In: Finance
the relative merits between credit risk modeling and qualitative analysis (human judgment). What is the danger of relying too much on models? How important is it to have a balance?
Credit risk modelling is useful as :
1) It can predict credit risk and chance of default through data
analytics.
2) The prediction accuracy only improves through a greater amount
of data and instances in a credit risk model and thus the model
accuracy can only improve with time.
3) Since the model is objective in nature, so there is no chance of
bias or subjectivity.
Major disadvantages are:
1) It can not model qualitative factors well.
2) Human behavioural models are difficult to make and tune.
3) Data collection might be expensive and modelling could be costly
also as well as difficult in certain circumstances.
The advantages of human judgement are:
1) It can make a judgement of the qualitative factors well
2) Can understand and infer human behaviour and preferences.
Major disadvantages are:
1) It can lead to bias and human subjectivity creeping in the
decision model.
2) The scoring cannot be done quantitatively and concrete
conclusion may be difficult.
Reliance on too much models may lead to ignoring the human
factors and thus skew the decision. Also models may be biased due
to data limitations or the kind of data availability as too many
defaults or a high number of defaults may lead to reduction in the
predictive ability.
The importance of balance is paramount as it would combine the
quantitative factors and the human judgement optimally to
ultimately enhance the credit decision making process.