As the U.S. economic expansion enters its seventh year, it is constructive to consider how commercial property revenues perform relative to U.S. unemployment rates and to evaluate property revenue volatility among the five major property types (hotels, multifamily, industrial, office, and retail). To that end, Standard & Poor’s Ratings Services performed an historical revenue volatility analysis for each of the five major property types.

Our findings indicate that hotel revenues are the most sensitive to changes in the unemployment rate, which makes them more susceptible to default. Like the overall U.S. economy, the U.S. lodging sector is now entering its seventh consecutive year of growth as measured by revenue per available room (RevPAR). However, the growth rate appears to be slowing, and in several larger markets, RevPAR has declined thus far in 2016. As hotel concentrations have increased in commercial mortgage-backed securities (CMBS) pools since the financial crisis, we believe our analysis indicates that a relatively more cautious approach to evaluating the lodging sector is warranted, and supports the use of lower loan recovery rates relative to the other major property types.

 To perform this analysis, we measured the correlation between national and regional rents/RevPAR for each of the major commercial property types with national unemployment rates over the past 18 years. Next, we calculated the annual percentage change in revenue for each property type using CMBS loan data over the same time horizon to analyze revenue volatility during various time periods. Finally, to evaluate the degree to which revenue volatility impacts loan performance, we examined debt service coverage (DSC) levels as well as default and recovery rates by property type for the underlying commercial loans.

 Of The Five Major Property Types, Hotels Have The Highest Correlation Between Rents/RevPAR And Unemployment Rates

Almost 1:1 (inverse) correlation between national RevPAR and unemployment rates

To test the sensitivity of commercial property rent/RevPAR data to the national unemployment rate, we ran a simple linear regression between these two variables, with the change in rent/RevPAR as the ‘y’, or dependent variable, and the change in the unemployment rate as the ‘x’, or independent variable. The results showed that hotels have almost a 1:1 correlation with unemployment (-0.94 correlation), followed by multifamily (-0.88), industrial (-0.67), office (-0.62), and then retail (-0.32). Chart 1 shows the clear inverse relationship between rent/RevPAR growth and national unemployment for the five property types.

Article 7 Chart 1

Hotels also have the strongest correlation between local regional rents/RevPAR and national unemployment rates

We also tested the correlation between local rents/RevPAR and national unemployment. Some markets display a very high (inverse) correlation, while other markets that may benefit from insulating factors provided by specific industries within a region (such as government, global tourism, or oil exploration) show lower correlation coefficients. The property types that typically have longer lease terms (industrial, office, and retail) than both hotel and multifamily properties also show a weaker relationship between rents and unemployment. Not surprisingly, the results again show that within local markets, the strongest correlation between rent/RevPAR (the dependent variable) and unemployment (the independent variable) was for hotel properties. The correlation between RevPAR and unemployment for the 25 local markets sampled averaged -0.80 for hotel properties; in 19 of the 22 markets where data were available for all five property types, hotels show the highest correlation to unemployment based on our annual percentage change metrics (see table 1). Among the remaining major property types, multifamily again demonstrates the highest correlation within the local markets and on a national level.

Table 1 – Correlations Between Annual Percentage Change In Rents/RevPAR And Annual Percentage Change In The National Unemployment Rate: 1997-2015(i)

Favorably, January payrolls increased in 325 of the 387 metros that are tracked by the U.S. Bureau of Labor Statistics, which paints a near-term positive story for commercial real estate markets. However, the correlation between commercial property rents/RevPAR and unemployment leaves us somewhat concerned, as the underwriting for hotels and multifamily properties, which show the strongest correlations to unemployment, must be based on a long-term sustainable level in order to mitigate any sudden economic setbacks.


With RevPAR Growth Showing Signs Of Slowing, And High Lodging Concentrations in Recent U.S. CMBS Deals, The Strong Hotel Correlations Are Concerning

Although national U.S. lodging sector RevPAR growth continues, the slowed pace of growth, together with hotels’ high concentrations in recent U.S. CMBS deals, have caused the strong hotel correlations to become a concern, in our view.

