Hacker News Reader

Hacker News

Comment by Dwolb | Hacker News Reader

Parent Full thread

Dwolb

There’s interest rate risk, credit risk, and prepayment risk with the securities they buy.

On a Treasury or guaranteed bond, there is no credit risk.

On a Treasury there is no prepayment risk.

Therefore if you hedge out the interest rate risk, you’re essentially left with 0 risk.

0 risk = 0 or near 0 premium.

e.g. there’s no point in doing the trade if you hedge.

View on HN

Yes that’s the point. Treasuries are cash, they are not meant to make money for banks, they are meant to be a place for banks to put money when they don’t have anything else to do with it. Banks are supposed to make money from the premium between the base interest rate and the rate on the loans they make. The implied contract when you deposit money in a bank is that the bank has a dependable business model as a lender. They are not supposed to be gambling.

SVB got greedy. They thought that zirp would continue forever and they made long term bets on that basis and lost. Banks should not be making long bets on macroeconomic conditions: that’s not a bank, that’s a hedge fund.

> Treasuries are cash,

Obviously not.

[edit to add: this seems a controversial comment, it’s being upvoted and downvoted wildly]

Some brokerages break down various types of position (equity, bond, etc). Mine in particular has a "Cash and equivalents" section. Money market accounts go there, but so do treasury funds and such.

I suspect the upvotes are because you're technically correct, treasuries are not literally cash. The downvotes are because grandparent obviously meant that treasuries are considered cash equivalents (whether or not they actually should be, given liquidity concerns).

Treasuries (without further qualification) are not cash equivalents. (And it’s not because of liquidity concerns.)

Treasury bills may be cash equivalents. They present (almost) no interest risk.

Long-term treasury notes/bonds is what was being discussed. There is interest risk in that case and neither your broker nor anyone else would consider them cash equivalents (unless they are already close to maturity).

The trouble is that other US government interventions also eliminated a big chunk of the "anything else to do with it". In particular, most mortgages in the US are fixed rate for the entire duration of their term through government backing, which almost entirely eliminates one of the big sectors of loans that banks in other countries can use to make money on the premium from interest payments with less duration mismatch than 10-year or 30-year bonds. In most other countries, mortgages are either variable rate or only fixed rate for a relatively short period. Also, a substantial proportion of all US mortgages in existence apparently locked in their interest rates during the time period in 2020 and 2021 when they were at record lows.

I'm not as familiar with how mortgage markets work in other countries, but at least in the US fixed rate mortgages are almost always risk balanced against fixed rate investmentsor wrapped up in securities that move risk off the banks' books.

A bank would indeed be crazy to hold onto fixed rate 30-year loans with interest rates at near zero.

Yeah, in the US fixed rate mortgages are handed off to the government who guarantees them and bundles them up into securities... that SVB then ended up buying a bunch of because it's not like there was exactly a wide variety of investment options available that they could back their deposits with that actually paid meaningful interest.

At least based on this article, it even sounds like this is what they wanted banks to do! The reverse-repo market was created because the Fed wanted banks to buy securities directly from them when interest rates where increased, sounds like SVB and others just followed that playbook

Maturity transformation (borrowing short and lending long) is still a core function of banking (and historically one of the key sources of profit. Manhunt this risk in a bank portfolio is one of the central jobs of management.

The naïveté was thinking the deposit base wouldn’t shrink (after growing 3x in as many years)

You don't need to hedge out all the risk. You can for instance hedge the risk that rates go up more than one percent and take the risk when it changes less, in return for a moderate yield. This way you limit your losses when things go wrong and probably you'll avoid bankruptcy.

Besides that, financial institutions don't only make money on re-investing deposits. They charge all kinds of fees as well, which make for a steady income.

Oh you can get a lot more creative than that. You can hedge out "catastrophic" interest rate rises while taking your chances on smaller increases. In this case, catastrophic can easily be estimated by taking the duration of the book and the equity cushion and applying a margin of safety. Or you can simply ladder Treasury durations that inherently don't have enough duration to cause problems. SVB apparently did none of that.

There is exactly one investment strategy that presents no risk of ruin for a bank: depositing all customer deposits in the bank's account at the fed. That investment strategy is so discouraged by regulators that they literally denied the application for a charter from an organization whose stated investment strategy was exactly that. (google the narrow bank if this is unfamiliar to you).

For any other investment of any amount of your deposits, a run of sufficient size combined with an investment loss of sufficient size will ruin a bank.

That being said, SVB had by a wide margin the flightiest deposits and the most IR risk. But it was an error of magnitude, not category. There is no run-proof fractional reserve bank.

There is always a point as the risk is not binary. SVB should have hedged against big interest jump that would bankrupt them, not against any risk.

Would pre-2018 regulations (when they were eased for banks with deposits less than $250B) have prevented this?

I think so. I believe that this is the reason EU banks have been ok so far, as they do need to hedge interest rate risk.

EU banks are ok? Didn't Credite Suisse just go under and had to be forcedly saved by its main competitor?

Credite Suisse's problems predate the interest rate issues by years.

I have no idea about the substantive discussion, but just to point out: Switzerland is not in the EU.

Credit Suisse is Swiss. Switzerland is not a member of EU and they don't use euro.

I appreciate the correction but that probably doesn't change the point being made, as the economy of Switzerland is obviously closely tied to that of the EU (when I read EU I normally think Europe, not just European Union, which is my mistake, but in conversations like this I think "Europe" would be the better term after all).

Yeah, I almost added CS to the original post, but their problems were relatively unique to them, rather than being an extreme case of a common problem, like regional banks in the US (who didn't need to hedge interest rate risk).

Aside from where the bank is chartered...

I would suggest a read of https://en.wikipedia.org/wiki/Credit_Suisse#Controversies and consider that each time that happened they lost money and trust.

At this point, they have neither.

If you bought treasuries at the prevailing rate and swapped them at the same rate then the market price of the swap will be zero. You'll still have to pay credit and funding premia though. If, however, you only hedge once the prevailing rate moves against you, you'll pay a market premium in the form of a spread.

Interest isn't just compensation for risk. It's also compensation for temporarily giving up control of your money.

Reply on news.ycombinator.com