Thursday, 18 March 2021

Market Commentary: Greensill -- The Critical Difference between Insurance and a Guarantee and Why It Matters

 

An Unhappy Outcome

Since I haven’t seen anything on this topic in re Greensill, I thought I’d offer a few thoughts on how the fundamental difference between (1) a guarantee of payment and (2) an insurance policy affects the Greensill “situation”

And how it might motivate actions by participants in this unhappy event.

The difference between these two instruments is frequently misunderstood, including by supposed finance professionals. Hopefully, this post will fill in any extant knowledge gaps.

A guarantee of payment (as opposed to a guarantee of collection) is a legally binding obligation by the guarantor to make payment to the guaranteed party if the debtor does not make a scheduled payment. Proof of the debtor’s non payment is generally fairly “easy” to make. Usually then the guarantor makes payment without undue delay.

An insurance contract is a legally binding obligation by the insurance company to pay the policyholder if the policyholder submits a valid claim.

Keep those last two words in mind. 

The insurance company reviews the policy conditions, the insured’s (or policyholder’s) actions, and makes the initial determination of the validity of the claim. Some policyholders have been known to complain that such assessments seem to move at a glacial pace.

As that should imply, the insurance company has more legal defenses against payment than a guarantor. And its payment is not as fast given the time to review the claim.

The insurance policy spells out the conditions for validity.

For example, in obtaining the policy, did the policyholder make a material misrepresentation or fail to disclose material information that would reasonably have caused the insurance company to refuse to write the policy? In such a case the entire policy is invalid.

Did the policyholder fail to take reasonable steps to prevent the loss?

For example, if he left his Maybach unlocked with the key in the ignition and his insurance company knew this fact, they would likely decline the claim for theft.

If she routinely stored gasoline in her villa and filed a claim for fire damage and the insurance company knew this fact, the result would be the same.

Did the policyholder take reasonable steps to mitigate damages?

When the fire broke out, did she call the fire department? Or just let the villa burn down?

If his trade counterparty was in financial difficulty and he should have been aware, did he shorten payment terms, ask for collateral, lower his credit limit for aggregate outstandings?

There may also be other specific policy exclusions: strike, riot, civil commotion, actions of political entities, foreign exchange controls, etc.

We can therefore expect that Tokio Marine and other insurance companies will be carefully reviewing their obligations under any outstanding policies on Greensill related debt. 

I saw in today's FT (23 March) that Tokio Marine had opined that the policies might not be valid

Today (2 April) the FT reported that Grant Thorton acting as administrator for Greensill had been unable to verify certain invoices underpinning loans to Liberty Commodities - part of Mr. Gupta's group.  

Actually, the article says that several firms whose names appeared on invoices denied any commercial relationship with Liberty.  You can guess that this means that any "insurance" on these invoices is invalid.

One would of course have to review the actual policies and the respective governing laws to determine the defenses the insurance companies might have.

But I wonder if it’s possible that policies issued in excess of underwriting limits might be one? 

Part of that might turn on whether Mr. Brereton was working for Greensill (as an insurance broker) or for Tokio Marine (as its employed underwriter).

As well one can imagine Credit Suisse fund managers' angst over the difference between insurance and a guarantee as well as potential liabilities that might arise from potential "defects" in disclosures in selling documents vis-a-vis disgruntled clients whose attorneys will be going over said documents carefully.

Keep up to date on developments.  

The FT continues to follow the Greensill saga with an interesting article on Mr. Brereton earlier this week.


Sunday, 14 March 2021

Market Commentary: Tesla "Loses" One-Third of Its Value

 

Make Sure Your Weighing Machine is Properly Calibrated

Just a few days ago, I read courtesy of Reuters that Tesla had lost one-third of its value

Shocked, I rushed to read how such a loss had occurred.

Had Brother Musk misplaced or “lost” the “code” to Tesla’s Bitcoin account?

Did meteors strike Tesla’s factories, wiping out needed capital assets?

