decision to engage in entrepreneurship is usually one that involves taking a risk to receive or obtain a profit (Smith and Chimucheka, 2014). In this study, entrepreneurship is measured by assessing whether anyone in a household is engaged in any form of household entrepreneurial activity, which ranges across agricultural businesses, service-related business, transportation, rendering of a service, the sale of a product as a professional or an informal service. This expanded definition of entrepreneurship accounts for a number of activities that do not require formal business registration.

1. Independent variable: Remittances

In this research, the independent variable is measured by assessing whether a household has access to remittances (also referred to as remittance receipt). A dummy variable that accounts for access to remittances is included from the survey to capture whether the household receives income from children living elsewhere. As a central variable within this investigation, it is expected that remittance receipt will be associated with a higher likelihood of being engaged in entrepreneurial activities as this can act as a source of financing for business activities.

1. Control variables

   • Household size: refers to the number of people in a household. It is expected that larger households will be more likely to engage in entrepreneurial activities, as larger household may offer more support, mentoring and more household income (if there are more household income earners) as well as risk sharing and cost reduction opportunities (Pittino et al., 2020).

   • Age: refers to the age of the respondents. It is expected that older household heads are more likely to engage in entrepreneurial activities, given that the ability to accumulate both financial and non-financial resources increases with age (Démurger and Xu, 2011).

   • Gender: The role of gender seems to indicate that females are less likely to engage in entrepreneurial activities as a result of lack of access to credit and start-up capital (Brixiová and Kangoye, 2016). However, some evidence does indicate that when remittances are involved, females are more likely to engage in entrepreneurial activities than those without access to remittances (Acosta, 2007). Female-headed households that receive remittances may be more inclined to engage in entrepreneurial activities, though the extent of this relationship remains uncertain.

   • Marital status: Marital status has an effect on entrepreneurship as evidence by Démurger and Xu (2011) who indicated that married couples are more likely to be engaged in self-employment, as work can be shared. It is therefore expected that households headed by a married couple are more likely to engage in entrepreneurial activities.

   • Education: Education has been found to be closely linked with the likelihood of engaging in entrepreneurial activities (Marchetta, 2012), while this may not be the case with tertiary education as individuals with higher education levels are more likely to be formally employed (Thomas and Inkpen, 2013). Categorical variables for varying education levels are included in the model and it is expected that having secondary education will be associated with greater entrepreneurial activities, while tertiary education may present an ambiguous effect on entrepreneurial activities.

   • Access to credit: Despite that remittances can help ease liquidity constraints, overall access to credit can also allow for households to invest in as well as start entrepreneurial activities (Duflo et al., 2013; Vasco, 2013). We therefore expect that access to credit is associated with a greater likelihood of engaging in entrepreneurial activities.

Variables for access to infrastructure have also been included as poor infrastructure limits investments into entrepreneurial initiatives.

• Access to water and electricity: Access to infrastructure has also been included as poor infrastructure limits investments into productive assets. To account for this, dummy variables for piped water and electricity access are included in our analysis (Vasco, 2013). It is expected that increased infrastructure will be associated with a higher chance of being involved in entrepreneurial activities.

• Internet access: A dummy variable for internet access is included as internet access acts as a means of reducing information asymmetry within business environments and can assist in the reduction of informational rents for entrepreneurial activities (Asongu and Odhiambo, 2020). Access to internet can also act as a means for entrepreneurs to advertise and sell their services through social media, which acts as another reason for the inclusion of this variable. It is expected that that increased internet penetration or usage will be associated with a higher chance of engaging in entrepreneurial activities.

• Urban household: A dummy variable for urban household classification has been included as urban households have been found to engage in more entrepreneurial activities than rural households (Hamdouch and Wahba, 2015). This is likely the case as urban residences have greater access to resources and financial services to undertake entrepreneurial activities (Kakhkharov, 2019). It is expected that urban residences will be associated with a higher likelihood of engaging in entrepreneurial activities.

