COCO bonds: Good finance or puffed-up accounting?

W e recently read about a new financial instrument that has become popular with some shortsighted managers over the last few years. Specifically, we’re talking about “contingent convertible bonds,” known as COCO bonds.

Writing in the Sept. 9, 2003, issue of Business Week, David Henry strongly criticized the generally accepted accounting principles treatment of them, especially the earnings-per-share calculation. Because APB Opinion 14 disallows breaking a convertible into components, managers issue COCOs, but account for them like regular bonds. Henry rightfully reported that the contingent convertibility has led literalist accountants to not acknowledge any dilutive effect in earnings per share from the possibility of conversion.

He is very right in reporting that there is a real but unreported risk of dilution from the sudden expansion of the outstanding shares in the EPS denominator that happens if and when conversion occurs. His article also includes various quotes from issuers and investment bankers to the effect that COCOs are great because they produce lower apparent interest rates and higher reported earnings per share.

We clearly agree with the tenor of David’s analysis — he is right on target. However, we don’t think that he, and certainly not COCO proponents, really comprehend the full weakness of current GAAP.

A regular convertible bond is not all that complicated.

First, it creates a stream of contractual cash flows, always including principal and often including interest, unless it is a zero. Second, the bond gives the holder the right to exchange it for common shares. A mere moment’s contemplation shows that the investor really owns two different instruments: a stream of fixed cash flows, plus a call option for a fixed number of shares where the strike price is to be paid by delivering the bond.

A COCO is not much different, at least in its basic form. Where it differs is that the conversion right is triggered if — and only if — the stock price tops a stated market value.

The same analysis holds: The so-called “borrower” actually issues two securities. One is the stream of contractual principal and interest cash flows, where the latter amount equals the face rate times the face value. The second piece is a convoluted (and more speculative) call option that may allow shares to be acquired below their market value in the future.

In effect, the deal is a lump-sum purchase, not unlike what happens when someone buys a big-screen television set for $1,000 and receives a DVD player as a “gift.” If the DVD player’s value is, say, $100, then the TV’s real value is probably $900, not $1,000. Nothing valuable is free, including call options imbedded in COCO bonds.

For example, suppose that XYZ Co. issues 20-year, 6 percent COCO bonds at their $100 million face value when the market rate for regular bonds is 7.5 percent. The covenant creates a $6 million annual cash payment that is reported under GAAP as interest on the income statement; in addition, the balance sheet reports a $100 million liability, thereby implying a 6 percent cost of capital.

By applying the market rate to the cash flows, the bonds have a present value of only $89.3 million, implying that the tantalizingly “free” conversion option is worth roughly $10.7 million. If the real rate of 7.5 percent is applied to the initial debt of $89.3 million, the annual cost is $6.7 million, clearly a more complete and more useful description of the issuer’s capital cost. Thus, GAAP once again creates an opportunity to pump up EPS.

In addition, XYZ has created a derivative liability of $10.7 million for the options that carries the threat of diluting the shareholders’ investments. By closing their eyes to the conversion feature, the APB (and the Financial Accounting Standards Board, through its failure to act) promulgated bad accounting.

Henry also describes the puffery inherent in GAAP by relating how Omnicom issued $900 million of zero-rate COCOs at face value and recorded them with a debit to cash and a credit to bonds payable. Because there was no discount (wink, wink), the company is reporting no interest over their life. My goodness, free money!

Of course, it is no such thing — management created (and hid) a discount by granting options to the bonds’ buyers. For Omnicom, the future cash flows discounted at a market rate of 8.5 percent have a present value of only $822 million, which means that the market discount is $78 million.

This amount is usefully reportable as interest over the bonds’ 30-year life, but remains totally unreported under GAAP. The other untold story is a $78 million call option liability that could (and should) be separately reported because its value describes the potential dilution of the existing stockholders’ interests.

Only a foolish person would believe that the GAAP picture is believed or used by sophisticated analysts. Perhaps that explains why managers of hundreds of companies, including General Motors, have scrambled to create the bogus illusion of cheap, even free, money.

According to Henry’s article, GM issued $4.3 billion of 6.25 percent, 30-year COCO bonds when the market rate was 8.5 percent. Using these facts, we estimate a market value of $3.26 billion for the bonds and observe that GM’s managers put $1.04 billion of their shareholders’ wealth at risk merely to produce the illusion of a lower cost of capital. What nonsense flows from bad GAAP!

So, what’s the remedy? The traditional response is to run to FASB to demand a new standard. Of course, as we have seen in recent years, the investment banking community stays one step ahead, finding new loopholes faster than FASB can close them down, always knowing that it will find bigger fools among managers who are ready to spend vast sums structuring transactions to pump up reported earnings simply because they don’t understand the rudiments of finance.

Another solution involves improved disclosures of the real underlying economics, leading to more complete and precise measures of the effects of issuing COCOs — and regular convertibles, for that matter. The added precision will help experts usefully discount the issuer’s stock value to reflect the shift of wealth from shareholders to the bond holders.

As we see it, managers who pursue Quality Financial Reporting will not wait for FASB, but will voluntarily disclose the deals’ underlying economics, not just when they happen, but as long as the bonds and options are outstanding.

Then again, the best practice would be to simply say, “No,” when investment bankers throw out the bait of cheap or free money. Even the simplest among us knows that there is no such thing. Those who fall for this temptation are basically declaring that they think the capital markets are composed of fools. Of course, their actions really proclaim to the world that they themselves are, at best, foolish — or, at worst, deliberately deceitful. In either case, the consequence must be a discounted stock price, not a higher one.

So, there we have it again — an argument in favor of telling the truth that relies on economic reasoning, not ethics. We continue to wonder how long it’s going to take for this idea to catch on. The delay can’t be blamed on us. It rests squarely on accountants, standard-setters and investment bankers who feed heaping bowls of COCO-Puff accounting to the capital markets as though they consist of only gullible investors who gobble down every artificially sweetened financial statement, blissfully ignorant of the empty calories, slavishly believing that everything is fine as long as management complies with the minutiae contained within GAAP.

Frankly, we’ve grown quite tired of that kind of arrogant behavior. The passivity of auditors, standard-setters and regulators is equally aggravating. We hope that you can see the need for a revolution as clearly as we do. It’s long past time for financial reports to be healthy and nutritious, instead of junk food.