A quick run through Silicon Valley Bank’s risk management snafus

Rick Bookstaber is the co-founder of Fabric, a risk management consultancy. Before that he was head of risk management at the University of California, Bridgewater, Salomon Brothers and Morgan Stanley.

With the benefit of hindsight, there is hardly a risk management failure — whether of a bank or a hedge fund — that can’t be explained in a few paragraphs, however complex and opaque it might seem.

And it always comes down to a few points: too much risk, too little diversification, and playing fast and loose with the rules (whether that be regulation, governance, or the pricing of assets).

The Silicon Valley Bank crisis is still unfolding. Despite the government’s efforts to draw a line under the mess we aren’t quite yet in postmortem mode. But let’s see the word count needed to assess the risk management failures from our current vantage point.

1. Risk

SVB loaded up on technology start-up clients. Technology is the most volatile sector in the equity market, and start-ups are the most risky enterprises.

2. Diversification

Diversification comes from holding a broad bunch of assets, varied by region, industry and strategy. When one part of the portfolio is down, there is a chance another will counterbalance that by being up. No luck there for SVB. Their clients were extremely concentrated in one region (more like, one city), and in one industry.

3. Hedging

For a bank, unwanted risk number one is rates. Banks naturally get rate risk on their balance sheet, because it is filled with bonds. So you hedge that, usually with swaps. SVB did that for a while, but appears to have taken them off. Not a good idea generally, and for sure not a good idea in a rising rate environment.

4. Marking to market

You want to know what the current value of your portfolio is. So you reprice your assets based on their current market price. But SVB didn’t have to re-mark their Treasuries unless they sold them. So the balance sheet impact was not known until the tremors of the crisis hit. Also not marked to market: the valuation of the companies in their VC clients’ portfolios. This is hard to do, but you can do better than zero.

5. Risk management oversight

If you are a $250bn bank — well, almost $250bn, and I’ll get to that in a bit — you should have professional risk management. SVB had one, but only until April 2022. They hired a new one in January, too late to have helped them out of the mess.

6. A new lesson: the importance of the network.

Not part of the usual risk discussion, but SVB clients all had pretty much one degree of separation, sharing the same central nodes. To wit: Peter Thiel’s Founders Fund and other high-profile venture capital firms advised their companies to pull money from the bank. It’s like having all of your power lines on one master switch.

Bank runs always come down to crowd behaviour. For example, in It’s a Wonderful Life just one person, Jimmy Stewart’s George Bailey, turned the crowd around. Here, SVP was set up to have one person — maybe a few — wave the crowd on.

7. A George Santos problem.

The fabulist congressman posted 40 expenses for his campaign at $199. This is a magic number, because the election committee only has to keep receipts for expenses that exceed $200.

The magic number for banks is $250bn. If your assets stay under that number, you skirt an increased layer of regulatory scrutiny. SVB didn’t stay at $249bn, but pretty close. It was flying $20-30bn under the radar, and CEO Greg Becker lobbied to have thresholds raised.

Granted, no one wants the overhead of regulation. Regulation that, among many other things, checks on the risk of your balance sheet.