Planetary Alignment

There was a story on the news the other night about a rare occurrence that has allowed astronomers during parts of May to see 6 planets all lined up in a row in the morning sky.

Compared to that, my topic today is just coincidence.

But it IS the case that we “went live” as TRE Buyers on June 1, 2009. We had been lurking and studying for a couple of months before that, but today is the end of our second year of active trading on TRE.

As it happens, we bought our 500th auction today, which is a significant milestone for us.

And to continue the coincidence theme, this is our 100th TRE Observer post.

Two years is not a long time, really. But it’s a big part of the life of the Exchange.

The first auction we bought was #255, which will mean something to those who follow TRE activity.

And, as I write this (at 4:00 pm) we have bought as many auctions so far TODAY as we did in our first MONTH as an active Buyer.

In fact, one of the auctions we’ve bought today was bought from the Seller of our second auction. And today’s auction included one of the same Account Debtors as did that June 2009 auction.

This blog is in the process of analyzing various risks that deserve to be compensated in buying auctions on TRE and, as such, the current conversation inevitably tends toward the negative: concentrating on problems and difficulties. I don’t apologize for that: it’s reasonable and necessary.

But neither would I modify the words of the first TRE Observer post written on May 28, 2009:

“TRE is new and it is far from perfect. It will inevitably be required to make adjustments as experience teaches its operators and its users some valuable (and some potentially expensive) lessons.

All beginnings are hard.

But TRE is a game-changer. It is disruptive.

Those whose businesses face disruption will scoff at first and then resist but they will ultimately adjust to the new reality because they will have no choice.”

Many would have bet at that time that TRE wouldn’t still be around today. They would have been wrong.

Not only is TRE still around but there is another online invoice trading platform about to go live in the US; there is one that is live and growing in the UK; there is one just getting started in Sweden; and, there is one that is live in Germany.

What's that about the highest form of flattery?

Some of these markets might thrive. Some might fail. But my bet is that the idea is here to stay. And that the industry will ultimately adjust to it.

Next stop 1,000!

Appropriate Compensation # 4: Financial Statement Risk

The issue of the reliability of TRE Seller financial statements was brought up very early in the history of this blog.

Our post of June 23, 2009 concluded:

“All else equal, the lower the level of assurance that the Seller’s financial data is reliable, the higher the appropriate Buyer’s risk premium.”

In our discussion at that time we pointed out that the Reliability Task Force of the American Institute of CPAs, under the criteria of its March 2008 paper on the subject, would include unaudited, unreviewed, management-generated statements in the category of having “no assurance” of reliability.

On the other hand, we noted that “the fact that an accountant has not certified the accuracy of the data presented in a financial statement doesn’t mean that the data is not accurate”.

There’s nothing in that June 2009 post that, nearly two years later, I would change.

But we do now have two more years of actual experience analyzing the financial statements submitted by TRE Sellers. And that additional experience does allow us to make some more observations.

One interesting phenomenon is that the percentage of TRE Sellers providing audited financial statements has actually fallen over time. The percentage was not high in earlier days but today it is almost zero. In fact, very few Sellers now provide even reviewed statements.

One reason, I suspect, is that TRE doesn’t REQUIRE audited, reviewed or compiled statements. And if a Seller does not HAVE to spend the money on such statements, its cost of obtaining financing is reduced, perhaps significantly, by eliminating that expense.

That’s only true, of course, if the Buyers don’t require a higher risk premium to compensate for the lack of assurance that a third-party review would provide.

Having had feedback from some Buyers on this subject I think I can say that there is a segment of the Buyer population that is looking quite closely at the Seller financials and using that information as a key element in the decision to buy or not to buy.

On the other hand I think it is clear that there are many Buyers who do not spend any significant time analyzing Seller financials, presumably on the assumption that the Account Debtor is the more important element in the decision process.

[One clear indicator of that is the speed with which the auctions of some first-time Sellers are bought. It has been clear in many instances that the Buyer had no time between posting and purchase to examine the due diligence materials available.]

That is perilous not only for the Buyer but also, ultimately, for TRE.

If any element of risk is materially mispriced, sooner or later it will generate losses that have not been adequately anticipated. Those financial losses will generate loss of confidence, which might actually prove the bigger long-term threat.

Since there is a Buyer for nearly every TRE auction, including those offered by Sellers whose financials some would find unacceptable or lacking in credibility, we have to conclude that Buyers who are NOT guided by the analysis of Seller financial capacity are probably setting the marginal pricing levels.

