«bath papers in interna onal development and well-being ISSN 2040-3151 Compe ng visions of ﬁnancial inclusion in Kenya: The ri revealed by mobile ...»
3 Methodology This research set out to examine changing use of financial services among low income people in Kenya. Research focused at mixed distances from three towns chosen to cross-cut Kenya’s district poverty rankings (according to GOK, KNBS 2006). These were Mathira and Nyamira in the agro-ecologically higher potential zones in particular where cash crops of tea and coffee are grown; and Kitui which is semi-arid and experiences crop failure and food insecurity on a frequent basis.
Supply-side research involved interviews with managers or their representatives of 59 service providers covering banks, savings and credit cooperatives (SACCOs), microfinance institutions (MFIs) and other NGOs or companies offering or facilitating group-based approaches. On the demand side the methods were both quantitative and qualitative. The survey component offered two functions: first to provide a randomly selected sample frame from which to purposefully select respondents for in-depth qualitative interviews and, second, to take advantage of that contact to conduct a survey which could capture broad patterns of usage so that the picture relative to the national overview could be established. The sample of 337 was 7|P a g e Bath Papers in International Development and Well-Being Paper Number 30
from 194 households, of whom 56% were female respondents, and main income sources were:
own agriculture, livestock and fishing (35%); employment in agriculture, casual labour or domestic chores (21%); own business (20%); public or private sector employment (11%) and pensions or transfers (11%). 56% fell below the $2.50 per day poverty line with 20% below $1.25 per day.4 The qualitative interviews focused on understanding how people used the services they did and why. The methodology is inductive and interpretive. It first uses quantitative data to highlight patterns which are explored and interpreted through the qualitative data to understand the reasons for and meaning of these interactions for users. The evidence is presented here in an integrated way. I first offer an overview of the financial market in terms of access in order to identify its key features. Second, I review the evidence on the development of the types of service in turn. I then turn to a discussion of what this evidence offers for interpreting patterns of financial practices.
4 Financial services and financial practices
4.1 Overview of the financial landscape The rates of service use across the formal and informal sectors are given in Table 1 (final column). This also gives comparable figures obtained from the nationally representative surveys for 2009 and 2013 (FSD Kenya and Central Bank of Kenya 2009; 2013).
First, the data confirms the high penetration of MMT, doubling the level found nationally in the 2009 FinAccess survey establishing it now as the most used financial service.
Second, informal financial groups (ROSCAs and ASCAs5) are heavily used. Membership of either of these is 51% in the landscapes survey, 36% for FinAccess 2009 and 31% in 2013. Rotating savings and credit associations (ROSCAs) are well-known (Anderson and Baland 2002; Johnson 2004; Kimuyu 1999; Nelson 1995) and known by a range of names in local languages and as “merry go rounds” in English. Accumulating savings and credit associations (ASCAs) are strong in this dataset reflecting donor programmes which have promoted them in all of the research areas. Moreover, 42% of the financial groups reported having multiple functions of both a ROSCA and ASCA or welfare component (30%) where support is given in cases of death, serious The Poverty Assessment Tool was used to estimate expenditure level and compare it to the poverty line, see www.poverytools.org.
The most basic form of a ROSCAs involves people contributing to a fund which is usually (but not necessarily) given to one person on each occasion, until everyone in the group has received the money in turn. The order of rotation may be determined by ballot, by age or seniority or other social systems of preferment. More sophisticated versions involve the payout being auctioned with the person willing to take the largest discount receiving the payout - the remainder being divided amongst those who have not yet won the payout. The ASCA develops this basic form by allowing funds to accumulate in a fund which is not paid out or bid for at each meeting, rather those who wish to take loans do so with interest, the fund therefore grows (as long as funds are repaid) and savers can receive dividends based on their savings.
8|P a g e Competing visions of financial inclusion in Kenya: the rift revealed by mobile money transfer Johnson illness or hospitalisation, assisting not only with money contributions but also with organising and carrying out funeral arrangements.
Table 1: Financial service access (% using)
The third most used service in this survey is banks (38.5%) although with a rather higher rate of use than in the FinAccess 2009 and 2013 surveys - reflecting the location of the study in higher potential areas. Beyond these top three most used services the levels of use between the Landscapes survey and the national data differ more substantially and may be explained by the focus in two relatively high potential areas. The falling figure for family and friends is intriguing given the rise of MMT, but we noted that there was frequent discussion of using borrowing from family and friends in the qualitative interviews even if this borrowing was not currently outstanding and hence perhaps not captured in survey instruments whereas ‘loan’ is understood as having a greater degree of formality.
Probit regression analysis (see Annex 2 of Johnson, et al. 2012) indicates that gender (being male) and higher levels of education are the most significant predictors of bank and MMT access, with income mildly associated with bank access and more significant for MMT access.
For financial groups, gender (being female), income and region are significant. Income has an inverse ‘U’ relationship and use declines with higher incomes.6 This suggests that service use Similar analysis on the FinAccess datasets 2006 and 2009 (Johnson and Arnold 2012) is consistent in the importance of education, gender and income related variables and also finds age, type of employment and assets to be positively correlated with bank access while being female and rural are negative. For MMT age and gender were not significant but rurality was, suggesting that the rural focus of this study may influence the significance of these variables. Informal groups were more likely to be used by women and those with mobile phones and in business.
