Archive for the ‘Debt Affordability’ Category


Traditional credit scoring is built on the inherent assumption that past behaviour predicts future behaviour or, in simpler terms, that if you have always repaid your loans in the past you are likely to continue to do so into the future. 


This method has worked so well because it is generally true that the type of person who has met their obligations in the past is also the type of person who will attempt to do so in the future.  However, behaviour is driven by the intention to act as well as the ability to do so.  Therefore, good or bad intentions are only partially responsible for actions. 


The traditional method falls short though in that it fails to take into account an individual’s ability to repay.  Good intentions and money management skills may have allowed a customer to meet their existing obligations but it is possible that the new loan or a change n the external environment may have been enough to push them over the edge.


Affordability checks, therefore, differ from risk checks in that they test an individual’s ability to repay debt.  Calculating the ability to repay a loan is theoretically easier than calculating the intention to repay a loan.  Simply put, a customer is able to repay a loan when they have more money available to meet their debt obligations than they need to keep those obligations up-to-date.


Affordability:    Available Cash ≥ Cost of Debt Obligations


So, a lack of affordability can originate from two events – a decrease in available cash or an increase in the cost of debt obligations.  The relative availability of cash is not something over which the bank has much control.  So, the only way in which the bank can directly impact the affordability of a customer is by changing the cost of debt obligations – usually by providing further credit but also by changing pricing. 


Therefore, whenever the bank is increasing the debt burden of a customer they need to calculate the implications that this strategy will have on affordability.  Increases in the debt burden can take several forms but include new loans (originations), credit limit increases (account management), debt restructuring (collections), etc.


Affordability and Bad Debt

Although subtly different, risk and affordability have the same outcome measure – bad debt provisions or bad debt write-offs.  In fact, a lack of affordability can cause a low risk account to display very similar behaviours to a high risk account and to arrive at the same end.


We have seen that a lack of affordability can originate from one of two events – a decrease in available cash or an increase in the cost of debt obligations.  In turn, each of these events can be attributed to either internal or external factors and, depending on which of these is dominant in any one case, the strategies employed will change accordingly.


The matrix below combines the two possible causes of a negative change in affordability and provides some examples of each.




When a portfolio experiences a large and widespread increase in bad debt – or the behaviour that usually precedes bad debt – it is quite possible that this has been caused by a deterioration in the external forces that affect affordability rather than by a deterioration in risk –  a large rise in interest rates for example.


Internal and external factors can act in conjunction so it is important to identify which of them is to blame for the deterioration of customer’s affordability.  Because external factors are by definition outside of the control of the individual, they are unlikely to be predicable using individual-level data.  Instead, external factors usually need to be predicted based on wider economic trends.  Internal factors are more closely correlated to individual behaviour and so should be predictable using individual-level data.


We could imagine a customer (represented by the circle in the diagram below) as falling somewhere within a range of “affordability” – from not being able to meet any of their debt obligations to being able to meet them all easily (represented by the pyramid).  A change in internal behaviour will move that customer closer towards one of those extremes.


The external factors act as a hurdle (represented by the line), with customers above the line being able to afford their current debt obligations and customers who fall below the line being unable to do so.  As the external environment worsens, the hurdle moves up the pyramid erasing the affordability of previously sound customers.




In the illustration above it is clear how a customer who can currently afford their debt obligation (bold circle, bold line) could find themselves unable to afford their debt in the future should their internal situation worsen (loss of commission income, loss of work, etc.) in conjunction with an worsening of external factors (increase in interest rates, new debt taken on, etc.).


The A significant worsening in external factors can therefore make a sudden impact across a portfolio while a worsening in internal factors is likely to be much more isolated in its impact. 


Measuring Affordability

Unfortunately, it is not as easy to measure affordability in practice as it is in theory.  The components of the affordability calculation are exceedingly difficult to obtain indirectly.  This means that a bank – or other lending institution – is usually forced to work with an estimate of affordability which is usually arrived at using an amalgamation of customer-supplied (direct) data and third-party (indirect) data. 


Unfortunately, customer-supplied data is prone to intentional and unintentional manipulation while third-party data is seldom complete.  It is unlikely, even for a customer with a rich credit history, for a third-party to know all of the relevant in-flows and out-flows.  This scenario is further complicated once joint-incomes and joint-debts are considered.  The relative availability of data will vary within and across markets – affected by factors such as credit bureau sophistication, privacy laws, etc.  It is important to know the relative availability of affordability data as it is the main factor dictating the level of predictions that are possible. 


As it is easier to prove the a customer can not afford a loan than it is to prove that they can afford one,  the relative availability of data and systems will dictate the type of affordability decisions that can be made.  For example, where an organisation only has access to basic, customer-supplied data it will be impossible to rule out manipulation of the data for the customer’s benefit.  However, it is reasonable to assume that whatever manipulation may have occurred would have been undertaken so as to improve the apparent level of affordability of that client.  Therefore, where the data provided is sufficient to show a lack of affordability it can be taken at face value that the customer should not receive the loan.  However, where the data provided shows the customer to be in a position to afford the loan it may not be sufficient to guarantee that the real situation is also as such.  So, the only decision the organisation will be able to make are those where a lack of affordability can be clearly proven.  Should that same organisation later gain access to third-party data, it might become possible to identify customers who probably can’t afford the loan, etc.




In reality, the affordability decision is always negatively framed – as with a statistical test where the hypothesis is either rejected or not rejected, never accepted.  In our example, should there have been insufficient evidence to show a lack of affordability, the bank would not have proven affordability, it would simply have failed to prove a lack thereof.  The fact that the bank may subsequently choose to proceed with the loan does not change this.  It is impossible to truly prove affordability.  The possibility of unknown data will always exist – perhaps the customer has a gambling problem, perhaps they know that they will be retrenched in a month, etc.


This might be a subtle point that seems little more than semantics, but it is important to the strategy setting process.    Unless the data and systems available are very sophisticated, affordability checks should always be a means of identifying otherwise good accounts that should be declined, never as a means of identifying otherwise risky accounts that should be approved.


Measuring Affordability in Practice

There are three important questions to answer – when should affordability tests be performed, what should they measure and how accurate should they aim to be.


Affordability is affected by a change in available cash or a change in the total debt burden.  As only the size debt burden is within the control – at least to a degree – of the bank, this is the most important trigger for an affordability test.


So, whenever the bank plans to increase the debt burden by offering further credit, it needs to test the proposed strategy’s likely impact on the target customers’ affordability.  Such a test would need to consider the impact that the strategy will likely have on internal factors as well as the impact that any anticipated changes in the external factors could have on the same customers.  


Having identified what should be measured, the final step is to determine the achievable level of accuracy.  It is never possible to prove affordability beyond doubt, so an affordability test will always be looking to find sufficient evidence of a lack of affordability.  Where little information exists, only the clearest cases can be identified – those where the customer can definitely not afford the loan – while, with slightly more information, it might be possible to identify those likely, but not guaranteed, to be unable to afford the loan, etc.  This is not only an important step to managed expectations but also to ensure the result sof the affordability test are used in the right context.



Read Full Post »