Formality and Innovation in French-Speaking Sub-Saharan African SME: Cases of Cameroon and Senegal

Abstract

Despite the importance of public policies in favor of the formalization of enterprises in French-speaking Sub-Saharan Africa, the productive fabric remains marked by a strong predominance of informal enterprises whose weight tends to limit the propensity of enterprises to innovate. In this context, becoming formal for an enterprise can improve the innovation capacity of enterprises. This article aims to analyze the role of formality on product, process, organizational and commercial innovations in Cameroon and Senegal. The results obtained using a sample of 1369 firms from data collected by the International Development Research Centre (IDRC) and logistic regression show that formal firms have a better innovation capacity. But the role of formality on innovation tends to be less important for Cameroonian firms. These results show that the Cameroonian authorities must intensify measures in favor of the formalization of enterprises to boost the potential for innovation within enterprises.

1. Introduction

For several years, innovation, which has become a full-fledged field of research in economics and management, has been the subject of abundant literature. Indeed, innovation enables Small and Medium-sized Enterprises (SME) to cope, by improving their competitiveness and preserving comparative advantage [1, 2, 3] to the uncertainty of the environment caused by the globalization of economies and the constant evolution of the market for their products due to the rapid dissemination of information [4]. Thus, an innovation is defined as the introduction on the market of a new or significantly modified product or the introduction of production processes, of methods concerning the provision of services or the delivery of products, of support activity new or significantly modified or the introduction of new or significantly improved solutions concerning the organization or marketing [5] impact the performance of SME [6, 7, 8]. However, this impact depends on the level of development of the economies.

In low-income countries in general and in Francophone Sub-Saharan Africa (FSA) in particular, the productive fabric is dominated by SME, the majority of which are informal. Although there are several obstacles to the expansion of these enterprises in these economies, including the weak incentive of government policies, access to finance [9] which tend to keep them informal, they are an important source of productivity [10]. The contribution of the informal sector is predominant in the national production of these economies. For example, in Cameroon, the informal sector represents 57.6% of national production and employs nearly 90% of workers [11]. Similarly in Senegal, the informal sector contributes 51.5% to national production and employs 48.8% of the active population. A major determinant of this economic dynamic is the innovation capacity of these firms because innovation allows each of them to achieve a positive performance [7, 12]. However, if formality plays an essential role in the performance of SME [13]; it can also strongly determine their capacity to innovate. Establishing the relationship between formality and innovation is relevant.

In the literature, innovation is the subject of several studies [14, 15, 16, 17]. That the majority of this work is concentrated in developed economies [18, 19, 20, 21], we note an increase in developing economies [22, 23]. These essentially attest that innovation determines the performance of SME. Furthermore, that this work has addressed the issue of innovation by highlighting the comparison by activity sector and by the size of the firms, neglecting the importance of the nature of the company. Innovation having made the postulate that firm are homogeneous. However, in these economies, there are concurrently formal firms and informal, each with innovation capabilities specific to their nature [24]. Only a few studies have attempted to fill this gap. So we offer many extensions on the existing literature.

First, considering the study of [25] of firms in Kenya, the authors analyze the fact for a firm to start its activities by being casual about its capacity for innovation. However, while they attest that a firm’s informal past harms its propensity to innovate, their study does not address the firm’s ability to develop several innovations simultaneously. Concerning their study, we analyze not the fact of starting informal, but the fact for the SME to carry out its activities in being formal or informal on innovation. We are therefore following the work of [26] carried out in Ghana without however analyzing the consequences on performance. Our analysis, therefore, aims to be integrative by understanding the effect of innovation on industrial firms as well as on service and trade companies, since the latter is predominant in ASF. Second, not only that we explore the formality-innovation link in SME, but we also take into account that SME are not a homogenous firm size category because of different resources, capabilities and obstacles they encounter [27]. Thus, we separately analyze not only SME companies, but also large firms for the comparative purposes. Third, we propose a comparative analysis between two countries (Cameroon and Senegal) using direct measures of innovation. Although these two countries are French-speaking, they nevertheless belong to different economic zones¹, making the comparison relevant. Therefore, given the low level of integration within EMCCA, we can expect that the challenges of innovation within Cameroonian SME will be lower than in Senegal. Finally, considering that companies develop several forms of innovation simultaneously, we propose to analyze the relationship between formality and frequency of innovation. So this paper aims to provide answers to the following research questions:

1. RQ1. How formality affect product, process, organizational or commercial SME innovation?

2. RQ2. What effect formality has on SME frequency innovation?

To answer these questions, the objective of this paper is therefore to examine the relationship between formality and innovation of SME in Cameroon and Senegal.

The data used come from IRDC as part of the project on the determinants of business performance in ASF. We use binomial logit model to evaluate the role of formality to the capacity of SME to develop product, process, commercial or organizational innovation and other hand multinomial regression to study the effect of formality to the capacity of SME to develop simultaneous different types of innovation. The results show that formality significantly determines the capacity of SME innovation. Besides, the role of formality varies depending on the type of innovation, the sector of activity, and the country.

The rest of the paper is organized as follows. Section 2 presents a review of the literature. Section 3 is devoted to the data and methodology of the study. Section 4 presents the results and interpretations. Section 5 concludes.

2. Literature review

A very recent empirical literature analyzes the impact of the coexistence of the formal sector and the informal sector on the capacity to innovate companies in developing countries. The results of this literature present two main strands.

The first shows intuitively that the coexistence in the same economy of the informal and formal sectors is harmful to business innovation [28]. Indeed, according to Mendi and Mudida [25], informal status negatively affects the innovation-decision of firms. Thus the marginal impact of informality on the innovation of formal enterprises decreases with the intensity of competitive pressure from informal enterprises. In this sense [29] highlight an inverted U-shaped relationship between the propensity to innovate and the competitive pressure of companies in the informal sector. For Ref. [30], this informal competition harms the innovation of formal companies. For Ref. [28] this negative impact is more noticeable at the level of product innovation and the most affected firms are those that lack a collaborative strategy with informal companies. According to Ref. [13], the start of the innovation-decision derives from the fact that informal enterprises provide low-quality products compared to those offered by formal enterprises since they share the same market. In this context, the low-income levels of consumers lead them to prefer low-quality products making it unprofitable for formal firms that offer superior quality products. For Ref. [31] it is the limitation on access to factors of production, especially human capital that is called into question. Indeed, the predominance of the informal sector in the economy distorts the accumulation of human capital, because the immediate availability of jobs requiring low qualifications in the informal sector, can discourage the accumulation of human capital, which makes this rarer factor. Advocates of this thesis, however, have overlooked the possibilities of interaction between firms in the two sectors.

The second cleavage shows with supporting examples that this coexistence of the two formal and informal sectors is a source of innovation [32]. Indeed, it leads to collaborations to innovate between firms [33]. For [34] the rapid democratization of innovation leads both consumers and companies to share information on the development of new products. This democratization of innovation is made possible by crowdfunding financing, which thus resolves the thorny problem of access to corporate finance [35]. Links Mhula et al. [36] show that the interaction between actors in rural areas in South Africa leads formal companies to develop innovations so the characteristics are similar to informal innovations. For Wiliams and Kosta [37], formal firms that do not perceive informal competition as an obstacle significantly increase their market shares than those which see this competition as an obstacle. Informal competition thus becomes a factor in boosting the innovation efforts of formal enterprises to cope with the rise of the informal sector. For Avenyo et al. [28], it is more intra-industry competition that improves product innovation. However, these authors show that local competition affects the product innovation strategies of large firms less than those of medium-sized firms. Based on the above arguments on the positive impact of formality on firms’ innovation, we formulate the following hypothesis:

1. H1. The impact of formality on each type of innovation (product, process, organizational or commercial) is positive in SME.

