A Comment on Risk

The downward pressure on auction pricing that began in late August has continued and intensified through November. I’ve suggested before that this might have more to do with excess liquidity than with either improved quality of Sellers or Debtors or with decreased transaction risk.

While auction volume has continued to increase it appears that liquidity has increased at a more rapid rate. One indicator of that is the percentage of auctions sold at their specified “buy-out” prices.

On one recent day, there was only one auction that DID NOT sell at its buy-out price.

I suspect that there is a substantial Buyer out there who has given instructions to a trader that he is not to BID on auctions; but rather only to BUY them.

So, no matter that an auction might be had on better terms, this Buyer pays the “ask” price. And that, of course, provides a powerful incentive for Sellers to drop their “buy-out” or “ask” pricing.

Now, I might very well be wrong about this and I have no knowledge of any other Buyer’s actual trading dynamics. I’m just speculating based on the activity as I see it.

Viewed from the perspective of both the Exchange and the Seller community, of course, the current dynamic is very attractive.

For the Exchange, it makes the job of attracting potential Sellers, and of retaining current ones, easier. Lower cost of funds and rapid auction closings provide the sales force with a strong "story" when marketing TRE to prospects.

For the Sellers, obviously, lower pricing is always desirable.

And Buyers could ultimately benefit as well to the extent that more Sellers might be attracted to the Exchange, although this benefit is a lagging function and is obviously offset by the lower yields available at least in the near term.

In the long run, the Exchange cannot thrive on the basis of a pricing level that does not, on average, compensate Buyers for: a) cost of funds, b) cost of operations, and c) risk assumed.

The TRE Buyer community is very diverse, not only in terms of the primary businesses they represent but also in terms of their motivation for participating in the Exchange. A Buyer that is not primarily in the receivables finance business; whose interest in TRE might principally be as a place to “park” cash balances in the short term; will have a very different view of acceptable pricing than will a Buyer whose principal business is buying accounts receivable.

It’s really impossible for anyone except the managers of the Exchange – who will know all of the Buyers and their reasons for participation -- to comment on the likely range of Buyers’ cost of funds, cost of operations or motives for participation.

But the risk involved in an Exchange transaction is the same regardless of the identity or motivation of the Buyer.

I’ve commented in previous posts about the differences between the typical business model of a company whose principal activity is the purchase of receivables and the model on which TRE is based.

In several respects the risks of buying invoices on TRE have to be acknowledged to be higher than is typically the case in the receivables finance business. Here I’m thinking of issues like the lack of personal guarantees and of blanket first liens on Sellers’ receivables.

On the other hand, there are offsetting advantages to the TRE model that are not inconsequential. The lock-box arrangement not typically available to a spot-factor, the advantageous nature of the Louisiana law and a generally higher-quality Debtor community, are examples.

Each Buyer, or potential Buyer, will make his own assessment of the balance between those advantages and disadvantages.

But today I want to revisit a risk-related topic that I’ve commented on before that has become a larger issue in the current pricing environment. It does not arise from the TRE operating model but is significantly affected by its pricing model: specifically the impact of its fee structure on auctions of very short duration.

Given the relatively brief history of TRE it’s very hard to devise a risk analysis that would result in a conclusion regarding the probability of loss in transactions with certain given characteristics.

We really can’t say, for example: “the probability of loss in a transaction in which Seller A is owed money by Debtor B for the provision of goods or services with characteristic C, under the terms of an agreement with characteristic D ---is in the range of X% to Y%”

We’re just not there yet.

But there is a way to make a STATEMENT regarding risk even without actually suggesting an actual evaluation of risk. The statement I’m referring to is grounded in the traditional payback-period calculation from Financial Analysis 101.

Applied to a TRE auction it would take the form of:

“It would take X successful transactions (with given characteristics) to offset 1 such transaction in which a total loss was suffered.”

That says nothing about the actual risk of loss, of course, but it can provide a useful means of:

a) conceptualizing the consequences of a default, and

b) providing a relative measure of loss consequence between and among transactions.

Looking to the actual auction activity from a recent day I can pull an example of a transaction made at the low end of the typical range of expected duration and discount rate; one at the high end; and one in the middle.

The analysis has two steps:

1) using the likely duration and actual transaction parameters, calculate the expected net dollars returned to the Buyer upon full payment of the invoices purchased, and

2) divide the total funds advanced, including transaction fees and costs, by the expected net dollars returned.

