Determinants of Employability of Young People in Congo

The objective of this paper is to analyze the determinants of employability of young people in Congo. To achieve this aim, we have combined both methodological approaches: the microeconomic approach and macroeconomic approach. The approach microeconomic highlights the microeconomic determinants from binary and multinomial logit models applied to the survey data of National Institute of Statistics while the macroeconomic approach estimates the macroeconomic determinants using the autoregressive models applied to aggregate data from the World Bank. The results show that the employability of young people depends on both microeconomic and macroeconomic factors. Take about microeconomic, education, individual gender, age; location residence and affiliate to a work organization are the main determinants of the employability. A macroeconomic level, analyzes have shown that GDP and gross fixed capital formation as well as imports are determinants of the employability of young people.

Theoretical Economics Letters the labor market. The majority of these young graduates find themselves in a situation of unemployment which is sometimes structural and sometimes cyclical. The difficulties in their employability can be explained by several factors, including the phenomenon of training-employment mismatch and economic instability. According to the National Human Development Report (2020), several skills are hard to find in the labor market, including foreign language skills, advanced computer technology or programming skills and technical or practical skills as well as skills in planning and organization. In addition, the Congolese economy is experiencing unstable growth that can negatively affect the structure of the labor market, in terms of job creation. According to the World Bank This article is structured as follows. The first section presents a brief review of the literature on the determinants of employability. The second section describes the methodology adopted. The results are presented and discussed in the third section.

Literature Review
The determinants of employability have always been one of the most discussed topics in the economic literature both from a theoretical and empirical point of view (Schultz, 1961;Becker, 1964;Phelps, 1972;Harvey, 2001;Knight & Yorke, 2003).

Theoretical Review
At the microeconomic level, models of heterogeneity of human capital (Schultz, 1961;Becker, 1964) and models based on the imperfect information hypothesis, notably models based on the filter hypothesis (Phelps, 1972;Spence, 1973;Thurow, 1972) are among the first models to analyze the relationship between education and employability. Models based on the heterogeneity hypothesis postulate that the differentiation of earnings or productivity observed between individuals and the probability of being employable can be explained by the differentiation of the human capital endowment of each individual. However, these models presented a certain limitation in the information on the productive capacity of the individual. At the time of hiring, only the individual has knowledge of his productive capacity. On the other hand, the employer has no knowledge of the productive capacity of the individual. Human capital cannot be an indicator of the employer's appreciation of the productive capacity of the individual, but rather, it will serve as a filter and selection between individuals with a high endowment in high human capital and those with a high human capital endow- DOI: 10.4236/tel.2022.122018 346 Theoretical Economics Letters ment. Harvey (2001), Yorke (2003, 2004), Dacre Pool and Sewell (2007) have also theoretically analyzed the relationship between education and employability by going beyond generic and specific capacities. Harvey's "magic ball" model (Harvey, 2001) states that attempts to measure the effectiveness of employability by graduate employment rates involve a "magic ball" model of the impact of education on employment. The assumption is that higher education institutions provide opportunities for developing employability and therefore finding employment. There is a link between the opportunities for developing employability and the individual employability of graduates. This link is invariable, used as legitimate for practical use of the employment rate as a measure of the employability rate of an institution. Higher education institutions offer a range of student employability development opportunities including the development of attributes to obtain, retain and develop jobs or careers, self-presentation skills, necessary when seeking employment, to encourage training and the willingness to continue learning.
The USEM model of Knight and Yorke (2003) suggests that employability is a set of accomplishments, skills, understanding and personal qualities, which make individuals more likely to obtain employment and to be successful in the occu- Otherwise, other work has highlighted social capital as a determining factor of employability. As a concept, the historical premises of social capital date back to the 19th century with the works of Alexis De Tocqueville (1835) and Durkheim (1893). It was at the beginning of the 20th century that the first known reflection of this expression appeared in a work that discusses the importance of social capital in education and local communities (Hanifan, 1916). The author tries to define this concept, relating it to the relationships of friendship, sympathy, mutual aid, cooperation and solidarity that characterize the members of a community. According to Hanifan, these relationships constitute in themselves a useful wealth to promote the well-being of the members of the community. Jacobs (1984) uses it to illustrate the importance of networks within cities and Loury (2002) uses it to describe the various economic opportunities that young people from ethnic minorities face through their social ties.
In France, Bourdieu (1980) developed the first approach by defining social capital as: "the sum of resources, current or virtual, which accrue to an individual or to a group by virtue of having a lasting network of relationships, of more or less institutionalized mutual knowledge and recognition". Following the exam-  Loury (2002), Bourdieu starts from the postulate according to which, to achieve the goals pursued, the actors do not only use the material means and their personal skills, but also the social relationships they have in their families, their home community and any other organization to which they are a member.
At the macroeconomic level, several explanatory models of employability have been developed, in particular the theory of structural transformation of the economy and the external factors of employability. The work of Clark (1940), Lewis (1954 and Kuznets (1966) on the theory of structural transformation of the economy addresses the dynamics of employability by highlighting the importance of the agricultural and industrial sectors in the process of economic development. According to the theory of structural economic transformation, labor shifts from the less productive sector to the more productive sectors, that is, from the agricultural sector to the industrial sector and then to the service sector. According to McQuaid and Lindsay (2005), among the external factors, we find macroeconomic factors characterized, in particular by macroeconomic stability, medium and long-term viability of enterprises, the level and nature of labor demand in the national economy. Macroeconomic stability can be interpreted as keeping the economy on its potential growth path. The latter reduces the level of unemployment in the national economy, and in turn increases the level of employability. Through positive economic growth, businesses become more and more viable. They create more jobs in which individuals can claim their employability.

