Brad Setser Profile picture
Mar 15 18 tweets 5 min read
All banks borrow short and lend long, playing the (typical) shape of the yield curve (& drawing on the stability of most short-term funding most of the time)

But anything taken to the extreme can become a source of risk. Silicon Valley Bank clearly took things to extreme.

As @M_C_Klein shows in an excellent post at the Overshoot, SVB "overshot" the limits of a borrow short/ lend long model --

The overwhelming majority of its deposits were large and technically uninsured. Its insured deposit base would only have supported a small bank!
The safe thing to do in the face of a big deposit influx during the pandemic would have been to hold a lot (and I mean a lot) of short-term bills and 2 year Treasury notes ...

SVB tho went for the extra yield at the long-end of the curve.

As @jeuasommenulle and others have illustrated, SVB built up a huge hold to maturity bond portfolio -- over $90b of bonds that weren't marked to market.

That included per SVBs own disclosure over $80b of Agencies (effectively funded with short-term uninsured deposits).

That is a bit nuts.

US Agency backed mortgages have one globally unique feature. They get pre-paid when interest rates fall, but no one exercises their right refinance a fixed rate mortgage when interest rates rise. So in a world of rising rates, SVB had "wrong way" risk.

The interest rate on its short-term liabilities would increase -- while the duration on its assets (and the associated mark to market losses, which were hidden in the HTM book) would increase ...

This model only works, technically, if the bank holds "safe" (as in no credit risk) assets -- as Treasuries and Agencies have low/ zero risk weights and thus don't eat up the bank's equity capital base.

Risk wasn't captured in capital charges. It was on the funding side

The 2008 gallows financial humor springs to mind --

"Q: what is the difference between a big bank and a hedge fund?

"A: A hedge fund is better capitalized"

The losses on the not marked to market Agency portfolio (and a smaller muni portfolio) were equal to SVB's regulatory capital ...

Because of the associated "gamble for redemption" risks alone the regulators should have been all over it ...

Matt Klein's post also highlighted a second. more subtle fact --

Deposit rates have been rising much more slowly than the Fed's policy rate -- a lot of SVBs deposits were still paying zero interest when it got into trouble.

So trouble was building on a second front --

SVB was obviously illiquid (no one would, prior to Fed facility, lend against underwater Agencies at par). But it was about to start bleeding income as deposit rates rose over the interest income on its legacy HTM Agencies.

The Biden Administration was, correctly I think, worried about funding contagion -- looking up SVB's big deposits until the value of its assets was determined (through a sale) was legally and analytically justified ...

But it risked a run on all the regional banks that drew heavily on large insured deposits for funding and had been exempted from the stricter liquidity regulations applied to bigger banks.…

But it is also I think clear that a lot of longer-term "safe" bonds (Treasuries and Agencies) issued between 2015 and 2021 are in portfolios that aren't marked to market. In the US of course -- I was shocked by BoA's HTM book -- but also in places like Japan.

So with US short-term rates where they are and with US long-term rates where they were, a bunch of financial intermediaries are going to be bleeding income and unable to self capitalize easily.

The BoJ was onto this risk in its latest Financial Systems Report --

it is a known known in Rumsfeld's formulation.

The associated liquidity run risk should have been a known/ unknown.

But there is no doubt that there needs to be a lot more focus globally on the risk associated with long duration instruments that are underwater and funded with short-term liabilities whose price will adjust to a higher rate world even with a lag ...

interesting *to me* detail in the semafor story on SVB/ the systemic questions it poses…

• • •

Missing some Tweet in this thread? You can try to force a refresh

Keep Current with Brad Setser

Brad Setser Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!


Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @Brad_Setser

Mar 16
The fall in China's reported holdings of Treasuries in January wasn't offset by a rise in the (suspected) hidden holdings of China (reported in Belgium) or by a rise in Agencies. Visibly Chinese holdings did fall a bit ... but only a bit.

Lots of folks looks at the slide in China's US custodial holdings of Treasuries. Only pros look at China's total reported US portfolio, adjusted for the euroclear (Belgian) account. China has about $1 trillion in USD assets on top of its US custodied Treasury book.

2/3 Image
China almost certainly has USD assets that don't show up in the US data ...

But the broad story is that China's visible USD portfolio is still pretty close to the COFER (global average) USD share

The January dip tho is slightly at odds with the rise in China's fx reserves

3/3 Image
Read 4 tweets
Mar 16
Switzerland (the country) has two things that help here --

1) no shortage of fx liquidity (the SNB has $850b in reserves & access to the Fed's swap lines);

2) and no shortage of fiscal capacity if the government needs to supplement liquidity support with a capital injection
I understand the politics of bailouts (equity capital support imo, not just liquidity support) are toxic everywhere, Switzerland included.

But CS won't fail because of a lack of liquidity (after the SNB's actions). It could fail because of a lack of capital.
The Swiss are as wealthy as the Saudis (if not more so) -- the Swiss public could inject capital if that is what is needed ... the constraints here are self imposed, not fiscal or financial.

Just sayin'
Read 5 tweets
Mar 14
One side effect of an underfinanced program is that the Fund can only move forward with the support of those actually putting up the bulk of the funds ...
China looks to be refinancing if maturing exposure -- but seeking to trade up in credit quality in the process (more loans to the central bank, fewer loans to the government/ guaranteed by the Government)
"The foreign exchange reserves, held by the SBP, stood at $4.3 billion on March 3, 2023, after receiving $1.2 billion re-financing from Chinese commercial loans. Pakistan is expecting another $800 million from Chinese banks within a few days"

(think most of this was to the SBP)
Read 4 tweets
Mar 12
The outline of the US government's response to the failure of Silicon Valley Bank is now clear, and the US didn't mess around. Clearly there was real concern about deposit flight and funding driven contagion.

The response has three legs --

First, the FDIC invoked the systemic risk exception (with the support of Treasury and the Fed) to protect all deposits (tho not the equity or the bonds) of SVB and Signature Bank. This allows the deposit insurance fund to cover any gap between ...

the funds obtained from the sale of SVB's assets and is remaining deposits, and thus makes all deposits whole. The deposit insurance fund can be refilled out of a bank levy -- so there is no cost to taxpayers (this is part of the post Dodd Frank architecture).

Read 13 tweets
Mar 10
The collapse of Silicon Valley Bank highlighted the risks posed by "underwater" long duration bonds -- particularly when those bonds funded with short-term liabilities.

Who else in the global economy holds a lot of underwater bonds?

Asian insurers and policy banks ...

To be sure, the associated risks depend as much or more on an institution's funding structure as on the size of the mark to market loss on bonds in hold to maturity portfolios.

As long as the funding (for SVB, big deposits) are there, losses don't need to be realized.

Japan's insurers probably still have close to $700b in foreign bonds on their books; Taiwan's insurers have a comparable sum. These bonds in long-term portfolios that back promised payouts to holders of life insurance policies --

But they are heavily fx hedged ...

Read 10 tweets
Mar 10
The biggest holders of underwater bonds globally are, of course, central banks -- including the PBOC.

But central banks are different, China's above all.

The January data is finally out and China's central bank looks to have added significantly to its reserves.

Both the PBOC& the state banks added ~ $20b to their foreign assets in January -- so a combined $40b plus. The Chinese have suggested that the rise in PBOC reserves is the runoff of structures set up back in 2008 (to hide intervention then) but that isn't obvious ...

January data is now stale -- the dollar changed course in February and so did the CNY.

But if China really was intervening (to limit CNY appreciation) in January, it would not be a total surprise. The PBOC historically has resisted appreciation pressure.

Read 9 tweets

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!


0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy


3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!