The cross-currency basis may be one of the most poorly understood concepts in financial markets. There are two main reasons for this: the first is that the very notion of a cross-currency basis requires the breakdown of one of the fundamental tenets of finance theory: covered interest-rate parity; the second is that the cross-currency basis requires a grasp of both interest rates and foreign exchange – two frequently siloed areas of financial markets.
As such, although the degree of understanding among even seasoned market participants can often be low, the cross-currency basis offers a fascinating window into powerful dynamics that shape global markets.
THE ‘FUDGE FACTOR’
So what is the cross currency basis? Introductory finance theory suggests that forward foreign exchange (FX) rates between two currencies (i.e. the price at which one currency can be bought or sold against another, for settlement at a specified date in the future) should be exactly and easily calculable from the ‘spot’ – or current – FX rate and the respective interest rates in each currency (running to the settlement date of the forward). This is known as covered interest rate parity.
This was generally true for forward FX markets prior to 2007, but has since ceased to be the case for many global currency pairs. Indeed, it is currently the case that for the US dollar and yen, the interest rates offered in each currency are so bad an explainer of the three-month FX forward rate that you need to adjust one of the rates by almost 80 basis points to arrive at the observed market level.
It is from this mismatch that the concept of the cross-currency basis emerges. For those inclined to think along the lines of traditional finance textbooks, if you were attempting to reconcile the observed level of FX forwards with the theoretical level, taking spot FX and interest rates as inputs, you would need to adjust one of the interest-rate streams by the cross-currency basis to arrive at the observed level of FX forwards. In other words, the cross-currency basis is the ‘fudge factor’ that reconciles covered interest-rate parity theory with the real world.
It follows that a large arbitrage opportunity would appear to exist in highly liquid FX markets that should not exist for any extended period of time under the textbook definition. And yet in many cases this situation has persisted, and even intensified, for almost a decade (see chart of the USD/JPY three-month basis).
US DOLLARS AT A PREMIUM
While there are important exceptions to this rule, it has generally been the case that since 2008, global markets have assigned a premium to borrow US dollars in global currency and interest rate markets. This is most acute in the case of USD/JPY, and to a lesser extent in EUR/USD, but is also frequently observed across G-10 and emerging-market currencies. In short, if you were seeking to borrow simultaneously in US dollars and lending another currency, forward FX markets would typically ‘punish’ that demand by making it more expensive to borrow greenbacks than it should be given interest-rate differentials.
What actually causes the cross-currency basis to exist? A useful way of thinking about the drivers of a given cross-currency basis market is:
- What causes the basis to appear in the first instance (or, what initially causes deviation in a forward FX rate from its theoretical level)?
- What are the barriers in place that stop risk-takers from arbitraging the basis away?
On the first point, the classic driver of a non-zero cross-currency basis is the persistence of a certain type of cross-border bond issuance (i.e. where a bond is issued in a currency other than that of the issuer’s home market). The relative tightness of credit spreads in one currency versus another (due to the relative strength of demand for credit products between currency areas) causes large issuers to prioritise bond issuance in that currency at the expense of others.
In mechanical terms, an issuer will respond to strong demand for bonds in one currency by selling debt into that currency and then swapping the proceeds back to its desired liability currency. A classic example of this is so-called Samurai issuance, under which strong demand by Japanese investors for yield causes tightness in credit spreads in yen, which in turn leads to nimble global bond issuers (like banks, supranational agencies and large corporates) to issue a yen bond and use forward FX and longer-term cross-currency basis markets to ‘swap’ back their liability to US dollars, or their currency of choice.
The final step of hedging back to a home liability is what creates the non-zero cross-currency basis. This is because market participants who are unwilling to take the opposite side of such transactions have little option but to ‘punish’ repetitive behaviour by imposing on the trade an increasingly large spread above what interest rates would appear to require.
Another important driver of the cross-currency basis, which is of particular relevance today, is when a currency area’s banking system (in aggregate) opts to allocate to non-domestic assets. Japanese banks, flush with Japanese yen deposits, currently face a lack of demand from local corporates and households to borrow, relative to their own desire to lend. To fulfil this desire abroad, the banks use foreign exchange markets to lend yen and simultaneously borrow US dollars, which they use to provide loans and invest in offshore assets.
This process of lending in local currency to borrow in a non-local currency creates a cross-currency basis (in the absence of offsetting flow). Indeed, this dynamic reflects the same fundamental underlying driver as Samurai issuance: insufficient demand to borrow locally relative to the pool of assets available to lend.
