Infrastructure & Business Finance – Data & Risk Layers Key Part of Dance

Many so called purchase-to-pay networks want to provide some finance technique to the invoice volumes that run through them. This takes more than just technology and ties into credit underwriting and algorithms.  After all, even if Pepsi or Corning approve an invoice and leverage third party capital, you still have a potential dilution problem, albeit the risk can be very small depending on the supply chain.  Risk could take the form of offsets, credit memos, deductions, etc.

Click here to download your copy of the 2016 State of Supply Chain Finance Industry

Risk Layer

Risks such as default risk, invoice dilution, shipment failure, political and currency controls, etc. exist along the functional value chain.  In financing receivables, there are several risks to address:

  • Transaction Risk: Will the buyer pay, will they pay the full amount and when will they pay? Even financing from a buyer approved invoice, which should eliminate dilution and timing, you still have buyer insolvency risk, supplier quality issues, and supplier breach of contract.
  • Fraud – The risk the invoice is bogus.
  • Perfection – third party claim to purchased assets, eg. what if the receivable already is pledged to another lender, such as Citibank or CIT?
  • General Setoff – potential that a Buyer seeks to recover losses incurred in the event of a Seller bankruptcy by setting off payment against all payables to that Seller. For example, I buy coal from a borrower and they go bankrupt, can offset outstanding receivable with the cost of procuring new coal supply.
  • Servicer Risk – The potential that the program Servicer’s platform or processes are flawed, un-reliable, or enable the perpetration of fraud.

In the Purchase to Pay space, many vendors are focused on selling integration services around eProcurement, einvoicing, or business networks to the Global buy-side 2000, (ie, the major corporates and their supplier base). Of course many have added abilities to help their client offer some form of early payment to extinguish a payable early.

Data Layer

Networks are information advantaged but relationship and operationally light compared to Commercial Finance companies. They first start with an integration project around eInvoicing, eProcurement, EDI replacement, etc. And many realize they have valuable network data, data that bankers and commercial financiers do not have that could greatly enhance underwriting opportunities, risk based pricing, loss rate modeling, etc. For example, the combination of a purchase order, invoice and invoice approval provides a combination of invaluable information to a lender. It provides a lender with significantly more comfort that an invoice will be paid and at what level (specifically) it will be paid than any other form of guarantee before payment.

When that is tied back to payment data, where the platform has direct integration with a company’s ERP system, and once payment is remitted, they post back to them payment reference detail, it becomes more powerful.

The challenge is to build intelligent underwriting models that can take this data and make credit decisions. There are many questions to ask. Here are a few:

  • How to use the data contained in supplier networks to make a finance decision?
  • What data is required (historical data on trading relationship)?
  • Does the data extend across business cycles?
  • Does it include dispute and dispute resolution?
  • Can you actually track an invoice to a payment?
  • What risk mitigation techniques are required? e.g., trade credit insurance, cash dominion control of accounts, etc.

Of course, as we move into AnalyticsasaService world, with real-time data feeding risk and underwriting models, this all will get much more interesting.

Click here to download your copy of the 2016 State of Supply Chain Finance Industry

Share on Procurious

Discuss this:

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.