This paper uses a probit model to estimate the effect of remittances on household entrepreneurship. Because the key variables are binary, a linear probability model would be inappropriate, as it can produce predicted probabilities outside the [0,1] range and often violates OLS assumptions such as homoskedasticity. The probit model, based on the cumulative distribution function of the standard normal distribution, ensures predicted probabilities remain valid and statistically consistent.

A linear probability model would normally expect that:

With Y being a binary choice model, $b$_i cannot be interpreted through the use of marginal effects. The main model specification to examine the effect of remittances on households' decision to engage in entrepreneurial activities (Entrepreneur_i) can be described using the following Probit model:

Where Remittances_i is the dummy variable for the receipt of remittances from children living outside of the household, and X_i is a vector for household characteristics such as age, gender, and education. C_i is a vector for contextual infrastructure related variables such as water and electricity, and internet access for household infrastructure. This vector also includes a classification for whether a household is within a rural or urban area. The error term is modelled by ε_i.

Female-headed households will also be looked at using a similar model:

Whereby, FemaleHousehold_i represents a part of the interaction term between remittance receipt and female-headed households. Remittance receipt is likely endogenous in the entrepreneurship decision due to two main reasons. First, reverse causality may exist: entrepreneurial households may be more likely to attract financial support from migrant family members, either to sustain or expand their businesses. Second, omitted variables such as unobserved household risk preferences, access to informal credit, or social capital may influence both remittance receipt and entrepreneurship simultaneously. As these confounding factors are not directly observable in the dataset, the receipt of remittances may be correlated with the error term in the entrepreneurship equation, leading to biased estimates if not properly addressed. Due to the potential endogeneity of remittance receipt on household entrepreneurship, two estimation choices have been made to help address this. Firstly, remittance receipt will have to be jointly estimated using simultaneous Probit models. This will mean remittance receipt is modelled as:

Where X_i is a vector for household characteristics. P_i is a vector for contextual infrastructure related variables as well as the urban-rural split.

Secondly, these estimations will be made with the inclusion of an instrumental variable (i.e. bank account access). This variable is chosen to address the endogeneity of remittance receipt for two main reasons relating to relevance and exclusion restriction:

Regarding relevance, access to a bank account is often a necessary condition for receiving formal remittances, especially from abroad. Formal transfer methods such as international money transfer operators or wire transfers require a recipient to have a bank or mobile money account. Prior literature (e.g. Aggarwal et al., 2011; Demirgüç-Kunt et al., 2018) supports the strong relationship between financial access and the likelihood of remittance inflow.

In terms of exclusion restriction, although bank access might have potential indirect links to entrepreneurship, in our study context—largely rural or underserved areas—bank accounts are primarily used for receiving funds and basic transactions. They are not the primary source of business capital or entrepreneurship inputs. Furthermore, our model controls for access to infrastructure (i.e. electricity, internet, urban/rural location), isolating the instrument's effect from confounding pathways. Therefore, bank access affects entrepreneurship only through its impact on remittance receipt.

We acknowledge that banking access may be lower in rural areas. However, this heterogeneity is precisely what enables its use as an instrument. Within rural communities, some households still maintain access via mobile banking services, cooperatives, or rural banking branches. This within-group variation supports identification, and rural/urban classification is controlled for in all specifications to mitigate location-based confounding.

In essence, the model will be considered in this way where y₁ is the likelihood to engage in entrepreneurship and y₂ is the receipt of remittances. While a number of exogenous factors have been identified as to whether a household engages in entrepreneurship, as mentioned, the decision to send remittances itself may be endogenous to the likelihood of engaging in entrepreneurship. This results in our basic model being as follows:

As a result, the realised model becomes:

Where the indicator function becomes one [1] when the conditions are true and zero otherwise. x = [x₁ x₂] is made up of all the exogenous variables. Equation (3) consists of a reduced form equation for y₂ equal to 1 when remittances are received by the family. If remittance receipt is correlated with the error term, the Probit Maximum Likelihood Estimation (MLE) of Equation (5) becomes inconsistent, as a result, Equations (5) and (6) will need to be jointly estimated. As these two dependent variables are binomial by nature, a bivariate Probit model specification acts as a control for unobserved heterogeneity and endogeneity, making it a suitable estimation approach for this research question. An instrumental variable with a test of endogeneity is also conducted to address the issue of endogeneity.