2. H2. Simultaneously investing in different forms of innovation (frequency innovation) is positive related to formality in the all firm size.

It emerges from this review that although SME are very exposed to competition from the informal sector, the effect of this sectoral duality in developing economies on business innovation remains to be determined. As this effect can differentiate the degree of the economic integration of countries, this study, therefore, aims to contribute by taking the example of two countries belonging to two different economic unions.

3. Methodology

3.1 Data and variable measures

The empirical analysis of this study uses data collected as part of the Determinants of FSA Firm Performance Project funded by IDRC. As in Cameroon, data collection in Senegal was carried out based on the general census of enterprises and surveys undertaken among the informal sector. Thus, our sample is made up of both informal and formal enterprises. An informal enterprise is an enterprise that does not keep accounts according to regulations, does not pay employees’ social contributions, and does not file statistical tax returns.

After analysis and processing of the data, in particular the missing data, we have a total sample of 1369, i.e. 642 in Cameroon and 727 in Senegal. Table 1 in the appendix shows the distribution of firms. It appears that the sample is made up of 59.82% of very small enterprises (VSE), 14.90% of small enterprises (SE), 10.08% of medium-sized enterprises (ME), and 15.19% of large enterprises. (LE). Depending on the formality, 68.22% of Cameroonian businesses are informal and 31.78% formal against 60.11% informal and 39.89% formal in Senegal. The breakdown by sector of activity shows that 23.05% of Cameroonian companies are in the industrial sector against 29.99% in Senegal. Unlike Cameroon, where we find 46.11% of companies in the services, Senegal has 29.16%. In commerce, Senegalese companies are more important than in Cameroon with 38.65% against 29.60%.

		Cameroon					Senegal
Designation		Number of firm	Innovative; product firms	Innovative; process firms	Innovative firms; in organization	Innovative; marketing firms	Number of firm	Innovative; product firms	Innovative; process firms	Innovative firms; in organization	Innovative; marketing firms
Firm size	VSE	394	200	188	201	222	425	139	137	116	103
	SE	67	29	32	33	42	137	44	56	61	42
	ME	55	22	25	26	35	83	26	28	34	29
	LE	126	52	51	59	61	82	34	45	42	39
firm formality	Formal	204	104	118	177	219	437	130	167	164	64
	Informal	438	199	178	142	141	290	144	99	89	149
Sector of activity	Industry	148	80	90	79	91	218	91	87	68	53
	Commerce	198	78	66	77	98	281	85	81	86	77
	Service	296	141	136	159	168	212	66	91	92	78

Table 1.

Description of the sample.

By innovation surveys carried out around the world (CIS survey), firms were questioned on four main forms of innovation, namely product, process, organizational and commercial innovations. Thus, we estimate the probability for each company to develop each of these innovations on the one hand and the probability of implementing several forms of innovation on the other hand. From Chart 1 in the appendix, it emerges that it is commercial innovation which is the first form of innovation (46.11%) developed in Cameroon, while it is the last in Senegal (29.30%). The analysis according to the formality highlights a better capacity of informal enterprises to develop product innovations (30.99% in Cameroon and 19.80% in Senegal) and commercial (21.96% in Cameroon and 20.49% in Senegal). As for formal Cameroonian firms, they are more oriented towards organizational innovations (27.57%) and commercial (35.78); while formal Senegalese companies innovate the most in the process (23.19%) and organization (22.55%).

Table 2 in the appendix presents the description of the study variables. The probability of innovating is explained by two models. In the first model, where we explain the probability of the company to innovate in a product, process, marketing or organization each of the variables to be explained is binary. These variables were constructed based on a set of questions asked to firms. For product innovation, companies were asked were some of your product innovations introduced between 2011 and 2013 new to your market, or new to your business. Or significantly improved? For process innovations, between 2011 and 2013, did your company introduce any new features or significant improvements concerning your manufacturing processes? Your logistics methods? Your support or support activities? For organizational innovations, the question was whether between 2011 and 2013, the firm introduced innovations relating to new operating methods for the organization of procedures, new methods of work organization, and decision-making, external relations with other companies or organizations. Finally, for commercial innovations, it was asked whether between 2011 and 2013, the firm introduced marketing innovations relating to significant changes in the presentation of a product, the use of new techniques or new media for the promotion of products, significant new methods of selling or distribution or new pricing strategies for their products.

Variable	Définitions
Dependent variables
in_produ	Binary variable equal to 1 if the firm declares to have innovated a product and 0 if not.
in_proce	Binary variable equal to 1 if the firm declares to have innovated in process and 0 if not.
in_orga	Binary variable equal to 1 if the firm declares to have developed an organizational innovation and 0 if not.
in_come	Binary variable equal to 1 if the firm declares to have developed a business innovation and 0 if not.
freq_innov	Categorical variable equal to 0 for non-innovative firms, to 1 for firms that have developed only one form of innovation, to 2 for two forms, to 1 for three forms and to 4 for all four forms.
Explanatory variables
Formality_en	Binary variable equal to 1 if the firm is formal and 0 if not.
localisation	Binary variable equal to 1 if the firm is located in an urban center and 0 if not.
r_dev	Binary variable equal to 1 if the firm carries out R&D activities and 0 if not.
Lage	Logarithm of the age of the firm
taille_ent	Categorical variable equal to 0 for very small enterprises (VSE), 1 for small enterprises (SE), 2 for medium enterprises (ME) and 3 for large enterprises (LE)
Tic	Binary variable equal to 1 if the firm has at least one tick equipment and 0 if not.
Leffec	Logarithm of the number of employees
Ltic	Logarithm of the number of ict equipment in the firm
tranche_age	Manager’s age group
instruc_ma	Categorical variable equal to 0 for managers with no formal education, 1 for those with primary education, 2 for secondary (2) and 3 for higher.
sex_ma	Binary variable equal to 1 if the manager is male 0 if not.
fortec_ma	Binary variable equal to 1 if the manager has a technical training related to the main activity of the company and 0 if not.
relig_ma	Categorical variable equal to 0 for animist managers, equal to 1 for Muslim managers, equal to 2 for Catholics, equal to 3 for Protestants and equal to 4 for other Christians.
stat_ma	Categorical variable equal to 0 for widowed managers, equal to 1 for single managers, equal to 2 for divorced managers and equal to 3 for married managers.

Table 2.

Description of variables.

As for the probability of implementing several innovations, the distribution of this variable (frequency of innovation) shows that firms belong to five groups: those not innovating at all, those who innovate in one form of innovation, those developing two forms of innovation, those which implement three forms and those which manage to develop the four forms of innovation.

The explanatory variables group together the characteristics of the firm, in particular the location, size, number of employees, age, ICT [38, 39]. Besides, we have associated with the manager’s characteristics, namely: gender, marital status, religious affiliation, level of education, technical training, and age group [40, 41]. We have introduced the sectoral indicators which allow us to capture the specificities linked to each sector of activity of the company Table 3 in the appendix presents the descriptive statistics of the different variables.

Variables	Observations	Mean	Standard deviation
in_produ	1369	0.3995617	0.0132429
in_proce	1369	0.4105186	0.0133002
in_orga	1369	0.4178232	0.0133346
in_come	1369	0.4185537	0.0133379
freq_innov	1369	1.646457	0.0413806
r_dev	1369	0.0759679	0.0071633
formality_en	1369	0.3608473	0.0129844
localisation	1369	0.8071585	0.0106669
taille_ent	1369	0.806428	0.0304609
leffec	1369	1.778067	0.0444247
tic	1369	0.6639883	0.0127707
lage	1369	2.363425	0.0194964
ltic	1369	1.233775	0.0226089
sex_ma	1369	0.7998539	0.0108177
stat_ma	1369	2.352082	0.0268715
relig_ma	1369	1.517896	0.0249713
instruc_ma	1369	1.719503	0.0297926
fortec_ma	1369	0.5617239	0.0134151
tranche_age	1369	0.8239591	0.03189

Table 3.

