Garnham and Tett’s large article last week on the risks of the carry trade – or perhaps the absence of risk, as they hint the big carry traders are now insured v. a surge in yen/ dollar volatility (aside: but who is selling the insurance?) – raises a topic that has interested me for a while. The growing extra-territorial uses of certain currencies. This is sometimes called the "internationalization of a currency."
How to start?
Back in the old days, Japanese households saved in yen, and their yen were used to finance yen-denominated domestic mortgages and yen-denominated loans to Japanese business. Maybe some yen were lent out to Japanese firms looking to finance investment abroad or to emerging markets governments looking for financing (Samurai bonds), but the sums were pretty small.
Japanese savers didn’t generally hold their financial assets in currencies other than the yen. New Zealand banks didn't finance themselves by borrowing from Japanmese households. And households in say Latvia didn’t generally borrow in yen to finance the purchase of a home. That seems to be changing, and fast.
Now, you might say, back in the old days a lot of Latin Americans (and others) preferred to save in dollars than in their local currency, and either had dollar bank accounts in Miami (or Panama or Uruguay) or dollar-denominated deposits in Argentina or Peru. And lots of governments borrowed in dollars as well – whether by issuing an international bond in dollars or by issuing dollar denominated domestic debt. Ricardo Hausmann famously called this “original sin” (he thought some countries were born unable to borrow in their own currency) others prefer liability dollarization.
Or put, differently, the dollar has been an international currency for a long-time.
But the use of the dollar in say Latin America is in a sense different than Japanese households putting their savings into New Zealand dollars. Latins wanted to hold dollars even though dollar accounts generally paid a lower interest rate than local currency accounts. They were looking for safety, not yield.
Of course, there are examples of households taking on a bit of currency risk to get a bit more yield in the past as well. While looking for articles for this post, I discovered European banks sold a fair number of bonds denominated in Australian dollars to their retail clients in the 1980s.
But the scale of these kinds of trades seems to be growing. A fairly large number of households in Japan are looking for a bit more yield, even if it means less safety.
And conversely, households in Latvia (and Hungary) are looking for lower interest rates on mortgages even if it means more risk.
I guess that isn’t all that different from the past either – banks in Thailand famously thought borrowing in dollars was cheaper than borrowing in baht before the 1997 crisis, back when the baht was tied to the dollar.
In the case of Latvian yen mortgages, though, the yen/ euro isn’t fixed. More importantly, Latvian households, not banks, are taking the currency risk.
More generally, modern finance makes it possible – even easy -- for say a bank in Latvia to finance its local mortgage lending with Japanese deposits, not local deposits. It either borrows the yen it needs directly from Japanese banks, or, more likely swaps the euros from its euro deposits with a Japanese bank that has yen. Rather than financing local mortgages, Japanese saving can finance Latvians mortgages – with the currency risk shifted to the Latvians.
Conversely, a bunch of New Zealand banks seeming have discovered that it is easier to finance their lending not with New Zealand’s own savings, but by issuing kiwi denominated bonds in Japan (this presentation is a bit dated, but it provides a nice summary of growth in the uridashi market). The cheapest source of New Zealand dollar financing hapens to be households in a country where no one uses the New Zealand dollar for day to day transactions.
I learned a bit about this kind of thing while doing some work on Turkey a while back. The Turkish banks have lots of dollar deposits -- a legacy of Turkey's history of monetary instability. Short-term rates on lira in Turkey were also higher than long-term rates – which made short-term lira deposits an unattractive source of financing for long-term lending to households. Moreover, short-term deposits aren’t the best match for longer-term lending.
One solution: European banks issued long-term lira denominated bonds to European households looking for a bit of carry. The European banks then basically lent the lira they raised to the Turkish banking system, though the transaction would typically be structured as a swap (the Turkish banks got lira, the European banks got dollars – which could be swapped into euros). In effect, European households, not Turkish households, were the cheapest source of long-term financing for the Turkey. At least that was the case before the lira mini-crisis in May 2006. Current lira rates have put a damper in the growth of lira-denominated mortgages -- though there seems to be plenty of demand for short-term lira t-bills.
I understand why Japanese households like kiwi-denominated bonds. I even understand why Europeans were tempted to buy Turkish lira denominated bonds. There is nothing like a high coupon. I also understand why Hungarians like to borrow in Swiss francs and Estonians like to borrow in yen. Ask any macro hedge fund ….
What I initially didn’t quite understand is why European and Asian banks seem so keen to issue in say New Zealand dollars when kiwi interest rates are so much higher than interest rates in Europe or Asia. Garnham and Tett in the FT:
“the amount of bonds denominated in New Zealand dollars by European and Asian issuers has almost quadrupled in the past couple of years to record highs. This NZ$55bn (US$38bn, £19bn, €29bn) mountain of so-called “eurokiwi” and “uridashi” bonds towers over the country’s NZ$39bn gross domestic product – a pattern that is unusual in global markets. “
The amount of Icelandic krona bonds outstanding (Glacier bonds) is far smaller –but it is also growing fast to satisfy the demands created by carry traders. Here, the same basic question applies with even greater force. Why would a European bank opt to pay high Icelandic interest rates?
The answer, I think, is that the banks who raise kiwi or Icelandic krona swap the kiwi or krona that they have raised with the local banks. That certainly is the case for New Zealand's banks -- well known Japanese banks and securities houses issue bonds in New Zealand dollars and then swap the New Zealand dollars they have raised from their retail clients with New Zealand banks. The New Zealand banks finance the swap with dollars or some other currency that the New Zealand banks can easily borrow abroad (see this article in the bulletin of the Reserve Bank of New Zealand).
I bet the same applies with Iceland. Iceland's banks presumably borrow in dollars or euros abroad. They then swap their dollars or euros for the krona the European banks have raised in Europe. That is just a guess though -- one supported by some elliptical references in the reports put out by various Icelandic banks (see p. 5 of this Landsbanki report; Kaupthing has a nice report on the recent expansion of the Glacier bond market, but is silent on the swaps) but still fundamentally an informed guess.
And at this stage, I don’t really have a well formed opinion on whether or not all this cross border activity in the currencies of small high-yielding countries is a good thing or a bad thing.
Two potential concerns jump out at me. One is that financial technology has opened up new opportunities to borrow which will be overused and abused. The other is that the amount of currency risk various actors in the global economy are taking on– not necessarily just classic financial intermediaries – is rising.
I am less worried that international borrowers are tapping Japanese savings – whether yen savings to finance yen mortgages in Estonia or kiwi savings to finance lending in New Zealand – than that so much Japanese savings seems to be financing residential real estate and household credit. External debt though is still external debt. It utlimately has to be repaid out of future export revenues. Financing new houses -- or an increase in the value of the existing housing stock -- doesn't obviously generate future export receipts.
Then again, New Zealand banks using uridashi and swaps to tap Japanese savings to finance residential lending in New Zealand aren't doing anything conceptually different than US lenders tapping Chinese savings -- whether through Agency bonds or "private" MBS -- to finance US mortgages. In the first instance, Japanese savers take the currency risk; in the second, the PBoC does. The PBoC is willing to lend at a lower rate, but the basic issue is the same: does it make sense to take on large amounts of external debt to finance investment in a not-all-that tradable sector of the economy?