Price a private loan as if you had to sell it
Life and annuity companies have moved a large share of their portfolios into privately originated credit. The extra yield is real, but boards usually see it long before they see what the loans would fetch in a sale, and that ordering is where the risk hides.
Ten years ago a typical American life insurer's general account was a portfolio of public corporate bonds, with mortgage loans and private placements at the edges. Today a large share of the industry's new money goes into credit that no public market prices: direct lending, asset-based finance secured on consumer and equipment loans, privately rated structured notes, and loans originated by asset managers that also own, advise or reinsure insurance companies. Wrenfield is part of that shift. About 27 per cent of our general account sits in privately priced assets, and since 2022 a minority shareholder has managed $5.2 billion of it.
I support owning these assets. Annuity liabilities are long and fairly predictable, and an insurer that can hold a loan for years should be paid for doing so. My argument is narrower. The extra yield usually reaches the board months or years before anyone shows it what the loans would sell for, and in that interval a company can build a portfolio whose stated income depends on never having to find out.
Where the extra yield comes from
Any premium over public bonds with the same rating has four possible sources. Some of it pays for illiquidity: we cannot sell quickly, so we earn more. Some pays for complexity, because a loan with bespoke terms takes work to document and monitor. Some pays for credit risk the rating does not capture, since a private rating assigned to a single transaction has never been tested by a market. And some is simply a fee, shared or not shared between the originator and the holder.
The first two are worth having. The third is the one to worry about, and the fourth is the one to measure. When we reviewed our privately priced holdings in 2023, the premium over comparable public credit ranged from under half a point to more than three points, and the largest premiums sat disproportionately in privately rated structured notes whose rating had been commissioned by the arranger. That did not make those notes bad assets. It meant the board was being shown a yield without a breakdown of what it was paying for.
Statutory capital adds a second incentive. A US insurer's capital charge for a bond depends on its NAIC designation, which for most privately rated assets has followed the private rating. A structure rated single-A attracts a far smaller charge than the loans inside it would on their own. Regulators have been tightening that link, giving their securities valuation staff more discretion to challenge private ratings, and I welcome it. But no company should wait for a regulator to tell it whether its own capital is real.
Two marks and the space between them
At Wrenfield every privately priced position now carries two valuations. The first comes from the manager that originated or selected it, using its own models and, where possible, recent comparable trades. The second comes from a valuation firm chosen by our Audit Committee, not by the investment team and not by the manager, working from the same loan-level data. Finance owns the comparison.
Most of the time the two marks agree within a point or two. The interesting cases are the ones where they do not. In 2024 they differed by more than three points on six of the 61 positions in the mandate our minority shareholder manages. In four cases the gap narrowed after a discussion of assumptions. In two it did not, and one of those, a facility secured on point-of-sale consumer loans, was later written down by $38 million. The outside mark had been closer.
A second valuation costs money, around $1.1 million a year for us, and it irritates managers who consider their own marks better informed. Sometimes they are right. The purpose is not to replace the manager's judgement. It is to make disagreement visible to people who are not paid by either side of it.
The harder question is which price to use. Fair value in accounting terms is the price in an orderly transaction between willing parties. I also ask for a second figure: what the position would fetch if we had to sell it within a quarter, in a market where other insurers might be selling too. That figure is an estimate, often a wide one. Its use lies in forcing a conversation about how much of the portfolio we could turn into cash, and at what loss, if policyholders surrendered faster than we planned. We learned in 2023 how quickly that can happen, when $2.3 billion of surrenders arrived within eighteen months and the assets we could sell cheaply were, unsurprisingly, public bonds rather than the private loans that had been earning the extra yield.
Four lines for the investment committee pack
I would ask any life or annuity board to see four lines every quarter, whatever the size of its private credit book.
The first is the share of the general account without a daily public price, shown by asset class and by originator, including any manager that is a shareholder or an affiliate. The second is the spread over comparable public credit for each class after all fees, with the fees themselves shown through every layer of the structure. The third is the gap between the manager's mark and an outside mark, reported position by position wherever it exceeds a threshold the committee has set. The fourth is the estimated sale price for each class within a quarter, under a stress in which surrenders double.
None of these lines tells a board whether to own private credit. They tell it whether it understands what it owns. A committee that sees them will ask sharper questions about concentration, about who benefits from each new allocation and about how the portfolio would behave if the company needed cash in a hurry.
This discipline has a cost, and I should own it. Our portfolio yield is lower than it would have been had we accepted the full mandate our shareholder proposed in 2022, and some analysts compare our spread unfavourably with those of insurers owned by asset managers. My answer is that I would rather report a spread I can defend in a forced sale than one that depends on never facing one. Annuity owners lend us their savings on the understanding that the money will be there when they ask for it. The assets behind that promise should be priced as though they might be called upon, because sooner or later some of them will be.