A New Impediment to Balance of Payments Adjustment: Underwater Bonds

A few years back, Silicon Valley Bank (and a few other regional banks) got into trouble because they held too many long-dated government bonds.

Government bonds are generally a safe investment; advanced economies who borrow in their own currencies don’t usually default on their own debt. But the market value of long-term bonds fluctuates with interest rates, and low-yielding bonds bought before COVID and during the first year of COVID fell in value when inflation took off and the Federal Reserve started raising interest rates.

A 10-year bond bought at par with a 2 percent coupon back in 2018 (or a coupon well below that in 2020) isn’t going to be worth its face value in the open market now. A coupon of 2 percent or so is just too low a rate on a bond that still has a few years to maturity.

The same is true for long-term Agency bonds (the underlying mortgages now won’t be refinanced, so the long really is a long-term bond) and long-term corporate bonds.

This, though, isn’t just a problem for U.S. regional banks.

A lot of institutional investors around the world (insurers, pension funds, etc.) bought a lot of long-dated U.S. bonds from 2012 on. This was a period when U.S. bonds didn’t yield a lot but still offered more yield than was generally on offer in their home countries.

Taiwan, as often is the case, is the best example. Taiwan’s life insurers bought a ton of bonds as their own regulators allowed them to increase their foreign bond holdings as a share of their total assets.

They were able to attract additional inflows by offering a higher yield than was otherwise available on safe investments in their home market.

Nearly all those bonds were denominated in dollars. 98 percent of the foreign bonds of the largest insurer are denominated in dollars; and over 90 percent of the bods for all the insurers are in dollars. Nearly all of those bonds are all underwater.

A rough-and-ready estimate based on the superb Bertaut-Judson data set (I did a cumulative loss from the start of 2020) suggests that Taiwan’s bond portfolio has lost 12 percent of its value in dollar terms over the last five years as a result of interest rate changes.

The central bank’s Treasury-heavy portfolio took smaller losses than the insurers’ Agency and corporate bond heavy book. Realistically, the insurers’ approximately $600 billion foreign bond portfolio has a current market value 10-15 percent below its book value (the insurers have around $700 billion in foreign assets, but do hold a few equities, so not all their foreign assets are in bonds).

The estimated mark-to-market losses are large relative to the lifers’ approximately USD $100 billion in reported capital. Most of those losses have not been realized: the bonds can be put in hold to maturity portfolios.

Bottom line: Taiwan’s life insurers cannot sell their bonds now (and realize the mark-to-market loss) without depleting their regulatory capital.

That’s a problem, as the insurers are effectively locked into a relatively low-yield portfolio.

The lifer insurers foreign bond portfolio may have an average coupon of under 4 percent, which isn’t that different from the Taiwan dollar return the insurers have promised to their policy holders.

One critical implication: the insurers cannot hedge the bulk of their portfolio against the risk of future foreign currency moves without killing their income and running ongoing losses, which would also eat into their equity capital. There is a reason why Cathay (the largest insurer, with $180 billion in foreign assets) only hedges 40 percent of its foreign portfolio.

For a while the large, unhedged dollar bond portfolio of the insurers was actually helping their solvency. The Taiwan dollar depreciated from 29-30 TWD/USD before COVID to 32-33 after the Fed started raising rates (after appreciating during COVID when U.S. rates fell to zero), and the currency gains from holding dollar bonds against depreciating Taiwan dollar liabilities helped the insurers’ reported balance sheet.

But those gains are now basically gone—the Taiwan dollar is a bit below its pre-COVID average, and bank analysts (JP Morgan, for example) now estimate that most of the insurers have depleted their foreign currency volatility reserves (funds set aside to cover small moves in the exchange rate, which substituted for true hedging).

The lifers are thus damned if they hedge and damned if they don’t hedge. If they hedge most of the foreign bond portfolio, the net dollar yield on the portfolio won’t cover the promised TWD payout to the domestic policy holders.  