Descriptive statistics for variables.

3.2 Methods

Concerning the nature of the variables to be explained (binary and categorical), we use two different estimation techniques.

3.2.1 Effect of formality on firm innovation

The objective is to distinguish firms that have innovated from others, the dependent variable is dichotomous. It takes the value 1 when the company innovates and the value 0 otherwise. This is why the binomial logit is used to estimate the probability of innovating. Let $\mathit{\pi }$ be the firm’s decision to innovate, ${\mathit{\pi }}_{\mathit{i}}\mathit{=}\mathrm{1}$ if there is innovation and ${\mathit{\pi }}_{\mathit{i}}\mathit{=}\mathrm{0}$ if not. The prediction made through this model makes it possible to quantify the strength of the link between the explanatory variables and the explained variable representing the propensity to innovate [42]. The technique thus adopted does not impose restrictions on the conditions of normality of the explanatory variables, nor does it impose any restrictions on the discrete nature or not of these variables. We can therefore admit that the diversified nature of the explanatory variables, the hypothesis of the non-linearity of the relationship between the decision to innovate and the explanatory variables characterizing it, as well as the recognized flexibility of the logistics models, justify the option taken here to study the relationship between formality and the propensity to innovate.

We assume that a firm’s decision to innovate is a function of the likely use of its experiences in innovating. Let ${\mathit{\pi }}^{\mathit{\ast }}$the latent variable representing the firm’s propensity to innovate vary from $-\infty$ to $+\infty$. This variable is determined by explanatory variables describing the specificities of the innovation so that we have the following equation:

$${\pi }^{\ast }=\alpha +X_i\beta +{\epsilon }_i$$E1

Where $i$indicate the observation, $\beta$is the vector of parameters to be estimated, $X$ the matrix of independent variables, and the error term, $\epsilon$which asymptotically follows a normal distribution.

By considering that the probable utility of innovating is $U_A\left({\mathit{\pi }}^{\mathit{\ast }}\right)$ and the probable utility of not innovating is $U_N\left({\mathit{\pi }}^{\mathit{\ast }}\right)$ and the latent variable is ${\pi }^{\ast }$, we have:

$${\pi }_i=\left\{\begin{array}{c}1\mathit{\text{if}}{\pi }^{\ast }>0\mathit{\text{or}}{U}_A\left({\pi }^{\ast }\right)>U_N\left({\pi }^{\ast }\right)\\ 0\mathit{\text{if}}{\pi }^{\ast }<0\mathit{\text{or}}{U}_A\left({\pi }^{\ast }\right)<U_N\left({\pi }^{\ast }\right)\end{array}\right.$$E2

In our case, the utility of innovating $U\left({\pi }^{\ast }\right)$is assumed to be related to all of the specific characteristics of the innovation and of the company as defined in Eq. (1). These are the characteristics of the company, the characteristics of the manager to which we associate the sector indicators. Therefore, the probability that a company innovates, for a given value of $x$ can be expressed as follows:

$$Pr\left(\pi =1/x\right)=Pr\left({\pi }^{\ast }>0/x\right)$$E3

By integrating the structural model obtained in Eq. (1) into Eq. (3) and by rearranging the terms, the probability of innovating by a company becomes:

$$Pr\left(1/x\right)=Pr\left(\epsilon >-\left(\alpha +\mathit{\beta x}\right)\right)$$E4

Thus, the empirical model used to estimate the probability of innovating takes the form:

$$\mathit{\text{Prob inno}}=log\left(\frac{P_i}{1-P_i}\right)={\beta }_0+{\beta }_1\mathit{\text{formality}}\_{\mathit{en}}_i+\sum _{k=2}^K{\beta }_k{Z}_i+\text{ssect}\_{\mathrm{act}}_i+{\epsilon }_i$$E5

Where $\frac{P_i}{1-P_i}$ is the ratio between the probability that a company $i$ innovates and the probability that it does not innovate. Our variable of interest $\mathit{\text{formality}}\_{\mathit{en}}_i$ corresponds to whether the firm is formal or not. $Z_i$ is the vector of control variables that groups together the characteristics of the firm and the manager. Also, sectoral indicators are introduced. ${\epsilon }_i$ the error term comprising all the variables that can explain the probability of innovating and which have not been considered here. Finally, it should be noted that Eq. (5) is estimated for each type of innovation, namely product, process, organizational, and commercial.

3.2.2 Effect of formality on firm frequency innovation

To analyze the role of formality on the frequency of innovation, we use a multinomial logit model. It is a model that conceptualizes utilities in terms of a firm’s likelihood of developing some form of innovation. Thus, the choice of a form of innovation $j$ depends on the gain in terms of utility that the firm achieves compared to other forms of innovation. Thus, the firm to make a choice ranks the utility functions of all categories of innovation and chooses the highest one. The utility of modality j depends on company ${\gamma }_i$ and takes the form:

$$U_j=X_j\beta +{\epsilon }_i$$E6

In other words, modeling a choice $j$ for a company $n$ in $m+1$ modalities takes form:

$$\mathit{pro}\mathrm{b}\left(y_n=j\right)=p\left(u_j>u_0,u_j>u_1\dots u_j>u_k>u_m\right)$$

$$=p\left(X_{\mathit{nj}}\beta +{\epsilon }_j>X_{\mathit{nk}}\beta +{\epsilon }_k\right)\mathit{\text{for}}k\ne j\mathit{\text{and}}k=0,1,\dots .$$

$$=p\left({\epsilon }_j-{\epsilon }_k>X_{\mathit{nk}}\beta -X_{\mathit{nj}}\beta \right)$$

$$=p\left({\epsilon }_j^{\ast }>X_j^{\ast }\right)\text{with},{\epsilon }_j^{\ast }={\epsilon }_j-{\epsilon }_k\text{et},X_j^{\ast }=X_k-X_{j.}$$E7

Considering Eq. (14), the probability becomes:

$$\mathit{pro}\mathrm{b}\left(y_n=j\right)=\frac{exp\left(X_{\mathit{nj}}{\beta }_j\right)}{{\sum }_{k=0}^{m}exp\left(X_{\mathit{nk}}{\beta }_j\right)}$$E8

Avec, $n=1,2,3\dots N,\mathit{\text{number of firms and}}$$k=1,2,3\dots m,$ choices.

By normalizing by ${\beta }_0=0$, we get:

$$\mathit{pro}\mathrm{b}\left(y_n=j\right)=\frac{exp\left(X_{\mathit{nj}}{\beta }_j\right)}{1+{\sum }_{k=1}^{m}exp\left(X_{\mathit{nk}}{\beta }_j\right)}$$E9

By considering that the parameters of this model are interpreted as a difference of ${\beta }_0$from the reference modality, we obtain:

$$\mathit{pro}\mathrm{b}\left(y_n=j\right)=\frac{\frac{exp\left(X_{\mathit{nj}}{\beta }_j\right)}{exp\left(X_n{\beta }_0\right)}}{{\sum }_{k=1}^m\frac{exp\left(X_{\mathit{nk}}{\beta }_k\right)}{exp\left(X_n{\beta }_0\right)}}$$E10

$$\mathit{pro}\mathrm{b}\left(y_n=j\right)=\frac{exp\left(X_n\left({\beta }_j-{\beta }_0\right)\right)}{{\sum }_{k=1}^m\left(X_n\left({\beta }_k-{\beta }_0\right)\right)}$$E11

$$\mathit{pro}\mathrm{b}\left(y_n=j\right)=\frac{exp\left(X_n{\beta }_j^{\ast }\right)}{{\sum }_{k=1}^{m}exp\left(X_n{\beta }_k^{\ast }\right)}$$E12

with and ${\beta }_j^{\ast }={\beta }_j-{\beta }_0$ et ${\beta }_k^{\ast }={\beta }_k-{\beta }_0$

