Banks, trading companies, leasing companies, and multinational corporations have currency convertibility risk even if they don't attempt to quantify the risk on their balance sheets. Buyers and sellers of currency convertibility protection must not only have a feel for pricing credit derivatives, they must be economists with a penchant for econometrics. These are negotiated transactions. Price, terms, conditions and size are all negotiated directly between counterparties. As there are so few counterparties for this type of protection, the broker market is usually ineffective. It is much more effective to contact well-known counterparties in the credit derivatives market and negotiate the transactions and market levels directly. This is a supply and demand driven market and prices vary from counterparty to counterparty.
Sovereign whim can drive this market. Buyers of convertibility protection have a knock-in spot trade. The knock-in is independent of currency levels, sovereign dynamics trigger the event. Credit spreads don't matter, models don't matter, intuition doesn't count. There is no exact mathematical model. The market defines the price.
Default Risk
Convertibility protection is sometimes combined with default risk. When convertibility is lumped with default protection language, a bank can purchase the hedge imbedded in macro country default protection risk. It is difficult to determine whether the buyer or seller got a fair price for the convertibility protection when risks are bundled. This is especially confusing when a convertibility event is used as a credit-default trigger event.
Convertibility default risk is usually greater than the risk of default by a sovereign entity on its debt obligations. The fact that most convertibility premiums trade above default options is intuitively correct, but there is no way to model it. In principal, I can create a balance sheet for a country--subject to data limitations--and come up with a default price. With currency convertibility, there are other issues involved. An econometric model that forecasts events is a better approach to model this type of risk.
A common misconception is that convertibility protection should be priced like a foreign exchange knock-in option. Actually it is nothing like a foreign exchange knock-in option. Knock-in options have an exchange rate trigger. That is not the case for convertibility options. The convertibility option only gives the protection buyer the right to convert at the then current foreign exchange spot rate. It is similar to a contingent option to exchange at the spot rate where the trigger is a government or Central Bank action.
Other market professionals sometimes describe convertibility options as a sovereign linked currency swaption. This is a false analogy. This isn't a swaption. It is similar to, but different from a sovereign default option.
There are important differences between default options and convertibility options. These differences are currently reflected in the fact that convertibility options usually trade at a premium to default options. But that needn't always be the case. The spread between a default option and a convertibility option may contract as a default event approaches, if there is certainty that both events will occur simultaneously. If a default on sovereign debt seems imminent, but convertibility of currency exists and seems continuing--albeit at a probable depreciated value--the convertibility option will trade below the default option.
A further complication is the lack of understanding of convertibility issues which keeps many potential market makers out of this arena. Supply and demand factors seem to dominate the pricing of convertibility protection. Some banks are booking convertibility premiums in the banking book, which they do not mark to market. They view the sale of convertibility protection as the functional equivalent of a loan.
Mark To Market
Other banks are not prepared to do this and must grapple with a difficult mark-to-market decision. This is an opaque market. There is no economic measure other than what the dealers quote and it's usually a one way market. One can poll dealers for a mark-to-market. One can also develop a model. As of this writing, I am not aware of anyone who has developed a model for marking convertibility protection to market, which reflects market levels.
Nonetheless, one can try reasonable justifications to defend a mark-to-market price. For instance, one might swap local currency domestically traded sovereign debt to dollars and then compare it with euro dollar debt for the same sovereign. That sounds easy, but there may not be a well developed swap market in the currency. A further difficulty is that many countries do not have the depth and breadth to their debt markets to enable such a comparison.
In the late 1990s, the Bank of Boston bought Brazilian convertibility protection. The bank was comfortable with Brazilian risk, but compared to its overall balance sheet, Brazilian risk was a high percentage relative to other banks. Its Brazilian risk rivaled that of Citibank. The key to good portfolio management is diversification, and Bank of Boston was over concentrated with Brazil risk. It attempted to reduce this risk.
Bank of Boston issued U.S. dollar denominated one year maturity CD's in Brazil which traded at LIBOR plus 80 basis points to LIBOR plus 100bps. The CD's had convertibility risk. If Bank of Boston was unable to pay in dollars because of convertibility restrictions, the CD's would pay off in Brazilian Reals. Meanwhile, Bank of Boston's U.S. issued dollar denominated CD's traded at LIBOR flat. This would imply that convertibility should have traded at 80 to 100 basis points. But Bank of Boston paid 115 basis points for this protection. The price reflected the fact that it was difficult for Bank of Boston to find counterparties. This suggests that the mark-to-market price for currency convertibility protection should include a liquidity premium.
Alternatives
The above method for finding a market clearing price for convertibility protection actually works--subject to assumptions about the liquidity premium--and seems as defensible as relying on true market levels. Benchmark instruments don't exist in many markets, however, so an alternative method must be used. Unfortunately, all of the alternatives have strengths and weaknesses. The following is a summary of the ways one might mark-to-market convertibility options:
* Observe market levels--the market defines the price, of
course. The difficulty is that this is a one-sided market and
we may not be able to get accurate marks at all times.
Broker levels are wholly unreliable for this product.
Arbitrarily say that convertibility trades at some percentage
spread above credit-default protection, say 150%. This has
the appeal of being easy and consistent. Greater illiquidity,
supply and demand factors, and higher probability of event
risk seem to dictate that convertibility will trade wider, but
there are scenarios where it can trade inside the
credit-default spread. This would be when a default exists,
but deeply devalued currency can still be converted. Even
when this appears to work, this formula method is not true
across maturities.
* Access domestically traded securities in the local currency
and look at a Eurodollar sovereign issue in dollars. Swap the
domestically traded security to dollars and compare the
spread difference. This spread difference should reflect the
convertibility premium for the currency. This also means
that there must be priceable local instruments as well as a
Eurodollar issue we can price and reference in the maturity
in which we are interested. At the very least, we must be
able to build a convertibility curve. For many countries,
there is a dearth of local currency and Eurodollar debt,
which makes this method virtually impossible to implement.
* Find a dollar issue of a foreign bank, which collects local
deposits. This bank must convert local currency to dollars.
Compare that with where they issue in the U.S. market.
The difference is convertibility risk premium. Example: The
dolar denominated CD issued by a U.S. local bank branch
that collects local deposits trades at LIBOR plus 125bps.
The obligations of the U.S. bank itself trade at about
LIBOR plus 20bps for one year, so the implied
convertibility premium is about 105bps.
* Find a convertibility linked note and back out the
convertibility premium. This method works, but one may
not be able to find such a convenient issue.
* Find a Sovereign with convertibility risk and strip out the
convertibility risk. The hard dollar Sovereign level in the market reflects the Sovereign risk premium. The remaining
spread is what the market implies is the convertibility
premium. The challenge with this method is the
unavailability of suitable debt instruments for many
emerging market countries.
If one must mark convertibility options to market, the best solution may be to adopt a policy based on a pragmatic combination of a few of the above reasonable methodologies. The key is to continue to observe market levels reflected from primary counterparties, however, since supply and demand factors dominate this market. Reality checks are essential.
This week's Learning Curve was written by Janet Tavakoli, founder of Tavakoli Structured Finance.
The article is adapted from Tavakoli's book Credit Derivatives and Synthetic Structures, 2nd Edition, John Wiley & Sons, 2001.
