1Finance is the most regulated part of the economy. This is particularly true in developing economy, like Haiti where there is a state institutional weakness. Both banks and microfinance organizations (MFO) and other financial actors (transfer, change, etc.) are strongly regulated and centralized. Central Bank plays the role of governing this regulation. It creates rules, communicates them to related actors, and reinforces these rules through sanctions when actors do not comply with them.
2Institutions used to govern financial system flow through a linear and extended stream. National regulators are connected to local actors in the one hand, and national actors and regulators are pressed by international regulation system in the other hand. The finance sector is then globally institutionalized, even more after the 2008 financial crisis. Actually, anti-money laundering rules are an example of international rules that complete the regulation of national financial systems.
3This institutional stream is adapted to the official financial system, which is centralized. In poor countries where financial inclusion is limited, many people are not served by official financial actors. As to adapt themselves to the global financiarized world, the excluded people often innovate by creating appropriate financial organizations. Such form of financial auto-inclusion mechanisms become current, in many countries. In Africa and Asia, self-help groups are well-documented (Patish, 2001, 2005; Gahia and Nandhi, 2007). In Haiti, several financial auto-inclusion initiatives exist, but they remain under-studied.
4Haitian credit unions (locally called Caisses Populaires or Caisses d’Épargne et de Crédit, CEC) has been created since 1946 as a community's response to the financial exclusion. During the last thirty years, microfinance has expanded largely in the country (Paul, Daméus and Fleuristin, 2012) but because of its commercialization and institutional mimicry on banking rules, it ends in a mission drift. Poor Haitians in rural and suburban areas remain excluded from traditional finance which includes banks and MFO regulated by the central bank.
5Haitian microfinance failed to include the poor (Lenève and Cadet, 2013), and to provide them appropriate services package (Paul, 2014), despite Central Bank designed financial inclusion strategy (BRH, 2015). In this financial exclusion context, within the financiarized poor Haitian economy, excluded people, with local NGO assistance, created their own local financial organizations called Association de Base de Cotisations et de Prêts (local savings and credit associations, ABCP for their short name in French). As they grow and constitute together an alternative to the traditional finance sector (with 10 316 associations totalizing 244 005 members, according to updated data by Paul, 2019), they represent an appropriate field to study how institutions is linked to access to credit. These relatively new organizations created their own rules or institutions to regulate both their operations and their saving and credit activities, with difference with traditional finance sector.
6In this paper, we surveyed a sample of ABCP and ABCP members to identify what are the core institutional settings (also called institutional capital) that make these organizations sustainable and what is the correlation between this institutional capital and members’ access to credit. In this new institutional economic perspective, we develop an econometric model test the following assumption.
7In the second following section, we review the literature on alternative finance. The third section presents the research methodology. The last three sections present the results, discussion and conclusion respectively.
8Most of the literature on the finance sector focuses on banks and microfinance. The traditional actors on those two so-called formal branches of financial intermediation dominate the scientific literature. Alternative or decentralized finance existing in rural and suburban areas are under-documented. Existing researches that focuses on alternative financial activities are often considered as related to microfinance literature. In fact, microfinance organizations often use them as methodology to attend the poor. In countries like Haiti, through an institutional perspective, community actors such self-help groups and local credit and saving associations constitute a growing alternative finance sector as they operate by their own institutional arrangements, with no official regulation of financial national authorities.
9A self-help group (SHG), both by definition and in practice, is a group of individual members who by free association, come together for a common collective purpose (Patish, 2005). It compounds individual members who have certain social binding factors, because they know to each other, come from the same village, community and even neighbourhood (Patish, ibid.).
10Self-help groups (SHGs) pre-existed to microfinance. Patish (2001) wrote, “The purpose around which an SHG is initially formed varies depending on the kind of program being implemented and the need for collective working”. This purpose can vary from managing a collective resource to promoting a social cause. In the context of microfinance, the author who works in India observed that “SHGs are formed (and sometimes old SHGs established with another purpose are converted) to foster savings and credit” (Patish, ibid.).