If there are Buyers willing to bid without giving effect to the financials then the Buyers who DO take the financials into account are faced with a binary decision.

They are either willing to buy or are not willing to buy based on their view of Seller capacity.

It is not yet possible, in my opinion, to implement an incremental pricing strategy that might, for instance, require an additional 50 basis points of monthly discount to compensate for a relative lack of confidence in one Seller’s financial statements versus those of another Seller.

Anyone who has analyzed the financial statements of TRE Sellers, which are largely privately-owned small to mid-sized firms, will have encountered some systemic issues that arise from the nature of that business structure. The accounting for owners’ contributions and distributions, for example will tend to penalize liquidity ratios and distort equity accounts. The desire to minimize taxable income, sometimes at the expense of a true picture of profitability, is also frequently apparent.

And management-generated statements for companies whose management doesn’t devote much attention to such matters can produce results that might charitably be termed idiosyncratic.

One of the most interesting phenomena is the change in statements over time by Sellers who are forced to actually produce and provide statements on a quarterly basis in order to remain in compliance with their agreements with TRE. I’ve noticed a number of cases in which it’s clear that increased attention produces improved product over time. Messy “legacy” issues on the books are cleaned up. Items are reclassified to more appropriately reflect the character of insider obligations. And so forth.

But the overall conclusion remains the same, in my opinion. There is meaningful, if not directly measurable, risk that the financial statements of the TRE Seller community, as a whole, would be found wanting if audited.

Given the obligation of Sellers to repurchase invoices not paid by Account Debtors that financial statement risk should command a pricing response.

Appropriate Compensation #3

In our last post we suggested, with a number of caveats, that a reasonable full-cycle credit loss assumption for the traditional factoring business might average something in the range of 10% to 12.5% of gross revenues.

Two of the critical caveats were that :

1) actual experience would vary substantially over the course of the credit cycle, and

2) different customer segments of the factoring industry faced widely varying levels of probable loss: for instance, loss experience in staffing and transportation might differ significantly from loss expectation in manufacturing and construction-related industries.

It’s been interesting that my invitation for comments or differing opinions on a baseline loss level has elicited little response. Readers of this blog have become more and more willing to venture opinions, especially contrary opinions, over the course of its history. No response received proposed a specific adjustment to the suggested range.

I don’t know whether to interpret the relative silence as tacit agreement that it’s just a tough subject to document; or that the conclusion seems to be about right; or that it’s just not worth a response.

But I’m going to assume for the moment that it strikes readers as at least in the ballpark. And I’ll go on with the exercise.

Having ventured an opinion on a baseline, it is now time to move to discussion of issues specific to TRE that would cause a loss estimate specific to TRE to vary from the baseline. There are a number of these but we have to acknowledge that it is going to be impossible to attach an incremental risk measure to individual issues.

The best we are going to be able to do is to lay out the issues, consider the risk implications of each independently, and then try to assess and quantify the risk impact of all the issues as a single adjustment.

Neither our individual assessments nor any overall suggestion of differential risk is going to be defensible on the basis of hard data.

I cannot propose any way to reliably quantify the incremental risk of assessing Seller suitability on the basis of management-generated financial statements vs independently reviewed statements or audited statements, for example. But I doubt that anyone would argue that unaudited statements should be given the same level of credence as audited ones. There is SOME incremental risk, no matter how difficult it might be to quantify.

With all that said, it is still appropriate for each Buyer to CONSIDER both the baseline level of risk in the business of purchasing receivables AND the incremental risks attributable to the TRE process.

It’s also only fair to add that there are some elements of the TRE process that might act to REDUCE risk, at least as compared to certain invoice purchasing models, and that those need to be considered and provided for as well.

So this process will involve identifying and discussing a number of individual issues over a series of posts and then a discussion of and suggestion for quantifying an adjustment factor for application to TRE transactions.

Since I’ve used the financial statement issue as an example I'll address that one in the first post of the series.

I invite any reader with either information or thoughts on the relative risks of relying on management-generated financial statements versus compiled, reviewed or audited statements, to share that information either publicly or privately with the understanding that I will respect all requests for confidentiality.

Appropriate Compensation # 2

It has been about six weeks since I wrote the first post in what is intended to be a series on the issue of “appropriate compensation” for the risk of loss faced by TRE Buyers.

I acknowledged at the time that this would not be an easy or straightforward task. And that has proven to be an understatement.

