Just passing through.

Photographer: Scott Olson/Getty Images.

How Cheap Are Buffett's Euro Bonds?

Matt Levine is a Bloomberg View columnist. He was an editor of Dealbreaker, an investment banker at Goldman Sachs, a mergers and acquisitions lawyer at Wachtell, Lipton, Rosen & Katz and a clerk for the U.S. Court of Appeals for the Third Circuit.
Read More.
a | A

I had sort of hoped that Berkshire Hathaway, which once claimed to have issued the world's first negative-coupon bonds, would issue actual negative-interest bonds this week, but it was not to be. Berkshire sold 3 billion euros of bonds today, and all their yields are positive:

Berkshire Hathaway sold 750 million euros of eight-year notes with a coupon of 0.75 percent, 1.25 billion euros of 12-year bonds at 1.125 percent and 1 billion euros of 20-year securities at 1.625 percent.

Based on the issue prices, the actual yields are 0.80 percent for the 8-year, 1.24 percent for the 12-year and 1.648 percent for the 20-year. Sadly, even 0.80 is greater than zero. It is, however, less than some other numbers. For instance, Berkshire Hathaway has some U.S. dollar bonds due in February 2023, just about eight years from now. Those have a coupon of 3 percent, and are now yielding about 2.56 percent.  So Berkshire's lame old 8-year bonds pay more than three times as much interest as its shiny new 8-year bonds.

Is that the right way to think about it? It seems a little too easy. I mean, the thing is, Berkshire is paying dollars on the existing bonds, and euros on the new ones. There is a reason that dollar interest rates are higher than euro interest rates. Or rather, there are a lot of reasons, and those reasons are to some extent subject to debate, but there is one simple arithmetic no-arbitrage pseudo-reason that dollar interest rates are higher than euro interest rates, and it is: because the market says that paying back euros in eight years will be more expensive than paying back dollars in eight years.  

This is arguably a strange thing for the market to say! But it says it with a certain amount of specificity, insofar as you can get quoted prices for cross-currency interest-rate swaps that will let you convert euros into dollars or vice versa. So Berkshire is issuing bonds in euros today. It will get 3 billion euros now, and will pay interest in euros for 8/12/20 years, and will pay back a total of 3 billion euros when the respective bonds mature. It could have swapped the payments back into dollars, turning the euro bonds into synthetic dollar bonds.

It probably didn't do that -- I mean, I have no idea, but it seems unlikely -- but we can know roughly what those swaps would look like, because, again, they are traded things. And those swaps come with an interest rate, which is the market price of turning euro borrowing into dollar borrowing -- of translating between euros and dollars. So we can construct hypothetical Berkshire Hathaway bond issuances. So here, very roughly, is how Berkshire's deal today translates into dollars (and, just to be silly, Swiss francs):

On this basis, is Berkshire saving money? Well, remember Berkshire has an 8-year dollar bond outstanding that yields about 2.56 percent. 2.78 percent is more than 2.56 percent. Berkshire is paying more to issue in euros, on a totally hypothetical swapped basis. On the other hand, that 20-year euro bond at 1.65 percent (3.59 percent swapped to dollars) seems attractive: Based on its existing long-term bonds, Berkshire's cost of issuing a new 20-year directly in dollars might be more like 3.7 or 3.8 percent.  

Is that the right way to think about it? I don't know, it sort of depends on how much you trust financial markets to tell you things. The swaps market is telling Berkshire that its 0.80 percent euro bonds are "equivalent" to 2.78 percent dollar bonds, which isn't as exciting as the headline number. On the other hand, that's pretty theoretical. Who cares what the swaps market thinks? Berkshire borrowed euros, and is now paying interest on those euros at a very attractive sub-1-percent rate.

