|
FINANCIAL DIARIES We
begin with a half-page Overview
which summarises the contents of this entry. This is followed by a
five-page Introduction
which explains the study, and finally by the Diary
Summaries themselves. There are 205 pages of
summaries broken down into 4 sections of approximately equal length . The Institute for Development Policy and Management (IDPM) at the University of Manchester is currently conducting research into ‘Finance for the Poor and Poorest’ as part of the ‘Finance and Development’ project supported by DFID. The project runs for the three years 1999 to 2001. The ‘Finance for the Poor and Poorest’ component is managed by Professor David Hulme in Manchester, Stuart Rutherford in Bangladesh, and Orlanda Ruthven in India. The objective of the research is to shed more light on how poor households manage their finances over a full year, and in particular, how and why they make use of financial services and devices. We believe that this is the first time such an exercise has been attempted. We anticipate therefore that the results will be of interest to a wide range of readers. These include microfinance practitioners and their supporters (such as international donors), policy makers in developing countries and in aid agencies, and academics interested in studying how poor households cope with their poverty. The Bangladesh diaries were completed at the end of October 2000. What we present below is a first view of some of the basic data, set out in the form of tabulated summaries of all 42 diaries. An introduction sets out the context, the research questions, and an outline of the methodology. Background: The research into ‘Finance for the Poor and Poorest’ comprises work in two countries (India and Bangladesh) using three research instruments, of which the ‘Financial Diaries’ are one. The other two are a ‘Snapshot’ study and an ‘Innovative Institutions’ study. The three instruments relate as follows. The ‘diaries’ and the ‘snapshot’ provide two views of how the poor, very poor and near-poor use financial services and devices. Whereas the diaries look at a small number of households (42 in each country) over an extended period of one year, the snapshots look at all the households in one village or one slum at one moment in time. Thus while the diaries illuminate how households use financial services over time, the snapshot gives us an idea of the quantum and variety of financial services and devices in use at a particular time in a particular environment. Both instruments are designed to study both the ‘how’ and the ‘why’ of financial services use. The Institutional Study is quite different. It takes a small number of innovative microfinance providers and reviews their services to see to what extent they address the financial services preferences of poor households as revealed by the diaries and snapshots, and how they do it. The research questions: For the
diaries, the key research question was: Why and how do poor households use financial services and devices over an annual cycle? In particular:
The research environment: The research was carried out in three rural and three urban neighbourhoods, for a total of six neighbourhoods from each of which seven respondent households were selected. (Note: To respect confidentiality the names of the research areas, but not of the respondent households, have been changed). The rural neighbourhoods are in a village area (which we have named ‘Grampur’) in north central Bangladesh. ‘Grampur’ is dominated by rice cultivation but with several other vegetable, fruit and tree crops. There are few non-agricultural employment opportunities other than local services such as transport, and the retailers in the markets, though there are two government-financed agricultural research stations with extensive landholdings which offer modest salaried employment to a few. There is a busy local market, and more important markets can be reached along a paved main highway. The capital, Dhaka, is about four hours away by bus. The area escaped serious flooding in the national floods of 1988 and 1998. There is a formal bank in the local market, several MFIs (both national and local), and a branch of an insurance company that offers basic low-cost life cover. Most homes in all three neighbourhoods are mud-walled and roofed with tin sheet. Better off households have more than one room, while the poor may have woven-bamboo instead of mud walls, and thinly-thatched roofs. Many but not all homesteads have a drinking water hand-pump. Most households are Muslim, with a few Hindus. Of the three neighbourhoods, one, ‘Rastapur’ lies alongside the main road and has good quality irrigable paddy land. The second, ‘Nichepur’, lies three kilometres back from the roadside to which it is linked by a paved secondary road. It too is an irrigated rice growing area and as in Rastapur its inhabitants farm, either as owner-operators, or as sharecroppers or lessees, or as land-poor or landless labourers. ‘Uporpur’ lies further back from the main road, and paved-road access arrived only very recently, during the research year. Here the land is mixed, with some paddy land and other higher land on which pineapples and trees can be grown. Trading in timber is a popular occupation for both landed and landless households. The three urban neighbourhoods are all on the western side of Dhaka. However, they are separated by several kilometres and have sharp differences of environment. The neighbourhood referred to as the ‘Beri Bad’ is a stretch of flood-protection embankment constructed after the catastrophic 1988 floods, on which many landless immigrants, above all from the very poor southern districts of Bangladesh, have squatted. During the research year the area underwent drastic change, as the government demolished all the dwellings and rebuilt the embankment in order to construct a paved highway on it. Some households received cash compensation for the disruption. Occupations and income sources vary widely. Some well-established households have built extra rooms which they let out to more recent arrivals. Some men are rickshaw or auto-rickshaw or private car or truck drivers. Men women and children day labour in brickfields (dry season only) or as brick breakers (paid by the piece). Some women work as maidservants. Other households get by with hawking clothing or other goods, recycling waste, or as construction workers, both skilled and