More than two and a half years since the onset of the pandemic, there is growing conviction that the office sector may remain relatively more impaired than any other sector within the sphere of commercial real estate (CRE), even as social distancing measures have been largely removed. Although the lodging sector was more acutely affected initially, as evidenced by higher commercial mortgage-backed securities (CMBS) delinquency and forbearance rates, this sector has rebounded faster than preliminary expectations. Even the brick-and-mortar retail sector--despite preexisting pressure from ecommerce, which only intensified after the pandemic and still leaves many retail malls vulnerable--was able to recover as social distancing measures eased and in-person consumer spending rebounded significantly. While both these sectors have benefited from increased utilization and foot traffic, many office buildings, both in central business districts (CBDs) and suburban settings, remain deeply underutilized, suggesting that once-deemed "temporary" workplace arrangements have become more permanent (or at the very least hybrid) in nature.

In light of this, along with market feedback regarding, in particular, the outlook for conduit loans backed by class B quality office assets, S&P Global Ratings has taken a closer look at how office valuation declines may affect our rated universe of CMBS conduits. We present scenario analyses that focus on various assumptions for office value declines (which translate to higher default probabilities and lower recoveries) to determine how enhanced term and maturity default risks may affect credit ratings.

Long Lease Terms Shroud Current Office Supply/Demand Dynamics

Office buildings enjoy some of the longest tenant lease terms in CRE--often 10 years or more--which likely masks the true scale of their current underutilization. These longer lease terms have given office landlords time and been key in keeping overall CMBS office loan special servicing and delinquency rates at relatively low levels (3.3% and 1.6%, respectively, as of September 2022). Despite this advantage, U.S. office vacancy rates have increased steadily, to a pandemic-era high of 17.8% as of third-quarter 2022, according to Cushman & Wakefield. In New York City's three main Manhattan office markets, which collectively account for 11% of the total U.S. office market, the overall vacancy rate stood at 21.9% for the third quarter, compared to the historical average of 11.5% (for more on the Manhattan office market, see "U.S. CMBS: Remote Work And The Evolution Of Manhattan's Office Market," Feb. 3, 2022). Much of the increase in vacancy rates is being driven by increased supply in the form of new developments that were in the pipeline prior to the pandemic. However, a certain percentage of office tenants have likely begun to give up space as they "right-size" their real estate footprint with employees continuing to work remotely. This will add pressure in the form of lower renewal rates, which translate to higher vacancy rates and, eventually, lower asking rents. This new equilibrium will take years to fully form as larger shares of leases reach their expiration dates.

In the case of New York City, more than 70% of office leases signed before the pandemic have yet to come up for renewal, according to "Work From Home And The Office Real Estate Apocalypse," Sept. 26, 2022. The rate was over 60% across the U.S. Meanwhile, according to Green Street's Commercial Property Price Index (CPPI), U.S. office property prices in the 50 major markets it tracks had decreased 4% over the past 12 months as of September 2022 and were 9% lower than just prior to the pandemic. A more recent report from Green Street ("Still Chasing the Puck," Oct. 7, 2022) states that office value declines since the pandemic might actually be twice as much.

Chart 1

Highly capital-intensive properties

Office buildings, particularly ones located in the CBD of primary or gateway markets, are acutely capital intensive. In addition to funding tenant allowances, building owners periodically have to fund extensive base building improvements to cover structural elements and common areas, such as façade, HVAC (heating, ventilation, and air conditioning), lobby, and elevator system upgrades. These costs, while necessary to retain and attract new tenants, have always been burdensome for landlords as they eat away at profits. Adding further pressure, in the case of New York City's Local Law 97 passed in 2019, most office buildings are required to reduce their aggregate greenhouse gas emissions by 40% by 2030 or pay a fine. While these challenges predate the pandemic, they remain a reality and dovetail with the view that landlords, to the extent feasible, will need to invest heavily to ensure their buildings have the physical infrastructure that tenants are seeking in a post-pandemic world (better HVAC, more flexible floor plan arrangements, significant amenity packages, concierge areas, etc.).

