A village-level study conducted in Udaipur district of Rajasthan attempted to map the financial status of the rural population and the funds flow indicated that the overall asset-savings-income profile of the rural poor was not alarming. However, most of the assets and savings are liquid, forcing the poor to borrow at high cost. The study reveals the failure of financial institutions to penetrate the savings and loan market. It also reconfirms earlier findings that health-related expenses are one of the major causes of indebtedness amongst the poor.

M S SRIRAM, SMITA PARHI

C

To fill the gaps between inflows and outflows, the poor need institutional mechanisms that help them manage the flows. In most villages the moneylender performs this gap filling function. There has always been a debate on the role of the moneylender in the local economy. There is no doubt that the moneylender provides access to credit, and there are arguments that the image of the moneylender is unnecessarily tarnished in the literature [Sharma and Chamala, 2003]. There are counter arguments on whether this fits with the development intervention to be undertaken [Chavan 2003]. There are arguments that because of traditional relationships of trust, it is almost impossible to replace the moneylender, but possible to redefine the relationship by providing an atmosphere for formal competition [Sriram 2002]. It is important to understand the roles of each of the players providing finance for the poor and how the poor manage money.

The most commonly used measure of poverty is based on income or consumption levels. People are considered poor if their consumption or income level falls below a minimum needed to meet the basic needs. This level is defined as the “poverty line”. This definition varies across time and place. Each country uses a definition appropriate to its level of development, societal norms and values.

Information on consumption and income is obtained through surveys, during which households are asked questions on their spending habits and sources of income. Such surveys are conducted more or less regularly in most countries. In India, the Planning Commission estimates the proportion and number of poor separately for rural and urban sectors at the national and state levels based on the recommendations of a task force. The task force had defined the poverty line as the cost of an all-India average consumption basket at which the calorie norms are met [GoI 2002]. The norms were 2,400 calories per capita per day for rural areas and 2,100 calories for urban areas. These calorie norms were expressed in monetary terms as Rs 49.09 and Rs 56.64 per capita per month for rural and urban areas, respectively at 1973-74 prices. Subsequently a study group and expert committee estimated the proportion and number of poor and defined the poverty line as Rs 131.80 and Rs 152.10 per capita per month for rural and urban areas, respectively at 1987-88 prices. These figures were updated again with the consumer price indices (CPI) in 1994-95. The updated numbers are Rs 228 and Rs 305 per capita per month, for rural and urban areas, respectively [Pradhan and Subramanian 2001; GoI 1993].

But the debate on definition of “poor” hardly ends here. There are arguments about alternative measures of poverty. Internationally a popular measure used to define poverty is to use a dollar a day income at purchasing power parity as an upper cut-off. Even within the category of poor, there are further slices

– extremely poor, vulnerable poor and economically active poor. For the purposes of this study, we do not debate the definition. We accept the poverty line, as defined by the state. We do not consider the definition as critical because financial services needs are on a continuum and errors on the margins do not alter our findings or recommendations.

This paper attempts to understand and map the financial flows of the poor and how do they manage a little money available to them. The paper is organised into five sections. Section II looks at the literature. Section III has the geographical setting, methodology, sample size, design and administration of the questionnaire. Section IV contains findings of the study. We conclude with Section V – discussing the issues that need to be addressed at a larger scale and also how this study can be taken forward, while identifying the limitations of the current study.

Financial services play an important role in assisting the poor in managing their money and in improving their economic status. The state has intervened in this segment to address the issues of inequity from time to time. It has not only created institutional mechanisms, but also has had targeted schemes that help the poor to come out of poverty. However, most of the efforts have been supply-driven and have looked at the credit and not the savings needs of the poor. The microfinance institutions (MFIs) have addressed the issue of financial flows and have succeeded to some extent. But still reliable financial services are not widely available. The offering of credit by MFIs is pigeonholed into the ‘grameen’ type with little flexibility and the self-help group (SHG) type with more flexibility. The loan products available in the formal sector do not address the needs of the poor. Therefore, there is still a gap in the needs of the poor and the offerings [Fisher and Sriram 2002]. While the literature on credit services is vast, here we focus on literature pertaining to savings and management of financial flows of the poor.