Did Lucid leapfrog Tesla's self-driving technology?

I read on.

Rather the article was about the decline in the price of Tesla stock.

The writer of the headline apparently is a naive adherent of the efficient market theory conflating stock prices with value.

So what is the point?

There is a difference between the price of a stock and its (intrinsic) value.

Many tragedies in the investment world have occurred because of a conflation of the two.

Market sentiment plays a large part in the price of a stock.

One day an NMC or a Wirecard are flying high. The next day they are not.

When 911 occurred, prices on the NYSE dropped dramatically forcing the closure of the market.

In both cases there was a wide gap between value (reality) and price (sentiment).

Prior to the price decline NMC and Wirecard had high prices, but no value, unless one were to count negative numbers.

In the second case, the stocks on the NYSE as a general group did not suffer any real loss of value. Their prices just diverged from value.

Earlier this year, one of my colleagues gave me a JPMorgan research piece on Tesla which posited a value of some USD 160 or so a share.

JPM had computed the value using multiple “different” methods, though as Aswath might tell you many of these seemingly independent methods are really fundamentally linked.

I found it entertaining but not convincing reading. The JPM research piece not the Professor's

A sum of the parts analysis in a distressed sale might have been more illuminating.


Thursday, 11 March 2021

Market Commentary: Bill Gates on Biden's USD 1.9 Trillion Covid Relief Bill

Answers to All Your Questions

Announcing a new feature here at SAM: AA’s trenchant commentary on news and developments in the "market".

I saw on the internet about two weeks ago a Fareed Zakaria interview in which he asked Bill Gates to opine on the Biden USD 1.9 trillion stimulus plan.

I was surprised.

Prior to that, I hadn’t known that Bill Gates was an expert on economics.

As Phil Rosenzweig can tell you, success in one field, particularly one in which an individual makes billions, automatically confers unique knowledge in almost every other field on that individual or at least the appearance of such knowledge.

Often such knowledge is attributed by folks who one hopes should know better.

Given Fareed’s academic and professional focus on foreign affairs, I was surprised that he did not seize the opportunity with Bill to heal an unfortunate rift in the Middle East by asking Bill to provide the definitive analysis of the meaning of “غَدِيْر خُمّ “ and “أَهْل ٱلْكِسَاء‎ “.

Or perhaps give his solution to the Korea issue.

Sadly, for whatever reason, he did not.

One or is that two for the “missed opportunities” file?

I, of course, would have had my own set of different questions.

Before outlining these, I need to make a material disclosure.

Devoted readers of this blog (I’m counting bots so I can use the plural) know that there is a bit of bad blood between Bill and me.

Sometime back I was expecting advice from him on what I should be having for dinner, hoping to draw on another area of his wide ranging expertise after my foray in a mall bookstore's business books section.

Advice that sadly never came.

Madame Arqala, as she so often does, did rescue me on that occasion.

Despite a bit of lingering rancor on that failure, I would have straightaway asked Bill what strategy he would employ as Arsenal’s new head coach to ensure that they repeatedly won the Premier League, the Champions League, etc.

All in the hopes that Brother Stan was watching. Or might see the interview later on the VAR.

I'd probably have moved on from there to ask him to opine on a sharp difference between my elder wiser brother (expert in many thing Asian though clearly not on pizza) and me over the best pizza:  deep dish or thin crust. 

Or perhaps why the last two words in the fourth verse of Surah 112 did not have the same terminal vowels. 

Eventually I’d probably have asked Bill to comment on SolarWinds and Microsoft Exchange.

Why these events happened?

What Microsoft could or should have done to prevent them?

Perhaps, an area where his skills might be more profitably employed.

Friday, 26 February 2021

Thursday, 25 February 2021

Citibank's USD 902 Million Revlon Payment - What's Not in the News

 


You’ve probably been reading about the mistaken USD 900 million payment made by Citibank acting as agent for Revlon.

