WOMEN'S ENTREPRENEURIAL INTENT IN BURKINA FASO

In Burkina Faso, women make up nearly 52% of the population, but they account for less than 20% of business start-ups. This paper seeks to identify factors that explain the entrepreneurial intent of Burkinabe women. It is based on socio-cognitive career theory and a quantitative approach involving a sample of 935 women. The results of the multinomial logistic regression show that five variables (perception of skills, fear of failure, education level, household size and household income level) explain women's intention to start a business with as a reference modality (1: having the intention). The main contribution is of a methodological nature, and through the choice of reference modalities allows for greater precision on the influence of the variables and the categories or sub-groups of women influenced.

For the latter, an individual generally tends to choose activities for which he or she believes he or she has the necessary skills and abilities. 66.5% of Burkinabè entrepreneurs feel they have experience in the current activity of their enterprises before the creation of their business. (Dialla, 2004). Similarly, 57.3% of creators mentioned as their motivation for creating the desire to use the experience they had previously acquired on their own account (Ouédraogo, 1999). Entrepreneurial self-efficiency, a concept that can be equated with perceived feasibility (Shapero and Sokol, 1982) or perceived behavioural control (Ajzen, 1991) is very important in explaining entrepreneurial intent (Barbosa et al, 2007;Zhao et al, 2005). Several works (Kolveired, 1996 In the context of this work, the perception of competences refers to the knowledge, skills and experience deemed sufficient by an individual to start a business. Thus we can formulate the following hypothesis.

H1. The perception of skills to start a business positively influences women's intention to start a business in Burkina
Faso.

Fear of failure (Fearfail) and women's entrepreneurial intent:
In some cultures, the relationship to failure is fatal. The failure of an entrepreneurial experience is often perceived as the failure of the entrepreneur himself (Liles, 1974) and as a result he loses his legitimacy, credibility and reputation as an entrepreneur. It is equated with personal incompetence that is severely sanctioned at all levels: psychological, financial, social, family, legal and professional (Diamane and Koubaa, 2015). In a study carried out in Algeria on the influence of individual and environmental factors, 63% of people agree that fear of failure prevents them from starting a business (Benredjem, 2009). Song Naba and Toé (2015) estimate that, in general, the fear of failure is very low in African countries. It is 23% in Cameroon, 13% in Uganda, 14% in Botswana and 25% in South Africa. However, cautionary behaviour is more frequent among women than among men, but the proportion varies according to age and level of education (Bernard et al, 2013). In the same vein, Iyer (1995) claims that men interviewed in his intention survey are less afraid of failure and more confident than women. Thus, the following hypothesis is formulated.

H2. Fear of failure negatively influences women's entrepreneurial intent in Burkina Faso.
Women's age and entrepreneurial intent:-Previous studies clearly highlight age as an important factor in determining entrepreneurial intention (Brockhaus, 1982;Reynolds, 1995;Rotefoss&Kolvereid, 2005). This relationship is curvilinear, with a peak around the age of 35, which is even stronger for women. However, in some countries, such as Spain, the opposite effect is even observed, with women over the age of 55 showing an increased propensity to start a business.  As a result of the opportunities provided by higher education, women with higher levels of education are less exposed to necessity entrepreneurship (Lima et al, 2014).
Extensive data from the United States indicate that groups with lower levels of education show less interest in an entrepreneurial career (Reynolds, 1995;Reynolds and Miller, 1990). However, the effect of education level does not appear to be linear. In this study we formulate the following hypothesis 4.

H4.
The level of education of Burkinabè women positively influences their entrepreneurial intention.