It is believed that poor know how to manage their flows [Rutherford 2000]. They need money in lumps and finding ways to meet such requirements is a challenge. Savings is nothing but the choice of not consuming cash. This is a fundamental and unavoidable first step in money management. We should look at issues pertaining to savings and credit together, to understand the needs of the poor [Rutherford 2002].

The poor save for a variety of reasons. Risk management is one of them. The risks are temporary risks, permanent risks like life cycle risks, structural risks and crisis risks. These risks disrupt the functioning of the household’s economic portfolio to such an extent that the income required to fulfil basic needs can no longer be maintained. But there are only few products offered by the institutions and hence this needs to be addressed. People look more for safety and liquidity and less for returns in case of savings. Crisis risks require products that are liquid and easily available [Mutesasira 1999]. The savings at home, loans from moneylenders and emergency loans from SHGs are most frequently used in overcoming crisis risks. Life cycle risks need large lump sum money within a short notice. Regular savings help in building large sums, are often useful in dealing with such risks. The rotating savings and credit associations (Rosca) and the accumulating savings and credit association (ASCA) promote regular savings that become available at a time of need. The structural risks are long-term in nature and are difficult to manage. Moneylenders provide loans during a job loss or other long-term losses. But, they look for assets to be pawned and ultimately the regular consumption needs have to be cut down, children have to be withdrawn from the schools or they have to migrate to other areas in search of employment. Therefore, credit and savings are equally important and one cannot survive without the presence of the other.

There are some recent studies focusing on financial flows of the poor. The MicroSave-Africa has done a series of studies to provide financial toolkits for bankers and others. These studies recognise the growing interest in introducing savings products in MFIs. The MicroSave and the consultative group to assist the poor (CGAP) collaborated to study the dynamics of institutional change in transformation of a microcredit institution to a MFI [Wright, Christen and Martin 2000]. They studied Association for Social Advancement (ASA), which is an important model for microcredit institutions planning to introduce savings products. The ASA was a microcredit institution working only on credit delivery and recovery system based on grameen methodology. Loan sizes and disbursements were standardised, operating systems were simple and only compulsory savings were collected. Though the facility of compulsory savings was available earlier, the ASA decided to provide access to more savings products. For them, this provided relatively cheap capital, increased outreach, increased lending while maintaining the portfolio quality and reducing poverty and vulnerability. The ASA’s primary motivation to develop a savings programme was to provide its clients improved financial services the contact and stakes would improve the quality of its own portfolio and reduce dropouts. But the ASA realised that there was a gap between the saving products offered by the institutions and the need of the people. They found that sometimes, flexible savings/voluntary savings do not necessarily generate inexpensive sources of capital.

higher growth rates, poverty reduction was of the same order as Rajasthan. In addition to expenses on healthcare and social expenses, debt bondage also played a role in people slipping to poverty. These studies conclude that there were multiple reasons for households to slip into poverty. The authors argued that falling into poverty is not just the converse of escaping from poverty but more than that.

It is evident that there is considerable interest amongst scholars in examining the financial flows of the poor. Our study is different from what we have reviewed. It focuses on regions recognised as backward. The objective of our study is twofold. First, we understand the financial flows of the poor with the help of an empirical analysis. Second, we understand the finer nuances of savings habits and credit behaviours. The paper presents the results of a pilot study conducted in Alsigarh village in Rajasthan. This study is a precursor to an all-India study.

The consideration for us was to choose between a case study method and a survey method to carry out this research. Since the larger objective of the study was to roll out an investigation across the country, we found that a survey would give a common template to make regional comparisons and draw generalisations. A questionnaire was designed to capture data on various parameters. The design ensured that we use significant events in the last decade as time markers to gather financial data on how these events were managed. We also had asset purchase and sale as additional markers. These helped us in associating the financial flows – savings, borrowings (both formal and informal) with the ups and downs of a family, and in triangulating the indebtedness data. Sample selection: choice of the area and village: This study has its focus on families defined as poor. All families under the “below poverty line (BPL)” category fell into our focus population. It is not our intention to debate the methodology adopted by the state in defining the “poor”. For this study we take the definition as given. As the idea of the study is to look at how poor cope with their financial flows, we believe that even if there is an error in poverty classification, it would do no harm to the overall design. This is based on the presumption that the findings would be used for developing financial products that would be offered to a continuum of clients from the very poor upwards. The artificial boundary of a poverty line is only helpful in drawing the sample.

