Multiple Borrowing - PowerPoint Presentation

Slide1

Multiple Borrowing among Microfinance ClientsResults from an Area Study

Prepared by Ronald T. Chua and Erwin R. TiongsonJuly 2012Slide2

Background and introductionOverview and Research ObjectivesSlide3

IntroductionThis PPT provides a summary of the findings from a study of multiple borrowing in an urban community.

As a pilot effort that draws information from a variety of sources, it potentially enriches our understanding of the dimensions, the prevalence, and correlates of multiple borrowing. As a pilot effort, however, it is also suffers from several technical weaknesses. As explained in the following slides, the results should be interpreted with caution.An accompanying report describes details of the methodology and the analyses conducted.Slide4

Background: Notes from 2009-2011 consultations

Multiple borrowing = “Over-indebtedness”“Over-borrowing”“Multiple indebtedness”“Loan recycling”“Client poaching”

“Loan pushing”“Credit pollution”“Over-supply of loans"Slide5

Background: Notes from 2009-2011 consultationsThere are several

dimensions of multiple borrowing Different notions of “multiple borrowing”Borrowing from multiple MFIs? Or from multiple sources, both formal and informal? Or multiple loans from the same MFI? All of the above?The incidence

of multiple borrowing is not clearHow prevalent is multiple borrowing?The causes

and consequences also unclearOn one hand

Loans used to pay off other loansMultiple borrowing may lead to rising levels of delinquency and defaultOn the other hand

Households may need multiple sources of credit Households may be perfectly able to manage their financesSlide6

Research ObjectivesClarify dimensions

of multiple borrowingMultiple borrowing across MFIs, within MFIs, across loan sources (MFI, non-MFI, informal) Estimate the incidence of multiple borrowingMultiple borrowers in percent of MFI clients, in percent of all households, etc.

Understand the correlates of multiple borrowingStatistical profile

Repayment recordSlide7

Research designData Sources: Key FeaturesSlide8

Research Design Commonwealth area of Quezon City as the area of study, including the following

barangays: Commonwealth, Payatas, Holy Spirit, Bagong Silangan, Fairview, Batasan HillsBased on a series of consultations in 2009-2011

Several MFIs are known to operate in this communityAnecdotal evidence suggested the existence of multiple borrowingSlide9

Sources: Google Maps; TNS-Global Methodological ReportSlide10

Research Design Technical considerationsCredibility

of household-level information: How reliable are self-reported levels of indebtedness?The sources of multiple loans: Should we include both formal and informal sources of finance? The unit of measurement: How do we treat individuals in the same household?Slide11

Research Design Sources of Information

Household SurveyRandom sample, geographically representative800 householdsPatterned after FIES, APIS, and credit modulesComprehensive credit information: MFI, bank, informal, government, etc.

Branch Client Data from MFIs Operating in the Commonwealth Area

Consolidated information using search algorithm based on name and address and/or birth dateFocus Group Discussions

(FGDs)Slide12

Research Design Sources of Information Dimensions of Multiple Borrowing

Borrowing from several MFIs (branch client data)Multiple loans from the same MFI (branch client data)Multiple loans from several sources of finance, including formal and informal sources (household survey)Sequential versus simultaneous multiple borrowing (FGD)Slide13

Research Design Sources of Information:

CaveatThey are analyzed separately and cannot be consolidated.They vary in representativeness

e.g., FGD participants not representative of general populationThey vary in scope and

comprehensivenesse.g., branch client data do not provide information on other sources of financing

They are subject to errorErrors include misreporting, sampling errors, encoding errors, etc.

They are cross-sectional (from a single point in time) and provide little information about

dynamics

(i.e., how things evolve over time) and

causality

(e.g., if multiple borrowing is correlated with financial distress: did multiple borrowing cause distress or was it distress that initially led to multiple borrowing?)Slide14

Research Design Additional features: Branch

Client DataAs agreed with partners, sources of branch client data are not publicly revealedRestricted access to branch client data (principal researchers)Slide15

Selected Research findings1. Household SurveySlide16

Household Survey: Selected Findings

57 percent of households: at least one outstanding loan

Only 6.8 percent of households are MFI clientsSlide17

Household Survey: Selected Findings

This chart refers to all households. Of the MFI client households alone, 77 percent are “multiple borrowers”,

all sources of finance considered.Slide18

Household Survey: Selected Findings

These charts show multiple borrowers by number of loans (

all sources of finance considered) For multiple borrowers

as a group, 52 percent have 3 or more loans. For multiple borrowers MFI clients as a group, 71 percent have 3 or more loansSlide19

Household Survey: Selected FindingsSlide20

Household Survey: Selected Findings

How Many Times Did You Miss Payments The Last 3 Months?

There is no evidence that multiple borrowing is associated with higher delinquency, compared with average indebted household.

Caveat: based on self-reported information

This is from cross-sectional

information or information from a single point in timeSlide21

Selected Cross-Checks

Source: Commonwealth survey; World Bank Global Findex database; authors’ calculations.

