Blog Post < Previous | Next >
SAn Francisco Local Government
Exploring the prospective commercial real estate mortgage crisis
A prospective commercial real estate crisis has been on everyone's mind since COVID hit, triggering the onset of a new Work From Home (WFH) era. WFH has caused a massive rise in office vacancy rates. The latter are rising everywhere, especially in the following cities:
- San Francisco, vacancy rate rose from 6% in 2020 Q1 to 25% in 2023 Q1;
- New York/Manhattan, vacancy rate rose from 9% in 2019 Q4 to 21% in 2023 Q1;
- Portland/Central Business District, vacancy rate rose from 10% in 2019 Q1 to 21% in 2023Q1.
In some of the graphs above, you can see office rent trends. In several cities, rent prices are already declining (i.e. San Francisco). In many others, they are flat. It may be only a matter of time until office rents adjust downward everywhere.
The above trends (rising office vacancies and flat to declining rents) caused by WFH are the foundation of a prospective commercial real estate mortgage crisis.
What is the scale of the prospective commercial real estate mortgage crisis?
Within the table above I am benchmarking the commercial real estate mortgage crisis with the S & L Crisis during the first half of the 1990s and the Great Recession & Housing Crisis of 2007 - 2009.
- The S & L Crisis was associated with $394 billion in S & L assets representing 8.4% of total banking industry assets at the time ($4.7 trillion).
- The Great Recession was associated with over $2.2 trillion in residential mortgages representing 17.2% of total banking industry assets.
- The prospective commercial real estate crisis is associated with $1.15 trillion in assets which amount to just 4.9% of total banking industry assets (FDIC data 12/31/2022).
I have to mention a couple of caveats.
The first caveat, not the entire residential mortgage sector was underwater during the Great Recession. The subprime mortgage market probably represented only a small portion of the $2.2 trillion in residential mortgages. However, this entire sector was destabilized by the subprime mortgage crisis. The fragility of this credit sector was in part due to residential mortgages' loan-to-value (LTV) ratios being most often way too high at close to 100%.
Second caveat, I have not found granular data to capture commercial real estate mortgages on office buildings only. I relied on FDIC data as of 12/31/2022. And, it had a reasonably good proxy for such mortgages by focusing on mortgages where the main source of repayment is from leases on the property (I refer to such mortgages as CREM1). So, the figure of $1.15 trillion would include mortgages not only on office buildings, but also on shopping malls, and other commercial buildings being leased. I
When we compare the exposure to commercial real estate mortgages (as specified) vs. residential mortgages during the Great Recession, there is an enormous difference in credit risk between the two. Within commercial real estate lending the standard LTV ratio is 75%. As mentioned earlier, during the Housing Bubble the LTV ratio in residential mortgage lending was often near 100%.
The prospective commercial real estate mortgage crisis will accelerate banking concentration
The US has by far the most fragmented banking system in the developed World. The top 5 US banks account for only about 40% of banking industry assets and deposits (FDIC data as of 12/31/2022). This level of banking concentration is far lower than in Japan, Australia, Canada, France, Switzerland, and anywhere else that has a developed banking system. Within these countries, the top 5 banks typically control 90% of banking assets and deposits.
The US banking industry has experienced an ongoing rapid consolidation. Back in 1986, the US had over 18,000 banks. Nowadays, it has just above 4,700 ones (FDIC data 12/31/2022). Below I highlight the trends in the declining number of banks by focusing on three different periods:
- the overall long-term trend when the number of banks has contracted by - 3.6% a year;
- the S&L crisis when the number of banks contracted much faster by - 4.4% per year;
- the Great Recession, when the number of banks contracted by - 2.7% a year.
The graph above indicates that the pace of banking concentration after the S & L crisis slowed down. Even during the Great Recession, this slowdown persisted. One can expect the pace of banking concentration to greatly accelerate going forward due to the future commercial real estate mortgage crisis and its aftermath.
