CALIFORNIA FIRST LEASING CORP (CFNB) stock intrinsic value fair value estimate & owner earnings analysis

California First Leasing is no longer really a leasing company. After winding down its national bank following 2017, it registered in 2022 as an internally managed closed-end investment company, and today almost everything it owns is a portfolio of US stocks. At June 30, 2026 total investments were $521M against a cost of $200M; about $50M was in money-market funds, and the leasing book had shrunk to $1.9M (under 0.5% of assets). It has no borrowings. The company is a regular taxable corporation, not a pass-through fund, so it carries a $91M deferred tax liability on its unrealized gains, and its reported net assets of $432.8M, or $2,410 per share on 179,568 shares, are already after that tax. What changed since our previous analysis is the portfolio, not the business model. Net asset value per share rose 68.6% in fiscal 2026, from $1,429 to $2,410, because the holdings are heavily concentrated: five semiconductor companies make up about 55% of net assets, with Micron at 19%, Applied Materials 15%, Marvell 9% and AMD 7%, alongside Alphabet, Goldman Sachs and Exxon Mobil. The earnings of a fund like this are its investment returns, so the per-share figures we track (about $85, $177, $109 and $981 in fiscal 2023 through 2026) say little about a steady earning level. Recurring income is small: about $4.6M of dividends, interest and lease income against $2.5M of expenses, so the fund roughly covers its costs and the owners' return is whatever the stocks do. Management is paid fixed salaries totaling about $1.35M with no incentive plan, which keeps costs low and avoids performance fees. We value it on what a minority owner can actually receive, not on the portfolio alone. There are two adjustments. First, the deferred tax is only paid when gains are realized. Portfolio turnover has been 7% to 24% a year, so much of that $507 per share of tax is years away, and in today's money it is worth perhaps $230 per share less than its face value. That puts economic value near $2,640. Second, outside shareholders have no say over when they get their money. The CEO, Patrick Paddon, holds 60%, and directors and officers together hold 92%. Dividends are rare (the only one in four years was $20 in fiscal 2024), and the main way cash has come back is tender offers priced at the market rather than at asset value. The fiscal 2025 buyback of 330,000 pre-split shares averaged about $18.72 against a $28.59 net asset value, roughly 35% below it. The shares have traded 25% to 37% below net asset value at every year-end we can see. Base case ($1,850): the portfolio earns roughly a broad equity return, with no assumed continuation of the semiconductor run, and the deferred-tax benefit above holds. A minority holder realizes value at about a 30% discount, in line with the company's own record of buybacks and market prices. Bear case ($900): a semiconductor downturn of the depth the industry had in 2008, when chip stocks fell more than half. We cut the semiconductor holdings 55% and the other stocks 35%, cash holds, and taxes absorb 28% of the loss. That takes net asset value to about $1,560, and with the discount widening toward 45%, a minority holder could realize about $900. Bull case ($2,250): the portfolio holds its value and the discount narrows to about 15%. The evidenced route is the company's tender offers, with odd-lot holders getting priority after the reverse split, or a liquidation. Nothing here assumes the semiconductor run continues. In none of the cases is there debt to repay; all value is equity. The structure carries risk that no case fully models. The reverse split, approved in February 2026 and done in June 2026, aims to cut the holder count below 100 so the company can leave the Investment Company Act. If it does, the protections and filings that come with registration may end. The company has said it plans to keep reporting, including its holdings, only while the stock trades on the OTC market. The float is a few percent of the shares, and the stock is very thinly traded. Net asset value moves daily with a handful of chip stocks, so the June 30 figure is a snapshot, not an earning power. Margin-of-safety price. We are confident about what the company owns, since the holdings are listed, audited and marked to market. We are not confident about what those holdings will be worth, because more than half is one cyclical industry after a 68% year. The worst plausible outcome is a chip-stock collapse while the controlling family stops tendering and deregisters, leaving a minority holder with shares worth well under half of asset value and no catalyst. The company would survive this easily, since it has no debt and its assets are liquid, but the minority holder would have to wait. The things we are most likely to be wrong about are how far the semiconductor holdings can fall and whether the discount keeps widening. If both go the wrong way, the shares are worth about $900. During the holding period the business pays essentially nothing, so the whole return has to come from asset value growth plus the discount narrowing at a tender. The margin-of-safety price is $1,100: from that price, an exit in about three years at the company's usual 30% discount, with assets growing at a normal equity rate, implies a bit over 20% a year. The bear case implies a loss of about 18% from that price, with hard, liquid assets and no debt behind the shares. Above that price, the semiconductor concentration and the lack of any say over the exit leave too little room.