Tuesday, June 28, 2011

Pardon the Interruption

I intended to complete the list of posts identifying sources of incremental risk assumed in trading on the TRE platform before either attempting to quantify a risk premium or suggesting risk-mitigating changes in practice.

So, pardon the interruption, but I want to pause here for just one post to make a suggestion. It’s one that I think would help mitigate risk from two sources: the lack of a personal guarantee and the lack of a full notification procedure.

First, let me acknowledge that I understand the reasons why TRE has adopted the policies in place.

As we’ve pointed out, almost all TRE Sellers would be asked to provide personal guarantees by other funding sources. The lack of that requirement by TRE is thought to be of significant help in attracting Sellers. On the notification issue, I understand that a true full-notification policy would require a very substantial commitment of additional time and manpower AND it would substantially slow down the process of bringing auctions to market.

I do get it.

But the fact is that the combination of these two issues produces a potentially tempting opportunity for Sellers who either fall on hard times , or who are simply dishonest, to post invoices for sale that do not meet the criteria set forth in the Master Program Agreement.

The lack of a full notification criterion undoubtedly increases the risk that invoices will be improperly posted for sale. And the lack of a personal guarantee reduces the risk to the Seller of doing so.

There is, I think, a fairly straight-forward way to mitigate (if not eliminate) this risk.

The principals of TRE Sellers could be required to execute a CONTINGENT personal guarantee agreement that would have no effect EXCEPT in the event that invoices not meeting the criteria of the Master Program Agreement were posted for sale on the TRE platform.

So long as invoices, even those that might become problematic, actually meet the specified criteria, the obligations to cure defaulted payments would remain corporate only. However, if the invoices were improperly posted, which is a condition under the control of management, TRE and the Buyers could look beyond the corporate assets of the Seller to the personal assets of the owners.

Obviously, such a change in policy would meet with Seller resistance and would likely cause some potential Sellers to fall by the wayside.

However, if the principals of a Seller are not willing to guarantee their own adherence to the invoice posting requirements; essentially assuring TRE and the Buyers of their own honesty; they might be Sellers we could all do without in the long run anyway.

Thursday, June 23, 2011

Appropriate Compensation #6: The Notification Issue

No single issue has been discussed in as many of our posts as the issue of invoice notification versus verification.

See the following for prior comments:

 An Inconvenient Truth Part One: July 6, 2009
 An Inconvenient Truth Part Two: July 8, 2009
 Whose Ox is Gored? : October 19, 2009
 Half a Bubble Off Plumb: November 30, 2009

Why has this issue been given such attention?

Because, in my view, it represents the single largest potential threat to TRE Buyers. And, as such, the compensation it deserves is meaningful.

That compensation should be considered both at the level of the entire TRE platform as compared to systems that DO include full notification and among TRE auctions and Sellers, where differential risks can often be identified.

Let’s review the general issue.

In a “full notification” invoice purchase the Account Debtor will be asked to confirm that:

1) It has a contractual relationship with the Seller,

2) The invoice presented has been generated under the terms of that contract,

3) The work done by the Seller has been completed AND meets the terms of the contract,

4) The amount of the invoice is correct,

5) The payment terms are correct, and

6) Payment will be made as instructed in the notification document.

Problems might (and do) still arise, of course, even given all of the assurances listed above, but the general approach should go a long way to minimize surprises.

TRE does not use a “full notification” system. Instead, it uses an invoice “verification” system.

Under the TRE verification system the Account Debtor is asked to verify that an invoice matching the one posted by the Seller for sale DOES exist in its Accounts Payable system. That is: there is an invoice from the Seller in the Debtor’s AP system that has matching identification and amount details.

Essentially this tells us that the Debtor has SUBMITTED an invoice. It does not necessarily tell us that the invoice is valid, that the goods or services meet contract requirements or that the Debtor actually acknowledges the obligation.

TRE is understandably reluctant to make public in any detail the exact mechanics of the verification system it uses. Too much transparency would risk allowing Sellers to “game” the system.

