It is a bit of a truism, possibly wrong, that disasters are very good learning experiences. As such, it behooves us to consider the case of Silicon Valley Bank (SVB). It was the 17th largest bank in the US before it went spectacularly bust in a very very short time. Normally in financial markets, the dynamic is that of a slow-motion train crash. This was a train crash.
So, what went wrong? These genius’ bought long dated bonds with yields at multi-generational lows, at the low point in the interest rate cycle. And they were double smart too, because instead of being dull and buying T-bonds, hell anyone could think of that, they bought agency debt. Similar credit, better yield. A few basis points at the very least. And what was the root cause of this generous funding of US mortgages? Basically they got too much money on deposit and the spread between what they paid on the deposits and what they would have received on t-bills was just too small. I mean a 25bp spread would not even give them revenue of half a billion dollars. Close to half a billion obviously, but still not great.
Now most of us, when punting large on long dated interest rates are vaguely concerned about the L part of the P/L. Not SVB. These super smart folk were the knowing beneficiaries of what can only be described as magic. Look away those of you in the Funds Management arena because this magic is not available to you, ok, maybe PE, but that is not the point. In banking there is a particular kind of magic whereby if you own a long-dated agency bond, and the price of the bond has declined, you have a choice. You can recognise the loss, possibly causing your removal from your executive position due to your lousy stewardship of other peoples money, or… , you can shake your magic 8-ball, and if the stars are correctly aligned (it is magic after all), you can forget about the loss, announce the magic words – “I am holding this bond to maturitycadabra” and suddenly the losses disappear. OK, I did make up the “cadabra” bit, but incredibly the rest is true. So, the double genius’ with magic effectively can have their cake and eat it. What could possibly go wrong?
It is reassuring that in reality magic is similar to that described in fantasy novels, i.e. there are normally conditions associated with it. In this case, if a depositor wanted their money back, then you may have to sell the magic bond, but sadly at the market price, thereby recognising the loss. If you had too many depositors looking for their money back, you may have to sell lots of bonds at a loss, overwhelming your available capital and the bank goes bust. I am not going to go into the banking as a confidence trick thesis, as that is record is old, however the magic above has a certain smell of it. And in truth, possibly for good reason. If you do not have to disclose your short-term losses, your depositors will not get worried, so they will not withdraw their funds, and when the bonds mature, everything is hunky dory. No pesky bank runs etc which is generally perceived to be a good thing. Everyone is a winner.
So, what we are looking at to keep the show on the road is simply described persistent information asymmetry. So long as the depositors do not know what the bank knows, nothing to worry about. Except of course if your depositors have possibly the greatest access to data in the history of the planet, and they , as it happens, are exceptionally good at accessing data and retrieving information, it is sort of what they do, then possibly worry. In which case, all those finance majors who went to tech finally make good. They can work out that if SVB has limited capital and large losses, then who withdraws first, wins. Oh, and then you tell all your mates, ‘cause it is cool to be first!
Hence train crash, not slow-motion. I would love to say that there are some lessons to be learnt here, but these folk were just plain dumb. Well paid however. Perhaps that is the lesson, perhaps not.