In attempting to establish a credit loss benchmark for the factoring business as a whole there are a number of difficulties that have to be acknowledged. Among them are:

 The industry is highly fragmented and relatively few of its participants are public companies with public reporting requirements.

 For most of the public companies that do engage in factoring activities, those activities are a relatively minor part of their overall business and their financial statements do not disaggregate results in a way that allows the isolation of factoring income and expenses. [Note: New FASB reporting standards will apparently allow factoring results to remain unclear due to the exclusion related to carrying values and the duration exemption.]

 The information that IS available on an industry-wide basis does not allow the isolation of credit loss figures that are mitigated by the employment of credit insurance.

 Disaggregated loss experience that IS available makes it clear that there are some segments of the factoring market that tend to generate much higher losses than average and some that generate much lower losses than average. This is the case both in terms of size of business and industry segment.

 Available information supports the intuitively reasonable notion that losses are cyclical with the change in overall economic conditions. So the most recent periods provide loss experience that is probably at the top end of a reasonable longer-term benchmark.

 Available information is not reported in a uniform manner: some reports measure losses as a percentage of income; some measure them as a percentage of average receivables owned; some report as a percentage of total receivables purchased; and some report as a percentage of average net funds employed.

Presentation of a full analysis of the data I’ve gathered and studied is beyond the scope of this post. And I could not present it here in a way that is fully documented in any case.

What I CAN do is to say that I’ve spent a reasonable amount of time on this question and gathered and analyzed what information I have been able to find in a pretty diligent search.

I can also say that, no matter what difficulties there might be in quantifying the risk numbers, it is appropriate to actively ATTEMPT to quantify them and to consciously incorporate a risk premium into the pricing of all factoring transactions, including those that occur via TRE.

So I’ll share with you what I think is reasonable, without suggesting that the research and analysis is actually robust enough to withstand a rigorous technical challenge.

There are undoubtedly others whose databases are better than mine. And there are certainly others whose primary activity is analyzing such data, while mine is not. And I invite anyone to provide alternate or additional information, either publicly or privately.

And finally, these figures are not those that I would suggest are appropriate for TRE transactions. These are meant as a starting point; a baseline for the typical factoring business BEFORE adjusting for the specific TRE approach to the business.

So, with all of that as preface, these are my suggestions at this point:

1. The nature of the TRE transaction requires that whatever adjustment for risk is chosen, it needs to be applied in the pricing of each transaction. There isn’t really a portfolio-level adjustment mechanism when you’re operating in a spot-factoring auction market.

2. I think the most useful way to state and to adjust for loss expectation for our purposes is on the basis of a percentage of gross revenue. I think that allows the most straightforward means of adjusting pricing parameters at the level of an individual transaction.

3. It is necessary to decide whether to choose a single loss expectation figure based on an assumption about a full credit cycle and a typical transaction. This would be in the nature of a fixed reserve allowance against which actual losses would be charged as incurred. It would presumably be higher than actual losses during relatively good times and lower during relatively bad times.

4. If the single reserve figure does not appear appropriate, then a range of loss expectations over the course of a credit cycle might be established, and at any point in time the actual allowance assumed would reflect the current state of the credit markets.

5. It is also necessary to decide whether to choose a single loss expectation figure for all transactions or to vary the allowance based on the character of the transaction. One level of adjustment might, for instance, be based on industry e.g. adopting a different allowance for staffing industry transactions vs manufacturing industry transactions or transportation industry transactions.

6. These are relatively “high level” adjustments, of course. Specific underwriting criteria and other matters affecting risk would also obviously need to be reflected in the pricing.

7. Based on the data I’ve seen it appears to me that a baseline risk adjustment that reflects the “typical” factoring transaction would range over the course of a credit cycle from somewhere in the mid single digits; say 5% of gross income; to a level much higher than that when we’re in the worst phase of a contraction; say 25% of gross income (or even more in some recent cases).

8. It appears that, over the course of a cycle, there are generally more years with experience towards the lower end of the range than at the peak of contraction. So, if we were to pick an average to use over the cycle, it would probably fall below the mid-point of the range.

9. It seems to me that an average credit loss allowance in the range of 10% to 12.5% of gross income in typical factoring transactions, over the course of a credit cycle, is not unreasonable.

This series of posts will continue with discussions of the differences between the “typical” factoring transactions and TRE transactions, focusing on the distinctions between the two that affect risk and loss expectations.

To make the point again: I welcome any comments on this issue and especially any data that might help make the analysis more robust.