Compare with the Consumer Finance Survey:

4 percent of households own credit cardsSlide22

Selected Research findings2. Branch Client DataSlide23

The Framework

MFI1Database Client 1Client 2

…Client 2000

MFI2

Database

MFI3

Database

MFI4

Database

MFI5

Database

MFI6

Database

MFI7

Database

MFI8

Database

MFI9

Database

MFI10

Database

MFI11

Database

MFI12

Database

MFI13

Database Slide24

Branch Client Data: Selected Results

14 percent

Incidence of Multiple Borrowing

On average, 14 percent of MFI clients

Substantial variation across MFIs (4-26 percent)Slide25

Branch Client Data: Selected ResultsIncidence: Within-MFI Multiple Borrowing“Within-MFI” multiple borrowing also exists, i.e., multiple loans from the same MFI

Most cases appear to be known to the MFISome may not be known (borrowing from several branches)On average, where they exist, within-MFI multiple borrowers represent 5 percent of all MFI clientsSome overlap with multiple borrowing as defined previously (14 percent)Slide26

Branch Client Data: Selected Results

Number of LoansMost multiple borrowers have loans from only 2 MFIsAbout 15 percent of multiple borrowers have loans from 3 or more MFIsSlide27

Branch Client Data: Selected ResultsShare of Outstanding LoansLoan sizes tend to be

uniform. As a result, the share of loans accounted for by multiple borrowers tend to be proportional to the incidence of multiple borrowing (14 percent)However, among multiple borrowers who also take out multiple loans from the same MFI (within-MFI multiple borrowing), their share of all outstanding 2nd

or 3rd loans within the same MFI tends to be large (40 percent).Slide28

Branch Client Data: Selected ResultsCorrelatesIn general, no strong evidence linking multiple borrowing to specific lengths of membership, loan cycles, client age, and other demographic patterns.

There are some suggestive patterns. Multiple borrowers with two or three loans tend to be on longer loan cycles. Multiple borrowers also tend to be somewhat olderSlide29

Branch Client Data: Selected ResultsDelinquencyThere is no evidence that multiple borrowing is associated with delinquent payment, at least not among the active clients.

This should be interpreted with caution. This indicates the average observable outcome to date and does not at all address the possible impact of economic or income shocks and whether multiple borrowers can fall into delinquency as a result. There are large differences in the availability of information on missing payments. The data are not recorded consistently across MFIs.May reflect business opportunities in a particular urban community. The economics may differ in other communities, including rural communities.Slide30

Selected Research findings3. Focus Group DiscussionsSlide31

FGDs: Selected FindingsParticipants consisted of:

MFI members known or who have admitted to having loans from at least one other MFIBranch staffMFI clients were selected by MFI staff. Only criterion was that they have borrowed or have current loans from at least two MFIsEach group consisted of 8 to 12 discussantsSlide32

FGDs: Selected FindingsProfiles of FGD participants

A number of participants are center chiefs or group leaders in the MFIsHave had many years of borrowing experienceLong time residents of Commonwealth (one had been resident in area for 30 years)Borrowing behavior reported

Borrowed sequentially from several MFIs Borrowed simultaneously several MFIsSlide33

FGDs: Selected FindingsOn proliferation of lending sources

Clients observed that there has been an increase in the number of lenders in the area recently, as recent as the past year (2010-2011)Clients appreciated the availability of more choicesSome cited concerns about their own inability to repay so expressed preference to stay with one MFIReasons for borrowing from multiple sources

To maximize benefits. No one single MFI offers a whole range ofObtain sufficiently large consolidated loan from

several MFIs as one MFI loan is not sufficient for their needsTo meet emergency needs

To try out other sourcesTo stagger payments (multiple loans from the same MFI are paid on the same day; multiple loans from multiple loans can be staggered)Slide34

SummaryThis is a heroic attempt at a summarySlide35

CaveatThis is not nationally representativeThis is from cross-sectional data that mask the changes over time

The data are subject to errorData are from multiple sources and not fully consistentSlide36

Summary and Concluding ObservationsMultiple borrowing exists, it exists in various forms, and is not small. Though the data sources

cannot be fully consolidated, they suggest the following incidence, along several dimensions:5 percent if multiple borrowing means taking out multiple loans from the same MFI14 percent if multiple borrowing means borrowing from several MFIs, though there is substantial variation across MFIs77 percent if multiple borrowing means borrowing from one MFI as well as from any other source of financeSlide37

Summary and Concluding ObservationsMultiple borrowers take out multiple loans for a variety of reasons. Some report that a single loan is insufficient to pay for a major consumption expenditure or to invest in a business activity. Some are after the auxiliary services attached to various loans.

We are unable to find a distinctive statistical profile of multiple borrowers. There is some evidence to suggest that they are older and are on longer loan cycles, but otherwise there is (to date) no strong statistical links between multiple borrowing and individual demographic characteristics as well as characteristics of the loan itself (size, loan cycle, reported loan use). Slide38

Summary and Concluding ObservationsThere is no evidence that multiple borrowing is associated with delinquent payment, at least not among the active clients.

This should be interpreted with caution. This indicates the average observable outcome to date. We are unable to say anything meaningful about the likelihood of falling into delinquency. There are large differences in the availability of information on missing payments and the data are not fully consistent across MFIs. Furthermore, the sample represents a particular urban community. It is not clear whether this relationship holds more generally.