A contingent convertible is a bond that converts into equity after some triggering event such as the decline in the debt equity ratio below a threshold level. For example, the first cocobond was issued in November 2009 by Lloyds Bank. The bond converts into equity as soon as the bank's Core Tier 1 leverage ratio falls below five per cent.

The issuance of contingent capital has been proposed as a method to avoid a new financial crisis and avoid government bailouts of banks "too big to fail". Interest in such securities has been made stronger after the observation that in the recent crisis governments bailed out subordinated debt holders.

There are, however, a number of concerns with cocobonds.

First, how should the trigger be set? In the case of Lloyds the trigger is based on regulatory capital ratios, which are only computed once every quarter.

Such a mechanism will not work in a situation where a bank capital structure deteriorates rapidly. For example, Citibank had a Tier 1 capital ratio that was measured at 11.8 per cent in December 2008, at the height of the financial crisis. This means debt holders are not likely to get their money back, which makes the bonds very expensive, and reduces their attractiveness as an alternative to simply issuing equity.

An alternative is to set the trigger based on market values of equity, not book values. For example the bond with $1,000 nominal value is convertible into 200 shares whenever the bank's stock price falls below five dollars. So the conversion price is in this example equal to the trigger price: five dollars.

In theory, this should make the bonds risk-free as the investor can sell the 200 shares at five dollars and get his money back.

However, market-based triggers create other problems. Specifically they create incentives for bondholders to short the stock to trigger conversion and dilute the equity holders. What makes manipulation easy here is that the short seller does not have to cover his short position by buying in the open market: the issuers will deliver the shares to the bondholder after conversion.

Even without manipulation through short sales, market panic that drives the stock price below "fair" value and triggers conversion can create the same diluting effect. Moreover, in a financial crisis, stock prices tend to experience strong downward jumps that make the numerical example above purely theoretical: when the stock hits $5 on a specific moment and conversion is triggered, by the time the investor has received his 200 shares, stock prices may well be significantly less than $5. In other words, the bondholder is again exposed to risk, which will make the bonds expensive and therefore unattractive to issuers.

In a recent paper George Pennacchi, Christian Wolff and I have proposed a new type of cocobond that we have baptised as 'COERC', a Call Option Enhanced Reverse Convertible. We believe that our design deals with the three concerns about cocobonds.

First, the trigger is based on market values, not "old" capital ratios. Second, the design protects equity holders against dilution caused by manipulation and/or market panic. Third, the COERC is far less risky than the typical cocobond structure proposed by others. As a result banks will be able to issue them without paying large risk premiums.

How does a COERC work?

Going back to our numerical example: the bond converts again into equity whenever the stock price falls below five dollars. However, now the conversion price is much lower than the trigger price, say one dollar per share. That means that the bond holder will now receive 1,000 shares instead of 200 shares.

Of course, this will create massive dilution to shareholders if this happens. However, under our structure the shareholders will have a right to buy back the shares from the bondholders at one dollar, thereby undoing the dilution. Shareholders will do this by making a rights issue at one dollar per share as soon as the conversion is triggered.

So what, effectively, will happen is that equity holders will be coerced into refinancing the bank and repaying the debt. So the equity issue is not imposed by the government or the regulator, but by the concern about massive dilution.

Note that under our structure the shareholders don't have to worry about manipulation or short-selling as they can always undo any dilution created through these activities. Second, because the equity holders have a large incentive to repay the debt holders, COERCs are much less risky than other bonds or non-convertible bonds. Our simulations show that credit spreads are generally small so that the instrument should be interesting to issuers as well as risk-averse fixed income investors.

Avoiding the next financial crisis can happen in two ways: putting restrictions on leverage and risk taking by banks which will be at the expense of economic growth. Or, thinking about innovative ways in financing banks without lowering their value and growth opportunities. We believe our COERC idea is consistent with the second approach.

Theo Vermaelen is Professor of Finance at INSEAD

This article was published in The Wall Street Journal on August 20, 2010