Josh Younger is a lecturer at Columbia Law School and a former JPMorgan analyst. Brad Setser is a senior fellow at the Council on Foreign Relations, and a former US Treasury official.
As FT Alphaville recently detailed, Taiwan is an interesting place, having amassed an absolutely massive treasure trove of US dollars. We now want to explore what is happening with those dollars — and the weird phenomena it can cause.
To quickly recap the last post, Taiwan has stealthily become the fifth largest foreign creditor in the world, with a net international investment position (external claims minus liabilities) of $1.7tn. That puts it in league with China, Germany, and Japan, despite being an economic minnow compared to those countries.
Among developed countries, only Hong Kong, which pegs its currency to the US dollar, and Norway, with its massive oil reserves, come close to matching Taiwan in net foreign assets relative to the size of their economies.
Taiwan’s foreign assets have also ended up in an unusual place — insurance companies. Taiwan has one of the largest life insurance industry in the world compared to its economy and population. At more than $1.1tn in total assets as of the end of 2023, that’s more than 140 per cent of Taiwan’s annual economic output, and works out to $47,000 per person.
How did the Taiwanese life insurance industry get so enormous? They offered policyholders attractive alternatives to their bank deposits — higher yields but many of the same liquidity features.
Perhaps most strikingly, these insurers are not only massive but also running an equally massive currency mismatch. As the last post pointed out, two-thirds of Taiwan’s insurance industry assets are overseas, and predominantly denominated in dollars, while the opposite is true for their liabilities, roughly 80 per cent of which are denominated in local currency.
In other words, Taiwanese insurers invest the proceeds of their premium income predominantly in foreign currency assets while making promises to redeem those policies predominantly in local currency. The gap can be bridged using foreign exchange derivatives or, increasingly, not at all.
The consequence of this transformation on a massive scale are important. We discuss them in the post referenced above. But the question we are after here is: what exactly are they doing with those dollars?
Call me maybe?
Taiwanese insurers are classic yield-seekers — looking for the highest returns on the longest maturity assets they can track down. In the years leading up to the 2008 crisis, they went after subprime mortgage-related securitisations and other complex financial products. After those blew up, they went to town on callable bonds.
Callable bonds are similar to more vanilla instruments in most respects but, critically, they give the issuer the right to call the bond at face value prior to its legal maturity. That’s the equivalent of issuers borrowing money and buying a call option on their own debt at the same time.
When rates go up, and the bonds in question are trading at a discount to their face value, this option is mostly worthless (who wants to buy something worth 90 cents for $1?). But when rates go down, and those bonds are trading at a premium, that option can turn out to be quite valuable. The investor is in practice short that option, and that means potentially substantial losses. To compensate themselves for taking this risk, the buyers of callable debt demand a higher yield.
The complexities of modelling and managing the risks embedded in callable bonds means they rarely trade and are difficult to value. But for life insurance companies trying to fund expensive liabilities, it’s easier to focus on the higher yield than the relative illiquidity.
These investments became wildly popular in Taiwan, and the industry clamour for more was rewarded with the creation of “Formosa bonds”.
Formosa bonds are listed in Taiwan but issued by foreign companies and denominated in US dollar. Their name — a 16th century-vintage Portuguese term for Taiwan meaning “beautiful island” — was chosen via public referendum, having edged out alternatives like “High-Tech Island Bond,” “Best Bond,” and “New Dragon Bond.”
At this point there are more than $100bn such bonds outstanding, the vast majority of which were issued between 2014 and 2018.
That’s a lot of corporate debt held by a few large investors and in a very illiquid format. But what does all of this have to do with mortgages?
it’s important to recognise that issuing Formosa bonds is often not really about borrowing. It’s a trade. Large, sophisticated international banks and other financial institutions recognised that captive demand for Formosa bonds led them to trade at elevated prices; that is, the issuer was able to buy call options via issuance of Formosa bonds for less than the cost to do so directly via derivatives.
Whenever such a wedge appears, it is profitable to buy the actual call by issuing Formosa bonds and sell a synthetic call constructed with options on interest rates. And that’s precisely what many dealers did. As a result, the options-selling by Taiwanese life insurance companies made its way rather quickly into the market writ large.
That trade was initially quite profitable. But, as dealers competed with each other for market share, the resulting flood of options selling depressed prices for synthetic calls as well. In other words, the price for those options implied a lower and lower level of volatility over time.
So, what does this have to do with mortgages? Well, American mortgage borrowers are also buying an option — an option to refinance their debt if rates decline. Just like Formosa issuers, that results in a somewhat higher rate on mortgage debt than other comparable instruments.
Mortgage lenders choose what rate to offer on a new loan in part by pricing in the value of the refinancing option as inferred from interest rate derivative markets. All else equal, the cheaper those options are as standalone derivative contracts, the lower the rate mortgage lenders are willing to offer.
Putting it all together, Taiwanese insurance companies had written a lot of policies. But they had to offer high guaranteed rates of return to do so. To hit those hurdles, they looked offshore and aggressively accumulating us dollar-denominated callable debt. Those callable bond issuers in turn used derivatives to carve up their liabilities into their component parts and sold off the embedded options at a profit.
That put downward pressure on long-dated interest rate option prices which fed through into mortgage rates and, ultimately, to us homeowners.
When Taiwan sneezes, US homebuyers catch a cold
What happens when this vol tap is turned off? That’s precisely what happened in 2022.
As rates rose rapidly, callable bond prices declined. Taiwanese life insurers — which by this point were nursing substantial losses — lost their appetite for callable debt. Although there’s been a trickle of issuance, the stock of such debt outstanding is basically unchanged over the past two years.
As the supply of callable bonds was choked off, so too was the supply of volatility. No longer did dealers “rinse and repeat” by issuing Formosa bonds and selling them for scrap. That led prices on long-dated interest rate options to rise steadily, essentially reversing the impact on mortgage rates of artificially low prices.
Depending on how you measure it, the increase in implied volatility likely raised American mortgage rates by between one-quarter to three-eighths of a point.
That might not move the needle on a new home purchase or refinancing on its own. But it’s definitely something you’d notice when locking a rate. And, when viewed in the context of the $14tn market for US residential mortgages, the sums involved are far from trivial.
Perhaps more importantly, the commingling of the interests of Taiwanese insurance companies and US homebuyers through the mechanism of derivatives markets is a classic case of the interconnectedness and complexity that has come to define modern financial markets.
