Banks are stuck holding Elon Musk's Twitter debt

When Elon Musk signed a binding agreement in April to buy Twitter for $44 billion, Wall Street's biggest banks thought they had won the underwriting mandate of the year. A syndicate led by Morgan Stanley, Bank of America and Barclays agreed to provide $13 billion of debt to fund the leveraged buyout. Six months on, with a court-ordered deadline to close the deal on 28 October, that commitment has become the most painful hung debt since the leveraged buyout boom collapsed in 2008.

The banks are suffering from commitment risk in a market where rates have moved fast. When they signed their commitment letters in the spring, benchmark rates were a fraction of today's levels and high-yield spreads were comfortably tight. Tech valuations had not yet suffered their mid-year slump. The debt package consists of $6.5 billion of term loans, $3 billion of secured bonds and $3 billion of unsecured notes, and by agreeing to cap the interest rates on it the banks in effect sold a large, unhedged put option on the credit market.

Six months of brinkmanship

For the lenders, a collapse of the deal would have been the cheapest outcome. Mr Musk agreed in April to pay $54.20 a share, tried in July to walk away, and was sued by Twitter in Delaware's Court of Chancery to force him to close. On 4 October he offered to complete the purchase on the original terms, and the court paused the trial that was due to start this month, giving the parties until 28 October to finish the job. The banks got a closing date instead of an exit.

The rest of the price comes from Mr Musk and his co-investors as equity, and he sold several billion dollars' worth of Tesla shares in April and again in August to raise his part. Equity ranks behind the banks' debt, so it cushions the lenders against losses. It does nothing for the price at which that debt can be sold, which is set by the yields investors demand today.

Nobody wants the paper

Leveraged finance normally works on an originate-to-distribute model. Banks hold the debt for a few weeks between signing and closing, split the loans and bonds into tranches and sell them to institutional investors such as pension funds, collateralised loan obligations (CLOs) and private credit managers. The banks collect generous fees and keep little long-term credit risk on their own books.

For Twitter, that machinery has broken down. The market for leveraged loans and junk bonds has deteriorated so far that buyers want yields above 11 per cent on senior secured paper and well over 13 per cent on unsecured debt. Because the Twitter commitments carry interest-rate caps agreed in the easy-money era, the debt can be sold only at a deep discount to par. Traders reckon the unsecured notes would struggle to clear at 60 to 70 cents on the dollar, and the secured debt nearer 80 cents.

The largest natural buyers face constraints of their own. CLOs, which buy most American leveraged loans, typically limit their holdings of loans rated in the CCC category to 7.5 per cent of their portfolios, and a buyout this heavily geared risks ratings low enough to crowd that bucket. The funds are also finding it harder to raise new money, because the investors who buy CLO tranches want wider spreads too.

Twitter's own accounts make the paper harder to place. Interest on $13 billion at the yields buyers now demand would exceed $1 billion a year, against revenue of about $5 billion in 2021, a year in which the company reported a net loss. Credit investors are being asked to lend at yesterday's prices to a business that has rarely made an annual profit.

Stuck on the books

Facing possible mark-to-market losses of between $1 billion and $2 billion, the banks have a grim choice. They can try to sell the debt now and take losses big enough to wipe out whole quarters of leveraged finance earnings. Or they can keep the $13 billion on their own balance sheets, fund it with costly commercial paper and deposits, and hope the market recovers.

Keeping the debt is a headache for bank treasurers. Holding $13 billion of illiquid, high-risk loans eats up risk-weighted assets under the Basel III capital rules, tying up capital that could otherwise back more profitable corporate lending. And under mark-to-market accounting the banks will still have to take sizeable write-downs on the hung commitments every quarter.

Echoes of 2007

Banks were last caught this way in the summer of 2007, when the credit market shut with debt for buyouts such as Chrysler and First Data still on their books. The Clear Channel buyout ended up in court in 2008 when its banks resisted funding it, and the overhang choked new buyout lending until it had been sold off at discounts. The pattern is repeating on a smaller scale: in September, banks that underwrote the buyout of Citrix sold much of its debt at steep discounts and booked losses to clear the decks.

Private credit fills the gap

The winners are the lenders that never promised to sell anything. Direct-lending funds, which hold loans to maturity instead of distributing them, have been taking buyout financing from the banks all year, and with syndication desks nursing hung deals they can now demand higher spreads, tighter covenants and bigger fees. Buyout firms pay more for their debt, and the risk migrates from regulated banks to funds whose investors are locked in for years.

The case against gloom is that marks are not defaults. Banks can hold the loans, collect interest well above their funding costs and sell later if markets recover. That works only if the borrower's cash flow covers the interest, which is exactly what Twitter's history leaves in doubt. Third-quarter results, which the big American banks start reporting on 14 October, will show how much of the pain has already been marked.

The Twitter deal probably closes the window for leveraged finance in this cycle. The banks will fund Mr Musk's takeover because a contract signed in a very different market obliges them to, and the losses they take on it will keep private equity firms short of bank debt well into 2023. Buyout sponsors hoping for a quick reopening of the loan market are likely to be disappointed, and direct lenders will take a larger share of whatever deals do get done over the next year.