Assessment of PMMY Report

Research insights on design and implementation
Version: Final Version, April 2018

This report was prepared by Dvara Research in collaboration with IFMR LEAD
Corresponding author: Vaishnavi Prathap, Dvara Research
If you have any queries on the report please reach out to: communications.research [at]
dvara.com
2
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ............................................................................................................................................ 4
ABOUT THE AUTHORS .............................................................................................................................................. 4
LIST OF TABLES ........................................................................................................................................................... 5
LIST OF FIGURES ......................................................................................................................................................... 6
LIST OF ACRONYMS.................................................................................................................................................... 7
1. INTRODUCTION ...................................................................................................................................... 8
1.1 Policy design and theory of change ........................................................................................................... 11

1.2 Study objective and research design ......................................................................................................... 16
2. ANALYSIS................................................................................................................................................. 18
2.1 Macroeconomic trends in MSME lending ................................................................................................ 19

2.2 Characteristics of MLIs.................................................................................................................................... 23
2.2.1 Market share and growth trends ........................................................................................................ 23
2.2.2 Product range and underwriting rules ............................................................................................ 32
2.3 Features and uses of PMMY loans .............................................................................................................. 36

2.3.1 Insights on loan features from MUDRA data ................................................................................. 36
2.3.2 Insights on loan features and uses from survey of MSME borrowers ................................ 43
2.4 Lender specialization and market fragmentation ............................................................................... 54

2.4.1 Evidence of specialization from MUDRA data............................................................................... 54
2.4.2 Implications for origination and market structure..................................................................... 68
2.5 Regional patterns in the PMMY portfolio ................................................................................................ 71

2.5.1 Concentration in Tier 1 and Tier 2 centres .................................................................................... 73
2.5.2 Regional drivers of credit concentration ........................................................................................ 76
3. RECOMMENDATIONS .......................................................................................................................... 84
3.1 Framework for evaluation ............................................................................................................................. 84

3.2 Guidelines for high-quality origination .................................................................................................... 86
3.3 Suggestions for additional MIS formats ................................................................................................... 88
3.4 Proposals for further research .................................................................................................................... 89
ACKNOWLEDGEMENTS
This research was made possible by the generous contributions of several institutions and
individuals. We appreciate the timely guidance received from Mr. Pankaj Jain, Joint Secretary,
Department of Financial Services and Mr. Jiji Mammen, CEO, MUDRA Ltd., and thank the entire
team at DFS and MUDRA for their extraordinary support during this study. We acknowledge
critical inputs received from Mr. Aniruddha Bagchi, DGM, NCGTC, and the officers of MOSPI Data
Centre and MSME-DI in Kolkata.
We gratefully acknowledge the patient and enthusiastic participation of survey respondents—

including enterprise owners, frontline loan officers and branch managers of financial institutions.
We sincerely thank our colleagues at IFMR Trust and IFMR LEAD for critical insights and feedback
during this study.
ABOUT THE AUTHORS

Dvara Research (formerly known as IFMR Finance Foundation) is an
Indian financial systems policy research institution promoted by Dvara
Trust. Our mission is to ensure that every individual and every enterprise
has complete access to financial services. We strongly believe in the power
of finance to unlock transformative opportunities for individuals, households, enterprises and
local governments. Dvara Research works on the themes of (a) developing and disseminating
financial systems design that achieve both inclusion as well as stability, (b) advancing customer
protection particularly for low-income households, and (c) building household finance as an
academic discipline to enrich our understanding of household financial behaviour through
rigorous analysis. We engage in extensive policy and regulatory research on these themes.
IFMR LEAD is a non-profit research organization conducting

high-quality scalable action research and outreach in
development economics and finance. Our vision is to foster
improved financial access, better legal and physical infrastructure, and targeted social services
for individuals, households, and enterprises, to help them attain their socioeconomic aspirations.
The Financial Inclusion vertical at IFMR LEAD partners with financial service providers,
technology providers and policymakers to design and test products and services that can
promote greater access to, uptake and utilization of financial services among low-income
households and other vulnerable sections of society.
IFMR LEAD undertook both the primary data collection components of this research and contributed
to the analysis of surveys and qualitative interviews featured in this report.
4
LIST OF TABLES
Table 1 Market share of various MLI types in PMMY portfolio .................................................................. 28
Table 2 Product range of MLIs interviewed ....................................................................................................... 34
Table 3 Key characteristics of PMMY loans from MUDRA data.................................................................. 39
Table 4 Sample composition for survey of MSME borrowers..................................................................... 44
Table 5 Key characteristics of loans from primary survey .......................................................................... 45
Table 6 Characteristics of borrowers’ enterprises, from primary survey ............................................. 48
Table 7 Characteristics of borrowers’ households, from primary survey ............................................. 49
Table 8 Customer experience with loan processing and service ............................................................... 52
Table 9 Composition of MLIs’ portfolios by loan size..................................................................................... 57
Table 10 YOY changes in MLIs’ portfolio composition ................................................................................. 63
Table 11 Tier 1 and Tier 2 credit concentration by MLI type ..................................................................... 74
Table 12 Description of variables featured in regional analysis................................................................ 76
Table 13 Correlation between measures of regional distribution in PMMY with regional

characteristics ................................................................................................................................................................. 79

characteristics, by MLI type ...................................................................................................................................... 82
LIST OF FIGURES
Figure 1 MSMEs perceptions of key obstacles to their growth .................................................................. 10
Figure 2 Theory of change for PMMY .................................................................................................................... 12
Figure 3 Number of MSME loans upto Rs. 10 lakhs by SCBs upto Rs. 10 lakhs ................................... 20
Figure 4 Total outstanding on MSME loans upto Rs. 10 lakhs by SCBs .................................................. 20
Figure 5 Trends in GLP and Off-B/S portfolio of NBFC-MFIs ...................................................................... 21
Figure 6 Trends in Off-B/S origination of all microfinance institutions ................................................. 21
Figure 7 Share in PMMY portfolio by MLI type (Accounts) ......................................................................... 24
Figure 8 Share in PMMY portfolio by MLI type (Disbursed amount) ...................................................... 24
Figure 9 YOY change in aggregate PMMY portfolio by MLI type (Accounts, FY17) ........................... 26
Figure 10 YOY change in aggregate PMMY portfolio by MLI type (Disbursed amount, FY17) ..... 26
Figure 11 Multiple uses of loans by borrowers ................................................................................................ 46
Figure 12 Median expenses in average, high and low months ................................................................... 51
Figure 13 Median revenues in average, high and low months ................................................................... 51
Figure 14 Profitability of sample enterprises .................................................................................................... 52
Figure 15 Composition of aggregate portfolios by loan size and MLI type ........................................... 55
Figure 16 Steps in loan processing and designated work allocation, by MLI type............................. 70
Figure 17 Regional disparities in access to PMMY loans .............................................................................. 72
Figure 18 Tier 1 and Tier 2 regions’ share of loans vs. share of MSMEs ................................................ 73
Figure 19 Example of high rank correlation ...................................................................................................... 77
Figure 20 Example of low rank correlation ........................................................................................................ 78
6
LIST OF ACRONYMS
BC Banking Correspondent
CGFMU Credit Guarantee Fund for Micro Units
CGTMSE Credit Guarantee Fund Trust for Micro and Small Units
CIBIL Credit Information Bureau (India) Limited
GLP Gross Lending Portfolio
GVA Gross Value Added
JLG Joint Liability Group
KYC Know Your Customer guidelines
MCLR Marginal Cost of Funds Based Lending Rate
MFI Micro Finance Institution
MFIN Micro Finance Institutions Network
MLI Member Lender Institution
MSME Micro, Small and Medium Enterprises
MUDRA Micro Units Development & Refinance Agency Ltd.
NBFC Non-Banking Financial Company
NBFC-MFI Non-Banking Financial Company - Microfinance Institution
NSSO National Sample Survey Office
PLR Prime Lending Rate
PMJDY Pradhan Mantri Jan Dhan Yojana
PMJDY-OD Pradhan Mantri Jan Dhan Yojana – Overdraft facility
PMMY Pradhan Mantri Mudra Yojana
PSL Priority Sector Lending
RBI Reserve Bank of India
RIDF Rural Infrastructure Development Fund
RRB Regional Rural Bank
SCB Scheduled Commercial Banks
SFB Small Finance Bank
SHG Self-Help Group
1. INTRODUCTION
Micro, Small and Medium enterprises have played a critical role in India’s economic transition
away from agriculture and allied activities—which contributed nearly 51% of GDP in 1950-51—
towards high growth in the non-farm sector. By 2013-14, the former accounted for only 18% of
GDP, while service activities grew rapidly to make the dominant contribution (57%), followed by
industry and manufacturing (25%) 1 . The NSSO 73rd round Survey of Unincorporated Non-
Agricultural Enterprises (excluding construction) in 2015-16 estimated nearly 6.34 crore
enterprises employing more than 11.13 crore workers in India. We find that 36.3% of these
enterprises were engaged in either wholesale or retail trade, 32.6% in a variety of services and
31% in manufacturing activities.
An overwhelming majority of unincorporated enterprises were small and informal—96% of them

were sole proprietorships, and were usually located within or just outside the owner’s household
premises. Only 31% were registered with any industry and trade association or development
board, indicating the informal and unorganized nature of their operations. Own-account
enterprises (comprising 84% of all enterprises) are those run by a single household member
without paid workers and are often only one of multiple income sources for low-income
households. The mean Gross Value Added of own-account enterprises was Rs. 7,980 per month
while the mean GVA of firms classified as establishments was Rs. 53,425 per month2.
The MSME sector is therefore strongly linked to both macroeconomic progress as well as
household financial well-being. At the level of the macro economy, this sector has long been
regarded as the “nursery of entrepreneurship”, featuring a heterogeneous set of business models,
practices and technologies. MSMEs are also viewed as playing a critical role in industry value
chains and as a source of employment opportunities for both entrepreneurs and their hired
workers 3 . At the level of individual households, the expansion in non-farm employment
opportunities is believed to have increased the average income of rural households and helped to
mitigate the income risk of farm-based households through diversification4. Further, Foster and
Rosenzweig (2004) have argued that the benefits from non-farm growth accrue more to poor
1
Sector-wise contribution of GDP of India. (2017, March). Retrieved from http://statisticstimes.com/economy/sectorwise-gdp-
contribution-of-india.php
2
National Sample Survey Office (NSSO). (2017). Key Indicators of Unincorporated Non-Agricultural Enterprises (Excluding
Construction) in India (NSS 73rd Round).
3
Government of India. (2010). Report of Prime Minister’s Task Force on Micro, Small and Medium Enterprises.
4
Morduch, J. (1995). Income smoothing and consumption smoothing. The journal of economic perspectives, 9(3), 103-114.
8
households5 (in contrast to the benefits from agricultural growth, which accrue more to landed
households).
There is limited data on the state of access to finance for MSMEs in India. The Economic Census
2013-14 reports only 11.8 lakh enterprises (2% of all enterprises enumerated) with institutional
loans as a major source of finance and a further 3.12 lakhs (0.53%) with SHG loans as their major
source. The NSSO 67th round Survey of Unincorporated Non-Agricultural Enterprises estimated
that, in 2010-2011, only 4.5% enterprises had outstanding formal loans, including loans from
central and state government lending institutions. Yet another 4.5% of enterprises had at least
one loan outstanding to an informal source but none from a formal source, suggesting that the
overall use of credit facilities and MSME sector credit depth is likely to be quite low.
It is important to note here that these low estimates have persisted despite several initiatives to
fund MSMEs. The most notable of these are the inclusion of MSMEs under Priority Sector Lending
norms, with an explicit sub-target6 and the CGTMSE (credit guarantee) facility, both of which are
aimed at lenders, and seek to expand the supply of credit to under-served enterprise segments.
PMMY is similar, in the sense that the scheme is designed to address and relieve supply-side
constraints and accelerate the flow of credit. In such a design, it is perhaps implicitly assumed
that a large segment of MSMEs are under-served and in fact, not only would they be able to
productively deploy additional borrowed funds, they also hold enough equity to maintain healthy
leverage during the tenure of loans.
However, the Indian MSME sector faces several constraints to growth—beyond access to credit—
and the success of any credit-based intervention is critically dependent on progress on other
fronts as well. These include a lack of access to markets and value chains, the unmet demand for
better infrastructure, difficulties in managing both skilled and unskilled workforce, technology or
environmental constraints and finally, barriers to accessing regulatory facilitation7. Surveys of
MSMEs—inquiring on the primary challenges they face—reveals that delayed or inadequate
credit ranked below larger concerns for market linkages and infrastructure but nevertheless, was
the primary constraint for nearly 20% of MSMEs (Figure 1).
5
Foster, A. D., & Rosenzweig, M. R. (2004). Agricultural productivity growth, rural economic diversity, and economic reforms:
India, 1970–2000. Economic Development and Cultural Change, 52(3), 509-542.
6
A comparison of MSME-PSL portfolio and PMMY portfolio size of reporting SCBs reveals that lending under PMMY (i.e.
unsecured loans under Rs. 10 lakhs) constitutes only 7.5% of their corresponding MSME-PSL lending. This suggests very little
overlap between the portfolios and further that without intervention, PSL alone is unlikely to adequately incentivize lending to
micro enterprises.
7
Grant Thornton. (2011). Vision 2020:Implications for MSMEs. Federation of Indian Chambers of Commerce and Industry
(FICCI). New Delhi, India and Reserve Bank of India. (2008). Report of the Working Group on Rehabilitation of Sick SMEs
Figure 1 MSMEs perceptions of key obstacles to their growth
Inadequate market linkages
Lack of infrastructure
Delayed/inadequate finance
Lack of skills
Obsolete technology
Delay in govt. clearances
0% 20% 40% 60% 80%

