ISSN (Online) - 2349-8846
-A A +A

Involuntary Exclusion and the Formal Financial Sector

 Ramesh Golait (rgolait@rbi.org.in) and Monika Sethi are assistant advisers at the Department of Economic and Policy Research, RBI, while Shobhit Goel is a student of economics at the IIT, Kharagpur.

Financial inclusion is a policy priority in India, with the focus on the supply-side of the financial inclusion drive and programmes such as the Pradhan Mantri Jan-Dhan Yojana. Insufficient attention, however, has been paid to the use of banking services by people at the bottom of the pyramid in order to understand what constrains them from using the formal financial services on offer. This study looks at the causes of involuntary exclusion from formal financial services in the slums of Delhi.

The views presented here are personal and the usual disclaimers apply. We thank the anonymous referee for comments. Gopinath Tulasi (tgopinath9@live.com) is a member of faculty, Reserve Bank Staff College, Chennai.

A more focused and structured approach to financial inclusion has been followed in India since 2005, when the Reserve Bank of India (RBI) decided to promote financial inclusion by implementing various policies from the supply-side (Joshi 2014).1 However, determinants of demand for financial services, especially in emerging market economies (Cole et al 2009), are not sufficiently understood. This is, therefore, the ideal juncture to study demand-side barriers to the use of financial services on offer, especially among people at the bottom of the pyramid in India. This article tries to address demand-side constraints on financial inclusion in Delhi slums.2 We have two objectives: to understand the constraints causing involuntary exclusion in India, and to estimate the incremental increase in the probability of availing banking services3 if these constraints are addressed.

Against this backdrop, this article attempts, inter alia, to gauge usage of financial services in identified slums of the Delhi–National Capital Region (NCR)4 by conducting a primary survey and using the survey data for an econometric exercise to identify demand constraints and draw policy inferences, if any. The article is organised in five sections. Section 1 briefly covers the logistics and summary results of the survey. Section 2 describes the methodology for estimating determinants of demand for financial services. Section 3 presents the results of the econometric exercise. Section 4 contains broad policy inferences, and Section 5 presents the conclusion.

1 Survey Results

A questionnaire was designed for the survey, its primary motivation being to compile and measure core and sub-core indicators of financial inclusion from the demand-side as formulated by the World Bank.5 An attempt was made to go beyond measuring these indicators and collect relevant data on financial behaviour and the level of financial literacy, focusing on awareness about know your customer (KYC) norms, the concept of inflation, online banking and mobile banking, and the ability to do simple arithmetic and calculate simple interest and compound interest. The questionnaire thus designed was administered in the sampled slums in Delhi. Census 2011 data on the number of slum households in Delhi availing banking services were studied, and slum locations with poorer use of banking facilities were selected as the sample. Thus, 25 slum locations spread across almost all nine districts6 of Delhi were located, covering about 600 slum dwellers (respondents). Face-to-face interviews were conducted with these respondents.7 The data thus collected were analysed, and the salient features are presented below in three subsections.

1.1 Socio-civic and economic characteristics: Most of the respondents reported lack of education, shortage of basic civic amenities and low economic status (Table 1, p 68). Sixty-six percent of the respondents were migrants. More than two-thirds of the respondents reported that their children do go to school. Regarding proof of identity, almost 93% of the respondents reported holding Aadhaar cards, but only 39% reported that the card was bank-linked. Respondents do seem to have access to the formal banking system (Table 1).

 

1.2 Core and sub-core indicators: The core and sub-core indicators of the World Bank’s Global Findex, compiled from the survey of Delhi slums, are presented in Annexure 2 (p 72). Quite a high percentage of slum households had a bank account, but operations in these accounts were limited. Delhi’s per capita income is thrice the all-India average, but the survey reveals that slum dwellers in Delhi have meagre incomes with low savings, pointing to the highly skewed income distribution. The study revealed a low propensity to save and borrow, and an element of indifference to formal and informal financial institutions among the poor in Delhi slums. Insurance coverage is significantly low and the majority cited high costs as the reason for not having insurance.

