Wharton Magazine

Jump to

Final Exam

In real estate, it’s all about location, location, location. No matter where you live, though, you’re welcome to tackle our latest Final Exam challenge.

Think you could still ace your way through Wharton? Well, here’s your chance to prove it.

In each issue of Wharton Magazine, we’ll test your knowledge with a question taken straight from an actual Wharton course exam. Submit the correct answer and you might just walk off with a great prize—a Wharton Executive Education program.

This issue’s Final Exam question comes from Todd Sinai, Associate Professor of Real Estate and Business and Public Policy. Professor Sinai sent along a question about everyone’s favorite topic: Mortgage-backed securities. Good luck.

The Basics

Commercial Mortgage-Backed Securities (CMBS) are created by pooling commercial loans varying in size, property type and location, and then transferring those pools to a trust. The trust then slices the pool of mortgages into tranches, which are underwritten and issued as a series of bonds. (Note: A tranche is a single issue of a security released at different times. The originator is the entity that made the original loans.)

Sales of CMBS averaged approximately $75 billion a year from 2000 to 2004. Sales were $179 billion in 2005, $221 billion in 2006 and $248 billion in 2007. But according to Bloomberg.com, sales of CMBS dropped to $11.15 billion in 2008.

The Question:

Is the following statement true, false or uncertain? Bond purchasers should be willing to pay more for a CMBS tranche if the originator of the underlying mortgages buys all of the most subordinate tranche.

The Answer:

Professor Sinai offers the long-awaited answer here:

The statement is true. 

But before I get to the explanation, let me take care of a little exam-writing housekeeping. “Bond purchasers should be willing to pay more” is shorthand for “are these CMBS tranches more risky?” Why? Well, bond purchasers will pay a higher price–and thus accept a lower yield–for two possible reasons. One is that the coupon is higher than the alternative. The second is that there is less risk. The coupon is the same whether or not the originator buys the most subordinate tranche since the underlying bonds are the same, only the purchaser is different. Thus, in order for the statement to be true, there must be less risk if the originator buys the most subordinate tranche.

The risk that bond purchasers should fear, and which was highlighted in the recent collapse of the CDO markets, is that entities that make loans but who sell off all the risk–who do not have any “skin in the game”–tend to make riskier loans. In this case, those entities are the loan originators, who issue mortgages, sell those mortgages into a CMBS, and do not hold any of the resulting CMBS bonds. The worry should be that loan originators may be originating loans that they know are more likely to underperform but which meet any stated underwriting standards and aren’t easily observably sub-par.

“But,” you might ask, “aren’t there institutions that maintain the quality of the collateral in a CMBS?” There are potentially three: The senior/subordinated structure, ratings agencies and the end investor in CMBS bonds. First, the end investor typically does not know much about the underlying real estate. That’s the whole point of CMBS: To enable non-real estate specialists to invest in real estate debt by making them more uniform, rated and easy to trade. That lack of real estate knowledge, plus the inherent practicalities of how loans are made, means that the originators of the loans know a lot more about the quality (or riskiness) of those loans than the end investors, or even the entity that structures the CMBS.

The senior/subordinated structure refers to the way that the cash flows from mortgages are allocated to bond tranches. In a senior/subordinated structure, one class of bondholders holds the residual claim on the cash flows from the pool of mortgages. They are the last to be paid. If those bondholders–who tend to be specialists in real estate and have a fair degree of influence over what loans are allowed in the pool–are willing to buy the most junior tranche, the logic goes, then other investors should feel comfortable buying the more senior tranches, knowing the junior bondholders will get wiped out before they themselves take any loss. By taking the first-loss position, the junior bondholders ostensibly protect the more senior tranches, plus they have the incentive and wherewithal to properly research the underlying mortgages. However, that third party still may not have as much information about the loans as the originator. And, if they have a higher tolerance for risk than the investors in the more senior tranches, they might not try hard enough to seek such information. Plus, if they sell the junior bond tranche–often it was into a CDO–then they don’t hold the risk at all.

The ratings agencies are supposed to be another line of defense. They are supposed to, for each bond tranche, rate the risk that the principal in that tranche will not be repaid. However, recent research demonstrates that ratings standards deteriorated significantly. As the observable quality of the collateral declined, ratings did not. And the amount of subordination–the size of the first-loss piece–was reduced, lowering the protection for the more senior tranches. And, if ratings agencies were not able to ascertain the true quality of the underlying real estate mortgages, this problem would be compounded.

A solution to this problem of asymmetric information is for the originator to own the most subordinate tranche. This carries the senior/subordinated structure to the logical extreme–the entity with the most information about the true quality of the underlying real estate should be in the first-loss position. This reduces, if not eliminates, the incentive for originators to issue riskier loans and pretend they are not, since they are on the hook for the risk. Thus CMBS investors should feel more confident in the bonds and be willing to pay more for a given tranche since it should have less risk. 

 

Share |