Table 2 highlights the descriptive statistics of the variables within the dataset for respondents that completed the questionnaire's remittance module on children living elsewhere. The results show that 37% of households engage in household entrepreneurship, with only 16.6% receiving remittances. The distribution of household size indicates that 49% of the households in the sample have more than 6 people residing in a household. The distribution of age shows that most of the household members (71.2%) are between the ages of 18 and 35 years, and most of the respondents (69.3%) are married. On average, 65.9% of the households are male-headed, while 34.1% of the households are female-headed. The frequency distribution of the education level of the sample shows that 88.77% of the respondents were found to have no formal education. Furthermore, only 28.1% of the households have access to credit. The infrastructure variables indicate that only 8.4% of the household have access to piped water, with 9.7% having access to electricity and 18.1% of the households having access to internet. On average 87.2% of the households are in rural areas.

The results of the regression analysis estimated using MLE using a simultaneous bivariate Probit estimation technique are presented in Table 3. The discussion of the results is done under two conditions, first based on the significance of the estimated coefficients followed by the interpretation of the predicted probabilities. The results are split into two sections with section one assessing the whole sample (i.e. Table 3), whereas section two undertakes a gendered analysis with the inclusion of the remittances and gender interaction term (see Table 4). Due to some missing data points in the education variable – the sample size for estimation was reduced to 18,011 respondents. The first column of each table focuses on the likelihood of receiving remittances, whereas column 2 focuses on the likelihood of engaging in entrepreneurship. Predicted probabilities are looked at in the third column, allowing for the assessment of marginal effects for binary regressors. A Wu-Hausman test for endogeneity was performed using an Ordinary Least Squares (OLS) procedure, which is included in each table. The Wu-Hausman tests indicated that remittance receipt is endogenous (p < 0.05), validating the need for an instrumental variable approach. This provides justification for instrumenting remittances with bank account access. While the main estimation is carried out using a bivariate Probit model, the Wu-Hausman result provides initial evidence that instrumenting remittances is necessary.

The determinants of remittances in Table 3 (Column 1) indicate that female-headed households have a higher probability of receiving remittances, which is significant to a 1% level. When assessing access to education, only primary education was found to increase the probability of receiving remittances at the 5% level. Access to electricity was also found to increase the likelihood of receiving remittances at the 1% level and urban households were found to decrease the likelihood of receiving remittances at the 5% level. In contrast, access to household credit was found to decrease the likelihood of receiving remittances. Household age and size were also found to increase the likelihood to receiving remittances at the 1% level.

In the first model (Table 3), remittances were found to have a positive effect on the likelihood of engaging in entrepreneurial activities and this was significant to the 5% level with a predicted probability of 0.584. This falls in line with key studies and provides strong support for the effects of remittances on entrepreneurship (Yavuz and Bahadir, 2021; Banda, 2020; Woodruff and Zenteno, 2007). Education level did not have a significant effect on the likelihood to engage in entrepreneurship, which is inconsistent with Marchetta (2012) who determined that education is closely linked with the likelihood of engaging in entrepreneurial activities. On the other hand, the positive and significant relationship between household credit and the likelihood to engage in entrepreneurship is present at the 1% level, with a predicted probability of 0.451. This falls in line with the literature suggesting that access to credit can be a strong source of helping to ease liquidity constraints (Duflo et al., 2013; Vasco, 2013) and consequently engage in household entrepreneurship. Access to piped water increased the likelihood of engaging in entrepreneurship at the 1% level with a predicted probability of 0.460. This also re-emphasises the role of access to basic resources to pursue entrepreneurial activities. Contextual variables such as being in an urban household indicated a positive relationship with the likelihood of engaging in entrepreneurial activities at the 1% level. This was found to have a predictive probability of 0.518, consistent with the expectation that urban presence provides access to resources conducive to entrepreneurial activitiy. This is likely due to access to resources that will allow for these entrepreneurial activities (Hamdouch and Wahba, 2015). Alternatively, the pursuit of household entrepreneurship is also likely to be factored in by the presence of high urban unemployment (Thomas and Inkpen, 2013). On the other hand, average age was negatively related to the likelihood of engaging in entrepreneurial activity at the 1% level. This goes against hypothesised expectations, as higher household age was expected to lead to an increased likelihood of engaging in entrepreneurship. However, this is likely a result of Malawi's youthful population, which is also represented in the sample. In the context of higher rates of youth unemployment, it is plausible that younger people are more likely to engage in entrepreneurial activities as a means of earning an income.