U.S. RevPAR showing signs of weakening

The RevPAR gain of 6.3% in 2015 was primarily a result of average daily rate (ADR) increases as hotel operators lifted rates in an environment of record high occupancy levels. The 2015 occupancy level of 65.6% was the highest in two decades. Memories of the unprecedented 17% decline in RevPAR during 2009 are fading, and many prognosticators, including Standard & Poor’s Corporate Ratings lodging analysts, forecast RevPAR gains to continue through 2017, albeit at a moderately lower growth rate than the 5%-8% annual increases gained since 2010.

Article 14 Chart 3

While our U.S. lodging sector outlook for 2016 and 2017 remains positive, pockets of weakness have begun to emerge, with RevPAR only increasing by 2.8% in February and 2.4% in January, and a larger share of the top 25 markets recording year-over-year declines than we have seen in many years. In both January and February, that figure grew to nine of the top 25 lodging markets, including New York, Houston, Miami, and Chicago. In addition, supply is set to grow in the next couple of years after hovering at less than 1.0% between 2011 and 2014, with major urban markets experiencing declines in international visitors as the dollar strengthens and room sharing services like Airbnb beginning to have an incremental negative effect in certain markets.

U.S. CMBS lodging exposure increased sharply in recent vintages

As RevPAR has increased, U.S. CMBS exposure to the lodging sector has also risen. The conduit sector’s share of hotel exposure increased to 17% in 2015 offerings from 3% in 2010 (see chart 3), with some deals as high as in the low-to-mid 20% range (some deals with exposure as low as the mid-single digits were exceptions). In addition, hotels were the most securitized property type in 2015 single-borrower deals, accounting for roughly 27% of issuance by balance. While investors and other market participants have sounded cautionary tones on the sector at conferences and in meetings, the trend of elevated exposures to the sector has shown no signs of reversing. Several recently announced large hotel company/portfolio purchases may also support single-borrower issuance.

Article 14 Chart 4

Lodging sector may experience less accommodative refinancing conditions

Beyond this recent issuance trend, it is noteworthy that 16% ($26 billion) of 2016 and 2017 maturing loans are backed by hotels. While admittedly incomplete, our data sample indicates that hotel debt yields have benefitted from several years of strong revenue gains, as the majority of maturing hotel loans have current debt yields over 10% (see table 2). We note that some other property types, which have been slower to rebound, have higher percentages of loans with debt yields under 8%.

Table 2 – Amount Maturing And Debt Yield Summary 2015-2017(i)

Notwithstanding the relatively low percentage of hotel loans with less than 8% debt yields maturing in the near term in our sample, we note that rising concentrations of hotels, combined with revenue volatility and a potential for rising interest rates, could lead to less accommodative refinancing conditions for the sector when newly originated loans approach their maturity dates.


Hotels Show The Most Revenue And Cash Flow Volatility To The Downside

Beyond the increase in CMBS exposure, the lack of long-term leases, high capital investment needs, operational risk, and exposure to event risk all contribute to a much higher level of net cash flow volatility for hotels relative to the other major property types. This volatility stems from the low operating margins for hotels, which typically range from 20% to 25% for full-service hotels and 30%-35% for limited-service hotels. Thus, changes in hotel revenue can, and often do, result in net cash flow changes twice that rate, eroding DSC at a faster pace than the other major property types. So while hotels were one of the first property types to rebound after the recent recession, they were also the first to exhibit weakness in 2008 as economic conditions began to deteriorate.

We illustrate this volatility in chart 4 by creating a performance tracker that shows the annual percentage change in the weighted average revenue per unit for loans from each of the five major property types. Our sample includes well over 60,000 loans that were securitized within Standard & Poor’s-rated deals or were pari-passu with loans in Standard & Poor’s rated deals between 1998 and 2014:

Article 14 Chart 5

We highlighted several points in chart 4:

  • Hotels clearly emerged as the most volatile property type in the data sample for the time period covered. Most prominently, hotel revenues within our securitized population dropped by over 15% between 2008 and 2009. This compares with some relatively smaller declines for the other property types during 2009-2011 in the period around the Great Recession; for example, roughly 5% and 4% declines for retail and industrial, respectively, in the 2009-2010 period, a 2.5% decline for office in 2011, and a 2% drop for multifamily in 2009.
  • Hotels also experienced an average revenue drop in 2001 and 2002 after 9/11, while none of the other major property types exhibited an average revenue decline in this period.
  • The longer-term leases of industrial, office, and retail tend to limit gains in the stronger performance periods and constrain declines during more distressed periods, resulting in more of a performance lag than hotels. While multifamily is sensitive to employment, stable vacancy levels may be providing more stable gross revenue performance.

