Why do banks need special assistance in that situation? If you depend on your input prices not going up by that much for a short period, you're hosed anyway...
[1] 1.09^(1/365) - 1 vs 1.01^(1/365) - 1
As a rates trader, your positions are denominated in the millions of dollars. Some back of the envelope math:
1mm of 10y treasuries (repo'd at EOD) is roughly $1k of risk (dv01 -- if rates move by 1bp, your pnl fluctuates by $1k). To finance that position, you're paying $236 at 9% vs $27 at 1%.
Now say you've got 1mm of 2y treasuries. That's roughly $0.2k dv01. Unless 2y rates move 1.25bp overnight, you're losing money just financing the position.
In short, there's a massive difference between paying 2.3bps and paying 0.27bps.
No, think of it as paying $236,000 for something that normally costs $27,000.
You can think of this business as providing banking services to large corporations, who aren't able to simply put money in a savings account at their local credit union.
It's doesn't collapse in the short run, but it can harm the ability of these corporations to access their funds. Short term financing disruptions can have major impacts across all markets.
Show me a business where a sudden 10x increase in costs isn't painful...
While dealing with assets in the billions and lending out at higher interest all the time.
>Show me a business where a sudden 10x increase in costs isn't painful...
All the ones where one input went up 10x temporarily and survived, or just the most characteristic examples?
Yes, those costs were computed on a book long 1 billion dollars of notional. There's a difference between notional and risk though. 1 billion dollars of 10y treasuries will be ~1mm dv01. Both matter.
> lending out at higher interest all the time
This isn't how rates trading works. Entering positions costs money. You're not making loans. If you hold treasuries, you're earning the coupon on them for the period that you own them. It gets more complicated when you factor in repo, but typically you pay GC repo when you're long.
This article is about repurchase agreements (repos). They are a form of collateralized loan used in rates trading. If you need to borrow money for a short period, you can do so in the repo market. You sell treasuries to a counterparty, agreeing to buy them back at a later date. You agree to pay an interest rate called a repo rate for this transaction. This is the rate the article is discussing.
Someone borrowing through the repo market is doing it to satisfy liquidity needs. They need the liquidity for some ongoing concern or investment or trade, or something (I don't know the details or relevant term, but it doesn't matter for the point).
That venture has some ROI. The temporarily higher repo rates have to be compared to that. That venture's costs are temporarily going from .002 percent to .02 percent of capital invested. Either way, a small portion. I don't see the emergency beyond "I wish this made the higher profit I am accustomed to".
The difference between rates markets and, say, equities markets is that you think of your risk in basis points and not percentage points. A 10bp change in equities prices on a $1 billion book will cause $1mm change in pnl. However the $1mm dv01 treasury book (with notional $1 billion) will make or lose $10mm with a 10bp change in interest rates. You can see why basis point moves can be much more meaningful in the rates space.
Let's consider market makers now. They are mandated to make money by provisioning liquidity, not by taking views. This means that their return is supposed to come from earning spread and not from being positioned the right way on market moves. This often requires holding and financing hedged positions overnight. If your financing costs are >2bp, there's a pretty solid chance you're going to lose money on the positions you're holding, considering that you hedged them specifically to try to minimize fluctuations. It's not just a matter of making less profit as much as it is one of losing money entirely. Were such increases in financing costs to become prolonged or more frequent, there would be a deleterious effect on rates markets in general.
The answer for finance is just the same as for computer science, you are doing things many times with big volumes, so tiny differences ass up.