To understand whether remittances may have differing effects on gendered household entrepreneurship, the following section focuses specifically on female-headed households. This involves the re-estimation of the results in Table 3, which includes an interaction term for remittance receipt and having a female head of the house (see Table 4). Consistent with the results in Table 3, female-headed households are observed to be associated with a higher likelihood of receiving remittances in the selection equation. Similar to the basic estimations in Table 3, female-headed households are less likely to engage in entrepreneurial activities with a predictive probability of 0.309 at 1% level of significance. The interaction between remittances and female-headed household (Remittances × Female HH) is observed to be positive which suggests female-headed households, with access to remittances, have higher likelihood of engaging in entrepreneurial activities compared to male-headed households with a predictive probability of 0.546 at 1% significance level. A possible explanation of this result is that female-headed households that receive remittances may have more bargaining power in the context of how remittances are spent (Pickbourn, 2016), and can therefore channel remittances towards entrepreneurial activities. This may also relate to differences in household spending priorities. Female-headed households are more likely to allocate remittances toward household well-being, education, and investment (Chant, 2007; Doss, 2013). When such priorities include small-scale enterprises, remittances may become an enabling factor. In contrast, male-headed households may prioritize different forms of expenditure, including consumption or asset accumulation. Moreover, given the higher economic vulnerability of female-headed households in Malawi (FAO, 2011), the receipt of remittances may be a more decisive factor in enabling entrepreneurship out of necessity, rather than choice. In this light, remittances function not only as capital but also as a catalyst for female economic agency particularly when no alternative income sources exist. This finding is consistent with equality feminism, whereby remittance receipts by female-headed households help remove the social and economic barriers that restrict women's economic participation.

Other changes were also observed across the two regression results, with one notable change in the context of household credit. The likelihood of receiving remittances in a household with access to credit changed from negative and significant (in Table 3) to positive and significant (in Table 4). However, when assessing the likelihood of engaging in household entrepreneurship, this changes from being positive and significant to being negative and insignificant in the gendered analysis. There are several factors that may be at play to explain this change, such as the form of credit being afforded to female households and the occurrence of gendered poverty. On the one hand, access to microfinance services has been found to increase the likelihood of engaging in entrepreneurship (Ngono, 2021), while on the other hand, poorer households that receive microfinance are much less likely to benefit from it as they often channel the financing to other commitments such as food and schooling (Ukanwa et al., 2018).

There were also some changes in the context of electricity access and access to piped water. Under the likelihood of receiving remittances, access to electricity changed from being significant in the first regression analysis to being insignificant in the second regression analysis. Furthermore, households with access to piped water were unlikely to receive remittances (see Table 3) yet, the gendered analysis found the relationship between access to piped water and the likelihood of receiving remittances to be positive and insignificant. In contrast, when examining the likelihood of engaging in entrepreneurship, access to piped water shifted from positive and significant in the first regression analysis, to negative and insignificant in the second. Given that these are both variables tied to infrastructure access, the lack of access to basic infrastructure can be viewed as a barrier to entrepreneurial activity which becomes more pronounced in female-headed households.

In  Appendix 2 and  3, two additional regression analyses were conducted to evaluate the impact of internet access on entrepreneurial activity (with the whole sample and the gendered results respectively). The internet variable was not included under the main empirical investigation as it would have led to a significant reduction in the sample size, as numerous data points were missing. Furthermore, this reduction in sample size could have influenced the instrumental variable, which was not identified as a strong instrument in this particular case. The findings of the secondary analysis revealed a positive relationship between internet access and entrepreneurial activity ( Appendix 2) while indicating a negative relationship between overall household entrepreneurship and remittance receipt ( Appendix 3).