You could also build a hypothetical shorter-term bond, say a 5-year. You could do this either based on Berkshire's existing dollar bonds (it has bonds due in August 2019 and October 2020), or based on its new euro bonds. But you'll get different answers. Here's a very casual glimpse of how that might look:

One thing that this tells you is that while short-term rates in Europe are quite low, on swapped to dollar terms it is not especially attractive for U.S. companies to issue short-term debt in Europe: Berkshire would pay up a lot to do a euro deal swapped to dollars versus a straight dollar deal. (Conversely, it could issue in the U.S. on terms that are "equivalent" to negative euro or Swiss franc rates.) Contrast the situation for longer-term bonds: Again, a swapped 20-year euro deal was probably more attractive for Berkshire than a straight dollar deal.

So consider Felix Salmon's view that "companies are pushing out the maturity of their new bonds until the yield becomes positive, and then they issue," because negative interest rates are a psychological barrier for investors. There might be something to that, but a simpler explanation is that longer-tenor bonds in Europe are a better deal on a purely financial basis. If Berkshire or Apple wants 5-year money, it should borrow in dollars. If it wants 20-year money, then euro rates are relatively more attractive.

So this is all weird fun with numbers, but what does it mean? Well, imagine you are a company. Should you issue bonds in euros, or dollars, or francs, or what? There are lots ways to answer that question that have nothing to do with these numbers. If you are a French utility company and most of your revenue is in euros, probably borrow in euros! (For free, basically! ) If you are an American multinational company with tons of sales in Switzerland, maybe borrow in Swiss francs, why not. Borrowing in different currencies might diversify your investor base or tap pools of demand you can't find in your home currency or better match your revenue or give your treasurer something fun to do or whatever.

On the other hand, if you are a company that lives in the U.S., and your revenue and expenses are primarily in dollars, and you just want dollars, what should you do? Generalizing wildly from Berkshire's experience, it seems like the cheapest way to get dollars is to borrow in dollars, at least for short and medium terms. Borrowing in euros and turning them into dollars -- on a fully hedged basis, where you turn them into dollars via swaps -- costs more than just borrowing dollars directly, though that relationship reverses in the longer term.

But of course you don't have to swap the thing into dollars. You could borrow, buy dollars at today's exchange rate, spend the dollars and then hope that you'll have enough money in 5 or 8 or 20 years to pay back the euros you borrowed. The benefit of this is: super low interest rates! The disadvantage is that you're taking currency risk: If the euro appreciates significantly against the dollar, the euro bonds will cost you a lot more dollars than you got, and you'll wish you had hedged.

In other words, all the U.S. corporate treasurers borrowing in euros when they really want dollars are just engaged in a straightforward carry trade, betting that the euro won't appreciate as much as the swaps market says it will. If they're right, they will save a lot on interest. If they're wrong, they will have to pay back a lot more than they borrowed. Judging from the euro borrowing binge, there seem to be a lot of those treasurers.  

Of course, I don't want to tell you whether the euro will appreciate against the dollar over the next 8 or 12 or 20 years. I have no idea. I will say that the treasurers have some good arguments: The European Central Bank is launching its quantitative easing program, trying to weaken the euro, while U.S. Federal Reserve is talking about raising rates, which tends to strengthen the dollar. The interest rate differential between the dollar and the euro seems to say something about future monetary policy that lots of people don't really believe. At least some people think that the euro probably won't appreciate too much against the dollar, at least until they pay back their bonds. One of those people might be Warren Buffett.

  1. Those bonds are CUSIP 084670BJ6, the 3 percent of 2023. All price, etc., information is from Bloomberg, retrieved over the course of the late morning and early afternoon today, and you should not assume that it is precise to any particular instant.

  2. Wait no that's wrong. I mean, that describes what the market is saying about the forward exchange rate, but the forward exchange rate is not the future exchange rate. Take it away Wikipedia:

    The forward exchange rate is determined by a parity relationship among the spot exchange rate and differences in interest rates between two countries, which reflects an economic equilibrium in the foreign exchange market under which arbitrage opportunities are eliminated. When in equilibrium, and when interest rates vary across two countries, the parity condition implies that the forward rate includes a premium or discount reflecting the interest rate differential. Forward exchange rates have important theoretical implications for forecasting future spot exchange rates. Financial economists have put forth a hypothesis that the forward rate accurately predicts the future spot rate, for which empirical evidence is mixed.