Empirical Review
The relationship between microeconomic factors and employability is one of the traditional themes of economic analysis. In view of a large number of works, only a few are selected to shed light on this research. The work carried out by Contreras et al. (2011) in Chile on the determinants of workforce participation and employability showed a positive relationship between human capital and employability. Indeed, their estimation results from the probit model for the youngest women show that the studies carried out, measured by the number of years of schooling, increase the probability of employability. Employability is particularly strong for individuals with at least 12 years of schooling. Likewise, Berntson et al. (2006)  Moreover, by analyzing the job search methods used by job seekers, authors such as Epiphane and Martinelli (1997) and Forsé (1997) observe that the process used to gain access to a job depends on the social characteristics of the individual, including the socioeconomic status of the job previously held, social origin, sex, age and marital status. They also claim that these job search patterns predict the socioeconomic status of the job obtained. For example, compared to other job seekers, the unemployed make more use of family relationships to find a job, young graduates and trainees are recommended more by their schools, while those who already have a job are recommended by their employer (Forsé, 1997).
In addition, the work of Wapoh (2013) on the modes of access to employment of graduates in Côte d'Ivoire shows that women use personal relationships more than men.
Works on the analysis of the relationship between macroeconomic factors and employability that highlights the opportunity factors of the job offer. According to Keynesian employment theory, business investment and consumer spending constitute effective demand. This encourages employers to produce goods and services in order to meet this demand. Thus, investment is a lever for job creation. Investment stimulates the economic growth of a country. Indeed, in countries where the level of investment is low, the productive capacity of their economy is also low. And this translates into lower economic growth rates and consequently lower job creation rates. Anyanwu (2013)  The causality test for Malaysia shows that government spending causes employment growth and that employment growth in turn causes capital. For the case of the Philippine, the causality shows that the total of exports and imports by a one-way relationship, but the error correction term is not significant.
Much empirical work has analyzed the relationship between gross domestic production or economic growth and employment. These works whose theoretical anchoring is the Keynesian theory of employment and Okun's law show of Using several methods, including the ordinary least squares method and the error correction method for different OECD countries, Dopke (2001) has shown that the service sector is also considered to be a determining factor of l elasticity of employment in OECD countries. Indeed, according to the author, the share of services in GDP growth is one of the potential determinants of employment intensity. Dopke finds that increasing the level of the service sector leads to an increase in the elasticity of employment with respect to growth. Using a staged lag model of panel data for Africa, the work of Kamgnia (2006) shows overall that economic growth exerts a positive and significant, albeit small, effect on the labor force in Africa. Importantly, the magnitude of this effect increases significantly over a one-year period, then decreases to the point of becoming negative and non-significant after a two-year delay. In addition, the employment variable delayed by one year has a positive effect on the level of employment in progress.
