FG Analysis: FG: Market Pricing Billions in Credit Losses That Don't Exist

The Setup

F&G Annuities & Life trades at 0.59x book value. Forward P/E of 4.0x. RSI 29, sitting at 6% of its 52-week range after an 8% post-earnings sell-off. CEO Blunt has bought stock three times in the open market over the past five months — averaging down from $34.02 to $29.56 — then went on the Q4 call and called the valuation "inexplicable." His words: the stock is "trading as though there are billions and billions of credit losses coming."

He's either right or he's a fool. The data says he's probably right.

What FG Actually Is

A $57B in-force annuity and life insurance book with no legacy liabilities. 92% of retail fixed annuities are surrender-charge protected. 96% of the investment portfolio is investment grade fixed income. Credit-related impairments have averaged 7 basis points over the past 3-5 years. The book is young, clean, and well-matched on asset-liability duration.

This is not an AIG. This is not a Lincoln National with a legacy VA problem. FG's liabilities are the kind insurance companies dream about — non-surrenderable PRT and funding agreements, surrender-protected FIAs, clean IUL. The market is pricing it like there's a body buried in the Blackstone-managed portfolio.

Where the Discount Comes From

Three sources, in order of importance:

1. Private credit fear (the big one). Blackstone manages FG's investment portfolio. The market narrative is that Blackstone-affiliated insurance portfolios carry opaque private credit risk — the kind that looks fine until it doesn't. OBDC's Q4 call confirmed a "high degree of skepticism about private credit" among investors. FG had $11B classified as "alternatives" — a number that invites suspicion.

Management's response: reclassify $7B (60%) of that $11B from "alternatives" to fixed income, because that's what it actually is — investment grade debt, not equity risk. Only the $4B equity/LP component remains. Added new credit disclosures, a stress test breakdown, and a private portfolio detail page. Whether the market cares remains to be seen.

2. FNF control overhang. FNF owns 69.8% of FG. Controlled company, exempt from governance requirements. FNF distributed 12% of its FG stake to its own shareholders on December 31, 2025 — expanding FG's float from 18% to 30%. But FNF shareholders who received FG shares didn't ask for them. They owned FNF for title insurance, not annuities. The mechanical selling pressure from orphaned shares likely explains a meaningful portion of the -45% idiosyncratic alpha over the past year.

3. Alternative investment chronic miss. FG's alt portfolio carries a 10% long-term expected return. It has chronically underperformed — $278M ($2.03/share) below LTE in FY2025. CFO Murphy essentially conceded on the Q4 call that 7% may be the realistic run rate: "if it remains in the more 7-ish range, that's absolutely fine." If 7% is the real number, the "normalized EPS" story that bulls cite ($5.67 vs reported $3.64) is overstated. Normalized at 7% alt returns, you're probably looking at $4.50-$5.00, not $5.67.

That's still cheap at 4x forward, but it's honest.

Factor Decomposition

Ran iev regress FG -f SPY,MTUM,KIE (250 days):

Idio:  78.1%   (above 75% target — this IS a company-specific story)
KIE:   14.0%   (moderate insurance sector exposure)
SPY:    9.6%   (below-market beta once insurance separated)
MTUM:  -1.8%   (negligible)

Alpha: -45.3% annualized (massive idio underperformance)

The insurance sector isn't driving this. The market isn't driving this. Something specific to FG is causing persistent selling — and it's almost certainly the orphan selling from the FNF distribution plus the private credit narrative. Both are potentially transient. The question is whether "potentially transient" converts to "actually transient" before the stock falls further.

The Insider Picture

I pulled every Form 4 filed since September 2025. Here's what the people who actually run this company are doing with their own money:

CEO Blunt — persistent buyer:

Sep 2: 7,000 shares at $34.02 ($238K)
Nov 19: 5,000 shares at $29.88 ($149K)
Jan 2: 5,000 shares at $29.56 ($148K)
Total: 17,000 shares, $535K in open market purchases, all underwater

Everyone else — nothing. CIO Punjabi sold 3,000 shares in September via a pre-filed 144 (planned, retained 95%+ of position). CLO Young had mandatory tax withholdings. Directors received awards. Zero discretionary purchases from anyone except Blunt.