Mediatisation (Social Valorisation) of entrepreneurship and women's entrepreneurial intent:
Role modelling and research to date suggests that women in particular require greater exposure to the success of women entrepreneurs (McMurray, 2001). Scott and Twomey (1988) have shown that repeated exposure to role models is likely to be a key contributor to the genesis of positive behavioural intent.
The social valorisation of entrepreneurship often assimilated to the context (environment) plays an important role in the development and pursuit of new opportunities (Audretsch, 2002). Context or environment can be understood as everything that is situated around the individual and can influence him or her. The social valorisation of entrepreneurship is understood through mediatisation (Nbmedia). In this sense, Amari et al (2014), consider that the desire to start a business comes from the individual himself, but refers to the social, family, political, economic and cultural environment. Basu and Virick, (2008; Linan and Chen, 2006; etc.) have shown that this variable significantly explains intention. This leads us to formulate the following hypothesis5.

Women's income level and entrepreneurial intent:
The initial capital of a new creation comes from personal income and family equity in general: social capital, human capital, financial capital and survival capital (Baccari E and Maoufoud S, 2008, Anderson and Miller, 2003). Hundley (2006) has shown that family income is positively correlated with the rate of self-employment. The same type of positive relationship between income levels and the self-employment rate was also observed among youth by Dunn and Holtz-Eakin (2000). In addition, Erkko and Zoltan (2010) found a positive relationship between an individual's family income and their aspirations to grow their business.
However, in a comparative study in Brazil between two income classes, Liman et al (2014) find that low-income (lower class) women have more self-efficacy and a higher intention rate than their higher class counterparts. We believe that income level can affect the level of intention and make the following hypothesis 6.

Household size and women's entrepreneurial intent:
Women starting their own businesses are often forced to cope with family obligations (Bonnetier, 2005). As Ronsen (2014) shows, it is not only the presence of children in the family that should be considered when studying women's entrepreneurship, it is indeed the whole family framework that will influence the woman's decision to start a business, mainly the dependants and spouse.
Fofana et al (2020), argue that motherhood and the associated family burdens have less and less influence on the motivation of women in their entrepreneurial initiatives. Their study shows that family size is really not relevant in explaining intention. In a similar vein, Bernard et al (2013) conclude that having one or more children does not play a significant role in being an entrepreneur, nor is it significant in the survival rate of businesses started or taken over by women.
While women entrepreneurs perceive entrepreneurship as a tool to adjust their careers to their family duties (Fofana et al., 2020), it is clear that these duties increase with family size. Thus we formulate the following hypothesis 7.

Research Methodology:-
This section presents the approach of the research we conducted.
Through this study, we seek to identify the factors that explain women's entrepreneurial intent in Burkina Faso. The dependent variable has three modalities (Yes= 1; No = 0 and Don't know = -1) to the question whether the individual intends to start a business. The explanatory variables are a mixture of binary and numerical variables. When the dependent variable has several nonordered categories (K > 2), we speak of multinomial logistic regression. We are then in the context of multinomial logistic regression. Several models (probit, logit, etc.) are often used in this case. The choice between the two models is difficult to justify from a theoretical point of view and depends on the application. There is practically no difference between these two laws (Aurier and Mejia, 2014). The option for the logistics law is simply motivated by its simplicity within this framework. A very interesting property of this regression is that it allows the estimation of an odd ratio which provides information on the strength and direction of the association between the explanatory variable and the variable to be explained.
Multinomial logistic regression consists in designating a reference category, the last one (those who intend in our) for example to fix the ideas, and in expressing each logit (or log-odds) of the (K-1) modalities with respect to this reference using a linear combination of the predictive variables.
The objective is to model the probability of an individual belonging to aYkmodality. We have an explanatory variable Y to K modalities and we try to model the probabilities.
( = ), = 1, … − 1, = 1, … , The approach consists of setting a reference modality, for example modality K, and modelling the probabilities. ( ) according to where = ( 1 ,•••, ) • Women's intention to start a business or not is thus modelled on the basis of a regression model whose variable to be explained is ( ) is polytomous.
X is the vector of the explanatory variables associated with the intention to create or not to create a business and is the vector of the associated coefficients. The vector of variables is composed of: perception of skills (Suskill), fear of failure (Fearfail), age (Age), level of education (Bfreduc), media coverage (Nbmedia), income level (Bfhhinc), household size (Hhsize). ε is the error term.