Source: RBI Report of Working Group on rehabilitation of Sick MSMEs, 2008
While PMMY does not (and cannot) address all of these constraints, several other initiatives do,
and PMMY is located within a larger agenda for reform to unlock economic growth. First, this
fiscal year saw the introduction of the Goods and Services Tax regime, to replace layers of varying
indirect taxation with a unified all-India tax structure. This move is expected to increase resource
mobility and expand market access for producers. Second, the identity verification and basic
payments infrastructure enabled by the JAM trinity (PMJDY accounts and Aadhaar number linked
to mobile numbers) create an information highway and payments highway that can be leveraged
to improve the outreach and quality of financial services. A third and related financial sector
development is the trend of disintermediation in financial services and formation of partnerships
between say, origination specialists and large financial institutions acting as risk aggregators. An
encouraging outcome of these partnerships could be the emergence of agile, specialized lenders
who are well-placed to close market gaps in terms of outreach, underwriting or product
innovation.
This study is an early-stage assessment of PMMY and does not set out to evaluate the impact of
the scheme. Instead, our focus in this report is to study the design and implementation of PMMY
and to generate insights on early performance trends. The report is divided in three chapters. The
rest of Chapter 1 outlines a theory of change and describes the research design. Chapter 2
presents results from the analysis, further sub-divided into five themes. The final chapter
concludes with a range of recommendations pertaining to aspects of loan origination, portfolio
risk management and overall performance monitoring.
10
1.1 Policy design and theory of change
The Pradhan Mantri Mudra Yojana was launched in April 2015, as an initiative to increase
unsecured lending to MSMEs with credit needs upto Rs. 10 lakhs. PMMY offers unsecured loans
for MSMEs requiring credit for investments in existing businesses, as well as for new startups.
Loans upto Rs. 50,000 are categorized as Shishu, from Rs. 50,000 upto Rs. 5 lakhs as Kishor and
further upto Rs. 10 lakhs as Tarun loans. In addition, overdraft facilities extended (to households)
against Pradhan Mantri Jan Dhan Yojana (PMJDY) savings accounts are also reported under this
scheme.
The loans under PMMY are made by Member Lending Institutions (MLIs), which include a
number of Scheduled Commercial Banks from both public and private sectors, Regional Rural
Banks (RRBs), Micro Finance Institutions (MFI) and Non-Banking Finance Companies (NBFC),
many of whom were previously lending to this target segment. Further, by direction from the RBI
since 2015, all lending to MSMEs lower than Rs. 10 lakhs by SCBs is required to be
uncollateralized 8 . As a result of these confounding factors, a straightforward time-series
comparison of loans to MSMEs before and after the scheme is misleading as a measure of impact,
because it is then unclear how to identify the effects of PMMY distinct from other underlying
drivers of MLIs’ original business.
To better understand the design of PMMY and its interaction with the existing lending network,
we develop a theory of change that maps the relationships between each of the scheme’s features,
their corresponding immediate or intermediate effects on MLIs or borrowers, and the type of
financial system impact or outcomes that we may expect as a result of these effects. We believe
the theory of change would serve as a relevant framework to comprehensively track scheme
performance and to retool for efficacy where needed.
Scheme features, effects and outcomes
We identify the unique contribution of PMMY as the introduction of four scheme features
designed to address specific constraints in the supply of enterprise credit to MSMEs. These
scheme features, and their corresponding effects and outcomes are mapped in Figure 2.
The first feature of PMMY is the enrolment of lenders as MLIs into the program, and ongoing
conversations with high-level management to emphasize desired outcomes and mutually agree
on lending targets for each year.
8
Reserve Bank of India. (2016). Master Direction - Lending to Micro, Small & Medium Enterprises (MSME) Sector.
Figure 2 Theory of change for PMMY
12
This may appear to be a rather simple or procedural feature of the scheme but in fact, follows in
a series of focused financial inclusion ‘missions’ initiated by successive governments and
regulators. Researchers studying the impacts of previous large-scale financial inclusion
initiatives 9 conclude that they were, primarily, effective in increasing the flow of finance to
intended beneficiaries. These first-order impacts (and their implications for high-quality
implementation of a large policy initiative) are often over-looked in the search for subsequent
impacts on economic indicators. The proposed theory of change suggests that the effectiveness
of the scheme in delivering first-order impact through the lending channel will mediate the
realization of all other outcomes. It is therefore critical to monitor the additional flow of credit
arising from PMMY, to understand the relative responsiveness of different institutions or
channels, and to study the changing patterns in financial access and financial depth as a result.
The nature of targets and supervision will also determine the nature of outcomes. For example,
we highlight in Figure 2 that setting MLI-level targets may be effective in increasing aggregate
disbursements, but unless MLIs are also specifically directed or encouraged to expand their
outreach or improve service qualities, many other corresponding outcomes such as lowering
regional inequalities, quicker loan processing or financial inclusion of MSMEs may not be
achieved.
The second scheme feature is the provision of refinance 10 support (facilitated by MUDRA) to
MLIs, conditional on their willingness to pass this on as lower interest rates to borrowers11. Banks
primarily finance their lending business through deposit mobilization and therefore enjoy a
lower cost of funds than non-banks, who finance their business through either debt or equity.
While the borrowing profile of MFIs in particular has improved substantially through the use of
assignments, NCDs and securitization, borrowings from banks for on-lending is the primary
source of debt finance used by MFIs (ICRA 2016) 12. The average cost of funds for MFIs is in the
range of 14-15%, depending on their size (MFIN 2016)13.
In this context, the availability of MUDRA refinance (or indeed, any other interventions by MUDRA
to lower MFIs’ cost of funds) could serve to either lower costs on the current size of lending or
9
Burgess, R., & Pande, R. (2005). Do rural banks matter? Evidence from the Indian social banking experiment. The American
economic review, 95(3), 780-795. DOI: 10.1257/0002828054201242 and Butler, A., & Boudot, C. (2016). No Policy is an Island:
Finance and Food Security in India. NSE-IFF Conference on Household Finance.
10
The rate at which MLIs avail refinance is linked to the average cost of funds from RIDF, with a markup of upto 0.75% for banks
and 3% for non-banks.
11
The terms of refinance require that the underlying loans be priced within a specified margin on MLIs’ cost of funds. For SCBs
this is no higher than MCLR+1%, for RRBs and Co-operative Banks no higher than base rate+3.5%, and for NBFC-MFIs no
higher than PLR+6%
12
ICRA. (2016). Indian Micro Finance Penetration levels increasing; Borrower – Lender discipline holds the key for sustainable
growth
Review for the year ended March 31, 2016 and Industry Outlook. New Delhi, India.
13
MFIN. (2016). Micrometer – data as of 31st March 2016. (Issue No. 17).
expand the amount of lending that is possible. However, we caution that these outcomes may be
realized only if the take-up of these facilities is high, and the restrictions on interest margin are
complied with.
The third feature of PMMY is the creation of a specialized credit guarantee facility for unsecured
loans less than Rs. 10 lakhs, intended to encourage uncollateralized lending to new or thin-file
clients. The Credit Guarantee Fund for Micro Units (CGFMU) is a portfolio guarantee facility
extended to eligible MLIs on unsecured loans lower than Rs. 10 lakhs14. All MLIs are eligible to
transact with CGFMU, and guarantee is currently available at a uniform fee of 1-1.5%. CGFMU
compensates 50% of the losses between 5-15% of amount in default.
This facility can be expected to have multiple effects. First, the availability of credit guarantee
could encourage lenders to expand their size of lending to MSMEs and extend loans with
confidence to thin-file clients on the basis of project viability rather than collateral15. Second, to
the extent that the credit guarantee lowers the cost to MLIs of managing credit risk, it could also
lower the risk premiums embedded in loan pricing 16 . Third, a large risk aggregator such as
CGFMU could effectively provide an indirect layer of portfolio risk insurance to those MLIs
(especially NBFCs and RRBs) facing high concentration risk, in the absence of market-prevalent
mechanisms to do so. Used in this manner, the credit guarantee could ultimately serve to increase
MLIs’ resilience to local or sector-specific shocks, and increase market stability. However, like
refinance, these benefits are likely to be realized only if MLIs participate in the guarantee
program.
Apart from these three key features, the issuance of MUDRA debit cards and other directives from
time-to-time could also directly influence the design of loan products offered, or ensure that more
appropriate suite of products is offered to MSMEs based on their need17.
Figure 2 describes the linkages between scheme features and expected outcomes, and outlines
the channels of impact through combinations of intermediate effects. Here, we define scheme
effects as immediate and short-term changes in the PMMY portfolio occurring as a result of one
or more of PMMY’s scheme features, while outcomes are longer-term impacts on the financial
landscape. Favourable outcomes materialize conditional upon on the size and characteristics of
14
With the introduction of CGFMU for unsecured loans lower than Rs. 10 lakhs, these loans will (over a period) no longer be
eligible under CGTMSE. The CGTMSE will remain a transaction-level guarantee scheme for MSME loans larger than Rs. 10
lakhs. This division allows the terms of guarantee under each to differ significantly, and for CGFMU to more suitably address the
higher credit risk latent in unsecured lending.
15
International Finance Corporation, World Bank Group. (2012). Micro, Small and Medium Enterprise Finance in India.
16
Ghatak, S. (2010). Micro, small and medium enterprises (MSMEs) in India: an appraisal. Journal of technology management
& innovation, 6(1), 66-76.
17
Stein, P., Goland, T., & Schiff, R. (2010). Two trillion and counting. International Finance Corporation and McKinsey &
Company.
14
intermediate effects. For example, if we observe systematic increases in aggregate loan
disbursements under PMMY (which is marked in the figure as an intermediate effect), this would
eventually reflect as improved credit depth ratios for the MSME segment, which is marked as an
expected outcome. However, additional outcomes are unlikely to materialize unless several other
intermediate effects are also observed. If loans are systematically targeted to underserved
regions, then regional inequality in access could be reduced. If significant trends of new client
acquisition or client graduation are observed, overall financial access for MSMEs is likely to
improve over the medium-term. If PMMY’s scheme features are effective at lowering interest
rates or encouraging better product design that meets the needs of small firms, then this is likely
to improve not only access to finance, but also the quality of services for MSMEs over the long-
term. Therefore, periodic review of intermediate effects from the early stage of scheme
implementation would be helpful to predict eventual outcomes over the medium and long-term.
This framework is particularly helpful to trace the dependencies of desired outcomes on

immediate, measurable changes that should be observable during scheme implementation. For
example, if policymakers wish to achieve lasting improvements in the terms of enterprise credit
even for thin-file clients and informal enterprises, this would require a coordinated set of
intermediate effects to be appropriately encouraged and periodically monitored. We recommend
anchoring future monitoring and evaluation efforts to this type of framework, and ensuring that
both immediate and longer-term effects and outcomes are robustly measured. Further
modalities are discussed in Section 3.1.
1.2 Study objective and research design
This research study was conducted in 2016-17, alongside the second year of PMMY’s operations
and at a time when several components of the scheme were still evolving towards full realization
(including the continuous enrolment of new MLIs and the ongoing implementation of CGFMU).
Being too early for a full-scale evaluation of scheme impacts, this study instead aims to generate
rich, early insights into various aspects of scheme design and implementation.
The study objectives are:
a. To outline the theory of change for PMMY considering all scheme components, objectives
and desired outcomes, and propose an analytical framework for monitoring of
intermediate effects and measurement of long-term outcomes.
b. Study both macro and micro aspects of scheme implementation. Macro aspects include
the variation by institutional or regional factors, while micro aspects include various
factors affecting loan decisions and the frontline interaction between lenders and
borrowers.
c. Understand the range of financial products available to MSMEs as well as the
characteristics of MSME borrowers, end uses of loan funds, and their overall customer
experiences.
d. Recommend early-stage improvements to better achieve desired outcomes.
The analysis was split into three components. The first component is a deep-dive analysis of
PMMY performance from the administrative data collected by MUDRA. This data includes
disbursement metrics and key portfolio characteristics for all MLIs and all loan size categories,
by state and district and for each year between 2015-2017. Disbursement metrics are number of
loan accounts, total amount sanctioned and total amount disbursed. Performance metrics include
percentage of loans or disbursements made to new customers, women, minorities, through
MUDRA cards, and to PMJDY overdraft accounts. Additional performance metrics representing
the spatial distribution of the portfolio and PMMY outreach relative to regional credit demand
were also estimated by overlaying PMMY data with other publicly available national data sources.
However, the administrative data—even when merged with national surveys— does not
adequately address research questions on the characteristics of MSME borrowers, uses of loan
funds and any continuing barriers to accessing adequate and affordable credit.
We therefore complement the all-India data analysis with two small-scale field surveys in three
financially well-developed districts— Kolkata in West Bengal, Ludhiana in Punjab and Rajkot in
Gujarat. The first is a survey of 176 active MSME borrowers designed to understand the contract
16
features, transaction costs and eventual use of enterprise loans as well as the business, financial
and socioeconomic characteristics of borrowers. The survey will not serve as a representative
picture of PMMY borrower characteristics but offers granular insights on loan contract features
and delivery channel factors as well as their suitability for borrowers and the intended use of
funds.
The third research component is also a small-scale primary survey in the same districts as above.
21 MLIs active in these districts were chosen, and semi-structured qualitative surveys were
administered to the respective frontline loan officers tasked with identifying suitable borrowers
and managing the last-mile delivery of the scheme. The interviews focused on understanding the
underwriting guidelines and loan processing for PMMY loans vis-a-vis other MSME loans, as well
as the service climate for and perspectives of frontline loan officers.
The primary surveys were administered by professionally trained survey teams at IFMR LEAD
and data quality was verified through industry best practices including a rigorous system of
logical triangulations and back-checks. Further details on questionnaire design, sampling
protocols and sample composition are discussed in Sections 2.2 and 2.3.
2. ANALYSIS
This chapter presents the results from three research components, distributed across five
themes. The results are drawn from the analysis of PMMY administrative data collected by
MUDRA (i.e. secondary data analysis), and the analysis of primary data collected from MSME
borrowers and MLIs’ frontline loan officers respectively (i.e. primary data analysis). Each of the
five sections in this chapter link to specific channels of impact in the theory of change, and
combine all relevant evidence from both primary and secondary data analysis.
The themes in each section are:
2.1 Analysis of macroeconomic time-series in MSME lending by both banks and non-banks,
seeking to understand the immediate effects of PMMY on higher aggregate disbursement
(the first scheme effect mentioned in Figure 2)
2.2 A study of the characteristics of MLIs including their participation, portfolio composition,
product range and underwriting rules that determine loan origination. This links to
several components in the theory of change such as the drivers of growth in
disbursement, drivers of new client acquisition and the range of products available to
MSMEs.
2.3 The third section discusses results from the primary survey of MSME borrowers. After a
brief overview of sample composition and loan uses, the findings on operational and
financial characteristics of enterprises with access to credit as well as the customer
experience during loan tenure are discussed. These results relate to the components of
the theory of change concerned with the types of products and loan use, loan processing
time, and new client outreach.
2.4 The fourth section dives deeper both administrative (secondary) data and qualitative
(primary) data to study trends in MLI specialization, and the implications of a fragmented
market structure for inclusion and stability.
2.5 The final section explores regional patterns in disbursements under PMMY and identifies
specific regional characteristics correlated with credit concentration. The results signify
critical implications for regional inequality.
18
2.1 Macroeconomic trends in MSME lending
PMMY was launched in April 2015 and in the first two years, nearly Rs. 3.1 lakh crores in
disbursements had been reported by MLIs under this scheme. In this section, we study aggregate
lending trends available in the public domain to understand whether this represents a substantial
expansion of credit vis-à-vis pre-PMMY levels of lending, and if yes, the channels through which
this effect was achieved. The analysis is limited and preliminary, and there is scope to collate a
wider range of variables for further analysis.
First, we study data published by the RBI on the outstanding credit of SCBs, where we subset only
loans lower than Rs. 10 lakhs and made to individuals or firms engaged in Industry, Trading,
Transport or Professional service occupations. Annual estimates since 2009 are reported by loan
size in Figure 3 and Figure 4. The most striking observation is a consistent upward trend between
2009-2017 across all loan sizes, both in terms of loan accounts and in terms of credit outstanding.
The number of smaller loans (less than Rs. 2 lakhs) made by SCBs rose significantly since 2015
(when PMMY was announced) as well as the total credit disbursed on the largest loans, i.e.
between Rs. 5-10 lakhs.
It is unclear whether this trend is consistent, especially since both instances of post-PMMY
increase occur in only either figure but not in both. Further, even notable increases from this
analysis cannot be rigorously attributed to PMMY alone and as well, it may not be fair to conclude
that where credit outstanding did not increase, PMMY had no effect on those categories. For
example, where credit was extended as working capital loans or credit lines—with a high
frequency of transactions within these accounts but low balances on average—the more
appropriate metric would be total credit disbursed through the year (flow of credit) rather than
credit outstanding at the end of the year (stock of credit). Alternatively, the type of loan would
need to be captured and accounted for.
Next, we study data on lending by microfinance institutions, collected and reported by MFIN.
Figure 5 shows a trend of steady growth in MFIs’ Gross Lending Portfolio (GLP) between 2011-
2016 and therefore, any increase in overall lending since PMMY cannot be directly attributed to
the scheme itself. The data reported also indicates that in FY16, only 64% of GLP was in the form
of non-agriculture enterprise loans, and that the remaining might include loans made towards
cultivation, livestock rearing, household consumption smoothing or other expenses.
Unfortunately, the break-up by loan use is not reported for every year, and we are unable study
the composition of the portfolio by purpose has shifted after PMMY.
Figure 3 Number of MSME loans upto Rs. 10 lakhs by SCBs
Upto Rs. 25k Rs. 25k- 2L Rs. 2- 5L Rs. 5- 10L
90
Lakhs
80
70
60
50
40
30
20
10
0
2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 H1 2016-
Source: RBI DBIE Quarterly BSR-1 Table No. 3.4 17
Figure 4 Total outstanding on MSME loans upto Rs. 10 lakhs by SCBs

Upto Rs. 25k Rs. 25k- 2L Rs. 2- 5L Rs. 5- 10L
90000
Rs. crores
80000
70000
60000
50000
40000
30000
20000
10000
0
2009-10 2010-11 2011-12 2012-13 2013-14 2014-15 2015-16 H1 2016-
Source: RBI DBIE Quarterly BSR-1 Table No. 3.4 17
20
Interestingly, the MFIN data also reports the percentage of GLP that is maintained off-balance
sheet, either through direct assignment or securitization, but due to certain reporting
inconsistencies we are unable to assemble a continuous series for FY16 and FY17. We find a sharp
increase in off-balance sheet origination in FY16 (the first year of PMMY), but we also learn that
this dropped to previous levels by FY17 (Figure 6).
Figure 5 Trends in GLP and Off-B/S portfolio of NBFC-MFIs
GLP GLP (Non-agri enterprise loans) % GLP Off-Balance sheet
60,000 100%
Rs. crores
50,000
80%
40,000
60%
30,000
40%
20,000
23% 21%
20%
16%
12% 20%
10,000
- 0%
2011-12 2012-13 2013-14 2014-15 2015-16
Source: MFIN Micrometer 2015-16
Non-agri GLP is only available for 2015-16
Figure 6 Trends in Off-B/S origination of all microfinance institutions
60%
40%
22%
20% 16% 15%
0%
2014-15 2015-16 2016-17
Source: MFIN Micrometer 2016-17
There is therefore no conclusive evidence that PMMY led to an increase in loan origination of
MFIs, but there is mixed evidence of an effect on banks. However, the channels through which
PMMY had an effect on banks in unclear. For example, if banks did increase their disbursements
of larger Kishor loans in a manner that may not be captured in stock measures such as credit
outstanding or GLP, did the increases accrue to their existing clients or to new clients? Did banks
directly originate new clients or did they leverage the originating capabilities of NBFCs to bring
more loans on-book? And most importantly, are the observed increases an effect of refinance or
credit guarantee facilities (in which case the increases will accrue more from those MLIs who
participated), or did they result purely from high-level pressure and supervision from MUDRA?
There is very limited data available in the public domain for an independent study to explore
these questions, but doing so would contribute richly to how a large policy intervention interacts
with a complex and inter-connected financial system to deliver benefits to citizens.
22
2.2 Characteristics of MLIs
In this section, we consider the types of MLIs participating in PMMY, their relative market shares
and key indicators of portfolio composition. This first set of findings build a richer understanding
of the types of MLIs participating in PMMY, while the second set of findings describe the range of
products they offer to MSMEs and the underwriting rules that govern the origination of loans that
are eventually reported under PMMY.
2.2.1 Market share and growth trends

Participating MLIs are classified in the MUDRA data by their institution type. A broad-level
classification renders three MLI types: Scheduled Commercial Banks, Regional Rural Banks and
Non-banks, including MFIs, NBFCs and SFBs. A second-level classification further categorizes
SCBs into State banks, Nationalized banks, Private banks and Foreign banks.
Analyzing the data available from MUDRA we find that overall, SCBs contributed roughly a-third
of the loan accounts in both years, while Non-banks contributed roughly two-thirds and RRBs less
than 5%. The proportions reverse when measured in terms of amounts, where SCBs contributed
roughly 60% of the disbursements in both years.
We note here that overdrafts extended against PMJDY savings accounts are also reported under
the Shishu category by State banks, Nationalized banks and RRBs. While large in number, these
accounts hold relatively smaller balances and as a result, excluding PMJDY OD accounts lowers
the market share of SCBs when calculated in terms of accounts (Figure 7), but not when calculated
in terms of disbursed amounts (Figure 8). Panels A and B of Table 1 report on Shishu loans
including and excluding PMJDY OD accounts while Panels C and D report the on Kishor and Tarun
loans respectively. The discussion going forward excludes PMJDY OD accounts unless otherwise
specified.
Table 1 reports the total lending and market share for both financial years (FY16 and FY17), and
adjusted estimates for FY17. The adjustment accounts for the ongoing enrolment of new MLIs in
FY17 and in order to present a fairer comparison of portfolio trends year-on-year, recalculates
estimates by excluding 16 MLIs who reported data in FY17 but not in FY16.
Figure 7 Share in PMMY portfolio by MLI type (Accounts)
450 Non-banks RRBs SCBs

Lakhs 400
350
300
250
200
150
100
50
-
Incl. PMJDY OD Excl. PMJDY OD Incl. PMJDY OD Excl. PMJDY OD
FY16 FY16 FY17 FY17
Figure 8 Share in PMMY portfolio by MLI type (Disbursed amount)
200,000 Non-banks RRBs SCBs