However, it did notify Buyers in 2009 that it would not verify 100% of the invoices posted for sale on the TRE platform. Instead, after gaining a certain (unspecified) amount of experience with a Seller/Debtor pair, it would employ a statistical sampling process that it said would at least match the best practices of the factoring industry in terms of the percentage of invoices verified in given Seller/Debtor relationships.

There are Sellers whose auctions routinely consist of dozens of individual invoices and some of those Sellers post auctions of invoices due from the same Debtors quite frequently. It is understandable that asking the AP department of a Debtor to frequently verify dozens, or even hundreds of individual invoices might generate a little “push back” from the Debtor after a while.

While I have no knowledge of the actual sampling system employed by TRE, I’m willing to accept the idea that sampling can be an appropriate practice IF, of course, you’ve already accepted the verification versus notification alternative.

But that’s not really the issue here. The issue is: “what level of additional income should a TRE Buyer receive to compensate for the added risk inherent in accepting the verification versus notification alternative?”

There are two points that can be made immediately:

1. As we've said on other points of risk assessment, we just don’t have the data to measure this risk factor with any meaningful level of statistical reliability; any "answer" will be a judgment call; and

2. The answer is not “zero”. There IS an incremental risk to the TRE Buyers that should be compensated and that, over time, MUST be compensated if the exchange is to thrive.

So, how can we approach the problem?

Is the experience of TRE, itself, a useful guide?

In terms of quantifying the appropriate risk premium I don’t actually think the experience of TRE to date would be particularly meaningful even if complete data were available.

First, the level of transaction volume is still in its early growth phase and too small to provide a reasonable denominator against which to measure default/loss experience.

Second, the public default information involves ongoing litigation whose outcome continues to be uncertain. So the numerator is also problematic.

Third, while the details are not public, it is certain that TRE has made changes to its practices on the basis of what it’s learned in the early cases of default.

That being said, however, we CAN identify the kinds of things that cause a verification process to carry with it a higher level of risk than a notification process.

These tend to fall into two categories:

1. Falsification of, or fraudulent offering of, invoices: by completely fabricating the invoice or by materially misstating its amount or payment terms or by selling invoices that have already been sold to others or that are subject to the specific claims of others.

2. Diversion of payments either by actively causing payments to be made to an incorrect party or passively by accepting and retaining payments mistakenly received.

There are a lot of variations on these themes, of course, but most situations will fall into one of these two broad categories.

I think it is unquestionably easier to sell false or fraudulent invoices in a factoring system that is based on verification than it is in one based on full notification.

The case of diversion is not as straightforward. Diversion certainly occurs in both its active and passive forms in notification systems as well as in verification systems.

It can be argued that active diversion is easier in a verification system because the only point of contact between the factor and the debtor is at the level of an AP person who might have no knowledge of the substance of the relationship between the Seller and the Debtor.

I suspect that the chances of a payment being sent improperly to an invoice seller are probably just as high in either system. On balance, given the level of uncertainties in trying to analyze any of these issues, I’d probably call the diversion risk a “wash” for our purposes.

But the falsification or fraudulent offering issue (including the various forms involving collusion) is, I would argue, a significant differential risk factor. Buyers deserve to be compensated for assuming that risk.

What would represent adequate compensation?

In this case, as in the case of personal guarantees, we have no evidence to offer in support of a specific incremental return requirement.

In this case, as in that one, however, both the incremental risk and the appropriate return premium are “significant”.

There ARE auction situations on TRE in which this risk factor is significantly mitigated; such as in the case of Ariba Network invoices. There are also cases where the Seller documentation provided with an auctioned invoice gives the Buyer substantial comfort that the invoice is valid.

We'll discuss ways in which TRE and its Sellers might provide risk-mitigating information with auctions in a later post.

But, in general, our view is that an invoice verification process carries significantly higher risk to the Buyer than would a notification process and its pricing should reflect that reality.

Friday, June 10, 2011

Appropriate Compensation # 5: It's Still Not Personal

In our post of June 22, 2009, entitled “Liability: It’s Not Personal” we discussed the fact that TRE Sellers are not required to provide the personal guarantee of their owner(s) as additional security for the performance of their responsibilities to TRE and its Buyers.

