As the coronavirus (COVID-19) continues to spread across the country, the U.S. lodging industry has experienced a rapid, systemic shock over the past few weeks. While performance data is limited at this point, weekly revenue per available room (RevPAR) declined by 11.6% in the first week of March and dropped 32.5% in the second week (per STR Inc. data), and all chain scales and locations have been broadly affected. The unprecedented COVID-19 containment efforts have essentially brought lodging demand to a halt.

S&P Global Ratings acknowledges a high degree of uncertainty about the rate of spread and peak of the coronavirus outbreak. Some government authorities estimate the pandemic will peak in June or August, and we are using this assumption in assessing the economic and credit implications of the pandemic. We believe measures to contain COVID-19 have pushed the global economy into recession (see "COVID-19 Macroeconomic Update: The Global Recession Is Here And Now," and "COVID-19 Credit Update: The Sudden Economic Stop Will Bring Intense Credit Pressure," both published March 17, 2020), which could also negatively affect employment levels, housing, and commercial real estate markets. As the situation evolves, we will update our assumptions and estimates accordingly.

With the prospect of a recession on the horizon, we expect the impact to the lodging sector to be exacerbated as hotel revenues exhibit nearly a 1:1 inverse correlation with unemployment rates and demand is highly correlated with the performance of the economy.

U.S. CMBS Exposure To Hotels Has Waned, But Remains Significant

At year-end 2019, the U.S. lodging industry had experienced 10 consecutive years of RevPAR growth. As RevPAR steadily climbed after the 2009 recession, U.S. commercial mortgage-backed securities (CMBS) exposure to the lodging sector also increased. Between 2015 and 2018, about 15%-17% of CMBS conduit collateral was backed by lodging properties before dropping to 12% in 2019. On the single-asset single-borrower (SASB) side, the percentage of lodging transactions over the past five years rose sharply and peaked at over 40% in 2017 and 2018, before falling to 25% last year and to 16% thus far in 2020. Nevertheless, lodging remains a significant part of collateral pools (see charts below):

Chart 1

Chart 2

Risk Factors Across Lodging Sector

The lodging sector is unique in that it is almost always the first commercial property type to feel the impact of changes in the economy, both positive and negative. The hotel sector does not benefit from long-term leases, which makes it relatively more volatile versus other major CMBS property types and extremely vulnerable to events that affect travel. Catastrophes such as 9/11 and economic downturns like those experienced in 1991 during the Gulf War and the 2009 recession resulted in RevPAR declines across the nation as travel and demand for lodging subsided during these periods.

In 2001, RevPAR declined by about 7% due to the events of 9/11 which further strained a sector that was already weakened by the bursting of the dot-com bubble. RevPAR declined further in 2002 by about 3% and took about two years to return to pre-2001 levels. Notably, during 2001, while people temporarily avoided flying, drive-to lodging destinations prospered as leisure travelers sought to benefit from discounted rates.

During the 2009 recession, RevPAR for the sector started turning negative in the second half of 2008 and was negative for 19 consecutive months. Monthly RevPAR declines averaged about 16% in 2009 before finally turning positive in mid-2010. After dropping by a whopping 17% in 2009, RevPAR for the U.S. did not reach the pre-recession 2008 level (unadjusted for inflation) until about 2012, and many major markets took even longer to rebound. During the 2009 recession, leisure travel again remained strong and helped mitigate the drop-off in corporate transient and group demand. Nevertheless, the lodging sector exhibited some of the worst credit metrics of all property types through the Great Recession. Based on lifetime loan performance data through June 2012, the cumulative default rate for S&P Global Ratings-rated CMBS backed by hotels was 24.8%, the worst performer of all major property types (the next worst property type was office at 16.7%), with an average loss severity of 43.9% (see North American CMBS Default And Loss Study: Positive Credit Metrics Carry Over Into First Half-2012, published Sept. 24, 2012).