The above indicators for Delhi are compared with those for all-India and the world, drawn from the World Bank’s Little Data Book on Financial Inclusion, 2015 (Figure 1).

(i) Ownership of a bank account is highest in Delhi but formal borrowing is lowest in Delhi, and formal saving is lower than the world average but slightly higher than the all-India average.

(ii) The use of informal channels for savings is higher in Delhi than the all-India average.

(iii) The use of formal accounts to receive wages or government payment in Delhi is almost double the all-India average, but lower than the world average.

1.3 Financial literacy indicators: Financial literacy is a critical aspect of financial inclusion from the demand-side. The survey results indicate that financial literacy and knowledge among the slum households of Delhi is abysmal. Though they do not seem to understand the formal concept of inflation and the role of the RBI therein, they are aware of the impact of inflation on their lives (76%). Hardly any respondent was able to calculate compound interest, not even the graduates. More than 90% of slum households were unaware of KYC norms and their recent simplification. However, the majority of respondents were able to compute simple arithmetic (68%) and had knowledge of simple interest (72%). Interestingly, about half the respondents reported being aware of online banking and mobile banking, and 79% reported being aware of insurance.

A gender divide in financial literacy was observed. In each of the financial literacy indicators computed, illiteracy was more pronounced among females than among males (Figure 2).

2 Determinants of Demand for Financial Services

In the literature, two leading theoretical underpinnings explain low demand for formal financial services. First, formal financial services entail relatively larger fixed transaction costs, and as a result, people who are excluded cannot afford these formal financial services at market prices (Beck et al 2007). Second, individual and household characteristics are important barriers to demand for formal financial services. These individual or household characteristics cause involuntary exclusion or voluntary (self) exclusion. The characteristics causing involuntary exclusion include income, education, gender, migration status and financial literacy. Characteristics causing self-exclusion are lack of interest, no need for formal financial services, mistrust, fear, refusal to be in debt and preference for informal financial services.

This article is premised on the second theoretical underpinning for the following reason: fixed transaction costs associated with provision of financial services are set to decrease dramatically in India, thanks to technological breakthroughs and institutional innovation in the form of payment banks and small finance banks. Therefore, the first theoretical underpinning is less likely to be of relevance in India. The second theoretical view is tested through a primary survey and an econometric exercise involving data thus collected.

2.1 Methodology: Since the data collected from the primary survey is binary in nature, in keeping with standard methodology, the general linear models (GLiM)8 approach is used. Using GLiM, a separate model is estimated for each of the banking services, namely, ownership of a bank account, formal savings, formal borrowings, formal remittances and insurance. For each of these models, there are seven independent/predictor variables, namely, income, gender, education, migration, occupation, civic amenities index and financial literacy index. Theoretically, income, education, and financial literacy are expected to influence demand for banking services positively. With regard to gender, literature expects that males demand more of banking services than females. Literature and theory are ambiguous about the influence of occupation. Civic amenities, on the other hand, are supposed to influence demand for these banking services positively. Finally, migrants are expected to demand relatively more of remittance services. In addition, a composite index of usage of banking services is regressed on the above seven determinants of demand for such banking services to study demand barriers at the aggregate level.

Accordingly, the model estimated separately for the five binary dependent variables (i) is as follows:

where,

x1: Income; x2: Gender; x3: Migration; x4: Education;

x5: Dummy variables for occupation;9

x6: Civic amenities index;10

x7: Financial literacy index;11

pi: Probability of positive response for the dependent variable i (ownership of bank account, formal savings, borrowing, insurance, formal remittances).

Further, the choice between logit and probit link functions is made, based on model fit criteria, including Akaike information criterion (AIC), Bayesian information criterion (BIC) and chi-square. The probit model performed marginally better as compared to the logit model. Therefore, in this article, the probit link function has been used for the five banking services model. For the Financial Services Composite Index,12 the identity link function of the Gaussian family has been used, as the dependent variable is a composite index that assumes value between 0 and 1. This link function is the GLiM equivalent of linear regression model. A maximum-likelihood estimation is conducted as a series of GLiMs with probit as the link. Further, marginal effects are calculated from the point coefficients estimated in the models.

