Recent Posts

Tuesday, July 2, 2024

Primer: Currency Risks For Banks

Yet another unedited section from my banking primer manuscript. My feeling is that this section is packing in too much information, and might be trimmed. The technical appendix may be too technical, but I will look at that later.

Although the major banks have global operations and currency trading is a massive financial market, this book largely ignores the complications created by banks operating in multiple currencies. The first reason is that the author has no useful experience in that area. The second is that currency risk is not a significant source of risk for well-managed banks. If a sensible bank is operating in two currencies, it is best understood as two banks operating in one currency, with one bank acting as parent. (From a regulatory perspective, the fact that the home base is in a different jurisdiction matters, but this text is not delving deep enough into details for that to be a concern.)

Currency risk is defined as the risk of generating losses based on changes to the exchange rate between two currencies (i.e., the price of a currency in terms of another). Currency risk is not the risk associated with a bank relying on transferring funding from one currency to another. This cross-currency financing risk was a major factor in the 2008 Financial Crisis, but it is not “currency risk” as it understood from a risk management perspective. This distinction matters because there is considerable folklore about banks running currency risks, and the people spreading that folklore make the mistake of treating the cross-currency financing risk as being a currency risk.

Currency Risk

Currencies are extremely annoying because there are two ways of quoting the price. The standard convention is to write how many units of one currency you get in exchange for one unit of another. For example, we can say that it takes C$1.10 (C$ is the Canadian dollar, or CAD using the ISO code) to buy U$1 (U.S. dollar, or USD). Alternatively, we can quote the exchange rate as U$0.9091 buys C$1. Things get complicated when the exchange rate moves. Imagine that the Canadian dollar weakens versus the U.S. dollar. The new exchange rate is C$1.20 per U$1, or U$0.833 per C$1. Using the first convention, the quoted price went up, and in the second, it went down. (This is why I write “weaken/strengthen” when discussing exchange rate moves, since “up/down” are ambiguous.)

For each currency pair, there is a standard direction for quoting it, and you will be laughed at for quoting it the other way around. Which currency ends up being the “base currency” depends upon the historical circumstances of when it started being traded (e.g., was London or New York the main financial centre). This causes considerable grief for anyone writing code that is supposed to deal with cross-currency calculations as well as display output in a manner that does not cause traders to have one of their episodes.

Measuring currency risk is relatively straightforward. You pick a base currency for your portfolio (including liabilities), and then see what the gain/loss is on the portfolio if the quoted exchange rate moves by 1% in one direction. (In the absence of unusual currency derivatives, the gain/loss for a 1% move in the other direction is going to be the negative of this.) For a bank, the “portfolio” in the definition would be the bank’s balance sheet, but it may also be the balance sheet of operating units. For a fund, the “liability” may be fund’s benchmark, and not accounting liabilities on a balance sheet. The benchmark matters – you would expect a different currency exposure for a fund where it has 50% weighting in two currencies than one that has 100% weighting in one currency.

Like an interest rate shock, you are likely going to be interested in the gains/losses for relatively large exchange rate movements, and you need to come up with risk scenarios.

Separation of Church and State

If you are looking at a bond portfolio with competent risk management after the mid-1990s, it is very easy to guess the foreign exchange risk: zero (0). The way fund management firms operate is that split the accounting into internal portfolios where the currency risk is always zero for fixed income funds. Any currency risks the firm takes is accounted for in a separate currency overlay portfolio. For example, if a firm just had one bond fund that is allowed to buy foreign currency bonds, the one fund that investors buy units in is accounted for internally as two portfolios: the pure fixed income portfolio, and a currency overlay portfolio. In practice, most fund companies have multiple funds, and so there might be one currency overlay portfolio sitting on top of the separate bond portfolios.