As as well as the decision by the US District Court, Southern District of New York, that the lenders to Revlon who received Citibank’s erroneous payments of principal were not legally obligated to return the funds.

Let’s take a closer look as I think some details of the story haven’t been highlighter. 

Here’s the link to the 105 page decision.

As a starter it's a good introduction.

It’s well worth a read as it explains the background and facts of the erroneous payment and the legal precedent Banque Worms vs Bank of America International (Banque Worms) which “led” the court to its judgment.

Banque Worms establishes the right of creditor who receives funds in error from a debtor or the debtor’s agent to retain the funds and apply them against the debtor’s outstandings as long as the lender had no knowledge of the error when it received the payment.

This precedent thus overrides the concept of equity that typically governs payments received in error: if one receives funds in error, one is obligated to return the funds.

Background and Some Clarifications

The “USD 900 million” payment was actually for approximately USD 902 million: approximately USD 894 million in principal and USD 7.8 million in interest. Citibank received funds from Revlon for the (authorized) interest payment, but used its own funds for the (erroneous) principal payment.

Citibank were able to recover about USD 390 million from the principal payment, reducing its loss to some USD 504 million. 

The USD 894 million principal payment was made to two groups of lenders. Allocations are based on my calculations using data in the court decision.

  1. The first group received USD 341 million. They were involved in a non-cash exchange of one Revlon debt for another (a “roll up”). Based on the terms of the exchange, they were not to receive any principal payments. Thus, they would not be able to resort to the Banque Worms precedent cited above. This group returned the full USD 341 million in principal payments they received.

  2. The second group received USD 553 million. This group was not involved in the debt exchange. Therefore, they had no right to any interest or principal. Under Banque Worms they had a legal defense to retaining the funds because they did not “know” the transfer was in error when it was received. An important part of the refusenik lenders’ defense was that they were unaware of the “roll up” involving the first group. Some lenders in this group returned USD 49 million either because they or they agents were unaware of the possible Banque Worms defense or decided to return funds despite such knowledge. 

At first blush, this sounds like a case of human error.

A “fat” finger, a misplaced decimal point, or an error in calculation.

But there’s more to the story than that.

Each lender received a principal payment in the exact amount of its principal balance as well as the exact amount of interest due it to the date of the payment – which was not an interest payment date.

The error was due to a combination of factors – human plus limitations in Citibank’s perhaps inaptly-named Flexcube system.

Citi was attempting to perform a complicated and uncommon transaction. 

  1. The first “bit” of complexity was that this was a “roll-up” transaction. Some lenders were exchanging existing Revlon debt (positions in the 2016 Revlon loan) for positions in other Revlon debt.

  2. The second “bit” of complexity was that not all of the lenders in the 2016 Revlon debt were participating in the “roll-up”.

The lenders participating in the roll-up—whose positions were managed by Angelo Gordon (the Angelo Gordon Lenders)--were entitled to a cash payment of interest on the existing debt up to the date of the “roll-up”. They were, however, not entitled to any cash payment for principal. Revlon owed this group USD 341 million in principal.

Those lenders not participating in the roll-up were not entitled to any cash payment of either interest or principal. Revlon owed this group USD 553 million in principal.

What were the problems? And what were the “issues” surrounding Citibank’s systems?

Citi’s system could not process an interest payment to just the Angelo Gordon Lenders.

All lenders had to be paid interest.

To get the exchange done, Revlon agreed to pay (cash) interest to all lenders. Some USD 7.8 million.

If my calculations are right, Revlon paid approximately USD 4.9 million in interest to the non Angelo Gordon lenders. Interest that it did not have to pay on that day.  

In effect it prepaid this amount of interest. 

Small beer perhaps, but perhaps not so small for a company in Revlon's financial position.  

As well as, perhaps, opening itself to the charge of creditor preference were the egg to fall off the proverbial spoon. 

Further complicating matters, to achieve the “roll up”, the principal balances of the Angelo Gordon Lenders in the 2016 Revlon term loan had to be reduced.