Using the ordinary least squares method, the work of Ningahe et al. (2015) overall concluded that economic growth has a positive effect on the volume of employment in Cameroon. Indeed, an economic growth rate of around 1% leads to an increase in the total volume of employment of around 0.41%, the authors stress. Observations of the values of sectoral elasticities or the respective reactions of agricultural employment, industrial employment and service employment in relation to the variation in GDP have shown that the elasticity of agricultural employment in relation to production is very low and statistically non-zero.
It takes a value of 0.29%. At the same time, when it comes to the reaction of secondary sector employment, a 1% increase in the economic growth rate translates into an increase of 0.52%. The tertiary sector saw its level of employment increase by 2.11% following an increase in economic growth of 1%. In the same vein, the work of Fulgani and Narayan (2011) on the trend and determinants of service sector employment intensity in India using ordinary least squares method showed that among the macroeconomic factors of employment intensity in the service sector, a distinction is made between investment and public expenditure. In contrast, other work on the relationship between macroeconomic factors and employability has shown negative effects of macroeconomic factors on employability (Temitope, 2014;Pleic & Berry, 2009). GDP has no effect on employment. However, the pre and post liberalization dummy variable has a negative effect on employment.
Other authors like Ozgur et al. (2018), Zonzilos (2000), Anton and Terelli (2001), Kitov (2011) analyzed the effect of macroeconomic factors on unemployment. In analyzing the effect of economic growth on unemployment, Ozgur et al. (2018) showed from an analysis of cointegration in the countries of Eastern Europe over the period from 1992 to 2014, that a 1% increase in economic growth led to a decrease in the rate unemployment rate of 0.08%. Zonzilos (2000) for the same relationship in Greece from 1965-1999 comes to the same conclusion that an increase in production of 1% leads to a decrease in unemployment of −0.28%. Similarly, Anton and Tirelli (2001) examined the relationship between growth and unemployment in OECD countries over a period 1995-2000 using a structural VAR model, and found a relationship negative between economic growth and unemployment. These conclusions are valid for the work of Kitov (2011) on the relationship between GDP per capita and employment in the American, French, British, Australian, Canadian and Spanish economies. Using the ordinary least squares method, Misini (2017)  ship between the unemployment rate and gross domestic product as a proxy for economic growth and shows that the relationship between two variables is significant at the 5% level in Kosovo. Indeed, nominal GDP negatively affects the unemployment rate, any increase of 1% of GDP leads to a decrease in unemployment of −0.43%. In a study on the euro area, Gomez-Salvador and Leiner-Killinger (2008) find that economic conditions, represented by economic growth, are negatively correlated with the youth unemployment rate, that is, the youth unemployment rate increases when the economic situation deteriorates and inversely. From this review, it emerges that employability depends on both microeconomic and macroeconomic factors. Microeconomic factors relate primarily to education, social capital and socio-demographic characteristics. However, it is important to note the absence of work with the variable "membership of a trade union or work organization" as a determining factor of employability. Thus, this research fills this gap of knowledge observed on the existing literature, which leads us to use the variable "membership of a union" as an explanatory variable. Macroeconomic factors mainly concern macroeconomic stability, gross fixed capital formation and international trade.