Rs.crores
180,000
160,000
140,000
120,000
100,000
80,000
60,000
40,000
20,000
-
Incl. PMJDY OD Excl. PMJDY OD Incl. PMJDY OD Excl. PMJDY OD
FY16 FY16 FY17 FY17
From Table 1 we identify specific institution types that dominate each of the Shishu, Kishor or
Tarun loan categories respectively. For example, NBFC-MFIs alone accounted for 63.9% and
57.3% of non-PMJDY Shishu loans in both years and as well, a small number of MLIs reporting as
Small Finance Banks accounted for more non-PMJDY loans than any other bank category
(comprising a much higher number of MLIs).
A notable shift in the Shishu portfolio occurs with the inclusion of two important Private Banks
in FY17, while they were not earlier reporting under PMMY in FY16. Private banks accounted for
8.9% of non-PMJDY Shishu loan accounts and 9.6% of corresponding disbursements in FY16, but
increased to 24% and 29.6% respectively in FY17. Comparing with the adjusted estimate, we
24
learn that only a 4-percentage point increase is attributable to organic loan growth by the original
cohort of MLIs reporting in the first year, while the residual increase is entirely driven by the data
reported by newly enrolled institutions. This highlights the need for more careful cohort analysis,
or institution-wise analysis for program management and internal decision-making on PMMY, as
these adjustments bear significantly on performance metrics. We discuss related
recommendations in Section 3.3.
While non-banks (and institutions with non-bank origins or SFBs) dominate the smallest loan
category, the traditional banks assume dominance with increasing loan size. Consider Kishor
loans, where Nationalized banks alone accounted for nearly half the portfolio, whether measured
in terms of number of accounts or amounts disbursed. State banks, Private banks and RRBs
followed, while non-bank institutions accounted for less than 2% of credit disbursed under the
Kishor category in FY16. Note here that the inclusion of new non-bank MLIs in FY17 significantly
increased their overall share from 1.9% in FY16 to 8% in FY17. This increase is directly
attributable to the new MLIs, since the adjusted estimate for FY17 is 2%, indicating that the
original cohort of reporting non-bank MLIs only grew their share by 10 basis points.
Similarly, Panel D reveals that Public Sector banks accounted for more than 40% of Tarun loans
in both years, followed by State and Private banks. RRBs, who were significant lenders of Kishor
loans, contributed less than 5% of Tarun loans. It is clear that with increasing loan size, the
dominance of large banks also increases. While MFIs are worthy contenders in the smallest loan
category, their role is smaller than RRBs and non-MFIs NBFCs in the Kishor category. Subsequent
analysis will explore this trend, and discuss the potential for increasing non-bank origination of
larger ticket loans.
Finally, we compare market shares of MLIs between FY16 and adjusted estimates for FY17. This
comparison helps to understand organic shifts in the portfolio driven by the same cohort of MLIs
in both years, rather than by the introduction of a new data. To do this, we compare the growth
rates for each MLI type between FY16 and adjusted FY17, and report separately for each loan size
and by accounts in Figure 9 or amounts in Figure 10. This allows us to see shifts within MLI types,
rather than changes in overall market share.
Figure 9 YOY change in aggregate PMMY portfolio by MLI type (Accounts, FY17)
Shishu Kishor Tarun
200%
150%
100%
50%
0%
State Banks Nationalized Private RRBs NBFC-MFIs Non-NBFC SFBs
banks banks MFIs
-50%
-100%
Figure 10 YOY change in aggregate PMMY portfolio by MLI type (Disbursed amount, FY17)
Shishu Kishor Tarun
200%
150%
100%
50%
0%
State banks Nationalized Private RRBs NBFC-MFIs Non-NBFC SFBs
banks banks MFIs
-50%
-100%
We notice that only two categories reported growth in their Shishu portfolio (Private banks and
non-NBFC MFIs), whereas State banks, RRBs and NBFC-MFIs reported nearly unchanged
portfolios and Nationalized banks reported a sharp withdrawal from the Shishu category. We are
cautious in interpreting these withdrawals, for two reasons. First, the ‘enrolment and
supervision’ scheme feature we described in Section 1 laid specific emphasis on originating
26
Shishu loans in the first year, and this emphasis was somewhat relaxed by the second year of the
scheme. The removal of additional pressure may have triggered the re-direction of resources as
per MLIs’ internal priorities. Second, FY17 also saw a slowdown in lending activity following
demonetization and the unavailability of human resources or other economic conditions could
have caused banks to pause lending. At this time, we flag large withdrawals by Nationalized banks
in FY17 for further analysis and continued tracking over time.
Consistent with the above reasoning, the removal of specific emphasis on Shishu loans seemed to
have also triggered a shift towards Kishor and Tarun loans, especially for SCBs and RRBs. While
State banks and Nationalized banks reported higher growth in Tarun loans, RRBs and MFIs
consolidated growth in Kishor while a few ventured into the Tarun category (which however
remains a tiny share of their portfolios).
Overall, this is an encouraging trend, but with some concerns. First, the market for large ticket
loans remains heavily dominated by banks and to the extent that non-bank entities are able (or
can be supported) to enter this market, they are likely to increase both outreach and competition.
Second, if PMMY serves to increase banks’ origination of larger loans or, systematically helps
deserving MSMEs graduate to larger loans or better terms of credit, that could also result in better
firm productivity and economic growth.
The only notable concern, as we flagged earlier, is the withdrawal of State banks and Nationalized
banks from Shishu lending, especially when it is unclear whether these clients have been
graduated or excluded. We study this trend at a more granular level in a later section dealing with
strategic specialization. A striking implication is that while increasing specialization allows
individual MLIs to optimally re-allocate their resources and lower the costs of delivery, this could
also pose persistent barriers to access or graduation for borrowers in fragmented or under-
served markets.
Table 1 Market share of various MLI types in PMMY portfolio
Panel A: Shishu Loans (incl. PMJDY OD)
Adj. 16-
2015-16 2016-17 17*
MARKET SHARE IN TOTAL ACCOUNTS
Scheduled Commercial Banks 23.7% 30.7% 21.7%

State Banks 2.9% 2.4% 2.9%
Public Sector Commercial Banks 12.6% 5.3% 6.5%
Private Sector Commercial Banks 8.2% 23.1% 12.4%
Foreign Banks 0.0% 0.0% 0.0%
Regional Rural Banks 3.2% 2.5% 3.1%
MFIs, NBFCs & Others 73.1% 66.7% 75.2%

NBFC-MFIs 59.1% 55.3% 62.0%
Other MFIs 2.3% 3.1% 3.8%
Non-Banking Financial Companies 0.0% 0.5% 0.0%
Small Finance Banks 11.7% 7.8% 9.5%
MARKET SHARE IN TOTAL DISBURSEMENTS (IN RS. CRORES)

State Banks 2.2% 1.5% 1.9%

NBFC-MFIs 57.0% 49.6% 59.6%
Other MFIs 3.0% 2.5% 3.3%
28
Panel B: Shishu loans excluding PMJDY OD accounts
2015-16 2016-17 Adj. 16-17*


State Banks 1.4% 1.3% 1.5%

NBFC-MFIs 63.9% 57.3% 64.8%
Other MFIs 2.5% 3.2% 4.0%

State Banks 2.2% 1.5% 1.9%

NBFC-MFIs 57.2% 49.6% 59.5%
Other MFIs 3.0% 2.5% 3.3%
Panel C: Kishor loans
2015-16 2016-17 Adj. 16-17*


State Banks 12.6% 12.8% 13.4%

NBFC-MFIs 3.5% 4.1% 4.2%
Other MFIs 0.0% 0.0% 0.0%

State Banks 16.7% 17.3% 18.4%

NBFC-MFIs 1.5% 1.7% 1.8%
Other MFIs 0.0% 0.0% 0.0%
30
Panel D: Tarun loans
2015-16 2016-17 Adj. 16-17*


State Banks 28.0% 28.6% 30.2%

NBFC-MFIs 0.2% 0.2% 0.2%
Other MFIs 0.0% 0.0% 0.0%

State Banks 29.4% 30.5% 32.1%

NBFC-MFIs 0.3% 0.2% 0.3%
Other MFIs 0.0% 0.0% 0.0%
2.2.2 Product range and underwriting rules
The first part of this section explored, using PMMY administrative data, the role of various
institution types in driving origination in each loan category. In this section, we further explore
the characteristics of each MLI type, with a particular focus on their customer orientation, product
range, origination and underwriting processes and specific relaxations for PMMY loans (if any).
The data for this analysis comes from 21 detailed qualitative interviews with a sample of frontline
loan officers of 15 unique MLIs, including SCBs, RRBs, MFIs and SFBs.
The qualitative survey sought out frontline bank staff in-charge of loans to MSMEs, including
PMMY loans. All MLIs interviewed confirmed that MSME lending was not a new vertical, and that
MSMEs were serviced by their branches in the pre-PMMY period. For each MLI, we noted down
the range of products (credit and non-credit) on offer for MSME clients, as well as their product
features. A consolidated summary is presented in Table 2.
We find that range and types of product offerings for MSMEs varied widely by institution type,
and this is a concern particularly where firms have access to only some MLI types, but remain
excluded by others. For example, MFIs and SFBs in our sample reported that loan amounts could
not exceed Rs. 50,000 on group loans on Rs. 2-3 lakhs on individual loans. Some SFBs were
developing secured loan products to offer larger loans against property, but these still remain a
very small portion of their MSME credit. The same institutions also reported much higher average
interest rates on their loans, anywhere from 10-14 percentage points higher than SCBs or RRBs.
A second concern is the near absence of flexible loans or credit lines for MSMEs served by MFIs.
All MFIs and SFBs in our sample only described term loans with fixed, equal repayments which
disregard underlying cashflow volatility and the needs of firms seeking credit for working capital
management rather than as finance for fixed assets. In fact, tiny or micro businesses who manage
their working capital through fixed term loans bear a disproportionately high interest burden, in
addition to the costs of managing or deploying excess funds at the time of disbursement. It would
be more efficient to underwrite the same amount as a credit line, and allow firms to transact
indefinitely within this limit and only pay interest proportional to the credit effectively used.
A third concern from Table 2 is the limited availability of non-credit products through non-bank
channels. MFIs are only able to offer savings products through BC partnerships, and those are
often limited to cross-sale to existing loan clients. Newly licensed SFBs have the opportunity to
quickly expand access to a range of savings products (including term deposits and recurring
deposits to accumulate small savings) to a large client base. Recall that the number of Shishu loans
originated by SFBs was greater than those by any other bank type and under the new license, all
of these clients should now have access to savings alongside credit.
32
A similar product gap is revealed when comparing the insurance products available through non-
bank MLIs. MFIs typically bundle a credit life insurance cover with group and individual loans
wherein the life of the borrower and borrower’s spouse are covered, but only upto the value of
the loan and only during loan tenure. In effect, this insures the lender from risk of default but
benefits to the borrower only amount to a waiver of loan dues and a small payout. In this first
part of this section we learned that a majority of borrowers under PMMY were served by MFIs or
SFBs, and we now understand that almost all of them are likely un-insured or under-insured. Even
in the case of banks, where multiple insurance products were reported, there is no data available
to understand how many MSME clients are appropriately insured.
Overlaying these insights on findings from the first part of this section reinforce that a majority
of MSME credit is delivered through MFIs and SFBs, where credit is both limited and expensive
and as well, credit is often the only available product and savings, insurance and investment
options are scarce. Together, these insights describe a financial landscape for MSMEs (even after
the launch of PMMY) that is expensive, incomplete, and often unsuitable and herein lies the
biggest challenge for PMMY in powering MSME growth, business health and household financial
well-being. While PMMY will remain a credit-focused scheme, policy-making and supervision will
need to recognize the risks of furthering the product gap by emphasizing loans over all other
financial products.
The implications of an expensive, incomplete and unsuitable range of products are particularly
salient when considered within the theory of change. In Figure 2 we hypothesized that some of
the immediate effects of PMMY might be to lower the cost of borrowing for MSMEs and that
through the use of MUDRA card or through access to a wider range of products, borrowers could
enjoy better matched products at better terms. However, if the PMMY portfolio in effect inherits
the features of the pre-PMMY financial landscape, these intermediate effects (and corresponding
impacts) are unlikely to materialize.
Table 2 Product range of MLIs interviewed
SCBs RRBs MFIs SFBs
Group loan Yes, incl. in partnership Upto Rs. 50,000 Upto Rs. 40,000
with BCs, MFIs
Individual unsecured loan
Term loan/fixed repayment Upto Rs.10L Upto Rs.10L Upto Rs.2L Upto Rs.3L
at 9-12% at 12-14% p.a. at 24-26% at 19- 24%
Flexible terms/WC loan Upto Rs. 10 lakh Upto Rs. 10 Lakh No No
Individual secured loan

Against business investments Yes, hypothecated Yes, hypothecated No No
(incl. transport vehicle) through loan tenure through loan tenure
Against personal property Yes Yes No Upto Rs. 10L at ~15%
Against gold Yes Yes Yes (some MFIs) Yes
Other financial products

Savings SB, Current A/C, FD, RD SB, Current A/C, FD, RD Limited, through BC Limited SB, Current A/C,
partnerships FD, RD (may not be
available in all branches)
Insurance Life, Personal Accident, Life, Personal Accident, Life Life
Property, Liability Property, Liability (limited to loan amount) (limited to loan amount)
34
In addition to a review of products, our qualitative interviews also captured details of loan
processing and underwriting or sanctioning guidelines. In a later section, we study the processes
and workflow differences between MLI types, but here we briefly describe specific issues with
regard to underwriting.
We notice in Table 1 that all SCBs reported a relatively small number of Shishu loans, given their
widespread branch network. From our qualitative interviews, we understand that most SCBs
took limited efforts to attract new Shishu borrowers, with the exception of responding to walk-in
inquiries or occasional awareness camps. Nevertheless, we understand that for those clients who
do submit loan applications, SCBs in our sample were willing to relax underwriting requirements
to an extent (in comparison to their standard requirements for all MSME loans).
Most importantly, SCBs were willing to waive requirements for business financial documents,
valuation of goods and assets and threshold liquidity or leverage ratios for firms who they viewed
as ‘only served through PMMY’. In interviews, respondents alluded to the credit guarantee facility
as a comfort in these cases and that therefore, rigorous documentation for informal businesses
was not requested. Instead, they employed on a simple 3-part rule: KYC documentation,
firm/shop registration documentation and a minimum CIBIL score of 600-700. We find that
similar rules were extensively used by many banks, and is likely a result of internal directives to
permit approval of more Shishu loan applications.
We find no comparable relaxations made by MFIs or SFBs. In these MLIs, Shishu loans were
typically made through joint-liability groups (where only KYC documentation, group membership
and currently outstanding loans were verified), while Kishor loans were rare, and only offered to
business clients with a long and disciplined association with the lender. Individual loans were
underwritten based on detailed business appraisal including an analysis of cashflows using sales
receipts, debt servicing capacity and owner’s financial position. MFIs and SFBs in our sample did
not report the use of credit guarantee facility. While it does appear that through joint-liability
lending they were able to reach out to formerly excluded clients, their selection criteria for Kishor
loans is currently quite restrictive.
Combining these insights with the relative dominance of these MLI types in the PMMY portfolio,
we conclude that both MFIs and SCBs take efforts to simplify access requirements for Shishu loans
but as loan sizes increase (and recall that banks contribute a larger share of the larger loans),
requirements for financial documentation can be a significant barrier to access. MFIs and NBFCs
have demonstrated the ability to underwrite business cashflows without extensive formal
documentation, but their capacity to do so remains limited and as a result, their contribution to
Kishor and and Tarun portfolios remains small.
2.3 Features and uses of PMMY loans
The previous sections discussed MLIs’ relative contributions and the key characteristics of the
financial landscape for MSMEs. In this section, we use available data to gain insights into the types
of loans made under PMMY as well as the characteristics of PMMY borrowers. To complement the
all-India data available from MUDRA, we conducted a small-scale sample survey that adds rich
detail on the characteristics and experiences of PMMY borrowers, but does not match the
representative breadth of MUDRA data. However, the primary surveys offer critical insights on
how beneficiaries of this scheme interact with the financial system.
2.3.1 Insights on loan features from MUDRA data