The $m+1$ the probabilities are therefore estimated as a function of $m\left(m+1\right)/2$ differences in ${\beta }_k-{\beta }_j$. The ratio of probabilities leads to a simple linear function of the form:

$$\frac{\mathit{\text{prob}}\left(Y_n=j\right)}{\mathit{\text{prob}}\left(Y_n=p\right)}=\frac{\frac{exp\left(X_n{\beta }_j\right)}{{\sum }_{k=0}^{m}exp\left(X_n{\beta }_k\right)}}{\frac{exp\left(X_n{\beta }_p\right)}{{\sum }_{k=0}^{m}exp\left(X_n{\beta }_k\right)}}=exp\left(X_n\left({\beta }_j-{\beta }_p\right)\right)$$E13

Related to the logarithm, the equation results in a simple linear function:

The categorical variable frequency of innovation (freq_innov) measuring the ability to implement several forms of innovation being defined according to five modalities, the empirical equation to be estimated is written:

$$\text{freq}\_\text{innov}=P\left[j/X_i\right]=\frac{\mathit{exp}\left(X_i{\beta }_j\right)}{{\sum }_{h=0}^2\mathit{exp}\left(X_i{\beta }_h\right)}$$E14

with,

$$j=\left\{\begin{array}{l}0\text{non}-\text{innovative firms}\\ 1\text{innovative firms in}\mathrm{one}\text{form of innovation}\\ 2\text{innovative firms in}\mathrm{two}\text{form}s\mathit{\text{of}}\text{innovation}\\ 3\text{innovative firms in three forms of innovation}\\ 4\text{innovative firms in four forms of innovation}\end{array}\right.$$

Statistics show that since non-innovative firms are the most numerous in the base, they are considered as a reference group. Thus, we get:

$$\left\{\begin{array}{l}P\left[0/X_i\right]=\frac{1}{1+{\sum }_{h=1}^2\mathit{exp}\left(X_i{\beta }_h\right)},\mathit{\text{for}}j=0\\ P\left[j/X_i\right]=\frac{1}{1+{\sum }_{h=0}^2\mathit{exp}\left(X_i{\beta }_h\right)},\mathit{\text{for}}j=1,2,3,4\end{array}\right.$$E15

Where $X_i$ is the vector of the explanatory variables. The final model to be estimated is as follows:

$$\text{freq}\_\text{innov}=\mathit{log}\left(\frac{P\left[j/X_i\right]}{P\left[0/X_i\right]}\right)=X_i{\beta }_j={\beta }_{j0}+{\beta }_{j1}\mathit{\text{formality}}\_{\mathit{en}}_i+\sum _{k=2}^K{\beta }_{\mathit{jk}}{Z}_i+\mathit{\text{ssect}}\_\mathit{act}+{\epsilon }_i$$E16

Where $ssect\_\mathit{act}$ represents the sectoral indicators.

We estimate this equation using a multinomial logit whose structure allows us to determine the influence of formality on the probability of belonging to each of the categories of companies mentioned above (j = 0,1,2,3,4) compared to those having not innovated during the study period (j = 0). Note that the explained variable being an intensity variable, we also estimated the ordered model as is often the case. The results being the same, in particular the positive effect of formality on the frequency of firm innovation, we only present the results of the unordered model.

4. Results and discussions

This section presents the results obtained from the models previously presented. First, we discuss the role of formality on the propensity of firms to develop each type of innovation. Second, we discuss the relationship between formality of firm frequency of innovation. The robustness of the results has been verified using a variance inflation test (VIF) (Table 4).

4.1 Formal firms have a better propensity to innovate

Tables 5 and 6 in the appendix present the results of the determinants of product and process innovation as well as organizational and commercial innovations for each of the countries and the overall sample of companies. Globally, the results show a positive relationship between formality and the propensity of companies to innovate in both Cameroon and Senegal. In other words, formal enterprises compared to informal enterprises have a better capacity for innovation. Thus, the fact that a company is formal increases its probability of innovating in products by 43.86% and the process by 51.66% in Senegal against 35.85% in product and 36.16% in process in Cameroon. In contrast, when considering organizational and commercial innovations, formality increases the probability of innovating by 51.33 and 57.15% in Cameroon against 48.70 and 45.07% in Senegal respectively. When the sectors of activity are taken into account (Tables 7–10), formality remains an important factor in the capacity for innovation of companies, although its role is different. In the case of product innovation, we note that formal Senegalese companies have a better propensity to innovate regardless of the sector of activity. In contrast in Cameroon, formality does not seem to be a determinant of product innovation. By considering process innovation, the results attest that the formality contributes to the innovation of Cameroonian firms in the industrial and commercial sectors. In Senegal, it appears that it improves the propensity to innovate in the three sectors of activity. As for organizational innovation, formality appears to be decisive in industry and service companies in Cameroon. On the other hand in Senegal, it determines the innovation capacity of trade and service companies. Finally, about commercial innovations, the results show that formality is a determining factor in Cameroon in trade and Senegal in trade and services.

Cameroun			Senegal
Variable	VIF	1/VIF	Variables	VIF	1/VIF
Lage	4.71	0.129740	Lage	4.22	0.236735
tranche_age	4.7	0.140963	tranche_age	3.77	0.265563
formality_en	4.76	0.210199	formality_en	2.46	0.407226
leffec	4.49	0.222811	Leffec	1.75	0.570786
Ltic	1.47	0.679462	instruc_ma	1.73	0.576618
Tic	1.47	0.680618	Ltic	1.58	0.633923
instruc_ma	1.29	0.772606	taille_ent	1.30	0.772003
sex_ma	1.14	0.879130	r_dev	1.28	0.782305
fortec_ma	1.12	0.896573	stat_ma	1.23	0.815709
r_dev	1.1	0.910413	localisation	1.16	0.862589
relig_ma	1.07	0.936015	sex_ma	1.14	0.874633
stat_ma	1.05	0.948345	fortec_ma	1.13	0.886794
localisation	1.05	0.954916	tic_a	1.12	0.893054
taille_ent	1.02	0.979067	relig_ma	1.10	0.905723

Table 4.

Multi-collinearity test of the variables.

	Product innovation			Process innovation
Variables	Cameroon	Senegal	Full sample	Cameroon	Senegal	Full sample
r_dev	1.159***	1.165***	1.216***	1.937***	0.964**	1.588***
	(0.348)	(0.353)	(0.240)	(0.421)	(0.396)	(0.282)
formality_en	0.691*	0.870***	0.170**	0.707*	1.197***	0.433**
	(0.395)	(0.256)	(0.192)	(0.403)	(0.254)	(0.193)
localisation	1.255***		0.537***	0.981***		0.419***
	(0.311)		(0.159)	(0.309)		(0.161)
VSE		0.274**	0.307*
		(0.228)	(0.170)
ME		0.443*	0.364*		0.588**
		(0.290)	(0.203)		(0.300)
LE	0.383*
	(0.218)
Leffec	0.258***	0.152**	0.0878*	0.377***	0.335***	0.206***
	(0.0939)	(0.0760)	(0.0509)	(0.0983)	(0.0788)	(0.0527)
Tic			0.308**		0.0653**	0.363***
			(0.134)		(0.262)	(0.138)
lage			0.287*
			(0.166)
Ltic	0.410**		0.250***			0.157*
	(0.176)		(0.0834)			(0.0857)
sex_ma	0.327*
	(0.193)
stat_ma
relig_ma			0.123*
			(0.0662)
instruc_ma		0.325***		0.193*		0.171***
		(0.0919)		(0.108)		(0.0650)
fortec_ma		0.784***	0.285**		0.503***
		(0.183)	(0.120)		(0.183)
tranche_age			−0.248**		−0.284**	−0.298***
			(0.0974)		(0.135)	(0.0998)
Constant	−1.781*	−2.372***	−2.035***	−1.714*	−2.719***	−2.017***
	(0.995)	(0.588)	(0.407)	(1.012)	(0.602)	(0.417)
Observations	642	727	1369	642	727	1369

Table 5.

