Last week, Royal Bank of Canada issued $1-billion worth of "non-viable contingent capital" bonds, known colloquially as NVCC bonds or, most enjoyably, as CoCos. Ideally, CoCos encourage investors to monitor banks for excess leverage. However, as new research by Jingya Li and Mark Reesor of the University of Western Ontario and Adam Metzler of Wilfrid Laurier University shows, if the conditions are right, CoCos may provide an incentive to cheat.
CoCos are supposed to shift the job of supplying troubled banks with equity from government to investors. Back in the heady days of 2008, a big problem in financial markets was that banks had too much debt and not enough capital. The results of this were not good, and governments spent lots of money to keep banks afloat.
One story as to how things got this way is the usual people we trust to monitor what a company is doing, shareholders, are not particularly good at monitoring what banks do. Shareholders are at the bottom of the capital structure and therefore like volatility because volatility increases the size of their returns. Thing is, we don't want volatile banks.
The Basel III banking rules have a solution to this problem. In order for non-common share instruments to count as loss absorbing or Tier 1 capital, those instruments have to either be written off or converted into equity when the bank reaches the point of non-viability. CoCos are debt that converts to equity when a bank gets into trouble and, therefore, can be counted as Tier 1 capital.
In essence, on the occurrence of a "triggering event" – which is almost always either a regulator telling the bank it is no longer viable (a "regulatory trigger") or based on some change in the bank's debt-to-equity ratio (an "accounting trigger") – CoCos will convert to equity. The stocks will be issued at either a fixed price or the market price of the bank's shares, thereby giving the distressed bank a jolt of capital to absorb losses. This is why CoCos are often called "bail-in" debt; bondholders recapitalize troubled banks so governments don't have to.
There are some nice incentives here. CoCo investors want to realize the yield from their bond and will lose that yield if they get converted into equity and knocked down the capital structure. Therefore, CoCo investors should have better incentives than volatility-loving shareholders to monitor a bank's leverage and suppress volatility.
But not always. Ms. Li's work shows that certain CoCo structures offer investors very different incentives as the CoCos approach conversion. Where the conversion price of the CoCo is based on the market price of the institution's common stock and the CoCo uses an accounting trigger, investors have an incentive to short the underlying shares in order to depress the share price prior to conversion. When the CoCos convert they will then convert at an artificially low price. When the price of the underlying shares rebounds – either because CoCo investors have closed their short positions or because the share price returns to reflect fundamental value – the CoCo investors will see a return on their newly converted shares.
Depending on how close to conversion CoCo investors open their short, Ms. Li calculates the return to short-selling at conversion to be between 3.7 per cent and 7.8 per cent.
In other words, there's an incentive for CoCo investors to drive down the price of an institution's stock as conversion approaches, making an already precarious time for the bank even more dangerous. At worst, this could lead a financial institution into a death spiral, where falling price begets falling investor confidence making it prohibitively expensive for a bank to raise further capital.
It would be easy to decry this short selling as market manipulation, but the legalities are more complicated. One way to hedge exposure to CoCos is to be short the underlying shares. As a financial institution's leverage increases, the price of both the CoCos and the shares will decrease, meaning that one can offset the loss on the CoCo by the gain on a short sale of the institution's securities. Since there's no rule that hedging needs to be perfect, clever holders of large amounts of CoCos could theoretically engage in market manipulation by aggressively short selling to drive down the price of the underlying shares, and have a credible defense that their actions were merely a hedge. The line between illegitimate market manipulation and legitimate hedging is a fine one.
Moreover, even legitimate hedging may pose a problem. Ms. Li's work estimates that while normal junior bondholders need to be short about 2.5 per cent of a bank's equity to perfectly hedge their stake, CoCo bondholders need to be short 5.61 per cent of equity to perfectly hedge their holdings assuming that CoCos constitute an average amount of junior debt for a Canadian bank. Given that a 5.61 per cent short position on the part of CoCo investors would turn investors a neat profit on conversion and would also hedge any loss at conversion, CoCo investors would be imprudent not to fully hedge their holdings as conversion approaches. With only a few percentage points of the equity value of a bank short at any given time, even this seemingly small short position could potentially influence the price.
Simply, our laws governing market manipulation are inadequate to deal with short-selling incentives as CoCos approach conversion.
Ms. Li's research shows that not all structures offer equal short-selling incentives. CoCos that use a trailing average with a price floor for the conversion price would decrease short selling incentives. The trailing average would increase the conversion price above the suppressed price, while the price floor would cap the amount of profit short sellers could make. Second, a regulatory trigger is more difficult for investors to game. Under the accounting trigger, an investor only needs to predict when a bank's publicly available leverage ratio will cross the triggering threshold. Under the regulatory trigger, regulators can time the conversion to thwart short-sellers.
On the above measures, the RBC issuance grades out pretty well. It uses a regulatory trigger and has a price floor, though it does not have a trailing average for the conversion price. While the issuance suggests some short-selling incentives near conversion, they are much smaller than they could be.
Still, the best solution to this issue appears to be a regulatory one and Ms. Li's paper offers advice that the Office of the Superintendent of Financial Institutions would be wise to consider. While current OSFI guidance mandates that all CoCos include a regulatory trigger, Ms. Li's paper offers evidence that accounting triggers should be more closely scrutinized or, perhaps, prohibited. Similarly, it calls into question the wisdom of OSFI's guidance that the conversion terms must reference the market price of common equity at or before a triggering event. OSFI should think about mandating both a trailing average price and a floor price to limit short-selling incentives.
Regulation is hard, especially with new securities. But CoCos are a good idea, and RBC has come up with a good structure, and researchers like Ms. Li are providing data showing both the potential costs and benefits of CoCos. OSFI should modify its guidance to make sure that CoCos help prevent the next crisis, not cause it.