This study aimed to assess the effect of remittances on the likelihood of female-headed households engaging in entrepreneurship in Malawi using the Fifth Integrated Household Survey 2019–2020 over a sample of 18,011 respondents. In the context of a country where most businesses consist of microenterprises that often struggle to access capital, this study considered remittances as an alternative source of foreign investment capable of providing much-needed funding. It further assessed this effect within female-headed households to understand whether distinct dynamics emerge when incorporating a gendered perspective. As such, the study applied the liberal feminist theory, more specifically the sub-domain of equality feminism, as an appropriate lens to view the role of remittances on entrepreneurship from a gendered perspective. The data was analysed using a bivariate Probit regression in two models. The first regression analysis assessed the likelihood of households receiving remittances and the likelihood of engaging in entrepreneurial activity across the whole sample. The second regression analysis included a gendered interaction term to determine whether female-headed households were more likely to receive remittances and engage in entrepreneurial activity.

While entrepreneurship is a key means of economic development in Malawi, only 37.7% of the households analysed were engaged in entrepreneurial activities, with 16.6% of the households analysed receiving household remittances. The empirical results indicate that remittance receipt has a positive and significant effect on entrepreneurial activity, while female-headed households are associated with lower likelihood of engaging in such activity. On the contrary, female-headed households who receive remittances were observed to be associated with higher probability of engaging in entrepreneurial activity. This highlights the effect of remittances in removing the social and economic barriers that restrict women's economic participation.

However, the findings must be interpreted within the context of the sample's rural composition, where 87.2% of households are rural. This rural skew raises valid concerns regarding the feasibility of entrepreneurship for rural women, who often face compounded barriers such as limited infrastructure, poor transport networks, unreliable electricity, and restricted access to financial institutions. While rural households in the sample were more likely to receive remittances, the ability to convert these into sustainable entrepreneurial ventures may be constrained by these structural deficits. This observation does not contradict the positive effect observed; rather, it reflects the complex interplay between remittance receipt and the enabling environment. The gendered effect of remittances remains significant, but its potential may be hampered by rural underdevelopment a factor future policy must explicitly address.

Based on these findings, we further the discourse on the role of remittances on the entrepreneurial activities of female-headed households where past studies of Ajide and Osinubi (2022) only reported the complementary role of remittances, along with foreign aid, in improving the level of entrepreneurial development in Africa. Moreso, studies that have established the link between remittances and conducting business in Sub-Saharan African countries (Asongu et al., 2019) did not consider the gendered effect of remittances on entrepreneurial activities. Therefore, this study made an important theoretical contribution towards equality feminism in uncovering that remittances are an economic resources female entrepreneurs draw upon to advance their economic position and engagement when they have agency over the use of remittances. From this perspective, the spending choices and bargaining power within households could offer further insights on the use of remittances for entrepreneurship, particularly for female-owned businesses. Policy makers should consider the intersection between gender, remittance receipts, and economic development when considering interventions to improve entrepreneurial activities within developing economies.

Looking more closely at the gendered dynamics, one of the key issues relates to the need for stronger support for female entrepreneurs in Malawi from a policy perspective. This is particularly the case when considering that the composition of female-owned businesses is often the smallest. A necessary intervention could be in the form of more targeted funding and training programmes for existing female entrepreneurs. In this way, explicit capital is directed to female-owned businesses and ancillary support to ensure these businesses are more sustainable. For rural women in particular, practical solutions must address the systemic barriers to productive entrepreneurship. Beyond providing better underlying infrastructure, there is a need for improving rural financial literacy, expanding access to mobile money platforms to facilitate easier receipt and use of remittances, and promoting community-based financing mechanisms such as Village Savings and Loans Associations (VSLAs), which can act as informal credit markets. Some evidence of this has already been found, as is the case in the study by Kaumba et al. (2023), which highlights the positive impact of VSLAs on the performance of micro and small enterprises in Malawi, suggesting their potential as effective tools for enhancing rural women's entrepreneurial capacity. These interventions can empower rural women to better translate remittance inflows into viable business activities.

