Argentina vs Austria, the smackdown of the century (bonds)

A few years ago, in the halcyon era of zirp, there was a mini-trend for “century bonds”. Even Oxford university got into the act. But the two most high-profile 100-year bonds were issued by Austria and Argentina.

The two countries couldn’t be more dissimilar. Austria is a dour, competent, fiscally conservative and drama-free developed country. Even people who like Austria think its flag being among the world’s oldest is one of the most interesting things about it. Seventy per cent of young Austrians say their favourite drink is tap water.

Argentina is, well, Argentina.

But in 2017 Argentina’s new president Mauricio Macri was keen to show that the country had changed, and celebrated its return to the fixed income market by selling a $2.75bn bond maturing in 2117, with a yield of 7.9 per cent. The then-finance minister crowed that “such an issuance is possible thanks to our recovery of the world’s credibility and confidence in Argentina and in the future of our economy”.

It went about as well as you might expect.

Macri was out by 2019 and Argentina defaulted on and restructured the century bond in 2020. Whoops. It’s almost as if lending money to a country where a board game titled Eternal Debt has the strapline “Can you beat the IMF?” was a best-seller might be a bit risky.

Argentina’s creditors received on average 54.5 cents on the dollar. But the 2117 bond was actually sold at 90 cents on the dollar to entice investors, and paid at least a few chunky coupons before it was restructured and exchanged for a lower-value new bond, which would have ameliorated the pain somewhat.

Sovereign debt veteran Paul McNamara of GAM kindly ran the numbers for Alphaville. It turns out that if you bought at issuance and reinvested the coupons back into the 2117 bond you would have lost about 30 per cent going into the restructuring, and roughly 53 per cent on the other side of it. (NB, the gap is unusually large, which means the Bloomberg data might be a bit shonky.)

If you held on to the restructured exchange bonds that mature in 2046 and reinvested the new coupon payments into that, you’d today be staring at a ca 63.7 per cent loss.

Austria has obviously not defaulted on the €3.5bn century bond it issued just a few months after Argentina in 2017 (later upped to €6bn). But remarkably, investors could have been better off buying the Argie paper. Here’s a chart of the price of the Austrian 2117s:

That’s an almost 40 per cent loss in price since inception. The miserly semi-annual 2.1 per cent coupon payments helps only a little, crimping the total loss is 31.27 per cent, according to Bloomberg data. If you were unlucky/foolish enough to buy the Austrian century bond at its peak price/record low yield in 2020 then you’re looking at close to a 75 per cent loss.

There are few better examples of the explosive power of duration when the interest rate cycle turns.

Regular FTAV readers probably won’t need reminding (and certainly not after 2022), but a bond’s duration is different from its maturity. Here’s a quick explainer for anyone else who is curious about some basic bond maths. Hopefully this description (and the bond math) is accurate, but let us know any snafus in the comments.

There are two (related) variants of duration that are often used interchangeably, but show different things. “Macauley duration” is how long it will take investors to recoup their principal through interest payments — a pleasingly elegant concept — but finance types mostly talk about “modified duration”.

Modified duration is derived from Macauley duration, and is the mathematical measure of how sensitive a bond is to fluctuations in interest rates. Every 1 per cent change in interest rates moves the value of the bond equal to the modified duration.

That’s because bonds become less valuable when rates rise, and more valuable when they fall. For example, if you buy a 2 per cent bond when rates are 1 per cent, it obviously becomes less valuable if rates go up to 3 per cent and newer bonds offer greater returns.

The longer the maturity, the greater the duration, as a bond that pays 2 per cent a year for two years is less vulnerable to the ebb and flow of interest rates than one maturing in 10 years. In practice, both Macauley and modified durations are often very similar (and both are often expressed in years, even if it’s only really appropriate for the former).

So low yields and long duration equals mega duration, while short-term junk bonds have minimal duration. For example, a 10-year Treasury bond (with semi-annual coupon payments) issued today at the current 4.79 per cent yield has a modified duration of 7.87 and a Macauley duration of 8.06. A two-year bond yielding 10 per cent at par has a modified duration of 1.77 and a Macauley duration of 1.86 (you can run your own numbers with this nifty online calculator).

The price impact of yield changes can be sizeable if the duration is high. If you invested $1mn into the 10-year Treasury bond today it would be worth over $1.1mn if yields dropped below 3 per cent again, or about $850,000 if yields approach the 7 per cent suggested by Jamie Dimon (we used these separate online calculators, because life was too short to do the numbers ourselves).

When it was issued Argentina’s century bond only had a modified duration of about 8 — thanks to its high coupon payments it was less sensitive to interest rate movements than its 100-year maturity might suggest. (it was pretty sensitive to the issuer being Argentina, however!)

In contrast, Austria’s century bond had a duration at issuance of almost 42. And when interest rates were floored after the pandemic and yields tumbled, the duration peaked at over 60 (when every €100 worth of bonds traded at a price of almost €240 for a yield of 0.38 per cent in March 2020). When yields started to reverse it all went horribly wrong.

However, Argentina and Austria’s century bonds can’t lay a hand on the ultimate duration champ: the UK’s £4.4bn inflation-linked gilt maturing in 2073, with a duration of over 50 when it was sold in late 2021 due to the combination of a 0.125 per cent coupon and the long maturity.

This has obliterated the inflation compensation it has offered to investors. The total loss is now a gobsmacking 79.7 per cent. That’s even worse than Alphaville faves like Cathie Wood’s ARK ETF, bitcoin and THG since then . . . 

Or rather: