Microfinance Focus, December 20, 2010: Something strange is happening with microcredit securitizations in India.  Two pre-eminent ratings companies, CRISIL (a subsidiary of Standard & Poors) and Fitch Ratings have taken diametrically opposing views on the credit quality of microcredit-backed securities.
Last month Fitch released a report, stating that “these transactions are unlikely to receive the highest Long- or Short-term ratings.”  Meanwhile, CRISIL continues to  issue high ratings to a number of microcredit securitizations, most recently rating a Rs. 54 crore ($12 million) pool issued by Asmitha Microfin as P1+(so), the highest rating that can be assigned to short-term securitizations.  This rating continues to stand, despite the fact that Asmitha’s debt has been downgraded since then and continues to remain under a ratings watch for further possible downgrades.
This disparity of views is unusual enough.  What’s particularly odd is that the two agencies are literally on opposing sides of one key issue, which we first highlighted earlier this year:  can the credit quality of an MFI’s securitized assets be independent from the credit quality of the MFI itself?
Fitch’s position is clear:  "the intricate servicing arrangements, coupled with a lack of institutional back-up collection mechanisms, make the rating of microfinance securitisations to be inextricably linked to that of the originator/servicer. In the absence of suitable mitigants with respect to such counterparties, Fitch will link the rating of the transactions to the rating of the originator/servicer."  Fitch does not describe what these mitigants are, but certainly, since the value of the asset itself is not independent from the originator, over-collateralisation would not be one of such mitigants.
On the other side of the spectrum, CRISIL, though it hasn’t made an explicit statement on this issue, has nevertheless left no doubt as to what its position is.  CRISIL’s assignment of P1+(so) rating in the face of a BBB rating of the originator is a clear statement that originator credit ratings are unimportant for microcredit-backed securities.  With the recent downgrade of Asmitha, this apparent disconnect has gotten larger, since to-date CRISIL has left the securitization rating untouched.
So who is right?  Whom should investors believe – Fitch or CRISIL?  The Asmitha rating may provide a good case study to answer that question.
The Rating and the Ordinance
In a perverse twist of irony, CRISIL issued the Asmitha securitization rating on Oct. 14th – the very day when the AP Ordinance was promulgated.  Since then, Asmitha’s collections in AP have effectively collapsed.  In downgrading the MFI’s debt, CRISIL recites some rather startling statistics:  Asmitha’s exposure to AP is 48% of its outstanding portfolio, while collections there remain at a distressing 15%, as compared to 99% prior to the Ordinance.
The rated security has an 11-month tenure, and provides 10% credit collateral, with a further 26% in subordinated receivables.   If the credit collateral is constituted of assets independent of both Asmitha and its portfolio, then in simple terms, an investor could not lose more than 90% of the principal.   The rest of the deal, including the subordinated receivables1, is affected by the collection problems in AP.   Assuming the deal has the same level of concentration in AP as Asmitha’s overall portfolio, the status quo situation suggests an expected repayment of about 74%2 plus an additional 10% from credit collateral, in other words a net loss of 16%.
That said, continuation of the status quo is the least likely outcome.  The political situation in AP may be resolved, which would presumably lead to some improvement in collection rates in the state.  But things could also get worse.
At this point, it’s difficult to imagine any scenario that wouldn’t involve extensive permanent defaults.  Such was the experience from the repayment crisis in the Krishna district in 2006, which had similar political overtones.  And the security could handle some defaults:  in order to pay security-holders in full, the final repayment rate in AP at the time of maturity must be at least 58% (or 41%, if the credit collateral is surrendered), with other states continuing to produce 99% repayments.  That scenario is certainly plausible and might even be likely, but is it the near-certainty that the P1+(so) rating implies?  Far from it.
First, the political situation in AP is yet to be solved – at a minimum, little is likely to happen before RBI’s Malegam committee submits its report, which is expected in mid-January, implying a minimum of 3 months of non-payment in AP and leaving only 8 months in the security’s term to collect the necessary sum.  Additional political wrangling will take further time.  And even once the political issues are settled and Asmitha goes back into the field, it will take more weeks and even months to reestablish contact with clients and convince them restart payments.  To imagine this as a quick process is optimistic, and anyone who has dealt with such widespread delinquency would know that such optimism is unwarranted.  And that’s the optimistic scenario.
Other possibilities leave even less room for optimism.  As a result of its huge exposure to the state, Asmitha has been under a serious liquidity strain.  In its downgrade action, CRISIL states that Asmitha has reduced its disbursements outside AP to 40% of total collections, adding that “a significant slowdown in disbursements made outside of AP could also result in a gradual drop in collections from those states, thereby leading to a further weakening in the company’s asset quality.”  In other words, the 99% repayment rate outside AP may not last.  Indeed, given the level of Asmitha’s exposure to AP, one cannot operate under the assumption that its future as an organization is assured, and if that happened, security-holders would find out directly whether their assets are independent of the issuer or not.  There is also the possibility that Asmitha (or more likely, its new bank owners, if it comes to that) may choose to abandon AP or its most-affected areas altogether, focusing their energies on the rest of the portfolio.  What exactly would be security-holders’ options then?
This is not exactly the range of scenarios one would expect to be looking at for a top-rated security.  The timing of the issuance and Asmitha’s concentration in AP suggests that this specific deal is probably among the most heavily affected securitizations, since the payment stream was impacted literally at the moment of rating.  But even if other microfinance securities are better-positioned, the issues in this case raise the question whether top ratings are appropriate for microfinance securities in the first place.  The crisis in AP is not the equivalent of a 100-year flood.  For a few years now, a number of commentators had suggested that a crisis of such magnitude was a very real risk.  A top rating, which implies the highest level of repayment assurance, simply cannot ignore the probability and potential consequences of such crises occurring.
Unhidden risks
The argument we made in our earlier article – that there are hidden risks in microcredit securitizations not captured in their ratings – is now being buttressed by numbers. The position taken by Fitch is very close to the one we laid out, so in this confusion of rating methodologies, we are of the view that Fitch is closer to the truth.
However, there is one other reason to believe this to be the case.  Unlike CRISIL, Fitch is largely an outsider to microfinance ratings in India, and has moreover done no ratings of microfinance securitizations.  Meanwhile, CRISIL has devoted significant energy to cover the Indian microfinance market, collecting fees for its efforts.  These fees and industry familiarity aren’t innocent – it has been widely contended that these were important factors in the overambitious mortgage security ratings by CRISIL’s and Fitch’s parent companies in the US.  These concerns eventually led to new regulations and oversight of the rating agencies laid out in the Wall Street Reform Act.
What this ratings confusion ultimately means is for investors and regulators to figure out.  In the meantime, we will continue to watch and report on this market.
----------------------------

References
[1] Note that the deal has an ultimate payment structure, meaning that the rating agency takes into consideration only repayments by the maturity of the transaction. While Asmitha is still expected to release principal payments month after month, CRISIL may take the position that its rating was only based on eventual payment by maturity, which gives CRISIL more flexibility.  We, however, base our analysis only on whether these payments will be made by maturity or not, without regard to their actual timing.
[2] 74% repayment rate calculated as:  (15% repayment in AP (48% of the deal)  + 99% repayment in rest of country (52% of the deal)) * 126% overcollateralizon level.