Shedding Light on SIV's

Structured Investment Vehicles -- otherwise known as SIV's -- have attracted considerable attention in the media over the past 5 months. But investors and the general public are still, to a large degree, in the dark concerning what exactly SIV's are, what they are invested in and to what degree they pose a threat to the overall economy. Information available to the general public on SIVs concerning real operating and financial data is sparse, which has provided a fertile breeding ground for tales of financial meltdown, Armageddon and woe, and less occasionally, an article that states everything is fine. (articles which investors have seen on a daily basis almost over the past few weeks from the financial media).

The lack of publicly available information concerning SIVs can be illustrated by the fact that, if one searches both Google and Edgar (Edgar contains filings with the SEC), an interested investor cannot find a prospectus or even a Company website for the largest SIV, Cullinan Finance, associated with HSBC -- which, according to Wikipedia, holds an estimated $27Bn of assets (and note that Wikipedia is the first hit on the results page for a search on Google for Cullinan Finance, at the time of this writing, and only two paragraphs about the SIV are included in Wikipedia). And there is not a separate company website for Gordian Knot, the largest SIV sponsor according to the Wall Street Journal -- at least the author is not able to find one in the first three pages of a search for "Gordian Knot" on Google. The lack of public information is due to SEC regulations, in which the prospectuses and documentation for the SIV's are only made available to registered and qualified investment funds.

The purpose of this article is to shed light on Structured Investment Vehicles -- SIVs -- and assess overall credit risk of the SIV universe, in generally quantifiable terms. Overall, an assessment of the SIV industry shows that the total potential loss from investments in SIVs are most likely not a large risk for the overall economy, even with a moderate to moderately severe meltdown in housing, but there are several caveats to this statement. The lower rated investments of the SIVs and even the higher tranches of the SIVs have a moderate probability of suffering credit losses however.

This article will take information available from the major ratings agencies (S&P, Moody's and Fitch Ratings), the Federal Reserve and the major publications -- the Wall Street Journal and the Financial Times, for example, in order to present the reader with a background on SIV's and bring some light to the issue -- mainly the credit risk exposure, which is most likely the most negative issue for the general economy. Note that the author has worked in both banking and in money management but does not directly have experience buying, forming or selling SIVs, and cannot guarantee the accuracy of all the numbers found in this article. Further the reader is encouraged to do his or her own due diligence with regards to the overall issues of SIVs.

What are SIVs?

SIV's are, according to Fitch Ratings, "Bankruptcy remote" investment vehicles (large pdf warning) which held approximately $400Bn of assets at 9/07. The key idea for SIVs is that credit risk is intended to be almost eliminated, by investing in high grade, mainly AAA and AA rated assets and a smaller percentage of A and BBB assets, diversified across investment instrument and geographic region . (Note, as will be discussed below, that "intended" is different than "has" in terms of elimination of credit risk). Leveraged is utilized. One can say with some confidence that most rated SIVs have invested in mainly "intended" high quality assets, as S&P, Moody's and Fitch have kept a close eye on the (nominal) asset quality of the SIV's, as SIV's must get a rating from at least one of the major rating agencies in order to be attractive to investors. The rating agencies have set an expected asset allocation in terms of credit quality and also geographic diversification for SIV investments: Standard & Poor's regulations concerning acceptable investments for SIV's can be found here:

S&P's Limits on Asset Allocation for SIV's:

Table 3 Example of Single Obligor Limits
Asset rating	Maximum holding of a single asset as a percentage of the portfolio
AAA	15
AA	10
A	5
BBB	3
BB	1

In order to receive a AAA or high rating from S&P on the senior portion of the SIV (the senior portion of the SIV is in comparison to the lower tranches of the SIV, which will be described below), the SIV must comply with the above asset allocation, which will mean a high credit quality -- if the assets the SIV holds are, in reality, mainly "AAA" or high A
(which, according to most definitions, means less than 0.1% default probability).

Further supporting the assertion that SIV have invested in generally low credit risk assets is an update from Fitch Ratings at 9/07, which stated that 58% of the invested assets in its SIV universe were invested in AAA rated assets, 32% in AA rated assets, 9% in A rated assets and 1% in BBB rated assets.

What Could Go Wrong?

The potential trouble is that some SIVs have invested in classes of assets which nominally were rated as "AAA" but were not, in reality, AAA or even close to AAA. These assets therefore face credit risk and the SIVs pose potential losses to investors, and if the losses are significant enough, a potential threat to the overall economy. The main concern for assets classes in terms of credit risk are subprime mortgage securities, as these clearly (to anyone who has picked up a newspaper in the past 5 months) face significant credit risk. But also at issue (but with less concern than subprime MBS) are non-conforming mortgages -- such as jumbo mortgages, which don't have the Fannie Mae or Freddie Mac guarantees but can still be rated AAA to A, and student loans without government guarantees, not high quality corporate debt, and credit card and auto securities also without federal guarantees. These assets have credit quality issues to varying degrees that are not reflected in the nominal AAA to high A ratings by the major credit agencies. Not at risk, in terms of credit quality, are conforming, federally guaranteed Mortgage Backed Securities, sovereign debt (for the most part) and high grade corporate debt (for the most part). With the Federal Guarantee these should be credit worthy in everything but a complete and total financial meltdown (worse than the Great Depression). The riskiness of the investment in the SIV is most directly related to the degree in which the SIV has invested in assets which do have credit issues.

Have SIV's Invested in Subprime Mortgage Backed Securities?

