Financial Exclusion and Social Integration in Sri Lanka

Financial Exclusion and Social Integration in Sri Lanka Paper for the Conference on Finance and Business Development, April 5-6, 2001, at the University of Manchester Introduction “There are millions of entrepreneurs who lack access to financial services.” (Otero and Rhyne, 1992: 2) The main aim of this paper is to explore the causes of greater, and lesser, use of bank services among poor people in Sri Lanka. This paper is part of an ongoing research project with both quantitative and qualitative elements. In the course of this paper I develop a theoretical framework within which one can analyse the role of subjective positionings in people’s decisions about savings and credit. In this way, the paper offers an innovative extension of previous work on the psychology of savings and credit (work reviewed in the first section of the paper). The second section gives some background to the present study and provides an overview of the extent of credit and savings flows, relative to household income, among a national random sample survey of Sri Lankan people. Within this context it is shown that many poor people in 1996/97 were highly indebted and had substantial dissavings relative to current income. In the third section, qualitative research (which is still in progress) on the subjective attitudes of poor people and bank staff is explored. 400 questionnaires were used to explore the attitudes of people living in two poor neighbourhoods of Colombo and in two towns, one in Matara and the other Ehaliyagoda. We found few age or gender differences in the exploratory attitude survey, but regression analysis among the 400 respondents does show that women’s subjective positioning tends to be less in favour of using banks, compared with men. The poorest respondents were also wary of using banks more. These intriguing findings will be explored further in the research programme. In the conclusion some points are made toward a theoretical and normative analysis of the possible ways that financial exclusion occurs. The policy implications are different if exclusion occurs through targetting that fails; through rejection of desirous poor people from banking services; or from mutual separation. Several new research questions are opened up during this exploration, and in 2001 a further survey among rural residents will help to clarify the predominance of each of the three types of financial exclusion. I conclude the paper with a summary and the suggestion that explanatory models of ex post indebtedness based entirely on records of events, such as income flows and wealth levels, may not be adequate to the task of explaining change in the uptake of banking services in a rapidly changing culture such as that of Sri Lanka. Examining attitudes may contribute to the interpretation of changes that are now underway. Section 1: Review of Literature This section reviews the overall situation of Sri Lankan people’s access to bank credit, non-bank credit, and bank savings vehicles. Then I present a theoretical framework based mainly on French and UK research which enables the study of attitudes to be integrated with structural explanatory factors influencing the uptake of banking services. In this way I expand the question of ‘access’ to banks, as posed by Otero and Rhyne on page 1 of this paper, to allow for the subjective attitude of potential customers. In Sri Lanka bank borrowing comprised about 54% of the flow of total borrowing during the 1996/97 period when the Consumer Finance and Socio-Economic Survey was undertaken among a random sample of 8,900 households (Central Bank of Sri Lanka, 1999a). Figure 1 and Table 1 show the breakdown of borrowing among several competing sources of funds. Table 2 shows a similar breakdown of the flow of gross savings from households to banks and non-bank destinations such as stock markets. Figure 1 The details of one-month and six-month recall data on sources of credit appear below. Table 1: Debt Incurred by Source, Sri Lanka, 1996/97 (Rs. 000s) Source Flow (One-Month Recall) Flow (Six-Month Recall) Commercial Banks 8722 51% 28979 44% Other, Interest Free 3790 22% 10773 16% Moneylenders 1605 9% 7389 11% Formal Sector Employers 1294 8% 5420 8% Rural Banks 423 2% 2727 4% Thrift Societies 361 2% 1733 3% Other, Not Interest Free 239 1% 1275 2% Financial Institutions 228 1% 1702 3% Regional Rural Development Banks 143 1% 1090 2% Non-Governmental Organisations 118 1% 1107 2% Finance Companies 25 0% 3780 6% Co-Operatives 13 0% 271 0% Total 16961 100% 66246 100% Source: Central Bank of Sri Lanka, 1999a: 689, 721. Table 2: Flows of Gross Savings by Destination, Sri Lanka, 1996/97, Rs. 000s Details of the Breakdown of Savings & Investments 1996/97 Rs. '000s Destination Flow (One-Month Recall) Savings account deposits 6658 Savings account withdrawals 8312 Fixed Deposits- In 1239 Fixed Deposits- Outdraws 383 Shares Bought 126 Current Account Deposits 8683 Current Account Withdrawals 7484 Total Deposits 16706 Total Withdrawals 16216 RELATIVE SIZE OF DEPOSITS ONLY: Savings account deposits 6658 40% Fixed Deposits- In 1239 7% Shares Bought 126 1% Current Account Deposits 8683 52% Total Deposits 16706 100% Hulme, Montgomery and Bhattacharya (1996) provide some policy background to this situation. In 1996/97, non-governmental agency (NGO) lending was just a 5% share of overall household credit flows whilst commercial banks were 44 to 50% and moneylender credit was around 10%. Loans from friends and family without interest were another large chunk of the total flows of consumer credit. Hulme et al. argue that thrift and cooperative societies became popular during the 1980s in the face of the failure of the main banks (Bank of Ceylon and People’s Bank) to reach the majority of the rural poor (1996: 180). Hulme et al. stress that “ A consequence of the reluctance of private banks and the difficulties of state-controlled banks to service the needs of small borrowers is that informal finance remains of great significance in Sri Lanka” (ibid.). Whilst the NGO and Cooperative sector is small in absolute terms, as shown in Figure 1, it is important to the people who access it. NGO and Co-operative finance are also subject to different regulations to the commercial banks and, for the purposes of this paper, are considered as outside of the main banking system. Another newer source of micro-credit, the Samurdhi banking system linked to the Samurdhi welfare benefit scheme, is discussed later in the paper but is also not regulated in the same way as the commercial banks. Detailed micro research on attitudes among Sri Lankan people about savings and banking is limited, so I use European empirical literature as a starting point for the framework of analysis. In the UK, Lunt and Livingstone have argued individuals’ attitudes to savings and credit influence their banking behaviour in addition to the effects of a person’s income and the wealth or income of their household (Lunt and Livingstone, 1991; 1992). (Emergence of economic and social phenomena at one unit of analysis based upon underlying factors at a different level of analysis is offered by Lawson as a major contribution of realist methodology to socioeconomic analysis; Lawson, 1997.) Nevertheless, the work of Lunt and other economic psychologists (see for instance Farnham, 1985; Wärneryd, 1989; Lindqvist, 1981) has used three innovative methods to capture individual factors that are convincingly shown to be relevant to savings and borrowing. Firstly they have used focus groups to indicate what discourses of savings were present, and in the UK it was possible to discern age-group differences in these discourses (Lunt, 1996; and for an Indian case study, see Arvain and **, **). Prudence, pro-savings and anti-debt attitudes were more predominant among the older generation in the 1981 sample (Lunt, 1996) and a more carefree attitude to debt, later called ‘profligacy’ by other authors in this area, was observed among young people. Gender differences were, apparently, minimal, and ethnic group differences were not evaluated by Lunt or Livingstone. Secondly, the authors used questionnaires with a combination of ex post questions about the quantum of debt and types of savings along with attitude scales to gauge the extent and positioning of different attitudes in the UK population (van Veldhoven and Groenland, 1993; Lindqvist, 1981). Factor analysis of the attitude scales revealed several factors differentiating the sample (Lunt and Livingstone, 1992 book), and these factors in turn were used with the socio-demographic variables in a regression analysis of debt and savings outcomes (Lunt and Livingstone, 1991), showing that attitudes do matter to economic decisions of households. The R2 of multiple regression could be raised from 45% to 65% through the introduction of attitudinal factors among respondents. Interestingly, only one respondent per household was contacted and therefore no question of intra-household opinion differentiation could be raised. Lindqvist (1981) found