3 Results

The marginal effects of the independent variables for each of the banking services are presented in Table 2. The results are robust and indicate that there are factors inhibiting demand for banking services. The factors influencing probability of demand for various banking services and the extent of such influence are analysed below.

3.1 Owning a bank account: Income, education, and financial literacy positively influence the probability of owning a bank account. The positive influence of income and financial literacy is statistically significant, while that of education, civic amenities and migration status is statistically insignificant. In particular, moving to a higher bracket of income (illustratively, from an income bracket of less than ₹5,000 a month to an income bracket of ₹5,000 to ₹10,000) increases the probability of owning a bank account by 4.3%. Further, enhancing financial literacy improves the probability of owning a bank account as the marginal effect is estimated at 0.5, which corresponds to an average 0.5% increase in the probability of owning a bank account for a 1% increase in financial literacy index score. The probability of ownership of bank account is higher among females than among males, though this finding is not highly statistically significant. The effect of occupation, on the whole, is not statistically significant, which is confirmed by the step-wise model determination method.

3.2 Formal savings: Income, education, civic amenities and financial literacy positively influence the probability of formal savings with a bank/financial institution, though such an influence is statistically insignificant in the case of income. This could be attributed to the fact that over 75% of the respondents belonged to the meagre income bracket of below ₹10,000 a month, with hardly any worthwhile savings. Improving financial literacy, on the other hand, increases the probability of engaging in formal savings as the marginal effect of financial literacy on formal savings is statistically significant and on average a 1% increase in financial literacy index score leads to an increase in the probability of undertaking formal savings by 0.32%. Similarly, up-scaling education augments the probability of formal savings by 2%. Augmenting civic amenities is also estimated to improve the probability of promoting formal savings, and a 1% increase in the civic amenities index score increases the probability of formal savings by 0.12%. Interestingly, females displayed a higher, though statistically insignificant, probability of engaging in formal savings than males. As regards occupation, a petty trader has a 10% higher relative probability of engaging in formal savings than a transport worker, and a 5% higher probability than a casual labourer. The probability of engaging in formal saving is neutral to migration status.

3.3 Formal borrowings: Income positively influences the probability of borrowings from banks/financial institutions. Illustratively, migrating from the income bracket of below ₹5,000 to a bracket of ₹5,000 to ₹10,000 increases the probability of borrowings from banks by 3.8%. Interestingly, females have a higher probability of borrowing (by 5%) than males. The probability of borrowing among non-migrants is higher (by 3.9%) than among migrants. A plausible explanation for this is the fact that non-migrants are expected to possess better documents regarding proof of residency, and can provide local references and thus are more likely to be accepted for a loan application. All these findings are statistically significant. Similarly, the marginal effect of the financial literacy index on the probability of borrowings from banks and financial institutions is estimated to be positive (0.09) though it is not highly significant. As regards occupation, typically, the petty trader and vendor has a higher relative probability of borrowing from banks than any other occupation considered in the survey, barring the transport worker, though these findings are statistically insignificant.

3.4 Formal remittances: Income, education and financial literacy positively influence the probability of availing of formal remittance services, though the impact of income and financial literacy is statistically insignificant. Improving education, in particular, is estimated to increase the probability of availing of formal remittance services by 12.6%. The probability of availing of formal remittance services among migrants is higher by 12.6% than among non-migrants. Further, males have a higher probability (by 5.8%) than females of using formal channels of remittance. The civic amenities index has a negative marginal effect on the probability of using formal remittance services (– 0.159) and is statistically significant, which implies that a 1% decrease in the civic amenities index, other things being equal,leads on average to an increase in the probability of availing of formal remittance services by 0.159%. Though this may seem counter-intuitive at first, it could perhaps be rationalised that a higher score on the civic amenities index is an indicator that the respondent is well-settled and therefore less likely to have family members living outside Delhi. As regards the influence of occupation on demand for formal remittance services, the results are broadly ambiguous, though petty traders seem to have a higher probability of availing of remittance facilities from banks than any other profession considered in the survey, barring transport workers.