While we wanted to base the study in some of the most backward districts in India, the choice of Udaipur was made purposively. The selection of Udaipur was driven not only by its general backwardness, but also the presence of Seva Mandir, a big non-governmental organisation (NGO) working in that area. The exact location – village Alsigarh, block Girwa was selected in consultation with Seva Mandir. We then worked with the list of BPL families supplied by the local authorities. The village had around 1,500 families of which 390 were classified as BPL households. Since 390 was a large number we restricted our study to one hamlet – Kherai Phala that had 156 families, with 49 in the BPL list. While the intention was to cover all the families in the hamlet, we ultimately ended up studying only 36 families due to limitations of time.

Details of the data pertaining to the village are summarised in Table 1. Design of questionnaire: For collecting household data, a detailed questionnaire was designed, with a view to capture financial flows of families over a long horizon of time. The base data were the demographic and asset profile of a household. Other data were built around this to get the financial history of the household. We collected details of income, indebtedness and savings. We sought inputs from local resource persons to include questions/ asset in the checklists specific to the geographical region.

We collected information on the income flows, agricultural land, physical assets, saving habits, loan transactions and the details of the events that happened in the family in the last 10 years. Although the questionnaire was not divided into different stages, each question collected specific information. This collectively gave an idea of the financial flows of a family. In the first part we collected data on the general family details, including income, inward and outward remittances. The second part collected information on landholding and details of other physical assets, including dwelling and livestock details. In this process we captured the information on financial transactions while purchasing or selling assets, the mode of financing and the purpose of purchase. The third part focused on the physical assets, where we captured the information on mode of financing, purpose of purchase, and its value. If any asset has been sold, we found the amount realised from the sale. By seeking this information, we tried to understand the process of acquisition and sale of assets and the circumstances under which they are acquired or sold. In the fourth part, we captured savings and indebtedness details of the family. We also asked the respondents to rank the sources with whom they had savings and loan transactions to get a feedback on their comfort levels, details on accessibility, costs, security and liquidity of the products they used. We also asked them the amount of maximum savings and loans and the source where it has been parked or drawn in the last 10 years. This roughly gave us an idea of the reach of the financial institutions and at the same time told us about the extent of convenience and faith the poor placed on these sources. It helped us find which of the formal or informal source provided most acceptable product. Similar details were collected on indebtedness. In the last part we collected details of the events that occurred in the last 10 years – such as marriage of the children, health expenses and purchase of assets or funeral expenses. These event details capture the financial flows involved with birth, death, education, marriage and emergencies. This gave insights into how such events are financed and managed. Administration of the questionnaire: Alsigarh is 35 kms away from Udaipur and is connected by buses and jeeps. This is a tribal

Size of the Population (in number) Male Female

Adult 1329 1311 Children* 540 539 No of BPL households 390 Land available in the village 554.86 hectares

Distance (in km) of infrastructure facilities (0 if located in the village)

outmigration (2) Low levels of wages for jobs in the village

Natural calamities Drought faced in the last two years

Major sources of (1) Village moneylenders money transaction (2) Shop keepers in the village (3) Family and relatives

Note: Includes children between 6 and 15 years of age.

area; all the dwellings in the village are dispersed. The distance between the houses made our data collection more effective as there was no disturbance from other families. This expanded the attention span of the respondent greatly. There was, however, a small problem about our identity. This was because a recent survey was conducted by the state to review the list of BPL households. So, it was difficult to get data in the initial days. However, we established that we were not from the government, but were associated with Seva Mandir, the data started flowing in freely.

