By Darrell Wheeler, Vivek Tiwari and Joe Yu, Amherst Securities

Vivek Tiwari

Darrell Wheeler

In the past year, ownership of several commercial mortgage-backed securities special servicers has changed hands to opportunistic investors or funds. We consider whether this ownership change has altered CMBS servicer behavior in this article and look at the post-2009 troubled loan experience with a focus on the resulting appraisals, note sales, modifications or liquidations. The article considers the servicer’s economic position within the CMBS transaction at the time the action was taken and we do find that so far special servicers have been inclined to liquidate the loans once the BB+ balance was written down to less than 25% of the original balance (when the controlling right is transferred).  We would caution that this result is based on a limited sample of 52 loan resolutions, but it does suggest that CMBS investors could anticipate an increase in liquidations as subordinate classes get extinguished. The article also compares liquidation and modification rates among the major CMBS special servicers and finds some interesting nuances about their different behaviors.  

Originally, these special servicers were staffed with workout specialists that had clear mandates to service the commercial mortgages and maximize recoveries for the CMBS trust, and this behavior was frequently reinforced by the special servicer’s ownership of the transaction’s noninvestment grade classes. However as the recession progressed, the capital structure of these CMBS transaction has experienced collateral write downs that has started to eliminate the servicer’s interest (“B” Unrated Tranche in Exhibit 1). Post recession many of the special servicers have been sold to new owners that may have more aggressive real estate-like return targets or alternative investment motivations beyond simply servicing the pool. In Exhibit 1, we show some of the major ownership transitions that have taken place over the past few years.

The table lists several major ownership changes that have transitioned many of these servicing operations from pension fund or insurance company to more opportunistic less institutional buyers. With these recent changes investor have become concerned that ownership of these servicers has changed hands over to a more real estate related firms that may potentially take advantage of the fair value option special servicers retain on asset dispositions. This option enables special servicers and controlling class holders (which in turn could be special servicers themselves) to buy a defaulted but not foreclosed loan at fair value. Generally, the trustees have to verify the purchase price established by the buyer but the fair value determination can be based on a recent appraisal. This leaves investors concerned that the new owners of special servicers may be more inclined to liquidate assets at undervalued price levels rather than fully marketing the assets.

Several investors have also highlighted that some special servicers now seemed to have ownership relationships with various commercial real estate brokers. This fee business was likely considered to be a natural progression as liquidation portfolios swelled during the recession and the servicer saw how much money they were potentially leaving on the table. These real estate brokerage fees can typically run from .5% up to as much as 6% when the buyer is not represented and could also potentially help the servicer to direct which investor gets to buy a loan. In fact, C-III Capital Partners has acquired NAI Global (which through its affiliates completes over $45 billion in real estate transactions a year). While Island Capital directly owns ICG Realty LLC which is another commercial real estate broker and in October announced a strategic relationship/investment with Grubb & Ellis, a commercial real estate broker. Exhibit 1 which lists the consolidation that has happened within special servicing industry also provides the related commercial real estate brokerage connections that we currently know. The potential for fees to drive a propensity to liquidation is a servicer conflict that should require servicer disclosure of related real estate commissions earned on loan dispositions.

Exhibit 1. Special Servicer Ownership Changes 2009 to To Date
Source: Amherst Securities Group LP, Commercial Mortgage Alert

Investors are particularly concerned over mid sized loans( between $10 to 50 million), which are not in the public spotlight. October’s remittance report for BACM 2002-2 provides a good cautionary example in two loans that were quickly liquidated at very high severities. The loans in question are 20555 and 20255 Victor Parkway in Livonia, MI, which were a $24.17 million and $22.59 million office loans fully occupied by Quicken Loans which was vacating in August, 2011. The defaulted loans were transferred to C-III, the special servicer, in March and quickly had a new appraisal in June for $3.9 million on the 20555 Victor property and $4.5 million for the 20255 Victor property. At that time the loan caught our attention (as we reunderwrote the loan in our CMBS ALIAS system), but based on our ALIAS comparable sales data we could not really determine that the large potential appraisal reduction was unjustified for an empty property. The required appraisal reduction was implemented in August when Quicken vacated and stopped paying rent, and now in this month’s remittance we see that 20555 and 20255 have suffered 85.5% and 89.3% losses, respectively from a discounted pay off from the borrower. The speed and severity of this resolution catches our attention as a servicer typically allows some time for reletting a property and improving income before disposing of an empty property. The servicer would typically weigh the properties’ carry cost versus a quick dark value disposition with their consideration that a tenant is unlikely to be found. But in this case the evaluation of leasing options seems to have been cut short by the deeply discounted October payoff by the owner, who is listed in the prospectus as Michael Kojaian, who is chairman of Grubb & Ellis a commercial real estate broker. On October 17th, C-III announced a strategic investment in Grubb & Ellis, so there appears to be a clear relationship1 between the borrower and the servicer, which hopefully did not play any role in the loan workout.

