capexus @capexus

@Hardov__ Paying in stock turned a $433 million mistake into $5.7 billion of dilution.

Buffett bought Dexter Shoe in 1993 for $433 million. He paid in Berkshire stock. The price kept rising for 21 years. 1993. 25,203 Berkshire shares go to the sellers. 2007. Dexter is worthless. The shares, $3.5 billion. 2014. The same shares, about $5.7 billion. Cash is paid once and ends there. Stock hands the seller a slice of every future gain. Buffett gave away 1.6% of Berkshire for a shoe company. A cash price is fixed the day you sign. A stock price is set by everything you build afterward.

Hardov @Hardov__

9 hours ago

x.com/i/article/2111…

2 0 8 454 7

@mojunel The $31,980 never paid for the brief, it pays for data and chat.

@kobaHUB Solar modules got 90% cheaper, and Suntech still went bankrupt in 2013.

India's banks added Rs 25.3 lakh crore of personal loans in four years. Credit to industry grew by Rs 10.0 lakh crore. 2013. Industry held 45.8% of non-food bank credit. 2025. About 22%. Manufacturing slid from 17.3% of GDP in 2006 to 13.1%. The rulebook explains the drift. A home loan carries a 20% to 40% risk weight. A loan to an unrated factory carries 100%. The same capital backs 2.5 to 7 times more mortgages. Banks lend where the rules say risk is cheap. A country ends up owning what its capital rules prefer.

🇦🇺Craig Tindale @ctindale

6 days ago

x.com/i/article/2109…

27 58 243 24K 259

@Hardov__ The $15.2 billion loss sat there for months. Depositors only picked the date.

@ridark_eth Paper fills skip slippage, so the backtest pays prices nobody would actually get.

@mojunel The 58% drop after publication is the real number. Printed edges get arbitraged away.

SVB's CEO sold $3.6 million of shares on February 27, 2023. Eleven days later the bank was gone, and so was the stock. The sale ran under a trading plan adopted on January 26. It went through at about $287 a share. Legal and common, and it still became part of the story. March 9. The shares fell 60%, to $106.04. March 10. Regulators closed the bank that morning. The same day, per the Fed's review, staff got cash bonuses. March 17. The holding company filed for bankruptcy. Bonuses follow the calendar. Shareholders follow the balance sheet.

capexus @capexus

5 days ago

x.com/i/article/2109…

2 0 27 5K 17

@Hardov__ In 2011 rare earth prices multiplied, since nearly all refining sat in one country.

The Fed's own review of SVB reads like a coroner's report. Its worst entries date from the months rates were rising. March 2022. Rates start climbing and the hedges start coming off. July 2022. The last of them is removed. April 2022. The rate model is changed and the risk looks smaller. The same month the Chief Risk Officer leaves. The seat stays empty for about eight months. At failure there were 31 open findings, about three times its peers. Rising rates were weather that every bank faced. Taking the hedges off was a choice made inside.

capexus @capexus

5 days ago

x.com/i/article/2109…

2 0 27 5K 17

@kobaHUB Railways in 1845 and fiber in 2000 were also built years before the revenue.

The Fed started lending against bonds at par on March 12, 2023. A bond the market priced near $83 counted as $100. That same Sunday every SVB depositor was covered, insured or not. From Monday they had full access to all of it. Shareholders were not protected and lost everything. Covering the uninsured at SVB and Signature cost $16.7 billion. A special assessment on large banks pays for it. First Citizens bought most of SVB at a $16.5 billion discount. The bond never defaulted. Its price broke, and other banks paid the bill.

capexus @capexus

5 days ago

x.com/i/article/2109…

2 0 27 5K 17

The 10-year Treasury closed 2021 at 1.52%. Silicon Valley Bank went out ten years and more for 1.58%. Short-term paper paid close to nothing. So the bank reached along the curve for a little more yield. Of its $91.3 billion book, $86.0 billion was due after ten years. Average duration was 6.2 years. Each one point rise in rates took roughly 6% off the price. On $91 billion, that is more than $5 billion per point. Then the Fed hiked eight times in under a year. One test before you reach for yield. What does a single point of rates cost you?

capexus @capexus

5 days ago

x.com/i/article/2109…

2 0 27 5K 17

Silicon Valley Bank filed most of its bonds as held to maturity. That label let a $15.2 billion loss live in the footnotes. Bonds under that label are carried at cost. The rule assumes the bank collects every payment and never sells. While that held, the market price stayed out of capital. End of 2022: cost $91.3 billion, fair value $76.2 billion. A year earlier the same gap was $1.3 billion. Equity at that date: $16.0 billion. A label can change where a loss is written. It cannot change who pays when the cash runs out.

capexus @capexus

5 days ago

x.com/i/article/2109…

2 0 27 5K 17

Silicon Valley Bank tripled its deposits in two years. The growth itself is where the failure started. End of 2019: $61.8 billion of deposits. End of 2021: $189.2 billion, much of it startup cash. No bank can triple its lending that fast and stay careful. So the cash went into bonds instead. The held-to-maturity book took $6.8 billion in 2020. In 2021 it took $85.5 billion, $20.4 billion in one quarter. By 2022 securities were 55% of assets, against 25% at peers. A deposit is a promise to pay on demand. A ten-year bond is a promise to wait.

capexus @capexus

5 days ago

x.com/i/article/2109…

2 0 27 5K 17

SVB sold about $21 billion of bonds on March 8, 2023. The math was sound. The message ended the bank. The sale booked a $1.8 billion loss after tax. It planned $2.25 billion of new capital. Management saw the loss earned back in about three years. Depositors heard three things at once. Cash was short, the bonds were worth less, and far more of them remained. Moody's cut the rating one notch that same evening. A balance sheet can be repaired over three years. A depositor can leave in a single day. Full breakdown below.

capexus @capexus

5 days ago

x.com/i/article/2109…

2 0 27 5K 17

@Hardov__ Silicon Valley Bank held safe Treasuries in 2023 and still failed on the marks.

Washington Mutual lost $19 billion of deposits over 16 days. Silicon Valley Bank lost $42 billion in a single Thursday. March 9, 2023. The shares fell 60% to $106.04. Founders Fund advised its startups to move their money out. It left through online banking, not teller windows. By the close the bank was about $958 million short. Over $100 billion more was queued for the next morning. Together that was about 85% of all deposits. WaMu's run in 2008 took sixteen days. SVB's took one business day. Full breakdown below.

capexus @capexus

5 days ago

x.com/i/article/2109…

2 0 27 5K 17

Silicon Valley Bank's bonds paid every coupon on time. The bank still lost $42 billion of deposits in one day. 2021. It put $85.5 billion into government-backed bonds at about 1.6%. By the end of 2022 every $100 of them was worth $83. That hidden loss equaled 95% of the bank's capital. 94% of its deposits were uninsured. On March 9, 2023 they left together. WaMu lost $19 billion over 16 days in 2008. A bond can keep every promise and still sink its owner. It only takes one morning when the money is needed back.

capexus @capexus

5 days ago

x.com/i/article/2109…

2 0 27 5K 17