using Swedish random-sample data for a date in the 1970s, that there was no significant factor influencing total savings or the flow of savings (R2 = 5% at the highest). However he also found that personal attitudes did matter substantially to the credit situation of a household. Using path analysis he discovered that income had indirect effects on savings and credit outcomes which were mediated by attitudinal factors. Because some effects offset each other, a single regression analysis was not adequate to deducing the relevant effects. Decomposition successfully teased out positive direct effects of income and negative indirect effects via attitudes. This promising path analysis methodology has not yet been applied to other data sets in so far as our research has so far discovered. In a third innovation, Viaud and Roland-Lévy (2000) have attempted to use the survey method and proximity analysis to analyse interview data by categorising the various responses. These authors ask the question whether people identify with a specific discourse, such as a high-spending low-saving profligacy discourse, by representing themselves in ways consistent with their overall social position, or whether instead representations chosen by respondents [i.e. their narration of identity during the interview] disengage with structural positions and focus, instead, on cohesiveness to an image, style, or discourse with which they wish to be associated. The latter finding would contradict both methodological individualism and a structural determinism that is perhaps also implicit in the Lunt and Livingstone’s predictive regression methodology (Lunt & Livingstone, 1991). Viaud and Roland-Lévy innovatively categorised the various activities and attitudes to savings and borrowing of fifty French interview respondents by using multi-dimensional scaling to create a matrix which could be examined for proximities among the locations of the respondents. They identified four ‘types’ of respondent: profligate; prudent; accumulating; and abstemious. Figure 2 summarises the framework offered as empirical findings by Viaud and Roland-Lévy. Figure 2: Four Types of Bank Users in France, Based on Multi-Dimensional Scaling and Proximity Analysis those who dislike saving those who like saving but like borrowing - and borrowing - ‘profligate’ ‘accumulators’ those who dislike saving those who like saving and dislike borrowing - but dislike borrowing - ‘abstemious’, abstaining ‘prudent’ from banking These four types combined tendencies to like or dislike saving and to like or dislike borrowing, and on the whole it seems that among the 50 cases people acted in accord with their own preferences. Thus support for the positioning of respondents in consistent ways was found, yet the small scale nature of the study restricted the authors from drawing any firm conclusion about whether in France consumers were acting in ways that might have been predicted by their structural location. Instead, the authors concentrated on stressing the link between the subjectivity (identity formation through and with discourse) and observed behaviour (reports of which bank was used, which type of savings was held, and so on). The authors’ claim that they were testing whether representations were consistent with positionings was perhaps an area for further research rather than an area on which they had definitive findings to offer. The hypothesis that representations may not be consistent with structural positions is of interest for policy makers and social economists in various country contexts, however. For instance, in recent years two banks in Sri Lanka have each been associated with specific ethnic groups (Buddhist and Catholic, respectively), and one might ask whether personal and household ethnic identities are associated with a preference for a particular bank (and its associated discourses, advertising, methods of interviewing, terms and conditions, and so on). In India and many other countries with secular legal systems, Muslim banks have also sprung up over the last two decades in response to a liberalisation wave that was intended to increase competitiveness by allowing free entry and competitive variation in the modes of operation of banks. Ironically, the ethnic identification of banks may imply a reduction of competitiveness – but only if people’s ethnic position implies both ethnic identification (which is not obvious, since ethnicities are complex) – especially if customers develop loyalty and/or an unwillingness to move to banks that are of an alien or ‘other’ ethnic identity. It is also possible that ethnic identification is a way of moving toward a market niche for a small bank. Thus the French research was small-scale and, arguably, explored the possibilities in interesting ways, whilst opening up many new avenues for social research on banking at the same time. Our present research aims to analyse four types of qualitative information to shed light on the research questions examined in the UK and French research outlined above: open-ended questionnaire responses translated into English; interview transcripts translated into English; loan and savings account application forms, again translated; and advertisements by banks and NGOs of financial services in the Sri Lankan context. There is overlap here with the work of Kemper (2001) and one can also draw upon a Foucaultian framework that would examine how people shape themselves to fit a discourse which might enable them to get respect / admission / approval from people in positions of power who shape the discourse. Based upon earlier research in India and Ghana, I expect to see people lying on forms, or leaving parts blank, in so far as that is acceptable to bank officers who wish (for predominating reasons) to lend to them or allow them to open a savings account. One might also expect to see images and styles in bank advertising that appeal to underlying social trends in Sri Lankan society, even while those trends are being affected by globalisation and the export of western culture and other cultural influences new to the island. Already three tendencies have been observed in Sri Lankan banking advertisements. Firstly, many of the advertisements refer to ancient Buddhist tales to imply that a good Buddhist will save with the bank to gain social status and an improved moral positioning. For women that may imply being a good modest mother figure while for men, and heads of household in general, Buddhist tales often give great importance to being prudent and sober. Sobriety is strongly advocated by the monks and temple staff. A second theme is a modernist one, rather than a religious/ethnic one. Many advertisements encourage the reader to associate their banking activities with their ability to access high technology and have a white-collar job. As found in my Indian research (Olsen, 2000: chapter 7; and a separate survey of labour-market stereotypes, Olsen, 2000b), it seems that Sri Lankan rural and non-white-collar workers are actively emulating the white-collar sector and desire to engage in modern, clean, high-prestige service-sector activities such as computing and having a credit card or cashcard for the cashpoint. Further research on the advertisements will illustrate the modernisation theme. Thirdly, some advertisements appeal to a profit-oriented populace who are keen on free markets and competition. As in the West, Sri Lankan advertisements claim that banks give good interest rates to savers and charge low or reasonable (ie competitive) interest on loans. Our exploratory survey indicated that hardly anyone believes these claims, with 70% of respondents feeling that banks should pay higher interest to savers and 80% of respondents feeling that banks should charge lower interest rates to borrowers. There is little prior research in these areas in south Asia. It is possible that the modernist theme appeals to accumulating and profligate people (using Viaud and Roland-Lévy’s framework), the Buddhist theme appeals to the prudent and abstemious people, and the profit-oriented theme appeals to the accumulating and prudent people. This is an empirical question but one might also expect to find a dialectical relationship between attitudes and decisions of consumers, rather than a deterministic one. Finally we should mention a vast literature on the way that personal earnings, permanent income, and the life-cycle of household overall earnings influence the aggregate savings of households. Mavrotas and Kelly offer a review, including an introduction to the newest element in this approach which is to integrate people’s understanding of their state pension rights into their savings plans (Mavrotas and Kelly, 1999; see also Schmidt-Hebbel et al., 1992, for a review and international empirical estimates of income effects. It has been pointed out recently that as the state pension schemes change and other social-security benefits are increased or decreased, people react by adjusting their savings and their overall portfolio of stores of value (including non-monetary forms