The 0.36 P/B looks optically cheap against Guild at 0.99 and PennyMac at 0.80, but because tangible book per share is negative, the discount is not evidence of a mispriced franchise — it more plausibly reflects the market pricing in some probability that the remaining ~$201M of book value, largely intangible, gets written down. Investors should note the P/E of -2.19 and the 0.18 P/S offer no cross-check either, so the equity book itself becomes the only reference point, and that reference point is being eroded each quarter.

In a normal bank, a 5% equity-to-assets ratio would be a tailwind to ROE through leverage, but here it acts as an amplifier of losses because the spread earned on the balance sheet is negative before any operating expense. The 87.8% fee-income share means reported profitability is almost entirely a function of origination volume and gain-on-sale margins, so ROE quality is closer to a brokerage cycle than a spread business, and the one profitable quarter in the series (0.2% in 2024Q3) coincided with a 19.7% efficiency ratio rather than any improvement in spread economics.

The improvement from 102.7% in 2024Q1 to 76.7% in 2025Q4 demonstrated genuine operating leverage during a recovery in volume, but the swing back to 88.9% suggests the cost base remains difficult to shrink in line with revenue declines, which is typical of a workforce-heavy origination model. Because fee income fell from $366.9M in 2025Q4 to $284.4M in 2026Q2 while the efficiency ratio rose 12 points, investors should monitor whether management can hold expenses flat through the next volume trough; the persistently negative NIM means no amount of spread recovery will rescue the ratio.

With roughly $203M of accumulated net losses across the window and only one profitable quarter, retained earnings cannot replenish the buffer, so book value is being sustained by the pace of intangibles amortization and any reserve releases rather than by earnings power. The dividend yield, which fell to nothing in 2026Q1 and 2026Q2 after a 0.9% reading in 2025Q1, is consistent with capital preservation becoming the governing priority; regulatory capital ratios are not disclosed in the supplied data, so the distance to minimums warrants further investigation.

Guild and PennyMac generate positive ROE while funding a comparable mortgage franchise, which suggests their warehouse and servicing economics produce a positive spread where loanDepot's does not — a structural difference, not simply a matter of origination volume. The market seems to be pricing this distinction: LDI's multiple has compressed from 3.53 to 0.36 over ten quarters while peers hold 0.80-1.28, implying investors see the discount as a solvency question rather than a cycle-positioning opportunity.