The data reported to MUDRA includes aggregated totals (at an MLI-level and region-level) of
specific loan tags for each loan category. Loans are tagged based on borrower characteristics,
including the gender, caste and minority status of borrower, whether the borrower is new-to-
bank, whether the loan was issued with MUDRA card or as a PMJDY overdraft, and whether the
loan is tagged under various government sponsored schemes. In this report, we analyze patterns
in the proportion of loans that are made to women, minorities, new-to-bank clients and through
MUDRA cards. In addition, we also estimate the average loan ticket-size for each MLI and loan
size category as the average disbursement per loan account reported. These results are reported
in Table 3 Panels A-F.
Panel A reports the average loan ticket-size by each MLI type and loan size category. We observe
that within the Shishu category (excluding PMJDY OD), RRBs and State banks made the largest
loans on average (greater than Rs. 30,000) while NBFC-MFIs made the smallest loans, at an
average around Rs. 19,000.
The Kishor category, by design, includes loans ranging from Rs. 50,000 to Rs. 5 lakhs and resulting
from a wide range, we observe a twin-peak distribution. For NBFC-MFIs & SFBs, the average loan
size was around Rs. 70,000-80,000, while average loan sizes available from all banks (including
SCBs and RRBs) and non-MFI NBFCs were between Rs.1.8-3 lakhs. Loans in the Tarun category
tended to have the least variation across MLI types, with average size ranging between Rs. 6.5-8
lakhs.
We also observed the average loan ticket-size changing between the years, perhaps reflecting
compositional changes in portfolio such as the expansion or reduction in the average scale of
finance available to borrowers within each category, or the inclusion of new borrowers who
received smaller initial loans. The size of the average Shishu loan made by Nationalized banks
36
increased from Rs. 27,283 to Rs. 38,121. Recall that Nationalized Banks reduced their overall
exposure in the Shishu category at the same time, perhaps making fewer but larger loans. Another
notable change is the reduction in average Kishor loan size by RRBs from Rs. 1,84,086 in FY16 to
Rs. 1,39,892 in FY17. The same is evident from Table 1, where RRB’s number of Kishor loans
registered a very small increase from 8,03,756 to 8,14,855 while the total disbursements
decreased from Rs. 2,917 crores to Rs. 2,733 crores. It is unclear whether this is an outcome of
resource limitations in FY17, or the result of making smaller loans to many “new” clients in place
of a portion of existing clients.
Table 3 also reports the proportion of accounts tagged as new-to-bank, as loans to women and
minorities, and loans with a MUDRA card. We are careful to interpret these estimates with caution
as all are self-reported by MLIs where definitions are inconsistent and for all practical purposes,
open to discretionary interpretation. It is also possible that some tags are not captured entirely
and loans are only tagged upon discretion, or upto the achievement of targets.
With these caveats, we note in Panel B that a large proportion of loans by all SCBs were tagged as
new to bank in both years. Many SCBs also reported a trend of a lower proportion of new-to-bank
loans in the Shishu category by the second year, matched by a higher proportion in the Kishor
category, suggesting a strategy to pursue inclusion through mid-size loans. However, the
proportions themselves seem questionably high and we later describe a more rigorous research
exercise to obtain more accurate estimates of inclusion and graduation under PMMY, and by
different MLIs (Section 3.4).
Panels C and D report the proportion of loans made to women and minorities respectively. The
data confirms that through the JLG model, MFIs, SFBs and some Private banks have successfully
reached out to women borrowers, who represent an overwhelming majority of loans in these
categories. However, outside of these MLIs, the representation of women clients in larger size
loans, or loans from SCBs & RRBs is low, and upto 30% at best.
Similarly, the proportion of loans made to minority communities is also low on average. It is
unclear whether both these tags are recorded and reported accurately under PMMY, and this
concern over-rides any inferences from this data. For example, it is likely that MFIs over-estimate
the proportion of loans eventually used by women entrepreneurs since loans may be borrowed
in the name of women but used for enterprises owned by men in the household. On the other
hand, banks may under-report the proportion of loans used by women and minorities, unless this
tag is internally monitored and verified.
Panels E and F report the proportion of loans made through MUDRA cards and as PMJDY
overdrafts, both of which are only made by State, Nationalized and Regional Rural banks. We find
that a very small proportion of loans are made through the MUDRA card, even for MLIs authorized
to do so. While MUDRA cards allow borrowers more control over cash management and timely
loan use, they also imply an additional fixed cost on each loan.
We also note that the proportion of loans issued with a MUDRA card are lower in the second year.
This could result from a large number of loan renewals to the same clients, therefore requiring a
lesser number of new cards. However, this would be inconsistent with Panel B, where a large
proportion of new clients are reported. In order to clarify reporting on this tag and enable
meaningful decision-making, we recommend issuing a clarification on this tag and discuss further
in Section 3.
In summary, the data reported to MUDRA informs only a primary characterization of borrowers
based on the segmentation of loan sizes and lender types. We understand the average loan size
for each MLI, and to the extent that loan size and other contract features are correlated with
certain types of loan uses, we might extend our understanding of whether loans are used for
working capital, consumption smoothing or for capital investments. We also understand that the
PMMY portfolio inherits social inequalities from the larger financial system, and barriers for
access to credit especially for women continue to persist. Recall that the theory of change in
Figure 2 hypothesizes that the reduction in inequalities may not be automatic, and will be
conditional on focused efforts to bridge regional inequalities. The same hypothesis may be
extended to gender or other inequality as well.
We understand the limitations of data collection from MLIs voluntary reporting, and complement
the administrative data with our own sample survey of MSME borrowers. We discuss the survey
and findings in the next part of this section.
38
Table 3 Key characteristics of PMMY loans from MUDRA data
Panel A: Loan average ticketsize
2015-16 2016-17 Adj. 16-17
Shishu Kishor Tarun Shishu Kishor Tarun Shishu Kishor Tarun
SCBs 25,156 208,845 726,940 30,302 207,502 751,892 27,334 207,878 751,916
State Banks 31,299 263,522 763,080 28,128 258,200 797,955 28,138 258,200 797,955
Nationalized Banks 27,283 188,264 716,500 38,117 187,224 749,180 38,121 187,224 749,180
Private Banks 22,265 233,305 703,945 29,368 236,722 687,553 23,952 239,694 687,592
Foreign Banks 303,729 814,474 334,091 816,552 334,091 816,552
RRBs 36,289 184,086 735,005 33,533 139,892 731,286 33,538 139,892 731,286
Non-Banks 19,009 80,625 783,897 20,840 172,382 689,242 20,812 79,430 785,000
NBFC-MFIs 18,452 82,247 784,284 20,611 79,237 789,124 20,577 79,219 789,124
Other MFIs 25,087 197,406 18,633 18,633
Other NBFCs 22,594 272,390 684,701
SFBs 20,627 71,859 728,571 23,209 81,026 659,524 23,209 81,026 659,524
Panel B: Percentage of loans with tag "New to bank"
2015-16 2016-17 Adj. 16-17
SCBs
State Banks 36.6% 43.4% 39.7% 14.5% 58.1% 58.6% 14.5% 58.1% 58.6%
Nationalized Banks 59.8% 71.9% 73.0% 38.3% 66.1% 60.9% 38.3% 66.1% 60.9%
Private Banks 60.4% 48.9% 20.1% 22.0% 66.5% 35.0% 47.4% 67.7% 35.0%
Foreign Banks 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%
RRBs 53.6% 44.6% 39.5% 49.1% 52.5% 33.9% 49.1% 52.5% 33.9%
Non-Banks
NBFC-MFIs 25.8% 5.7% 4.7% 19.7% 12.3% 8.5% 19.8% 12.0% 8.5%
Other MFIs 14.2% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%
Other NBFCs 0.1% 51.3% 61.4%
SFBs 28.7% 54.6% 0.0% 33.6% 31.6% 64.3% 33.6% 31.6% 64.3%
Panel C: Percentage of loans with tag "Women"

2015-16 2016-17 Adj. 16-17
SCBs
State Banks 10.8% 14.7% 8.9% 5.9% 9.9% 2.4% 5.9% 9.9% 2.4%
Private Banks 78.4% 18.4% 8.7% 94.2% 16.4% 9.4% 87.6% 16.7% 9.4%
Foreign Banks 0.0% 9.8% 12.5% 0.0% 15.9% 13.1% 0.0% 15.9% 13.1%
RRBs 20.9% 20.5% 16.8% 30.3% 31.0% 16.7% 30.3% 31.0% 16.7%
Non-Banks
NBFC-MFIs 99.7% 71.6% 74.5% 81.0% 72.6% 72.9% 83.0% 72.5% 72.9%
Other MFIs 100.0% 100.0% 0.0% 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
Other NBFCs 99.9% 3.0% 3.0%
SFBs 100.0% 63.7% 0.0% 71.5% 5.4% 0.0% 71.5% 5.4% 0.0%
40
Panel D: Percentage of loans with tag "Minorities"
2015-16 2016-17 Adj. 16-17
SCBs
State Banks 3.7% 5.7% 3.9% 1.7% 5.5% 2.7% 1.7% 5.5% 2.7%
Private Banks 15.2% 11.6% 7.7% 15.2% 12.6% 9.0% 10.0% 12.6% 9.0%
Foreign Banks 0.0% 0.0% 0.0% 0 0 0 0.0% 0.0% 0.0%
RRBs 10.2% 10.8% 17.9% 19.4% 22.3% 21.0% 19.4% 22.3% 21.0%
Non-Banks
NBFC-MFIs 15.1% 0.1% 0.3% 14.6% 0.7% 0.2% 14.4% 0.7% 0.2%
Other MFIs 10.0% 6.5% 0.0% 0 0 0 0.0% 0.0% 0.0%
Other NBFCs 0.0% 4.9% 3.5%
SFBs 0.6% 2.1% 0.0% 5.4% 18.6% 11.9% 5.4% 18.6% 11.9%
Panel E: Percentage of loans with tag "MUDRA card"

2015-16 2016-17 Adj. 16-17
SCBs
State Banks 6.3% 3.2% 3.3% 0.3% 0.5% 0.5% 0.3% 0.5% 0.5%
Private Banks
Foreign Banks
RRBs 8.7% 0.5% 1.3% 2.0% 0.3% 1.0% 2.0% 0.3% 1.0%
Non-Banks
NBFC-MFIs
Other MFIs
Other NBFCs
SFBs
Panel F: Percentage of loans with tag "PMJDY OD"
2015-16 2016-17 Adj. 16-17
SCBs
State Banks 53.7% 48.6% 48.6%
Nationalized Banks 41.5% 41.7% 41.7%
Private Banks
Foreign Banks
RRBs 21.9% 12.1% 12.1%
Non-Banks
NBFC-MFIs 0.0% 0.4% 0.5%
Other MFIs
Other NBFCs
SFBs
42
2.3.2 Insights on loan features and uses from survey of MSME borrowers
The primary survey of MSME borrowers was located in regions with relatively high access to
credit under PMMY that also had a competitive financial landscape (i.e. active participation by
multiple lenders and institution types). Accordingly, three districts were chosen— Kolkata in
West Bengal, Ludhiana in Punjab and Rajkot in Gujarat— and urban and semi-rural MSME
clusters in these districts were identified. Within each cluster, micro and small enterprises were
sampled using a combination of systematic random sampling and purposive selection. The
criteria for purposive selection excluded enterprises without an outstanding business loan during
the survey period.
Most borrowers were unaware or unable to report whether their loan was under PMMY. We
therefore sampled likely PMMY borrowers based on the type of enterprises, type of lender and in
most cases, the loan amount. Based on the loan sizes reported we classified respondents by their
designated PMMY category. The sample includes few businesses with non-PMMY loans (loan size
above Rs. 10 lakhs) as well and were retain them in the sample to provide a comparison for loan
usage and customer experiences.
A trained survey team administered a detailed questionnaire to the owner or most

knowledgeable informant of each enterprise. The questionnaire was designed to understand
primarily loan access, use and transaction process. In addition, we also collected details of
business ownership, activity, scale of operations, revenue and expense cash-flows and volatility,
physical and financial asset ownership and basic socioeconomic characteristics of business
owners’ households. A total of 176 interviews were administered. 97% of survey respondents
owned a single business and had only one outstanding enterprise loan, while the remaining
reported multiple businesses or loans.
The 176 enterprises sampled in the primary survey were chosen from Kolkata, Ludhiana and
Rajkot districts. A majority of enterprises were engaged in trading activities (57.4%), followed by
a variety of services (29.5%) and only 13% of the sample were engaged in manufacturing
activities. In terms of loan size, we classify sampled enterprises by four PMMY loan categories
instead of the usual three— borrowers of Kishor loans are further sub-classified at a cut-off of Rs.
2 lakhs as Kishor 1 (Rs. 50,000 – Rs. 2 lakhs) and Kishor 2 (Rs. 2-5 lakhs). We find that the
pooled sample is well-balanced across all loan sizes.
Table 4 Sample composition for survey of MSME borrowers
All Kolkata Ludhiana Rajkot
Number of surveyed enterprises 176 72 42 62
Primary activity
Manufacturing 13.1% 13.9% 4.8% 17.7%
Trading 57.4% 62.5% 81.0% 35.5%
Services 29.5% 23.6% 14.3% 46.8%
PMMY loan category
Shishu
28.4% 52.8% 16.7% 8.1%
(Upto Rs. 50,000)
Kishor 1
32.4% 27.8% 42.9% 30.6%
(Rs. 50,001 – 2 lakhs)
Kishor 2
14.8% 6.9% 16.7% 22.6%
(Rs. 2 – 5 lakhs)
Tarun
14.2% 9.7% 11.9% 21.0%
(Rs. 5 -10 lakhs)
Above Rs. 10 lakhs 10.2% 2.8% 11.9% 17.7%
The analysis of survey data in this section begins with a quick description of sample composition
and nature of financial access in the survey locations. However, the key results from the survey
focus not on describing the receivers of PMMY loans but on aspects of business cashflows, loan
use and customer satisfaction.
Table 5 describes the credit landscape in survey areas. As expected, Nationalized and Private
banks contribute anywhere between 50-80% of loans, with their share being the lowest for the
smallest loan sizes. MFIs made 46% of the Shishu loans reported in the sample, but had almost
no participation in any other loan category. All co-operative bank loans in the sample were
reported in Rajkot district, where these banks were an active competitor to Nationalized and
Private banks.
In comparison with Table 3, we see that the average loan sizes are slightly higher than the all-
India average. This could be due to the fact that the areas studied are moderately high access
markets and large centers of economic activity and thus received higher credit than average.
44
Table 5 Key characteristics of loans from primary survey
Shishu Kishor 1 Kishor 2 Tarun

% Sample loans in sample, by MLI type:
Nationalized Bank 20.0% 26.3% 32.0% 12.5%
Private Bank 32.0% 56.1% 36.0% 50.0%
Microfinance Institution 46.0% 5.3% 0.0% 0.0%
Cooperative Bank 2.0% 12.3% 32.0% 37.5%
Median loan amount (in Rs.) 35,000 100,000 310,000 800,000
Median loan amount by MLI type:
Nationalized Bank 50,000 135,000 300,000 800,000
Private Bank 32,500 100,000 400,000 850,000
NBFC-MFI 30,000 100,000 - -
Cooperative Bank 50,000 100,000 300,000 700,000
% Loans requiring collateral*, or collateral verification
All sources 22.0% 24.5% 36.0% 41.7%
By MLI type:
Nationalized Bank 30.0% 33.3% 50.0% 33.3%
Private Bank 25.0% 21.9% 44.4% 66.7%
NBFC-MFI 17.4% 0.0%
Cooperative Bank 0.00% 28.6% 12.5% 11.1%

*
Includes both household and business assets
Survey respondents also noted that their currently outstanding loans from banks often required
some form of collateral or collateral verification. It is unclear whether the verification included
both business and personal assets and further, if any personal assets were pledged against loans
which are likely reported under PMMY. However, we observe a higher proportion of banks’
customers reporting lenders’ interest in collateral than those borrowing from MFIs.
Through the survey of enterprises, we also inquired on the eventual uses of loan funds.
Encouragingly, a large majority of borrowers reported direct use of funds for both capital and
working capital needs attached to their business— in fact, 99% enterprises reported at least one
business-related use. Note that Figure 11 reports the percentage of enterprises reporting use of
loan funds on each purpose, and the same loan might be used for multiple purposes. Purposes are
not mutually exclusive, and the estimates do not total to 100%. However, we do find the diversion
of some loan funds towards non-business uses, most commonly to finance household purchases,
consumption smoothing or meet emergent expenses.
While it is nearly impossible for lenders to exercise control over the end use of loan funds, these
estimates are helpful to understand the extent to which funds disbursed under PMMY are
deployed in productive enterprise activity, and the likely scale of economic benefits that might
accrue. Understanding diversion of funds is also helpful if it indicates households’ struggle to
manage cash-in-hand. Product innovations such as the MUDRA card allow the borrow a higher
degree of control over their use of funds.
Figure 11 Multiple uses of loans by borrowers
100%
80%
60%
40%
20%
0%
46
We now turn our attention to a few operational and financial characteristics of MSME borrowers.
As reported earlier, a majority of enterprises were engaged in trading or small service activities,
and manufacturing enterprises are more frequent only in the largest loan size category (see Table
6). We also learn that a majority of the enterprises we interviewed were sole proprietorships, but
around 30-40% of those receiving larger loans were jointly owned and managed by multiple
people within the household. Formal inter-household partnership firms were rare in the study
locations.
Two other results from Table 6 link to results earlier discuss from Table 3. We find that the
proportion of women-owned enterprises in the sample is quite low, even among Shishu loans,
suggesting that while many JLG loans may be disbursed to women they may not be exclusively
deployed in women-owned enterprises. This highlights the need to exercise caution when
interpreting aggregate estimates of “loans to women” from PMMY administrative data. Similarly,
we also find that many enterprises in our sample were well-established— the median age of
enterprises was 11 years— and we find little evidence, in this sample, for a massive boost to
entrepreneurship through increased access to credit. A more realistic estimate is, we find, that
around 14% of enterprises are likely to be “new” but even here, it is impossible to establish a
direct causal link with PMMY.
Finally, on sample characteristics, we discuss characteristics of borrowers’ households. For most

MSME borrowers, the income to household from the business was their primary and dominant
source of income. Only 5% respondents reported an additional source of agricultural or livestock
income, while 11% reported additional income from regular wage employment of any household
member. We expect that in remote, rural regions (where this sample has very little
representation) we might observe more income diversification through cultivation, livestock
farming and wage labour, and fewer opportunities for regular wage income.
Respondents reported fairly low access to financial products (for either household or business
use), indicating a wide product gap even among active clients of formal financial institutions.
Curiously, life insurance was the most frequently reported financial product, followed by savings
deposits as liquid or fixed deposits. Respondents also reported ownership of a variety of physical
assets (likely also a form of alternative savings, in the absence of access to financial savings
products)— we find that more than 64% borrowers owned land, and 42% borrowers reported
ownership of gold, and both might be used as collateral against formal credit.
Table 6 Characteristics of borrowers’ enterprises, from primary survey
Sample Shishu Kishor 1 Kishor 2 Tarun > Rs. 10L
Women-owned enterprises % MSMEs 10.2% 10.0% 10.5% 7.7% 20.0% 0.0%

Median age of enterprise Years 11.5 16.5 8 11.5 10 11.5
Established in 2015 or later % MSMEs 14.2% 12.0% 12.3% 15.4% 12.0% 27.8%
Primary activity
Manufacturing % MSMEs 13.1% 10.0% 12.3% 15.4% 20.0% 11.1%
Trading % MSMEs 57.4% 56.0% 57.9% 57.7% 60.0% 55.6%
Services % MSMEs 29.5% 34.0% 29.8% 26.9% 20.0% 33.3%
Nature of business ownership

Single owner % MSMEs 67.6% 92.0% 56.1% 69.2% 60.0% 44.4%
Partnership/Family owned % MSMEs 32.4% 8.0% 43.9% 30.8% 40.0% 55.6%
Number of employees Median 2 2 2 3 4 2
48
Table 7 Characteristics of borrowers’ households, from primary survey
Sample Shishu Kishor 1 Kishor 2 Tarun
Median Household size 4 4 4 4 4
Median annual income

From enterprise 150,000 100,000 150,000 190,000 200,000
From all sources 196,200 118,200 150,000 200,000 230,000
% households with:
Agri or Livestock income 5% 0% 9% 0% 0%
Regular wage income 11% 14% 12% 8% 12%
Receiving welfare
14% 12% 25% 8% 8%
payments
Receiving pension 2% 4% 4% 0% 0%
Physical asset ownership (% households)

Land where household
64% 78% 63% 46% 56%
resides
Land for business use 10% 6% 18% 8% 4%
Gold 42% 57% 46% 16% 32%
Bicycle 19% 46% 14% 0% 0%
Motorcycle 51% 36% 46% 69% 60%
Four wheelers 6% 2% 6% 8% 16%
Furniture 32% 48% 33% 15% 24%
Machinery/equipment 27% 24% 27% 23% 28%
Small tools 14% 22% 9% 15% 16%
Financial product ownership (% households)