In that post we made a general statement:

Most firms that buy individual invoices routinely obtain a personal guarantee of the seller’s obligations from one or more individuals associated with the seller.”

As we revisit that topic in this series on the issue of appropriate pricing of risk I think it’s time to firm up that earlier generality.

I’ll draw on two current sources for data relating to the prevalence of personal guarantee requirements:

1) the semi-annual report of the “Pepperdine Private Capital Markets Project” for summer 2011 by Dr. John Paglia of Pepperdine University (PPCM), and

2) the “Annual Asset-Based Lending and Factoring Survey Highlights, 2010” (April 27, 2011) of The Commercial Finance Association, (CFA).

The PPCM data is collected from a broader cross-section of the market that the CFA data and the CFA data is published publicly in “highlight” rather than detailed form but, for our purposes, those differences are not really critical. In fact the results of the two fit together quite reasonably.

According to PPCM data, the requirement for a personal guarantee in asset-based lending transactions varies substantially and as a function of the size of the loan. The smaller the loan the more likely is the requirement for a personal guarantee.

For loans under $1 million the median response revealed a 100% requirement for a guarantee. That figure falls to 90% as loan size increases to $5 million and 80% as it increases to $10 million. At a $50 million loan size, no personal guarantee requirements are reported.

PPCM survey responses from those in the factoring business are not given in such detail. PPCM reports that 87.5% of factoring industry respondents require personal guarantees while 12.5% do not. I think it’s reasonable to assume that the size of transactions and relationships is also a substantial variable in this sample and that the 12.5% of the deals done without personal guarantees would be found at the larger end of the size distribution.

So, as a general conclusion, PPCM suggests that the smaller transactions; presumably involving the smaller businesses; have a very high incidence of personal guarantee requirement; whether the transaction is an asset-based loan or a factoring arrangement.

The CFA survey draws data from a smaller, more-targeted sample of firms. In its own words it “decided to base the industry surveys on data reported by almost 40 of the largest asset-based lenders and factors…”

The “highlights” version of the CFA report does not provide data on the requirement of personal guarantees by asset-based lenders but it does provide that data for the respondents in the factoring survey.

In the CFA sample only 31% of respondents reported requiring full or partial “recourse”: a portion of which presumably includes personal guarantee requirements. While this might seem at odds with the PPCM data, it is not necessarily.

The CFA data is drawn specifically from the largest providers of funds. Presumably the largest transactions would be heavily represented among the largest funding providers. And, as we saw in the PPCM data, as deal size increases the requirement for personal guarantees decreases.

These two current reflections of the market for both asset-based and factoring transactions strongly suggest that personal guarantees are required as a matter of course in at least most of the smaller transactions in both the asset-based lending and factoring markets.

The issue of what is “small” is not as clear in the CFA data as it is in PPCM, but I think we would be justified in concluding that most, if not all, of the transactions that trade on TRE would; if they were made in the conventional marketplace; be subject to personal guarantee requirements.

One of the marketing points that TRE makes to prospective Sellers is that personal guarantees are NOT required, which is itself a statement about the otherwise general prevalence of the guarantees.

The “value” of those guarantees is not readily calculated – at least not with data that I have available. My own experience dealing principally with small sellers is that personal guarantees are of greater value in curing defaults than are “all asset liens” and certainly more valuable than “specific asset liens”.

In fact, given the length of time required to litigate a claim, obtain a judgment and then actually execute on a judgment it’s often the case that there’s nothing to be gotten from what remains of a business by the time a judgment creditor can actually try to seize anything.

On the other hand, the owners of businesses against which judgments have been secured, often have personal assets that really CAN satisfy claims. But if they haven’t provided a guarantee, the creditor has little effective means of reaching those assets.

This is another one of those cases where quantifying the value of a risk parameter is a difficult task if we’re looking for real analytical rigor. But I suspect we’d find broad agreement in the industry with the assertion that the premium required to compensate for the lack of guarantees should be “significant”.

What “significant” means, I think, has to be considered as a part of an overall evaluation of a transaction or relationship and I suspect that it changes with respect to a client over time and with experience.

But this is certainly an important item on our list of issues deserving a pricing increment in (at least) most TRE transactions.