of hoarding, such as animals, land, and durable goods) to keep their portfolio at an optimum. The present research is intended to complement this body of established theory. One addition the present research makes is to offer, as a problem, the possibility of disagreements among household members about the individuation of ownership of collateral or savings, and disagreements about the target savings level and optimal debt level. Once the qualitative data are complete, intra-household gender and generational issues can be raised. A final area of relevant literature on uptake of savings options is the transactions cost literature. The issue of transactions costs as experienced by poor people in third-world contexts were reviewed by Mosley in Hulme and Mosley (1997, Volume 1) although it is recognised that transactions costs are notoriously difficult to measure. Many respondents have commented on transactions costs in our research, and they mention repeatedly the time it takes to organise a transaction apart from the direct costs associated with travelling and getting certificates. The opportunity costs of time are part of the felt costs of making a transaction, but they differ (in money terms) depending on the earning capacity of the person. One approach would place market values on time, making rich people have higher transactions costs per unit of time than do poor people. A different approach might allow for the subjective rate of time preference or subjective valuations on the returns to time spent in different activities (such as watching animals, gathering goods from common land, or checking on irrigation) many of which do not have monetary remuneration to enable an ‘estimate’ of opportunity costs. These unresolved problems in comparing opportunity costs for poor versus non-poor people have inhibited new institutionalist economists from undertaking much detailed empirical work (with some notable exceptions, such as Hulme and Mosley, ibid)., but they have not inhibited them from developing relevant theoretical models. Again the present research complements this body of work. The factors causing people to save less or more, and to borrow less and more, from the banking system are all changing over time. We have mentioned several in the review of relevant literature: personal attachments to discourses of modernisation that make banking look desirable; the return on saving and the direct and indirect costs of borrowing; prudent or frugal habits and routines that may match up with a discourse of sobriety; and the earnings and wealth position of a person, and their household, which may change and thus cause the desired savings to exceed the actual savings (or debt, or both) of the person / household. Our related quantitative survey research tries to tease out the more easily observed of these factors across Sri Lanka’s varying districts and to distinguish financial exclusion due to the absence of bank branches from financial exclusion that is more due to personal preferences or lack of funds among the poor (Olsen and Kirkpatrick, 2001a and 2001b). This literature review has focused on monetary, psychological and discursive factors at the cost of attending to the underlying labour relations. It is worth noting that micro-finance and credit for small business proprietors arguably fall within the consumer-credit remit. Consumer credit for productive purposes in turn is linked to developments in labour markets. Specifically, lending to micro-entrepreneurs aims to keep them in the informal sector or help them progress to becoming small businesses; it may also aim to help them employ other people on flexible or family-labour arrangements. Savings arrangements for the poor, too, have often aimed to encourage these people to use their savings for investment purposes. Whether they invest in human capital or in financial and economic capital, it seems wiser within capitalism to spend lump sums in such ways rather than merely on expensive luxury goods. In these ways the micro-finance literature often presumes that poor people need to make arrangements for their own survival by creating jobs and avoiding luxury spending. (Gulli, 1998, reviews this situation, whilst Rogaly, 1996, has argued that it is unfair to place upon poor women the expectation that they would use lump-sum funds for anything besides emergency spending or investment in their own family’s human capital.) Feminists break through this contradiction by arguing for a lessening of the strong border lines between ‘productive’ and ‘consumption’ spending. Because consumption spending is aimed at the reproduction of a flourishing household and family life, it is linked to social reproduction, and that in turn is linked to production since satisfied people are able to produce more both for markets and for home consumption – and also produce other people. Many ‘productive’ sectors are also integrally involved in social reproduction, notably food but also the transport, education, capital-goods, manufacturing, and even the banking sector itself. Thus reproduction can barely be separated from production. Instead, say some feminists, the judgement that some social classes are qualified to spend money on luxuries while others are expected to spend money on investments is a reflection of values implicit in capitalism. Within capitalism freedom, in this view, involves wider ranges of choices for the well-off social classes and highly constrained choices for poor classes. This debate, highly relevant to policy makers wanting to avoid the social exclusion of poor people, is commented upon in lucid ways by Sayer (1995, 2000) who argues that we must dissociate analysis of markets, including the financial markets, from analysis of capitalism itself. Whilst radicals, feminists, and marxists may want to question the underlying class structure of capitalism, that is still a different question from asking how financial markets should operate in a society with a complex division of labour. The two are not the same question. Indeed, Sayer argues, it is worth asking what to do in immediate policy contexts that will benefit people. This pragmatic approach offers empirical work in addition to what some marxist feminists presently do, which is to critically analyse only the faults of the capitalist system and sometimes to equate capitalism with the presence of markets and globalisation. Sayer’s critique of the usual modes of radical left thinking is a helpful guide to how several kinds of normative thinking are possible during and after solid empirical research in a specific historical context. The specific labour relations in Sri Lanka have been changing in interesting ways during the last two decades. Many authors have commented on how scarce and treasured government jobs are; during liberalisation however private-sector jobs have increased massively compared with the older plantation-sector employment and farming livelihood strategies. Statistical records show a strong feminisation of the Sri Lankan labour force, both in agriculture and in non-agricultural sectors (Central Bank of Sri Lanka, 1999b). Specifically, there has been a rise in the female labour force participation rates from 28% rural and 22% urban in 1990 to 38% rural and 27% urban in 1998; among males these rates rose from 51% to 68% (rural) and from 48% to 65% (urban) over the same period (ibid.: 15). Public-sector employees fell from 22% to 14% of the employed labour force and private-sector employees rose from 34% to 41% during the same period 1990-1998 (ibid.: 16). Unemployment fell during this period. There has been a growing involvement of women in informal-sector micro-enterprise during the liberalisation period 1977-present, as seen also in Mexico which has also had long-term structural adjustment policies since about 1982 with emergent feminisation of labour and employment (Chant, 1994: 206-8). Financial exclusion of the poor in the UK is generally considered to mean a lack of access to banking services. It has been interpreted as being caused by the closure of bank branches and building society offices (e.g. Kaur, et al., 1996) and thus ignores the possibility of informal-sector lending offering a substitute for bank services in remote areas. In this paper we adapt this kind of framework for the possibility that informal-sector arrangements (dowry, moneylending, intra-family loans) offer serious competition to bank services at this time in Sri Lanka; analysis of discourses surrounding financial transactions then requires sensitivity to people’s attitudes and practices in money transactions generally, not just with banks. This approach reverses the insight of Bouman (1994), who argued that informal finance gives insight into the underlying society; we would argue that the underlying social formations and ideologies give insight into how people choose between the informal and formal financial sectors – but that the empirical research requires detailed qualitative analysis. However our research programme also allows for spatial aspects of financial exclusion (Olsen and Kirkpatrick, 2001a and b) Figures 3A and 3B: Flow of Debt From Moneylenders and from Banks in Sri Lanka Source: Central Bank of Sri Lanka, 1999a. Notes to Figure 3: Spending units were grouped into 16 ranked income groups, and within each group the total amount of loans from ten sources over a one-month recall period were averaged. In this sense, the data record the flow of borrowing rather than the stock of debt. In Figure 3A we have separated debt from moneylenders out of the total, and presented it as a percentage. In Figure 3B the sum of debt from commercial banks, rural banks, and development banks is presented as a percentage of the total flow. The other sources in the total include loans from friends and family on zero interest; non-governmental organisations; co-operative societies, and others. See Appendix 3 for caveats regarding this data source. Section 2: Introduction to the Empirical Research In 2000 four sites for exploratory research were chosen: two in urban Colombo and two in rural areas. In Colombo, respondents in Rathmalana are re-settled ex-slum dwellers originating from central Colombo. These are the poorest group overall of the four, whilst respondents from Dehiwala, also in Colombo, included numerous stallholders and market traders of varying levels of income and wealth. In Eheliyagoda, a town about 8 km from Colombo, a wide range of respondents were obtained from very poor to very well-off; and in Matara town in the southern tip of Sri Lanka a range of town residents (but just five farmers) were selected. In each area, quotas were set for 100 respondents to include half women and half men, representation of each main age group (15-25, 26-45, and over 45) and representation of a variety of local occupations. Each respondent was asked to respond to a short questionnaire. This questionnaire contained some precoded categorical answers, which are reported on here, as well as blank space for qualitative commentaries. In total 406 questionnaires were received. The exploratory questionnaire lacked details of household or personal income so the rest of this section presents some background data on the associations between income, credit and savings for Sri Lanka as a whole. Before presenting this background, a review of welfare benefits in the country will be useful since they form an important source of transfer income for many poor people. In 1995/96 the receipt of three welfare benefits (foodstamps, Janasaviya, and Samurdhi benefits) together amounted to 11% of household income, according to a national random sample survey of 30,000 households covering all districts except the two that had substantial violence in the north and east of the country (Department of Census & Statistics, 2000: 15). The targeted benefits of these three schemes comprised 25% of the income of the lowest income decile, and 18%, 14, and 12% of the next three income deciles, showing a high involvement in the schemes among Sri Lanka’s poor (ibid., p. 14). Samurdhi scheme eligibility thus reached 66% of the lowest decile (and 61%, 58%, 51%, as shown in ibid., p. 15) who could then participate in joint savings scheme and borrowing groups if their locale had access to the support services of the Samurdhi administration. According to an independent report dated 1999, the Samurdhi financial services have reached many people but have been run without proper financial administration since they are kept outside of the list of banks which have to comply with central bank directives such as reserve requirements and official auditing (Gunatilake and Salih, 1999). Audits of the schemes show a danger of overcommitment of funds and potential for local biases in the allocation of funds (according to the Economic Research Department of the Central Bank; memorandum, no date). Members of the borrowing activities of Samurdhi are made to save a small regular amount, and they cannot have access to these savings; such requirements have been criticised as discouraging the involvement of poor people in international comparative studies (e.g. Gulli, 1998). Indeed in the Sri Lankan case there are concerns that many non-poor people have had access to the Samurdhi scheme, blunting its anti-poverty focus. For instance in 1995/96 4% of the income of households in the top three income deciles came from the three main welfare benefits (Department of Census, 2000: 14), and 13% of these households were recorded as being recipients (ibid., p. 15). The Central Bank memorandum was critical of the access of non-poor recipients to the Samurdhi banking scheme nationally. Access of each income decile to loans in total, across Sri Lanka during 1996/97, is reported in Figure 4 below. Figure 6 shows the breakdown of savings, relative to income, for the same deciles. Figure 4: Total Flow of Debt to Households in 1996/97, by Income Decile Source: Central Bank of Sri Lanka, 1999c. Figure 5: Total Flow of Savings (Adjusted) of Households in Sri Lanka, by Income Decile Source: Central Bank of Sri Lanka, 1999c. By way of contrast, in Figure *, data for the UK show also the relatively high indebtedness of the lowest income decile. If mortgage debt were included, this figure would show a rising curve with income, but the high debts of the poorest group are not primarily mortgage debt. Figure *: Consumer Debt Excluding First Mortgage, 1988, UK Source: PAS Business Surveys (1988) Consumers' Use of Credit Survey, London, OFT. Figure 6 focuses more directly on the two subjectivities that come into contact when someone experiences financial exclusion. Figure 6: Subjective Positionings of Poor People and Staff Regarding Financial Exclusion/Inclusion Subjectivity of Respondent Toward Banking System: (below ↓ ) Whether bank staff and/or regulations exclude poor people: YES, exclude: NO, do not exclude: Whether the poor person her- or himself wishes to avoid using the banking system* AVOID: mutual separation AVOID: self-exclusion USE: bank rejection of poor customers USE: financial inclusion *A likely variant is that they are trapped into an exclusive relationship with a moneylender. To be analysed after all the rural fieldwork and interviews are done. Other alternatives include family lending and micro-finance institution sources including Samurdhi. The four possibilities have very different policy implications. In the next section we examine the qualitative data for clues about whether the potential bank customers are voluntarily avoiding banks or are being rejected by banks. Section 3: Qualitative Findings (to be augmented later with analysis of 50 interviews) Investigation of the 406 survey questionnaires reveals strong support among the (admittedly non-randomly chosen) respondents for the use of seetu by poor people. This usually small-scale type of revolving savings society involves groups of which half are net borrowers, and half are net savers, over a period of mutual monthly payments. Since each member gets the whole pot of money once during the period, seetu clearly acts as a substitute for small-scale savings aimed at lump-sum expenditure. Beyond this area of agreement, however, widely divergent views of dowry, use of the banks, and the co-signing system (locally called the use of guarantor; in India this is known as surety) emerged, as Table 3 shows. Table 3: Breakdown of Opinion Among Residents of Four Poor Areas in Sri Lanka Frequencies Gender Difference Ethnic Difference Is seetu a good system for poor people, in your view? no = 17%; yes = 81% no significant gender difference (Chi-square value .62, signif. = .43) no significant ethnic difference (Chi-square value 1.95, signif. = .377) In your opinion should people be able to borrow more money than they do now? less = 18%; same = 23%; more = 59% no significant gender difference (Gamma signif. = .356) Chi-square = 16***; Gamma signif = .019**; English speakers most likely to say ‘more’. Should a poor person be asked to give a guarantor or co-signer who would sign their loan application? it is bad = 39%; have doubts = 20%; it is good = 41% no significant gender difference (Gamma signif. = .957) Chi-Square 1.69, not significant; no significant ethnic difference (Gamma signif. = .801) In your opinion is dowry a way to get savings for the young couple? do not believe in dowry = 18%; dowry is not savings = 11%; dowry is savings = 71% no significant gender dfiference (Chi-Squared value 1.68; signif. = .43) no significant ethnic difference (Chi-squared = 1.87, signif = .758) In your opinion is buying jewellery a way to save money, and should people be helped to buy jewellery and gold more? more gold as savings = 50%; don’t know = 31%; less gold as savings = 19% no significant gender difference (Gamma signif. = .490) no significant ethnic difference (Gamma signif. = .333) Do you consider yourself as a religious person? no = 8%; yes = 92% Women significantly more likely to say yes; Chi-square = 2.829,signif = .093* no significant ethnic difference (chi-square value = 1.65, signif. .439) Do you consider dowry as an essential thing in marriage for yourself? no = 78%; yes = 23% Women significantly more likely to agree; Chi-square = 11.37, signif = .001*** no significant ethnic difference (chi-squared value .327, signif. = .849) What is your opinion on Western culture? Do you think it is a bood thing or do you think Western culture is a bad thing? bad = 44%; somewhat good = 3%; good = 44%; do not know = 8% Women significantly more likely to think it ‘bad’; Gamma value -.184 with significance .059* ambiguous result; Chi-square = 13, signif. = .01; but 22% of cells have expected count <5; Gamma not significant (signif. = .233) Differences by gender and favourite language have been tested for using a chi-square or gamma test. *** indicates significance < .01, ** signif. < .05, * signif < .10. Table 4 shows the overall distribution of respondents according to local ascribed social status, occupation, language group, gender, and age-group. Just a handful of farmers were in the sample so we cannot comment on the views of agriculturalists in this paper. Table 4: Socio-demographic Characteristics of Sample from Four Poor Areas Breakdown by gender Breakdown by other variable local status level poor in local terms = 24%; middle group = 62%; and well-off in local terms = 15% No significant gender difference 32% poor in Dehiwala; 5% in Rathmanalana; 31% in Ehyliyagoda; and 28% in Matara occupational group* 51% of the sample were permanent employees 55% of men and 48% of women were permament employees. 75% of Dehiwala sample were permanent employees; 31% of Rathmalana sample; 39% of Ehyliyagoda sample; 61% of Matara sample. language groups (What is your favourite language for speaking?) 327 Sinhala; 46 Tamil; and 33 English as first language no significant gender difference Percentage Sinhala speaking 74% to 96% in all areas; % speaking Tamil ranged from 20% in Dehiwala down to 4% in Eheliyagoda; the remainder English (no Urdu) gender* 203 females; 203 males 50% female and 50% male Somewhat fewer Tamil females than in other groups * In each locale, the interviewer was asked to get equal representation of the two genders and a representative selection of people from different age and occupational groups. The aim was to represent the potential clientele of banks in that locale, including the very-poor as well as the less-poor and non-poor. Our expectation is that structural social locations, such as occupational group and gender, would influence a person’s attitudes and hence their understanding of what banks are for and what financial activities the person should undertake. Regression analysis of the patterns soon revealed few gender, ethnic, occupational, or age-related patterns in the exploratory attitude survey, with one exception. Females were far more likely than males to point out problems with Western culture. The question on Western culture has its origins in surveys of youth (Hettige, Fernando, et al., 2000) and voters (Hettige, Mayer, et al., 2000) which, through extensive piloting, reached the conclusion that attitudes to commercialism and individualism were important to voters’ attitudes although it is very difficult in a structured survey format to elicit these viewpoints. Our exploratory questionnaires were intended to elicit qualitative answers, and these have yet to be analysed in full, but categorisation of the qualitative answers was possible on the ‘pro/anti western culture’ question because of the strong polarity of views that existed at the time in Sri Lanka. We found women more likely than men to point out problems with Westernisation, as shown in Table 3 (column 3). This pattern was in turn linked to a hesitation or suspicion of bank borrowing, as shown in the path diagram in Figure 7. Figure 7: Analysis of Attitudes to Banking Among Residents of Four Poor Locales Female + Perceive Problems with Western Culture – Age (no discernible effects ) + Think the Banks – Should Lend More Poorest Group + to People + Most Well-Off Group - Think Guarantor System is a Bad One For Poor People Sinhala Main Language + Think Poor People Should be Asked by Banks to Provide a Guarantor NOTE: Dotted lines refer to inverse associations and solid lines refer to positive associations. All are measured in terms of the way a variable affects the odds of the outcome occurring, and we used logistic regression. Details are in Table A1. Source: Field data collected by Wendy Olsen and Nishara Fernando with research assistants, 2000/01. As a starting point for the analysis of structured viewpoints, the exploratory survey offers an intriguing initial finding. Women and the poorest of the poor, who are precisely the groups focused upon in our study of financial exclusion, tend to have a specific type of rationale for being wary of banks lending more to poor people. Our analysis of the open-ended answers from these respondents, and of additional in-depth interviews during 2001, will (we hope) provide answers to this mystery. Initial hypotheses would be (1) that the individualism implicit in western consumer emulation of rich people’s habits may be avoided by poor people and women who are aware that they need to spend less money than they earn (rather than more) if their finances over a lifetime are to sustain life and the development of their human capital. In other words getting heavily into debt to finance a consumer lifestyle may be bad for a poor person. We call this the frugality hypothesis. A separate paper takes up prudent and frugal behaviour in more detail (Olsen and Kirkpatrick, 2001a) (2) A second hypothesis is that poor people and women have alternative sources for borrowing which they prefer to banks, and their choice is more one of voluntarily keeping themselves out of the time-consuming and prescriptive routines of becoming a bank borrower. We call this the bank-avoidance hypothesis. Whilst the other paper takes this attitude up in more detail (with reference to other research on abstemiousness and the psychology of avoiding indebtedness) we wish to stress the complexity of examining bank-avoidance. Persons may be tied to an employer, a trader, a landlord or a moneylender and through these ties be discouraged from taking up competing sources of loans (Basu, 1986). In Sri Lanka one case study examined such ties in detail (Risseeuw, 1991) and showed that women not only avoided banks, but also wished to create non-individualistic collective organisations to depersonalise the pressure associated with challenging traders’ local monopolies. Elsewhere women have also repeatedly expressed interest in collective action, forming groups and acting in concert rather than individualistically (Purushotaman, 1997) and theorists have described this as empowerment-with rather than individual empowerment (Mayoux, 1998). Since evidence elsewhere shows this tendency to be rather concentrated among women, a gender question has now arisen from the empirical research. However we wish to make individualism and collectivism a topic for further empirical research involving both men and women, and the interim findings shown in Figure 7 are thus influencing the ongoing field research. In the open-ended questions on the exploratory questionnaires, we found several examples that illustrate bank rejection and financial inclusion, and these are extracted below in Table 5. Table 5: Examples of Subjective Views of Bank Rejection Bank rejection: Dehiwala respondents said: Id 72- there are more rules in government banks but not in the private banks (specially when we need to have loans) Id 63- it is very hard to show the documents and also pawning the deeds Id — 36- difficult to get loans because they ask guarantors and deeds A Rathmalana respondent said: Ser.No: 45; The people’s expenditure is very high in now a days. People, who like me, they can’t have even a loan from the banks. Mutual separation: Respondents in Matara, in south Sri Lanka, said: Id 23 — the main problem is some social norms and specially religion. Apart from that people refuse the bank because of the low interest for the savings Id 05- the majority of Sri Lankans are living under pressure of poverty. So, how can they become the customers of banks. Id 83- the main problem is no education of the people. Next is the poverty and another problem is filling unnecessary documents. A respondent in Ehaliyagoda said: Ser.No: 01; Waste the time and do not care about the customers and majority of Sri Lankans are poor. How can they become bank customers. On the other hand, people do not have enough time and enough money to save with the bank. Financial inclusion: As yet there is little evidence yet for a situation of mutual subjectivities favouring financial inclusion of the poor in the banking system. Self-exclusion: In Dehiwala, urban Colombo, people said: Ser.No: 36; This method is good for anyone to get quick money. It is very reasonable way to get loans people like us. (Businessman) Ser.No: 63; We can have a loan as soon as possible from the moneylenders. So moneylending is somewhat good job according to my view. Ser.No: 72; Some time people are facing a lot of financial difficulties. Then they can have a loan