CMBS lows offer high-quality lessons for future

CMBS defaults study by BAML highlights criteria to make deals safer, reports Lauren Parr

Bank of America Merrill Lynch analysts have recommended that regulators look to lessons learned from defaulted CMBS loans to create a set of criteria for future high-quality real estate loans, which could in turn be used to design a high-quality CMBS product.

BAML’s Historical drivers of CRE loan performance report identified recurring factors that created defaults and losses among European CMBS loans, including vintage, leverage and property type.

It found that all €3.5bn of the European CMBS losses to date were concentrated in 150 loans made between 2005 to 2007 (see fig 1 below), and suffered as a result of the increase in property values at that time and the scale of the subsequent correction.

However, the report also says that weaker underwriting standards at the market’s peak  may have contributed to the poorer performance of the 2005-07 vintages.

Some examples of this involve loan documentation, which should entail good risk assessment of individual properties and loans, full understanding of the tenant(s), sufficient provision for maintenance and insurance of the property etc.

Defaults rise with LTV levels

In terms of capital value, the report found both default rate and loss severity of loans increased as the initial loan-to-value (LTV) ratio increased. Among loans with 60-70% initial LTV ratios, 10.4% experienced a loss. This rose to 20.2% and 25.6% for loans with initial LTV ratios of 70-80% and 80-90% respectively (see fig 2).

Frequency of losses: Higher LTV ratio loans incurred greater losses

Regarding debt service, the lower the initial interest cover ratio (ICR), the more likely loans were to default. Just 2 of the 68 loans with initial ICRs of 2.5 or greater  experienced a loss, while almost half of the 26 loans with an initial ICR of between 1.0 and 1.1 experienced a loss (see fig 3).

REC 05.15 - p 10 chart 3

BAML said while no two commercial properties or loans are alike, its findings suggest that there are factors that broadly contribute to good credit performance, which could be used to minimise the likelihood of future loan defaults.

Applied to CMBS, they could be used to design a product that qualifies for the ‘high-
quality securitisation’ type 1 label regulators are developing. Currently under Solvency II, all CMBS deals are assigned to the ‘type 2’ group, which is more costly to hold than ‘type 1’ deals and so limits the investor pool.

Alexander Batchvarov, head of international structured finance research at Bank of America Merrill Lynch Global Research, says: “You cannot argue that commercial real estate is not part of the real economy, so to brush aside CMBS outright is wrong. That’s why it is important to create criteria that allow for at least some CMBS to enter simple, transparent and standardised securitisations. There is a lack of risk sensitivity in the current regulatory framework. Solvency II has created a massive artificial cliff, which is unjustified.”

In simple terms, regulators insist on high granularity in the assets backing securitised loans. “But it doesn’t mean that a pool comprised of one building let to five AA and AAA tenants isn’t high quality,” he says.

“I can create a high-quality, AAA tranche from any asset pool, even if properties are highly concentrated in a securitisation, by putting the attachment point for AAA at 60% and each property at a 50% LTV level. But it won’t be a high-quality securitisation for every single tranche, because the lower tranches could be affected by a single property default.”

Two ways to guarantee quality

He argues that there are two levels of high quality: one is to create an AAA tranche with a high credit enhancement so that default is unlikely, and even if it does happen, there is a high level of recovery, with the AAA note holders always repaid in full. The second approach is to impose credit criteria for the underlying loans, to create a high-quality securitisation throughout the entire capital structure.

Selection criteria to make sure underlying loans are of high quality relate to LTV levels, debt service coverage ratios and refinancing risks, according to Batchvarov.

If an LTV threshold were useful, BAML’s analysis suggests 60%, because above that,  default rates increased sharply in the past.

It also supports the idea of mitigating the volatility of property values by introducing ‘through-the-cycle’ property valuations, as promoted by the Real Estate Finance Group in last year’s Vision for Real Estate Finance in the UK report. The proposal involves sizing banks’ regulatory capital for property loans using an LTV ratio based on the long-term property valuation.

Another recommendation is for lenders to focus on stability of ICRs by selecting properties with granular pools of tenants (with limited exposure to a single business) or higher credit quality tenants, or both. Credit could also be given for scheduled amortisation, which benefits both the ICR and LTV levels of the loan, it said.

Ultimately, the importance of establishing a set of criteria for high-quality commercial real estate loans is that the high-quality label will create a greater breadth and depth of the CMBS investor base.

“Of course commercial real estate has a different set of risks, but if you can establish parameters for determining the performance of CRE loans, you should be able to create high-quality CMBS,” says Batchvarov.


BAML sampled wide range of European loans

BAML’s report analysed the performance of more than 1,000 commercial property loans across 20 European countries, totaling €157bn. These loans were securitised in 184 European CMBS transactions between 2000 and 2013, using data provided by Trepp.

The UK was the largest market covered  in the study, with 459 loans totalling €72bn, followed by Germany at €52bn (321 loans); the Netherlands, at €10bn (33 loans);
France at €8bn (94 loans); and Italy, at €7bn (58 loans).

By type of use, office properties were the largest sector, at €44bn, followed by mixed-
use property, at €35bn; retail property, at €29bn; multi-family property, at €27bn; healthcare property, at €5bn; and industrial property, at €4bn.


Multi-family loans were as safe as houses

BAML’s research found that loans backed by multi-family property were least likely to default. Just 7% of multi-family housing loans originated from 2005 to 2007 suffered a loss and among the five multi-family housing loans that did suffer a loss, the average loss severity was 18.6%.

REC 05.15 - p 11This compares to loans backed by industrial or distribution property at the other end of the spectrum, which incurred the highest losses.

Among the 2005 to 2007 vintages, 10 industrial loans (19% of the total) suffered a loss, while the average severity of that loss was as much as 50%.

In certain types of property, rental income tends to be either more stable or more volatile, which BAML believes has a bearing on CMBS loan default rates. For instance, income tends to be more volatile in the hotel sector than it is in the residential rented sector.

As such, hotel loans incurred the most defaults, at 22% over the 2005 to 2007 vintages. However, the resulting loss severity for hotel loans was below the average loss severity, at 14%.