Sample:
The data used in this work is taken from the Global Enterpreneurship Monitor (GEM) Burkina survey database conducted in 2015. This survey covered 2,850 individuals, respecting the distribution of the Burkinabe national population by age and gender.
As our study focused on women's entrepreneurial intent, we initially eliminated men from the sample. This left 1147 women (out of 2326 individuals). We then eliminated the women who did not respond to the three response modalities on intention (i.e. yes, no and don't know). Finally, our analysis focused on a sample of 935 women. The average household size is 9.8 people. 37, 64% had an annual income between 0 and 364 US dollars and only 9% had an income above 1820 US dollars. 36.14% are between 25 and 34 years old and 6.31% are between 55 and 64 years old. 65.77% of the sample did not go to school, 10.3% have secondary education and only 0.11% have higher education.

Analysis of the data:
We ran a multinomial logistic regression to determine the relationship between the independent variables and the probability of intending to go into business. The ability of the logit model to process and propose an interpretation of the coefficients for the explanatory variables is very interesting. The Stata/SE14.0 software was used for this purpose. To interpret the results we used odds ratios.
The choice of the reference modality mainly affects the reading of the coefficients. The reference modality defined by the software is 1. As the initial choice of the reference modality was not restrictive, we have subsequently changed, characterising in our case the modality (0) in relation to (-1). There is no need in this case to repeat the regressions. We obtain the logit (logarithm of the odds) between any 2 modalities of the dependent variable by simple differentiation.

Research Results And Discussion:-
The following table1 presents the results of the logistic regression. The tests are significant at the threshold of 1% (3 stars); 5% (2 stars) and 10% (1 star). On the other hand, household size and level of household income are not significant in the case of modality (-1) i.e. women who, when asked if they intend to, answered "don't know" and the level of education is also not significant in the case of women who chose modality 0 (no) compared to the reference modality 1 (yes).
By characterising the modality (0) in relation to (-1) the following result is obtained:

Link between perception of skills and intention to start a business (Suskill):
The regression results show that women who choose modality -1 feel less competent compared to those who choose modality 1. Otherwise women who say they do not know if they intend to start a business in the next 6 months perceive their skills less than those who say they intend to start a business.
In addition, women who do not intend to go into business feel less competent than those who do.
Finally, by characterising the modality 0 by -1, we discover that the more women's perception of skills (Suskill) increases, the more they choose 0 in relation to (-1).
This result shows that the more women perceive their skills, the more precise they are in their choice. And that the choice of intention increases with the level of perceived competence.
This result confirms SakolaDjika (2019), who finds that in Burkina Faso, an individual who perceives entrepreneurial skills is 1.75 times more likely to have entrepreneurial intentions than one who does not. As a reminder, this variable is understood through experience in the field of activity among others. These results are consistent with those of Wilson et al (2009), Giacomin et al (2010) who found that the more experience an individual has in a trade, the more competent they feel to practice the trade and therefore to undertake in that field.

Fear of failure:
The more women are afraid of failure, the more they choose modality -1 (don't know) over modality 1. Then, the more they fear failure, the more they choose not to have the intention in relation to the reference modality (Yes: have the intention). In other words, an increase in the fear of failure makes women more likely to choose not to have the intention rather than to have the intention.
By characterising the modality 0 in relation to -1, we realise that the more the fear of failure increases, the more women choose 0 in relation to (-1). Moreover, the effect of this variable is more exacerbated in (0 vs. -1) than in (0 vs. 1). This means that the more women apprehend their fear of failure, the more visibility they have regarding their choice, in this case it is to have no intention. This result is consistent with that of Diamane and Koubaa, (2015) and Benredjem, (2009) among others. It is also justified by the fact that cautious behaviour is more frequent among women than among men (Bernard et al, 2013).

Age :
The results of studies on the influence of age are contradictory. In the present study, age does not explain the probability of intending to start a business among women in Burkina Faso.

