In a fascinating piece of statistical detective work published in the latest BIS Quarterly Review, three of its foremost staff members argue that the use of FX swaps and forwards is hiding a massive quantity of debt because current accounting conventions place it off  balance sheet. They estimate that, in respect of US dollar debt, this financial dark matter is worth some USD10.7 trillion. Their conclusion is that FX swaps and forwards should really be accounted for on the balance sheet like repo.

The BIS paper is well worth reading and, as one would expect from these three authors, is a masterclass on how to make the best use of available statistics. Unfortunately, their basic premise is flawed.

First, a technical criticism. The paper claims that currency swaps and FX swaps are the same. I spend a lot of time trying to explain to students that, just because these instruments are both called swaps, it does not mean they are the same. Currency swaps are not analogous to FX swaps. There are fundamental differences.

A currency swap is an exchange of a series of interest payments in different currencies throughout the life of the swap and an exchange of principal amounts in the same currencies at maturity. At the end, a currency swap leaves each party with the currency they bought through the swap. If they want to get back into their original currency, they will need to transact another FX deal. Consequently, a currency swap creates exchange rate risk (which is typically used to hedge existing exchange rate risk).

An FX swap, on the other hand, creates no exchange rate risk for its counterparties. This is because currencies are sold and repurchase at fixed spot and forward rates. In other words, when you sell one currency for another through an FX swap, you know from the start how much of your original currency you will get back. An FX swap is a liquidity management tool, not a risk management tool.

I fear the BIS paper is working on the basis of a common but fallacious definition of a currency swap, in which there is also an initial exchange of currencies. But this structure, which is indeed equivalent to an FX swap, is actual a combination of a currency swap and a spot deal, with the currency swap acting as a hedge. The combination is found with new bond issues, when the proceeds of the new issue need to be swapped. But where existing issues are being hedge, the spot deal is unnecessary. It is therefore not an essential component of a currency swap.

A correctly-defined currency swap is analogous, not to an FX swap, but to an outright FX forward, which creates the same exchange rate risk. Indeed, there was once a market called long-term FX (LTFX), which offered multi-year outright forwards as an alternative to currency swaps. Currency swaps and outright forwards differ only in the distribution of cash flows. In an outright forward, all cash is exchanged at maturity. In a currency swap, cash is exchanged throughout the life of the transaction. An outright forward is an exchange of two forward deposits or zero-coupon bonds in different currencies. A currency swap is like an exchange of two interesting-bearing bonds in different currencies.

However, to be frank, the misunderstanding about currency swaps and FX swaps is not fatal to the BIS paper. I raise it merely to set the record straight in case any of my students should read the paper.

The basic premise of the BIS paper is that currency swaps and FX forwards are incorrectly classified as derivatives, which inappropriately shifts them off balance sheet, whereas they are “functionally equivalent to borrowing and lending in the cash market” and should similarly be captured on the balance sheet.

To illustrate the point, the paper offers three alternative ways of funding the purchase of a foreign currency security without running currency risk:

• Buy the foreign currency spot, use it to purchase the security and hedge the currency risk by selling the proceeds of the security forward.
• Use an FX swap in which the spot leg exchanges domestic for foreign currency to buy the security and the forward leg converts the proceeds of the security back into domestic currency.
• Fund the security by borrowing foreign currency through the repo market.

The first two transactions are currently accounted for simply as a substitution of a domestic currency asset by a foreign currency asset, with no gross change in the size of the balance sheet, which means the transactions are off balance sheet. The third transaction, the repo, sees the security bought by the investor and repoed out as collateral to raise funding, with the security remaining on the balance sheet of the investor, where it is joined by the cash proceeds of the repo, resulting in a gross increase in the size of the balance sheet. (The security stays on the balance sheet of the investor because he continues to be exposed to the risk of the security and to earn return on it because of his contractual commitment under the repo to repurchase at a fixed price.)