Despite rising headwinds, delinquency rates remain low

For much of 2020 and 2021, the CRE market, along with the residential real estate and broader financial markets, benefited from fiscal stimuli, monetary policy easing, and historically low interest rates. As investors searched for yield, they moved to risk-on mode and made 2021 a record-setting year for the U.S. private-label CMBS market, with issuance reaching a post-Great Financial Crisis high of $110 billion. More recently, multi-decade record inflation, rising interest rates, and the uncertainty brought on by the Russia-Ukraine war have led to tighter financing conditions (which are likely to persist), with benchmark rates rising alongside widening credit spreads. The concern here is twofold: a market with reduced liquidity (a particular concern for a sector characterized by interest-only loans with balloon maturities) and growing risks of a broad-based recession in 2023 as successive rate hikes take hold and the financial health of office tenants, upon whom landlords depend, get tested.

But it's not all doom and gloom. As previously noted, office delinquency rates remain low even by historical standards. Many office buildings and most in CBD markets are multi-tenanted, meaning lease expiration or rollover schedules are often staggered, allowing landlords time to secure new tenants. In addition, many loans require excess cash flow to be escrowed for releasing purposes if major tenants have given notice that they will not renew. Office buildings, especially in primary markets, are major assets, often with significant land value given their close proximity to transit infrastructure, suggesting landlords have every incentive to protect their equity. Lastly, many companies have started mandating return-to-office policies ranging from hybrid to full-time, increasing office utilization rates. Cities less dependent on public transportation, such as in the Sunbelt, have experienced higher office utilization rates throughout the pandemic than their mass-transit-dependent counterparts, and this trend may very well continue.

Scenario Analysis

Considering the various headwinds affecting the office property sector, we conducted various scenario analyses to test the resilience of our U.S. conduit CMBS ratings to potential adverse value movements.

U.S. CMBS conduit rated universe

Our rated U.S. CMBS portfolio included 143 private-label conduit transactions as of September 2022 (excluding Freddie Mac transactions). In aggregate, these were secured by a total office loan count of approximately 1,128 and had an outstanding trust loan balance of $37.4 billion (see table 1).

Table 1

To put this in context, the 143 rated conduits contain approximately 6,229 total loans, comprising an aggregate trust loan balance of approximately $118 billion. So, office loan exposure accounts for about 18% by count and 32% by trust loan balance (see chart 2 for the full breakdown by property type).

Chart 2

In addition, roughly 87% of the office loan exposure ($32.4 billion) comes from 2018 or later vintages, which is understandable given the relatively heavy issuance volume in the last years of the prior decade. This indicates that many loans with 10-year terms have years before they reach their maturity date.

In chart 3, we focus on the office loan exposure for transactions in each vintage year and further bifurcate them by the quality/class of the office properties serving as collateral for the loans. We do this by referencing our assigned S&P Global Ratings office cap rates from our new-issue reports or most recent surveillance reviews, which take into consideration whether we deem the underlying office property as class 'A' or class 'B' in terms of quality. Key drivers for designating asset quality would include the in-place rents a given property commands and how those rents compare to other properties of similar quality in a given market. Other signifiers for quality include the age of the property and amenities offered at the property. We also reference our own historical dataset and/or rely on third-party vendors or appraisals to cross reference our assumptions.

Chart 3

As chart 3 shows, class B office loans (based on count) are typically far more common than their class A counterparts. This makes intuitive sense, as conduit pools benefit from the diversity gained from adding relatively smaller middle-market loans, often in the form of class B office assets located in secondary or tertiary markets. Chart 3 also exhibits how the ratio of class A and class B office loans evolve as transactions season. For conduits within five years since their issuance (2017–2022), the ratio of class A to class B office loans is roughly 1:2. However, the ratio of class B office loans far exceeds class A for older vintages. This suggests that class A office loans refinance more frequently than their class B counterparts, as they presumably benefited from higher appreciation and a greater likelihood for cash-out refinancing or sale/acquisition.