The low-return transaction example, because of its short duration and low discount rate, produces a small dollar-amount of expected return. If a Buyer were to suffer a total loss on an auction with these exact characteristics, it would take 489 successful transactions with the same characteristics to recapture the loss.

The high-return transaction example has a very long expected duration and an unusually high discount rate. It would take only 7 successful transactions with these characteristics to offset a total loss on one.

The mid-range example, which is more characteristic of the typical transaction, has a moderate expected duration and a discount rate somewhat higher than the current average (but closer to the historical average) than either of our “extreme” examples. In that case it would take 45 successful transactions to recapture one total loss.

To my knowledge there has been no default on auctions involving the Seller/Debtor pairings whose auctions I’ve used in these examples.

It SEEMS clear that the Buyer community assesses the relative risks among them quite differently. But I wonder…..

Let’s take the case of the low-end example above. I don’t know who is buying those auctions or others with similar characteristics. But I wonder if they actually consider the relative risk to be roughly 1/11 that of the mid-range example?

I don’t know.

But I do know that it’s wise to remember two things:

1) The typical receivables transaction is of very short duration when compared to most other fixed-income investments; meaning that the ratio of dollars earned to dollars risked in a single transaction is low, and,

2) The direction of errors in projecting (percentage)returns on an invoice-purchase transaction is almost always negative. Other than earning a slightly higher return than expected if durations extend, you can’t really get surprised on the upside.

As time passes and experience is accumulated we’ll be better able to quantify relative risk using appropriate statistical means. Until then we’ll have to do the best we can with what we’ve got.

But I think it pays to at least acknowledge the implicit judgments being made until we get to the point where explicit ones can be supported.

Expanding the Product Line

The Receivables Exchange announced a new initiative last week: a facility to be operated separate from but in tandem with its current on-line market for the receivables of small-to-medium sized businesses.

The new facility is called the “Corporate Auctions” program, as distinguished from the current “SMB” program. It targets Sellers in the Fortune 1000 category of size and quality.

This is yet another big step for TRE and has significant implications for both current and prospective TRE Members on both the Buy and Sell sides of the market.

At the time of its announcement TRE also announced that a Fortune 10 company has signed up as the first Seller in the Corporate Auction program.

The identity of that Seller was not made public but I can say that you can find multiple products of that company in every American household (and most households and businesses throughout the developed world, for that matter).

TRE Buyers on the SMB platform will have to be separately approved to buy on the Corporate platform; and the size, character, bidding methods and pricing parameters will certainly be different on the Corporate platform than on the SMB platform.

The average transaction size will certainly be significantly larger in the Corporate program. And the returns to the Buyers will be significantly lower; reflecting the credit quality and the reliability of the financial data available on the Sellers and Account Debtors in that program.

After all, the great majority of the SMB Sellers are private companies providing Buyers only internally generated, unaudited financial information.

In the Corporate program the Sellers will very likely be large public companies with complete, easily accessed, audited financial statements already studied and commented upon by professional financial analysts.

Clearly, these markets will appeal to different types of Buyers with different risk appetites, costs of capital and motivations for participating.

This is an unambiguously good thing for the Exchange. Especially with the announcement of a “whale” as an initial Corporate program Seller.

But it is also an unambiguously good thing for those Members of the Exchange that stick to the SMB program.

Why is that?

First, this is an additional revenue stream for TRE. Obviously TRE is still in the stage of its evolution when it is burning through venture capital cash as it aims for break-even operations. The added revenue stream of the new program presumably advances the date on which break-even can be projected and thus it reduces the risk of all current participants, whether owners, Buyers or Sellers.

Second, to the extent that some types of Buyers are unable, because of the character of the SMB receivables, to participate in that program, this new initiative provides the potential for those Buyers to enter the market. This will bring not only their cash and their names to the Exchange but it will broaden the universe of financial market participants with actual Exchange experience. The more broadly the Exchange is known and understood in the financial markets the better, in the long run, for us all.

Third, to the extent that there are Buyers now active on the SMB platform whose actual needs and motivations are more aligned with the Corporate-type risk/reward profile, those Buyers now have an option to get what they really want, leaving the SMB program to those whose appetite and objectives more closely match it.

Fourth, while this might be considered a “stretch”, I see no reason why TRE could not go searching down the Corporate program supply chain for additional SMB Sellers. If we know that there is funding moving down the supply chain from a Corporate program Seller, that knowledge affects the assessment of the risk of buying paper due from a third or fourth-tier supplier of labor or material in that chain.

