This paper reviews various analysis conducted in order to
obtain a relationship between student loans and loan defaults furthermore
suggest ways to minimise the default risk on these loans. Empirical research conducted
by scholars suggest that education loan defaults are mainly influenced by various factors such as security, borrower
margin, and repayment periods. On the other hand, the presence of guarantor or co-borrower and
collateral significantly reduce default loss rates. Other determinants of the
default risk could be socioeconomic characteristics of borrowers and their
regional locations which act as important factors associated with education
loan defaults. In order to mitigate risk, banks can follow certain pricing
strategies that have been suggested by various empirical analysis that involve
segmentation of borrowers on various characteristics.
Our main objective is to study the performance of loans over
time and identify key risk factors of such loans across various geographies and
Banks act as a backbone of an economy, they play an
important role in promotion of education and skill development. In an emerging
economy such as india, education loans are essential for pursuing higher
education. An education loan scheme provides financial assistance of Rs 1
million for pursuing higher education and Rs 2 million for studies abroad. The
interest rate is around 13%, depending upon the amount of the loan. Repayment
commences one year after completion of the course or six months after securing
a job, whichever is earlier. The maturity period of the loan for studies in
India (up to Rs 1 million) and studies abroad (Rs 2 million) is 5–7 years.
Generally, no security is required for loans up to Rs 400 000. But for
loan amounts ranging from Rs 400 000 to Rs 750 000, banks may seek
third party guarantee. For loans above Rs 750 000, tangible collateral
security of suitable value, along with the assignment of the future income of
the student for payment of instalments, is required. The loans for vocational
courses are unsecured loans generally in the range of Rs 20 000 to Rs
150 000 for those pursuing courses that have a tenure ranging from 2 to 3
months to 3 years. The moratorium period ranges from six months to one year.
As per the data collected by Bandhopadhya,A.(2016), the
gross NPA for student loans is around 6%, whereas for the retail counterparts
it is 2%-3% (2012-13).
The accumulation of NPA with the bank has led to a sharp
decline in the growth rate of education loans across various commercial banks
in recent years.
The various causes of default of loans are
– Idiosyncratic borrower specific problems; These include
risk arising out of repayment problems, collateral risk, academic failure,
financial problems etc.
– Systematic factors ;
These involve various external factors such as unavailability of jobs
due to economic slowdown, recession, lack of quality education etc.
– wrong selection of beneficiaries
– ineffective follow up of advances
– failure of debt collection framework in banks
One of the earliest researches down on the subject of
student loans was done by Boyes,
Hoffman, and Low (1989) wherein research
demonstrated a method of successfully determining
the probability of individual default
risk using the data on borrower specific personal characteristics, economic
variables and financial variables from credit card applicants.
It was by Greene (1992) when he successfully developed a
statistical scoring model for discrete choice and explained its utility in
predicting consumer loan default and loan approvals.
Scoring models help
predict the future default and survival probability of a customer and thereby assist
in the lending process. Fritz, Luxenburger, and Miehe (2007) describe the
retail score card development process through a linear combination of several
input variables such as socio-demographic information, borrower financial
information, account data, collateral characteristics, and external credit
history data.This was followed by Roszbach (2004) wherein he developed a
bivariate Tobit model to predict future defaults and loan survival time for new
retail applicants. Such models allow banks to better predict the risk of a
customer and make more realistic evaluations of the returns.
Flint (1997), Knapp and Seaks (1992), Volkwein and Szelest
(1995), and Woo (2002) focused on the individual’s socio economic background
and its relation to defaulting of student loans. Gross, Cekic, Hossler, and
Hillman (2009) conducted a survey of studies of student loan default.
Summarising research between the 1990s to the 2007s factors influencing
the student loan defult can be classified as under:
a) students characteristics
category (type, area, educational outcomes etc.);
c) level of student debt
d) students’ employment and income and total debt position.
Lochner, Stinebrickner, and Suleymanoglu (2013) have used
survey and administrative data from the Canadian Student Loan Programme and
have considered demographic characteristics (age, gender, and aboriginal
status), educational background, income and other financial resources in their
study. They find that income level, access to savings and family support,
educational attainment, and various demographic factors have influence on
student loan repayment behaviour.
