Fantasy Planning

Introduction

“He who fails to plan is planning to fail.”

Winston Churchill.

It never ceases to amaze me how little credit grantors actually plan. Speaking specifically about the credit risk environment, plans should be created for at least a 12 month cycle, which typically matches the company’s fiscal year.

Whereas credit risk managers are often very good at budgeting to ensure that they get funding for that extra analyst and specific scorecard re-developed over the next budget cycle, they are often less successful at what really matters: planning for their credit risk activities.

This has often been the case for as long as I have been in the industry. However, a new phenomenon has arisen over the past few years, which is the topic of today’s blog: ‘Fantasy Planning.’

Introducing Fantasy Planning

I had never heard of this phrase, until a colleague alerted me to this fascinating conference paper

“Numerous man-made disasters have revealed fantasy plans – safety artefacts which do not represent the reality of operational risk.

In contrast to “drift,” where operations gradually become less safe, fantasy planning describe protections that have never been fully implemented, understood, or operated as intended.

The theory of fantasy planning originally comes from sociologist Lee Clarke. Clarke described how the oil spill contingency plan for the Port of Valdez, Alaska, claimed that approximately 130 k barrels of oil could be recovered from the sea after a spill. This was tested in 1989, when the Exxon Valdez tanker ran aground, spilling 260 k barrels of oil. Despite the plan’s claims, no amount of oil even close to that amount had ever been recovered from open waters.

In the city of San Bruno, California, 2010, a buried gas pipeline owned by Pacific Gas and Electric (PG&E) ruptured in a massive explosion. Hayes and Hopkins in their book ‘Nightmare Pipeline Failures’ noted that despite PG&E’s strong commitment to safety and use of bespoke risk modelling techniques, PG&E’s understanding of risk was not aligned to “actual” field risk.

Both the oil spill contingency plan and the IMS had little grounding in reality. They were, in part, fantasies – but not deliberately by people.

Most concerning is that sometimes risk protections may never be implemented or capable of being effective.

This paper explores why, contrary to efforts towards safety, organisations create systems that describe a physical reality that may never exist. It also speculates on ways to close the gap between safety systems and operational reality. Ultimately, this paper seeks to sensitise risk practitioners to the problem of fantasy planning, and the vulnerability it creates.”

Fantasy Planning in the Credit Risk Department

Does this strike a chord with you?

Whilst the conference paper examples and definition of ‘Fantasy Planning’ are focused on operational risk, this concept can be extrapolated to the credit risk department of any credit grantor.

As previously mentioned, credit risk managers are often very adept at budgeting for costs and revenues, but less successful at creating a plan as to how the financial objectives are to be achieved operationally. By signing up for the financial targets of the new fiscal year, the credit risk manager is alluding to having an actual plan, however this is often not the case!

This disconnect between accepting financial targets, and lack of an operational credit risk plan has often struck me as curious, as the domain of credit risk lends itself to planning, if the will is there to do this.

Examples of True Planning

The credit risk department of a company is perfect for implementing true planning, whilst avoiding all forms of ‘fantasy planning.’

In parallel with the company’s budget cycle, the credit risk manager needs to put in place all of the components which create a credit risk operational plan.

Portfolio Chronology Log

Logically the first step is to consult the Portfolio Chronology Log, which is a straightforward ‘dear diary’ style log that is maintained and accessed by the entire risk team.

As a rule of thumb, the more information that is included, the more accurate the Log will be. The sources and types of information should be wide ranging and from both internal and external sources

• Marketing campaigns
• Sales drives
• New product launches
• New scorecards implementation
• New scorecard cut-offs
• Scorecard re-alignments
• Changes to policy rules
• National events
• Strikes
• Financial crises
• Interest rate changes
• Inflation rate changes
• Exchange rate changes
• Changes in national regulations
• Software upgrades
• Software changes
• Credit bureau changes

The Portfolio Chronology Log makes life a lot simpler, as the risk manager can see at a glance what will need to be addressed in the upcoming fiscal year; for example, scorecard re-developments or re-alignments etc.

Liaise with other departments

Typically the credit grantors marketing department is highly plan-driven and will have created a plan covering the upcoming year’s marketing initiatives and new account drives.

By working closely with the marketing department (which is very rare in my experience), credit risk managers can get a better understanding of what is going to happen and when. This can then be accounted for on the risk side.

For example, it is probably not the best idea to roll out brand new application scorecards for the largest new accounts drive of the year.
Create a Testing Plan
Modern credit risk management is all about A/B testing so that a credit risk team can ‘test and learn.’ This means running multiple champion/challenger strategies over the course of a year.

Without a proper testing plan, scheduling when new strategies will be designed, reviewed and rolled out, it is not possible to manage a comprehensive set of tests.

Two other tasks are vital components of a testing plan:

Re-randomisations

The fundamental tenet of an A/B test is that the testing groups are random, ensuring that any differences in observed results are due to the strategies themselves. Lack of randomness in the applications or accounts that are being tested will result in biased results.

It is best practice to re-randomise the base on a 12-15 month cycle, so this task needs to be planned for. Re-randomisations cannot occur halfway through a champion/challenger test year!

Scorecard Re-developments/Re-alignments

These need to also be planned for and implemented prior to new A/B tests going live, as otherwise score cut-offs will be affected, resulting in biased test results.

Based on all of these considerations, and tasks that need to be in place prior to a new A/B testing cycle being implemented, it is best practice to use 12-15 months as the testing cycle period. This allows for the strategies to run their due course over a year, plus a leeway for re-randomisations, re-alignments and analysis of the various A/B test results, in order to crown the new champion strategies.

Summary

From my own personal experience, working with many credit grantors in many different countries, Fantasy Planning is rife within the industry. Targets are set and committed to, but no credit risk operational risk plan is ever completed, which is a sub-optimal situation to be in.

By not implementing a credit risk operational plan, outlining what is going to occur and when, many companies resort to knee jerk reactions and “strategy of the month” A/B testing, which never solves the problem and can be viewed as a ticking plaster approach at best.

By ‘flying by the seat of your pants,’ it is no wonder that so many credit risk departments miss their annual targets.

About the Author

Stephen Leonard is the founder of ADEPT Decisions and has held a wide range of roles in the banking and risk industry since 1985.

About ADEPT Decisions

We disrupt the status quo in the lending industry by providing lenders 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.