Savings deposits 12% 10% 11% 20% 12%
Life Insurance 35% 38% 38% 20% 36%
Health Insurance 5% 6% 2% 4% 8%
Asset Insurance 2% 2% 2% 0% 4%
Credit card 6% 2% 2% 8% 20%
Pension 1% 2% 0% 0% 0%
Mutual funds and
0% 0% 0% 0% 0%
investments
The analysis so far described a sample of borrower enterprises in urban and semi-urban districts
with high economic activity, matched with vibrant financial markets with multiple institutions
and institution types. The characteristics of enterprises themselves closely mirror trends in the
overall MSME landscape for India, in terms of ownership, business activity and size. Enterprises
sampled do seem to have received higher credit than average, and reported access to a variety of
financial products. However, the access is in no way comprehensive or adequate, given both
business’ and households’ complex financial needs. The rest of this section will explore trends in
business cashflows and profitability and then focus on aspects of loan processing and service that
relate to customer satisfaction.
Business cashflows and profitability
The survey asked detailed questions in all items of business expense and revenue in an average
month, as well as the volatility experienced in months of high expense or revenue and low
expense or revenue. The data is insightful to understand both cashflow volatility as well as
dynamic working capital management by firms to avoid revenue deficit in certain months. These
results are summarized in the graphs below.
Figure 12 described the median experience of borrowers by loan category in terms of expense18
volatility. We find that borrowers in the Kishor 1 category had the highest absolute expenses on
average, while Shishu borrowers had the lowest expenses in the average month. Figure 13
describes the experience with revenues, and we notice that revenues are more volatile than
expenses across all categories—the blue markers represent median revenues in an average
month, while the vertical black lines represent how high or low the revenue could be in peak and
low months respectively. Note that enterprises with larger loan sizes experienced the lowest
volatility, while enterprises with smaller loans experienced more volatility. If we assume that
enterprises with smaller loans also have a smaller scale of operations and lower managerial
capacity, then it is likely that the burden of volatility management is disproportionately higher
for smaller firms.
Figure 14 considers the interaction of expense and revenue in various scenarios and its’
implication for business profitability and debt servicing capacity. First, we note that 88% all
MSME borrowers in the sample were profitable on average, while the remaining 12% did not
appear strictly profitable in the current year. It is unclear how unprofitable enterprises managed
18
Total expense and profit estimates exclude wages or other payments deducted by the owner.
50
loan repayments. Also note that the enterprises with larger loans also had a higher likelihood of
being profitable on average.
Figure 12 Median expenses in average, high and low months
40,000
35,000
30,000
25,000
20,000
15,000
10,000
5,000
0
Figure 13 Median revenues in average, high and low months
40,000
35,000
30,000
25,000
20,000
15,000
10,000
5,000
0
Figure 14 Profitability of sample enterprises
% profitable enterprises Profit margin ("Average" month)
Profit margin ("High" month) Profit margin ("Low" month)
100%
80%
60%
40%
20%
0%
Note: Gross profits excludes wages and other payments deducted by the owner.
Yet another interesting observation from Figure 14 is that while enterprises in all categories
experienced both expense and revenue volatility, their average profit margin in all scenarios
remained reasonably consistent. We take this as indication of dynamic working capital
management by savvy firms, although it would be interesting to further explore what types of
firms were able to successfully maintain a consistent profit margin even under volatile conditions,
and if the type of finance had any role to play in supporting a profitable strategy.
Factors in customer satisfaction
We also interviewed enterprises to understand customer perceptions of MLI’s and features of

loans. We learned that the loan amount required or available was the primary motivation for
choice of lending institution, followed by recommendations of a particular MLI from business
contacts. One of every four enterprises reported that, at the time of loan application, they did not
seek out alternative offers from any other lender, while another 54% reported that they sought
upto three offers.
Table 8 Customer experience with loan processing and service

All sample Nationalized Private Microfinance
Banks Banks Institutions
Information submitted* during loan appraisal
Credit history 58.3% 57.9% 47.6% 50.0%
Repayment delays 48.6% 52.6% 43.9% 3.8%
Value of assets 54.3% 71.1% 46.3% 7.7%
Income/expense seasonality 78.9% 86.8% 72.0% 69.2%
Financial statements 57.1% 68.4% 53.7% 7.7%
References 42.3% 55.3% 30.5% 0.0%
52
Site visit 70.5% 63.2% 79.3% 84.6%
Time to disbursement
Less than 15 days 43.2% 21.1% 42.7% 73.1%
Between 16-30 days 38.1% 47.4% 35.4% 26.9%
Between 31-60 days 16.5% 21.1% 20.7% 0.0%
More than 60 days 2.3% 7.9% 1.2% 0.0%
Number of visits to branch 6.7 6.3 5.8 2.0
Overall satisfaction
Not fully satisfied 17.6% 34.2% 17.1% 0.0%
Unlikely to borrow again 21.6% 31.6% 30.5% 0.0%
*This is self-reported by borrowers based on their experience during loan appraisal, and may not include any data
collected by lenders on borrowers through external sources.
Table 8 describes the various aspects of borrower appraisal (subject to respondent recall) as well
as the onerous waiting time till loan disbursement. Infact, 28% borrowers reported that in
addition to long waiting times, they had to make 6 or more visits to the branch during the loan
appraisal process.
It appears that a majority of lenders in this sample (across all institution types) asked to verify
the value of assets, and this could include both personal and business assets. It is unclear from
borrower reporting whether the assets were valued but not actually pledged as collateral. (It is
unlikely that surveys based on respondent recall will be able to rigorously distinguish between
the two, given the complexity of loan documentation and imperfect understanding and recall of
respondents in this regard).
Overall, 82% of borrowers in our sample were satisfied with their experience and relationship
with their MLIs. Interestingly, 100% of borrowers from MFIs reported satisfaction, while banks’
clients were more likely to express dissatisfaction and dissent. One of every five borrowers
reported that they were unlikely to borrow again from their current lender. While we
understand that MFIs offer significant advantages in terms of loan processing time, and often
service clients who would otherwise remain excluded— and both of these may go a long way in
explaining high customer satisfaction— there remains scope for improvement for all MLIs in
terms of satisfying customer needs and further, the financial needs of the entire MSME sector.
The implications of customer dissatisfaction are far-reaching. The theory of change implicitly
assumes that customers once ‘formally included’ will remain so, and that the benefits of financial
inclusion can compound and accrue over time. However if, as the scheme achieves more maturity,
we observed MSMEs moving in and out of formality or not seeing their needs suitably met, there
emerges a risk that the benefits of inclusion may not be realized to their full extent.
2.4 Lender specialization and market fragmentation
The analysis in earlier sections presented evidence of increasing lender specialization in the
second year of PMMY, i.e. when the initial emphasis on originating Shishu loans was lifted, many
SCBs shifted their focus away from smaller loans and instead towards making (larger) Kishor and
Tarun loans. In this section, we explore this trend at a more granular level— at the level of each
institution, rather than institution type. We also discuss pertinent evidence from qualitative
interviews of frontline loan officers on how their workflow and loan processing is organized, and
how this reflects an inherent institutional strategy favouring particular loan sizes. We conclude
by drawing out the implications for market fragmentation and ultimately, for MSME borrowers.
2.4.1 Evidence of specialization from MUDRA data

Primary evidence of lender specialization comes from the administrative data on PMMY itself
(Figure 15). Supporting previous evidence on MLI specialization, we find that most MLI portfolios
were heavily in favour of either smaller or larger loans in both years, and very few MLIs’ appeared
to cater both loan sizes. The graph highlights that within the bank category, State banks,
Nationalized banks and RRBs have portfolios dominated by Kishor or Tarun lending, while
Private banks stand out with a different composition. Among banks, Private banks alone have an
aggregate composition dominated by Shishu loans. Similarly, among non-banks, aggregate
portfolios can be expected to be dominated by Shishu loans, with the only exception being non-
MFI NBFCs who, in this case, are MSME specialist lenders.
It is likely that the MLI types whose portfolios are atypical from their peers are making a strategic
choice to do so. For example, we know that many Private banks have entered into agreements
with NBFC-MFIs to co-originate JLG loans, and these loans will be held on banks’ balance sheets.
Similarly, while many NBFC-MFIs find it profitable to have an exclusive and limited focus on
small-ticket JLG loans, non-MFI NBFCs have an entirely different specialization, and are in the
business of financing enterprises in specific trades and have built up informational and technical
advantages to do so in a competitive manner.
We further explore evidence on strategic specialization, by studying portfolio trends at a more

granular level. Table 9 reports the proportion of Shishu, Kishor and Tarun loans respectively for
each MLI, while Table 10 YOY changes in MLIs’ portfolio compositionTable 10 YOY changes in
MLIs’ portfolio composition reports the year-or-year growth between FY16 and FY17 in each
loan category by MLI. Supporting previous evidence on MLI specialization, we find that most MLI
portfolios were heavily in favour of either smaller or larger loans in both years, and very few
MLIs’ appeared to cater both loan sizes.
54
Figure 15 Composition of aggregate portfolios by loan size and MLI type
Shishu Kishor Tarun
State Banks
Public Sector Commercial Banks
Private Sector Commercial Banks
Foreign Banks
Regional Rural Banks
NBFC-MFIs
Other MFIs
Non-Banking Financial Companies
Small Finance Banks
0% 20% 40% 60% 80% 100%
Defining a non-specialist as an MLI with atleast 10% of their PMMY loans in all three loan sizes,
we identify only twelve non-specialist MLIs out of a total of 137. Of these, five were State banks
and one was an RRB with only 13 loans in FY17. The remaining six were Private Banks with a
total contribution of 99,216 loans in FY16 and only 30,109 loans in FY17.
A striking trend, as shown from data in Table 10, is that almost all State and Nationalized banks
decreased their exposure in the Shishu segment (in terms of both accounts and disbursement
amount), and that many of these banks reported corresponding increases in Kishor and Tarun.
An exception here is state banks grew exposure in all three categories. Among 21 Nationalized
banks, 16 grew Kishor or Tarun portfolios while decreasing Shishu exposure.
Among Private banks and RRBs, we find mixed trends. Of 17 Private banks reporting in both
years, 5 banks reduced exposure in all three categories while 6 banks increased exposure in all
three. Only 1 bank was observed with a moderate increase in Tarun accompanied by a
corresponding decline of nearly one-third of Shishu loans, while a private bank reported
aggressive growth in Kishor and Tarun loans but no Shishu loans. Among RRBs, a total of 56 banks
reported PMMY loans in both years. All RRBs made Shishu and Kishor loans, but only 45 reported
greater than 30 Tarun loans. Here, we find that 14 RRBs decreased exposure in all loan categories,
but only 4 RRBs increased exposure in all three. 24 RRBs were observed increasing Kishor and/or
Tarun exposure while decreasing Shishu exposure.
Among non-banks we observe an overall trend of consolidation towards Shishu loans. 35 non-
bank entities reported in both years, of which 24 MLIs made exclusively Shishu loans. Of these
24, only 4 MLIs reported systematic decreases in Shishu exposure while the reported high growth
in this segment. Only 1 SFB reported decreases across the board in all categories.
We are cautious to infer these time trends as purely strategic, for a number of reasons. We
consider the possibility that many lenders faced significant resource constraints 19 after
demonetization, and resorted to making fewer larger loans in order to meet disbursement targets
in time. However, since only 22 MLIs reported decreases across all categories, this suggests that
the decreases are perhaps more reflective of resource constraints and credit rationing rather than
blanket liquidity constraints. Given this, the trends observed among different MLIs can be
considered indicative of their strategic alignment in favour of larger or smaller ticket loans
respectively (and as their likely choice of lending strategy in the absence of express direction to
do otherwise).
From this analysis, we infer that MLIs participating in PMMY (and indeed, all MLIs) are loan-size
specialists by choice, perhaps as part of a larger strategy to optimize internal processes and
capitalize on particular strengths or successes. Sahasranaman and George (2013)20 documented
the costs of rural credit delivery through different channels and find that costs vary significantly
based on whether the loans are originated by the banks through their rural branches (41.5% of
the loan amount) or in partnership with a microfinance institution (in which case the costs are as
low as 13.8%). Surprisingly though, even Nationalized and Private banks were specialists, while
many State banks are not. It is unclear whether State banks’ non-specialization is a result of
management mandate or an outcome from their sheer scale of outreach. In the case of the latter,
it might be that their wide rural outreach manifests in significant cost advantages.
The trend of MLI specialization interacts uniquely with the theory of change. Figure 2 assumes
that credit risk is the only (or the most significant) barrier for thin-file clients to receive credit
and therefore, the introduction of a well-functioning credit guarantee accompanied by a nudge
from the supervisor should be sufficient to smoothen barriers to graduation. However, lender
specialization and market fragmentation may render this linkage ineffective, since MLIs
specialization choices may over-ride the benefits of a credit guarantee. Thin-file clients may
remain constrained to microfinance institutions, and their ability to graduate to a higher loan size
is then dependent on the emergence of a particular type of enterprise lending specialists who are
able to underwrite thin-file clients. Increasing market fragmentation decreases consumers’
effective choice and restricts both mobility and competition.
19
Several NBFC-MFIs and other non-banks faced liquidity constraints both in terms of ability to borrow and as well as in internal
redeployment of funds from repayments collected. State and Nationalized banks may have faced other constraints with regard
to work allocation and availability of personnel to process a large number of loans. For all of these reasons, small-ticket lending
appears to have slowed.
20
Sahasranaman, A. & George, D. (2013). Cost of Delivering Rural Credit in India. IFMR Finance Foundation Working Paper.
Retrieved from http://foundation.ifmr.co.in/2013/04/23/cost-of-delivering-rural-credit-in-india/
56
Table 9 Composition of MLIs’ portfolios by loan size
Member Lending Institution 2015-16 2016-17
Shishu Kishor Tarun Shishu Kishor Tarun
SCHEDULED COMMERCIAL BANKS
State Banks
MLI 1 55.8% 27.1% 17.1% 29.0% 31.2% 39.7%
MLI 2 43.4% 42.6% 14.0% 32.5% 51.2% 16.3%
MLI 3 22.0% 61.5% 16.5% 24.9% 58.8% 16.3%
MLI 4 57.7% 28.5% 13.7% 34.3% 40.8% 24.9%
MLI 5 54.4% 37.2% 8.4% 41.3% 46.2% 12.5%
MLI 6 54.3% 30.8% 14.9% 50.6% 33.7% 15.8%

MLI 1 86.5% 12.1% 1.3% 67.6% 29.6% 2.8%
MLI 2 60.8% 34.3% 4.9% 34.5% 55.6% 9.9%
MLI 4 58.8% 37.4% 3.9% 32.7% 59.3% 8.0%
MLI 5 44.9% 48.2% 6.8% 32.0% 59.8% 8.2%
MLI 6 62.7% 32.1% 5.2% 50.3% 40.7% 9.0%
MLI 7 74.3% 19.5% 6.3% 73.2% 21.2% 5.6%
MLI 8 73.3% 19.2% 7.4% 40.5% 40.6% 18.8%
MLI 9 91.6% 6.9% 1.5% 88.1% 10.2% 1.8%
MLI 10 70.4% 25.2% 4.4% 41.2% 46.8% 12.0%
MLI 11 59.9% 33.7% 6.5% 40.9% 49.8% 9.3%
MLI 12 75.8% 21.9% 2.4% 43.4% 53.3% 3.4%
MLI 13 55.0% 39.1% 6.0% 56.6% 38.3% 5.1%
MLI 14 74.0% 22.1% 3.9% 42.9% 40.9% 16.2%
MLI 15 57.6% 36.6% 5.9% 36.7% 55.2% 8.2%

MLI 16 54.1% 36.9% 9.0% 25.1% 57.3% 17.6%
MLI 17 54.1% 37.5% 8.3% 40.2% 52.3% 7.6%
MLI 18 76.6% 21.0% 2.4% 48.0% 46.5% 5.5%
MLI 19 57.8% 36.8% 5.4% 36.2% 53.9% 9.9%
MLI 20 65.6% 25.5% 8.9% 32.5% 52.8% 14.7%
MLI 21 55.7% 37.7% 6.6% 35.2% 52.7% 12.1%
MLI 1 26.4% 46.7% 26.9% 22.7% 50.8% 26.5%
MLI 2 20.2% 52.3% 27.5% 19.0% 49.6% 31.3%
MLI 3 15.5% 48.2% 36.3% 33.7% 44.8% 21.5%
MLI 4 50.6% 28.3% 21.1% 80.1% 11.1% 8.8%
MLI 5 0.0% 40.0% 60.0% 85.0% 6.6% 8.4%
MLI 6 25.6% 56.5% 17.9% 5.4% 61.3% 33.3%
MLI 7 93.3% 5.2% 1.4% 96.1% 2.8% 1.0%
MLI 8 93.2% 4.8% 2.0% 95.1% 2.8% 2.1%
MLI 9 28.8% 67.6% 3.6% 85.1% 11.9% 3.0%
MLI 10 97.9% 1.6% 0.5% 94.4% 4.5% 1.0%
MLI 11 20.8% 49.8% 29.4% 42.7% 37.1% 20.1%
MLI 12 26.7% 66.7% 6.7% 88.8% 10.3% 0.9%
MLI 13 28.7% 63.1% 8.2% 24.6% 68.2% 7.3%
MLI 14 89.4% 9.8% 0.8% 90.2% 8.7% 1.2%
MLI 15 3.5% 55.3% 41.3% 2.1% 70.0% 27.9%
MLI 16 18.4% 62.0% 19.6% 14.5% 63.2% 22.3%
58
MLI 17 37.7% 32.2% 30.1% 20.8% 54.5% 24.8%

Foreign Banks
MLI 1 0.0% 70.6% 29.4% 0.0% 40.0% 60.0%
MLI 2 0.0% 0.0% 100.0% 0.0% 0.0% 100.0%
REGIONAL RURAL BANKS

MLI 1 46.5% 52.7% 0.8% 68.9% 30.7% 0.4%
MLI 2 25.5% 73.9% 0.6% 50.4% 47.6% 2.0%
MLI 3 68.1% 30.6% 1.3% 29.5% 69.5% 1.1%
MLI 4 79.1% 20.2% 0.7% 74.8% 24.5% 0.7%
MLI 5 94.7% 4.7% 0.6% 85.4% 14.0% 0.6%
MLI 6 88.9% 10.8% 0.3% 84.4% 15.4% 0.3%
MLI 7 49.0% 34.9% 16.1% 61.5% 15.4% 23.1%
MLI 8 89.1% 8.5% 2.4% 82.4% 14.4% 3.3%
MLI 9 73.4% 22.2% 4.4% 63.1% 34.8% 2.0%
MLI 10 54.5% 41.5% 4.1% 27.9% 64.9% 7.3%
MLI 11 77.3% 22.5% 0.2% 50.4% 49.1% 0.5%
MLI 12 71.6% 25.2% 3.3% 56.1% 39.0% 4.9%
MLI 13 84.8% 14.6% 0.6% 40.1% 59.3% 0.6%
MLI 14 73.7% 25.8% 0.5% 40.2% 59.4% 0.3%
MLI 15 83.6% 16.3% 0.1% 44.0% 55.8% 0.2%
MLI 16 71.0% 28.7% 0.4% 92.9% 7.0% 0.1%
MLI 17 34.7% 54.4% 10.9% 23.6% 75.9% 0.5%
MLI 18 79.2% 19.4% 1.4% 77.2% 21.3% 1.5%
MLI 19 82.5% 16.4% 1.1% 63.5% 34.5% 2.0%
MLI 20 82.2% 16.7% 1.1% 93.7% 4.5% 1.8%
MLI 21 31.0% 62.9% 6.1% 31.0% 62.9% 6.1%
MLI 22 25.9% 68.2% 5.8% 30.1% 62.9% 6.9%