The US banking industry could concentrate a great deal. It has over 3,700 banks with less than $1 billion in assets. They account for 79% of the total number of banks, but only a very small fraction of banking industry assets. These banks may not have the economies of scale to survive.
Within a decade, the US banking system could have only about 1,000 banks (vs. over 4,700 nowadays) as all the banks with less than $1 billion in assets get merged into bigger banks. Even so, the US banking industry would still be far more fragmented than any of its foreign counterparts.
How many banks are in trouble?
It depends on how you slice the data (FDIC 12/31/2022). As a first cut, I look at banks that have commercial real estate mortgages (CREM) that are equal to more than 3 x their equity capital. This is an actual OCC regulation that applies to all banks.
The table and graph above count the number of banks that fall within various pairing criteria. The number of banks in trouble, as defined, peaks when looking at banks that have less than 10% in Equity Capital and more than 30% in CREM as a % of total assets.
Next, let's look at the banks that have a large exposure to CREM1 (where the source of repayment comes from leases). CREM1 represents the higher-risk component of commercial real estate mortgages.
The table below focuses on banks that have exposure to CREM1 greater than 25% of assets.
None of the large banks (assets > $100 billion) have a CREM1 exposure greater than 25% of assets. It is within small banks with $1 billion to $10 billion in assets that you see the greatest concentration of such troubled banks (85 out of 829 banks or 10.3% of such bank asset category).
Next, let's focus on the mix of such banks by asset category. Now, we see that the smallest banks with assets of less than $1 billion account for over 58% of such troubled banks (or 129 out of the total of 222 of such troubled banks).
Next, let's explore how exposed are these banks to uninsured deposits (FDIC 12/31/2022). A high exposure to CREM on the asset side combined with high exposure to uninsured deposits on the liability side makes for a lethal combination. Such banks are far more vulnerable to bank failure.
Within the graph above, I focused on banks with CREM greater than 30% of assets and Equity representing less than 10% of assets. We already know that there are 298 such banks. And, I look at different levels of exposure to uninsured deposits ranging up to > 60% of the total balance sheet. As shown, the number of such banks drops rapidly as you increase the exposure to uninsured deposits. They drop from 298 when you do not use an uninsured deposit criteria to 67 if you use > 40% and only 21 if you use > 50%.
As the next data cut, I'll focus on identifying the troubled banks with the following criteria:
- CREM1/Assets > 25%
- CREM/Assets > 30%
- Uninsured deposits/Assets > 50%
- Equity/Assets < 9%
As shown above, when I use the mentioned criteria we end up with only 8 banks out of 4,715 ones (FDIC data 12/31/2022). All of them are within the $1 to $10 billion in assets category.
As explored within the above section, figuring out how many banks are in trouble relative to their exposure to the commercial real estate mortgage sector is very sensitive to what criteria you are choosing.
This CRE mortgage crisis will put downward pressure on the valuation of banks exposed to this sector. Among the 8 weak banks I identified above, only one of them had publicly traded stocks. All the other ones were private. This is probably the case for the majority of banks with less than $10 billion in assets (or 4,555 out of 4,715 banks in the US!).
However, when looking at overall bank valuations using broad-based bank indices, the valuation of publicly traded bank stocks is already way down.
The two NASDAQ bank indices shown above are down by around - 25% since the beginning of the year (as of May 24, 2023). I suspect that the actual valuations of the smaller private banks exposed to the commercial real estate mortgage sector have decreased even more.
Exploring the Mechanics of a commercial real estate mortgage crisis at the individual bank level
To explore the mechanics of a commercial real estate mortgage crisis, I built a simple model. I use the following starting assumptions:
- Interest rate 3.0% (before COVID such mortgage rate levels were common)
- LTV 75%
- Amortization 15 years
- Office vacancy rate 5% (before COVID not unusual)
Using the above, I calculate the resulting lease payment to service the associated mortgage on a commercial building.