Research Methodology
The literature review revealed two types of factors of employability: microeconomic factors and macroeconomic factors. As a result, we have adopted two methodological approaches not only to capture the effect of each category of factors, but also taking into account our objectives and the nature of the data.
The macroeconomic approach will be based on the ARDL model (Autoregressive Distributed Lag) or the autoregressive model with staggered delays introduced by Pesaran et al. (2001) while the microeconomic approach highlights the binary logistic model which originates from choice theory, and therefore from the utility functions of individuals (Van De Vyvere, 1995;Hakim, 2001).

Theoretical Model of Employability
The macroeconomic methodological approach used is inspired by that of Greenway et al. (1999) derived from the Cobb Douglas function.
With i and t representing branches and time respectively, Q output at constant prices, A is an index of technological change which measures technical efficiency, L is unskilled labor, H is skilled labor and K is stock of capital at constant prices.
It is assumed that the goods and labor markets are in perfect competition. In this equation the variables λ, α and β represent the shares of each factor in the production sharing, and γ is a coefficient that allows other factors to affect the efficiency of the production process. The first-order condition of profit maximi-DOI: 10.4236/tel.2022.122018 352 Theoretical Economics Letters zation implies that the firm employs the factors of production to the point where the marginal product of each factor equals its price. This gives the following equations for the wages of unskilled and skilled workers, respectively: w l , w h , p, and represent respectively the wages of unskilled workers, the wages of skilled workers the price of the good produced and the two-factor marginal productivity (H and L).
By making the ratio between two Equations (2) and (3), we obtain the follow- Using this last equation the relative demand for labor is written: In a Cobb-Douglas function, the technological bias in favor of the relative demand for skilled labor corresponds to an increase in the ratio (α/β), all other things being equal. Technical progress increases the relative demand for skilled labor because it pushes the marginal productivity of H to the detriment of that of L.
From Equation (4), we can write the following equation: By replacing H by its expression The derived demand for low-skilled labor can be obtained by taking the log of Equation (7): The demand for skilled labor can be defined in the same way: The demand for skilled and unskilled labor is negatively related to the relative wage w h /w l and w l /w h , respectively. The demand for skilled and unskilled labor are positively related to output but negatively related to technological progress where LC is a measure of trade openness, in our case we use imports in value (m). The expansion of trade openness can increase international competition from imports in domestic markets, and lead to greater exposure to international markets, thus causing an impact on factor demand. T represents the time trend. Replacing A t by its expression in the two equations found previously (8) and (9), we obtain for an industry i at time t the following equations: At the end of the presentation of the theoretical model, we will present the methodology used in the estimation of the macroeconomic determinants of employability.

Estimation Methodology (ARDL Model)
This article uses the Autoregressive Distributed Lag (ARDL) approach to analyze the macroeconomic determinants of employability. The most widely used cointegration techniques are those in two stages of Engle and Granger (1987), the approach developed by Johansen (1995) and Johanson and Juselius (1990). However, these usual cointegration tests recommend the use of integrated series of the same order I (0) or I (1). In addition, they are suitable for large samples. In order to remedy these shortcomings, Pesaran and Pesaran (1997) and Pesaran et al. (2001) have developed a new approach that is more flexible and less restrictive than the previous techniques, the ARDL (Autoregressive Distributed Lag) model or the autoregressive model with staggered delays. This model makes it possible, on the one hand, to test long-term relationships by using limits tests "bounds test" on series which are not integrated of the same order and on the other hand, to obtain better estimates on small samples (Narayan, 2004).
Starting from the theoretical model and drawing on the empirical work proposed by Anyanwu (2013)  a 0 and β 0 are constants; a 1 , a 2 , a 3 , β 1 , β 2 and β 3 are coefficients of variables.
By adopting the ARDL approach, the two models look like this: With Δ: first difference operator; a 1 -a 4 : the Error Correction Models (ECM) representation; θ 1 -θ 4 : long-term relationships; p is the number of delays in the explained variables RE (employment ratio) and TC (unemployment rate); q is the number of lags of the explanatory variables. In addition, we apply the "bounds tests" approach in order to find out whether there is a long-term equilibrium between the variables. Thus, a Fisher test (the test statistic is the F-statistics) is implemented to verify the following hypotheses: the null hypothesis (H0) is the absence of a long-term equilibrium relationship: θ 1 = θ 2 = θ 3 = θ 4 = 0, while the alternative hypothesis H1 is the presence of the long-term uniform relation between the studied series: θ 1 # θ 2 # θ 3 # θ 4 # 0. ARDL Bounds test presents two sets of critical values. The first set corresponds to the case where all the variables are integrated of order I (0), while the second set corresponds to the integrated series of order I (1) (Pesaran et al., 2001).
The data used to estimate the macroeconomic determinants of employability come from aggregate data from the World Bank. They cover a period from 1991 to 2019, that is 29 years. This period simply depends on the availability of data.

