Credit Risk Measurement

Credit risk measurement
Currently the Bank relies virtually exclusively on subjective analysis or so-called

‘subjective/expert systems’ to assess the credit risk on loans. To this effect, credit analysts
use information on various borrower characteristics such as borrower character (reputation),
capital (leverage), capacity (volatility of earnings), collateral and cycle (economic
conditions), the so called the 5 “Cs” of credit, to reach a largely subjective judgement (i.e.,
that of an expert) as to whether or not to grant credit. In addition to these five “Cs,” an expert
might take into account the level of interest rates. Negative relationship between high loan
rates and expected loan returns occurs because of adverse selection and risk-shifting.1
Although many banks still use expert systems as part of their credit decision process, these
systems face two main problems: inconsistency and subjectivity. Potentially, the subjective
weights applied to the five Cs by an expert can vary from borrower to borrower if the expert
so chooses. This makes comparability of rankings and decisions very difficult for an
individual monitoring an expert’s decision and for other experts in general. As a result, quite
different standards can be applied by credit officers, within any given bank, to similar types
of borrowers. It can be argued that loan committees or multilayered signature authorities are
key mechanisms in avoiding such consistency problems, but it is unclear how effectively they
impose common standards in practice. This disparity in ability across experts has led to the
development of computerized expert systems, such as artificial neural networks, that attempt
to incorporate the knowledge of the best human experts.
There has been little progress towards systems that are more objectively based. A natural
extension towards objective assessment of credit risk on loans and advances is to use
accounting-based credit-scoring system – univariate and/or multivariate models. In the
former, the credit decision maker compares various key accounting ratios of potential
borrowers with industry or group norms. When using multivariate models, the key accounting
variables are combined and weighed to produce either a credit risk score or a probability of
default measure. If the credit risk score, or probability, attains a value above a critical
benchmark, a loan applicant is either rejected or subjected to increased scrutiny.
The idea is pre-identifying certain key factors that determine the probability of default (as
opposed to repayment), and combine or weight them into a quantitative score. In some
1
When loan rates rise beyond some point, good borrowers drop out of the loan market; they prefer to self-
finance their investment projects (adverse selection). The remaining borrowers, who have limited liability and
limited equity at stake, have the incentive to shift into riskier projects (risk shifting). In good times, they will be
able to repay the bank. If times turn bad and they default, they will have limited downside loss.
cases, the score can be literally interpreted as a probability of default; in others, the score
can be used as a classification system: it places a potential borrower into either a good or a
bad group, based on a score and a cut-off point.
There are at least four methodological approaches to developing multivariate credit-scoring

systems:
- The linear probability models

- The logit models
- The probit models
- The discriminant analysis model.
Logit analysis uses a set of accounting variables to predict the probability of borrower
default, assuming that the probability of default is logistically distributed i.e., the cumulative
probability of default takes a logistic functional form and is, by definition, constrained to fall
between 0 and 1.
The most common form of discriminant analysis seeks to find a linear function of accounting
and market variables that best distinguishes between two loan borrower classification groups
- repayment and non-repayment. This requires an analysis of a set of variables to maximize
the between group variance while minimizing the within group variance among these
variables.
While in many cases multivariate accounting-based credit-scoring models have been shown
to perform quite well over many different time periods and across many different countries,
they have been subject to at least three criticisms. First, that being based on book value
accounting data (which in turn is measured at discrete intervals), these models may fail to
pick up more subtle and fast-moving changes in borrower conditions, i.e., those that would be
reflected in capital market data and values. Second, the world is inherently non-linear, such
that linear discriminant analysis and the linear probability models may fail to forecast as
accurately as those that relax the underlying assumption of linearity among explanatory
variables. Third, the credit-scoring bankruptcy prediction models are often only tenuously
linked to an underlying theoretical model. As such, there have been a number of new
approaches ± most of an exploratory nature, that have been proposed as alternatives to
traditional credit-scoring and bankruptcy prediction models.
The Basel II provided banks with two options to objectively measuring credit risk, each one
containing its own unique set of problems: using risk categories based on credit rating
agencies (CRAs) or using the bank’s own credit risk models. In the absence of credit rating
agencies, in less developed financial markets like in Ethiopia, banks have no option but to use
the relatively more complicated internal risk-based approach (IRB). This allows a bank to
make its own assessment of risk, rather than rely on CRAs. A bank can use its own data to
determine the risk level of its loans by analyzing: the probability of default in one year, the
bank’s exposure and losses if there is a default, and when the borrower will likely repay if it
does not default. Within IRB, there are two further options: advanced IRB and foundation
IRB. The latter entails more oversight than the former. The foundation approach allows banks
to estimate the probability of default for each asset but requires the bank supervisors to
estimate the three other factors. The advanced approach, which is only available for the most
sophisticated banks, permits a bank to make most of the estimates, provided the bank
supervisors approve them.
According to the Basel The Bank must have an adequate credit risk management process,
including prudent policies and processes to identify, measure, evaluate, monitor, report and
control or mitigate credit risk on a timely basis, and covering the full credit life cycle – credit
underwriting, credit evaluation and the ongoing management of the bank’s portfolios.
Principles 17 and 18 of the Core Principles for Effective Banking Supervision (Basel Core Principles) 8
emphasise that banks must have an adequate credit risk management process, including prudent
policies and processes to identify, measure, evaluate, monitor, report and control or mitigate credit
risk on a timely basis, and covering the full credit life cycle (credit underwriting, credit evaluation and
the ongoing management of the bank’s portfolios). Additionally, adequate policies and processes
must be in place for the timely identification and management of problem assets and the
maintenance of adequate provisions and reserves in accordance with the applicable accounting
framework.

Credit Risk Measurement

Uploaded by

Copyright:

Available Formats

Credit Risk Measurement

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

Credit Risk Measurement

Uploaded by

Copyright:

Available Formats

Credit risk measurement

Currently the Bank relies virtually exclusively on subjective analysis or so-called

There are at least four methodological approaches to developing multivariate credit-scoring

- The linear probability models

You might also like