It follows from the above discussion that the most pressing question is: how many SIV's have invested in subprime Mortgage Backed Securities? We can say with some (but not 100%) certainty that most SIV's do not have significant subprime exposure as Fitch Ratings announced at 9/2007 that only 0.9% of all the SIV assets in Fitch's SIV universe were subprime RMBS
Fitch's assessment should cover a substantial portion of the (see: http://www.fitchratings.com/web_content/presentations/2007/sivs/fitch_siv_update_outlook_sep07.pdfSIV university as SIVs are incentivized to receive ratings from Fitch, in order to be sold to investors. We can double check this information from Fitch, as Citigroup has announced the exposure of its sponsored SIVs to subprime assets. According to Citi, none of Citigroup's sponsored SIVs hold direct exposure to subprime debt, but have $70M of "indirect" exposure (out of a total of over $100Bn of SIV assets, so only 0.07% of the total asset level). Note that Citigroup SIV's comprise approximately 25% of the total value of all SIV's outstanding. Of concern are the few (so far) reports of SIV's which held high concentrations of Subprime MBS which may not be captured by Fitch or Citigroup -- such as the report concerned the SIV "Mainsail II" in August of 07, a SIV with an approximate $4.5Bn in terms of assets, which appears to be issued by the New Jersey based Hedge Fund Solent Capital (which was underwritten by Barclay's Capital but Barclay does not appear to have credit risk exposure to Solent): In the case of the Mainsail SIV, the investors in this fund are likely to face significant losses.

Is the Low Average Net Asset Value of the Fitch SIV Universe a Threat to the Financial System?

To answer the above question, in short, the relatively low NAV does not likely represent a threat to the financial system. This will likely be contrary to what most readers have read from the press in recent months so I will step through the reasoning as follows (and note again all readers are encouraged to follow links and do their own due diligence with regards to this issue).

The Net Asset Value calculation for SIVs has caused significant confusion so an explanation is needed. A (one can say) notorious chart is presented in Fitch Ratings latest presentation on SIVs, which shows that SIV Net Asset Value (NAV) to in the Fitch Universe have an "NAV Percentage" of approximately 70% at 10.07, compared to 105% at July of 2007. See page 12: here: (large pdf warning) (the chart to the left represents the decline in NAV value of the SIVs) This is an alarming drop in 3 months, and further, the definition of NAV% is not fully defined in the presentation so has led to some dire predictions of credit losses. As an example, a few bloggers (that the author has seen) have concluded that an NAV of 70% means that on average AAA rated SIVs are underwater by 30% of the total $400Bn outstanding, so face on total a potential credit loss of $120Bn. However, this is not accurate, as the NAV % is defined as the excess of total assets over the liabilities. Generally the SIVs are leveraged at a high rate -- at 10x to 14x to 1. So the total assets are actually down (if one uses 12x leverage) 2.5%, so total losses are expected at a 30% NAV of $10Bn. This last statement is tricky, but absolutely essential, so here's a review, at an NAV of 70% for the SIV universe refers to approximately 1/12th of the total SIV assets at 12x leverage, or a potential loss of 30% * 400/12 = $10Bn. $10Bn is not a significant shock to the financial system -- however it is not pleasant either, of course. But also, of course, the NAV could continue to drop, depending on the problems in the respective asset classes in which the SIV has invested. If the NAV drops to 0% -- which is possible if confidence collapses across most high grade asset classes, but not extremely likely as in crises investors tend to flee to quality, so at least some of the AAA and AA assets that the SIV holds should appreciate in value -- then the credit loss to the financial system would be 8.3% of $400Bn or $33.2Bn. These numbers are not extremely disturbing to the financial system as a whole -- but of course not good, but should be noted that the losses are disturbing to the investors in the SIV.

Other Risks Stemming From SIVs Not Directly Related to Credit Risk:

This conclusion that SIV's are not a probable threat to the overall financial system as argued above holds several caveats. First, if the SIV are all forced to liquidate at the same time, this could -- and probably would -- drive down the price of the assets they hold and cause more severe losses for investors and the financial system (one can imagine in this scenario that everyone running for the exit in a movie theater at once during a fire, clogging the exit). The extent of these losses would be determined at lower amount than the full $400Bn as described in the NAV section above, but still could become substantial.

Second, there could be accounting fraud at the SIVs -- so investors could have been mislead as to the amount of, for example, subprime or other risky assets that the SIVs hold. This would increase losses -- but it is difficult to determine the probability that SIVs have had these issues. The probability of this threat is determined by the confidence one places in the competence of the auditing firms and management teams (overall the author thinks this risk is relatively minor).

Third, if there is a overall general meltdown of the economy, then SIVs will also be affected as will all other asset classes.

Fourth, there could be risks stemming from other asset classes that also impact the overall financial system in addition to (a very pessimistic case) $50Bn or so loss from the SIV asset class. Note that this article only focused on SIVs and not subprime MBS as a whole, and not Credit Derivatives or Asset Back Securities as a whole, which are separate classes of assets and have their own financial risks and issues. A spread of investor fear across all collateralized asset backed classes has not been assessed in this article -- likely, the risk here would be greater if all these asset classes posed significant credit risk separately, as to generate fear that the banking system could not handle the overall, eventual losses. But as of this point in time the author is not aware of separate assessments for credit risks in other areas of the Asset Backed Securities markets.

Conclusions:

Concerning the SIV universe as a separate asset class, credit risk appears to be relatively well contained from the subprime fallout, based on documents from the Rating Agencies and Major Banks as discussed above. Subprime exposure at across most of the known SIV universe appears low according to Fitch Ratings, Standard & Poors and Citigroup. Further, losses should be calculated based on a correct understanding of the Net Asset Value of SIVs, which is significantly lower in almost all circumstances from a triple digit, billion dollar loss, which has been thrown around on some blogs. But this statement -- that SIV credit risk is contained -- holds several caveats, as was noted above, namely liquidity risk, general risks stemming from an overall financial meltdown and risks stemming from other Asset Backed Securities classes.