Citi’s system required this be achieved using a principal payment.

To do this without the movement of cash, the principal payment would have to be made using internal Citi “wash accounts” (General Ledger accounts). The amount of the principal payment would be funded by debiting an internal Citi G/L account. Then the principal payments to each roll-up lender would be “paid” into an internal Citi G/L account. Then (presumably) each lender’s principal payment would be be used to purchase an equivalent amount in the other Revlon facility.

But there was the same problem as before.

Citi’s system could not apply this mechanism to only the Angelo Gordon Lenders.

All lenders would have to be “repaid”.

That means that the non participating lenders’ shares of the 2016 loan would have to be "repaid" using the Citibank G/Ls as above. But with a key difference. In the final step instead of “purchasing” shares in another Revlon loan, these lenders would “repurchase” their shares in the 2016 loan.

That wasn’t the only wrinkle.

To ensure that interest payments would be made in cash and that the principal payments would not, Citi’s ABTF (Asset Based and Transitional Finance Team) had to check three boxes on the Flexcube payment system: FUND, FRONT, and PRINCIPAL to override the standard settlement instructions. They also had to input the related G/L account numbers in each of the three fields associated with these boxes

You can see a “screen shot” of Citi’s Flexcube system on page 13 Figure 1.

This seems a rather complicated system. Unnecessarily complicated.

One that is not intuitive. What does FRONT mean? What does FUND? And thus subject to mistake.

Citi’s agents (Wipro) checked the box “override default settlement instructions” next to PRINCIPAL as well they input the internal G/L in the required field.

They did not do the same for FUND and FRONT.

There are two possibilities here. 

The first case: If settlement instructions means the details of where Citi was to pay the funds (to the creditors) then it would appear Flexcube “ignored” the two override inputs for PRINCIPAL.

I would have expected it to decline the transaction with a note “incomplete” or “inconsistent instructions”.

Also if the standard settlement instructions for the principal payment were “overridden”, shouldn’t the system have been unable to make the payment as presumably it did not have other settlement instructions?  

The second and more likely case: Settlement instructions meant the accounts that Citi should debit (the debtor's), the system processed a payment against an internal Citi G/L account that was it seems (note that caveat) designed for non-cash transactions.  

With 10/10 hindsight, it would seem that ideally the system should have been programmed to refuse such a transaction.  

Or perhaps more likely programmed to raise a warning -- You are attempting to make an external cash payment against an internal non-cash account.

As well, it would have been ideal if the warning message associated with the final release of funds had been more explicit. For example, “You are about to release an external cash payment of USD 902 million: USD 894 million in principal and USD 7.8 million in interest”.

Somehow the fact that Citibank had not received USD 894 million from Revlon and was “going out of pocket” was not reflected in Flexcube.  

Typically agent banks require a borrower to remit funds to a special “loan agent” account to cover loan repayments and only release payments to lenders when receipt of funds is confirmed. Banque Worms is one reason why.  

There are of course other reasons. The Agent wants to be certain it has received the funds and that the funds are in an account under its control, not the borrower's account.

As indicated above, and with admitted 20/20 hindsight, it seems there were several steps in the process where a fail safe mechanism in Flexcube could have prevented the payment.

That being said, the human element is not innocent here.

Citi’s Flexcube manual contained detailed instructions.

The three Citi employees charged with input of the payment, checking of details, and authorization (release) apparently assumed that checking the PRINCIPAL box and inputting the G/L number was all that was required.

The fact that this wasn’t a common transaction, that Citi’s mistakenly named Flexcube required workarounds to complete the transaction, that the size of the transaction was "large", and that Revlon’s credit was weak should have prompted additional steps by Citi’s staff.

Tuesday, 2 February 2021

ذكرى وفاة أم كلثوم 3 فبراير 1975


 

"  لاَ يَرُوقُ الْوُجُودُ مِنْ دُونِها "

" لاَ أَرَى الْعَيْشَ مَا تَفَكَّرْتُ فِيهِا "