from the moneylenders without any regulation as not in banks. Ser.No: 80; We can’t say that money lending is bad business. Because we can get money for our needs from the moneylenders very easily. In Rathmalana, respondents said: Ser.No: 45; The both of bank and money lending are not a anti-social. This sector (money lending) will reach to every people in the country introducing quick systems to get loans etc. Ser.No: 19; Money lending is very useful for most people. Their lending procedure is better than banks. In Ehaliyagoda a respondent said: Ser.No: 68; We can take loans without any problems. But we have to pay a high interest rate. On the other hand, money lending is not a legal business There is thus evidence of three possibilities of financial exclusion in the exploratory study. The preponderant view seems to be toward the mutual separation and self-exclusion side, rather than bank rejection of potential clients. In other words the poor people and the banks seem to mutually agree to leave each other alone, especially since the moneylenders offer an active non-bank competition. It should be remembered that in this paper financial exclusion refers to the non-uptake of formal sector banking and official non-governmental lending and savings schemes, not to informal moneylending. Therefore we have the problem that the hidden sector is a crucial competitor to the banking sector. The hidden informal sector was described by Ghate and Ballon as the bulky mountain of debt beneath the visible tip of the formal-credit iceberg; (1998), yet arguably it needs a careful normative analysis in its own right. Several authors have attempted this in different contexts but in Sri Lanka the studies of informal-sector moneylending are few in number. Interestingly, informal sector credit is ignored completely in western industrial countries’ studies of debt, both empirical and theoretical. Further analysis of the 2001 interview data from Sri Lanka is expected to show poor respondents associating western consumerism with a relatively profligate lifestyle and individualistic or opportunistic thinking. Those who feel positive about one of these may feel attracted to the others, too. This pattern may get expressed in their (personal) higher uptake of bank loans and formal-sector savings schemes. In other words we expect the uptake of this discourse to be a causal mechanism leading toward either bank rejection or financial inclusion – depending on the reaction of bank staff acting in the context of bank regulations. (The policy implication if these people face bank rejection is that either the bank behaviours need to change or that the potential customer needs assistance to conform to bank regulations.) By contrast, we also expect that the view that banking is undesirable will be a causal mechanism for some people leading toward self-exclusion or mutual separation. We may find an agreeable truce in which undesiring poor people avoid the use of banks, and bank staff avoid such people. However we are aware of the danger of subjective self-exclusion by deserving respondents, such as poor people who have no non-bank savings vehicle, and we are concerned that anti-Western thinking might be a barrier to the uptake of financial services. Evidence given above indicates this may be most common among poor people and women. Indeed, if alternative sources of loans are monopolistic (e.g. moneylenders) and if alternative ways of saving are socially undesirable (e.g. dowry) then it may be worth raising the issue of what is good / what is bad in ‘western’ thinking among respondents who express this kind of discourse. Progress in the credit markets may involve subjective changes, and hence discursive changes, among potential customers as well as among bank staff. Our further interviews and detailed analysis will allow us to comment briefly upon this. A comparison with qualitative research on banking in India may be useful here, although the political and social contexts are very different there (for background, see Olsen, 1996). In rural south India four discourses of banking were observed (Olsen, forthcoming), and no mention was made of anti-western thinking or self-exclusion from banking. Instead, the poor rural residents tended to all share the assumption that banks are helpful but that bank rejection of poor and landless people was the main problem. A dominant mutual-benefit discourse expressed optimism that bank savings and lending (and private moneylending) provided a useful service of benefit to both the organisation/lender and to the customers. Competing discourses were: 1. avoiding the exploitative moneylender by taking up the bank loans instead (escaping exploitation); 2. transforming the banking system by creating collective organisations involving social capital and women’s collective action; and 3. seeing oneself as isolated, as unusual, and as exceptionally using a bank or NGO to evade the routines of both moneylending and banking, which in themselves were seen as barely offsetting the cruel regime of poverty in which one is rooted. These discourses range from the very optimistic and hopeful mutual-benefit discourse, associated among international elites with neoliberalism but rooted in rural India with a deep populism and faith in state banks . . . to the deep scepticism and pessimism of the financially excluded in the last discourse mentioned. Although this analysis was based upon interviews with women only, and is therefore incomplete, it suggests that deeply rooted subjectivities with great historical specificity are involved when one tries to make an empirical study of the discursive basis for financial exclusion and bank rejection. Section 4: Theoretical and Normative Analysis Sayer has argued for the inclusion of normative analysis in empirical research on unequal societies (Sayer, 1995; 2000: 238). The specific questions opened up during the present empirical research are many, but here we focus on two. Apart from being questions of immediate policy relevance, these two normative questions also raise broader ethical issues and create a basis for sustained debate about the aims of bank restructuring in Sri Lanka. The first question is whether subjective approaches that favour financial inclusion are morally good, even if several objections could be placed against them? For instance, one might argue that subjective inclusiveness is meaningless if it doesn’t work in practice. Indeed, if alternative non-bank sources of finance are persistently preferred to banks by poor people, then financial exclusion may not be a major policy issue. Alternatively (further exploring ethical implications of the research), a neoliberal might argue that inclusiveness if fine but only under conditions that make banking sustainable or profitable. A reformist or Keynesian might argue that some restrictions on inclusiveness are necessary in order to boost overall aggregate savings and investment, or their efficiency. Finally, governance questions may be raised by feminists because of the passive role customers play in formal-sector banking compared with the high accountability expected of some NGO micro-finance institutions and the government’s own allocation of budgets. In commercial banks, internal banking practices and value judgements are not subject to much democratic scrutiny and some feminists have begun to question whether purely financial principles, operating in a hidden, hierarchical framework within large commercial banks, are the best way for society to allocate its loanable funds – including those gathered as savings (Hutchinson and Burkitt, 1997, 1999). By contrast, community-based micro-finance institutions are sometimes lauded for their internal democratic workings (e.g. see Albee and Gamage, 1996; or Hulme, et al., 1996 for details). A second ethical question is whether bank avoidance or frugality are morally good. If so, are they good as a substantive social good (a situation of goodness) or only as a procedure social good (a means to good ends)? The complexity of the real situation must be taken into account instead of offering a glib answer. For instance, farmers’ bank avoidance due to moneylender-trader relationships raises a genuine clash between procedural and substantive normative ‘goods’: having the relationship keeps one out of banks but helps one do trading, and thus perhaps facilitates economic growth; however having a relationship but also having bank savings might be a substantively better situation (with risk spreading, diversification, and more competitiveness both on the money markets and in trade) whilst not being as efficient at achieving the overall aim of rapid economic growth. These difficult normative questions are the tip of the iceberg and Sayer is right (1995; 2000) to argue that this is an area for serious policy debate. In this section we have tried to indicate the kind of discussion one might want to have after the qualitative and quantitative research from this project is finished. Whether development policies should aim for long-term or short-term development