DBRS identified a sample of approximately 3,700 CMBS properties that were collateral for repeat loans (i.e. loans that have been securitized in multiple CMBS transactions over the past 20 years). The data available from these repeat loans can provide representative snapshots in time of the health of the CMBS market and are helpful in assessing the market’s long term sustainability. Specifically, we will use them to create apples-to-apples comparison of leverage and other financial metrics for CMBS loans through time. This commentary explores repeat loan trends throughout the history of CMBS. Broadly speaking, the results show that since the near collapse of the CMBS industry in the late 2000s, repeat loans have signaled some relative stability in the CMBS market – but the industry has yet to achieve a long-term stable footing.
Loans representing an original balance $42.0 billion have been re-issued into new CMBS transactions. Through refinancing or other purposes, repeat loans have found their way back into the CMBS universe with a combined balance of $57.0 billion (a 35% increase in leverage). This increased leverage is illustrated by the measure of whether a repeated loan is cash-in vs. cash-out . During the bubble years , observances of cash-in loans declined substantially, underscoring the increased risk in legacy CMBS transactions. This trend has been reversing itself since late 2007; but instances of loans that are cash-in still remain less than half of all repeat loans.
Cash-outs can be normal and healthy when accompanied by proportional cash flow growth. But when either loan balance growth is greater than value growth (LTV increase), or value growth is greater than cash flow growth (cap rate compression), DBRS recognizes a high likelihood that, all else equal, the repeat loan’s ability to perform and refinance in good standing is reduced relative to its prior term. The observed 35% increased leverage of repeat loans is partially justified by an offsetting 15% cash flow growth, but the balance was created by changes in underwriting standards (i.e. LTV increases) and cap rate compression (i.e. inflating property values).
Repeat Loan Examples
The experience of repeat loans after and including CMBS 2.0 is so far relatively strong, but there are points of concern as illustrated by three case studies below.
Case #1. Serrano Highlands
The Serrano Highlands loan is secured by a large multifamily property that has now been securitized three times in CMBS pools: first in 1998, then in 2004 and finally in 2007.
The loan experienced strong cash flow growth of 7% yoy from 1998 to 2004 (the first repeat), and then 5% yoy from 2004 to 2007 (the second repeat). But loan growth yoy was nearly 26% between 2004 and 2007 (the second repeat), compared with a 3% yoy change at the first repeat. At the same time DSCR decreased just 2% yoy in the second repeat despite the large increase in leverage. The massive loan increase of 100% in absolute terms between 2004 and 2007 was funded by a dramatic change in the loan’s debt constant. During the first repeat, the interest rate of the loan dropped substantially but the debt constant of the loan remained high (even increased), as the loan’s amortization term became significantly more conservative. During the second repeat the interest rate of the loan remained relative low (although it increased), but the debt constant dramatically reduced as amortization disappeared from the loan — the impact of which is best observed by the dramatic changes in this loan’s exit debt yield. This loan had significant cash outs at both of its repeats.
Case 2. Bay View Community MHC
The Bay View Community MHC loan is secured by a large mobile home community that was twice repeated in CMBS, originating in 2003, then reappearing in 2007 and finally 2013.
Unlike the first example, Serrano Highlands, the Bay View Community MHC loan did not experience dramatic changes in its cost of capital. However it did experience large increases in leverage that grew at a 12% yoy rate at the first repeat, and a 4% yoy rate at the second repeat. The large increase in leverage was accompanied both times by a decrease in the loan cap rate. Despite reported LTV declines the asset is still operating with a debt yield that is well below the original securitization in 2003, marking a significant increase in credit risk. This loan had a cash-out at both its repeats.
Case 3. 369 Lexington Avenue
The 369 Lexington Avenue loan is secured by a Manhattan office building which was first securitized in 2004, and then repeated twice in 2006 and 2013.
The loan experienced a large 21% yoy increase in its loan amount at the first repeat in 2006. Unlike the two previously described loans the change was driven not by debt constant or cap rate declines but by cash flow increases of 28% yoy in the same period. LTV in the first repeat thus remained the same as did DSCR. In February of 2012 the loan was transferred to the special servicer for maturity default. It was ultimately resolved without a loss and re-securitized in 2013. The 3rd origination of the loan in 2013 has roughly the same valuation but a lower leverage point and going in LTV of 56%. In terms of both going in and exit debt yields the loan is on firmer footing than it has been in its previous two terms. This loan had a cash-out at its first repeat but then a cash-in at its second (as might be expected for repeat loans that are distressed in their prior term).
CMBS Repeat Loans Cash-In/Out
In the next graph we summarize the trend towards cash-in or out loans. Looking at the green bars we see that just 18% of repeat loans in CMBS history involved a cash-in, a trend that changes through time approaching just 5% for the vintages that were marked by the highest leverage increases. Post-crisis CMBS repeat loans have trended towards having more instances of cash-in, but these instances remains below 50% of all repeat loans and have begun to trend downwards again more recently (see the fitted line that summarizes average percent of cash in).
Our measure of cash-out is, in and of itself, not always credit negative. DBRS expects a measure of cash-out when accompanied by strengthening collateral fundamentals such as cash flow growth through rent increases, growth in the NRA of the property, a change of hotel flag or a change in tenant profile.
In newer CMBS vintages this increase in loan amount has more often than not been accompanied by increased value growth – but not in equal part by increased property cash flow growth. As such, the CMBS industry has not achieved a sustainable basis for improved bond performance, but rather has seen a compressing cap rate environment facilitating unsupported leverage increases. This is demonstrated by the black bars in the graph. Since 2009 this trend, while still evident, has not yet become as pronounced as it was during the pre-crisis years (2005 to 2007).
