How algorithmic DCA allocation boosted loan collections at a Tier-1 Servicer

Our suite guides banks, funds and asset managers in the transformation of the Non Performing Asset into cash via different strategies


• Most NPL workout strategies are focused on manpower rather than effectiveness and efficiency

• This generates highly inefficient workout units that need optimisation in economic downturns

• Banks and Distressed Funds outsource the workout of default loans to Servicers, who outsource it too to Debt Collection Agencies and Asset Managers

• This creates an inefficient value chain where responsibility is diluted in too many outsourcings, where middlemen add little value while charging fees

The Challenge

• Menhir’s task was to identify in advance which loans would generate payments in the short term, and design an effective and efficient asset management strategy,based on the output of the algorithm

• Loan data quality was very low, Asset Managers lacked the methodology to implement well designed processes and stick to them

• The customer held a +4bn€ NPA portfolio, with both NPL and REO. A small mistake could generate huge impact

The Approach

Payment anticipation model: Menhir created a Machine Learning model capable of identifying paying customers before any management was performed

Allocation algorithm: Menhir designed a loan allocation algorithm to sort the assets between the available resources, to increase management intensity (with the same FTE), while reducing servicing costs

• As not all assets could undergo automatic allocation, Menhir segmented the AUM in two trenches:

     • Assets with Payment Anticipation = 3,441.7 M€

     • Assets with Payment Anticipation & Allocation Algorithm = 721.2 M€

The results

Monthly collection rate increased by 81%

Collection cost reduced 41%

Stay tuned.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.