Chart 4 looks at the exposure of class A and class B office loans by their share of trust balance, whereas chart 3 focused on loan count. While chart 3 showed the higher frequency of class B office loans, chart 4 provides a more balanced picture. Though there are fewer class A loans, they come with an average loan balance that's twice the amount of their class B counterparts ($50.1 million vs. $25.1 million). Like the higher frequency of class B office loans, this also makes sense considering how the top–10 loans of conduit pools often represent anywhere from 50% to sometimes 60% of total pool balance and have individual loan balances ranging from $40 million to $100 million depending on the overall size of the deal. Loans on large, institutionally owned, class A office towers in primary markets often serve as top–10 collateral in pools. In fact, many loans are large enough to facilitate their placement in single-borrower single-asset (SASB) transactions along with various other senior pari passu notes placed in multiple conduit transactions.

Chart 4

While there is a higher prevalence of class B office loans, especially as conduits season, we view the fact that class A office loans account for roughly half of overall office loan trust balance exposure as a credit positive given their historical higher payoff rate and the expectation that such higher quality assets will fare better in a post-pandemic environment.

Chart 5 provides a distribution by balance of current ratings for 130 conduits issued since 2012. We excluded 13 conduits originated prior to 2012 as most are legacy CMBS 1.0 product with non-investment-grade ratings ('BB+' or lower) due to adverse selection pervading the transactions at this stage in their life. More than 81% of our ratings are in super-senior 'AAA's represented in 128 conduits. Junior 'AAA's and 'AA' category ratings together account for 13.9% and are often associated with a rated transaction's A-S/A-M class. 'A' and 'BBB' category ratings make up 3.5% of our conduit portfolio. The remaining portion (1.6%) is in non-investment-grade territory, primarily reflecting prior downgrades. Viewed collectively, the rating distribution of our conduit portfolio is weighted toward the higher end of the capital structure for most transactions. Although we typically assign high-investment-grade ratings, we do periodically assign 'A', less frequently 'BBB', and, on occasion, 'BB' and 'B' category ratings.

Chart 5

Conduit sample set

As previously noted, our rated U.S. CMBS portfolio included 143 private-label conduits. Of these, we selected a sample set of 50 transactions based on highest exposure to office loans (see table 2 for a summary).

Table 2

The sample set has a weighted average office loan exposure of 44.6% and ranges from as high as 75% to as low as 34% for individual deals. For perspective, our overall conduit portfolio has an office loan exposure of 32% (see table 1). In addition, our sample set broadly mirrors our rated portfolio in terms of overall rating distribution. At issuance, our sample set of conduits had a weighted average S&P Global Ratings office net cash flow (NCF) variance and market appraisal value variance of -17.7% and -40.6%, respectively. These two assumptions are key variables in our CMBS rating methodology and serve as beginning reference points for our scenario stress testing.

Scenario approach

At its core, our CMBS methodology is recovery-based and relies on our determination of a long-term S&P Global Ratings value for an encumbered asset that informs our 'B' level rating, followed by progressively lower recovery threshold assumptions as you move up the capital structure (see "Rating Methodology And Assumptions For U.S. And Canadian CMBS" Sept. 5, 2012). In the case of conduit transactions, our methodology also incorporates a lifetime default probability factor, driven by each loan's S&P Global Ratings loan-to-value and debt service coverage ratios, as well as a transaction's effective loan count. As presented in the sample set, our office valuations were, on a weighted average basis, 40.6% below appraisal/market values at the time of issuance. Such variances or discounts are typical and reflective of our cash flow analysis, which is focused on long-term sustainable trends, often resulting in issuer cash flow haircuts and the use of our cap rates, which typically maintain 150 basis points or more in spread to prevailing market cap rates. A key takeaway here is how our conservative valuation approach provides our rating model with a form of inherent or built-in cushion to account for idiosyncratic risk. Accordingly, our scenario stress testing looks to further decrease our already discounted S&P Global Ratings office NCFs and values by varying degrees.