aims is a large dilemma continuously facing policy-makers in the year 2001. The reproduction of societies with substantial social and economic inequality usually includes the perpetuation of ideologies which legitimate inequality whilst fostering national unity or feelings of belonging within a divided society (Durkheim, 1947; Broadfoot, 1999). One view of the Sri Lankan situation from 1977 up to the mid-1980s was that liberalisation was a free-market, secular ideology opposed to the growing ‘primordialism’ (Sharma, 1988, ch. 7) of religious, ethnic and linguistic loyalties. In this context, Sri Lankan nationalism has had several streams of slogans aimed at fostering national unity, and for some these are Buddhist or Sinhala whilst for others they are secular notions of freedom within a united Sri Lankan state. Sharma uses the term ‘primordial’ for the type of discourse which appeals to ethnic identify to achieve a sense of social unity and integration. By contrast, as Sharma, Kearney and Foucault have pointed out, liberalism itself has in many places put forward a discourse of social integration centred on individuals’ freedom and citizens’ capacity to take up benefits available to all and equally serving to enable individuals to participate as equals in a growing, improving society (Foucault, 1976; Kearney, **). Critics argue that the liberal ideology masks the continual re-creation of poverty (Jones, 2001) and that it appeals to the masses to accept welfare benefits as a compensation for their low place in society (Caplan, 1985) without questioning the basic distribution of control over resources in society. Novak’s analysis of UK welfare systems, for instance, argues that during the Conservative government of the 1980s ‘the system of social security has been used to play a major part in this creation of a more coercive social policy and a more authoritarian state. . . In doing this the social security system does no more than it has always done. . . In the course of its history the state has at times had to respond flexibly to the political challenge created by poverty. But the social security system has never been intended to abolish poverty, nor has it been capable of doing so. The centrality of poverty to capitalism, and the importance of social security as a means of maintaining poverty and of “moralising” and regulating the poor, has and will always set the limits of its operation.’ (Novak, 1988: 200). There is, therefore, an ideological aspect to financial exclusion/inclusion which deserves empirical study because it will affect the implications and ramifications of uptake of specific forms of credit. In this paper we have noted gender differences in personal discourses about banking in Sri Lanka, and thus highlighted an aspect long neglected in the analysis of the legitimation and social-inclusion aspect of development banking. In this paper we have examined the financial exclusion of poor people in Sri Lanka and described an ideology of banking in which poor people are allowed to enter the system in two main ways, both of which are rich with power-laden discourses that influence the decisions of the poor as well as of bank staff. Firstly, poor people must comply with specific dictates of banks and other financial intermediaries if they wish to be ‘eligible for loans’ and in the target pool of clients. This compliance becomes voluntary and is internalised by poor people who aspire to emulate businesspeople. Secondly, poor people who take up the Samurdhi welfare scheme as a ‘right’ are welcome to join savings and loan schemes under this heading, but only after accepting that labels along the lines of ‘needy’ and ‘deserving’ apply to their household. This second route into financial inclusion may create barriers to the further progression of individual customers into the commercial banking system. Our further research, involving interviews with three types of people (farmers, poor urban residents and bank staff) will shed light both on the gendered implications of the Samurdhi scheme for saving and borrowing, and on the connections between the use of liberal discourse and men’s and women’s uptake of opportunities for saving and borrowing in the commercial banking system. Conclusion Two policy implications already stand out from this interim research report. (1) The situation in Sri Lanka exemplifies the international trend toward the growing individuation of loans and savings in developing countries. This can be seen as a modernisation of finance, deepening of bank finance, and increase in financial intermediation as more and more individuals make savings and get access to loans. (Adding women as individual customers potentially doubles the number of customers compared to perceiving households-as-customers.) As a result, banks may wish to enable women to offer social collateral for loans. They may want to encourage women to generate personal savings to act as collateral over the longer term. Meanwhile inheritance rules & property law may need to allow women to take individual control of financial collateral, such as jewellery and land, if women are not to be financially excluded through bank rejection during the modernisation process. This paper has pinpointed self-exclusion and bank rejection as analytical categories useful in empirical research about credit in different cultures. We have noted the empirical question of whether mutual agreed separation of potential customers from banks implies a self-exclusion that is detrimental to economic growth and national integration. More generally, though, (2) government-sponsored national surveys of household expenditure should allow for the possibility of individual responsibility for loans and savings. Surveys should record [either through the household member’s serial number or by taking their basic characteristics directly (age, gender, marital status and occupation)] who is the ‘taker’ of loans and the ‘owner’ of savings accounts. Discourse analysis underpins these recommendations. The authors express doubt over whether full individuation is ever desirable, but they recognise that in multicultural societies the partial individuation of financial services is likely to be occurring and should be recorded in survey data sets. Direct implications follow for micro-finance schemes, which often work on a fully individuated basis whilst some households nevertheless have un-individualised strategies. A side effect of our empirical focus on individuals is the comment that household-level research on the economics of savings may be weak in its analysis of the motives for saving. Individuals have a personal choice of discourses and they have gender, age-group, ethnic, and occupational identities that are, initially, given as structural locations. We pointed to UK and French research that suggested that individual aspirations and attitudes had significant effects on the quantities saved and borrowed, quite apart from household income levels. It remains to be seen whether this is the case in Sri Lanka and other developing countries but, if true, it may lead to the re-working of theories that relate household-level wealth and permanent income to household-level savings and debt aggregates. The theories will need both an individual-level and a household-level explanation, with an integrated approach to how these relate. Finally, the paper has commented on the tension between social integration and social inequality. We pointed out that welfare schemes, as well as attempts to reduce financial exclusion, have an ideological role in encouraging the poor to feel part of society and, while feeling integrated, to behave responsibly as members of society. Specifically in banking this means conforming to expectations in filling in loan application forms and being sure to make repayments on time and keep in communication with the lending agency. These are, one could say, the conditionalities of micro-lending for the poor. In addition some banks insist that poor people save with them, creating a basis in their own experience for judging the reliability of the borrower. One might argue that financial inclusion is conditional on conformity with western banking norms. However in the empirical work poor and female respondents were resistant to, or suspicious of, western norms in some (as yet unexplored) ways. Therefore we posed the frugality hypothesis and the bank-avoidance hypothesis for further empirical research. In each case, one needs to know the subjective self-placement of the potential bank user before one can know whether the person is simply avoiding getting into debt (prudent or frugal behaviour) or is avoiding using the bank itself (due to other arrangements that are preferable). This area of further research offers a rich vein of socio-economic exploration. Appendix 1 Table A1: Details of Logistic Regressions Dependent Variable: Age Age2 Gender Sinhala Poor Well-Off % of Cases Predicted Correctly Western culture problematic .98 1.00 1.34 1.03 2.57*** 1.03 63% “ .98 1.00 1.37 2.57*** 1.03 63% “ 1.35 2.63*** 1.03 