The Microeconomic Determinants of Employability
Within the framework of this research, employability is approached according to three dimensions within the meaning of Hillage and Pollard (1998), access to employment and job retention and job change. These three indicators are each qualitative and binary. The binary logistic model is best suited for estimating the microeconomic determinants of employability given the nature of the indicators.
In a choice decision process, the goal of the decision is to find a better solution among the possible alternatives to meet the objectives (Van De Vyvere, 1995;Hakim, 2001). In reality, there are two types of choice, the first type is the con- Based on the utility theory (Etner & Jaleva, 2015), the individual will choose, among the J, modalities at his disposal, the one which gives him the highest utility. In this perspective, the discrete choice model is described by a variable y, which indicates the observed choice , 1, , n i i J =  , and by the latent variables Uin which govern this choice: If the individual n chooses modality i, the probability P n (i) of observing this choice is defined as follows: By making assumptions on the joint probability distribution of the error terms , 1, , in i J ε =  , we can derive from it all the usual multinomial choice models. In general, the probability P n (i) is equal to: And ( ) 1 , , n n jn f ε ε  denotes the density function of random terms. For the analysis of disaggregated choices, the models most often used are the logit and probit models.
Drawing on the theoretical foundations of discrete choice models and the methodology used by Zalle et al. (2017) in their work on vocational training and employability of young people in urban areas in which the authors explain the occurrence or not of the "employability" event. Thus, for each individual in the sample, we observe whether he is employable or not and we ask: Thus, we can define the probability that an individual in the sample is employable as the mathematical expectation of the variable Y i since: The logit model defines the probability associated with the event Y i = 1 as the value of the distribution function of the logistic law at point X i β: The model to be estimated is given by: where Y i is the variable to be explained, X i is a vector of observable characteris- When the law of the parameters is known, the most common estimation method is that of maximum likelihood.
The data used in the microeconomic approach come from data from the survey on the transition to working life conducted in Congo in 2015 by the National Institute of Statistics among young people aged 15 to 29 years.

Analysis of Results of the Macroeconomic Determinants on Employability
The macroeconomic analysis successively covers descriptive analysis, stationarity and cointegration tests as well as the interpretation of the results of the estimates. Table 1  The first statistic in Table 1 represents the mean. Imports and gross current fixed capital formation are respectively US$4.28 billion and US$3.19 billion. The average current GDP per capita is US$969,296.9. The averages of the composite index, the unemployment rate and the employment ratio represent 14.66% respectively; 33.09% and 29.29%. The maximum and minimum of the variables are almost of the same order as the average of the variables.

Descriptive Analysis
The results of the unit root tests (   analyzes that all the variables are integrated of order 0 or 1 for the Zivot and Andrews (1992) test. Then none of them is integrated with an order greater than 1. Therefore, the ARDL (Autoregressive Distribution Lag) model is the most appropriate to estimate the dynamic equation of employability. Due to its low constraint, this technique is used more and more as an alternative to the usual cointegration tests because of the flexibility it offers in data processing. Indeed, the ARDL test does not require that the model variables are purely I (0) or I (1). It is also a technique that offers the possibility of dealing jointly with long-term dynamics and short-term adjustments. Thus, we adopted this approach to analyze the relationship between macroeconomic factors and employability. ARDL modeling with the appropriate offset will correct the two series correlation and endogenous issues. Another reason for using the ARDL approach is that it is more robust and performs better for small sample sizes than other cointegration techniques by providing better estimates (Narayan, 2004). The number of lags of the dependent variable and of the explanatory variables is selected using an information criterion.