The justification for this separation was justified by horrific examples that arose in earlier decades (mainly the 1980s). Portfolio managers who were supposed to be bond gurus blew up their portfolios making currency bets. Instead, portfolio managers are supposed to stay within their domain expertise. (Canadian firms are doctrinaire about this principle and refer to this as “separation of church and state.” The relative illiquidity of the Canadian bond market forces large bond funds to jump in an out of U.S. dollar markets to deal with large inflows, and so currency risk is a constant concern.) It is much easier to understand a firm’s investment strategies when the fixed income experts lose money on interest rate risk, and the foreign exchange (“forex” or FX) experts lose money on currency risk.

This separation does not imply that no foreign exchange risk is taken by the investment firm. Instead, it has the effect of moving the foreign exchange risk into an overlay portfolio that is generally managed by a different team. The firm will have a global risk budget, and that risk budget has to be split between the currency overlay team and the bond team. Since potential outperformance is proportional to the risk budget, fights for risk budget allocations are extremely bitter. The end result is that foreign exchange risk (when compared to benchmarks) ends up being relatively small versus the size of the fixed income balance sheet.

Banks

Well-run banks use the same risk management software and principles as investment funds. They also observe the separation of the church and state. To the extent that foreign exchange risks are taken, it is done by a team with a limited risk budget. Although this risk budget might produce losses that look scary to most onlookers, they are generally insignificant versus the size of the bank’s balance sheet. Guessing the direction of currencies is not a core business strength of a bank, and so the risk budget is only going to be sized relative to the importance of the capital markets trading group (including foreign exchange trading).

Although a bank can nudge around its foreign exchange risk exposure using derivatives, banks need to rely on matching assets and liabilities to manage rely. That is, borrowing in the currency that matches the currency of its assets. If the bank has subsidiaries operating in foreign currency zones, those subsidiaries are ideally self-financing.

Cross-Currency Financing

A typical situation for a global bank is that it has a large presence in its home country. It has a large bank branch base that draws in cheap deposit financing, and local investors are familiar with the bonds and so they can issue bonds at tight spreads. Its foreign currency operations are there mainly to conduct capital markets activity and support the overseas operations of corporations headquartered in the bank’s home market. It is less recognised and has a lesser ability to attract deposits, so its financing spreads are wider. This creates the desire to export the cheap local funding to its foreign subsidiary somehow. The depth of American capital markets has the result that there are a lot of non-U.S. banks with American subsidiaries in this situation.

What does not work is getting funding in the home currency and buying the foreign currency in the spot currency market (without taking exchange rate risk).

For example, imagine that a German bank wanted to buy a $100 million 2-year bond in the United States, and the dollar (USD) euro (EUR) exchange rate is conveniently 1:1. It could issue €100 million in bonds in Germany, and then exchange the EUR into $100 million. Although it has managed to buy the bonds, it is then exposed to exchange rate changes. If the USD falls so that 0.90 EUR is worth 1 USD, then the USD bond par value drops to being worth €90 million, which is €10 million less than the par value of the bond issued to finance the position.

The way to exchange funding across currencies is to use derivatives to hedge the currency risk. The cross-currency basis swap market is the core of the hedging system. Since these swaps can be confusing, their description has been moved to the technical appendix below. The important thing to observe is that when banks use these instruments properly, they are hedged against movements in the exchange rate.

Economic Folklore

In economic and market commentary, one often runs into market folklore about banks running large currency mismatches: they borrow in one currency and lend in another. This does happen – in countries that feature pegged or managed currencies.

If everyone believes that a currency peg will hold, there is apparently no need to hedge currency risk – the exchange is not supposed to move. People will just look at the lower interest rates on the senior foreign currency (that is the target of the peg) and borrow in that one. However, one a peg breaks, currencies typically have violent exchange rate moves. This creates extremely large market value differences between unmatched assets and liabilities that will overwhelm bank’s limited regulatory capital.

A country can either put into place tough regulations on currency mismatches or do whatever it takes to prevent their currency peg from breaking (which usually requires extremely painful austerity and deep depressions). Allowing mismatches to occur and then letting the exchange rate move is raw incompetence.

