Seeking Alpha
Special situations, bonds, portfolio strategy
Profile| Send Message|
( followers)  

It’s always interesting to analyze forecasts about future spreads and yields. The CMBS industry’s weekly newsletter, Commercial Mortgage Alert publishes predictions on future CMBS spreads every January & July. I have always liked to look at them, not because anyone has the ability to predict the future, but because the predictions often provide interesting insights into what the market participants are thinking.

The table below is the latest set of forecasts published. It shows predictions from ten industry participants on where recent vintage (i.e. 2006 to 2008 deals) CMBS spreads will be six months later. Some observations follow.

CMBS Spread Forecasts for Dec 2009 - table from Commercial Mortgage Alert included in an article by Malay Bansal

Let’s start with the averages. The average prediction suggests that, six months later, Super Senior AAAs will be tighter but BBB spreads will be wider. The credit curve will be steeper. That makes sense and reflects the belief of many that CMBS will face losses but not to an extent that it will impact super-senior AAA tranches in most deals. The range of 350 to 975 is wide, but not that much given the current uncertainty.

The BBB forecasts are more interesting – with a really wide range from 3500 to 17500. The majority expects spreads to widen, but the widest spread is mine. Why am I negative on recent vintage BBB and BBB- bonds? My reasons are not different from others, and so are not that interesting (and hence listed as a footnote at the end). What is interesting to me is a question I have been asked multiple times – if I am twice as negative as the average and why? With average prediction of 8473, and my prediction of 17500, it is a fair question. So, let’s examine that in a little more detail.

The current spread for BBB is listed as 6500 bps over swaps. If we take a single bond as a representative of the sector, that spread implies a yield of about 68.5%, a price of about 10.40, and Modified Duration of 1.6. The bond has a coupon of 6%. If you buy the bond at a price of 10.40, at the end of second year, you would have received 12.00 in coupon payments (2 years at 6%/year), which is all of original price you paid for the bond plus about 15%. In effect, the market is saying that it expects these bonds to receive just about two years of coupon, and not much else after that. In other words, the market price implies that it expects these bonds to be wiped out after about two years.

If you logically follow that belief, then after 6 months, an owner of the bond would have received 3.00 and should be expecting one and a half years more of cashflows. So the price of the bond would be roughly 3.00 lower than 10.40, or 7.40. Indeed, that price is not too far from what the average spread prediction would result in. In other words, the average spread prediction of 8475 implies that the market will be expecting about 1.5 years of cash flows. In comparison, my prediction of 17500 assumed expectation of only 1.0 year of cashflows. So, I am not twice as negative as the average, just a little more - the difference being 6 months in cashflows for seven year bonds. So, why such a big difference in spreads? That has to do with the current low prices of these bonds, and the fact that as bond prices decrease, duration decreases too. For a bond priced at 10.00 with a duration of 1.0, a 100 basis point change in yield results in a change in price of 1%, or just 0.10 (1% of 10.00). As duration decreases, it takes a larger change in spreads to make a difference in price. That bond math explains the big difference in spreads, but that is not the important point. The important observation here is that, whether expectation is for 12 or 18 (or something in between) months of remaining cash flows, it reflects a pretty negative expectation for bonds that will have about 7 years remaining to maturity.

Despite the negative expectations, there is another point that is important to keep in mind – it is a simple one but is often missed by many, especially in media reports – CMBS and property markets are linked, but are separate markets, and one can be way ahead of the other. It is entirely possible for commercial real estate to face problems at property level, but CMBS BBBs to tighten. That is another way of saying that, despite all the issues mentioned, it is entirely possible that the majority may turn out to be wrong. Note that there were two predictions for tighter spreads on BBB. There are scenarios possible under which losses in CMBS will be limited even with problems at property level, and that will logically lead to a significant rally in the BBB tranches. A lot will depend on government and industry steps, some of which are helping tighten spreads, but I do not yet see steps that make the more optimistic scenarios likely. Apparently, neither does the majority.

Here is perhaps the most important point of this article: if you are not sure and have doubts about extent of losses in BBB and BBB- classes, then AA and single-A classes at prices in 20s with subordinations of 9.8% and 7.5% (and same about 6% coupon), have to be a bargain, especially if you can do credit work and pick the right deals. Your views will have important implications for your views on large holders of CMBS - insurance companies, banks, REITS, and others including closed and open-end funds. As always, it’s not just the sector, but exactly what from that sector they own that matters.

Note: The views in this article and my spread predictions in the Commercial Mortgage Alert article referenced are solely my own.

___________________________________

Footnote:

Some Reasons for being negative on recent vintage BBB & BBB- CMBS

Consider the following facts and opinions:

  • Commercial property values are down 34.8% from Oct 2007 peak, as measured by the Moody’s/REAL CPPI index.
  • Average subordinations for 2007 BBB and BBB- were 4.20% and 3.10% respectively. That means losses of approximately 4.20% and 5.3% will result in complete write-down of these tranches.
  • Appraisal reductions can lead to interest shortfalls (via ASERs) and can stop cash flows to the subordinate bonds long before actual write-downs.
  • It remains very difficult to refinance maturing loans, especially those with larger sizes. 29% of all loans that matured this year remain outstanding. Of the 5 year loans that matured this year, only 50% were refinanced. Number of CMBS loans maturing increases dramatically from $18 Bn this year to 35 Bn next year, and 56 Bn two years later.
  • Loans going in special servicing generally require new appraisal and can result in appraisal reductions before actual losses. According to Fitch, loans in special servicing are expected to increase from current $50 Bn to 100 Bn (or about 14% of total outstanding universe on average) by year end. In addition to losses, the special servicing fee of 0.25% on loans in special servicing will reduce cash flow to bonds increasing interest shortfalls.
  • Among just the large loans, about 15 Bn with pro-forma underwriting with expectation of higher cash flows have not realized the expected increased cash flows. They are paying debt service at present, but may default in next 6 to 12 months once the interest reserves are completely used up.
  • Over next few years, partial IO loans will start to amortize after the initial 2 to 5 year interest-only period. Increased debt service may not be fully covered by property cash flow, leading to defaults. In 2007 vintage, about 32% loans were partial IO along with 53% that were interest only to maturity.
  • Delinquencies are still very low, even for loans with cash flow not covering debt service. In 2007 vintage, roughly 14.7% loans have debt service coverage level of less than 1.0.
  • Overall delinquencies are expected to increase. For example, JP research expects aggregate delinquency to reach 4% by year end, and 10% by end of 2010. DB research expects delinquencies to reach 6-7% by end of 2009 in aggregate, and high-teens to 20% for 2007 vintage. REIS expects delinquencies to possibly reach 7% by year-end. Last time delinquencies were higher than 6% was in 1991 after the S&L crisis.
  • Significant losses are expected in the recent vintage deals, even before maturity of loans. For example, DB research expects term losses, i.e. losses before maturity, to be 4.3 to 6.3% for the entire CMBS universe (with total cumulative losses of 10% including those at maturity), and 8.4% to 12.1% for the 2007 vintage. Fitch expects cumulative losses on 2006 to 2008 vintage deals of approximately 8% (maybe more than 14% for some 2007 deals), with 25% of loans possibly defaulting before maturity.

Source: Examining the CMBS Spread Forecasts