11In reality, SHG is one of the two microfinance approaches that have been experimented, according Gahia and Nandhi (2007), beside the “Grameen type” or microfinance institution model. This socioeconomic reality has been long before microfinance industry developed to instrumentalize credit to the poor. As a socioeconomic organization, SHG mobilizes social ties, creates appropriate rules, to develop financial inclusion in social context where institutions are unwritten, and considered as informal.
12SHG is designed to combine the safety and reliability of formal finance with the convenience and flexibility of informal finance, stated Gahia and Nandhi (2007). Empirical assessment on SHG-Bank Linkage conducted in India, based on data collected from bank branches in Andhra Pradesh and Karnataka, has shown financial viability, since SHG were found profitable more than banks, despite low interest rates (Seibel and Dave (2002). Such finding can be considered a support for the assumption that SHG, when existing properly as organizations, deserve interest for (socio)economic researchers. Early, Gahia and Nandhi (2008) stated about SHG, also known as “SHG Bank Linkage”, “it is expected to become a dominant form of financial access for the rural poor”. This may be actually the case in rural Haiti, where microfinance expansion (Paul, Daméus and Fleuristin, 2012) resulted in a mission drift through commercialization and using of traditional banking philosophy. This was recently observed in China, by Jia et al. (2016).
13Local credit and saving associations are more structured community-based entities whose purpose is clearly financial. Some of them do not accumulate overtime; they just rotate savings and credits. Rotating Savings and Credit Associations (ROSCAs) are common in many developing countries. In Haiti, they are also called “sòl” or “sang”. In Africa, the most common name is “tontine”.
14In contrary to ROSCAs which receives large attention among researchers, there are non-ROSCAs types of community financial associations. Frits JA Bouman, very early proposed to pay attention to these non-ROSCAs. He proposed to call these “ugly duckling” that defy the traditional financial organizations, Accumulating Savings and Credit Associations (ASCAs). ASCAs properly accumulate savings to fund credit to members. They appear to more complete as financial inclusion organizations, as they contribute to financial education and discipline through accumulation overtime. They also obtain loans from banks or other financial organizations and on-lend them to their members. Ajai Nair (2005) reported that these kinds of structured groups are efficient in delinquency, since they registered less than 5% by 2003.
15There is a hybrid form of local alternative finance organizations. The later combine both elements from ROSCAs and ASCAs with an emergency loan fund. In the case of Haiti, we find three types of self-help groups, as ABCP categories: Mutuelle Solidarité (MUSO), Banques Communautaires (BC), and Associations Villageoises d’Épargne et de Crédit (AVEC). There is also village banking (BV) as Finca’s microlending methodology (Hassan, 2002), but we do not analyze them in this research, since they are not community-based autonomous initiatives. All ABCP types have three basket funds identified with colors: 1) a green box for savings and credit activities, 2) a red box for solidarity gifts, in case of emergency or social expenses, 3) a blue transitional box in which funds from external organizations transit and accounted.
Figure 1. ABCP operating model
Source: KOFIP, 2002
16In this study, we used 3 steps data collection to capture information on ABCP in 3 different geographic departments among the 10 of the country. Through this method, we try to go beyond previous research (Lenève and Cadet, 2013) in order to test, through a representative sample, how effective are institutions designed in those community-based microfinance initiatives in creating financial inclusion for their members. We first identified the most common institutions or rules that govern interactions in those associations, through key informants interviews and focus groups during usual ABCP meetings. As a third step, we use semi-directed questionnaires to get information from ABCP members and the effects the institutions on their socioeconomic situation and perception.
17Research on determinants of access to credit often considers demographic and other variables such as age, gender, family size, education and experience including family background in entrepreneurship, and social capital proxy (Julien and Cadet, 2013). We also asked questions related to those variables.
18The model used to analyze the correlation between institutions and access to credit is based first on two levels (ABCP focus groups before one-in-one interview with members) identification of core institutional arrangements that regulate the financial intermediation created inside these local savings and credit associations. In a second time, the model analyzes the members’ perception of these institutions in relation with socioeconomic variables.