Estimation of the probability of product and process innovation.

*

p < 0.1.

**

p < 0.05.

***

p < 0.01.

Standard errors in parentheses.

	Organizational innovation			Marketing innovation
Variables	Cameroon	Senegal	Full sample	Cameroon	Senegal	Full sample
r_dev	1.369***	1.289***	1.362***	1.196***	0.684*	1.065***
	(0.402)	(0.403)	(0.280)	(0.398)	(0.352)	(0.257)
formality_en		1.086***	0.708***		1.378***	0.678***
		(0.253)	(0.195)		(0.268)	(0.197)
localisation	1.213***		0.551***	1.202***		0.598***
	(0.313)		(0.164)	(0.291)		(0.165)
PE		0.643***
		(0.226)
ME
GE	0.281**			0.396*
	(0.228)			(0.221)
leffec	0.338***	0.190**	0.170***	0.258***	0.257***	0.132**
	(0.102)	(0.0783)	(0.0534)	(0.0996)	(0.0825)	(0.0535)
tic			0.307**			0.590***
			(0.140)			(0.140)
lage
ltic	0.130**
	(0.180)
sex_ma
stat_ma			0.110*		0.239**	0.195***
			(0.0637)		(0.119)	(0.0635)
relig_ma				0.0414**		0.202***
				(0.0825)		(0.0682)
instruc_ma			0.211***			0.243***
			(0.0659)			(0.0664)
fortec_ma					0.395**
					(0.197)
tranche_age						−0.198**
						(0.101)
Constant	−2.071**	−2.038***	−1.956***	−0.177***	−1.999***	−1.886***
	(1.026)	(0.590)	(0.424)	(1.014)	(0.617)	(0.426)
Observations	642	727	1369	642	727	1369

Table 6.

Estimation of the probability of organizational and marketing innovation.

*

p < 0.1.

**

p < 0.05.

***

p < 0.01.

Standard errors in parentheses.

	Cameroon			Senegal
Variables	Industry	Trade	Service	Industry	Trade	Service
r_dev	2.838**	1.776**	0.878*			1.482***
	(1.272)	(0.868)	(0.450)			(0.536)
formality_en				1.304**	1.278***	1.112**
				(0.660)	(0.444)	(0.509)
localisation	1.804***	1.550**	0.903*	0.312***	0.331**	0.0301***
	(0.612)	(0.643)	(0.503)	(0.561)	(0.336)	(0.405)
VSE
ME	1.453*
	(0.782)
LE	0.815***	0.871**		0.500**	0.911*
	(0.518)	(0.421)		(0.657)	(0.530)
leffec	0.233***		0.325**	0.778***	0.552***
	(0.197)		(0.134)	(0.211)	(0.157)
tic		0.365**	0.0377***	0.293**	0.506*	0.933*
		(0.389)	(0.334)	(0.624)	(0.447)	(0.532)
lage				0.0425***	0.262**
				(0.534)	(0.403)
ltic	0.691*	0.571*	0.280***	0.214*	0.0954***	0.172**
	(0.410)	(0.311)	(0.293)	(0.289)	(0.205)	(0.199)
sex_ma			0.603**		0.909*
			(0.284)		(0.476)
stat_ma		0.251*
		(0.151)
relig_ma
instruc_ma	0.0849**		0.0643**	0.412**
	(0.273)		(0.150)	(0.199)
fortec_ma				1.328***		0.832**
				(0.500)		(0.389)
tranche_age
Constant	−0.327**	−4.264**	−1.426	−3.594**	−2.957***	−2.838**
	(2.293)	(2.072)	(1.447)	(1.435)	(1.133)	(1.165)
Observations	148	190	296	218	281	212

Table 7.

Estimation of the probability of product innovation by industry sector.

*

p < 0.1.

**

p < 0.05.

***

p < 0.01.

Standard errors in parentheses.

	Cameroon			Senegal
Variables	Industry	Commerce	Service	Industry	Commerce	Service
r_dev			1.839***	1.475*	2.605**	1.670**
			(0.545)	(0.857)	(1.126)	(0.709)
formality_en	0.129**	0.1638**		1.211**	0.794*	2.616***
	(0.868)	(0.325)		(0.583)	(0.422)	(0.551)
localisation	1.408**	0.755*	1.004*
	(0.579)	(0.618)	(0.544)
SE				0.208**	0.859**
				(0.442)	(0.377)
ME				0.401***	0.143***
				(0.621)	(0.476)
LE	0.451**	−0.845*		0.745**	0.833*
	(0.513)	(0.450)		(0.617)	(0.504)
Leffec	0.185***		0.394***	0.219**	0.292**	0.0737***
	(0.200)		(0.145)	(0.165)	(0.147)	(0.148)
tic
lage				−0.677*
				(0.406)
ltic	0.294*	0.553*	0.279**
	(0.416)	(0.330)	(0.317)
sex_ma	0.314**
	(0.486)
stat_ma	0.335**
	(0.170)
relig_ma
instruc_ma	0.279*	0.563**	0.0412*
	(0.278)	(0.240)	(0.155)
fortec_ma	0.00840**	0.378	−0.0695	0.0140***
	(0.409)	(0.361)	(0.284)	(0.371)
tranche_age
Constant	−1.929	−2.395	−1.935	−1.176	−2.498**	−2.741**
	(2.241)	(2.159)	(1.552)	(1.125)	(1.101)	(1.181)
r_dev	−0.173	1.336	1.839***	1.475*	2.605**	(1.126)
Observations	139	190	296	218	281	212

Table 8.

Estimation of the probability of process innovation by industry sector.

*

p < 0.1.

**

p < 0.05.

***

p < 0.01.

Standard errors in parentheses.

	Cameroon			Senegal
Variables	Industry	Commerce	Service	Industry	Commerce	Service
r_dev			2.050***
			(0.663)
formalite_ent	0.5590758*		03653788**		1.180***	1.628***
	(0.954)		(0.314)		(0.428)	(0.486)
localisation	3.056***		0.593**
	(0.864)		(0.480)
SE		1.205*	0.116*
		(0.630)	(0.481)
ME
LE		0.757*
		(0.424)
leffec	0.835***	0.323***		0.405***	0.669***	0.0324*
	(0.265)	(0.263)		(0.155)	(0.154)	(0.139)
tic	0.841*	0.239*	0.393***	0.228**	0.452*	0.196*
	(0.484)	(0.378)	(0.347)	(0.498)	(0.425)	(0.490)
lage				0.0429**	0.225*	0.509**
				(0.373)	(0.395)	(0.425)
ltic	0.535**	0.0420*	0.0763***	0.0607**	0.208***	0.202***
	(0.445)	(0.304)	(0.315)	(0.207)	(0.197)	(0.205)
sex_ma	0.343***	0.394*	0.455***
	(0.516)	(0.374)	(0.293)
stat_ma		0.0619**	0.308**
		(0.153)	(0.135)
relig_ma
instruc_ma	−0.193***	0.508**	0.00201*
	(0.336)	(0.223)	(0.162)
fortec_ma	0.832*			1.031***	0.0189**	0.134**
	(0.459)			(0.371)	(0.332)	(0.363)
tranche_age				−0.0116**
				(0.266)
Constant	−2.291	−3.572*	−0.732	−3.161***	−2.893**	−2.712**
	(2.369)	(2.056)	(1.643)	(1.144)	(1.134)	(1.120)
Observations	148	190	296	218	281	212

Table 9.

