Phantom Premium – Part 2: Quantifying the visibility premium in AAA CLOs

But the 70bps premium is the starting point, not the answer. The CLO market has long debated why triple A spreads remain where they are.

Some point to regulatory treatment, others to complexity, financing, yield, market technicals or liquidity. These explanations are usually discussed together but rarely separated or assigned a value.

First, we estimate the cost of call optionality at 10bps. A typical new issue CLO has a five-year reinvestment period and two years of call protection, after which it can be refinanced, reset or called at par. This creates reinvestment risk for the triple A investor when spreads tighten. Corporate bonds can also be callable but make-whole protection generally makes that option a far less valuable one to the issuer.

Because discount margin does not isolate the value of the refinancing option, we model triple A cashflows across a range of future spread scenarios using the broad framework described by Santander.² We assume the CLO is refinanced when the savings justify the transaction cost.

On that basis, a new-issue triple A tranche quoted at a discount margin of 125bps has an estimated OAS of approximately 115bps.

Second, we assign 10bps to tranche size and idiosyncrasy. Triple A CLO tranches are generally a few hundred million dollars, rather than the multibillion-dollar benchmark issues common in corporate credit. Their smaller size limits the buyer base, while the analysis required for each structure makes the bonds less interchangeable. Because there is no directly observable spread for this difference, we use the midpoint of a 5bps to 15bps range.³

At the midpoint of those ranges, the combined effect of the higher repo rate and additional haircut implies a financing disadvantage of approximately 60bps for a fully financed investor. We apply one quarter of that amount to reflect a marginal buyer that is partly financed.

Fourth, we assign 10bps to immediacy. A BWIC provides access to liquidity, but not the same certainty as a continuous market. Execution takes time, the number of buyers is not known in advance and the clearing level is established only when bids are submitted.

This allowance reflects that execution horizon and uncertainty. This leaves a residual of 25bps, or 20% of the triple A CLO spread. That is the Phantom Premium. It is the price investors charge when they cannot see where risk will clear, how quickly or in what size.

And the cost is ultimately borne by CLO equity. The residual may also capture capital treatment, extension risk and technical supply and demand. But the estimate is conservative in one important respect; financing and immediacy, which we account for separately, are themselves consequences of limited visibility.

The important conclusion is that a material premium remains after making reasonable allowances for the factors the market already prices, and that limited visibility appears to explain a meaningful part of it.

That limits participation by systematic and macro investors, who require reliable pricing, efficient financing and confidence that they can exit. Their absence narrows the marginal buyer base and concentrates liquidity in the most visible part of the market.

And the cost adds up. Triple A tranches represent roughly 65% of the USD 1.3trn CLO market, so at our 25bps estimate, the Phantom Premium costs CLO equity investors more than USD 2bn a year.

Over five to six years, that equates to approximately USD 11bn to USD 13bn of additional excess spread that could otherwise accrue to CLO equity before discounting.

Recovering even part of that cost requires more than an active BWIC market. BWICs account for more than half of historical CLO TRACE volume and move risk effectively, but they do not keep bonds visible between auctions.

Continuous secondary market support would give investors clearer pricing and greater confidence that they can exit, broadening the buyer base and, over time, narrowing the Phantom Premium.