Cash flow volatility can lead to volatile DSC and more defaults when DSC drops below 1.0x, especially for hotels

To account for their significant cash flow volatility, hotels are usually underwritten to a higher DSC level at origination relative to the other major property types. On a weighted average basis, the DSC at origination for hotel properties has remained above that of other property types in legacy deals (2008 and before), CMBS 2.0 (2010-2011), and CMBS 3.0 (2011-2016) (see table 3).


Table 3 – Securitized Debt Service Coverage By Property Type(i)


In our analysis, we calculated the number of loans within our sample, by property type, that dropped below a 1.0x DSC. Loans secured by hotels have a higher propensity (33%) to fall below a 1.0x DSC than the other property types (see table 4), despite typically being originated at a higher going-in DSC. More importantly, of the loans whose DSC fell below 1.0x at some time during their loan term, 47% ultimately defaulted, which we define as having gone 60 or more days delinquent. Office had the next highest propensity to default subsequent to DSC falling below 1.0x, but at a much lower 32%.


Table 4 – Default Rates By Property Type For Loans Below 1.0x DSC And Resulting Loss Severity


Table 4 also shows the loss severity rate for the loans that met our definition of a default. Retail and hotels had the highest severities, at just over 50%. For retail, aging properties in secondary and tertiary markets can lead to somewhat binary outcomes; thus, it is not surprising that underperforming loans would liquidate at the highest average severity. Multifamily was the top performer, at 37%.


Table 5 looks at the percentage of each property type by loan count that had a DSC less than 1.0x in any given year. Hotels had the highest percentage of loans with a DSC under 1.0x in nine of the 18 years examined. Notably, hotels demonstrated the highest percentage of loans below a 1.0x DSC between 2001 and 2003, largely because of the impact on the U.S. economy from the events of 9/11, in addition to the technology bust that occurred during this period. Similarly, the lodging sector again exhibited the largest increase in the percentage of loans with a DSC under 1.0x in 2008, as U.S. economic conditions started to weaken and then jumped by a staggering 8.8 percentage points in 2009 to 14.1% as the recession took hold.


Multifamily had the highest percentage of low DSC loans in the years leading up to and during the peak CMBS issuance years. This is likely due to the fact that homeownership levels peaked during this time, hovering around 68%-69% in 2004-2007, compared with a current level of approximately 64%. Also as previously noted, multifamily is the most aggressively underwritten property type, in terms of having a lower starting DSC, on average. So although the DSC for multifamily loans frequently fell below 1.0x, table 4 indicates that this didn’t lead to defaults nearly as frequently as it did for hotel loans. Additionally, multifamily properties show the lowest loss severity in the sample for loans that met our definition of default.


Table 5 – Percentage Of Loans Below 1.0x Debt Service Coverage By Property Type (1997-2014)


Hotels’ Downside Risk Necessitates A More Conservative Analysis Of Property Cash Flows And Credit Enhancement Levels

Despite currently healthy property fundamentals, we believe our findings clearly show that hotels are more volatile than the other major property types that collateralize CMBS transactions. While hotel demand is dependent on many factors, hotel sector revenues exhibit nearly a 1:1 inverse correlation with unemployment rates. If employment levels falter, hotel revenues will likely quickly follow. This volatility, coupled with the sector’s low cash flow margins, results in loan-level metrics that demonstrate more volatile DSCs, higher default rates when DSCs fall below 1.0x, and lower recovery rates relative to the other major CMBS property types.

 While the percentage of hotel loans contributed to CMBS transactions continues its upward trajectory, it is our view that caution is warranted when analyzing this property type. Our analysis of hotel properties aims to derive a sustainable net cash flow and value that can endure both upward and downward market fluctuations, typically reflecting more conservative RevPAR assumptions than what recent market conditions and property performance may indicate. Additionally, our stand-alone loan-to-value thresholds, which are used to determine expected principal recoveries in the event of default, are significantly more conservative for hotel properties (e.g., 35.0% at the ‘AAA’ rating level compared to 47.5%-50.0% for the other four property types). As a result, a collateral pool with a higher percentage of hotel loans will typically require higher credit enhancement levels up and down the capital structure based on our analysis.