CMBS Repeat Loan Characteristics
The chart below segregates the vintages of CMBS into four buckets that each have defining characteristics of repeat loans in terms of NCF growth, interest rate and cap rate migration, value growth, and leverage increase.
The pre-2003 vintages displayed robust cash grow growth for repeat loans. They also displayed significant downward interest rate migration, and accompanying moderate downward cap rate migration. This created an environment where property values grew faster than even the robust cash flow growth experienced. Leverage increased even faster than value, which significantly outpaced cash flow growth, making CMBS loans less able to handle a crisis or downturn in performance.
The 2003-2008 vintages (bubble vintages), displayed less robust cash flow growth than the pre-2003 vintage (albeit still positive and strong). Like the pre-2003 vintages, they continued to experience significant downward interest rate migration. Unlike the pre-2003 vintages, the downward cap rate migration outpaced the interest rate decline which contributed to value increases that were almost double their cash flow growth. Like the pre-2003 vintages, leverage outpaced value increase.
The 2009-2012 vintages (post bubble vintages), displayed suppressed cash flow growth. While downward interest rate migration was still prevalent, it slowed, and downward cap rate migration was less than the previous CMBS vintages (albeit still a downward migration). Value increases outpaced cash flow, but leverage increases for the first time in CMBS history did not outpace value growth. While this trend is somewhat encouraging, the fact that leverage increases still outpaced cash flow growth is a credit negative.
The 2013-2014 vintages, were similar to the 2009-2012 vintages, but the magnitude of all the migrations increased. Interest rates trended downwards at a quicker pace, as did cap rates. Value growth still significantly outpaced cash flow growth, but it also outpaced leverage increases. Leverage increases were still higher than cash flow growth. The average life of the loans that contributed to the 2013-2014 repeat loan bucket was 9.52 years. They were from vintages that faced a much higher interest rate environment. As such, upon repeat, their leverage and values were pressured to outpace cash flow growth.
Despite the fact that cap rate compression remains persistent in post crisis CMBS, repeat loans have leverage that are a far cry from pre-crisis trends. The chart below displays the magnitude of difference between current CMBS and post crisis CMBS.
It is rare in the history of CMBS to find a quarter when cash flow growth of repeat loans outpaced leverage. And yet, it is more common in post crisis CMBS to find quarters when value growth out-paced leverage. This is encouraging but not altogether sustainable. It is also not unexpected given the persistently downward trending interest rates, throughout CMBS history.
Still 2013 and 2014 are significantly less aggressive than the 2006-2007 vintages. On top of the sheer magnitude of leverage increase relative to cash flow growth, we can also see this difference reflected in the repeat special serviced loans. The chart below demonstrates the same repeat loan statistics of cash-in vs. cash-out relative to repeat loans that were under performing assets before being re-issued in CMBS. Under such circumstances we should except that the proportion of cash-in to exceed cash-out instances, a scenario which is mostly true today but not so for the top of the market vintages. In the chart below the number of repeat special serviced loans increases dramatically in 2012, 2013 and 2014; most market participants realize that the large increase is caused by the lagged effect of recession, as these loans become distressed due to recession, transfer to special servicing, and experience workouts before being re-issued in new CMBS transactions.
The chart below further demonstrates that prior to 2008, repeat loans that were once specially serviced rarely had cash flow growth that exceeded increases in leverage; whereas after 2008 they typically had cash flow growth (or declines) that were more favorable than their leverage increases (or declines).
Perhaps the starkest example comes from summarizing the vintages by our previously assigned categories. In the chart below, the 2003-2008 vintages had repeat loans that were once specially serviced and on average recorded a leverage increase (red bar). So far in the post bubble vintages this has not been the predominant experience.
The final assessment of CMBS new issuance lending as seen through the lens of repeat loans remains in murky territory. In order for it to be seen by DBRS as unequivocally credit positive on CMBS bond performance the cash flow growth of properties would need to exceed leverage increases – but it still does not and never has. What is positive is that value increases in post-crisis CMBS tend to exceed leverage increases for the first time in sustainable form in CMBS history. This is however not in small part due to interest rate declines and compressing cap rates.
Additionally, the fact that new issuance CMBS have repeat loans whose previous loan terms averaged 9.75 years, means that these loans are from the earlier vintages of CMBS where credit underwriting was tighter, and debt constants and the interest rate environment were higher. It is perhaps not fair to suggest that cash flow growth should exceed leverage growth for the time being – because most of the repeat loans were originated in starkly different interest rate environments. As the latter vintages of CMBS mature in 2016 and 2017 cash flow growth will need to exceed leverage increases for the industry to gain solid long term equilibrium. As it stands now the CMBS industry has not achieved a sustainable basis for improved bond performance, but has benefited from a compressing cap rate environment which is facilitating leverage increases that will be difficult to support in a higher rate environment.
 A “repeat loan” is counted as its collateral property leaves one CMBS transaction and enters another (in light of new issuance). DBRS excludes pari passu loans or loans that exist in multiple CMBS transactions at the same time. There have been approximately 3,700 instances of repeat loans in CMBS.
 When the issuance loan amount of a repeat loan increases over its previous issuance loan amount we call this a “cash-out”, to reflect the net cash realized by the owner(s) of the collateral property; likewise if a loan amount decreases we call it cash-in. Cash-in, reflects strong and sustainable lending practice especially when accompanied by cash flow growth, and especially in absence of past delinquency in the loans’ original life. It strongly indicates that the loan is capable of performing at its leverage point.
 Roughly speaking the 2004-2007 vintages.
 Defined as loans that had originally transferred to the special servicer and were subsequently originated in new CMBS issuance.