  3. That is: A bank exchanges the 3 billion euros into dollars today, and dollars back into euros in 8/12/20 years, and lets Berkshire make the interest payments in dollars and turns them into euros, all at exchange rates fixed in advance, so that Berkshire has effectively borrowed in dollars.

  4. Okay so this is:

    • Bloomberg function SWPM, choose "Cross-Currency Swap (Fixed-Fixed)" from the "Products" menu, choose EUR for Leg 1 and USD for Leg 2, put in the tenor and yield of Berkshire's bonds in Leg 1, put zero in for the premium, and use the Solver to solve for the Leg 2 coupon.
    • Obviously this is not, like, executable -- it pulls in various assumptions, includes no fees for banks, plus I did it at various times today so it's not based on a consistent yield curve or anything. It's extremely illustrative.
    • You can verrrrrrrrrrrrrrrrrry loosely approximate this arithmetically, by taking the spreads that Berkshire is paying in euros (27, 42, 57 basis points respectively for its 8/12/20-years, according to Bloomberg News), adding them to the relevant USD Libor swap rates (from Bloomberg page IRSB) and then adding the cross-currency basis swap differential (Bloomberg page XCCY). I get something like 2.7 percent, 3.1 percent and 3.4 percent, respectively, for the 8/12/20-years from this method, 11-24 basis points lower than the more accurate numbers from SWPM.
    • Probably just ignore the Swiss franc rates.
  5. This is the main reason I assume that Berkshire is not doing these bonds on a swapped basis: It'd be more expensive than just issuing dollar bonds!

  6. There's a 5.75 percent of 2040 -- a little under a 25-year bond -- yielding about 3.87 percent, or a little more than 125 basis points over swaps. The swap curve from 20 to 25 years is about 4 basis points, take off another 10 basis points for credit curve, maybe add back a bit for the new issue, and you get something over 3.7 percent for a 20-year.

  7. The bold numbers are the yields for direct issuance; the rest are swapped numbers. 

    I see Berkshire's dollar bonds of 2019 and 2020 yielding about 1.71 percent and 1.94 percent, respectively. The 1.85 percent in the table is just interpolated between those two, plus a little bit for a new issue.

    For the euro bonds, I just took the 5-year euro mid swap rate (about 0.29 percent) and added 17 basis points, a number I made up. (Berkshire's 8/12/20-year actual euro bonds came at spreads of 27/42/57 basis points to mid swaps, so I figured 10 basis points tighter than the 8-year seemed plausible.)

  8. Another thing I guess it tells you is that no-arbitrage arguments only go so far. Berkshire can issue at a rate A in dollars that implies a rate B in euros, or at a rate C in euros that implies a rate D in dollars, and all of those rates are different. Market segmentation, etc. etc.

  9. No, I kid, a zero-coupon two-year bond yielding 0.13 percent isn't free. 0.13 percent is still bigger than zero.

  10. This is called "no arbitrage," obviously.

    Of course even if you're a company with revenue in lots of different currencies, you are still making a currency bet by choosing to borrow in any particular one of them. It's just a safer bet than if you're an all-dollars company borrowing unhedged in Swiss francs.

  11. Here's a guy:

    “It’s so cheap to borrow in euros that it’s a no brainer,” said James Athey, a fund manager at Aberdeen Asset Management in London. “It’s something we expect to see more and more of to the point where, potentially, U.S. corporates are issuing far less in their own market and issuing far more in Europe.”

This column does not necessarily reflect the opinion of Bloomberg View's editorial board or Bloomberg LP, its owners and investors.

To contact the author on this story:
Matt Levine at mlevine51@bloomberg.net

To contact the editor on this story:
Zara Kessler at zkessler@bloomberg.net