The questions on which we had difficulty in getting data were about health-related problems and expenses. They were unwilling to talk about these issues. These details were collected in a circumspect manner. Data were not forthcoming on some sensitive issues as well. As this is a tribal area, there is a prevalence of bride price as against dowry in the plains. Getting data on bride price paid was difficult. Unlike traditions of other places, in this village a girl had the right to choose her own husband and for that she could stay at the prospective groom’s place before the formal wedding, to find out whether the house is suitable for her. If she does not like the place, she has the freedom to walk out. If the marriage happens, the groom has to pay the bride price and also bear all the expenses of the conduct of marriage. We had instances where a boy had married twice or thrice in a period of two to three years. Capturing these expenses, when the boy married more than once was difficult. The community as a whole bears a fair amount of expenses by way of gifts for the first wedding. However, this is not repeated for subsequent weddings if the first one fails. Collecting information on subsequent weddings was difficult, though there would have been significant expenses incurred on this event. Our data is inadequate on this count.

In this area people had a small piece of land, productivity was low and most of the produce was consumed. The levels of monetisation were also low. Imputing a value for self-consumption was therefore difficult. Using events as time markers were useful, but that gave us the data on financial flows at the event point. However, several respondents were unable to articulate their outstandings, due to low levels of awareness on aspects of repayment and the split between interest and principal.

The data was collected using men and women investigators. We found it was better to use women investigators for data collection. Using women helped us because:

However, the downside of collecting data exclusively from women put a question on accuracy. Ideally this data should have been triangulated by a short interview of the men. But due to constraints of time, this could not be done.1 The findings of this study should be interpreted in the light of this shortcoming.

General household and employment profile: We used data from 34 of the 36 households from which we collected information. The two households that were omitted were outliers, one being extremely poor and the other being at the other end of the spectrum. These 34 households had total 181 individuals – an average of around six persons per household. The basic demographics are given in Table 2. A third of the population was under the age 15.

sources. Many persons from the village go to Udaipur city to work with non-farm enterprises. Connection with the city has played a major role in diversification of livelihood opportunities. The new income streams discovered out of diversification from the present job has pumped in extra cash to the regular cash flow [Krishna 2003]. High debt had also forced them to come out of the village and look for alternatives that fetch them regular cash flows.

Sometimes the income is in kind. We captured this by converting the flows into monetary terms. For instance, grass and MFP collected, contributed significantly to the income flow of the household. These were monetised. In the upper end households where the income is more than Rs 4,000 per capita we found that more than one member of the family got regular work in Udaipur. Some of them also had land, adding to their flows.

Although we did not find households abandoning agriculture, Table 6 shows that agriculture is not lucrative and finding wage employment seems to be an alternative. The households falling in the lower income group, continued depending on agriculture, and were unable to move out of the poverty trap. Asset profile: The assets owned by the families are given in Table 7. From the list we see that apart from utensils, cots (charpai) and rudimentary farm implements, there is pretty little in the form of assets that the households had. The most significant asset in the households was silver. It was found during the field visit that most of the assets listed were not usually sold. People in the village prefer to borrow in times of crisis at fairly high rates of interest, rather than liquidate any assets and if they need to sell their assets they would first sell livestock but would not touch the jewellery. All respondents had a dwelling unit of their own. Some of them had two dwelling units, but the families used both. None of the families had leased out land, while several families had leased in land. Borrowings: The profile of borrowings is shown in Table 8. The maximum number of loan accounts was with moneylenders. However, the average size of a loan from moneylender was smaller than other sources. In all, borrowing from moneylender and other informal sources accounted for almost 85 per cent of the number of loans and 80 per cent of the amounts borrowed. Borrowing from relatives and from commercial banks had a significantly high average loan size. There was no significant difference between the source from which Group I and Group II had borrowed.4 It appears that SHG was not an option for Group I households. The formal sector has been unable to reach this segment of the population. The reasons might pertain to transaction size and costs. Even the SHGs were working with the upper end of the poor families.

When we compared loan amounts and borrower profiles, we found that the commercial banks have a bias towards making loans for productive assets (Table 9). The bank had given one loan for social consumption5 out of five loans made. The healthrelated expenses, contributed to higher expenditure.

The borrower portfolio was diverse for the moneylender. The moneylender had extended loans for consumption, social consumption, health expenses, buying assets, and also to meet charges for litigation. The moneylender loans for assets were mainly for the purchase of livestock. All SHG loans were for consumption. People borrowed mainly for consumption, social consumption and health-related expenses from the family sources. The community usually funded the social events in the village – the expectation was that the recipient would pitches in when there was a similar event in others’ family. Therefore, the borrowings for marriage and funerals were usually from informal sources. However, the community would not chip in if somebody got married for a second time.

