Best Practices for Investing in Cédulas de Crédito Bancário in Brazil

If you’ve not heard of Morada Bank, it was a spectacular bank failure wrought with fraud in April of 2011 that is still causing headaches for its investors and regulators. We believe that this fraud, perpetrated by Morada Bank and ultimately costing the Petrobras pension fund Petros’ R$72 million in losses, provides a heedful case study about what can go wrong with an investment in Brazil in general and in Cédulas de Crédito Bancário in particular. In this post, we will explore what happened, and try to understand how a rated national bank could have perpetrated this fraud. We will end with some thoughts on due diligence procedures and processing that will help prevent these types of artifices in the future.

At the center of scandal are the securities in the form of structured bonds called Cédulas de Crédito Bancário (CCB) backed by payroll loans or “consignado” loans underwritten by Banco Morada. CCBs are a form of collateralized debt that can be issued with or without guarantees.  The Banco Morada loans were subsequently sold to three companies (Allcred, Secred, and Morada SPE) that were effectively owned by the executives of Morada Bank. These three sister companies then used the payroll loans to collateralize the CCBs that they sold to Petros and other investors.

Valor Econômico, which reported on this story on March 16, 2012, said that Morada Bank transferred loans to Allcred, Secred, and Morada SPE to “clean up” its balance sheet and to allow it to reserve less regulatory capital as provisions against bad debt. It can be argued that these financial entities were effectively holding companies for Morada that allowed the bank’s executives to perpetrate this fraud and keep it undiscovered until April of last year. In all, this scam continued from 2007 until April 2011 when Morada was seized by the Banco Central for violating “disciplinary legal norms” and moreover because it was insolvent.

Basically, it was supposed to work like this: Morada underwrote the payroll loans that are collateralized by a lien on a government paycheck. These loans are considered less risky because the borrower doesn’t have to make a payment; it is automatically deducted before the borrower receives their paycheck and the payment is sent to the lender by the employer.  However, lenders must apply to employers to be allowed to enter this program.

The loans were credit enhanced (115% overcollateralization, a liquidity fund, and backed by Allcred’s equity shares) and rated AA+ by LF Ratings, a Brazilian rating agency. The custodian, Banco Bradesco, was to receive the payments from the loans from Morada not less than five days after they were made. LF Rating reaffirmed the ratings of the CCBs of Allcred AA+ and also gave a AA+ rating to Semear’s CCCBs, CCBs backed by Allcred’s CCBs, no less than two months before the Central Bank intervention in Morada.  LF Ratings also ratified their AA+ rating of the CCBs at this time.

This is the truly important part: Morada was transferring the loans to its cohorts but retaining control over the collection and servicing of the loans themselves. Morada also sold off the loans with better credit classifications than would have otherwise been given to the same loans had Morada kept them. According to Valor, information about transferred loans in future deals will be cross-checked by the Central Bank with the records at the issuing bank. But this seems impossible to know, as many of the loans pledged to the CCB were never found after intervention last April. This is according to Fabio Mazzeo, the president of Metrus, an institution for the workers of the São Paulo Metro that invested R$23 million in the CCBs.

But that is not all, the Valor article points out that Morada used a play right out of the PanAmericano fraudster’s handbook and diverted prepayments on loans that were pledged to other institutions and commingled those prepayments with their own income. This happened in spite of Morada’s legal obligation to send those monies to their rightful legal owners in order to mature the loans and return principal to the investors. However, the loans were sold “without recourse” in a “true sale” transaction and Morada had no economic interest or incentive to service the loans correctly.   Unsurprisingly, this practice ended up being a lot of money; on average, around 30% of all payroll loans contracts were prepaid prior to maturity during 2010.

In our opinion, if a for-profit institution wants to commit fraud and break the law, new laws and regulations are not going to prevent them from doing so. A better mousetrap will only beget a better mouse, and real change will only come when structured transactions bring in third-party servicers to remove the moral hazard and the asymmetrical information from the equation.