unskilled. Some keep livestock on the embankment slopes or on the nearby land outside the embankment. A few make their way each day to garments factory jobs. There are no formal banks in the immediate area, and relatively few MFIs serve the population. There are NGO- and government-run primary schools. Water and sanitation services are extremely primitive. We call the second urban neighbourhood ‘Manushpur’. This a clearly-demarked urban slum of about 10,000 people, settled partly on government and partly on private land. It is extremely dense, consisting of very narrow lanes of about one metre width (now mainly paved by NGOs) giving access to homes one-room deep and back-to-back, made of woven bamboo walls and tin sheet roofs. Public infrastructure has developed over the twenty years or so that the slum has existed, despite several devastating fires caused by open-flame cooking. There is a scattering of water-points and public latrines, mostly provided by NGOs, and there are mosques, primary schools and simple clinics (again mainly NGO-run). Occupations are as varied as in Beri Bad, but with no brick-breaking, much less livestock rearing, and with more people taking garments factory jobs or working as domestic servants in middle-class areas close by. There are far more MFIs serving the slum, including several major national ones. Banks are available nearby though not in the sum itself. There are many small shops in the slum, and some are now converting to ‘pucca’ (masonry) construction. The third area is ‘Sonaro Mohalla’ (also the site for the urban ‘snapshot’ study where a full description of the neighbourhood can be found). Sonaro is closer to central Dhaka and is surrounded by poor and middle-class residential and market areas. It is much more established than the other two neighbourhoods: a third of the inhabitants have been there more than 30 years, and almost none less than five years. Occupancy rights are jealousy guarded. We worked in three adjacent ‘pockets’. Of these, one is a collection of about 50 households squatting on public land in conditions not dissimilar to those described for Manushpur. This area is referred to as ‘City Corporation Slum’ in the diary summaries. Another is a very small poor slum (‘Alam slum’) on private land consisting of just 12 households in bamboo-and-tin rooms constructed by the landlord and let out at 950 taka each per month (US$19). The third (‘Safiq Colony’) is a better-quality version of the second, with masonry walls and cement floors, where each room costs 1,250 taka a month ($25). For Sonaro as a whole, the most common occupations are in transport (car, rickshaw and auto-rickshaw driving) and in the construction trades (masons, carpenters, labourers). There are also many garment workers and some other low-paid salaried jobs, some shopkeepers, and some hawkers. Methodology: The research team was drawn from people with considerable experience in poverty or microfinance studies in Bangladesh. Stuart Rutherford, who managed the research, lived in Bangladesh from 1984 to 1999. His Research Assistant, S K Sinha, teamed up with two former Assistants of Dr Imran Matin (now of CGAP at the World Bank) whose PhD thesis was written on microfinance in Bangladesh, his home country, and who authored the Innovative Institutions study. These two, Saiful Islam and Mohammad Eakub, were responsible for diary collection in the rural area (where Eakub is a native and where Saiful had worked previously), while Sinha and Saiful handled all the urban interviews. The research areas were selected according to a number of criteria. Of these the first was to ensure than the selected neighbourhoods were not untypical – economically, socially and culturally – of Bangladesh generally. The second was familiarity: all six neighbourhoods are areas where one or more of the research team had some knowledge either of the local economy, or of local microfinance, or of the local households, and about which we could draw on previous studies. Access was our third criterion: fortnightly interviews demanded that the team be based locally, so we decided that the rural neighbourhoods should be grouped in one area, and that the urban ones be within 20 minutes travel time of Sinha’s Dhaka home. Next, we sought variety: in both the rural and urban context we looked for a mix of environments among the three neighbourhoods (see the notes on ‘the research environment’ above). Finally, we wanted at least one of the diary neighbourhoods to be covered by the ‘snapshot’ study: this was achieved in Sonaro Mohalla, where Dr Maniruzzaman carried out the snapshot study. The diary respondents (the households whose financial activities we diarised) were selected to provide an unbiased selection of poor, upper poor and ‘non-poor’ households. In most cases this was achieved by careful wealth ranking of all the households in the area and then by random selection of one non-poor, two upper-poor and four poor households from the ranked lists. In some areas in the urban neighbourhoods this technique was not suitable because householders didn’t know each other well enough to allow for good wealth ranking. In those cases respondents were selected by a totally random pulling of household names from a hat. (Each diary summary, below, bears a note showing the technique used). In all cases drop-outs (of whom there were very few) were replaced by pre-selected names drawn at random from the respective wealth ranks. After the close of the diaries, the Teams re-ranked the respondent households into ‘non-poor’, ‘upper poor’ and ‘poor’ categories based on written criteria that they had evolved during the course of their work. In a few cases these rankings differ from the initial ‘wealth ranks’: for this reason both systems of ranking are shown at the head of each diary summary. The frequency of the diary visits was determined largely by our resources. We would have preferred a weekly visit, to reduce reliance on respondents’ recall. However, this would have proved too costly in terms of manpower and data-processing, and would have lost the unity that came from each respondent being dealt with exclusively by a team of just two researchers. It would also have caused more disruption to the respondents, without whose cooperation we would not have been able to proceed. It should be noted that most respondents are illiterate, so we couldn’t rely on their keeping actual written