Stories & Learnings from a Career in Credit Risk Part 5
Introduction
Last month we published a couple of light-hearted pieces about some of the funny or strange things I have encountered in my many years in the credit risk industry. I also provided some of the learnings gained from these experiences.
The two articles proved to be very popular with our readers, so we decided to start off the new year with two more of these humorous collections of stories.
Here are two more stories, from different parts of the world, and for obvious reasons all names and institutions have been anonymised.
“Sometimes reality is too complex. Stories give it form.”
Jean Luc Godard, film writer, film critic, film director
This collection of stories is focused on how lenders’ policies and operational environments can impact decisioning strategies and results. These stories highlight why a good credit risk manager will always investigate all possible causes of unexpected strategy results.
Exceptions that are Unexceptional
When working with a new client for the first time, I always found it to be a good idea to invest a few days reviewing their policies, product and systems parameters, reports and high-level operations. This groundwork, before the design of decisioning strategies had even commenced, often identified a lot of items that could not be explained.
Just like a laptop, lenders often accumulate legacy junk over the years and the actual reasons why these policies or procedures were put in place are lost in the sands of time. “We have always done it this way” becomes the explanation and so the legacy situation becomes self-perpetuating.
So, despite the known best practice, we had to take a small European card issuer live with their decision engine, but without any preparation meetings ever being held. I cannot remember the reason for this, perhaps it was lack of available time, or maybe it was my employer, which was in cost-cutting mode at the time.
We completed the design meetings, the client signed-off the strategies and they all went live the following month. Hooray, mission accomplished!
However, as we eagerly looked at the first month’s results, questions started to be raised. The results were well below expectations, but even more strangely, this crossed all strategies and all decision management areas. The lacklustre results were not caused by one bad strategy, all of the strategies were disappointing!
We started to review other reports and found to our astonishment that a very high number of accounts were excluded from the strategy. In fact approximately 15% of the total accounts were excluded, which is well above the norm. This explained why on a percentage basis, the results looked so weak and below all expectations.
The following week we spent three days onsite, two days to examine the exclusions at an account level and then one day to meet with the bank’s management to discuss the results. The bank’s management were ‘underwhelmed’ by the benefits that the decision engine had produced and so we were expecting a “frothy meeting.”
Having concluded our onsite investigation we went into the bank’s management meeting, and it started off ‘chilly,’ to say the least, and there were mutterings of “refunds” and “cancellations” being made.
We then explained the results and finally presented what the root cause was. Approximately 15% of the accounts were excluded from any treatment by the decision engine, as they had been flagged as ‘VIPs’. That’s right, the bank considered 15% of the accounts to be VIPs which were flagged for special treatment only.
When looking at a random selection of the VIPs, we identified friends, family members, journalists, police, in fact anyone that over the years the bank employees thought should be excluded from standard procedures. These accounts had accumulated over time to become significant in terms of number of exclusions.
There were red faces all around, as nobody could remember why the bank would consider 99% of these VIPs to actually be special and the following month the bank re-set the VIP flag and restricted it from ever being used again by the majority of staff.
The following month the results looked more in line with expectations and whilst there were a number of customer service challenges over the next few months, (“but you have always let me spend double my limit before” was a common complaint, along with “but you always let me pay what I want, when I want!”).
Learnings
- Account management strategies often result in a major change in the way business is conducted and so be prepared for challenges in the first few months of implementation.
- Never take shortcuts when designing and taking a strategy live. Preparation and research is key, and a day of preparation will save a week of investigations. A shortcut will only bite you in the proverbial!
- Sometimes implementing decisioning strategies means that longstanding operational policies need to be changed and the bank staff coached on the new processes.
- When reviewing unexpected results, always delve deeper into the background reports to find explanations. This can be tiresome but is essential.
Host Systems Parameter Settings and Operational Negation
A good risk manager will not only review all of the decision engine parameters that will impact the decision engine strategies, but also all of the appropriate parameters that can be set on the host systems. Host systems include the account processing system, the collections system, loan originations system, etc.
Whilst this can be tedious, it is best practice to review all of the appropriate parameters on a regular basis, in order to identify anything that has changed and also any settings that could be counterproductive.