19According to double check analysis from key informants and focus groups with samples of ABCP, the institutional arrangements that actually regulate ABCP financial intermediation include the following categories (Table 1). This methodology to identify the institutional capital existing in microfinancial organizations has been used for the first time by Paul, Garrabé and Daméus (2011).
Table 1. Institutions that govern ABCP internal interactions
Category
|
Institutions (obligations or prohibitions)
|
Personal characteristics
|
To be adult (at least 18 years old)
|
To be legally allowed to enjoy civil and political rights
|
To make proof of discretion about internal ABCP information
|
Obligations in relation with the association operations
|
To pay membership fees
|
To contribute regularly to the saving and credit fund
|
To contribute regularly to the solidarity fund
|
To participate regularly and punctually in the association meetings
|
Sanctions
|
To pay a penalty in case of lateness in association meetings (5 to 15 Haitian gourdes)
|
To pay a penalty in case non-motivated absence in association meetings (10 to 25 Haitian gourdes)
|
Exclusion from the association in case of indiscretion, legal sanction, etc.
|
Specific to credit activities
|
To reimburse in time the credit with indicated interest rate
|
To pay a 1-3% monthly interest rate on the credit amount
|
To accept sanction (no credit for a specific period) in case of delay in reimbursement schedule
|
Source: The Author, data from our interviews.
20We consider a participant i (where i=1, 2…I) earns a utility for each of its j choices (where j=1, 2…J). The Utility function includes a deterministic part of the Vij observable characteristics and an error term (εij). For the participants i this utility can be written as follows:
21The error term is supposed to be distributed identically and independently by a law of extreme value of type 1. The rule of decision, for each participant i, is to select the j alternative which maximizes its utility function.
22In this study, the appropriate model to be estimated is a binary model. The latest is interested in the choice or not to have access to credit in the ABCP (evaluated in the comparison between the number of demands and the number of credit is offered, with no consideration regarding the amount of credit). To estimate the binary model, we create an indicator variable Yij which takes the value 1, if the member had at least one time access to credit during his experience as ABCP members, and 0, if he did not. In this case, the endogenous variable of the model is a dichotomous variable. The linear multiple regression standard model can be written as:
23Estimate this binary model implies to be certain that the predictions will fall into the interval (0, 1). And, as the number of observations (97) is sufficiently high, it is possible to argue that the data are distributed according to a normal law; which allows to opt for a Probit model. The form of the equation to estimate is then:
24In this relation, F is a cumulative density function given by:
The parameters m and β of the model are estimated using methods
numerical maximization of the logarithm of the likelihood function
which is as follows:
25As suggested in the literature review, the vector of explanatory variables Includes characteristics related to client/member’s profile, institutions (Paul, Garrabé and Daméus, 2011) and variable-related members socioeconomic assets (Mayoukou and Kertous, 2015). Member's profile include: age, sex, education, marital status, family charge. Socioeconomic variables include type of economic activity, activity cashflow or their evolution, savings. We also add ‘previous connection to the traditional finance sector’ as a determinant because, in Haiti, this sector runs as a closed circle with barriers to entry, even for new clients (Jeanty, 2019). The institutions include rules that regulate the financial intermediation or their perception or acceptance as the member is exposed to them in a significant period of time (then the duration is important in the effect of the institution).
26The following table 2 presents ABCP members’ profile and variable statistics
Table 2. Variables description and statistics
Variable description
|
value
|
Expected sign
|
Gender
|
Gender of ABCP member, 1 if women, 0 if not
|
+
|
Marital status
|
Member’s marital status, 1 if married or in partnership, 0 if single
|
+
|
Education
|
Member's level of education, 1 if attended a school, 0 if analphabet
|
+
|
Log Charge
|
Log of total relative in charge of the ABCP member
|
-
|
Log Duration
|
Log of total years of experience as ABCP members
|
+/-
|
Log Savings
|
Log of total savings accumulates in the ABCP
|
+
|
Economic activity
|
Member’s type of economic activity, 1 if commerce, 0 if other (production, service)
|
+
|
Evolution of CA
|
Member’s progression of total commercial cash-flow, 1 if progression, 0 if not
|
+
|
Traditional Finance
|
Experience with traditional finance including banks and MFOs, 1 if yes, 0 if not
|
+
|
Other income
|
Existence of other source of income, 1 if yes, 0 if not
|
+
|
Institutions
|
Members’ perception of internal rules effect on its activities, 1 if positive, 0 if not
|
+
|
Source: The author
27As young people are excluded from ABCP credit activities, age were not revealed as a significant variable in model estimates.