diaries. In the event a fortnightly (twice-monthly) visit proved practicable and satisfactory. Gaining entry to the sites was easy in the rural areas where the Team was already known and had a good reputation. In the urban neighbourhoods a number of preliminary visits were made and conversations were held with key inhabitants. The purpose and outcome of the interviews were explained in simple, truthful terms – these interviews were being done for a University research project that sought a better understanding of how poor people use financial services, and would not lead to direct action or benefits aside from a small present to be given at the end of the process as a ‘thank-you’ for cooperation. The interviews followed a set pattern. The first full interview was designed to establish a friendly relationship. Basic demographic data was collected and notes made on current involvement in financial services. This led to a ‘Initial Questionnaire’ format being completed. From then on the twice-monthly interviews reviewed and recorded the half of the calendar month preceding the interview. As much detail as possible was taken about transactions involving financial services and devices (acts of saving, drawing down savings, borrowing, lending, repaying, paying insurance policies, and so on) and as much detail about general income and expenditure and the household’s situation as was required to make sense of the financial services transactions. Interviews were recorded in Bengali on the spot, and then copied into hard-bound notebooks (one per respondent household). Between interviews notes were made to facilitate the next session – transaction series to be followed up, contradictions to be resolved, implications to be explored, and so on. Where useful, data was checked with third parties – MFI records, managers of savings clubs, Bank Managers, neighbouring creditors or debtors, and so on. The bound books were typed up and returned to the Team for checking, then sent to a translator. The Manager (Stuart Rutherford) developed a running ‘transaction summary’ for each household and on the basis of this sent notes on the translated text back to the researchers for follow-up. After the last interview, a review was carried out for each case, and a ‘Final Interview Document’ prepared which collated all the data available and highlighted any remaining contradictions or ambiguities. This was used for a final, extended, interview in which the full two-person Team (or sometimes more) were involved. The diary summaries, below, are based on the corrected ‘Final Interview Documents’. Quality control of data was approached in several ways. Primarily we relied on high quality experienced native-language interviewers. We reinforced that with regular training and feed-back sessions: Stuart Rutherford worked with the team for a total of six weeks, and both David Hulme and Imran Matin made two visits to the team in Bangladesh. We didn’t overload our researchers: they had the time to follow up the diaries between visits. We kept the number of respondents small enough so that the researcher could hold details of each household in his memory. We used ‘triangulation’ – checks with third parties, as reported above – wherever possible. Incoming data was transcribed and tabulated regularly and speedily so that inconsistencies could be picked up quickly. Problems and constraints. Despite the precautions noted above, there were problems. A very small number of selected households proved uncooperative and were dropped and replaced (these cases are noted in the diary summaries). One urban household (code DBB06) left Dhaka altogether but we were able to follow them to their village home and complete a summary interview there. Five respondents were in some way unreliable – they told us untruths or were so ambiguous or so contradictory as to undermine our confidence (three cases), while one was so deaf and another so old and inarticulate that it was hard to make sense of their reports. In these cases the only remedy was to spend even more time with them. More generally, the sheer density and frequency of financial service relationships meant that different debtors or creditors got confused in the minds of the respondents, researchers, or both. For example, one Dhaka respondent with membership of multiple MFIs doesn’t know each MFI by name, calling them the ‘ten taka a week’ or the ‘fifty taka a month’ MFI. Since she didn’t hold her own passbooks it was often extremely difficult to disentangle them. Not unexpectedly, some respondents grew to trust us only slowly, so that cautious versions of the truth told in the early weeks were contradicted by more frank versions later: interest-free family loans became interest-bearing ones, ‘illness’ was gradually revealed as drug addiction, ‘losses in the timber trade’ were finally seen as gambling losses, and transactions that had been hidden at first were suddenly revealed months later, when memories were fading. Transaction types that the researchers did not anticipate but in the end proved common were hard to catch at first – this is particularly true of the use of ‘money guards’ for saving, for example. The structure of the diary summaries: The diary summaries presented below are structured as follows. The summaries are given in sequence, with the rural ones first. The headings for each summary comprise the code for the household and its ‘ranks’ (its wealth ranking where appropriate, and the rank assigned by the researchers), the names of the respondents, their neighbourhood, the date of the final interview and the names of the researchers who conducted it. A box follows which summaries in narrative form the economic and demographic situation of the household, their use of financial services and devices, and a thumbnail sketch of their fortunes during the year. A number of tables and notes follow. They are:
Finally comes the principal précis of the data, in the form of a Transaction Summary. This table takes each type of financial service or device used by the household in turn, sub-divided by each transaction or transaction set (for example, a loan received and repaid, or a series of compulsory savings to an MFI). In each cell a summarised ‘history’ of the transaction, with dates, values and reasons, is given. This table frequently contain information that sheds interesting light on the financial services behaviour and preferences of the household.
|