Only one loan from the family sources was for buying assets. Tables 9 and 10 indicate that people borrowed from moneylenders for asset purchase. Borrowing from moneylenders for emergency purposes, is understandable, but the larger share in asset purchase indicates that there is scope for formal institutions to step in. We should also note that the most frequent purpose for borrowing is “health-related”. Health expenses indicate not only “out of pocket” expenses, but loss of income because the person would

Asset List Number Approximate Value of the Asset ( Rs)

Physical assets

Cycle 01(01) 200 Scooter 01(01) 7000 Clock 09(08) 940 Watch 10(08) 2210 Radio 05(05) 2300 Cot (charpai) 24(20) 5100 Chairs 01(01) 50 Elec connections (number of bulb points) 15(15) 3500 Utensils (approx value) -17900 Farm implements 52(25) 11500 Pump 01(03) 8000 Jewellery (silver) (approx value) -213600 Jewellery (gold) (approx value) -1500 Mahua trees 08(08) -

Livestock

Cows 37(26) 33700 Bullocks 55(32) 66000 Buffaloes 08(08) 9200 Goat/sheep 81(30) 35650 Poultry 41(16) 6720

Land (area in acres)

Own irrigated land 0.375(02) 22000 Own rain-fed land 20.5(34) 81000 Own non-cultivable land 11.7(25) 232000 Leased rain-fed land 1.875(11) 66000 Leased non-cultivable land 0.375(02) 10000

Dwelling

Small 17(17) Medium 20(16) Large 01(01)

Note: Figures in parenthesis show the number of households have these assets.

Details of the Monetary Break-up of the Transactions Client Base Sources No of Loan Ammount Ave Loan Group I Group II Accounts (Rs) Size (Rs) 19 HHs 15 HHs

Commercial banks/ 05 49,000 9,800 02 03 post office (7.8) (17.33)

Moneylenders 42 134,100 3,193 18 24

(65.62) (47.45) SHG 04 2,700 675 00 04

(6.25) (0.95) Relatives 12 91,800 7,650 06 06

(18.75) (32.48) Any other 01 5,000 5,000 00 01

(1.56) (1.79) Total 64 2,82,600 4,415 26 38

Note: Figures in parentheses represent the percentages of borrowers under each loan source where the total acts as 100. The total does not add up to the number of households as some households have multiple loans. Some of the loans from family sources are interest bearing, and some are not. The moneylender charges around 3 per cent per month, payable monthly.

be incapacitated. These causes accounted for 16 per cent of the loan accounts while social consumption accounted for 49 per cent of the accounts. This is a cause for worry, and is similar to the findings by Krishna (2003), where the major causes for slipping into poverty were because of these factors. In spite of drought for last three years, the heavy expenses on social consumption are somewhat baffling. The data on significant events, their average costs and source of funding was collected. Table 10 has the details. Savings profile: Table 11 shows the savings of the poor in institutions. Most of the savings in the bank and in the post office were fixed deposits. There was one recurring deposit account. Savings in SHGs were on the monthly basis. Many members were irregular in their savings. Most women kept savings in SHGs and with informal sources. Informal savings included cash stashed away from the daily earnings. Even this was irregular as there is no regular income flow in the household. So whenever there was a little money available with the women either by selling the MFP, vegetables or bamboo, they preferred to save in the safe earthen container inside the house but away from their husbands’ eyes. From the data on financing of asset purchase and financing of significant events, it was evident that these savings are very sparingly used for outflows. Sale of assets and jewellery was not seen at all in the sample households. Savings are perceived to be a different compartment that was to be used sparingly.

The total assets owned by the poor were 10 times more than the total borrowings on an average. An overall look at the income, savings and borrowings data indicates that the level of indebtedness is not alarming (the figure). In almost all cases the overall borrowing was less than their annual income, and far less than the total worth of the assets they had. In this sense no respondent suffered from a negative net-worth. However, what seemed to be very prevalent is stashing money away in pots, as there were no alternatives available for savings. Formal sources were accessed only by a handful of people and they also seemed to have multiple accounts. This problem was faced both in the borrowing and the savings departments. It is interesting to note that most of the assets are illiquid assets.