Application of the Cointegration Test
The cointegration test of the long-term relationship between macroeconomic factors and employability is done using the new ARDL "Bounds tests" procedure used by Pesaran et al. (2001) within the framework of the ARDL models. Table 4 shows the ARDL Bounds Tests results for Model 1 and Model 2.
The results of the cointegration tests show that the Fisher Statistics (F-Statistic) are equal to 6.43 and 3.88 respectively for Model 1 and Model 2. These values should be compared with the critical values below and above the significance level of 1% and 5%. The test statistics are above the upper bound for both models at the 5% significance level. Therefore, we reject the null hypothesis (H0) of the absence of a long-term relationship and we conclude the existence of a long-term relationship between the different variables, so there is a cointegration relationship between the macroeconomic factors and employability for both models.  Source: Author from Eviews software. The empirical results presented above show that current GDP per capita (LPIBHC) or economic growth is highly significant at different thresholds (1%, 5% and 10%) and positively impacts the employment ratio. Indeed, a 1% increase in current GDP per capita tends to increase the employment ratio by 0.04%. These results are consistent with theoretical predictions and for the majority of the work carried out on this issue with regard to the role played by economic growth in an economy in general and on employability in particular. Indeed, Keynesian employment theory states that the growth in the production of goods and services enables employers to increase the level of employability in the labor market. Empirically, these results corroborate those of Pissarides and Vallanti (2004) in the context of panel models in the United States, Japan and Europe. Our authors have shown through that productivity growth has a positive effect on employment. Otherwise, the work of Sarra (2015) on the relationship between economic growth and employment intensity in developed countries over the period 1991-2011 revealed that elasticity estimates vary considerably from one country to another. Otherwise, employment elasticities tend to be higher in countries with higher standard of living and closed, macroeconomic policies aimed at reducing macroeconomic (price) volatility have a significant effect on the increase elasticities of employment. Likewise, growth in employment intensity tends to be higher in countries with a larger service sector.  Countries with a higher share of the urban population tend to have higher employment elasticities. The work of Kamgnia (2006) shows overall that economic growth has a positive and significant effect, albeit a small one, on the labor force in Africa. Importantly, the magnitude of this effect increases significantly over a one-year period, then decreases to the point of becoming negative and non-significant after a two-year delay.

Interpretation of the Results of the Estimation of the Long-Term Relation
Our results are also in the same direction of those of Ningahe et al. (2015) who show that economic growth has a positive effect on the volume of employment in Cameroon using the ordinary least squares method. Indeed, an economic growth rate of around 1% results in an increase in the total volume of employment of around 0.41%. Observations of the values of sectoral elasticities or the respective reactions of agricultural employment, industrial employment and service employment in relation to the variation in GDP showed that the elasticity to agricultural employment to the production is very low and statistically non-zero. It takes a value of 0.29%. In contrast, when it comes to the reaction of secondary sector employment, a 1% increase in the economic growth rate translates into a 0.52% increase in the level of employment. The tertiary sector saw its level of employment increase by 2.11% following an increase in economic growth of 1%. In contrast, the increase in current GDP per capita tends to lower the unemployment rate.
Gross fixed capital formation has the same effects as current gross domestic product per capita. According to Keynesian employment theory, business investment is a component of effective demand. This encourages employers to produce goods and services in order to meet this demand. Thus, investment is a lever for job creation. Investment stimulates the economic growth of a country.
In countries where the level of investment is relatively low, the productive capacity of their economy does not increase. And this translates into lower eco- author has shown that import penetration is negatively associated with the level of employment in the secondary sector. Indeed, a 1% increase in the variation in import penetration leads to a decrease of −0.045% at the 5% significance level. Table 6 shows the estimation results of Model 1 and Model 2.