MLI 23 81.0% 18.5% 0.4% 67.7% 31.4% 0.9%
MLI 24 68.3% 30.7% 1.1% 51.5% 46.8% 1.6%
MLI 25 92.0% 7.8% 0.3% 89.8% 8.5% 1.7%
MLI 26 57.4% 38.0% 4.6% 43.0% 50.9% 6.0%
MLI 27 95.6% 3.7% 0.7% 88.7% 9.3% 1.9%
MLI 28 95.0% 4.5% 0.5% 91.2% 8.2% 0.6%
MLI 29 82.9% 13.3% 3.7% 87.3% 8.7% 4.0%
MLI 30 40.3% 53.4% 6.3% 27.4% 66.5% 6.2%
MLI 31 53.3% 45.1% 1.6% 41.0% 57.3% 1.7%
MLI 32 51.7% 45.3% 3.0% 48.3% 49.3% 2.4%
MLI 33 67.7% 30.7% 1.6% 45.7% 51.6% 2.8%
MLI 34 74.7% 23.9% 1.3% 60.8% 37.3% 1.9%
MLI 35 58.7% 37.9% 3.3% 99.5% 0.5% 0.0%
MLI 36 34.9% 64.1% 1.0% 34.9% 63.3% 1.8%
MLI 37 25.2% 74.8% 0.0% 16.0% 79.3% 4.7%
MLI 38 67.7% 30.5% 1.9% 59.3% 37.4% 3.3%
MLI 39 66.7% 30.6% 2.8% 68.7% 23.5% 7.8%
MLI 40 63.2% 35.9% 0.9% 26.2% 72.8% 1.0%
MLI 41 81.4% 18.5% 0.0% 62.1% 36.9% 1.0%
MLI 42 23.8% 68.4% 7.8% 20.8% 68.9% 10.3%
MLI 43 85.9% 13.1% 0.9% 79.4% 19.6% 1.0%
MLI 44 60.6% 37.2% 2.2% 77.1% 21.8% 1.2%
MLI 45 74.3% 23.7% 1.9% 85.6% 13.8% 0.6%
MLI 46 70.3% 29.1% 0.5% 46.1% 53.1% 0.8%
60

MFIs, NBFCs & Others
NBFC-Micro Finance Institutions
MLI 1 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 2 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 3 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 4 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 5 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 6 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 7 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 8 98.2% 1.8% 0.0% 96.0% 4.0% 0.0%
MLI 9 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 10 99.8% 0.2% 0.0% 99.8% 0.2% 0.0%
MLI 11 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 12 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 13 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 14 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 15 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 16 97.4% 2.6% 0.0% 99.4% 0.6% 0.0%
MLI 17 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 18 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 19 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 20 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 21 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 22 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 23 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 24 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 25 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 26 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 27 100.0% 0.0% 0.0% 99.9% 0.1% 0.0%
MLI 28 98.9% 1.1% 0.0% 99.3% 0.7% 0.0%
MLI 29 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
MLI 30 99.9% 0.1% 0.0% 99.9% 0.1% 0.0%
MLI 31 99.6% 0.4% 0.0% 99.8% 0.2% 0.0%
Other MFIs
MLI 1 99.9% 0.1% 0.0% 100.0% 0.0% 0.0%
Small Finance Banks

MLI 1 100.0% 0.0% 0.0% 99.8% 0.2% 0.0%
MLI 2 99.0% 1.0% 0.0% 99.9% 0.1% 0.0%
MLI 3 99.2% 0.8% 0.0% 98.4% 1.6% 0.0%
MLI 4 100.0% 0.0% 0.0% 100.0% 0.0% 0.0%
62
Table 10 YOY changes in MLIs’ portfolio composition
Member Lending Institution Shishu Kishor Tarun
(Only MLIs reporting in both years, % YOY growth in terms of % YOY growth in terms of % YOY growth in terms of
and >30 loans in each category) Accounts Disbursement Accounts Disbursement Accounts Disbursement
SCHEDULED COMMERCIAL BANKS
State Banks
MLI 1 -71% -68% -36% -37% 28% 38%
MLI 2 -2% 2% 56% 58% 52% 56%
MLI 3 42% -2% 20% 12% 24% 26%
MLI 4 -57% -46% 3% 20% 31% 44%
MLI 5 -26% -29% 21% 21% 45% 46%
MLI 6 19% 5% 39% 35% 34% 40%

MLI 1 -73% -58% -15% -13% -27% -27%
MLI 2 -64% -51% 4% 8% 29% 35%
MLI 3 -239% -41% 15% 37% 16% 51%
MLI 4 -51% -51% 38% 51% 80% 81%
MLI 5 -43% -41% -1% -3% -4% -6%
MLI 6 -27% -4% 15% 52% 58% 83%
MLI 7 5% -2% 16% 9% -5% -5%
MLI 8 -64% -60% 39% 53% 67% 74%
MLI 9 -17% -30% 28% 22% 2% -5%
MLI 10 -52% -44% 53% 82% 127% 159%
MLI 11 -21% -13% 70% 57% 65% 77%
MLI 12 -48% 37% 119% 58% 28% 44%
MLI 13 -22% -14% -26% -32% -35% -34%
MLI 14 -65% -61% 11% 53% 151% 184%
MLI 15 -19% -17% 92% 88% 77% 77%
MLI 16 -63% -64% 23% 38% 55% 62%
MLI 17 -29% -31% 34% 32% -12% -3%
MLI 18 -66% -56% 20% 23% 22% 23%
MLI 19 -30% -37% 65% 28% 104% 125%
MLI 20 -47% -48% 121% 102% 76% 53%
MLI 21 -48% -26% 15% 21% 52% 44%

MLI 1 -34% -39% -16% -22% -25% -27%
MLI 2 11% 19% 12% 26% 34% 26%
MLI 3 -73% -65% -88% -90% -93% -93%
MLI 4 342% 349% 9% 7% 17% 18%
MLI 5 303% 306% 246% 228%
MLI 6 -99% -98% -93% -91% -88% -88%
MLI 7 22% 40% -36% -38% -14% -14%
MLI 8 39% 42% -20% -19% 41% 38%
MLI 9 1005% 1147% -34% -28% 216% 217%
MLI 10 2% 4% 207% 158% 117% 115%
MLI 11 -61% -51% -86% -86% -87% -86%
MLI 12
MLI 13 65% 46% 108% 74% 71% 55%
MLI 14 35% 30% 17% 83% 97% 75%
MLI 15 93% 106% 311% 349% 119% 97%
MLI 16 -25% -33% -3% 1% 8% 13%
MLI 17 -96% -58% -87% -89% -94% -94%
Foreign Banks
MLI 1 -70% -67% 7% 9%
MLI 2
REGIONAL RURAL BANKS

MLI 1 -21% -26% -69% -89% -73% -76%
MLI 2 11% 27% -64% -63% 86% 77%
MLI 3 -60% -56% 111% 74% -24% -32%
MLI 4 -17% -12% 6% 4% -10% -11%
MLI 5 -42% -9%
MLI 6 102% -18% 201% 164% 96% 80%
MLI 7 -96% -97% -99% -99% -95% -94%
64
MLI 8 -49% -49% -7% -3% -27% -30%
MLI 9 192% 75% 433% 172% 59% 56%
MLI 10 -74% -65% -21% -16% -10% -10%
MLI 11 -71% -67% -2% -5% 10% 22%
MLI 12 100% -10% 295% 15% 286% 74%
MLI 13 -49% -58% 334% 278% 8% 4%
MLI 14 -29% -36% 197% 26% -14% -3%
MLI 15 -68% -56% 108% 86%
MLI 16 -77% -55% -96% -96% -97% -97%
MLI 17 -42% 3% 19% 63% -96% -91%
MLI 18 -9% 4% 3% 7% 0% 0%
MLI 19 -51% -45% 33% 12% 19% 21%
MLI 20 -73% -72% -94% -82% -62% -68%
MLI 21 -1% -2% -1% -1% -2% -2%
MLI 22 -34% -75% -48% -43% -33% -31%
MLI 23 -2% 7% 99% 90%
MLI 24 -58% -51% -16% -9% -15% -11%
MLI 25 2% 14% 14% 83% 587% 526%
MLI 26 -26% -23% 33% 27% 29% 29%
MLI 27 -23% -22% 109% 108% 138% 117%
MLI 28 -48% -49% -2% -18% -34% -41%
MLI 29 22% 16% -25% -25%
MLI 30 -40% -36% 10% 1% -13% -10%
MLI 31 -43% -45% -6% -4% -23% -25%
MLI 32 35% -28% 57% 57% 16% 1%
MLI 33 -66% -61% -14% 10%
MLI 34 -63% -67% -30% -32% -37% -41%
MLI 35 -69% -61% -100% -100% -100% -100%
MLI 36 -37% -38% -38% -30% 13% 6%
MLI 37 -24% -24% 27% 46%
MLI 38 19% 7% 66% 33% 140% 132%
MLI 39 -1% -2% -26% -9% 172% 219%
MLI 40 -58% -52% 104% 115%
MLI 41 -25% -30% 96% 83%
MLI 42 -45% -52% -36% -35% -17% -18%
MLI 43 -18% -4% 32% 29% -8% -9%
MLI 44 146% -2% 13% 3% 0% -1%
MLI 45 59% 71% -19% 6% -56% -42%
MLI 46 -50% -49% 39% 90%
MLI 47 -51% -45% 21% 13% -31% -32%
MLI 48 -23% -17% 63% 46% 31% 25%
MLI 49 -41% -20% 49% 57% -99% -97%
MLI 50 -40% -22% 105% 147%
MLI 51 33% 44% 371% 326% -4% -24%
MLI 52 -55% -53% 4% -1% 21% 5%
MLI 53 -19% -59% 128% 8% -6% -2%
MLI 54 -58% -44% 143% 174% -6% -7%
MLI 55 -1% 4% 73% 57%
MLI 56 472% 195% 322% 130% 94% 58%
MFIs, NBFCs & Others

NBFC-Micro Finance Institutions
MLI 1 42% 62%
MLI 2 57% 81%
MLI 3 189% 208% 322% 357%
MLI 4 114% 134%
MLI 5 79% 89%
MLI 6 53% 80%
MLI 7 14% 29%
MLI 8 -32% -32% 56% 48% 23% 25%
MLI 9 18% 37%
MLI 10 -3% 16% -1% -18%
MLI 11 14% 30%
MLI 12 -46% -42%
66
MLI 13 111% 147%
MLI 14 122% 205%
MLI 15 96% 99%
MLI 16 284% 326% -14% 95%
MLI 17 120% 24%
MLI 18 19% 33%
MLI 19 -13% 17%
MLI 20 -27% -10%
MLI 21 129% 150% 82% 292%
MLI 22 -21% -6%
MLI 23 82% 123%
MLI 24 84% 125%
MLI 25 -67% -63%
MLI 26 11% 40%
MLI 27 1% 19%
MLI 28 3% 39% -33% -36%
MLI 29 107% 129%
MLI 30 29% 57% -1% -2%
MLI 31 6% 13% -37% -35%
Other MFIs
NON NBFC-Micro Finance Institutions 52% 13% -100% -100%
Small Finance Banks

MLI 1 -31% -28% 529% 960%
MLI 2 -12% -1% -94% -93%
MLI 3 -54% -48% -10% 12%
MLI 4 39% 70%
2.4.2 Implications for origination and market structure
The analysis in Section 2.4 so far presented evidence from administrative PMMY data on MLIs’
strategic specialization and portfolio consolidation towards their perceived strengths even
within a short period of two years. Along these lines, we discuss insights from the interviews of
frontline loan officers on how lending processes are organized in a segmented market, and draw
out critical implications for overall market structure and for consumers.
Our qualitative survey of frontline loan officers of both banks and MFIs inquired in detail on loan
work allocation, including client identification and the various levels of decision-making involved.
We learn that for almost all MLIs in our sample, regardless of institution type, the processing of
PMMY loans was combined with all MSME loans and in the case of small MFIs, it also shared
multiple steps with group loans too. There is a clear efficiency advantage to doing so, and this
suggests that MLIs with stronger systems and higher capacity to process MSME loans also have
the potential to scale the fastest in terms of PMMY.
Figure 16 walks through the broad steps in loan origination. We learn that for most bank branches
(both SCBs and RRBs), all decision-making and sanctioning powers with regard to PMMY loans is
self-contained within the last-mile branch, and the loan officer may be keen to build a service
history with the client before sanctioning credit. Some larger banks have dedicated MSME lending
branches and all loans sourced from a designated service area and referred to the specialist
branch. In contrast, MFIs prefer that limited autonomy be given to the last-mile node and instead,
loans are underwritten and sanctioned by a central officer at a regional office. A single credit
officer will service individual loan applications from multiple service centres and often this officer
will be tasked with verifying documentation, visiting business locations in-person and also
performing additional quality monitoring checks through the appraisal process.
We understand that performance incentives vary greatly between MLIs but in general, SCB and
RRB branches are assigned overall credit volume targets (for all MSME credit and not specific to
PMMY) and in combination with the fact that the origination process is almost self-contained
within each branch, this could lead to a perceived optimization of human resources by seeking to
make “easier” larger-ticket loans to “safe” borrowers. On the other hand, loan officers of MFIs are
incentivized on the basis of new group formations/new clients, but since the appraisal and
underwriting (of individual loans) is performed by a trained specialist at a regional office, the
cascading effect of incentives is more pronounced in encouraging more group loans where
underwriting criteria are also simpler and loan sanction is quicker.
68
We therefore find that the cascading effect of both cost21 and incentive structures is pronounced
and persistent in determining the allocation of resources and loan outcomes. SCBs and RRBs,
when they don’t have specialized MSME lending branches, have a high degree of autonomy in
loan-making and therefore may choose to achieve their assigned targets by optimizing their
resources in favour of fewer, larger-ticket loans. This explains why a majority of their loans fall in
the Kishor and Tarun category (despite a widespread branch network in rural and low-income
areas) and why, when Shishu loans are made they are often the result of persuasive walk-in
customers.
MFIs’ processes are such that group loans are still encouraged as the primary entry point for all
new customers and even those with pre-existing businesses are often required to complete 2-3
cycles as disciplined group borrowers before they become eligible for larger, individual loans.
Second, because the capacity to underwrite business cashflows within these institutions is often
limited to 1-2 senior staff in each regional office, they also have a limited capacity to rapidly scale
this vertical of their portfolio.
We find that for these reasons, the trend of specialization is both strategic and pre-determined by
institutional factors. It would be unreasonable to expect that any MLI would deviate significantly
from their loan-size specialization under PMMY unless there are powerful incentives and easy
channels to do so. Our concern is that this creates a fragmented market for MSME credit, and in
the absence of easily accessible means to graduate from one MLI to another, small firms may find
their growth significantly restricted by which MLI they are associated with, or have access to.
Consider, for example, that many micro and small firms were first formally “included” through
MFIs, with whom they built up a repayment history over time. Not all firms will grow uniformly
and infact, many firms might find it optimal to remain small. However, for firms that have the
appetite, competency and opportunity to make early investments and grow rapidly, credit
constraints imposed by MFIs could be limiting. Recall that of all MFIs’ loans under PMMY, less
than 1% of loans were greater than Rs. 50,000.
In a fragmented market, the availability and use of comprehensive credit reports by lenders, their
approach to underwriting clients who are experienced borrowers but effectively “new” to this
institution, and vibrant competition are critical to removing barriers to client mobility and
graduation.
21
Sahasranaman, A. & George, D. (2013). Cost of Delivering Rural Credit in India. IFMR Finance Foundation Working Paper.
Retrieved from http://foundation.ifmr.co.in/2013/04/23/cost-of-delivering-rural-credit-in-india/
Figure 16 Steps in loan processing and designated work allocation, by MLI type
Client identification Documentation & Loan decisioning Disbursement

Verification
Frontline bank officer or BC Credit officer Credit officer/Branch Frontline officer/Branch

manager manager
SCBs & RRBs
Walk-in or awareness camps, CIBIL score, financial
following which they are statements and ratios Sanction depends on credit Upto 14 days after loan
assisted with application form including asset-liability limit assigned to the branch sanction
and collection of KYC position and debt service
documents capacity
Loan officer/Sales officer Area manager/Senior loan Nodal credit officer Loan officer
officer
Area survey and group Decision to sanction is made Within 7 days
MFI & SFB formation KYC and residence by a nodal credit officer upon
Group loans verification, Compulsory verification of application forms
Group Training and Group and additional quality
Recognition Test monitoring
Loan officer/Sales officer Area manager/Senior loan Nodal credit officer Loan officer
officer
MFI & SFB Decision to sanction is made Within 7- 14 days
Individual loans Leads generated from existing
pool of clients, preferably with Business site visit, verification by a nodal credit officer upon
trading, transport or small of business cashflows and verification of application form,
manufacturing businesses and inventory in addition to credit financial analysis and
with a good repayment record reports additional quality monitoring
70
2.5 Regional patterns in the PMMY portfolio
The PMMY portfolio is spread across the length and breadth of the country and unsurprisingly,
there is significant variation is disbursements at both state and district levels. The data reported
to MUDRA is available for all MLIs and all loan categories at a state-level for FY16 and at both
state and district-levels for FY1722.
This section explores the adequacy of credit supply in various regions and how the flow of credit
is determined by demand-side and supply-side features. For example, when we rank districts in
terms of their PMMY credit supply in FY17, we find that the supply to the top 10 districts alone
(with a total of 31.5 lakh PMMY loans) was roughly equal to the supply to the lowest-ranked 355
districts at the other end of the distribution. These districts are mapped in Figure 17, where the
districts shaded orange received the highest share of loans under PMMY, the districts shaded blue
received loans but were ranked at lowest in the distribution. The districts are chosen such that
the number of loans received by 10 orange-shaded districts and 355 blue-shaded districts were
roughly the same. It is evident that the top districts are large centres of economic activity and the
loans were likely made to productive causes. However, the stark inequality raises a concern that
productive enterprises elsewhere in the country may not be as well-served.
In this section, we use PMMY data from FY17 and estimate two measures of regional performance.
The first is a measure of credit concentration, i.e. each district’s share in all-India number of
Shishu, Kishor and Tarun loans. The second is a measure of credit saturation estimated as the
number of PMMY loans in a district divided by the number of micro and small enterprises
estimated in the Sixth Economic Census 23 . The saturation measures serve as an important
comparison to the concentration measure, and adjusts for the likely concentration of loans in
enterprise clusters or regions with a higher ability to productively deploy credit24.
22
Two MLI types did not report data at district-level, instead reporting their data under district “Other”. These were Non-NBFC
MFIs and Non-MFI NBFCs, and they constitute only 5.02% of total loans and 5.05% of total disbursement amount in FY17.
23
Data for the Sixth Economic Census was collected in 2012-13.
24
A limitation of this measure is that we are only able to consider the number of MSMEs in a region (varied by available estimates
of their size), but not their GDP or size of credit demand. If district-level MSME GDP estimates were available, a second measure
of credit saturation or credit depth could be estimated as the total disbursed amount relative to the GDP.
Figure 17 Regional disparities in access to PMMY loans
The top 10 districts (marked in orange) received