Underscoring this point, a report by the Bank for International Settlements found an important relationship between the size of bank and institutional investors’ net foreign asset position and the level of global cross-currency basis. There are other significant drivers of the cross-currency basis that can affect day-to-day price movements, such as the relative degree of monetary stimulus between two currency areas, ‘crisis’-like funding issues and relative dynamics in domestic funding markets (such as the Libor versus overnight index swap basis). These, however, often share a common cause to the dynamics described above, and it is beyond the scope of this piece to explore them further.
LIMITS TO ARBITRAGE
On the second point, regarding arbitrage, it is worth noting that prior to the global financial crisis, even though there were also large flows in FX markets of a systematic and repetitive nature, the cross-currency basis rarely experienced the sustained deviation from zero that we have seen since.
There are many important reasons why limits to arbitrage exist in a stronger form today than in the past, the most important of which is regulatory change in the global banking system. This has both prioritised a reduction in balance-sheet size and penalised many forms of risk-taking activity, particularly those that are leverage-intensive and short term in nature.
Arbitraging the cross-currency basis in any meaningful size requires the simultaneous lending and borrowing of large balances of currencies, while taking on a large (derivative) notional exposure to interbank counterparties and other market players. And having done all of this, what may seem like large rewards to cross-currency basis watchers may not really be as appealing, in the context of the opportunity cost of the capital required to dedicate to the strategy.
Furthermore, for leveraged investors who are the traditional ‘police’ of such deviations, there are also drawbacks to seeking to profit from the eventual reversal of extreme moves in the cross currency basis. One is that so-called quarter- and year-end effects, whereby banks and other agents hoard cash ahead of reporting deadlines, mean periodic flare-ups in the basis – such as those seen at the end of September 2016 and December 2015 – appear to have become more common.
For a leveraged investor who needs to borrow the money that will be used to lend into basis markets, this can be a significant issue, as the cost of that borrowing periodically spikes (albeit for short periods). Uncertainty as to the extent to which this will happen – and to what degree – inhibits the willingness of leveraged investors to take part in such arbitrage opportunities.
POTENTIAL ALPHA SOURCE
But it remains the case that for cash-rich investors – particularly US dollar-rich investors such as asset managers, dollar-based banks and sovereign wealth funds – the presence of wide cross-currency basis represents an important opportunity and potential source of alpha. For example, one can use basis markets to obtain significant yield enhancement from high-quality, short-dated government bonds.
Indeed, a key part of the Old Mutual absolute return government bond team’s cash management process involves referencing the current levels of global cross-currency bases (as implied from short-dated FX forwards); choosing which G-10 government bonds to fund using repo; and deciding which bond purchases to fund through FX forward transactions (and indeed in which G-10 bond market to store the funds’ cash).
While we do not make direct leveraged investments on the cross-currency basis, we do use it to deepen our understanding of global markets and assess the relative value in the many ways that we can express macroeconomic views across the G-10 interest rate and foreign exchange universe.
* Note 1: For readers with a fixed income background – the cross-currency basis is also defined as the adjustment you need to make in order to equate the present value of one floating Libor stream in one currency equal to that of another. This is an equally big leap in terms of traditional finance theory, in the sense that anyone who has been introduced to interest rate swaps will be familiar with the idea that the float side of an interest rate swap (at Libor flat, i.e. with no spread) should be priced at par. And, therefore, the idea that paying one floating Libor stream (with no spread) in one currency versus receiving another floating Libor stream (with no spread) in another currency should be priced at anything but zero is alien. But this is, indeed, the case across global interest rate markets, whereby the Libor streams needs to be adjusted by the cross-currency basis in order to arrive at zero net present value.
* Note 2: How to track cross-currency basis. For anyone who wishes to track the history and evolution of global cross-currency bases for themselves, Bloomberg is an excellent source of data. JYBSC Curncy <go> is the three-month USD/JPY basis (likewise EUBSC Curncy and BPBSC Curncy track the same rate in EUR/USD and GBP/USD); and JYBS1 Curncy <go> is the one-year USD/JPY basis (EUBS1 / BPBS1 / ADBS1 for EUR/USD, GBP/USD, AUD/USD respectively). Note that as with many matters in fixed-income derivative markets, the way of expressing basis is not entirely intuitive: a premium to borrow dollars transmits itself in cross-currency basis markets as a negative basis. This is because cross-currency basis is expressed as the spread to the non-US side that would be required to equate the two currency streams, and therefore a negative basis applied to the non-US dollar side is analogous to a premium (or positive basis) to the US dollar side. This is just one of the foibles of markets with which people unfortunately need to become accustomed!