Interpretation of the Results of the Estimation of the Short-Term Relation
From the analysis of the results of this table, we appoint by D the first difference of the variables considered. The error correction term (−1) corresponds to the lagged residuals from the long-term equilibrium equation. Their estimated coefficients are negative and significant, confirming the existence of an error correction mechanism. These coefficients, which express the degree to which the variables (employment rate, unemployment rate) will be recalled towards the long-term target are estimated respectively of −0.9948 and −0.7184 for our ARDL models, thus translating a relatively rapid adjustment to long-term target.
The employment rate differential and the one-year and two-year lagged unemployment rate differential have a positive effect on their respective current values. In other words, the employment ratio improves the level of employment while the unemployment rate deteriorates employability. These results seem to be logical with regard to the theory of the segmentation of the labor market. Insiders in employment tend to stay as long as possible in the light of their experience in the profession, while unemployed outsiders will not have the opportunities for a job offer due to their inexperience for new workers and old unemployed workers. The work of Kamgnia (2006) illustrates this situation because the employment variable delayed by one year has a positive effect on the level of employment in progress. However, the effect of the three-year lagged employment rate differential and unemployment rate becomes negative.
The differential in current GDP per capita has a negative effect on the unemployment rate. This result is consistent with theoretical predictions, including Okun's Law, according to which changes in GDP lead to lower unemployment rates.
The differential in gross current fixed capital formation has a positive effect on the employment ratio and the composite employment index, and a negative DOI: 10.4236/tel.2022.122018 362 Theoretical Economics Letters  effect on the unemployment rate. However, its lagged values have opposite effects to those of the differential in gross current fixed capital formation.
The one-year lagged import differential has a positive effect on the employment ratio and a negative effect on the unemployment rate. The nature of these effects can be explained by the fact from the entry of imported products into Congolese territory to the wholesale and retail sale of products constitutes a chain promoting job creation and at the same time the reduction of unemployment. Otherwise, the purchase of inputs and production goods also promote job C. M. Makouezi, R. G. Ngobila DOI: 10.4236/tel.2022.122018 363 Theoretical Economics Letters creation and therefore lower unemployment. In contrast, the import differential has a negative effect on the employment ratio and on the composite employment index, and a positive effect on the unemployment rate. This result seems logical since the return to investment even in the case of reduced unemployment cannot be realized instantly even less so in the first year. The first few moments can be spent purchasing fixed assets and equipment and will continue to operate for years to come. Therefore, we can observe an increase of the unemployment rate and a decrease of the employment rate during the investment period.

Robustness Tests of the Results
The estimated error correction models are generally good, as shown by the obtained values of R 2 which are all close to 1, that is 0.9986 for Model 1 and 0.9988 for Model 2. The following Table 7 shows the results of normality tests of the residuals and the tests of stability of the coefficients.
According to the analysis (Table 7), it addresses that all the tests of Model 1 and Model 2 are all conclusive because the corresponding probabilities are all greater than 5% (0.05). We cannot therefore reject the null hypotheses at the 5% threshold according to which the residuals are homoscedastic and follow a normal distribution on the one hand, and the coefficients are stable on the other hand.

Interpretation of the Results of the Estimation of Microeconomic Determinants
Recall that the results are interpreted according to the three dimensions of employability, namely, access to employment, job retention and job change, according to Model 1, Model 2 and Model 3 as well than the composite employability index.

Likelihood of Access to Employment
We recall that in Model 1, the dependent variable is the ability to get a job (initial employment), conditioned among other things by the education system, and human capital was measured by the highest level of studies (diploma), vocational training as well as the field of study. These results are consistent with economic theory (Becker, 1964), and are also supported by the results of empirical research (Wittekind et al., 2010;Contreras et al., 2011), and de Lassassi and Hammouda (2012) on a sample composed only men. In addition, Harvey (2001) and Knight and Yorke (2003) highlight higher education as a determining factor of employability. However, the results of Lassassi and Hammouda are contrary on a sample composed of men and women, and those of Zalle et al. (2017) in a market where the informal sector occupies 75% of employment areas, but employing workers with a low level of education or absolutely no.
When we consider the feminine gender as the modality of reference, thus, the results show us that men are more likely to access a job than women in Model 1 and Model 3. In other words, women have more difficulty accessing a job than men. These results are similar to those of Lassassi and Hammouda (2012). In their work, the authors show that gender is the first factor in a person's availability to participate in economic life. Men are 18.18 times more likely to be employable than women. Likewise, Zalle et al. (2017) find the same results insofar as women are 19.57 times less likely to be employed than men. This discrimination in terms of gender could be due to the interruption of their career for reasons of maternity or taking care of sick children on the one hand, and to absenteeism mainly due to the difficulties of finding a balance between work and family responsibilities and/or domestic. In the case of this approach, the work of Erickson et al. (2000) has shown that family responsibilities are an important explanatory factor for absenteeism for men and women. According to the author, the correlation between hours spent on family and domestic tasks and absenteeism is positive for women and for men, although stronger for women.
The variable "belonging to a union or a work organization" has a positive effect on access to employment in Model 1 and in Model 3. Indeed, young people belonging to a work organization are more likely to "access a job than other young people". These results can be explained by the fact that associative movements play a large role in the employability of young people insofar as they put pressure on employers to recruit young people.
The education of the father and the education of the mother are significant at the 10% level in Model 3 with the opposite effects on access to employment. Indeed, the education of the father has a negative effect while that of the mother has a positive effect. The positive effect of mother's education theoretically corroborates the work of Granovetter (1973), Lin (1995) who highlight a positive  Kamanzi (2006) in that they establish a negative association between access to employment and the mother's level of higher education.