While the government has created a Ministry of Gender, Community Development and Social Welfare (MoGCDSW), there are limited examples of policies that improve female entrepreneurial outcomes, especially in the context of funding. For example, existing policies from the Malawi Government (2015), such as the National Gender Policy make specific mention of creating a women's fund to support women and other vulnerable groups in business as well as providing capacity building in entrepreneurship but there are limited examples of government driven initiatives doing this. Most of the current policies tend to lean towards poverty alleviation strategies, despite the disproportionate effects of poverty on female households and entrepreneurs. Should a proportion of government funding be intentionally directed towards female entrepreneurs, this would address the need for poverty alleviation, while simultaneously closing the gender funding gap in female-owned businesses, given that female entrepreneurs are fast growing in numbers and the importance of businesses in economic growth. One way in which government can assist female owned businesses is by implementing existing policies such as the Public Procurement Act (2003) of the Malawi Government (2003) in a more targeted manner. The Act makes specific mention to the state's commitment towards providing opportunities for small and medium enterprises through public service contracts but is not legislated in such a way that seeks out specific marginalised groups. Other existing tools such as the Micro and Small Enterprise Policy (2019) and Malawi Vision 2063 (2020) all make mention of the intention to improve entrepreneurial outcomes for SMEs, but they are often not clear strategies to undertake these changes, particularly for female entrepreneurs.

The research focuses on remittances from the use of national household data, and while this is a rich source of data, more could be learnt from designing a remittances survey that specifically addresses this topic as gaps in the data may limit the generalisability of the results. From a design perspective, the cross-sectional nature of this study limits the establishment of causality between remittance receipt and entrepreneurial activity. Another key factor is data access. On one hand, the survey's interruption during COVID-19 limited the sample size, which was further affected by missing data points for variables (such as internet access), thus preventing their inclusion in the main regression analysis. While the chosen estimation technique and use of instrumentation sought to mitigate endogeneity, some endogeneity could still be present in the results. From a methodological perspective, remittances were only considered from the perspective of children living elsewhere that sent money to their families. There are multiple avenues that can be used to channel remittances, which include spouses and other relatives, and this could also be looked at more holistically through a survey designed to specifically account for all the sources of remittance receipt. Despite these challenges, a large majority of the respondents who received remittances according to this survey (approximately 87%) were captured in the results of the study.

Future research could also place a stronger focus on the interaction between household spending on access to basic infrastructure and the engagement in female entrepreneurship. This line of inquiry is particularly relevant given the rural skew of the sample, and could help uncover how rural female-headed households might use remittances not just for business capital, but also to overcome infrastructural deficits that inhibit entrepreneurial success. This can allow for a better understanding of household entrepreneurship, as the lack of access to infrastructure is a known barrier to entrepreneurship.

These findings suggest several policy avenues. Financial inclusion efforts targeting remittance-receiving households, particularly in rural areas could be expanded through mobile banking services, savings cooperatives, or subsidized bank account access. Additionally, policies that encourage formal channels for remittance transfers may facilitate entrepreneurial use, especially when paired with small business training or matching grant schemes. Future programs may also consider targeted support for female-headed households, which, as shown in our findings, may exhibit distinct remittance-use behaviours.

Lastly, future research on Malawi should consider examining the role of remittances in the context of farm-related entrepreneurship and farm mechanisation as the agriculture sector plays a dominant role in the Malawian economy. Understanding how remittances affect smallholder farmers as a means of insurance and investment can also help inform policy on how best to deliver loans and insurance programmes to Malawian farmers as government moves towards diversifying the agricultural sector. Although the type of remittances (altruistic vs investment-seeking) has the potential to impact on entrepreneurial decision, such dynamics could not be examined in this paper owing to data limitations. This presents an avenue for future research.

The authors appreciate the constructive comments from three anonymous reviewers. All caveats apply.