The findings from these studies have further guided
researchers in the study and the framing of the hypotheses, the methodology,
and the choice of variables.
Bandhopadhya,A. (2016) examined education loan default in
regards of the various characteristics associated with the loan (loan amount,
interest rate and repayment period), and security positions (margin given,
security, etc.). The mathematical model included a multivariate statistical
technique to control multiple factors that contribute to default risk.
We also check how various borrower characteristics (age,
marital status, presence of guarantor/co-borrower, etc.), geographic locations
of borrowers (rural, semi-urban, urban, and metro), course related factors
(domestic vs. overseas education and placement record) and rating of the
education institutes explain risk of default in student loans.
Across various empirical researches conducted by scholars, experimentation
includes certain variables that affect the probability of defaulting student
loans such as age, presence of guarantor/co-borrower, marital status, etc.),
geographic locations of borrowers (semi-urban, urban, rural,and metro), course
related factors (domestic vs. overseas education and placement record) and
rating of the education institutes.
Using the multivariate model as per Bandopadhya ,A(2016), female
borrowers have a higher risk of defaulting on student loan as compared to the
male borrowers. as per calculations, male borrowers are 1.42 times safer than
their female counterparts.
In a separate regression, it was established that married borrowers are riskier than
Study of time period as a variable established that longer
the repayment period, the lower is the chance of default.
A stark difference was observed within the geographical differences
wherein borrowers from the urban and rural centres were riskier than the urban
borrowers. This captures the local situational factors on risk of default.
Efforts to test whether merit and placement records have
statistically significant risk reduction effect by introducing intercept
dummies as additional variables in the regressions was also attempted. The results
showcased coefficients to be negative, however they are not statistically
Therefore the study
establishes that in order to lower default rate the loan must be secured and
the borrower’s own contribution for the course must be higher. Moreover, the borrowers with security are 1.5 times more likely to remain solvent
than those without security.
The hypothesis of the paper is built on the framework set by
various researches done in the past and the related literature.
The study included a set of variables such as age, marital
status, gender, geographical regions etc. the hypothesis was if the following affect
the default risk, calculating the probability of the same and controlling them
may lead to a decrease in default risk of student loans.
The study was a review type research design. It comprised of
reviewing literature on the subject
matter of student loans and coming to a common conclusion and suggestion for
reduction of the default risk.
A review of various researches suggest that on order to reduce
credit risk in education loans banks need to focus on strengthening credit risk
assessment techniques, borrower risk assessment through credit rating,
portfolio monitoring, due diligence in lending and institute performance
measures. Merit, employability of course, and reputation of institutions should
matter in loan appraisal to reduce the default risk. Creating awareness among
the borrowers/co-borrowers for repayment of the dues as scheduled and building
a repayment culture among the students is also part of the social
responsibility for banks. Regular tracking of the student and follow-up may
also reduce the risk of default. Employers should be sensitised regarding
payment of equated monthly instalments (EMI) of education loan of their employees.
Moreover, by segmenting borrowers by probability of default
and loss given default in a multidimensional scale, banks can adopt better loss
mitigation and pricing strategy to resolve borrower problems. Borrowers with
high probability of default and high loss severity can be segmented from lower
credit risk borrowers.
Though the smaller loans are mostly unsecured, for bigger
amounts, banks may ask for securities (in the form of fixed deposits (FD), LIC
policies and property) and co-applicant as a guarantor to reduce the risk. We
recommend that banks use yearly cohort default rates measures (e.g. transition
matrix or NPA movements) to track the rating slippages to estimate the
portfolio credit risk. This is to be done across regions, course types, institution-wise,
and so on, to better understand and monitor portfolio risk. A portfolio
approach may enable a bank to better monitor the risky customers and would
allow for targeted collection efforts to resolve the default. Banks may
prioritise the collection process for high risk accounts earlier in the
delinquency cycle. Else, they may opt for credit guarantee protection from the