The level of education:
The higher their level of education, the more women choose modality -1 with reference to modality 1. Then, the higher their level of education, the less they choose 0 with reference to (1). Otherwise, the higher their level of education, the more women say they do not know if they intend to start a business compared to the choice to have the intention. Indeed, it should be pointed out that the higher their level of education, the more women have the opportunity to access salaried jobs. Therefore the choice of entrepreneurship would be a last resort (after unsuccessful attempts to get a job in the civil service). This result confirms Bayala (2015) who finds that 82% of creators have not attended school or have less than the Baccalaureate and that the rate of creation decreases with the level of education.
On the other hand, this variable is not significant in the case of women who choose not to have the intention (modality 0) compared to those who do (modality 1).

Media coverage:
It follows from the table of results that the media coverage of successful new businesses does not influence the likelihood of women's entrepreneurial intent in Burkina Faso regardless of the modality of reference. However, this can be put into perspective when we know that the Burkinabe population is essentially rural and that in this environment access to the media is not necessarily the most common thing, especially among women.

Household size and intention:-
As the size of their household increases, they are more likely to be unintentional. In other words, those who do not intend have larger households than those who intend.
This result can be explained by the demands of women's roles. Indeed, in Burkina Faso it is the woman who takes care of the house, as they say. The larger the size of the household, the less time she will have to devote to entrepreneurship. We can also see the effect of the pooling of resources. Indeed, the different members of the family each contribute in their own way to the running of the household. Thus the more members there are, the higher the contribution and the less the woman feels obliged to be entrepreneurial, given that women are more entrepreneurial out of necessity (Song-Naba and Toe, 2015). This reasoning is corroborated by the result on the link between household income and intention.

Income level and intention:
Finally, a higher household income encourages women not to intend to use the reference modality. In other words, the probability of women's entrepreneurial intention decreases as their income increases. This result suggests that women choose to be entrepreneurial when household income is insufficient to support the family. It confirms the authors' work that women undertake more out of necessity than opportunity. In a similar vein, Liman et al (2014) find that lowincome women have more self-efficacy and a higher intention rate than their higher income class counterparts.

Conclusion:-
This work focused on the factors that explain women's entrepreneurial intent in Burkina Faso. A multinomial logistic regression was performed on a sample of 935 women from the GEM (2015) survey data of Burkina Faso. Five variables explain women's intention to start a business with as a reference modality (1: having the intention).
Firstly, a reading of the 0 versus 1 modality shows that the more competent women feel, the more they choose the intention to go into business rather than the opposite. On the other hand, the more their fear of failure increases, the larger the size of their household, and the more their household income increases, the less they choose the intention to go into business.
Then by switching to modality -1 versus 1 we find that women who do not know whether they intend or not, perceive their entrepreneurial skills less than those who intend. They express a greater fear of failure compared to those who have the intention. The higher their level of education, the more they say they do not know.
Finally, by characterising the modality 0 in relation to -1 we can see that the more the level of women's perception of competence increases and the more the fear of failure increases, the more they choose not to have the intention (0) in relation to (-1). The effect of the fear of failure is more exacerbated in (0 vs. -1) than in (0 vs. 1).
These results, on the one hand, add to the knowledge about women's entrepreneurship, particularly women's entrepreneurial intent in Burkina Faso. The main contribution is methodological, and through the choice of modality allows for more precision on the influence of variables and the categories or sub-groups of women influenced. In managerial terms, this makes it possible to design and provide specific support for the needs of the sub-groups. Indeed, this work offers levers to build on in order to generate more intention, with the understanding that it is an important step towards creation.
For example, support should be more oriented towards competence building (training, internships, etc.), promoting entrepreneurship as a desired career and not "suffered 3 ".
However, the results of our study have limitations. Firstly, they concern the failure to take into account other variables such as the place of residence (urban or rural), occupation at the time of the survey (employed or unemployed), etc. These limitations are not always taken into account. Taking these variables into account in the framework of another study will undoubtedly contribute to a better understanding of the subject.