Emerging market commentators resist the implication that banks and regulators in their markets are incompetent. Instead, they tend to argue that bank currency mismatches are a fact of nature. This has the result that emerging market commentators are a regular source of incorrect predictions of doom for developed markets.

Post-Keynesians who have a deep ideological attachment to fixed exchange rates are another constituency pushing the “banks have to borrow in foreign currencies” story. They typically end up quoting the effects in models where bank balance sheets are how international transactions are financially intermediated.

It is relatively straightforward to verify that developed market banks do not run major currency risks. If you look at historical exchange rates, you can see they have moved a lot without any associated bank failures. (Euro exit risk is frightening because it is a new currency being created when a country leaves, and there is no good way of hedging that.)

The capital markets (including foreign direct investment as well as foreign official currency reserves) are how capital flows are intermediated, not banks. Developed equity markets are deep, and very little of equity exposure is currency hedged. Foreign official reserves represent a primary source of unhedged fixed income exposures, institutional investors in fixed income usually do so to hedge local currency actuarial liabilities and therefore hedge exposure. (Actuarial liabilities mainly include the liabilities of insurance firms and pension funds.) Individuals are much less likely to hedge, and this will matter for countries without a highly developed financial sector that creates large fixed income portfolios that are managed against actuarial liabilities.

Borrowers’ Currency Risk

Lending to clients whose cash inflows are denominated in a different currency than the loan is a stupid thing to do. In addition to the usual credit risks (impairment of cash flows due to economic changes, collateral value drop, etc.), the borrower may be unable to meet the payment obligations due to the currency value change. Once again, the usual culprit inspiring such behaviour are currencies where the exchange rate is managed.

A key qualifier in the statement about lending is the reference to cash flows. Exporters may generate most of their revenue in a foreign currency, while its expenses may be mainly in the local currency (e.g., wages, leases, etc.). Borrowing in the currency the revenues are denominated in reduces the overall risk of the firm. This is a common situation in Canada, as the natural trade flows are North-South across the Canada-U.S. border, as opposed to the East-West flows between Canadian provinces. (Trade barriers may be lower as well because of the Free Trade Agreement.) As such, foreign currency borrowing by non-financial firms may not be a true source of risk. (Financial intermediaries still have to match their assets/liabilities.)

Concluding Remarks

Banks running currency risks are mainly the domain of historical analysis and emerging markets. In the absence of exotic foreign exchange derivatives, currency risk is an easy risk to monitor and hedge. The cross-currency basis swap market is the glue that ties funding markets together.

Technical Appendix: Cross-Currency Basis Swaps

Banks can use cross-currency basis swaps to exchange funding between currencies. (Currency swaps are similar to cross-currency basis swaps but feature a fixed interest rate leg which is less attractive as it makes the structure asymmetric. Currency swaps are more well known, since central banks enter into currency swaps with each other to manage funding risks.)

A cross-currency basis swap is economically equivalent to the two counter-parties exchange loans to each other at the current exchange rate, pay each other a floating rate during the life of the loans, and then paying off the principal of the loans at the original exchange rate. The exchange of principal values is unusual for a derivatives transaction, and they generate more potential counter-party risk (that is managed in a few ways that this text skips over for simplicity).

For example, imagine that the current exchange rate is 1.1 USD per 1 EUR, and a German bank wants to get its hands on $110 million for five years. It will enter into a basis swap that has the following cash flows.

  • The German bank lends the American bank €100 million, and the American bank will pay the EUR floating rate on this amount for 5 years.

  • The American bank lends the German bank $110 million, and the German bank will pay the USD floating rate on this amount for 5 years.

  • On top of these floating exchange rate payments, there is a fixed spread attached to one floating leg payment. This spread is fixed at the inception of the swap, and this is the “price” for the swap quoted in the market (since every other payment property is fixed). For example, the German bank may pay an additional 10 basis points over the USD reference rate.