The entity acting as agent, in this case Morada, is the same one that is receiving cash flows and payment information. Therefore this agent is the sole gatekeeper for information and funds for the investor, or the principal. This payment system for the Allcred CCBs is common (it is also used for some of other structured finance vehicles such as a few FIDCs) and was set up as in Figure 1.

It comes as no surprise that Morada was tempted by the large moral hazard to exploit the principal, the one party in the transaction who has equity or “skin in the game.”  Morada perpetuated the shell game to mask shoddy underwriting and credit controls. Since Morada was never putting capital at risk, why not?

Figure 1

Best practices would dictate that the “servicer,” or collector and workout agent for payments from borrowers should be a legally separate (or at least ring-fenced) entity that can provide full disclosure to the investors whom are the principals in this transaction. Since the very nature of government-sponsored consignados makes having a separate servicer difficult to implement (remember that these banks have special agreements with government institutions that allow them to collect on payrolls before the borrow has access to the funds) a sensible solution would be to ring-fence a separate box or create a strict firewall within the bank that receives the payments form borrowers.

With that small change, Figure 2 illustrates how borrower funds and payment information could flow directly to the investors and help keep the underwriter honest by having a “check” upon their flow of capital and information. Of course, they could still create value by selling these loans into the structure at more than their cost to originate, but their long-term incentive shifts away from payment fraudulence and towards underwriting loans.


Figure 2

All this is made more relevant today by the news from the end of March 2012 that Valor Econômico that Banco Central is bearing down on institutions whom are using structured products to create off-balance sheet leverage. If thoughtful structural changes are implemented, it will help investors and improve transparency.  However, we fear that the well-intentioned Banco Central action will only serve to help the Banco Central. Regulations and laws cannot replace creating a best practices framework for structuring cash flows.

To summarize, investors must realize that CCBs are very flexible securities that can be issued with bespoke terms and different forms of guarantees on the collateral.  This flexibility can work for or against investors. They must perform thorough due diligence to discover all of the interested parties and understand the security’s terms and covenants completely. In this case, the servicing arrangement was a giant red flag. The covenants of the CCBs collateralized by Morada loans could have specified the types of payment mechanisms that we are advocating.

Another giant red flag was the fact that executives of Banco Morada were also the owners of the legally separate companies Allcred, Secred, and Morada SPE.  This is a typical configuration for executing fraud in Brazil. Usually, such fraud is masked more effectively through “laranjas”, holding companies that are set up with a figurehead executive but effectively controlled by another person.

A thorough due diligence would have looked at why Banco Morada itself did not issue the CCBs. This was another red flag. Furthermore, why did Morada transfer the loans to companies with no balance sheets?  According to the March 16, 2012 Valor article, the three shell companies issued the CCBs and then paid Banco Morada for the loans with the proceeds from the investors.  Also, the same article points out that Banco Morada actively sold portfolios of payroll loans to other banks.  Why were these loans sold separately and not included in one of the portfolios sold to another financial institution?

Investors must also realize that Brazilian custodians and administrators do not always protect the investors.  As we stated earlier in the article, Allcred executed a fiduciary assignment of payroll loans with a value equal 115% of the face value of the CCBs to over-collateralize the CCBs.  Second, there was a separate amortization account that required 200% that needed to be maintained at 200% of the next payment to investors.  The third layer of protection was equity shares of Allcred that were put in a bank account with Bradesco for additional collateral.  We have to question how all of these various levels of collateral disappeared so quickly and Banco Bradesco, one of the largest banks in Brazil and in the world, could do nothing to prevent the complete loss to the shareholders?

Finally, the covenants can be written with tougher terms for the transfer of the loans to the CCB and to specify an independent servicer for the CCB.  For instance, the terms can specify eligibility based information from Serasa or other credit agencies and can be retroactive.  The covenants can be written to allow the CCBs to “put back” the ineligible or fraudulent loans.

Given the sophistication of Brazilian transaction system, it seems possible that an elegant solution is available to break the principal-agent dilemma, significantly reduce the moral hazard, and protect investors in Brazilian ABS structures.

Speak Your Mind