Last week we looked at smaller regional banks with the greatest exposure to US commercial property loans. Now it’s worth reviewing what systemic risk — if any — could come from these loans.
This discussion isn’t all that new; see the many pieces written (by us and Unhedged) about US regional banks and commercial real estate over the past year. But the slide in office valuations is ongoing. And because New York Community Bancorp has given investors a scare, and as the Fed’s BTFP facility starts to wind down, it’s worth revisiting.
To summarise the latest takeaway from UBS: If smaller regional banks cut back on lending because of CRE pressures, they still aren’t responsible for a large enough share of credit provision that they would single-handedly slow the broader US economy.
Head of credit strategy Matthew Mish points out this week that most US loans are underwritten and held by the largest banks and investors (ie non-banks):
We have consistently downplayed the importance of US regionals in supplying credit. Our updated stock charts show non-banks and the largest banks (>250bn) hold 90% of C&I, 85% of mortgage, 75% of consumer and 60% of outstanding loans/debt (“the stock”). Over the latest one-year period, the largest banks in essence originated 100% of C&I, 100% of mortgage, 63% of consumer and 57% of CRE loans provided by the banks (“the flow”).
Of course, regulatory guidelines now impose an important constraint on banks as well. There are plenty of post-GFC examples of banks acting cautiously long before they run into any existential risks. So could banks sharply curb their activity in the absence of a crisis, slowing the economy? Possibly.
But the banks that are most affected are relatively concentrated in CRE markets, as Mish writes:
But . . . individual bank exposures are more nuanced, and . . . what regulators are looking at is an increasingly important consideration.
For more granularity we used the 2007 regulatory guidance, recently updated, to isolate banks that could face more scrutiny over CRE exposures (specifically those with CLD loans > 100% of total capital, or those with CRE > 300% of capital and >50% growth over 3 yrs).
We show the cohort of banks that meet either criteria in terms of share of bank loans and bank and non-bank loans/ debt. On the latter, this cohort holds about $1.5tn in assets and $1.1tn in net loans (~3%); by major sector they account for 1-2% of C&I and mortgage debt, less than 1% of consumer, and 10-25% of CRE loans (dependent on sub-sector).
So credit is more likely to tighten within other parts of commercial property markets, and not corporate lending, residential mortgage lending or consumer lending:
Larger banks could face some turbulence around stress-test time, but it wouldn’t only be the CRE meltdown that causes it, says UBS. With our emphasis:
A different approach to estimating banks at-risk based on insolvency tests, due to CRE stress and market value declines from higher rates, implies banks with $1-1.4tn in assets could be at-risk. But note we view these estimates as the upper end, arguably playing out over time, and not incorporating reactive monetary policy action.
In other words, a widespread global tightening of credit conditions would have to come from higher-for-longer interest rates (and presumably inflation), not a steep decline in office property values.
Still, somebody’s going to be holding the bag for the steep decline in US office values. So who are they? UBS breaks it down:
One takeaway from the chart above is that banks with less than $10bn in assets hold nearly 20 per cent of CRE loans. Their share is minimal in other loan markets; only in the market for mortgages do they seem to get past 5 per cent.
But non-banks still hold the most CRE debt, it seems, along with a large share of mortgages, corporate loans and auto loans. They’re even big in consumer lending, where large and mid-sized banks have a greater presence.
So what does it mean for the financial system that non-banks and investment funds are so exposed to this risk? From one perspective, this is the whole point of the originate-to-distribute model that regulators encouraged for banks after the GFC. While it’s not great for a fund investor or retiree to be left holding the bag for a busted loan, it shouldn’t in theory cause a credit crunch.
Except! Banks have also been doing a lot more lending to non-banks since the financial crisis. That means non-banks could themselves be an avenue for stress to feed back into the banking system.
So who exactly are the recipients of all this credit?
The three biggest recipients of bank credit commitments are 1) “other vehicles”, which UBS says includes closed-end funds, pensions, hedge funds and other investment funds 2) non-bank mortgage brokers, along with property lenders or lessors, and 3) corporate ringfencing or securitisation vehicles like special purpose entities, collateralised loan obligations and asset-backed securities.
This is interesting. It’s difficult to tell exactly how much (or whether) a commercial property bust could rebound back on to the banks from any of the categories above. Commercial real estate CLOs do exist, though the market is primarily one for corporate lending. The most prominent non-bank mortgage brokers lend against residential real estate, which is still a strong market.
UBS reckons that the biggest macro exposure for all of these non-banks is global corporate credit markets, which are still trucking along:
. . . our more constructive corporate debt outlook given robust US credit conditions, easing stress in riskier credit sectors, and strong EU credit technicals should placate some of the flashpoints for shadow banking risks.
In other words, too-high interest rates can cause plenty of problems: Corporate debt strains, bank bond-portfolio losses, job losses, etc. But on its own, the office-property bust shouldn’t cause a financial crisis.
