[adriancr]

They bought 10 year bonds at 1.5% yearly. For every 100$ they will get 116$ at maturity.

Right now there are 10 year bonds at 4% that will pay 148$ at maturity.

To be able to sell your 1.5% bonds right now you need to discount them sufficiently so they have the same value as the new 4% 10 year bonds. (otherwise why would anyone buy them)

I'd guess you'd need to discount 148$ - 116$ = 32$.

This means selling your 100$ bonds at 68$ right now to have buyers... Otherwise money is stuck for 10 years which is unfortunate if you ran out of available cash.

Is this wrong?

[loeg]

I believe I read that the average maturity of their holdings are ~6 years out of the full ten. It looks like T-notes with similar maturity are yielding around 3.9-4.0%.

> Right now there are 10 year bonds at 4% that will pay 148$ at maturity.

How are you calculating that? My impression is:

> Notes and bonds are issued to pay a fixed rate of interest called the coupon rate. A $10,000 treasury note with a seven percent coupon rate pays an investor $700 per year interest in two semi-annual payments of $350 each. The interest from notes and bonds paid out to investors is simple and does not compound

Over a 10-year duration, I think that 4% bond would pay $140 on $100. 6-year to maturity notes at 3.9% would pay, I believe, $123 on $100 today; and at 1.56%, $109.

I think you'd value the 1.56% notes by something like the ratio of the two values at maturity? About 89% of what you'd pay for a 3.9% note. ($100 / 0.886 => $112.87; $112.87 * 1.0936 => about $123.)

(I don't work in this sector and I might be mathing it wrong.)

[skippyboxedhero]

They have also been marked down already, that is what the big unrealised losses are about.

I am not 100% sure what the accounting rules for htm vs afs are anymore. I believe htm allows you to amortize losses over the term of the loan (which is, of course, still as controversial as it was in 2008). But SVB has already taken fairly substantial markdowns already on securities that were transferred into htm after they dropped significantly.

And the purpose of receivership is to preserve value for depositors. So the problem is that the losses have absorbed the firm's capital, not that other sources of funding have taken losses. A book of MBS is not going to be trading at a 30% discount to the mark a few weeks ago when their financial period ended. All of this stuff is liquid, unless their corporate lending was awful (unlikely) then there won't be a massive discount.

Btw, this did happen last year in the UK. The BoE essentially left the market to sort out problems caused by higher rates/falling bond prices, and hedge funds absolutely rinsed pension funds. Some made hundreds of millions in a few hours. This won't happen in this case because FDIC has stepped in and is running a proper auction.

[panarky]

The rule of thumb is every 100 bp increase in rates means a reduction in the market value of the security equal its years to maturity as a percentage.

So if rates are up 250 bp and there are 9 years remaining to maturity, that would be a 2.5 * 9% = 22.5% reduction in market value.

But I believe current yields on 10-year MBS are greater than 4%, the numbers I've seen put them at about 110 bp over 10-year Treasurys, which would make the reduction in market value even deeper.

[kgwgk]

> But I believe current yields on 10-year MBS are greater than 4%, the numbers I've seen put them at about 110 bp over 10-year Treasurys, which would make the reduction in market value even deeper.

Presumably they were yielding more than treasuries when they bought them as well. The relevant thing is whether the spread has narrowed or widened (too lazy to check and too coward to guess...).

[loeg]

I don't think 10-year T-notes are up 250 bps from 1.56% (might be mistaken -- looks like 235 bps to me), though 10-year MBS might be, and my impression is that SVB's average maturity is more like 6 years than 9. Those would both soften the impact on market value.