Probability of organizational innovation by sector of activity.

*

p < 0.1.

**

p < 0.05.

***

p < 0.01.

Standard errors in parentheses.

	Cameroon			Senegal
Variables	Industry	Commerce	Service	Industry	Commerce	Service
r_dev			1.279**		1.876*	1.061*
			(0.543)		(1.101)	(0.544)
formality_ent		1.088*			1.452***	2.001***
		(0.050)			(0.134)	(0.530)
localisation	0.0331*	1.001*	1.383***
	(0.571)	(0.540)	(0.506)
SE						−0.845*
						(0.458)
ME	1.121**	0.682***	0.163*			−1.030*
	(0.550)	(0.550)	(0.471)			(0.575)
GE	0.442*	−0.871**	0.0751
	(0.235)	(0.415)	(0.347)
leffec	0.634*	0.00762**	0.247*	0.323**	0.451***	−0.163**
	(0.461)	(0.253)	(0.143)	(0.151)	(0.147)	(0.148)
tic				0.0922***	0.187***	0.177**
				(0.479)	(0.441)	(0.509)
lage	1.021**		0.880***
	(0.449)		(0.718)
Ltic				0.125***	0.273*	0.192**
				(0.203)	(0.192)	(0.211)
sex_ma
stat_ma	0.535**	0.100**				0.559**
	(0.252)	(0.152)				(0.224)
relig_ma		0.454**
		(0.179)
instruc_ma		0.378*		0.310*	0.305*	0.0735***
		(0.211)		(0.171)	(0.175)	(0.175)
fortec_ma				1.345***	0.373	0.822**
				(0.359)	(0.321)	(0.390)
tranche_age
Constant	−0.559	−0.496	0.285	−3.494***	−1.882	−0.513
	(0.955)	(1.998)	(1.613)	(1.166)	(1.151)	(1.135)
Observations	139	190	296	218	281	212

Table 10.

Probability of marketing innovation by industry sector.

*

p < 0.1.

**

p < 0.05.

***

p < 0.01.

Standard errors in parentheses.

Beyond formality, our results show that several other factors contribute to the implementation of innovations within companies. In line with work carried out in Africa [43], our study shows that the size of the company is an important determinant of innovation, although its role is not systematic according to the forms of innovation. Besides, companies located in large urban centers have a better capacity for innovation. They seem to benefit from the exchange of information and collaborations with other companies in the innovation process. Also, our results show that the age of the company determines the capacity for innovation, with a more determining effect in Senegalese companies. However, we note that in the Senegalese industrial sector, its role is negative in the context of process innovations. This shows that in this sector, it is young companies that are more oriented towards innovation. We also note the important role of ICT in the innovation process in both Cameroon and Senegal. Our results also confirm the important role played by the number of employees in innovation. When taking into account the types of innovation, companies with a large number of employees exhibit a better capacity for innovation. Taking into account the sectors of activity, it emerges that the number of employees is decisive in the industry both in Cameroon and in Senegal regardless of the type of innovation. However, our results show a negative effect on the number of employees in Senegalese service companies that develop business innovations. This result thus shows that it is important for companies to emphasize the qualifications of employees, a source of innovation, and not only on the workforce, as is very often the case in French-speaking Sub-Saharan Africa.

Contrary to certain results obtained in developing countries and Africa in particular, our analysis underlines the determining role of R&D activities in the implementation of innovations in Cameroon and Senegal. Indeed, it emerges that companies that have an R&D activity have a better capacity for innovation. However, our analysis shows that the role of R&D is not systematic when we take into account the forms of innovation according to the sectors of activity. It appears that R&D seems to be decisive in the industrial sector as well as in services in Cameroon, while in Senegal R&D improves the product innovation capacity of service companies. In terms of process innovations, R&D plays an important role in service companies in Cameroon and all sectors in Senegal. Finally, by considering organizational and commercial innovations, we see that R&D is decisive mainly within service companies. Our study attests to the results obtained by Ref. [44] and shows that R&D activities must be taken into account when defining innovation policies in Africa [45]. But this must be done taking into account sector specificities insofar as R&D also plays a decisive role in service companies.

Finally, our work shows that firms that develop innovations in Africa as in other environments fundamentally rely on the characteristics of the manager as well as his qualifications and skills. In this sense, it emerges that the technical training of the manager increases the propensity of companies to innovate, with a greater effect in Senegal. We also note the role of manager education as well as age, the sometimes negative effect of which shows that companies with young managers have a better capacity for innovation. Finally, like [46] in Tanzania, our work emphasizes that gender determines innovation.

4.2 The frequency of innovation is positively linked to the formality of the firms

The results in Table 11 validate the idea that the frequency of innovation increases with the formality of the business. The graphs in Figure 1 which present for the four forms of innovation the evolution of the marginal effect as a function of the formality of the firm attest to the positive relationship between the formality of the firm and the observed innovation capacity previously. In the case of Cameroon, the fact that a company is formal increases its frequency of developing innovation by 14.60, 23.11% for two forms, 15.35% for three forms, and 18.69% for the four forms of innovation. As for Senegal, formality increases the frequency of innovating by 22.96% for one form, by 21.38% for two forms, by 13.62% for three forms, and by 19.40% for all four forms innovation. However, we see that this increase is less significant in Cameroon when the company develops from three forms of innovation.

Variables	Cameroon				Senegal
	One form	Two forms	Three forms	Four forms	One form	Two forms	Three forms	Four forms
r_dev						2.405**	2.480**	2.987***
						(1.080)	(1.072)	(1.052)
formalite_ent	0.243*	0.268**	0.496*	0.562**	1.688***	1.438***	1.610***	2.076***
	(0.692)	(0.622)	(0.623)	(0.601)	(0.364)	(0.350)	(0.422)	(0.407)
localisation	0.931**	1.780***	1.618***	1.837***	1.258***	1.019***
	(0.434)	(0.474)	(0.461)	(0.459)	(0.318)	(0.300)
PE					0.326*	0.391**
					(0.341)	(0.287)
ME						0.278*	0.501**	0.531***
						(0.398)	(0.470)	(0.442)
GE			1.012***	0.515*		0.0321***	0.427***	0.0725***
			(0.372)	(0.312)		(0.416)	(0.485)	(0.432)
leffec	0.315*	0.188**	0.547***	0.608***	−0.215*		0.484***	0.328***
	(0.177)	(0.164)	(0.157)	(0.151)	(0.118)		(0.123)	(0.119)
tic				0.261**			0.224**	0.151***
				(0.305)			(0.453)	(0.404)
lage			1.255*	0.0444**		0.521*
			(0.672)	(0.657)		(0.294)
ltic	0.158	0.0619**	0.269***	0.434***			0.0810***	0.255***
	(0.323)	(0.250)	(0.265)	(0.265)			(0.179)	(0.179)
sex_ma					0.625*	0.163**
					(0.345)	(0.376)
stat_ma
relig_ma		0.197**				0.346*	0.290**
		(0.126)				(0.255)	(0.283)
instruc_ma	0.154***	0.248*	0.201***	0.0757*
	(0.173)	(0.169)	(0.166)	(0.155)
fortec_ma			0.0254*	0.0745***			0.559*	0.951***
			(0.273)	(0.257)			(0.299)	(0.298)
tranche_age						−0.626***
						(0.226)
Constant	−1.764	−3.252**	−0.576	−2.831*	−1.555*	−3.766***	−4.284***	−3.253***
	(1.686)	(1.565)	(1.555)	(1.525)	(0.803)	(0.845)	(0.987)	(0.903)
Observations	642	642	642	642	727	727	727	727

Table 11.

