Not a banking regulation expert so going to mishmash my wording due to the lack of mark to market that the bank did not have the recognize the loses on the loans until they sold the treasuries. As a result, the capital reserve lose was recognized all at once when they sold the treasuries vs GSIBs who have had to continual mark their books to market resulting in hedging and managing their capital reserve. You see this with JPM who came out months ago saying they would not touch long date treasuries with a 10 foot pole because it was clear that interest rates would rise causing loses on these investments having to be realized.
The regional banks did not have to realize these loses because of mark to market rules and were forced to sell because deposit draw down was accelerating. Because SVB deposit base was shrinking so fast (mostly VC money being lite on fire by tech) and lack of new funding from VCs into the ecosystem, SVB had to shore up its capital reserve. The only way to do that was sell its loan book. If SVB had to mark to market, they would had to months ago change their capital reserve sooner and not all at once.
Arguably if they had to mark to market and loses were born over time instead all at once, you would not have had the mass panic causing the capital flight.
Lots of mistakes were made including how GS handled the equity raise which was comically bad. That does not change the situation where mark to market could have prevented the situation.
I don't have the specifics on liquidity but understood that is related but not wholly dependent on mark to market on loan portfolio to test ability to sell assets on reasonable prices to test if a bank can handle reasonably extreme market movements.