Some examples of ‘maverick’ host system parameters that were identified due to their negative impact on account management strategies include:
Limit Increase Amounts
When implementing a limit increase strategy, we identified that a client had a system parameter settling limiting the maximum credit line increase amount. This had the effect of impacting any larger limit increases by truncating the increase amount.
| Existing limit | GBP 2,500 |
| Increase % | 30% |
| Increase Amount | GBP 750 |
| Maximum Increase Parameter | GBP 500 |
| Truncated Increase Amount | GBP 500 |
| Expected New Limit | GBP 3,250 |
| Actual New Limit | GBP 3,000 |
| Impact of the Parameter | GBP -250 |
Often these parameters were set years ago and are never reviewed and increased due to inflation, competitive pressures or any other factors.
PIN Lottery
Then there was the client who had never reviewed the majority of their parameter settings after converting onto a brand-new host system. The host system had set the default setting for the ATM PIN attempts parameter to 999.
This meant that a customer could use their card and attempt to draw cash from an ATM up to 999 times every day! For most banks and customers, the accepted maximum is the ‘three strikes and you are out’ rule. Not this bank!
Needless to say, this default parameter setting was changed on the very day it was identified and the changes were fast-tracked beyond the 30 days it normally took the lethargic bank to make any changes!
One Cent and You are Out
In my early days as a bank trainee, I worked a rotation in the cards collections department, which was run by a very ‘old school’ manager who was ex-US Marine Corps (I think you can get the picture!). His ethos was “work hard, be hard, play fair” and so you can imagine some of the collections policies that were in place.
One of the most infuriating policies, which was enforced by a host system parameter setting was the zero tolerance on payment amounts. In other words, unless the delinquent amount was paid in full, the account would still be considered as delinquent.
For example:
| Minimum Payment Amount | USD 55.01 |
| Amount Paid | USD 55.00 |
| Minimum Payment Tolerance | 100% (USD 55.01) |
| Status | Failed the tolerance setting: still delinquent |
Most companies use a minimum payment tolerance parameter setting, which is usually set at around 95%. (If the amount paid falls above 95% of the minimum amount due, then the account will be set to up to date). For these lenders, the same account would thus be as follows:
| Minimum Payment Amount | USD 55.01 |
| Amount Paid | USD 55.00 |
| Minimum Payment Tolerance | 95% (USD 52.26) |
| Status | Passed the tolerance setting: now current |
The collectors were trained to always quote a rounded up minimum payment amount to cover for this situation, which in this account’s case would have been USD 56-60. However, it is in human nature to round down and so the number of accounts that rolled forward in delinquency due to the underpayment of just a few cents was significant every month.
I don’t know if this situation was ever changed, as my rotation ended, and I moved to another department within the bank. It was certainly sub-optimal in terms of the collectors’ morale and the bank’s customer service.
Learnings
- You cannot fight against human nature, so any policies that create a dysfunctional situation like this should always be reviewed and changed.
- Sometimes being so rules bound will also backfire on the manager, as sticking so dogmatically to the 100% rule resulted in worse monthly collections results and a higher collector turnover.
(Another example of a misused payment tolerance amount was the collections department that set the parameter to just 5%! In other words, customers only had to pay a mere 5% of the amount due to be considered up to date!
When we discovered this ‘anomaly’ the entire collections department feigned ignorance and confusion, from the management down to collectors. Meanwhile, they had all been pocketing large bonuses for a long time. Needless to say when the parameter was ‘corrected’ to 95%, the bonuses ended!)
Summary
As mentioned in the introduction, and to paraphrase the disclaimer always used in American TV series, “…no identification with actual persons (living or deceased), places, buildings, and products is intended or should be inferred.”
However, I have added learnings to each story, and these should be taken on board by the reader, as they have helped me greatly in this industry over the years.
About the Author
Stephen John Leonard is the founder of ADEPT Decisions and has held a wide range of roles in the banking and credit risk industry since 1985.
About ADEPT Decisions
We disrupt the status quo in the lending industry by providing fintechs with customer decisioning, credit risk consulting and advanced analytics to level the playing field, promote financial inclusion and support a new generation of financial products.
