The need to delta hedge CLO price risk
admin | April 24, 2020
This document is intended for institutional investors and is not subject to all of the independence and disclosure standards applicable to debt research reports prepared for retail investors.
Hedging CLOs is notoriously difficult. Spreads on most CLO debt show weak correlations over time with a range of simple, fixed hedging instruments across high yield credit and equity. But CLO debt trades and should be hedged much more like an option with hedge ratios that change significantly with the broad price of leveraged loans. Price action in CLOs so far through the coronavirus crisis is just the latest clear example.
CLOs from mid-February to today have traded at prices ranging from the high $50s to around $100. The pricing reflects the usual mix of considerations including the underlying leveraged loans, the CLO structure, the manager and general market conditions (see A model of CLO spreads). However, pricing clearly reflects the combination of loan portfolio market value and structure captured by market value overcollateralization, or MVOC. MVOC arguably reflects the market’s opinion of whether a deal will be able to liquidate its loan portfolio—whether through sale or maturity—and pay off debt.
At the highest levels of MVOC seen since mid-February, belonging almost entirely to ‘AAA’ classes, debt has traded near $100 (Exhibit 1). Prices start to soften for ‘AAA’ debt as MVOC declines from 145 toward 110. And as MVOC rounds 110 and heads lower, prices for most classes of debt start to fall at an accelerating pace.
Exhibit 1: CLO secondary prices have ranged widely since mid-February
The relationship between price and MVOC is clear after fitting a simple curve to trades seen since mid-February (Exhibit 2). The curve shows the price stability at high levels of MVOC, the slow decline from 145 to 110 and the acceleration below 110.
Exhibit 2: CLO debt prices vary clearly with MVOC
The sensitivity of CLO debt price to a 1-point change in MVOC reflects the option relationship between CLO equity and debt. CLO equity holds a put option on the market value or eventual par return from the loan portfolio. Losses to equity are limited to the par subordination below the first rated class, with the par subordination creating the strike price on the put. In a $100 CLO loan portfolio with $10 of subordination to ‘BB,’ for example, equity can only lose $10. The strike price on the put, consequently, is $90. As MVOC on ‘BB’ declines from (100/90), or 111, to (90/90), or 100, the put option becomes more valuable. The ‘BB’ class has written the put option to the equity class, so, as the option rises in value the price of the ‘BB’ should fall. This is true for other classes, too. More senior classes have written put options to the debt immediately junior in the capital structure, with these puts struck at progressively lower and lower strike prices reflecting rising par subordination.
The curve in the line between MVOC and price reflects the changing delta of the option, or the changing value of the option for a 1-point change in the value of the loan portfolio. For a CLO class with an MVOC of 100, the delta should be around 0.5, meaning a 1-point change in the value of the loan portfolio should come with a 0.5-point change in the value of the put. The market is arguably implying a 50/50 chance that the portfolio will return enough principal to fully pay off the debt. If the value of the loan portfolio goes up and MVOC rises significantly, the delta should drop, implying lower odds of a principal shortfall. And if the value of the loan portfolio goes down and MVOC falls significantly, the delta should quickly approach 1.0, implying almost certainty of shortfall. The value of debt then would fall 1-for-1 with the loan portfolio.
Hedging the market value of a CLO class should involve calculating its delta to the underlying portfolio of leveraged loans and selling short the amount of leveraged loans needed to offset the potential change in market value. Since the delta changes with each change in market value, so should the hedge ratio. The change is most rapid around an MVOC of 100.
A careful calculation of delta would also include the weighted average life or maturity of the hedged class. CLOs with longer weighted average lives would show deltas that change relatively more slowly around 100 since moves above or below 100 would have plenty of time to reverse. CLOs with short weighted average lives would show sharper moves in delta around 100 since those moves have less time to reverse.
There are practical limits to selling leveraged loans short. The loan market does not really offer a way to short sell loans. It may be more plausible to sell short leveraged loan ETFs such as BKLN or others, but an investor would have to weigh the cost of carry and the bid-ask involved in rebalancing the hedge.
Even for portfolios that commonly hedge CLOs with high yield credit default swaps or similar instruments, calculating a delta is still important. The daily correlation since mid-February between daily price changes in BKLN and HYG, a high yield ETF, has been 0.84. That is a reasonable correlation, although still fairly noisy. A calculated delta between CLOs with MVOCs below 110 and HYG should help stabilize the CLO position. As for CLOs with higher MVOCs, the delta to BKLN or HYG should be low. In fact, the ETFs could end up adding more noise than protection at higher MVOCs. The best hedge for CLOs at MVOCs above 125 may be no hedge at all.