28The data used in this paper were collected in the framework of KNFP (Konsèy Nasyonal Finansman Popilè, a promoter of a network of ABCP around the country), with funds from the Inclusive Finance project (USAID/DAI). The data collection was made in March 2019, in 3 of the 10 geographic departments of Haiti. The selected area sample was balanced to include the 3 types of ABCP (MUSO, BC, AVEC). Focus groups were conducted in 17 ABCP, and questionnaires were distributed to 97 members.
Figure 2. Repartitions of data sample
Source: The Author, from collected data.
29Collected data were processed with Gretl statistics (version 2018c). We used backward specification to arrive to the final estimated model.
30The results of econometric estimates describe significance for almost all determinants. However, some expected signs are different. A variable such as level of education was not retained in the model. In fact, the great majority of ABCP members (73%) have primary or secondary school level; with 14% with no school frequentation. The education level does not determine access to credit inside the associations, but membership and savings.
Table 3. Model estimates
Dependent variable : Access to credit
|
|
Coefficient
|
Erreur Std
|
z
|
p. critique
|
|
constant
|
6,88221
|
3,02841
|
2,273
|
0,0231
|
**
|
Gender
|
-2,02760
|
0,608794
|
-3,331
|
0,0009
|
***
|
Marital Status
|
1,76384
|
0,922987
|
1,911
|
0,0560
|
*
|
l_Charge
|
-2,43712
|
0,845070
|
-2,884
|
0,0039
|
***
|
l_Duration
|
2,00788
|
0,595974
|
3,369
|
0,0008
|
***
|
l_Savings
|
-0,509438
|
0,278812
|
-1,827
|
0,0677
|
*
|
Economic activity
|
1,86340
|
0,648736
|
2,872
|
0,0041
|
***
|
EvolCA
|
-1,76558
|
0,771341
|
-2,289
|
0,0221
|
**
|
Traditional Finance
|
0,948388
|
0,447406
|
2,120
|
0,0340
|
**
|
Other income source
|
-1,72733
|
0,773220
|
-2,234
|
0,0255
|
**
|
Institutions
|
2,19682
|
0,873931
|
2,514
|
0,0119
|
**
|
Dependent variable average
|
0,844156
|
|
Dependent variable st.D
|
0,365086
|
McFadden R square
|
0,441228
|
|
Adjusted R square
|
0,111086
|
Log de likelihood
|
-18,61769
|
|
Akaike criterion
|
59,23538
|
Schwarz criterion
|
85,01724
|
|
Hannan-Quinn
|
69,54790
|
* Significant at p < .01.
** Significant at p < .05.
*** Significant at p < .1.
31According to these empirical results, all the considered variables are significantly correlated with the access to credit dependent variable. However, some signs were unexpected. As found in previous similar research (Mayoukou and Bertous, 2015), our estimates show that ABCP members’ level of education has no significant relation with their access to credit. Members’ age (and log of age) is found as non-significant variable in determinant access to credit in ABCP. In fact, all the active members in ABCP are adults. Children who register in ABCP are participating through their parents.
32Unexpectedly, being a female ABCP member does no guaranty higher access to credit inside the associations. This result can be explained by the fact that more and more ABCP have both male and female members. And, women represent 67% of total ABCP, but credit is not distributed with gender discrimination inside the associations. Members marital status and family charge consistent with previous findings on credit access study (Mayoukou and Bertous, 2015).
33Institutions are significantly correlated to members’ access to credit with expected signs. This result confirms our hypothesis and brings an affirmative answer to our initial question. Institutional capital constituted by institutional arrangements that govern ABCP microfinancial intermediation is significantly correlated to members’ access to credit. And, the more the members participate in ABCP and submit themselves to those institutions, the more they have access, as the variable “duration” is also significantly and positively correlated to access to credit.
