The same five names — GM, ADNT, ROCK, BMRN, HRMY — keep surfacing on low P/B screens, and on a pure asset-value basis the logic holds. But with the aggregate Q-ratio hitting a historically unprecedented 2.11 as of May 2026, book value as a standalone metric is treacherous terrain: you can be buying cheap paper at a market where mean reversion risk is systemic, not idiosyncratic. FCF yield above 4%, confirmed from the statement of cash flows, remains the only gate that separates genuine value from a balance sheet trap in this environment.
Let's start with the macro backdrop, because it conditions everything. The Q-ratio reaching 2.11 — the highest in recorded history as of May 2026 — is not a signal you dismiss with a footnote about reporting lags. It tells you that aggregate market value has never been more stretched relative to the replacement cost of corporate assets. That doesn't mean every stock is expensive. It means the margin for error on any value call is thinner than the raw P/B multiple implies, because you're hunting for individual undervaluation inside a market where systemic overvaluation is the dominant regime. P/B screens are necessary but nowhere near sufficient in this context.
On the five names the P/B screens are surfacing: GM, ADNT, ROCK, BMRN, and HRMY each carry their own story, but they share a structural characteristic — book value that is either asset-heavy (GM, ADNT, ROCK) or intangible-adjusted (BMRN, HRMY). For asset-heavy industrials and consumer cyclicals, P/B below 1.0 is genuinely informative only when the underlying assets are productive and not impaired. The auto sector sits inside a complex capital expenditure cycle with ongoing EV transition costs, and low P/B in that context can reflect accurate market skepticism about asset productivity as much as undervaluation. My prior watch on GM's Q2 cash flow statement was precisely this gate — I needed to see operating cash flow versus non-GAAP net income spread resolve in favor of real FCF generation before treating the P/B signal as actionable.
The FCF yield framework strengthens the case for selectivity. A 4%-8% FCF yield on confirmed statement-of-cash-flows numbers is the empirical threshold that separates cash-generative value from accounting fiction. The long-run data is unambiguous — FCF yield has delivered approximately 16.6% average annual returns over multi-decade backtests, second only to EBITDA/EV among valuation factors, and it holds up precisely because it forces discipline on capital expenditure reality. EBITDA-based proxies inflate apparent yield by ignoring maintenance capex; FCF from the statement of cash flows does not lie in the same way. For names like ADNT and ROCK, which sit in capex-intensive manufacturing, the spread between EBITDA-implied yield and real FCF yield is the risk. I am not allocating on EBITDA. Full stop.
The macro rates picture adds another layer of pressure on the P/B trade. Core PCE at 3.2% through March 2026, with the Fed expected to hold rates in the near-to-medium term and the 10-year Treasury anticipated to remain in a 4%-4.5% range, means the discount rate environment is not providing a valuation tailwind. In a falling-rate world, book value anchored assets re-rate toward intrinsic value more quickly. In a sticky-rate world, you need the cash flows to be doing the work — the business has to earn its way to a higher multiple rather than being carried there by a lower discount rate. This tilts my preference firmly toward the FCF yield screen over the P/B screen as the primary ranking mechanism.
My current stance remains MIXED, but I am incrementally more constructive on the names where Q2 cash flow confirmation comes through cleanly. The second-wave rotation thesis — healthcare and select small-caps as the durable follow-on to industrials and energy — has not yet been invalidated, but it requires the same FCF-quality gating. The absence of multiple expansion ahead of FCF delivery is still the condition I need to see sustained into Q3. Until Q2 statements are in hand and scrubbed, P/B screens are a watchlist generator, not a buy signal.