31.5 lakh loans under PMMY in 2016-17.
That is roughly the same as the number of loans

received by 355 other districts (marked in blue)
in total, in the same year.
Patna
Nagpur Kolkata
Belgaum
Bangalore Kancheepuram
Mysore
Cuddalore
Coimbatore
Thanjavur
72
2.5.1 Concentration in Tier 1 and Tier 2 centres
We first try to understand whether credit is concentrated particularly in Tier 1 and Tier 2 centres.
We use the 3-tier classification25 recommended by the Sixth Central Pay Commission, which was
developed on the basis of population but has come to have a wider appeal as a broader indicator
of population density, urban development and economic opportunity. We would therefore also
expect that cities with a higher-ranked classification would also have a better economic and
financial profile. Since PMMY data is only available district-wise, districts with Tier 1 and Tier 2
cities are tagged, and we compare their respective shares in PMMY lending with their share in the
distribution of MSMEs.
We find that 8 Tier 1 districts alone account for nearly 5% of all Shishu loans, and 86 Tier 2
districts accounted for 25% of Shishu loans. These proportions are roughly in-line with the
presence of MSMEs in these respective districts. However, for Kishor and Tarun loans we observe
higher concentration, and that the degree of concentration is higher for larger loan sizes. For
example, Tier 1 districts received 6.3% of Kishor loans and 10.4% of Tarun loans, while Tier 2
districts received 30% of Kishor loans and 36.5% of Tarun loans.
Figure 18 Comparison of Tier 1 and Tier 2 regions’ share of loans and share of MSMEs
Tier 1 Tier 2 Others
Share of MSMEs
Share of Shishu loans
Share of Kishor loans
Share of Tarun loans
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
We also present these estimates disaggregated by MLI type in Table 11. For Shishu loans we find
that State & Nationalized banks made a disproportionately high share of their Shishu loans in Tier
25
According to this classification, Ahmedabad, Bangalore, Chennai, Delhi, Hyderabad, Kolkata, Mumbai and Pune are Tier 1
(“X”), 91 other cities are Tier 2 (“Y”) and the rest are not assigned to any tier.
1 locations than the average (15.8% compared to 4.6%), at the cost of credit flow to non-tier
locations. In comparison, we find that all other MLI types (including Private banks) disbursed a
large majority of Shishu loans in non-tier locations. Regional Rural banks, unsurprisingly, wered
absent from Tier 1 locations and consistently disbursed a dominant share of their loans (of all
size) in non-tier locations, likely by design and the localized nature of their operations.
Table 11 Tier 1 and Tier 2 credit concentration by MLI type

Tier 1 Tier 2 Non-tier
Number of districts 8 86 613
Share of MSMEs 4.7% 27.3% 68.0%
Shishu loans
Share of PMMY loans 4.6% 25.0% 70.5%
Share of PMMY loans by MLI type:
State & Nationalized banks 15.8% 26.0% 58.2%
Private banks 2.7% 21.0% 76.3%
Regional Rural banks 0.1% 22.7% 77.2%
Non-banks 4.1% 26.3% 69.6%
Kishor loans
Private banks 10.8% 44.1% 45.1%
Non-banks 25.5% 48.9% 25.6%
Tarun loans
Private banks 20.0% 48.9% 31.2%
Non-banks* 50.2% 34.2% 15.6%
*Non-banks made only 42 Tarun loans in FY17.
Non-banks, particularly MFIs, are known to focus on rural and low-income clients. Accordingly, a
large majority of their Shishu loans were made in Tier 2 or non-tier locations. However, we find
that one-fourth of non-banks’ Kishor loans and one-half of their Tarun loans are concentrated in
Tier 1 locations, but recall from earlier analysis that these loans contribute only 11% of non-
banks’ total disbursed credit in FY17. One possible reason for this could be that most non-banks
find it operationally convenient and profitable to operate within small, economic dense
geographies rather than build a wide branch network. As a result, their disbursements of Kishor
non-JLG loans may be concentrated in a few urban or semi-urban economic centres.
74
Private banks, similar to non-banks, reported a higher proportion of Kishor and Tarun loans in
Tier 1 and Tier 2 locations than the average, but less than 3% of their Shishu loans in Tier 1
locations. This trend is interesting because many Private banks, in partnership with MFIs, have
acquired large rural portfolios of Shishu loans, but remain driven by operational conveniences
and profitability concerns for their larger-ticket loans and as a result, these loans remain
concentrated in Tier 1 and Tier 2 locations. This trend in particular illustrates the effects of co-
origination for smaller loans, and the potential opportunity to use similar arrangements to
transform the lending landscape in non-tier locations.
It is important to reiterate here that concentration of credit is not a cause for concern in itself,
especially if the flow of credit is adequately responding to high demand and an ability to profitably
deploy credit in these regions. Indeed, the estimates by Tier 1, Tier 2 and other cities do suggest
that except for certain institutional deviations, the overall distribution of loans is matched by the
overall distribution of MSMEs and underlying economic activity. However, if the flow of credit in
some regions is heavily influenced by supply-side features (such as existing branch networks or
MLIs’ comfort in underwriting only certain client segments), and if PMMY’s design is insufficient
to over-ride historical inequalities, there is a risk that the scheme will yield only marginal benefits.
The next section furthers the regional analysis by replacing the broad Tier 1/Tier 2 classification
with specific regional demand and supply characteristics, and studies the drivers of both credit
concentration and market saturation.
2.5.2 Regional drivers of credit concentration
This analysis replaces the Tier 1/Tier 2 classification with regional credit demand and supply
characteristics (listed in Table 12), and estimate the responsiveness of credit distribution to these
characteristics. We estimate the relationships as correlation coefficients between, on the one side,
credit concentration or saturation at a district level and on the other side, a set of district
economic and financial characteristics. The choice of characteristics for this analysis is currently
limited by data availability.
Table 12 Description of variables featured in regional analysis

Variable Description
PMMY performance
Concentration Number of PMMY loans in a district divided by all PMMY loans,
calculated from MUDRA administrative data, FY17
Saturation Number of PMMY loans in a district divided by the number of

MSMEs, calculated from Sixth Economic Census 2013-14
Distribution of enterprises
Share of MSMEs Number of MSMEs in a district divided by all MSMEs
Share of GVA Total GVA by MSMEs in a district divided by all-India MSME

GVA. GVA estimates obtained from NSSO 67 Survey of
Unincorporated Non-Agricultural Enterprises, 2010-11
Financial sector development
CRISIL Inclusix score Obtained from CRISIL Inclusix Report Vol. 3, 2015
Access indicators
MSME access to formal loans Incidence of indebtedness for subset of MSMEs, calculated
from NSSO 67 Survey of Unincorporated Non-Agricultural
Enterprises, 2010-11
Household access to formal Incidence of indebtedness for households, calculated from

loans NSSO 70 All India Debt and Investment Survey, 2012-13
Institutional loans as major Proportion of MSMEs reporting institutional loans as a major

source of finance source of finance, calculated from the Sixth Economic Census
2013-14
SHG loans as a major source of Proportion of MSMEs reporting SHG loans as a major source
finance of finance, calculated from Sixth Economic Census 2013-14
76
We estimate Spearman correlation coefficients, which estimate the strength direction of
association between two ranked variables. This method is a better fit for non-linear distributions
and less sensitive to outliers, while the use of rankings also moderates the variation in the data.
The results therefore reflect the relationship between the relative ranking of districts along these
dimensions, and not the absolute values themselves.
Box 1. Interpreting Spearman correlation coefficients
As an illustration of how to interpret the coefficients, we visualize examples of high and low
correlation estimates. Figure 19 plots district-level share of Kishor loans on the vertical axis and
the district-level share of MSME profits on the horizontal axis, representing the relationship
between distribution of loans and productive deployment of credit.
Figure 19 Example of high rank correlation
The left panel plots values as-is, while the right panel plots the relative ranking of districts on
each variable. The Spearman coefficient estimates the strength and direction of correlation of the
ranks, as presented in the right panel, and returns an estimate of 0.788, indicating that districts
representing a higher share of MSME profits also received a higher proportion of Kishor loans.
In contrast, Figure 20 presents a relationship with the lowest correlation in our analysis. We
observe a non-linear, non-monotonic relationship in the left panel, while the dispersion is
amplified even further when presented as ranks in the right panel. It is evident that there is no
consistent relationship between the variables, and the same is also reflected in the magnitude of
the coefficient (0.086).
Figure 20 Example of low rank correlation
We analyse rank correlations between a range of regional characteristics and both concentration
and saturation measures. Table 13 reports estimates with concentration and saturation by
Shishu, Kishor and Tarun loans respectively, while Table 14 reports estimates of concentration
disaggregated by MLI type. Estimates of market saturation are not available by MLI type. Each
estimate is accompanied by the p-value and where p-values are lower than 0.05, estimates are
considered statistically significant. However, we only consider relationships as economically
significant if their coefficient is greater than 0.4, having observed that the distribution is too
dispersed for coefficients of a lower magnitude.
78
The first result from Table 13 is that the concentration of loans of all sizes (Shishu, Kishor or
Tarun) is highly and positively correlated with the presence and profitability of MSMEs. For
example, the correlation at the district-level between the share of Shishu loans and the share of
MSMEs was high at 0.669 (and higher for Kishor and Tarun loans), and similarly high when
considering share of MSME profits instead. These coefficients are the highest in the table, and we
interpret this as a positive indication of productive disbursement of credit overall.

characteristics
Concentration Saturation
Share of MSMEs 0.669 0.780 0.763 0.161 -0.086 -0.164
0.00 0.00 0.00 0.00 0.04 0.00
Share of GVA 0.602 0.788 0.787 0.252 0.228 0.114
0.00 0.00 0.00 0.00 0.00 0.01
CRISIL Inclusix score 0.370 0.572 0.551 0.189 0.428 0.224
0.00 0.00 0.00 0.00 0.00 0.00
Access indicators
MSME access to formal loans 0.213 0.390 0.333 0.087 0.387 0.219
0.00 0.00 0.00 0.04 0.00 0.00
Household access to formal loans 0.344 0.393 0.290 0.225 0.194 -0.059
0.00 0.00 0.00 0.00 0.00 0.15
Instl. loans as major source of finance 0.086 0.175 0.135 0.113 0.298 0.296
0.04 0.00 0.00 0.01 0.00 0.00
SHG loans as a major source of finance 0.375 0.261 0.168 0.374 0.199 0.052
0.00 0.00 0.00 0.00 0.00 0.20
This coefficient is relatively smaller for Shishu loans than for Kishor and Tarun but this is
unsurprising, given that many Shishu loans may be used for only semi-productive purposes, or
by home-based enterprises not counted in MSME surveys. However, looking at the saturation
measure instead of concentration produces very different results. The overall magnitude of
effects drops and while statistically significant, the effects remain too small to be economically
significant. In other words, while the presence of profitable economic activity is positively
correlated with the flow of credit, even districts with a high concentration of profitable activity
are hardly saturated.
The second striking result from Table 13 is the persistence of supply-side features in relation to
patterns in credit disbursement. We report correlations with the CRISIL Inclusix score, which is
a composite index of branch networks, savings penetration and credit penetration of SCBs and
MFIs. On credit concentration, we find that districts with a higher Inclusix score received a higher
share of Kishor and Tarun loans (0.572 and 0.551 respectively).
In both cases we observe that the correlations with credit concentration are significant, but with
saturation are (economically) insignificant. While districts with a higher number of MSMEs or a
better banking network are able to attract more loans, these features are yet inadequate to ensure
a sufficient amount of credit for all enterprises in the region. In fact, this result is indicative of
large market gaps that exist even in regions that might be thought of as financially or economically
well-developed. This suggests room for increasing credit saturation even in top economic centres
alongside, of course, lesser developed districts which are often the focus of inclusion efforts.
Table 13 also reports correlation coefficients with historical access indicators from the Sixth
Economic Census and the NSS 67th Round Survey of Non-Agricultural enterprises. Several of
these correlations are statistically significant with both concentration and saturation measures,
but their effect sizes are too small to be economically significant26 to be interpreted as a strong
driver of PMMY loan trends.
For comparison, we also report pooled estimates for each loan size (across all MLI types) from
Table 13, allowing us to learn which institutions are driving higher correlations. For example,
among the strongest effects in Table 13 was the high correlation between districts’ share of PMMY
loans and districts’ share of MSMEs and MSME profits. Table 14 reports correlations of regional
characteristics with credit concentration by MLI type and loan size. From Table 14 we learn that
the effect is strongest State & Nationalized banks (0.687 and 0.665 for Shishu loans and higher
for Kishor and Tarun), and weakest for Regional Rural Banks (at 0.47 and 0.318 for Shishu loans).
This is perhaps because State & Nationalized banks have widespread branch networks and the
ability to move resources toward more profitable branches, while RRBs lack this flexibility and
must serve the regions they are located in.
Similarly, decomposing the correlations with CRISIL Inclusix also reveals interesting patterns by
MLI type. In the Shishu segment, only State & Nationalized loans are (economic) significant in
relation to district-wise Inclusix scores, while private banks, RRBs and non-banks are not. Private
banks in particular may be unresponsive to the strength of branch networks that Inclusix
represents owing to their determined focus on growing exposure to group-based Shishu loans
either directly or through co-origination partnerships, since we also note that their
26
Statistically significant but economically not significant effects can be better understood using multivariate analysis robust to
the endogeneity between many of these variables.
80
corresponding coefficients for Kishor and Tarun loans are significant. Recall that this trend is
similar to the results observed by the Tier 1/Tier 2 classifications as well.
Overall, this analysis of regional variation and correlation estimates underscores the complexity
of addressing regional inequalities. While it is heartening to observe that the most dominant
effect on credit concentration is the underlying distribution of economic opportunities
themselves, there is also evidence that lending by atleast some MLI types systematically
reinforces existing patterns of regional access and inequality. Determined efforts such as Private
banks’ focus on microfinance Shishu loans act to significantly weaken self-perpetuating patterns
in inequality. While a more robust analysis of these drivers could reveal tremendous insight, we
find it is beyond the scope of this study.
Linking to the theory of change in Figure 2, we apply the key insight that the distribution of PMMY
loans is heavily determined by the concentration of economic activity and the historical ease of
credit delivery at the district level and therefore, regions without these will likely remain
underserved in the absence of corrective practitioner or policy actions.
Table 14 Correlation between measures of regional distribution in PMMY with regional characteristics, by MLI type
Shishu Kishor
Non- Non-
All S&N Private RRBs All S&N Private RRBs
banks banks
Share of MSMEs 0.669 0.687 0.589 0.470 0.598 0.780 0.799 0.586 0.528 0.547
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Share of GVA 0.602 0.665 0.525 0.318 0.559 0.788 0.772 0.738 0.461 0.545
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
CRISIL Inclusix score 0.370 0.467 0.232 0.136 0.339 0.572 0.570 0.547 0.344 0.380
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Access indicators
MSME access to formal loans 0.213 0.229 0.086 -0.021 0.230 0.390 0.368 0.408 0.185 0.165
0.00 0.00 0.04 0.62 0.00 0.00 0.00 0.00 0.00 0.00
Household access to formal loans 0.344 0.368 0.218 0.448 0.323 0.393 0.430 0.239 0.308 0.168
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Instl. loans as major source of finance 0.086 -0.038 0.076 0.004 0.056 0.175 0.120 0.160 0.096 -0.062
0.04 0.36 0.07 0.93 0.17 0.00 0.00 0.00 0.02 0.13
SHG loans as a major source of finance 0.375 0.151 0.284 0.213 0.358 0.261 0.263 0.170 0.204 0.131
0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
82
Tarun
Non-
All S&N Private RRBs
banks
Share of MSMEs 0.763 0.743 0.626 0.387 0.309
0.00 0.00 0.00 0.00 0.00
Share of GVA 0.787 0.747 0.754 0.350 0.369
0.00 0.00 0.00 0.00 0.00
CRISIL Inclusix score 0.551 0.506 0.586 0.250 0.224
0.00 0.00 0.00 0.00 0.00
Access indicators
MSME access to formal loans 0.333 0.286 0.403 0.172 0.098
0.00 0.00 0.00 0.00 0.02
Household access to formal loans 0.290 0.286 0.249 0.193 0.080
0.00 0.00 0.00 0.00 0.05
Instl. loans as major source of finance 0.135 0.114 0.116 0.208 -0.001
0.00 0.01 0.00 0.00 0.99
SHG loans as a major source of finance 0.168 0.183 0.070 0.239 0.119
0.00 0.00 0.09 0.00 0.00
3. RECOMMENDATIONS
3.1 Framework for evaluation
In Chapter 1, we outlined a hypothetical ‘theory of change’ framework that links policy inputs to
immediate effects and eventual long-term outcomes. Such a framework can serve as the
foundation for ongoing scheme supervision, target-setting and administrative data collection. In
this basic framework, we identified six immediate scheme-specific effects (higher disbursements
quicker loan processing, lower interest rates and new client acquisition) and five corresponding
financial sector outcomes potentially arising from successful scheme implementation. Broadly,
the set of desired financial sector outcomes from PMMY aims to achieve either service
additionality (such as the inclusion of formerly excluded firms, or relieving credit constraints and
enabling full service) or improvements in service quality (such as lower borrowing costs, quicker
processing time and better-matched products). At this time, we have focused the list of outcomes
to financial sector changes and excluded hypotheses on the effects on labour markets or
aggregate economic activity.
The primary objective of disaggregating immediate effects from desired policy outcomes is to
generate early insights on likely outcomes, identify gaps in design or implementation and to
inform early-stage design improvements. We recommend using various available administrative
data collection methods to continuously track scheme performance based on the set of
intermediate scheme effects, while also setting up nationally-representative and periodic data
collection efforts to measure progress on desired outcomes. Ideally, data on intermediate effects
should be available directly from MLIs, on an annual or more frequent basis. This could be a direct
input into scheme management and ongoing supervision. On the other hand, indicators of desired
financial sector outcomes or changes are better captured through nationally representative
surveys directly administered to small firms on a 3-5 year basis.
Monitoring intermediate effects
With regard to intermediate effects, we learned that MUDRA has built an extensive network with
MLIs for administrative data collection and at the current time, this includes data on number of
loan accounts in each category, amounts disbursed several social inclusion tags and a
rudimentary (although not robust) financial inclusion tag. Through this network, MUDRA
aggregates timely, accurate and relevant data on PMMY performance, and this data serves as a
critical input to ongoing monitoring and supervision. We recommend expanding the scope of
MUDRA’s administrative data to include:
• Broad categories of loan product type and loan features that may be accurately measured
from MLIs, such as whether the loan was extended as a term loan or flexible working
capital product, whether it was originated directly or co-originated through partnership
agreements, loan tenure, and interest rate.
• A standardized “new entrepreneurs/new accounts” tag for all MLIs.
Measuring performance on outcomes
It would be important to periodically review the impact of the scheme on the financial landscape
as a whole, and specific financial sector outcomes such as increased credit depth for small
enterprises, easier access to enterprise credit and increased formalization of enterprise financing.
To this end, we reviewed available public data sources and identified three national surveys that
generate rich and representative data on small enterprises:
• The Sixth Economic Census surveys all enterprises in India and was last conducted in
2012-13. While a rich source of data on the nature and activities of enterprises, it only
provides data on formal credit if it forms a major source of finance for enterprises and
thus any measure of access computed from this database proves to be a significant under-
estimate.
• The Ministry of Micro, Small and Medium Enterprises occasionally commissions a census
of registered and unregistered enterprises in the country. Ensuring that these census
estimates are comparable with the Economic Censuses, and including basic indicators of
financial access in both surveys will help to generate reliable estimates of headline
indicators.
• NSSO conducts a period Survey of Unincorporated Non-Agricultural Enterprises,
collecting detailed data on enterprises operational and financial characteristics. This
dataset is extremely detailed, and we recommend a comparison of key indicators over the
two most recent rounds (2010-11 and 2015-16) to study structural changes in the MSME
landscape and the effect of financial sector expansion.
Updating extensive and representative databases periodically could serve not only as a well-
established reference for trends and progress in enterprise credit (similar to the NSSO All India
Debt & Investment Survey is for household access to credit), but also as a means to reliably verify
and robustly attribute any accompanying changes in the labour market, occupational choice or
overall economic activity to specific interventions in enterprise credit. In the absence of reliable
data sources, it would be impossible to measure the broader equilibrium effects of a large credit
policy intervention such as PMMY.
3.2 Guidelines for high-quality origination
Moving on to the specifics of loan origination, our analysis reveals that the PMMY portfolio
inherits specific features from the overall patterns of credit distribution in India, including
regional inequalities, lender specialization and market fragmentation. Special efforts will
therefore be required to enlist and encourage lenders willing and able to reverse these trends
and bridge gaps in access.
3.2.1 Calibrate Refinance and Credit Guarantee to increase take-up
As incentives to boost MSME lending, it is also relevant to consider whether the current design of
policy levers (particularly refinance and credit guarantee) are suitable for and well-aligned with
the needs of high-performing MLIs, and whether they are effective in encouraging more lending
to target segments.
MUDRA financing (both as direct refinance or purchase of securitized assets, or as second-loss

