OPTIMISATION IN CREDIT WHERE CAN OPTIMISATION HELP YOU MAKE BETTER DECISIONS AND BOOST PROFITABILITY

Transcription

1 OPTIMISATION IN CREDIT WHERE CAN OPTIMISATION HELP YOU MAKE BETTER DECISIONS AND BOOST PROFITABILITY CSCC XIII Martin Benson Jaywing

2 Many business problems that arise in credit management can be tackled through constrained optimisation How can I assign loan prices to applications in a way that maximises profitability, but subject to meeting targets for lending and losses and being compliant with advertising regulation? How can I assign customers to collections strategies in a way that maximises cash collected but subject to operational constraints?

3 General Decision Optimisation Scenario Objective Constraints Optimisation Optimisation Algorithm Strategy Offers / Actions Assignment Decision Units

4 Loan Pricing Scenario Objective Hit Lending Targets Don t Breach Risk Appetite >50% Loans at Typical APR Optimisation Strategy Design Strategy Loan APR Assignment Loan Applications

5 Credit Card Limit Strategy Scenario Maximise Profitability Control Incremental Risk Manage Attrition Levels Optimisation Strategy Design Strategy Limit Increase Amount Assignment Credit Card Accounts

6 Pro-active Retention (Mortgages & CCs) Scenario Maximise Profitability Hit Retention Target Control Risk Manage Margins Optimisation Strategy Design Strategy Retention Offer Assignment Customers Due To Roll Off Discount Period

7 Collections Entry Streaming Scenario Maximise Cash Collected Minimum/Maximum Daily Call Volumes Hit Target Cure Rates Optimisation Strategy Design Strategy Contact Strategy Assignment Collections Entrants

8 Benefits of optimisation Optimisation is proven to return higher portfolio profitability than more traditional techniques, across a range of industries and use cases. It is the current state-of-the art in strategy design. Traditional Approach Optimised Strategy: Up to 20% uplift

9 SOLVING OPTIMISATION PROBLEMS

10 The Maths of Optimisation Maximise Subject to max o O f o some function c i o v i i = 1, over all possible parameter values subject to a set of constraints

11 Solving is difficult in general Subject to max o O f o c k o v k (k = 1, ) f x General solution techniques exist: Gradient methods Nelder-Mead algorithm Genetic algorithms, etc. But, they converge to local maxima, not global maximum (if they converge at all). Global Maximum Local Maxima x

12 Tractable formulation of decision problems max o i,j 0,1 I J Subject to i I j J o i,j c i,j,k i I j J o i,j f i,j (k = 1, ) For each decision unit (enumerated by I), should you make a particular offer (enumerated by J): yes=1 no=0 The optimisation goal is a sum of contribution values, each corresponding to assignment of an offer to a decision unit Also, assume that all constraints are linear in o i,j

13 Tractable but not trivial! max o i,j 0,1 I J Subject to i I j J o i,j c i,j,k i I j J o i,j f i,j (k = 1, ) Binary Integer Program - convergence to global optimum is guaranteed...eventually Still difficult - BIPs are NP-complete Most statistical software packages can solve BIPs using generic algorithms, but may take a very long time, even for small problems However, use of proprietary heuristic algorithms can enable even large problems to be solved quickly

14 PREDICTIVE MODELS

15 Modelling requirement Offers Decision Units A key requirement for optimisation is the ability to predict objective values for every decision unit, for every offer Model? Could attempt to model the relationship directly but it s not a good idea! Too many moving parts in the middle. Instead Objective

16 Loan Pricing example Scenario Objective Hit Lending Targets Don t Breach Risk Appetite >50% Loans at Typical APR Optimisation Strategy Design Strategy Loan APR Assignment Loan Applications

17 Profitability map Profit Per Loan Interest Revenue Fee Income Fee Amounts Loan Amount/Term APR Offered Expected Profit Expected Loss Cost of Capital Voluntary Attrition PD Customer Profile Prob. of Take Up Other Costs EAD, LGD Cost Information

18 Offers Decision Units Objective Regression Model Parameter Derived Profit Per Loan Interest Revenue Fee Income Fee Amounts Loan Amount/Term APR Offered Expected Profit Expected Loss Cost of Capital Voluntary Attrition PD Customer Profile Prob. of Take Up Other Costs EAD, LGD Cost Information

19 5.9% 6.9% 7.9% 8.9% 9.9% 10.9% 11.9% 12.9% 13.9% 14.9% Probability of Take-up Expected Profit 5.9% 6.9% 7.9% 8.9% 9.9% 10.9% 11.9% 12.9% 13.9% 14.9% Model errors drive decision errors Errors in predictive models 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Actual TU Rate TU Model 1 TU Model 2 create generate errors sub-optimal in objective decisions estimates TU Model 1 TU Model 2 APR APR

20 5.9% 6.9% 7.9% 8.9% 9.9% 10.9% 11.9% 12.9% 13.9% 14.9% 5.9% 6.9% 7.9% 8.9% 9.9% 10.9% 11.9% 12.9% 13.9% 14.9% Probability of Take-up Risk Band 1 Risk Band 2 Risk Band 3 Risk Band 4 Risk Band 5 Probability of Take-up Historic strategy can create modelling challenges Historic strategy may confound data 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Masking true offer impacts APR APR

21 5.9% 6.9% 7.9% 8.9% 9.9% 10.9% 11.9% 12.9% 13.9% 14.9% Probability of Take-up Extrapolation potentially drives under-performance 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% APR Extrapolation Over/under prediction of propensities Overestimation of the value of some offers Selection of suboptimal offers Strategy under delivers, models look misaligned

22 How to prevent problems Build models on the data that are available Constrain optimisation only to allow offers that are similar to previous experience Introduce small randomised control groups to support future modelling Updated strategy and testing generates richer data Iterate to unlock the full potential of optimisation in a controlled manner

23 SUMMARY

24 Summary Many business problems encountered in credit management can be addressed through the use of optimisation techniques Optimisation can deliver significant benefits over traditional approaches to strategy design. However, this is not straightforward, and success requires the right expertise and tools Careful consideration must be given to what models are required to support optimisation, and special care taken to ensure that areas of model weaknesses cannot compromise strategy performance Interested in finding out more? Read our White Paper:

25