62% People should get more loans 1.04 1.00 .81 1.63*** .56** .92 61% “ .80 1.60*** .56** .94 61% “ 1.60*** .55** .92 61% Think guarantor is bad for poor people .98 1.00 1.03 1.02 1.71** .56* 64% “ .98 1.00 1.03 1.70** .56* 65% “ 1.04 1.76** .56* 62% “ 1.76** .56* 62% Think the poor people should give a guarantor 1.01 1.00 .96 .91 .90 2.24*** 61% “ 1.01 1.00 .91 .89 2.22*** 61% “ .91 .88 2.23*** 61% “ .89 2.17*** 61% Note: Each row represents one regression. The table contains the odds ratios from logistic regression equations. Blank spaces indicate that a variable has been dropped because it was not significant. Using the Wald statistic, *** indicates significance < .01, ** signif. < .05, * signif < .10. The path diagram is based on indicating only those associations which are significant at least at the 10% level. Appendix 3: Caveats Regarding the Consumer Finance Survey 1996/97 of Sri Lanka The survey is not based on an all-Sri Lanka sample because of war activity in two Zones and in some locales within the selected Zones. Various other minor caveats can be found in Central Bank of Sri Lanka (1999c). References Albee, A., and N. Gamage (1996) Our Money Our MOvement: Building a Poor People’s Credit Union, London: IT Publications. Archer, M. (2000) Being Human: The Problem of Agency. Cambridge: Cambridge University Press. Archer, M., R. Bhaskar, A. Collier, T. Lawson, and A. Norrie, eds. (1998) Critical Realism: Essential Readings, London and NY: Routledge. Bauman, check date. “Informal Credit: An Aladdin’s Lamp of Rural Society”. Bhaskar, R. (1986) Scientific Realism and Human Emancipation, London: Verso. Bhaskar, R. (1993) Dialectic and the Pulse of Freedom, London: Routledge. Bhaskar, R. (1998) “Facts and Values: Theory and Practice”, chapter 17 in Archer, et al., eds., 1998. Broadfoot, P. (1999) “Assessment and the Emergence of Modern Society”, ch. 5 in Moon, B., and P. Murphy, eds. (1999) Curriculum in Context, London: Sage Caplan, P. (1985) “Women’s Voluntary Social Welfare Work in India: The Cultural Construction of Gender and Class”, Bulletin of Concerned Asian Scholars, 17:1, pp. 20-31. Central Bank memorandom. No date. Samurdhi Bank Societies (Samurdhi Banku Sangham). Study conducted by the Economic Research Department of the Central Bank, with cooperation by the Rural Credit Dept. Central Bank of Sri Lanka, (1999a) Report on Consumer Finances and Socio-Economic Survey, 1996/97, Part II: Statistical Tables, Colombo. Central Bank of Sri Lanka, (1999b) Economic and Social Statistics of Sri Lanka, Colombo. Central Bank of Sri Lanka, (1999c) Report on Consumer Finances and Socio-Economic Survey, 1996/97, Part I, Colombo. CENWOR, Centre for Women’s Research, Women in the Economy: Trends and Policy Issues, CENWOR, 1998. Chant, S. (1994) “Women, Work, and Household Survival Strategies in Mexico, 1982-1992: Past Trends, Current Tendencies and Future Research”, Bulletin of Latin American Studies, **, pp. 200-220. Durkheim, E. (1947) The Division of Labour in Society, New York: Free Press. Foucault, M. (1976) Two Lectures (trans. A. Sheridan), London: Allen Lane. Ghate, P., E. Ballon and V. Manalo. 1996. “Poverty Alleviation and Enterprise Development: The Need for a Differentiated Approach”, Journal of International Development, 8:2, March-April. Special issue: Sustainable Banking With the Poor. Gimenez, M. E. (1999) “For Structure: A Critique of Ontological Individualism”, Alethia: The Newsletter of the International Association for Critical Realism, 2:2, pp. 19-25. Gulli, H. (1998) Microfinance and Poverty: Questioning the Conventional Wisdom, Washington: Inter-American Development Bank. Gunadasa, H.W., P. Jayamanna, and W. Dayaratne, Report of the Internal Review of Social Mobilizer Programme of Moneragala District Integrated Rural Development Programme (MONDEP), July 1993. Gunatilaka, Ramana. (check year**) “Credit-Based, Participatory Poverty Alleviation Strategies in Sri Lanka: What Have We Learned?”, Institute of Policy Studies, Poverty and Income Distribution Series no. 2 (with UNDP support). Gunatilaka, Ramani, and Rozana Salih. (1999) “How Successful is Samurdhi’s Savings and Credit Programme in Reaching the Poor in Sri Lanka?”, Inst. of Policy Studies, November, Research Studies no. 3. Hettige, S. T., Markus Mayer, Premakumara de Silva, and Nishara Fernando, national research team, National Youth Survey: Overview Report on selected topics, Sri Lanka, 2000.UNDP, Friedrich-Ebert-Foundation, Jaffna Rehabilitation Project, and German Cultural Institute, 2000. Hettige, S. T., Nishara Fernando, Anton Piyarathna, Jinali Chandrika, and Anula Rathnayaka, Voter Aspirations in Sri Lanka, in collaboration with the Centre for Anthropological and Social Studies (CASS), Univ. of Colombo, with Friedrich-Ebert-Stiftung (FES), FES, 2000. Hutchinson , F., and Olsen, W.K. (2001). “The Unsustainable Ship Of Capitalist Money Management”, paper presented to the International Sustainable Development Research Conference, University of Manchester, 5-6 April 2001 Hutchinson, F. and Burkitt, B. (1997) The Political Economy of Social Credit and Guild Socialism London and New York. Routledge. Hutchinson, F., and Burkitt, B. (1999) “Towards a New Approach to Income Distribution and Environmental Sustainability”, Sustainable Development, 7, pp. 57-63. Jayawardena, Deshamanya N. U., “Central Banking as a Catalyst of Economic Growth”, Centre for Banking Studies, Occ. Paper no. 31, 1999. Central Bank of Sri Lanka. Jones, B.G. (2001) “Explaining Global Poverty: A Realist Critique of the Orthodox Approach”, chapter forthcoming in J. Cruickshank, ed., Critical Realism: The Difference It Makes, London: Routledge. Kaur, S.P., S. Lingayah, E. Mayo (1996) “Financial Exclusion in London”, London: New Economics Foundation. Kearney, R. P. ** Kemper, 2001. forthcoming book on Sri Lankan advertising and its impact on discourse. Lawson, T. 1997. Economics and Reality. London: Routledge. Lindqvist, A. (1981) “A Note on Determinants of Household Saving Behaviour”, Journal of Economic Psychology, 1, pp. 39-57. Lunt, P.K., and S. M. Livingstone (1992) Mass Consumption and Personal Identity: A Study of Everyday Economic Experience, Ballmoor: Open University Press. Lunt, P. K. and S. M. Livingstone, (1991). “Psychological, Social and Economic Determinants of Saving: Comparing Recurrent and Total Savings”, Journal of Economic Psychology, 12, pp. 621-641. Lunt, P.K. (1992). “Discourses of Saving”, Journal of Economic Psychology, 13, pp. *. Mavrotas, G., and R. Kelly (1999). “Financial Sector Development And Savings Mobilisation: An Assessment”, Working Paper no. 10, Institute for Development Policy and Management, University of Manchester. Ministry of Samurdhi, Youth Affairs and Sports, Progress 1998 and Programmes for 1999. Novak, T. (1988) Poverty and the State: An Historical Sociology, Milton Keynes: Open University Press. Olsen, W. K. (2000) stereotypes paper; see EAEPE site www.tuwien.ac.at/eaepe for details. Olsen, W.K., and C. Kirkpatrick (2001a), “Positions and Representations of Sri Lankan Poor People”, draft mimeo. Olsen, W.K., and C. Kirkpatrick (2001b), “Credit, Savings, and Poverty in Sri Lanka in the 1990s: A Multi-Level Statistical Analysis of Social Exclusion from Banking”, draft mimeo. Olsen, Wendy Kay 2000. The Limits to Conditionality: Grassroots Evidence from Rural India, WorldView Publications, Oxford, forthcoming. Otero, M., and Rhyne, E., eds. (1992) The New World of Microenterprise Finance: Building Health Financial Institutions for the Poor, London: IT Publications. PAS Business Surveys (1988) Consumers' Use of Credit Survey, London, OFT. Poverty Alleviation Project, IBRD, Framework for Poverty Reduction in Sri Lanka (Discussion paper), Sept. 1999. Risseeuw, C. (1991) Gender Transformation, Power and Resistance among Women in Sri Lanka: The Fish Don’t Talk About the Water, Delhi: Manohar. Rogaly, B. (1996) “Micro-Finance Evangelism, Destitute Women and the Hard Selling of a New Anti-Poverty Formula”, Development in Practice, 6:2, pp. 100-113. Sabaragamuwa Development Bank [Ratnapura and Kegalle Districts], Annual Report 1999. Sayer, A. (2000) Realism and Social Sciences, London: Routledge. Sayer, A. (1992) Method in Social Science: A Realist Approach, London: Routledge. Sayer, A. (1995) Radical Political Economy: A Critique, Oxford: Blackwell. Schmidt-Hebbel, K., S. B. Webb, and G. Corsetti (1992) “Household Saving in Developing Countries: First Cross-Country Evidence”, World Bank Economic Review, 6:3, pp. 529-547. Sharma, K. L. Society and Polity in Modern Sri Lanka. Delhi: South Asian Publ. Ltd, 1988. Van Veldhoven, G. M., and E. A. G. Groenland (1993) “Exploring Saving Behavoiur: A Framework and a Research Agenda”, Journal of Economic Psychology, 14, pp. 507-522. Viaud, J., and C. Roland-Lévy (2000) “A Positional and Representational Analysis of Consumption: Households When Facing Debt and Credit”, Journal of Economic Psychology, 21, pp. 411-432. Wärneryd, K. (1989) “On the Psychology of Saving: An Essay on Economic Behaviour”, Journal of Economic Psychology, 10, pp. 515-541.