guarantee provision to private-sector transactions) could have a tremendous impact on
improving the terms of finance for MLIs and eventually, the flow of credit to MSMEs. MUDRA
participation would also serve to increase market confidence in MSME loans as an asset class.
The credit guarantee (CGFMU) provides an efficient mechanism to manage credit risk, especially
for non-bank MLIs who tend to have higher concentration risk. However, we learn that take-up is
low, and that the costs of managing credit risk continue to remain embedded in loan APRs,
subsidized by the low cost of funds (in the case of banks) or absorbed by high capital reserves (in
the case of NBFCs). MFI-NBFCs in particular may be reluctant to participate in CGFMU given that
the current guarantee fee (around 1%) is higher than their credit losses in an average year.
However, this comparison ignores the destabilizing effects of higher-than-average credit losses
in the years that they do occur (say, in 2017 following demonetization or in 2011 following the
AP government ordinance), as well as the additional costs of equity incurred in shoring up capital
reserves to absorb risk.
We recommend that both MUDRA financing and CGFMU be viewed as interventions that could
potentially improve the rating of MSME lenders or MSME loan-backed assets, and any pricing or
feature considerations for both should incorporate insights from credit rating agencies with
sector expertise. For example, rating agencies are well-equipped to predict underlying risk in
each of these asset classes and to inform the ideal level of first-loss and second-loss default
guarantee thresholds in CGFMU. Second, rating agencies’ insights would be helpful to quantify the
equity implications for MLIs participating in CGFMU, i.e. if participation in a well-functioning
guarantee arrangement could deliver a higher rating for their assets at a lower capital ratio. We
believe that a re-calibration along these lines might be helpful to appropriately design and
increase take-up of these two important features of PMMY.
3.2.2 Including specialist originators while ensuring a balanced product

suite
In the analysis we learned that NBFC-MFIs were highly specialized— with a sharp focus on the
smallest loans— and played an important role in disrupting regional inequalities to credit access.
Recall that in Table 10, Shishu loans reported the lowest correlations between credit
concentration and historical indicators of institutional access to enterprise credit. These were
also the loans with the lowest concentration in districts with a higher share of MSMEs, suggesting
more widespread outreach. Similarly, we also learned that non-MFI NBFCs played a small but
significant role in originating loans in the Kishor and Tarun category, in comparison to banks. We
therefore recommend identifying and enrolling more specialist lenders under PMMY.
These include institutions who enjoy a sectoral advantage through the strength of their
experience or size of their networks as well as institutions developing new, technology-focused
underwriting and delivery capability. Further, it would also be important to enrol institutions
offering an adequate and comprehensive suite of products, including working capital credit,
vehicle finance and large equipment financing. The purpose of broader enrolment would be to
extend refinance, credit guarantee and any other facilities to all MLIs in a manner that would
encourage them to reach out to under-served enterprises.
3.2.3 Promote co-origination of MSME loans
Alongside enrolling new MLIs, MUDRA could also consider promoting co-origination
arrangements between MLIs, which would serve both to expand outreach and increase lending
efficiency. For example, MFIs have developed strong capabilities for client sourcing and loan
management, but often face significant constraints raising funds and making loans that do not
meet standard PSL eligibility criteria. For this and other reasons, many MFIs partner with SCBs
to originate loans on their behalf, for a fee. In our analysis, we find evidence that such
partnerships played a large role in supporting Private banks’ determined expansion in the Shishu
segment, and we believe that given opportunities and incentives, similar partnerships with
NBFCs could expand lending in Kishor and Tarun segments too.
3.2.4 Eliminate known barriers to graduation
These recommendations in this section relate to specific practices in loan underwriting we

learned about during qualitative interviews with frontline loan officers, and we believe small
changes in the lending guidelines could serve to improve customer experiences under PMMY.
First, we learned that many banks rely heavily on the CIBIL score as a primary selection
mechanism of borrower eligibility, unlike MFIs who rely more on the verification of outstanding
loans from either or all credit bureaus. This particular requirement for a CIBIL score proves quite
exclusionary for even mature MFI clients who have rich histories with MFI bureaus, but remain
unscored on CIBIL. While the RBI has now required complete integration of retail and
microfinance credit histories, this process is still underway and likely to take several years to
complete. In the interim, equipping banks’ frontline loan staff to use and interpret scores
Highmark and Equifax in addition to CIBIL could expand the universe of eligible borrowers. Banks
of course would remain free to determine appropriate cut-offs on each score as per their
underwriting norms, but this small step would ensure that firms not scored on CIBIL are given
the opportunity to apply for PMMY loans through banks at lower interest rates.
The second recommendation is related to the long turnaround times for loan processing reported
both by frontline loan officers at banks as well as MSME borrowers as clients of banks. Most banks
reported following relaxed underwriting norms for Shishu and smaller Kishor loans, but similar
processes as other loans, thus contributing to extended processing times. However, some
reported the use of a fast-track processing system, where MUDRA loan applications were tagged
for faster movement through internal systems. To the extent feasible, we encourage all banks to
follow a similar system, especially when approval processes can be streamlined or certain skips
stepped as waived by underwriting norms for small loans.
3.3 Suggestions for additional MIS formats

Our third set of recommendations relate to specific aspects of administrative data collection and
ongoing performance analysis by MUDRA.
3.3.1 Since the set of participating MLIs under PMMY will grow rapidly in the initial years, the
inclusion of data from new players inflates aggregate totals year-on-year and overestimates the
growth of lending in the economy. Similar to the analysis of administrative data in Table 1, we
recommend tracking portfolio performance for a fixed cohort of MLIs (in addition to portfolio
aggregates). This adjustment is necessary for more accurate estimates of changes in lending, and
allows the decomposition of growth as ‘organic’ (i.e. growth driven by participating MLIs) and
‘inorganic’ growth driven by the enrolment of new MLIs into the scheme.
3.3.2 We also recommend the incorporation of certain analysis formats into ongoing scheme
supervision that reveal more granular trends across both institutional and regional dimensions.
On the institutional side, we recommend tracking individual MLIs’ portfolio composition over
time and year-on-year change to detect systematic strategic shifts (corresponding to Table 9 and
Table 10).On the regional side, we recommend mapping district-level credit concentration,
market saturation and credit depth. Some MLIs (non-MFI NBFCs and non-NBFC MFIs) are yet to
report disaggregated data at the district-level and therefore, these reporting formats may need to
be incorporated in a phased manner.
3.3.3 The analysis reveals that the Kishor category comprises a wide variety of loans and loan
sizes, including individual personal or business loans as small as Rs. 60,000 from MFIs as well as
MSME loans of upto Rs. 5 lakhs from banks. As a result, tracking aggregate performance of
institutions under the Kishor category produces less meaningful insights, and ceases to
differentiate between lenders specializing at the edge of these ranges. For the purpose of
administrative reporting alone, we recommend splitting the Kishor category into loans from Rs.
50,000 to Rs. 2 lakhs and loans between Rs. 2-5 lakhs respectively.
3.3.4 Our final recommendation is with respect to harmonizing the definitions of loan tags as per
express guidelines. For example, the loan-level tag “MUDRA card” is to represent loans with a
linked MUDRA card or other debit cards, and not only newly issued MUDRA cards in that year.
This estimates the proportion of loans accessible through linked cards, and will be a critical input
to decision-making around this feature of the PMMY scheme. Similarly, we also recommend
standardizing the definition of the “new accounts” tag, perhaps by linking this to specific fields27
in customer credit reports themselves, rather than as discretionary input by the loan officers.
3.4 Proposals for further research

While the scope of this report was limited by a set of initial objectives, in the course of study we
were able to develop a comprehensive understanding of PMMY policy design challenges and
related academic literature on access to financial markets and enterprise growth, as well as the
strengths and weaknesses of available data sources on this subject in the Indian context. We
leverage this understanding to propose ideas for continued policy research that could support
and inform critical design improvements and enhanced supervision of PMMY.
27The credit report will return a unique value for loan applicants without a previous formal credit history,
or without a currently outstanding loan etc.
The following is a preliminary list of research questions.
3.4.1 Estimating inclusion and graduation using credit bureau records
Currently, the policy conversation on financial inclusion of MSMEs and the additionality achieved
through PMMY revolves around the “New Accounts” tag in the MUDRA data, which is known to
be reported inconsistently. We propose instead, as a one-time exercise, to leverage the strength
of credit bureaus to arrive at a robust estimate of inclusion and client graduation.
This would require anonymized but comprehensive credit histories of a randomly selected of
sample of PMMY borrowers, preferably drawn from a representative set of MLIs. This data could
then be used to identity the proportion of loans made to first-time formal borrowers as well the
proportion of loans made by banks to borrowers who were earlier served only by non-banks. The
latter represents client graduation into a lower interest rate.
3.4.2 A study of MSME credit depth to identify under-served districts
The regional analysis in this report was limited to measuring market saturation in terms of the
number of enterprises due to the unavailability of estimates valuing the district MSME GDP.
However, the NSSO 73rd round Survey of Unincorporated Non-Agricultural Enterprises 2015-16
will soon release unit-level data on the operational and financial characteristics of a
representative sample of MSMEs.
Kumar and Baby (2016) 28 argue that stark inequalities in credit access—measured through
district-level credit depth— can have a cascading effect on economic growth and inequalities over
time. We propose a replication of their exploratory analysis at first, with a specific focus on
understanding the variation in MSME sector credit depth by comparing estimates of MSME credit
through PMMY with estimates of MSME Gross Value Added from NSSO.
3.4.3 Understanding the effects of financial access on the MSME landscape
Leveraging new data available from the NSSO 73rd round Survey of Unincorporated Non-
Agricultural Enterprises on the operational and financial characteristics of MSMEs collected in
2015-16 and previously in 2010-11. A comparative analysis of both rounds of data could offer
insight into the changing nature of MSME operations and in particular, the effects of expanding
financial access on their investments, risk-taking and profitability.
28
Kumar and Baby (2016), “Determining Optimal Credit Depth Allocation at a District Level”, IFMR Finance Foundation Working
Paper series. Available at: www.bit.ly/creditdepth
3.4.4 Calibrating the design of CGFMU to improve take-up by all MLIs
The fourth proposal is related to understanding the underlying risk in Non-banks enterprise
lending portfolios with a view to informing the calibration of the credit guarantee facility. This
exercise should be undertaken in collaboration with the National Credit Guarantee Trust
Corporation and, in addition to the analysis of risk experiences, must also review market
prevalent forms of refinance and credit guarantee currently available to MLIs, and arrive at terms
that are both appropriate and competitive

Assessment of PMMY Report

Uploaded by

Copyright:

Available Formats

Assessment of PMMY Report

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Assessment of PMMY Report

Uploaded by

Copyright:

Available Formats

Research insights on design and implementation

Version: Final Version, April 2018

1.1 Policy design and theory of change ........................................................................................................... 11

2.1 Macroeconomic trends in MSME lending ................................................................................................ 19

2.2.2 Product range and underwriting rules ............................................................................................ 32

2.3 Features and uses of PMMY loans .............................................................................................................. 36

2.4 Lender specialization and market fragmentation ............................................................................... 54

2.4.2 Implications for origination and market structure..................................................................... 68

2.5 Regional patterns in the PMMY portfolio ................................................................................................ 71

2.5.2 Regional drivers of credit concentration ........................................................................................ 76

3.1 Framework for evaluation ............................................................................................................................. 84

We gratefully acknowledge the patient and enthusiastic participation of survey respondents—

ABOUT THE AUTHORS

IFMR LEAD is a non-profit research organization conducting

Table 1 Market share of various MLI types in PMMY portfolio .................................................................. 28

Table 2 Product range of MLIs interviewed ....................................................................................................... 34

Table 3 Key characteristics of PMMY loans from MUDRA data.................................................................. 39

Table 4 Sample composition for survey of MSME borrowers..................................................................... 44

Table 5 Key characteristics of loans from primary survey .......................................................................... 45

Table 6 Characteristics of borrowers’ enterprises, from primary survey ............................................. 48

Table 7 Characteristics of borrowers’ households, from primary survey ............................................. 49

Table 8 Customer experience with loan processing and service ............................................................... 52

Table 9 Composition of MLIs’ portfolios by loan size..................................................................................... 57

Table 10 YOY changes in MLIs’ portfolio composition ................................................................................. 63

Table 11 Tier 1 and Tier 2 credit concentration by MLI type ..................................................................... 74

Table 12 Description of variables featured in regional analysis................................................................ 76

Table 13 Correlation between measures of regional distribution in PMMY with regional

Table 14 Correlation between measures of regional distribution in PMMY with regional

Figure 1 MSMEs perceptions of key obstacles to their growth .................................................................. 10

Figure 2 Theory of change for PMMY .................................................................................................................... 12

Figure 5 Trends in GLP and Off-B/S portfolio of NBFC-MFIs ...................................................................... 21

Figure 6 Trends in Off-B/S origination of all microfinance institutions ................................................. 21

Figure 7 Share in PMMY portfolio by MLI type (Accounts) ......................................................................... 24

Figure 8 Share in PMMY portfolio by MLI type (Disbursed amount) ...................................................... 24

Figure 11 Multiple uses of loans by borrowers ................................................................................................ 46

Figure 12 Median expenses in average, high and low months ................................................................... 51

Figure 13 Median revenues in average, high and low months ................................................................... 51

Figure 14 Profitability of sample enterprises .................................................................................................... 52

Figure 17 Regional disparities in access to PMMY loans .............................................................................. 72

Figure 19 Example of high rank correlation ...................................................................................................... 77

Figure 20 Example of low rank correlation ........................................................................................................ 78

An overwhelming majority of unincorporated enterprises were small and informal—96% of them

Inadequate market linkages

Delay in govt. clearances

0% 20% 40% 60% 80%

Scheme features, effects and outcomes

This framework is particularly helpful to trace the dependencies of desired outcomes on

The study objectives are:

The themes in each section are:

Upto Rs. 25k Rs. 25k- 2L Rs. 2- 5L Rs. 5- 10L

Figure 4 Total outstanding on MSME loans upto Rs. 10 lakhs by SCBs

Figure 5 Trends in GLP and Off-B/S portfolio of NBFC-MFIs

GLP GLP (Non-agri enterprise loans) % GLP Off-Balance sheet

Figure 6 Trends in Off-B/S origination of all microfinance institutions

2.2.1 Market share and growth trends

450 Non-banks RRBs SCBs

Figure 8 Share in PMMY portfolio by MLI type (Disbursed amount)

200,000 Non-banks RRBs SCBs

Shishu Kishor Tarun