Estimation of the multinomial logit model.

*

p < 0.1.

**

p < 0.05.

***

p < 0.01.

Standard errors in parentheses.

Figure 1.

Marginal effects of formality on the probability of innovating.

The results in Table 11 validate the idea that the frequency of innovation increases with the formality of the firm. The graphs in Figure 1, which show the evolution of the marginal effect as a function of firm formality for the four forms of innovation, attest to the positive relationship between firm formality and innovation capacity observed earlier. In the case of Cameroon, the fact that a company is formal increases its frequency of developing an innovation by 14.60, 23.11% for two forms, 15.35% for three forms and 18.69% for all four forms of innovation. As for Senegal, formality increases the frequency of innovation by 22.96% for one form, 21.38% for two forms, 13.62% for three forms and 19.40% for all four forms of innovation. However, this increase is less significant in Cameroon when the company develops from three forms of innovation.

The results by sector of activity (Tables 12–14) confirm those previously obtained, in particular, that formality positively affects companies’ propensity to innovate. However, we note that the relationship between formality and frequency of innovation is more decisive in the case of Senegalese companies.

Variables	Cameroon				Senegal
	One form	Two forms	Three forms	Four forms	One form	Two forms	Three forms	Four forms
r_dev			−4.334**	−1.817*
			(1.936)	(1.001)
formality_en			1.040***	1.252***	1.763**	0.151***	0.996**	1.718*
			(1.379)	(1.229)	(0.854)	(0.820)	(0.913)	(0.939)
localisation	2.993***	3.234***			2.601***	1.411**
	(1.033)	(1.150)			(0.876)	(0.709)
SE
ME			2.823**
			(1.296)
GE				0.341***
				(0.153)
Leffec	0.696**		−0.671*		−0.513**		0.900***	0.737***
	(0.327)		(0.354)		(0.252)		(0.283)	(0.281)
Tic	0.971**		0.839*	0.861***	1.151*
	(0.673)		(0.711)	(0.676)	(0.642)
Lage	0.263*	0.937**		1.661***
	(1.551)	(1.883)		(1.493)
Ltic	1.111*		2.751***
	(0.650)		(0.782)
sex_ma					1.809**
					(0.811)
stat_ma
relig_ma			0.820**
			(0.383)
instruc_ma				0.838*			0.0410**	0.282***
				(0.433)			(0.277)	(0.285)
fortec_ma			1.107*	1.066*			1.937***	2.538***
			(0.649)	(0.582)			(0.751)	(0.798)
tranche_age			1.613**	0.962**
			(0.794)	(0.698)
Constant	3.475	3.272	−9.387	2.714	−0.907	−4.258**	−4.927**	−5.673***
	(3.640)	(4.324)	(873.6)	(3.598)	(1.516)	(1.716)	(2.026)	(2.187)
Observations	148	148	148	148	218	218	218	218

Table 12.

Multinomial logit model estimation of firms in the industrial sector.

*

p < 0.1.

**

p < 0.05.

***

p < 0.01.

Standard errors in parentheses.

Variables	Cameroon				Senegal
	One form	Two forms	Three forms	Four forms	One form	Two forms	Three forms	Four forms
r_dev
formality_ent		0.236***			1.468**	1.921***	0.989**	1.512**
		(0.151)			(0.671)	(0.593)	(0.769)	(0.633)
localisation	1.348	2.206**
	(0.939)	(0.887)
SE	1.904*	1.558*		2.254*		0.106***	0.649*
	(1.067)	(0.903)		(1.200)		(0.511)	(0.623)
ME	0.632**	2.060*		0.375*
	(0.931)	(1.169)		(0.709)
LE	−1.389**	−1.034*		−1.954***
	(0.696)	(0.568)		(0.700)
Leffec				0.763*			0.586**	0.775***
				(0.396)			(0.262)	(0.220)
Tic				0.0402*			0.317*	0.503**
				(0.585)			(0.845)	(0.597)
Lage		2.736**
		(1.378)
Ltic	0.852*						0.0447**	0.155***
	(0.510)						(0.324)	(0.301)
sex_ma	0.133**						0.0440***	0.912**
	(0.607)						(0.686)	(0.762)
stat_ma
relig_ma	0.677**	0.227*	0.765**		0.466***	0.662*
	(0.302)	(0.258)	(0.344)		(0.444)	(0.434)
instruc_ma		0.496***	0.911**
		(0.330)	(0.382)
fortec_ma					0.0688*
					(0.472)
tranche_age		−1.234*
		(0.681)
Constant	−3.850	−8.431**	−20.05	−4.519	−1.599	−3.709**	−4.832**	−2.948*
	(3.412)	(3.320)	(1093)	(3.209)	(1.618)	(1.569)	(2.049)	(1.753)
Observations	190	190	190	190	281	281	281	281

Table 13.

Estimation of the multinomial logit model for trade firms.

*

p < 0.1.

**

p < 0.05.

***

p < 0.01.

Standard errors in parentheses.

Variables	Cameroon				Senegal
	One form	Two forms	Three forms	Four forms	One form	Two forms	Three forms	Four forms
r_dev								3.069**
								(1.253)
formalite_ent		0. 086**	0.705***		3.495***	2.804***	4.933***	4.346***
		(0.621)	(0.327)		(0.833)	(0.777)	(1.044)	(1.011)
localisation	2.283**	1.189*	1.222*	2.037**	1.845***	1.667***	0.635***	0.511**
	(1.103)	(0.673)	(0.731)	(0.849)	(0.673)	(0.608)	(0.654)	(0.707)
PE	1.465*	0.196**	1.141*			1.020*
	(0.776)	(0.798)	(0.684)			(0.578)
ME		0.979**		0.232**				1.149**
		(0.658)		(0.784)				(0.988)
GE		0.109**		0.445**				0.735***
		(0.559)		(0.497)				(0.964)
Leffec			0.582**	0.593**	−0.656***			−0.491*
			(0.236)	(0.231)	(0.227)			(0.260)
Tic			0.0504***	0.116*
			(0.496)	(0.502)
Lage					0.436*	0.962***
					(0.563)	(0.593)
Ltic			0.207**		0.568*	0.510*	0.440***	0.156***
			(0.450)		(0.291)	(0.284)	(0.361)	(0.368)
sex_ma	0.586**	0.999**	0.650**	0.683*	0.435**	0.159***		0.510*
	(0.481)	(0.446)	(0.432)	(0.421)	(0.730)	(0.775)		(0.929)
stat_ma				0.450**				0.834**
				(0.189)				(0.363)
relig_ma
instruc_ma		0.151**	0.105***	0.0387***		0.154**		0.248**
		(0.258)	(0.243)	(0.236)		(0.236)		(0.306)
fortec_ma			0.0653**	0.0987***	1.100**			1.913***
			(0.434)	(0.413)	(0.520)			(0.725)
tranche_age						−0.791*
						(0.444)
Constant	−6.267**	−3.089	−0.279	−2.953	−2.150	−4.382***	−18.34	−3.790*
	(2.988)	(2.537)	(2.463)	(2.485)	(1.649)	(1.683)	(547.7)	(1.978)
Observations	296	296	296	296	212	212	212	212

Table 14.

Estimation of the multinomial logit model for service firms.

*

p < 0.1.

**

p < 0.05.

***

p < 0.01.

Standard errors in parentheses.