The problem, according to the BIS paper, is that accounting rules allow FX swaps and forwards to be treated as derivatives. But derivatives normally involve a net settlement and not payments of gross principal amounts. The paper is correct in arguing that FX swaps and forwards are not derivatives, but incorrect about the appropriate accounting rules.

A derivative should not be defined merely as an instrument whose value depends on prices derived from other instruments. The financial market is an ecosystem in which the value of everything is ultimately interconnected with everything else. In other words, on the basis of how value is derived, all financial instruments have a claim to be derivatives. It is just that the derivation is easier to see with derivatives because of their explicit contracts.

The real definition of a derivative is an instrument which never pays principal, only profit or loss, ie net not gross settlement. This definition allows a meaningful distinction to be drawn between derivatives and the other class of off-balance sheet of financial instrument, forwards. Forwards pay principal amounts but are off balance sheet because these payments are due at maturity, whereas balance sheets measure only current assets and liabilities. This means that FX swaps and currency swaps are forwards, not derivatives (they have derivative versions in the form of NDFs).

It is important to distinguish forwards from derivatives, despite both types of instrument being off-balance sheet. For most of its life, the credit risk on a forward is a net replacement cost, the same as a derivative. But at maturity, the payment of principal amounts loom and expose the parties to gross settlement risk. In other words, forwards are riskier than derivatives even though they perform the same tasks. This can be seen by comparing the risk profiles of a currency swap and a true derivative such as an interest rate swap. The risk on the currency swap rises as it ages, describing a concave curve, to reach a peak at maturity that can be a multiple of the contract amount. In contrast, the risk on an interest rate swap rises as it ages to a much lower peak (typically below 5% of notional principal amount) before falling to zero at maturity as the settlement of interest payments reduces the impact of further interest rate divergence.

The fact that forwards are riskier than derivatives might seem to support the BIS argument for changing the accounting treatment of forwards by bringing them onto the balance sheet. The problem with that argument is that you would be asking a balance sheet to do things which is not designed to do. Balance sheets are supposed to measure current assets and liabilities. The appropriate way of measuring future assets and liabilities is discounted adjustments in the form of net replacement cost because that is the risk that is actually being taken. If the counterparty fails, the non-defaulting party can sell off the currency he has bought, which means the risk of a net loss.

So, why is repo accounted for differently to FX forwards? The paper says, “in each case, the investor’s economic position is much the same”. I disagree. FX forwards do not increase leverage. In the case of outright forwards and (correctly-defined) currency swaps, you permanently lose one currency in exchange for another. Where’s the hidden debt here?

The case of FX swaps may appear more complicated, because they are analogous in structure and apparent purpose to repos in that they both consist of spot and forward legs at fixed prices and are both used for collateralized lending and borrowing. And in practice, these two transactions are often substitutes (the thin repo market in a number of Asian markets is partly due to competition from the FX swap market, which has the advantage of established liquidity and use of which avoids the cost of collateralization with securities). However, I would argue that an FX swap does not increase leverage, as parties lose the purchasing power of the currency they sell. Their purchasing power does not increase. Again, where’s the debt?[1]

In contrast, in a repo, the seller (cash borrower) converts a security to cash, while keeping the risk/return on the security. The cash increases his purchasing power, which allows him to leverage his position by buying another security.

Of course, things can go wrong with FX swaps. The other party can default, leaving the non-defaulting party with the wrong currency. The sale of that currency for the original can incur loses. However, such loses will be marginal.

Interestingly, if accounting rules do change, they are likely to go in the opposite direction to the BIS suggestion. In 2011, the IASB proposed to account for repo like a derivative and some countries do so. This involves derecognizing the collateral and removing it from the balance sheet but adding the replacement cost of the repurchase leg. That proposal was rejected by the market as likely to inject unwelcome volatility into balance sheets.

[1]  And the argument in the BIS paper that an FX swap is like an outright forward once the spot leg has settled ignores this does not erase the risk created by the spot leg and hedged by the forward leg.