4 ms·

I’m not going to be using this site to guide my opinions of VCs, I don’t generally outsource that much of my opinion-making to 3rd parties. But it does broadly

by fwlr 4y ago

I’m not going to be using this site to guide my opinions of VCs, I don’t generally outsource that much of my opinion-making to 3rd parties. But it does broadly reflect my opinions, you could get a very rough “general direction” picture of how I think from reading it. “It’s rational to join a bank run, VCs would have been negligent to their companies if they advised to stay” I know! Believe it or not I have studied a fair bit of game theory, enough to go beyond “it was a prisoners dilemma” and analyze the SVB collapse as an iterated prisoners dilemma with variable information/reputation and a small number of rounds (in this case the amount of information/reputation broadcasting very quickly drove the whole population to “defect always” and the ecosystem rapidly collapsed). I trust myself to be capable of assigning blame for the bank run correctly (broadly speaking: I. nobody started the run, everyone joined it; II. joining the run earlier is very slightly more blameful than joining the run later; III. being part of the run “loudly” is very significantly more blameful than being part of the run “quietly”). That is maybe 20%, while the other 80% and thus the real albatross here is the demand to be made whole afterwards. Roughly speaking the least blameful is “you’ll get 250k immediately, a substantial fraction soon, and almost all eventually; if this still kills your business let us know and we’ll try to arrange bridging finance”, while the most blameful is “the government must promise to make us whole on Monday morning or the contagion will spread; use taxpayer money if you have to, this is that important”.

lynx23 4y ago
You're using the word "defect" to describe someone pulling their money from a bank? Seriously? It is so easy to assign blame apparently...
- flandish 4y ago
  I agree. This all reads like a solid reminder that two things happen in that realm of highly VC funded systems: - Investment/money management often happens by people who are very new to this process and make swift sometimes stupid decisions. - Everyone else is clearly in the “protect profit” game, even if that means ignoring the clearly stupid things svb (and really the entire system) did/does. Mort backed securities, crap from the 08 markets, deregulation.. etc. A failing bank should not crash the economy nor should a bunch of startups failing as a result crash too. Poor startup decisions are one thing. But “too big to fail” things maybe need to be smaller or nationalized? I realize the site I’m on, and know I am still learning about this. I have just, in all my (short 43) years, felt the “too big to fail” reason pop up more than it seems it should because, to me, it means “gov must protect my profit, or I’ll deadman switch the economy.”
- fwlr 4y ago
  “Cooperate” and “defect” are technical terms in game theory, they don’t carry their usual connotations here.
shubb 4y ago
I worry that... given how much high risk free rates must be hurting VCs, it was rational for them to crash a bank to force the fed to ease policy. I'm not saying they deliberately did that. I'm just saying it would have been a smart reason to do the thing that they did. The VCs will benefit hugely from the consequences of these events, forget the compensation. This is what they call moral hazard.
yafbum 4y ago
You forget to consider, in your blame list, the people who created the unstable situation in the first place. If I put a pencil on a table, balanced on its tip, can we find which puff of wind to blame when it falls? The bank's leadership, for example, by choosing to focus heavily on one industry known for rash herd behaviors, both benefited from it some (lots of VCs recommending them) and now seems to have been undone by it. The bank's regulators might have also looked the other way, we'll find out more.
MacroChip 4y ago
It wasn't a prisoner's dilemma. Cooperating gets you almost nothing while defecting gives you benefit. It was nearly the opposite of the prisoner's dilemma. Couldn't another school of thought "blame" those who didn't participate loudly?
- fwlr 4y ago
  “Cooperating gets you almost nothing while defecting gives you benefit” That is the prisoners dilemma. What you have offered here is in fact an accurate statement of the payoff matrix for the prisoners dilemma in the case where the other prisoner cooperates.
  - hnfong 4y ago
    In prisoner's dilemma cooperating gives both parties the most benefit in aggregate, while defecting gives on party additional benefit at the expense of a steep disadvantage for the other party. Also note that if all parties defect they all end up in a rather undesirable state. For the SVB bank run, the main loser is the bank, who isn't party to this game (of whether to keep deposits or not). Even if everyone defected, the early defectors don't lose anything. And arguably, it can be foreseen that even late defectors don't lose anything either due to SVB marginally in the "too big to fail" category, or at least, the startup ecosystem being "too important to fail".

by fwlr 4y ago

lynx23 4y ago

You're using the word "defect" to describe someone pulling their money from a bank? Seriously? It is so easy to assign blame apparently...

flandish 4y ago
I agree. This all reads like a solid reminder that two things happen in that realm of highly VC funded systems: - Investment/money management often happens by people who are very new to this process and make swift sometimes stupid decisions. - Everyone else is clearly in the “protect profit” game, even if that means ignoring the clearly stupid things svb (and really the entire system) did/does. Mort backed securities, crap from the 08 markets, deregulation.. etc. A failing bank should not crash the economy nor should a bunch of startups failing as a result crash too. Poor startup decisions are one thing. But “too big to fail” things maybe need to be smaller or nationalized? I realize the site I’m on, and know I am still learning about this. I have just, in all my (short 43) years, felt the “too big to fail” reason pop up more than it seems it should because, to me, it means “gov must protect my profit, or I’ll deadman switch the economy.”
fwlr 4y ago
“Cooperate” and “defect” are technical terms in game theory, they don’t carry their usual connotations here.

shubb 4y ago

I worry that... given how much high risk free rates must be hurting VCs, it was rational for them to crash a bank to force the fed to ease policy. I'm not saying they deliberately did that. I'm just saying it would have been a smart reason to do the thing that they did. The VCs will benefit hugely from the consequences of these events, forget the compensation. This is what they call moral hazard.

yafbum 4y ago

You forget to consider, in your blame list, the people who created the unstable situation in the first place. If I put a pencil on a table, balanced on its tip, can we find which puff of wind to blame when it falls? The bank's leadership, for example, by choosing to focus heavily on one industry known for rash herd behaviors, both benefited from it some (lots of VCs recommending them) and now seems to have been undone by it. The bank's regulators might have also looked the other way, we'll find out more.

MacroChip 4y ago

It wasn't a prisoner's dilemma. Cooperating gets you almost nothing while defecting gives you benefit. It was nearly the opposite of the prisoner's dilemma. Couldn't another school of thought "blame" those who didn't participate loudly?

fwlr 4y ago
“Cooperating gets you almost nothing while defecting gives you benefit” That is the prisoners dilemma. What you have offered here is in fact an accurate statement of the payoff matrix for the prisoners dilemma in the case where the other prisoner cooperates.
- hnfong 4y ago
  In prisoner's dilemma cooperating gives both parties the most benefit in aggregate, while defecting gives on party additional benefit at the expense of a steep disadvantage for the other party. Also note that if all parties defect they all end up in a rather undesirable state. For the SVB bank run, the main loser is the bank, who isn't party to this game (of whether to keep deposits or not). Even if everyone defected, the early defectors don't lose anything. And arguably, it can be foreseen that even late defectors don't lose anything either due to SVB marginally in the "too big to fail" category, or at least, the startup ecosystem being "too important to fail".