34Being economically active, preferentially in commerce and having accumulated certain amount of savings in the ABCP are both correlated to access to credit. But, commercial activity reinforces this access, while the members who have accumulated savings tend to have less access to credit. This may be explained by their fewer propensities to ask credit. This seems to be the same when their activity is growing financially (variable evolutionCA) and if they have other sources of income.
35ABCP members who had (or continue to have) relation with traditional finance (MFOs or banks) have more access to credit in the associations. This can be explained by the fewer cost of credit in ABCP (2%) than in microfinance (3%). And, in contrary to microfinance and bank, credit revenues in ABCP return to members at the end of the credit cycle (4 years in average).
36Previous research on Haitian microfinance (Julien and Cadet, 2013) revealed that the poorest of the poor do have access to microcredit. This is particularly through, even in the metropolitan area where there is no particular geographical problem of accessibility to reach the poorest people.
37Institutions regulating alternative finance we studied are what Santos and Barrett (2011) called “informal formal credit arrangements”. Although most of research on access to credit do not focus on institutions that govern (micro)financial intermediation, our research fill a gap on the existing literature by proving a more complete understanding. Institutions are particularly determinant in having access to credit. This is consistent to existing literature on the relation between financial capital and institutional capital (Paul, 2011).
38The set of institutions, in order words internal rules representing an institutional capital accumulation, is particularly effective both in ABCP management and in creating financial inclusion. As a result, ABCP operations and transaction costs are close to zero (based on volunteering) when in MFO and banks these costs (may represent more than 50% of total costs (for example, MFO operating income absorbed by the operating costs ranged from 60 to 85% for 2017). The ratio of the amount of credit demands on total number of credit granted is 96.74% (as of 2018), when this rate is usually lower in traditional financial organizations in the country. In addition, ABCP credit policy is more gender sensitive (percentage of microcredit to women: 71.7% as of 2017) than MFO (percentage of microcredit to women: 68.8% as of 2017), according to the latest census of the microfinance industry in Haiti (USAID/DAI, 2018).
39Mayoukou and Bertous (2015) found that member duration of being MFO client had no incidence on demanding or receiving credit. In the contrarily, being ABCP member where there is both social and financial solidarity (through special red box for collective savings to assist members in emergency) for a longer time increase access to credit. In fact, with the time passing, new members have more time to show they can act in conformity with designed institutions in the ABCP. This support the idea that institutions create mutual “confiance” and sense of belonging, as stated by Nobel Laureate Douglass C. North (1991). As also supported by Nobel Laureate Elinor Ostrom, institutions govern and determine repetitive interactions between economic agents, in ABCP, members who do not comply with particular institutions (such as: discretion, savings, etc.) can be excluded to the associations.
40The purpose of this article was to identify institutional capital in the ABCP alternative finance intermediation and analyze the correlation between institutions and access to credit. With empirical data collected in rural and suburban microfinancial associations in Haiti, we show that these community-based associations know how to design their own institutions to govern their financial intermediation. And those institutions have strong positive and significant effect on access of their members to credit. As the traditional financial sector is highly regulated, this finding support the idea that community-based institutional capital is can be more effective for financial inclusion than official rules.
41As already shown by Elinor Ostrom in governing the commons, institutions created in these community-based financial associations are particularly effective in governing saving and credit activities, with more flexibility, solidarity and rewards for members than in other form of organizations. This result partly questions Mats Lundahl’s argumentation about the failure of community-based entrepreneurship in Haiti (Lundahl, 2010), because through institutional arrangements, ABCP not only support social capital accumulation but they create places for financial accumulation and business training, even though they cannot show particular results in terms of entrepreneurship evidence, apart from microbusiness activities among poor people.
42The more time ABCP members are exposed to institutions, the more they learn to comply with the designed rules and benefit from other members’ solidarity and mutual trust, and overall get access to credit at a low cost. This result creates new research avenues to study institutions at a microeconomic level, and with a larger sample.