V Concluding Notes

Mapping the financial flow of the poor requires careful investigation of the income and expenditure patterns and the most important is the involvement of the people themselves. This paper illustrates the results of a study conducted in one village of Rajasthan which was under the influence of drought for last three years and has experienced some rainfall this year. Being near to the city solved some of the important problems – particularly pertaining to wage employment and helped them diversify their livelihood sources. Even health-related expenses would be have been lower due to greater access in Udaipur. Although there are various studies conducted to identify the factors that drag people into the poverty trap, the major findings of this study are that the overall asset-savings-income profile of the poor in this village give a comfort while compared to the indebtedness. However, most of the assets and savings are illiquid, forcing the poor to borrow at high cost and service such loans. The data on income might be a bit distorted, as the year of study was a non-drought year after three years of continuous drought.

Table 9: Number of Loan Accounts Classified on Source and Purpose

Purposes Sources Gen Consumtn Soc Consumtn Health-related Buying Assets Litigation Other6

Commercial bank/post-office -01 (6.66) -03 (15.78) -01 (25.00) Moneylender 06 (66.67) 08 (53.33) 10 (62.5) 15 (78.94) 01 (100.00) 02 (50.00) SHG 01 (11.11) 01 (6.66) 01 (6.25) --01 (25.00) Family/relatives 02 (22.22) 05 (33.33) 04 (25.00) 1 (5.26) -- Any other source --01 (6.25) --- Total borrowers 09 15 16 19 01 04 Total loan amount 12700 (4.5) 138500 (49.00) 45600 (16.13) 74800 (26.46) 2000 (0.70) 9000 (3.18)

Table 10: Significant Events and How They Were Financed

Event Detail No of Events Ave Amt Savings Borrowings Source for Borrowings in the Past Years Spent (Rs) Used (Rs) (Rs)

Marriage of children 19(17) 13,432 1,895 11,537 Bank (1) moneylender (8) informal sources (10) Health problems of family members 31(20) 1,955 674 1,281 Moneylender (9), SHG (1), informal sources (19) Construction of house 10(06) 7,570 1,800 5,770 Moneylender (6), informal source (4) Purchase of agricultural land 07(04) 3,457 428 2,428 Moneylender (6) informal source (3) Funeral expense 04(04) 1,200 -1,200 Informal source (4) Other (includes events like Gauri Puja, 18(14) 3,989 906 3,083 Moneylender (10), SHG (1) and informal source (2)

Bhumi Puja) Savings (5)

Note: Figures in parenthesis in column two indicates the number of households that have borrowed in the last few years. Figures in parenthesis in the last column indicate the number of events that have been funded by a particular source.

Table 11: Number of Savings Accounts in Different Sources

Details of the Savings Transactions Clients Belonging to: No of Accounts Amount (Rs) Ave Saving (Rs) Group I 15 HHs Group II 19 HHs

Postal deposits 4 (11.11) 6,300 (11.60) 1575 03(03) 01(1) Commercial bank/coop accounts 3 (8.34) 27,500 (50.65) 9267 02(02) 01(1) SHG 16 (44.45) 5,090 (9.37) 318 06(04) 10(09) Informal sources 12 (33.34) 13,700 (25.23) 1142 04(03) 08(06) Investment in assets 1 (2.76) 1,700 (3.13) 1700 -01(01) Total 36 54,290 1508 15* 21*

Notes: Figures in parenthesis in columns pertaining to accounts and amount of savings indicate percentage of the total. Figures in parenthesis in the last two columns pertain to number of households.

* There are 12 households (HH) which have got 15 savings account in <Rs 4,000 category and 18 households which have 21 savings account in >Rs 4,000 category.

Economic and Political Weekly December 23, 2006

Figure: Mapping of Various Important Variables acrossthe sample area was affected by severe droughts in the past threethe Households

years. A significant gap was also found in the lack of data

collected on current expenditure.

80000 70000

Email: mssriram@iimahd.ernet.in

60000

EPW

Amount in Rs