3.6 Financial services index: The financial services index denotes composite use of the different financial services under study. The range of the index is from 0 to 1, with 0 signifying no services availed and movement towards 1 underscoring greater use of financial services. The results indicate that income, education, financial literacy and civic amenities positively influence the financial services index, though the impact of civic amenities is not statistically significant. In particular, moving from an income bracket of below ₹5,000 to the income bracket of ₹5,000 to ₹10,000 leads to an improvement in the financial services index by 2.9%. Similarly, financial literacy positively impacts the index, with a positive coefficient of 0.23, which is highly significant. In other words, an improvement in financial literacy index score by 1% leads to improvement in financial services index score by 0.23%. Education also has a positive impact, as reflected by the positive coefficient of 0.011, which, though not highly statistically significant, leads to an increase in the financial services index score by 1.1% for each successive higher level of education (Table 3).

As regards type of occupation, a petty trader has a statistically significant higher score by 4.4% and by 6.2% on the financial services index as compared to a worker in services and casual labour, respectively. Petty traders score higher on the index vis-à-vis transport workers by 3.3% though this difference is not highly statistically significant. With regard to gender, all other things being equal, females score higher on the index than males by 1.1%, though it is statistically insignificant. The probability of demand for various banking services seems to be neutral to migration status and the role of civic amenities in influencing these probabilities remained ambiguous.

4 Policy Inference

In view of the evidence of demand-side constraints that are limiting the use of financial services among people at the bottom of the economic pyramid, as evidenced by the results presented above, the article offers the following policy recommendations:

(i) Increased employment opportunities, which would provide better earning avenues to people, will help promote demand for financial services such as insurance, borrowing and ownership of an account. Identification of skill-gaps and decentralised skill-building initiatives in terms of vocational education aimed at augmenting employability of the excluded and thereby generating potential employment among them needs policy priority. The National Mission for Skill Development, which aims at consolidating skill-building initiatives across several ministries, is of relevance in this context and hence is a step in the right direction.

(ii) Further, the survey results indicate that respondents belonging to middle (₹5,000 to ₹10,000 per month) and upper-income (₹10,000 and above) categories are identified with higher education levels. Thus, increasing levels of general education among Delhi slum dwellers would augment their income-earning potential.

(iii) From a policy perspective, female-centric financial inclusion drives are not what is required, as is typically perceived, but a general empowerment of women in the integral sense of the term, with a focus on reducing the gender gap in their education, financial literacy and income.

(iv) There is a need to shift from one-size-fits-all schemes to tailored and targeted schemes developed at subregional levels as this could lead to greater financial inclusion. This view is supported by the finding that types of occupation and migration status have a differential impact on demand for certain financial services.

(v) There is a need to focus on increasing the awareness of people at the bottom of the pyramid regarding the financial products and services being offered and regarding the norms and regulations governing them.

5 Conclusions

Upscaling financial inclusion has been a policy priority in India. Accordingly, much ground has been covered with regard to calibrating the supply-side of the financial inclusion drive, especially in the form of RBI’s financial inclusion plan and Government of India’s Pradhan Mantri Jan-Dhan Yojana (PMJDY). Given its policy relevance, however, there is a need to assess use of banking services among the people, especially at the bottom of the pyramid, to be able to understand the constraints that prevent them from using the formal banking/financial services on offer. This article attempts to address this need, apart from compiling core and sub-core indicators of financial inclusion from the demand-side for Delhi slums, on the lines of the World Bank’s Global Findex,thereby adding a regional dimension to the relevant database. The salient but qualified finding is that slum households in Delhi have a low propensity to save and borrow, and are indifferent to formal and informal institutions. The article finds econometric evidence that there are constraints—in terms of income and occupation—that are holding Delhi slum dwellers back from using banking services. In order to enable effective financial inclusion, we recommend the removal of constraints on demand for formal financial services by enabling the lower strata of society to earn more. We also recommend customising financial products for identified occupations, improving financial literacy and being agnostic about gender.

Notes

2 Fundamentally, there are two aspects to demand-side constraints, namely involuntary exclusion and voluntary/self-exclusion. However, the ambit of this article is confined to involuntary exclusion.