Next, I sensitize the interest rate upward ranging from 3.0% to 6.0%. I also sensitize the office vacancy rate from 5% to 25%.
Using the original lease payment, I calculate the PV associated with higher rates and higher vacancy rates. The PV is much lower than the original amount of the mortgage. The resulting valuation is shown in Table 1 below. As shown, if the interest rate rises from 3.0% to 6.0% and the vacancy rate rises from 5% to 25%, the value of the building drops by - 35.2%. The value of the building represents the PV of the original lease payments adjusted for a higher vacancy rate and discounted by the higher mortgage interest rate.
Table 2, shows the resulting loss on the CRE mortgage given the original LTV of 25%. The zone highlighted in red shows the scenarios when the bank would actually incur a loss on this mortgage. Thus, when interest rates rise to 6.0% and vacancy rates rise to 25%, the resulting loss on this CRE mortgage is - 13.6%.
Table 3, shows the impact on the Equity/Assets ratio of that bank assuming that its starting Equity/Assets ratio is 9% (close to the banking industry median of 9.3% as of 12/31/2022. I also assume that the bank has high exposure to commercial real estate with a CREM/Assets ratio of 40%. Thus, when interest rates rise to 6.0% and vacancy rates rise to 25%, this bank's Equity/Assets ratio declines from 9.0% to 3.7%. Such a decline in Equity Capital would certainly result in a bank failure.
Below, tables 4, 5, and 6 replicate the same calculations as tables 1, 2, and 3 above. The only difference is that the calculations below add the impact of a decrease in rent levels of - 5%.
Leveraging the individual bank model to explore the banking industry exposure to commercial real estate mortgages
Here, we are giving it a second look using a different methodology. Using the model disclosed above and the following assumptions:
- LTV 75%
- Increase in mortgage rates from 3% to 6%
- Increase in vacancy rates from 5% to 25%
We next figure out by how much the Equity/Assets ratio would decline given the increase in interest rates and vacancy rates associated with different levels of exposure to commercial real estate mortgages.
As shown above, given the specified assumptions, if a bank's exposure to CREM equals 20% of assets under Scenario 1, its Equity/Assets ratio would drop by - 2.6 percentage points. Scenario 2 simply adds another assumption: a decline in rent level of - 5%.
Next, we figure out the number of banks associated with different levels of CREM exposure. Here I differentiate between the more risky commercial real estate mortgages where the repayment comes from leases (office building and shopping malls, I refer to those as CREM1) vs. all commercial real estate mortgages that include owner-occupied properties with lower credit risk (CREM).
As shown above, there are 469 banks with a 20% or greater exposure to CREM1. These banks, given the specified assumptions, could incur a drop in their Equity/Assets ratio of - 2.6 to - 3.4 percentage points or much more. Those are already material drops in Equity Capital that could often cause a bank failure.
One could argue that my assumptions are rather dire (increase in mortgage rates from 3% to 6% and vacancy rates from 5% to 25%). However, the assumptions are not unrealistic. Commercial mortgages are often fixed for the first 5 years. And, they incur a rate reset at the end of the 5th year. Commercial leases also have often terms of 5 years. Thus, consider a building that was financed in 2018. At the time of the mortgage rate reset in 2023, the interest rate would be a lot higher, the office vacancy rate would also be much higher, and the rent could be much lower. All of those would cause a marked drop in the value of the building.
Conclusion
As reviewed, the commercial real estate sector is likely to face ongoing headwinds including:
- high interest rates,
- high vacancy rates, and
- declining office rents.
All of those adverse trends are caused by WFH which will most probably maintain a certain level of permanence. WFH just makes too much sense from both economic and environmental standpoints to be ignored.
This potential crisis does not have the same scale as the S & L crisis and the Housing Crisis & Great Recession. This future crisis is unlikely to destabilize the banking system like the Housing Crisis did. However, lower office building values and related mortgage defaults will cause many banks to fail or merge. This will contribute to an acceleration in the ongoing consolidation of the US banking industry.