9|P a g e Bath Papers in International Development and Well-Being Paper Number 30 maybe somewhat segmented with groups being used by poorer people and bank services and mobile money being used by the better off. However, high levels of multiple use are also evident with 83% of bank users also using MMT, and 54% using groups and almost a half using all three services. Moreover, descriptive statistics on the profiles of service users indicate that the proportion of bank users who also use MMT rises as poverty falls, and follows the same inverse ‘U’ for those also using both banks and groups. But most interestingly, it rises with use of all three services: from 34% of those on less than $1.25 per day, to 45% of those between $1.25 per day and 50% of those on more than $2.50 per day. 7 Table 2: Multiple use of main services
Source: Own survey The national data set for 2013 gives similar ordering of results. 90% of bank users also use MMT;
lower but still significant proportions of bank users are also in a financial group - 41%, and both groups and MMT – 34%. FinAccess 2013 offers the opportunity to examine the relationship with financial capability indicators and attitudinal variables such as: being worried about having enough money in old age; going without basic things in order to save; not feeling in control of finances; making spending mistakes; having a budget and sticking to a budget. Probit regressions8 revealed that having a budget was significantly correlated with the use of all three types of service once socio-economic and demographic variables were controlled for.
Interestingly the use of financial groups was correlated with being worried about having enough money in old age which might suggest a seeking out of these as a means to enable improved fund management or as a means of creating social connections for support. However, having a budget remained significant across all combinations of multiple use, while this worry about old age did not. This suggests that it is not an aptitude for greater financial discipline that leads people to particular services or combinations of services, and hence that it is not obviously different types of people who are using them.
Descriptive data on characteristics of users of each service and multiple use is available in the Annex 1.
Available on request.
10 | P a g e Competing visions of financial inclusion in Kenya: the rift revealed by mobile money transfer Johnson Overall, given this shape of the financial landscape, the question is how the use of MMT can be understood relative to the use of banks and other informal financial groups and what this reflects in terms of financial practices involved. While MMT has overtaken both banks and groups, the latter appear to retain an important role. The next sections discuss these services in turn starting with MMT, followed by financial groups and finally banks.
4.2 Mobile money transfer 72% of the total sample reported having ever used the MMT service (i.e. more than the proportion registered) and many cited the lower cost, convenience, instantaneous nature, fee payment on withdrawal and extensive agent network, as advantages. 47% had sent funds to family and friends in the last 12 months using MMT and 58% had received funds.
Data on the last transactions made suggests a strong pattern of receipts from family, household and ‘other’ relatives (67%) with almost half of these from ‘other’ relatives. The pattern of sending was more strongly towards family and ‘other relatives’ (53%) than household members (18%). Proportions transacted with friends and for business are similar across both sending (15%) and receiving (14%). These data do support the view that there is a strong pattern of receipt of funds in the rural areas from spouses or children who are “sending money home” but it also suggests strong patterns of transactions with ‘other relatives’ and an important though smaller role for friends.9 The range of reasons for sending or receiving funds given by respondents is as diverse as any need to transfer funds. The reasons certainly embraced remittances from husbands working away from home and from children working in the town or city. But they also included sending to children who had gone away for education or were looking for jobs in Nairobi; money for investments in businesses; or for group contributions either one’s own contribution or assistance received from a relative – such as a niece – to make a contribution. They extended to transfers sent or received in relation to particular events such as a pre-wedding; wedding;
funeral; Christmas; birth of a child. Further reasons were assistance to others in paying for medical expenses; school fees; or for payments related to work for picking tea, casual or regular labour contracts; rental payments as well as business transactions of many kinds.
The qualitative evidence suggested that the role of close relatives such as siblings, aunts, uncles, nephews, nieces was strong. Obviously the boundary between these as family and ‘other’ relatives is a porous one depending on how families and households are formed and operating.
Transfers received in these contexts could be interpreted as instances of redistributive “sending money home”, and it is clear that they involve cases where repayment is not expected. As one young man who received funds from his brother for his brick making business explained, “I can’t refund him the money as he has work in Mombasa” (703/1). The themes in the reporting of gifts
from siblings in particular was that they were assistance for a child for clothes or in case of sickness, or because the recipient was “low on money”.
While the majority of qualitatively reported MMT transactions involved gifts rather than loans, it is certainly clear that MMT was also expediting inter-personal borrowing. Respondents reported cases of borrowing and lending with their relatives (daughter, sibling, and sister-in-law) and friends, while others talked more generally about the potential for borrowing, lending or paying debts using MMT. While only 5% reported having an outstanding loan from someone else at the time of the survey the prevalence of borrowing directly in cash as well as via MMT from friends was much clearer from the qualitative data. Approximately one-third of in-depth interviewees spontaneously talked about how they had borrowed from friends or family. The time frame for returning such borrowing is flexible and no interest is usually paid with repayment likely to be dependent on “paying the people who were hurrying [me] to get the money” (810/2).