3 Financial inclusion does not only signify owning a bank account or availing credit from banks. It is more composite, involving owning a bank account, engaging in formal savings, availing credit, remittance and insurance. Hence, all these banking/financial services are part of this study.

4 The use of banking services by slum households is relatively low and the share of Delhi population residing in slums is as high as 10.6%, according to Census 2011.

5 In recognition of the need for better data to support the financial inclusion agenda, the World Bank’s Development Research Group has built the Global Financial Inclusion Index. The Global Findex, as it is known, is the first internationally comparable public database of demand-side indicators. This database documents financial usage across gender, age, education, geographical regions and national income levels across countries and over time. The World Bank has tied up with Gallup Inc to interview at least 1,000 people per country (CAFRAL 2012).

6 North West Delhi, South Delhi, West Delhi, South West Delhi, North East Delhi, East Delhi, North Delhi, Central Delhi, New Delhi.

7 The RBI, New Delhi office, provided the resources, including transport and security, needed for the conduct of the survey. We thank Deepak Singhal, then Regional Director, RBI, New Delhi, for these resources. The services of the summer 2015 interns posted to the Department of Economic and Policy Research (DEPR), New Delhi office, were deployed in canvassing the questionnaire in identified slum locations in Delhi. We thank Roshni Garg, Kashish Verma, Archita Misra, Sadaf Beigh and Jaskirat Singh Sidhu.

8 Any standard econometric text could be consulted for a theoretical background of GLiM models.

9 The base (reference) occupation selected for the econometric exercise to avoid the dummy variable trap is petty trader and vendor.

10 See Annexure 1 for details.

11 See Annexure 1 for details.

12 See Annexure 1 for details.

References

CAFRAL (2012): “Workshop on Measuring Financial Inclusion from the Demand Side: A Background Paper,” Centre for Advanced Financial Research and Learning, Mumbai.

Cole, S, T Sampson and B Zia (2009): “Prices or Knowledge? What Drives Demand for Financial Services in Emerging Markets?”Working Paper 09–117, Harvard Business School.

Joshi, P D (2014): “Financial Intermediation for All: Economic Growth and Equity,” talk delivered at Dun & Bradstreet Financial Inclusion Conclave, Mumbai, 26 August.

OECD (2008): “Handbook on Constructing Composite Indicators: Methodology and User Guide,” Organisation for Economic Cooperation and Development.

Annexure 1: Methodology for Construction of Indices

The composite indices for civic amenities, financial literacy and financial services are compiled in accordance with the methodology prescribed in the OECD Handbook on Constructing Composite Indicators (OECD 2008). The civic amenities index is constructed with an aim to capture respondents’ access to various civic amenities. It takes into account the size and construction type of dwellings, installation of a water connection, sanitation facilities, cooking gas connection and ownership of any land. The financial literacy index aims to capture respondents’ awareness of available financial products and services and their knowledge of basic financial concepts of inflation, time value of money, equitable distribution and time consistency. The financial services index has been constructed with the aim of capturing overall use of various financial services, including ownership of a bank account, formal savings, formal remittances, formal borrowing and insurance.

Following multivariate analysis of the data, a modified version of the budget allocation process is used for weight allocation. This method is a computation and time-efficient technique suited to the construction of an index to capture specific parameters. As we have ordinal parameters, the weight for a parameter also gives the maximum possible score for that parameter. Therefore, the sum of the weights allocated for all parameters gives the maximum possible total score that is possible.

As the raw survey data suffered from missing response values for the different parameters for several reasons, it was necessary to account for this while calculating the index value. Not accounting for the missing response and computing the index out of the same possible total score is equivalent to considering missing response as equal as the lowest score option. This would have created a downward bias in the sample as the index score would have been artificially low for the respondents with missing responses. To avoid this, each respondent was evaluated only on the questions actually responded to and thus a variable base was to scale the score. Further, respondents who had answered too few questions were not allocated a score. For each index, the score was normalised to ensure that the range lay from 0 to 1.

 

Updated On : 13th Sep, 2017

Comments

(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top