Chart 3

While the lodging sector has endured a plethora of unfavorable conditions and demand-limiting events in the past, the duration and severity of the impact of a global pandemic on the lodging sector is unknown. It is an unexpected, untested phenomenon with few points of comparison. The virus has brought about unprecedented curtailment measures, almost all of which are resulting in an effective demise of demand from corporate, leisure, and group travelers. Since the COVID-19 outbreak, there has been a dramatic decline in airline passenger miles stemming from governmental restrictions on international travel and a major drop in domestic travel. In an effort to curtail the spread of the virus, group meetings, both corporate and social, have been canceled, corporate transient travel is restricted, and leisure travel is halted due not only to fear of travel, but also closures of major demand generators, including amusement parks, casinos, and ski resorts, as well as cancellation of virtually all major sporting events.

In addition to the sector's inherent revenue volatility, lodging properties have historically exhibited the lowest net cash flow (NCF) margins of any major property type. Therefore, changes in hotel revenue can result in NCF changes approximately twice that rate, eroding debt service coverage (DSC) at a faster pace than that of the other major property types. With only two weeks of data since the COVID-19 outbreak took hold in the U.S., it is not yet possible to project how RevPAR will fare in 2020. However, given the projection that the number of COVID-19 cases will not peak until summer, we expect that U.S. RevPAR will decline by double digits for the foreseeable future.

To gauge the potential impact of high-magnitude room revenue declines on hotel NCF and DSC, we evaluated the impact of a given RevPAR change on the NCF and DSC for both a typical full-service and limited-service hotel with a NCF margin of about 25% and 40%, respectively. We assumed about a 90% S&P Global Ratings' loan-to-value (LTV) ratio for each scenario, which is similar to the leverage point of recent SASB and conduit hotel loans, and assumed RevPAR would decline by 8%-40% from current levels, with a roughly equal decline in occupancy and average daily rate. We also assumed an interest-only (IO) loan given that most SASB loans are interest-only and due to the large percentage of IO loans in conduit transactions in recent years.

Table 1

For a full service hotel loan with a starting DSC of about 2x, a 22% RevPAR decline would result in a NCF drop of about 48%, which is the point at which DSC would fall to 1.0x. Due to its higher NCF margin, the limited service hotel could withstand a RevPAR drop of about 28% and a corresponding NCF decline of about 51% before DSC falls below 1x.

Clearly, at these high levels of RevPAR decline, we expect hotel NCFs will initially drop significantly. Assuming hotels remain open and have some level of occupancy, operators will be able to cut expenses initially by lowering staffing levels, closing certain food and beverage outlets, and eliminating non critical amenities; however, a minimum level of staffing would need to be maintained to be operational, and fixed expenses like taxes and insurance are immutable. Hence, we expect large near-term NCF declines across all hotel types. On a positive note, floating-rate loans with low interest rates should benefit from the decline in LIBOR rates over the last year, which may help offset some of the performance declines.

S&P Global Ratings' Approach To Hotel Cash Flow And Value

Given the inherent volatility of the lodging sector, our analysis of hotel revenue reflects more conservative RevPAR assumptions than what prior-year market conditions and property performance indicate. We have generally scaled back our RevPAR assumption two to three years in rising revenue environments and lowered RevPAR further in environments experiencing declining trends. We utilize expense levels that reflect long-term averages and reflect industry norms.

Our NCF variance to issuer NCF for the CMBS transactions backed by hotels we rated from 2015-2019 was about 18%. In addition to adjusting our NCF, we also use higher capitalization rates for hotels, resulting in higher value variances for this property type. For the deals we rated in 2015-2019, our variance to the appraised value averaged about negative 40%. In addition to adjusting our NCF and using higher capitalization rates for hotels relative to the other major property types, our analysis of loans backed by lodging properties incorporates lower recovery thresholds. Our stand-alone LTV 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 major 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.

Despite our conservative assumptions, our typical hotel analysis does not contemplate a prolonged shock or unprecedented level of stress where RevPAR remains at extremely depressed levels for long periods. While it is projected that COVID-19 cases will peak in the summer months and then stabilize, and lodging demand will certainly rebound at some point, the timing of both is unclear. We expect short-term volatility in hotel NCF and DSC levels as demand has already retreated drastically throughout the U.S. The American Hotel & Lodging Association has requested a federal bailout, which would help operators with expenses and potentially debt service on certain obligations; however, a definitive package has not been announced.

We will continue to monitor the situation as more data comes in, but given our conservative long-term cash flows, high capitalization rates, and conservative post-default recovery assumptions, we would want to make sure there is a fundamental shift in valuations before we consider taking specific rating actions in this sector.