In the industrial sector, formality improves the innovation capacity of Cameroonian companies which develop three and four forms of innovation. In Senegal, on the other hand, formality contributes to the propensity to innovate, both for companies that implement one or two forms of innovation and for those adopting three and four forms of innovation. In commercial enterprises, formality remains very decisive in the capacity for innovation of Senegalese companies. The results show that formality systematically impacts the frequency of innovation. In Cameroon, on the other hand, formality is only decisive for companies that adopt two types of innovation. Finally, in service companies, we see that formality is a factor of innovation. In Cameroon, it impacts the frequency of innovation of companies that develop two and three forms of innovation, while in Senegal, its role remains decisive for all forms of innovation.

The analysis of the other characteristics confirms the results obtained in the case of the different facets of innovation because it appears that the frequency of innovation depends on several factors. First, we observe that R&D improves the frequency of innovation only in Senegalese companies that implement at least two forms of innovation. This result suggests that despite resource constraints, companies consider it important to invest in improving manufacturing processes in support or support activities as well as logistics to hope to take advantage of the benefits and spinoffs related to this form of innovation, the impact of which is generally considerable [47]. Senegalese companies seem to be part of a dynamic contrary to that observed in the economies of the Maghreb since the work carried out in this context [48] shows that R&D is not a determining variable in the innovation process. Our analyzes, therefore, show that it is relevant to take into account forms of innovation rather than considering an aggregate measure of innovation.

In addition, we note that companies located in large urban centers have a better frequency of innovation. We also note that size plays a determining role. In the case of Cameroon, its role is decisive for companies that adopt at least three forms of innovation. In contrast, in Senegal, it improves the frequency of innovation for small businesses that adopt at most two forms of innovation as well as for medium and large businesses that develop at least two forms of innovation. The results also show that the number of employees improves the frequency of innovation, and we observe in the case of Senegal that its impact is more decisive for young companies that set up a single form of innovation. In addition, our analyses show that age and ICT improve the frequency of innovation. Like [49] in Nigeria, our analyses highlight the key role of manager characteristics in the capacity for innovation. It appears that the manager’s gender, level of education, age, and religious affiliation determines the frequency of innovation. Finally, the analysis according to the activity sectors also highlights the impact of the afore mentioned characteristics on the frequency of innovation with some important specificities. Particularly concerning the role of R&D.

4.3 Discussion

Our results generally show that there is a positive relationship between formality and business innovation. Indeed, it has been established that formality increases the innovation capacity of enterprises on the one hand and that informal enterprises have a low frequency of innovation on the other. These results thus reinforce the work in Africa which has shown that the productive fabric in this context is heterogeneous, marked by a formal and informal divide and therefore a better understanding of this duality can help boost business productivity and economic growth.

In the specific case of analyzes on innovation factors, our analyses essentially confirm the results obtained in Kenya by Mendi and Mudida [25] and Fu et al., [26] in Ghana. Mendi and Mudida [25] find that the fact that a business starts up informal has a low impact on its capacity for innovation. As for Fu et al. [26] in Ghana, they attest that formal companies have a better capacity for innovation, particularly about technological innovations and better productivity. Following on from them, our analyzes confirm this relationship even in the case of organizational and commercial innovations. We thus show that, if non-technological innovation in a dual productive fabric dominated by SMEs as is the case in Cameroon and Senegal, is strongly based on the manager’s qualifications and skills, on the other hand, the formalization of companies is essential because it gives access to qualified employees, to information and exchanges with other companies. In addition, our work attests to considerable sectoral differences since it is demonstrated that formality further increases the capacity for innovation in the industry, then in services, and slightly in trading companies.

Unlike the analyzes of Agwu et al. [50] whose work on African countries shows that there are no differences between countries in innovation trends within companies, we show that even if overall the innovation factors tend to be the same, on the other hand, it is important to carry out comparative studies between countries, especially when the latter belong to different economic zones. Indeed, it has turned out that R&D is a much more important factor in Senegalese companies, especially in the industrial and service sectors. In addition, Senegalese companies have a higher frequency of innovation insofar as they develop several innovations. This result shows that formalization promotes access to information through the collaborations that companies develop and allows them to favor complementarity in terms of innovations rather than substitution, which is a much more expensive process for small businesses firms. From this perspective, R&D seems to be an essential asset.

Finally, we show that the frequency of business innovation increases with formality. This result which is established in the two countries with a strong significance in Senegal shows that one way to improve the rate and the scale of the innovation of the companies being able to make it possible to reach the visions of development fixed by the two countries is the formalization of companies. Indeed, so far, innovations in business in Africa appear to be minor and adaptive. And very often the result of managers. Certainly, in recent years, policies have been adopted to facilitate the formalization of businesses, but surveys show that informal businesses remain very predominant in the productive fabric and that the two countries (Cameroon in particular) have changed very little in the doing business ranking. In Senegal, for example, despite the process of formalization and business support initiated by the International Labor Office and government authorities, we note that 9 out of ten workers are in informal employment [51].

5. Conclusion and recommendations

Business formalization and innovation remain major concerns in ASF’s economies. Indeed, the productive fabric in this space remains dominated by informal businesses, limited access to quality human resources, a low-incentive business climate, and very limited access to financing their activities. In such a context, understanding the role of formality in the innovation process appears essential in the face of the challenges of development and survival of companies but also in the context of the definition of public innovation policies. Research has shown that formal enterprises are not only the most likely to adopt and develop innovations, but also have better economic performance. Thus, the objective of this article was to analyze the relationship between formality and innovation in companies in Cameroon and Senegal. In particular, we have studied the role of formality on the capacity for innovation but also the frequency of innovation. Our results thus aim to provide a better understanding of this relationship in the context of French-speaking African countries which for the majority are oriented towards growth strategies (Plan Senegal Emergent, 2035 and Growth Strategy for Employment, 2035 for Cameroon).

Our results show that in this space, formality improves the innovation capacity of companies. Much more, it determines the frequency of innovation with greater effect in Senegal and industrial and service companies. Moreover, we have found that business characteristics including age, location, size, ICT, number of employees improve the capacity and frequency of business innovation. In addition, our analyzes have shown that it is important to continue R&D activities since the industrial and service companies that carry out their activities have a better frequency of innovation. Finally, our results confirmed the major role of manager characteristics and skills in the innovation process.

It is therefore essential to intensify initiatives aimed at the formalization of businesses if we want to promote the development of innovations in French-speaking Africa and move from simple copying of innovations and their adaptation to real innovations that can impact growth. On this issue, Cameroonian public authorities are even more concerned because not only does the rate of informality remain very high in Cameroon, but the impact of formality remains insignificant, especially when considering the frequency of innovation. It, therefore, seems essential not to be limited to accompanying measures towards formalization. Indeed, to move towards the objective of industrialization, Senegal and Cameroon must further promote incentive frameworks that would allow companies to legalize and formalize their activities. Certainly, several measures have already been developed and others are underway, but their impact remains mixed given the weight that the informal sector still represents. Studies on obstacles to the formalization of companies could lead to more appropriate policies.

Therefore, if we want to boost the rate of innovation, the process of supporting businesses towards formalization must emphasize measures aimed at supporting them financially. Such initiatives would allow companies to reduce the financial constraints linked to the formalization process. These measures are essential and would benefit more the small informal enterprises which are experiencing real cash flow problems. States can also go towards incentives for small businesses that decide to regularize their activities.

As it is considered difficult to put an end to informal activities, government authorities must therefore give greater importance to informal enterprises since their actions force formal enterprises to more innovation. The public authorities could perhaps imagine a model of collaboration between formal and informal enterprises which would allow developing more innovation and thus boost performance and growth and would cause informal enterprises to regularize. Finally, an important emphasis must be placed on employee training. Indeed, our results have shown that the number of employees determines the innovation capacity of companies. In this sense, companies must place significant emphasis on training them to develop their skills and thus take advantage of the opportunities in their environment to develop more innovations.

Acknowledgments

The authors thank Ms Beyala Marie for her comments and suggestions.