When biotech company Cephalon Inc. wanted to raise money and refinance old debt in June, it thumbed its nose at this year's low rates. In fact, it didn't want to pay any interest at all on its new bonds. And thanks to an accounting loophole, it was able to fashion a 0%, $750 million offering of so-called contingent convertible bonds that investors snapped up.
Dubbed co-cos for short, such bonds are the latest craze in corporate finance. This year more than 100 companies, including Comverse Technology and advertising-agency holding company Omnicom Group, have sold them in various forms. In June, General Motors sold the largest one yet, placing some $4.3 billion in bonds paying 6.25% to bolster its underfunded pension plan. At the time, standard GM debt was yielding about 8.5%. As with Cephalon, Omnicom is paying no interest after raising $600 million in June.
Raising money by selling bonds at little or no apparent cost is great for corporate bottom lines, but not so great for unsuspecting shareholders. The bonds carry below-market interest rates because investors, typically hedge funds, get a conversion option -- the right to swap their bonds for stock -- instead of interest payments. And, just as with the stock options granted to employees, companies don't need to treat the conversion options as an expense under generally accepted accounting principles (GAAP).
Even more critical for investors, companies -- when counting their outstanding shares -- don't have to include the shares that would be issued if the bonds were converted to stock. So shareholders can get hit with a nasty surprise. Accounting analysts at Bear, Stearns & Co. say companies issuing co-cos would see their earnings per share slip by an average of 6% if the conversion options were included in the calculation.
For their part, the companies say they aren't hiding anything. Instead, they're doing the deals to get cheap capital, they say. Robert S. "Chip" Merritt, senior director of investor relations at Cephalon, says the company has been straight with the market. "Any professional investor would understand that this conversion could occur and would have modeled for it appropriately," he says. However, Christopher Senyek, an accountant at Bear Stearns (BSC ) who looked at more than 100 of the contingent convertible issues, says public information on the deals is often "sketchy at best," making it hard to gauge their impact on stock prices.
Investors can blame clever investment bankers for this latest maneuver that stays one step ahead of the regulators. Merrill Lynch, Credit Suisse First Boston, and Banc of America Securities are the kings of the co-co so far. Tyco International Ltd. (TYC ) is believed to have issued the first co-co in 2000; it was underwritten by Merrill Lynch & Co. shortly after a proposed rule on convertibles from the Financial Accounting Standards Board omitted contingency conversions. Companies issued almost 120 co-cos in the next two years. This year the deals have mushroomed further, with 105 issued through July, according to Bear Stearns.
The advantage of co-cos is simple. With regular convertible bonds, companies must count the shares that might be issued when figuring their per-share GAAP earnings. But they can postpone the dilution if they add another condition, or contingency, to the standard convertible terms. For example, when CSFB underwrites a co-co, it usually allows the switch to stock only if the company's share price reaches 20% above the price used to figure how many shares will be given for each bond at conversion. In Cephalon's case, half of its bonds can be converted into stock at $56.50, but only when the stock, now at around $44, trades above $67.80. If these bonds were regular convertibles, they could be swapped for stock as soon as the price reached $56.50.
Issuing a co-co rather than a regular convertible postpones the day of reckoning for Cephalon. But if its stock reaches the trigger level, the company's shares would suffer a sudden dilution, clobbering its per-share earnings because bondholders could shift into shares. Bear Stearns estimates that if this occurred next year, the hit could be 15%. Already, that has happened to some companies. Counting its co-co shares for the first time in the second quarter, wireless outfit UTStarcom lopped 11% off its per-share earnings.
So investment bankers have devised another product to deal with this problem -- a hedge to offset the dilution. But the hedges are expensive. Cephalon, for example, paid CSFB some $258 million for one that would pay the company in cash or shares and help maintain the earnings per share if its bonds are converted. But that's a huge slice of the $750 million Cephalon raised by selling the bonds in the first place. So at the same time, CSFB agreed to return $178 million if Cephalon would take back some of the risk. The result: Cephalon bears the risk for conversions above $72.08, instead of from the $56.50 conversion price.
There's method to this apparent I-pay-you, you-pay-me madness. Cephalon counts the $258 million as an expense against its taxable income. Figured at the maximum 35% corporate tax rate, the result is a $90 million tax saving, enough to cover the $80 million difference between what Cephalon paid CSFB and what it got back. Cephalon says the $178 million it retrieved from CSFB doesn't affect its taxes because the government doesn't tax gains from dealings in a company's own stock.
Meanwhile, FASB is playing catch-up. It has toyed with the options issue since at least 1990 and might propose a rule before the end of the year that would close the contingency exception for counting the shares behind a convertible. But if the past is any guide, the bankers will have found the next loophole before the new rule takes effect, leaving investors with another variation of the by-now familiar riddle: How much did my company really earn?
By David Henry in New York