  • At the end of the 5 years, the loans are repaid at the original exchange rate. That is, the American bank must pay the German bank €100 million, and the German bank must pay back $110 million.

Since the exchange rate on the final payments is fixed, the future level of the traded exchange rate has no effect on the cash flows. That is even if the exchange rate moved violently, the exchange of principal amounts is unaffected.

The reader can then look at the above cash flow structure to realise that the German bank could buy a $110 million 5-year bond, enter into the above swap, and finance the position by issuing a €100 million bond. The maturity of the USD bond matches its USD principal repayment obligation under the swap, and the swap proceeds allow it to repay the EUR bond. The net profit/loss on the position would be based on the comparison of cash flows on the bonds and the swap. (If the bonds were floating rate, the profits would be based on the spread differentials and the fixed spread in the swap structure.)

In summary, the cross-currency basis swap market allows participants to trade floating rate funding between each other without taking currency risk across the lifetime of the swap. (If you dig into the structure of the example with the bond, we see that once the swap is entered into, it will have the opposite currency risk sensitivity to the bond purchased by the proceeds of the swap.) The cross-currency basis swap market is the mother’s milk of modern global financial capitalism.

The flip side of its importance is that strains in the cross-currency basis swap market links financing across all the major developed markets. Fears about the cross-currency basis swap represented the peak moment of fear in the 2008 Financial Crisis.

One of the ironies of the 2008 Financial Crisis was that aspects of it ended up being the exact opposite scenario of some the doom-monger economists who claim to have predicted it. The consensus economist worry was that the United States would face a financial crisis because its current account and fiscal deficits were (allegedly) unsustainable. Foreign financing would be withdrawn, and everybody would be sad. What happened is that pretty much everybody – particularly European banks – loaded up on toxic garbage securities that Wall Street financial engineers unloaded on them. These purchases were first financed by issuing USD-denominated paper in the American money markets as well as cross-currency basis swaps. Once the dismal outlook for those toxic assets became clear, American money market investors withdrew their funding from foreign banks. The foreigners tried to bridge the gap with more cross-currency basis swaps, but that market started to buckle. The authorities were faced with the prospect of a total meltdown of the global financial system, or the Fed had to engineer a bailout via sending USD to foreign central banks using currency swaps. These swaps gave the foreign central banks USD that they could provide to their local banks.

Finally, there are a few other derivatives hedging strategies that can achieve the same effect as a cross-currency basis swap – currency forwards and futures, and currency swaps (with fixed interest rate legs). However, those markets are junior markets with respect to the senior cross-currency basis swap market. If you dig into currency forward pricing, one will encounter “the basis,” which is the spread between the market forward price and the one predicted by interest rate differentials. That “basis” is determined by the quoted spread on cross-currency basis swaps.

References and Further Reading

  • This section is almost entirely the result of my experiences in finance. Cross-currency basis swaps is one of the markets that I analysed. Any decent book on risk management should cover the fundamentals of currency risk management. The challenge of currency risk management is that foreign exchange trading is that the risk management is easy, and currency traders take advantage of this to focus on trading the most complex instruments possible.

  • Despite its importance, it is hard to find good primers on the cross-currency basis swap market. It is easy to find descriptions of the cash flows, but such treatments rarely discuss why the market is important. It is the domain of very few large players who keep a low profile.

  • Almost all discussion of how foreign exchange rates relate to interest rate differentials is terrible and should be ritually burned. The only trustworthy sources are the ones describing how arbitrage works for currency forwards, but unless the cross-currency basis swap market is drawn in, it is only an approximation that financial academics use. Otherwise, the foreign exchange market is a market, and it is hard to profitably trade it in a consistent fashion. Currency traders do tend to be trend-followers (when compared to rates markets), but trend following only works until it doesn’t.

Email subscription: Go to https://bondeconomics.substack.com/ 

(c) Brian Romanchuk 2024

No comments:

Post a Comment

Note: Posts are manually moderated, with a varying delay. Some disappear.

The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.

Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.