Fifth, if TRE can cite the participation of some very high quality Corporate program Sellers, it will very likely help them to attract new SMB Sellers. Even though the programs are different, the impact of quality-by-association should not be ignored. The imprimatur of a Fortune 10 industrial conglomerate does mean something to the owner of a smaller, privately-owned business.

I am not sure when TRE will be able to make more information publically available about this new program. And until TRE makes it public, confidentiality agreements keep the rest of us from doing it.

But, from my point of view, this is as big a deal as the announcement of the Bain funding earlier this year.

Both events take TRE a big step closer to an assured future.

Both mean that evolution in the receivables finance market continues to advance.

In the Way of St. Augustine

St. Augustine of Hippo is famously known for pleading: “Lord, grant me chastity-- but not yet!”

Augustine clearly viewed chastity as a good thing but he wasn’t quite ready to enjoy its benefits.

In the receivables purchasing business, getting paid is a good thing. But getting paid much more quickly than expected – not so much.

The “good” in a payment pattern is more a function of predictability than absolute timing.

Regular readers might now anticipate my returning to a pet peeve. But hear me out.

An auction that I purchased recently was closed out last week: paid as agreed; which, in itself, is certainly a good thing.

But it wasn’t paid either “as expected”, or “as advertised”. And that’s not such a good thing.

When I first started buying receivables in my spot factoring business I made the mistake of telling a couple of prospective clients that I was looking for invoices that would pay in the range of 45 days, give or take. It’s been my good fortune to have gotten many of my clients via referral from other clients. And it’s amazing how often I’m told that a referred prospect’s payment expectation is about 45 days!

Word spread from those initial clients, of course, that that’s what I wanted to hear—so that’s what I’m invariably told. But those prospective sellers are most often saying “45 days” when the truth is that payments will take LONGER. They know it and I know it.

The phenomenon I’ve griped about in prior posts with respect to TRE auctions is having invoices presumably due in 45 or 60 days that get paid in 7 or 9 days.

Why do I gripe about that? Not because it happens once in a while -- that can be just the luck of the draw.

But recently the number of instances of significant disparity between the posted payment expectation and the actual payment experience has been increasing. This is the case both in auctions that I’ve bought and in those I’ve just been monitoring.

The trigger for this post is that the auction I referred to above, which was paid last week, was “due” to pay out in about 30 days. In fact, the actual weighted average duration was 3.7 days!

The result was that my actual net earnings on that auction were roughly zero. That was a “successful” but still unsatisfying transaction.

I had bought several auctions from that Seller prior to the one I’m describing and all had performed as expected. I have actually been quite happy with the experience.

And as annoyed as I was with the result of that one auction, I was prepared to assume it was just bad luck. That for some unknown and unusual reason the Account Debtor just paid very early.

And so I bought another auction from the same Seller even before this one had closed-out.

The first payment on the NEW auction was made ONE DAY after the auction closed. The invoices that were paid were not “due” until mid September.

With that additional experience, as much as I have been pleased with the initial auctions bought from this Seller, I’m now going to have to stay away from those auctions until there is evidence that these are anomalous situations and that a predictable relationship between the posted due dates and the actual payment expectations can be anticipated.

The reason, of course, is that acceptable pricing changes significantly as the duration of auctions changes. This is particularly true in cases of very rapid payment.

The marginal impact on return of payments expected in 45 days but received in 30 is actually pretty minimal. But the impact of a payment expected in 45 but received in 10 is quite significant.

I am not suggesting that the Seller in this case deliberately misstated the payment expectation. I don’t believe that is the case. Actually, I suspect that this transaction will have been as unsatisfactory for the Seller as for the Buyer.

Because of the TRE fee structure, the costs to the Seller on this auction will have been as high on a relative basis as the return to the Buyer was low.

But there is a small number of Sellers whose actual payment experience appears to be consistently more rapid than “expected”. That creates a credibility problem—-for me, at least. And I’m no longer willing to bid on those auctions.

Ultimately it is to everyone’s benefit for there to be a rational relationship between the posted “due date”, the Seller’s posted “expected payment date”, and the actual payment experience.

I’m not suggesting that there won’t always be outliers in payment patterns, in both directions. Of course there will.

But when the pattern itself is of “outlier” events – it has to be the posted due date or posted payment expectation that is called into question.