Job Retention Probabilities
Model 1 is not well specified because the probability associated with the chi2 statistic is greater than 10%, or 16.24%. This poor specification is due to the selection of the sample. By specifying the selection model, it addresses that eight variables are significant at the different thresholds.
The level of education, especially young people with a secondary professional level and a post-secondary professional level as well as the university level are more likely to keep their jobs than young people without level. These results show that there is a positive link between education and job retention. Young people with a high level can be considered as the insiders. Job retention for these young people can be explained by the fact that they benefit from good working conditions with the protection of their jobs and attractive remuneration.
Young people who belong to a trade union are more likely to keep their jobs than other young people. Trade unions are workers' organizations which defend the interests of workers. As a result, union members have an influence in the organization and especially on employers. The union has a look at the organization's policy in terms of recruiting new workers and the training policy for agents, and in the event of a recession in activities, employers will lay off workers who do not belong to a movement syndical.
The "age" variable is significant at the 1% level. This reflects a positive effect on job retention. This means that the older the individual increases, the more likely the individual is to increase their chances of keeping their job. From this perspective, job retention is an increasing function of age due to the cumulative effects of the acquisition of experience which improve the productivity of workers. These results confirm the results of Contreras et al. (2011) regardless of the genre.
The variables married and divorced have negative effects on job retention compared to single people. This result seems paradoxical insofar as newlyweds, for example, should in reality keep their jobs given the mutual social pressure and possibly the social pressure due to dependent children.

Probabilities of Changing Jobs
Place of residence (urban) is significant at the 1% level with a negative effect in all three models, except in the selection equation where it has no effect on job change. Young people living in urban areas are less likely to change jobs com- Union membership has a negative effect on job change. Young people in a work organization are less likely to change jobs compared to other young people.
In the Congolese context, job offers are much lower than job demand. Young people are not tempted to want to change their jobs for lack of job opportunities.
Therefore, they prefer to keep their jobs, a result which corroborates with the dimension of job retention.
The social characteristics of the individual, the level of education of the mother has a negative effect on job change. The more the mother's level of education increases, the young people are less likely to change their jobs. In contrast, young people whose parents have post-secondary vocational education are more likely to change their jobs. The second result is consistent with Lin's theory, but contrary to that of Kamanzi (2006).

Conclusion
The results of this article have shown that the employability of young people depends on microeconomic and macroeconomic factors. On the macroeconomic level, current GDP per capita, gross current fixed capital formation as well as imports have an effect on employability.
From a microeconomic point of view, the results show that both individual and environmental factors are determinants of the employability of young people.
Indeed, an individual's education, gender and age, place of residence as well as union membership all have an effect on employability.
In view of these lessons, it is appropriate that a number of perspectives are emerging on both the job supply side and the job demand side. On the job supply side, economic policies should prioritize support for the private sector and the promotion of entrepreneurship, in particular through a reallocation of resources in productive sectors with high labor capacity. On the job demand side, it is opportune to re-frame the education system in order to improve the training offer. In addition, huge efforts are expected from the public authorities to reduce gender inequalities in order to improve the employability of young people and in particular that of young women. Likewise, the public authorities should promote trade union action in organizations because without it, all decisions in the organization would be taken unilaterally by the employers. This could lead, in particular to breach of employment contracts and unfair dismissals which infringe employability.