Our scenario runs looked to further reduce our office NCFs and valuations by either 10% for class A quality office assets or 20% for class B quality assets (see table 3). This translates to a weighted average S&P Global Ratings office valuation haircut of about 14.1% for the sample pool, or roughly 1.5x the current declines observed by Green Street's office CPPI since the onset of the pandemic. While Green Street has indicated these declines are probably twice as much as of its October report, we note that S&P Global Ratings values were already deeply discounted at the time of issuance. The additional stresses being applied, though based on dated market values, would increase the sample set's initial weighted average market value discount to 49.0% from 40.6%.

Table 3

Testing runs in the scenarios initially apply NCF and value decline stresses for the smallest office loans in each conduit, then progress into the top 10, and finally all, office loans.

Ratings impact

To assess the impact on rated certificates, we ran all four scenarios in our model for the 50 transactions, applying our revised NCF and valuation assumptions on the affected properties (see chart 6). However, we first established for each transaction a set of model-indicated ratings for benchmark purposes. Our committee-assigned and published ratings for the various principal- and interest-paying classes within a given deal often reflect these model-indicated ratings, but there may be differences due to consideration of factors like bond liquidity support, volume of near-term maturities within an asset pool, credit subordination of the classes within the capital structure, and collateral performance (both historical and expected), etc. Referencing model-indicated ratings for benchmark purposes assures an apple-to-apple comparison of how ratings would perform in the various scenarios.

Chart 6

Observations

We observed the following:

• While there is a general correlation between office concentration and model-indicated rating movements (as expected), exceptions occur due to factors such as office quality makeup and available credit enhancement. These are averages, and individual pool results will vary.
• Super-senior 'AAA's were the least affected category of ratings, with a model-indicated average rating downgrade of 0.57 notches in Scenario 4 when all office loans are stressed. Our sample set included 273 classes from all 50 conduits.
• Junior 'AAA's (classes A-S/A-M) were less resilient (as expected), but still maintained a model-indicated average rating downgrade of 1.88 notches in Scenario 4. Our sample set included 16 classes from as many conduits.
• 'AA' category ratings had a model-indicated average rating downgrade of 0.44 notches in Scenario 1 and ranged as high as 3.26 in Scenario 4. Our sample set included 39 classes from 37 conduits. More than half (23 classes) referenced a given transaction's A-S/A-M class.
• 'A' category ratings showed the most model-indicated change on average, especially in Scenarios 3 and 4--2.38 and 4.69, respectively. Our sample set included 13 classes from as many conduits.
• 'BBB' category ratings had a model-indicated average rating downgrade of 0.75 notches in Scenario 1 and ranged as high as 3.50 in Scenario 4. Our sample set for 'BBB' category ratings, however, was limited (four classes from three conduits).
• 'BB' category ratings in the sample set were also limited (three classes from two conduits). Ratings in this category had a model-indicated average rating downgrade of 1.67 notches in Scenario 1 and ranged as high as 3.33 in Scenario 4.

The results indicate that super-senior classes appear well insulated to our chosen stresses, while classes in the middle of the stack, especially in the 'A' category, could see larger downward rating movements, especially if performance deterioration is more broad-based (as represented by our two harshest scenarios--3 and 4). The smaller sample sizes of 'BBB' and 'BB' ratings introduce some idiosyncratic results, although they also show greater movement relative to their more senior counterparts in the conduit capital stack. Lastly, we reiterate that our outstanding ratings on these 50 conduit transactions currently reflect base-case stressed office value declines averaging 40.6%. In this scenario analysis, we're layering in additional S&P Global Ratings value declines to bring the overall office value stress to 49.0% in the most severe scenario (Scenario 4).

Feedback is welcome

This scenario analysis is intended as an exercise in both transparency and benchmarking; any feedback is welcome. We remind market participants that any future rating actions would represent the outcome of a rating committee's assessment of individual office loan performance (in conduits, of course, it would be in the context of a diversified pool) and market trends, including tenant rollover, lease expiration timing, and a borrower's ability to fill vacancies should they arise. They would also assess the likelihood of the rated loans being refinanced by their maturity dates.

The authors would like to thank Jamil Joshi, Nikhilesh Kane, Sahil Kundra, Ambika Garg, Serena Columbo, Krista Poch, and Thomas Forte for their contributions to this report.