Then on October 19th, just five days after the discounted payoff, an article in Crain’s Detroit Business2 announces that Kojaian Management Co. has fully leased both buildings to Trinity Health under a ten year lease. Obviously, with a new quality long term tenant signed up within days of the discounted payoff the building is likely significantly worth more than the discounted payoff value of the loans. We would hope that special servicer had no idea that potential tenant was in the wings when they went forward with the discounted payoff and this would not be the first time servicer has been hoodwinked by borrower looking for a modification or discounted payoff. But the chain of events is suspect, a relationship between borrower and servicer appears to exist, and servicer did not appear to allow significant time to evaluate leasing options, so we expect that investors in this transaction will be contacting the trustee and master servicer to investigate the special servicer’s decisions.

Given the events on the BACM transaction there are obviously valid investor concerns over any increased propensity to liquidate, especially if it affects severities and bond repayments. Also, opaque details on loan liquidation process further compound investors’ problem understanding servicer behavior. Analyzing these potential behaviors is difficult as servicer actions are being influenced by market refinancing conditions, property investment yields and value and expectations for the economy which have clearly been volatile over the past 36 months. While many times there is just limited information in order to determine what factors drove a particular servicing decision. As strategists we do try to predict future outcomes, and the CMBS ALIAS system has been established to box in potential liquidation/modification outcomes for troubled CMBS loans and the increased information that investors continue to request would be helpful to this process. But at this point we have the data servicers currently report which we will use in the remainder of this report to see if we can explain servicer actions with some level of greater precision.


Exhibit 5. Loan Resolutions Sorted by BB+Status
Source: Amherst Securities Group LP, Intex Data Solutions

CMBS Loan Resolution Analysis

To consider if there is a detectable shift in servicer behavior we decided to look at servicer behavior while original B-piece first loss buyer may still have an economic interest and then as the BB+ class is eliminated or starts to take shortfall. To do this in Exhibit 2 we present resolutions results sorted by the status of the BB+ class.

Looking at the first table we see that “pre-write off state” servicer behavior seems to be best predicted by loan size: larger loans are more likely to be modified and smaller loans are more likely to be liquidated3. The average size of modified loans is $38 million while the average for liquidated loans is $9 million. We have suggested in the past that this trend indicates special servicer is modifying large loans, liquidating small ones and sitting on the fence about middle ones. This is why we designed our CMBS ALIAS system to underwrite specific potential loan modifications with additional third party data for all troubled loans over $10 million and consider potential modification outcomes for each.

However, looking at the two lower tables of loan resolutions where the BB+ class gets reduced or decreases below 25% of the its original balance we do see a disturbing propensity to liquidate loans and result in higher loan loss severities. In these cases the special servicers’ behavior does appear to be predictable, as they liquidated 17 loans and modified only 1 loan. This has us thinking that liquidations may pick up in later years as the BB+ class is eliminated, yet many of these liquidations are from some of the worst deals of 2006 and 2007 and appear to be smaller loans that were aggressively underwritten. So, further defaults in those deals are likely still being evaluated fairly and it may be too early for us to consider adjusting our current parameters for extensions. As we gather more resolution and workout data from recent vintage issuance and actually have time to see some modifications we will have to continue to evaluate the liquidation versus modification behavior.

But with this liquidation behavior we also wanted to see if there were any trends that could be detected by specific special servicer so we repeated the loan resolution table in Exhibit 5 below.

Initially, the numbers suggest behaviors are similar among all special servicers: larger loans get modified and smaller ones are liquidated. However, there are also some significant difference that Clarion/Torchlight serviced deals have undertaken the most modifications at 66% while CW Capital, LNR, and Midland have only modified 45%, 35% and 47% of their troubled loans. LNR appears to be the quickest to liquidate loan with a 25% liquidation rate, but that figure is not significantly different than other servicers and the speed of liquidation may actually be proven to minimize loan loss severities. This is the classic fast liquidation versus patient servicer debate and we suspect that the final optimal outcome may likely still have more to do with the economic recovery than the servicer’s specific approach.

Conclusions – Still Need More Data on Special Servicer Behavior

Overall, with only a 1,592 loan resolution sample we have to be cautious about our initial conclusions as the recent 24 months of experience could be very different than the experience in an unpredictable economy. While several recent vintage deals demonstrate a high liquidation rate once they eliminate the BB+ class most of the deals have only reached that write down level for a few months and not really experienced sufficient time to demonstrate loan outcomes other than liquidations. In many of those quick liquidation instances the servicer may have simply decided to liquidate what looked like hopeless situations on the few aggressively underwritten loans that were exposed by the recent recession. Yet, in the BACM 2002-2 transaction we did find one sample that suggests the servicer may have acted too quickly to take a loss, so we will have to continue to track data. If we see several more examples then we have to state that conflicts of interest in servicing are hurting investors and actions should be taken. But currently, we still have limited overall data set and would say that we still need additional data before making major changes to our expected loan resolution time frames.

In anticipation of changes in future activity we have now established this special servicer monitoring table format to follow their actions over the next twelve months and will evolve this approach to look at achieved severities versus Amherst anticipated severity on a servicer specific basis. So, while the initial results suggest the potential beginning of some suspect trends we expect to develop much more concrete conclusions in an additional twelve or twenty four months.