Saturday, December 9, 2017

Thoughts on convertible bonds

A convertible bond becomes “busted” when the issuer's stock price falls so far below the strike price as to render the conversion feature nugatory. When this happens, the convertible begins to trade like a regular bond. For instance, if bonds of similar size and credit quality trade at a 9-10% yield to maturity, the convert will typically trade at a price—almost certainly a big discount to par—that gives it a similar YTM.

Busted converts were great investments during the 2001-02 market downturn and again during the Global Financial Crisis. Their strong performance during those periods of stress piqued my interest, and I read as much as I could about the convert market, assuming that it would offer similar opportunities during the next crisis.

After a couple days of reading, I no longer think that will be the case. Actually, my reading suggests that converts will be relatively unattractive going forward, for several reasons:

The market has shrunk. Since 2007, the face value of outstanding convertible bonds has fallen from approximately $700bn to $500bn worldwide and from $300bn to $200bn in the United States. (By comparison, Apple's market cap is $870bn.) There are two reasons why the market is smaller: One, many companies issue converts to save money on interest, and historically low interest rates obviate that. Two, convertible arbitrage funds are major buyers of converts, and since the GFC they've been forced to use less leverage than they'd previously used, reducing their buying power.

With arbs using less leverage, there's less potential for forced selling to create bargains. Investor psychology has also changed. The convertible selloff after Lehman Brothers' bankruptcy was so ferocious in part because it took people by surprise. Nothing like that had happened before, at least not on the same scale. Now investors know it's within the realm of possibility, which I think makes it less likely to happen again. As the old cliché goes, a watched pot never boils.

Transaction costs during the post-Lehman selloff were exorbitant, with bid/ask spreads equal to 7% or more of the midpoint price for many converts:

(Source: The Handbook of Convertible Bonds by de Spiegeleer and Schoutens)

The market was chaotic and illiquid, with December 31, 2008 13f-hr filings showing that various hedge funds had one bond—Ciena's $300mm May 2013 issue—valued at anywhere from 48 to 53 cents on the dollar. Buying converts in 2008 was a great trade on paper, but trading costs made the true returns lower.

Convert issuers tend to be smaller, more speculative, lower-quality companies. Many of these companies also have debt that's senior to their converts—typically, converts are subordinated to all other existing and future debt issues. The GFC was painful but brief; in a protracted recession, converts would likely experience much higher defaults.

Tech industry over-represented

Technology companies are over-represented among convert issuers. I think this is because a convert is most attractive when the issuer's stock is volatile but it has low credit risk. Usually the two conditions don't go together: a company that's less likely to default will typically have a less volatile stock price. But tech companies often trade at high valuations because investors expect them to grow quickly. Hence the stocks are very sensitive to investors' beliefs about the future, and they can be quite volatile without implying anything about the companies' creditworthiness.

Issuers with net cash

In looking at dozens of convert issuers, I found a surprising number that had more cash than debt. Even more surprisingly, the net cash didn't necessarily make them safe credits.

Many tech companies issued convertibles during the dot-com bubble. In 2001, after the bubble had burst and the tech industry was suffering, some of these companies still had net cash. How their converts performed from that point on depended on how they had financed themselves. Some issuers that had raised money primarily through converts defaulted, while the ones that had raised money through a mix of equity and convert issuance almost all survived. In the latter group, many converts paid off at par even after the issuer's stock had fallen 95-99%.

A couple of tech issuers managed to stay afloat only by repurchasing most of their outstanding bonds at a big discount to par. These buybacks meant that they technically satisfied their obligations, but most of their bondholders didn't benefit.

While the dotcom-era tech issuers generally treated their creditors well, that wasn't the case during the GFC. Many issuers with net cash took creditor-unfriendly actions:

ADC Telecom bought back stock in late 2008 while its converts traded at 45 cents on the dollar.
Ciena made a big all-cash acquisition in 2009 as the GFC appeared to be ending, taking it from a net cash position to a significant net debt position. Integrating the acquisition was more difficult than expected and Ciena lost money for several years.
Energy Conversion Devices invested most of its cash in growth initiatives, only to find that its niche solar technology wasn't cost-competitive with polycrystalline solar cells.
TTM Industries made a big acquisition in 2009 that took it from having working capital in excess of all liabilities to having significant net debt, primarily bank loans that were senior to its converts.

Energy Conversion Devices went bankrupt within a few years. The other companies survived and repaid their converts, but the expectation that their net cash provided a margin of safety turned out to be wrong.

By looking at issuers' recent actions, investors may be able to find net-cash issuers that are more favorably disposed to creditors:

In 2002, Juniper Networks had $1.7bn of cash and $1.15bn of outstanding convertible bonds. That May, it acquired Siemens' networking business. Juniper had enough money to buy the business entirely for cash, but instead it paid in a mix of cash and stock, leaving it with a net cash position. A few months after the acquisition, it bought back ~18% of its converts at a discount to par. In 2003, it tendered for the remaining bonds at par.

In 2007, Arris Group acquired C-Cor. Like Juniper, Arris made the acquisition with a mix of cash and stock that left it with a net cash position. It later bought back some of its converts during the GFC and stated in its 2008 10-K that “Maintain[ing] a strong capital structure, mindful of our 2013 debt maturity” was a key part of its corporate strategy.

No comments:

Post a Comment