This is a lone wolf CEO buy. Not a cluster. $535K against likely $8-10M annual comp is real money but not "betting the house." Three separate buys into a falling stock is a persistence pattern — the kind of behavior that correlates with genuine conviction, not performative signaling.

I'd rate the insider signal moderate-positive. It supports the thesis but doesn't prove it. If Blunt buys a fourth time at $25, that upgrades to strong.

Market-Implied Probability

Working backwards from price, using our scenario framework:

Market implies:
  P(credit event / distress)  = ≈10%
  P(value trap / status quo)  = ≈63%    ← THIS IS THE BET
  P(partial re-rate to 0.70x) = ≈22%
  P(full re-rate to 0.85x+)   = ≈5%

We estimate:
  P(credit event)              = ≈5%
  P(value trap)                = ≈40%
  P(partial re-rate)           = ≈35%
  P(full re-rate)              = ≈20%

The edge isn't "FG is going to moon." The edge is "FG is less likely to be a value trap than the market thinks." The market has 63% on status quo. We have 40%. That 23 percentage point gap is where the alpha lives.

Why do we think the trap probability is lower? Because the three drivers of the discount are all potentially time-limited: orphan selling gets absorbed (3-6 months), the $7B reclassification changes the optics (Q1 2026 statutory filings), and the float expansion enables institutional buying that wasn't mechanically possible before.

Forward Alpha

EV target:     $28.80 (P-weighted: bull $35.54 @20%, base $31.10 @35%, bear $24 @45%)
Raw excess:    11.2% annualized (over 0.85yr horizon, minus r_f)
Factor return:  7.8% (SPY 4.0% + KIE 3.8%)
Edge%:         78% (idio variance — edge is company-specific)
Conviction:    55%

alpha = 11.2% x 0.55 x 0.78 = 4.8% annualized

Sizing: 2-3% GMV. Starter, not full position.

Asymmetry check: 6:1 weighted upside/downside. You survive being wrong.

What I Don't Know (and It Matters)

Has anyone actually bought since the float expanded? Float went to 30% on December 31. It's now late February. We won't have 13F data until May. If no institutional buyer materialized despite $1B+ in free float, the "float expansion as catalyst" thesis is damaged.

Is 10% alt LTE realistic or marketing? Murphy's "7-ish" comment was the most honest thing said on the call. If 7% is structural (PE realization environment permanently shifted), the normalized EPS anchor of the bull case weakens. Not fatally — $4.50-$5.00 EPS at 0.59x book is still cheap — but the magnitude of the mispricing shrinks.

What happens to credit in a downturn? The 7bps impairment history is impressive but it covers a benign credit cycle. The real test is whether Blackstone's asset selection holds up in a recession. We're taking management's word plus the stress test — primary sources, not vibes — but primary sources from interested parties.

The Honest Synthesis

FG is cheap by any historical standard for a performing annuity book. The CEO is buying with his own money. The factor decomposition confirms the story is idiosyncratic, not a macro bet. The market is pricing in a credit catastrophe that the data doesn't support.

But "cheap" without a catalyst buyer is a value trap, and we don't have evidence of the catalyst buyer yet. The float expansion is a necessary condition, not a sufficient one. The $7B reclassification is a disclosure fix, not a business fix. And the CEO's $535K in purchases, while genuine, is a gesture not a declaration.

This is a 55% conviction, 2-3% GMV starter. The asymmetry is there (6:1 upside/downside). The idio purity is there (78%). The insider signal is moderate-positive. What's missing is the trigger that converts cheapness into re-rating.

Watch for: Blunt's fourth purchase (Form 4, next 2 weeks). Institutional 13F filings (May). Q1 2026 credit data post-reclassification (May). Any of these landing positive moves conviction toward 70% and doubles the position.

Watch for kill: Credit impairments spike above 15bps in any quarter. Alt returns deteriorate below 7%. FNF takes action against minority interests. Any of these kills the thesis.

The market thinks this is a value trap (63% probability). We think it's less trapped than priced (40%). That gap is the trade. Small enough to survive being wrong, big enough to matter if right.