If they don’t hedge, any further Taiwan dollar appreciation will eat directly into the life insurers capital—or rather it would under normal accounting rules (the lifers want the flexibility to NOT have to mark their bond portfolio for either bond or currency market moves).

As a result, the lifers preferred solution to their dilemma is just to have Taiwan’s central bank block any further appreciation.* 

That is in fact what the Central Bank of the Republic of China (Taiwan) did in the month of May. $10 billion in reserve accumulation in a month is a lot, it works out to $120 billion a year—or an annualized rate of intervention equal to 15 percent of Taiwan’s GDP.

That is equal to Taiwan’s current account surplus. It is also enough to trigger all of the U.S. Treasury’s old warning lights, and also promised new measures that the Treasury expects to put forward, and almost automatically lead to a (warranted) designation for currency manipulation.

After all, the agreed-upon definition of currency manipulation in the IMF Articles is blocking balance of payments adjustment, and keeping the currency of a country with a massive BOP surplus artificially weak through direct FX intervention would be a textbook example of impeding adjustment.

And it sure seems like Taiwan’s central bank intervened at the end of June to weaken the Taiwan dollar and help the lifers report smaller end of quarter currency losses. And there are certainly reports that suggest that the Central Bank of China (Taiwan) more actively “smooths” the Taiwan dollar at around 29.

That’s Taiwan’s policy dilemma. To avoid an “Asian crisis in reverse” Taiwan’s central bank has to keep the Taiwan dollar weak. But keeping the Taiwan dollar weak dooms Taiwan to maintaining a massive external surplus and accumulating an ever-increasing stock of foreign assets that are being bought at artificially inflated prices. It basically means more future losses—and of course, it also means standing in the way of the adjustment in trade balances that the Trump Administration wants (though the Administration wants conflicting things.

While the size of the unhedged portfolio of Taiwan’s lifers makes Taiwan unique, the same dynamics are at play on a smaller scale across a number of Asian economies. Japan is the most obvious example. Japanese lifers hold a lot of underwater foreign bonds too, and that makes selling those bonds and taking advantage of the obvious opportunity provided by high nominal yields on twenty and thirty-year JGBs much more difficult.   The Bank of Japan may equally be reluctant to raise short-term rates because of the impact of higher rates on the banks legacy holdings of low-yielding bonds, which also have often been stuffed into hold to maturity portfolios.

If the analogy is pushed a bit further, the unrealized losses in China’s financial system from their lending to zombie property developers (who haven’t been wound down or properly recapitalized) creates pressure to keep interest rates in China down, and low rates on the CNY helps keep the currency weak in real terms.

The yuan is now down 15-20 percent in real terms over the last 3 years, a move that institutions like the IMF really should notice.

The solution is simple in theory but hard to implement in practice. Undercapitalized financial institutions like Taiwan’s lifers need to be recapitalized (likely with government money, which is always politically difficult; see my FT Alphaville piece) so they can reposition their bond portfolio and hold higher yielding bonds. This will enable them to both hedge and pay out their policyholders (alternatively, the policyholders can be forced to bear losses).

That is, of course, politically difficult, as the recapitalization ideally would be done ahead of any large currency move. In fact, it probably would need to be done congruent with a policy that allows a bit of managed appreciation.

No matter. The essential argument is simple: on a host of fundamental metrics, many Asian countries are deeply undervalued. Real exchange rates are very weak and current account surpluses are generally quite high, particularly with the fall in oil prices.

These economies have accumulated a lot of foreign bonds (as ongoing surpluses implied foreign asset accumulation) and those bonds are often both underwater in dollar terms and unhedged.

Thus, the currency appreciation that is warranted by balance of payments fundamentals would potentially put financial stability at risk.

The risk of an Asian financial crisis in reverse is thus very real and something that the IMF should make central to its surveillance as it refocuses around its core mandate, which, I would note, is balance of payments surveillance, not fiscal surveillance.

The modern “it is mostly fiscal” IMF sometimes needs reminding!

 

*/ The lifers other strategy it to time their hedge so that the hedges are on during any period of movement in the Taiwan dollar.