Tuesday, March 10, 2015

On Negative Bond & Bund Yields

Frances Coppola has written an insightful article titled "The ECB’s policy mix is poison for banks" (FT.com).

Here's one sentence which caught my attention:

If the expectation is first that inflation will continue to fall and second that central banks will continue to intervene in markets to try to prevent inflation falling, then yields will continue to fall without limit.

Well, "without limit" might be too bold a statement, but it is hard to tell where any limit eventually lies. For me, the important point to draw from this is that, as things stand, "short-selling" government bonds with a negative yield could be a "widowmaker" on a par with selling short Japanese government bonds (JGB) in the past two decades. But here I mainly talk about German bunds and some other AAA/AA bonds. Due to political instability and fragility in the construction of the euro system, we cannot extend this to any and all eurozone bonds. Prudence, not hope, should guide our decisions.

Here's Ms Coppola again:
Even in a deflationary environment, retail customers — who provide the majority of banks’ stable funding — are generally unwilling to accept negative rates on deposits, especially as physical cash is an alternative. This, of course, is the reason for the ECB’s minus 0.2 per cent limit. It cannot lower the interest rate on excess reserves much without hitting the “slightly negative lower bound” at which banks would exchange reserves for vaulted cash.

I must admit I had mostly overlooked the possibility that banks could exchange reserves for vault cash. Would be interesting to hear about any practical limits to this. For instance, would the authorities print 100-800 billion worth of new euro banknotes (large denominations) just like that? I doubt it.

Overall, Ms Coppola seems to possess an independent and searching mind. In my opinion, we need people like her if we are to make sense out of all what is going on in the world economy today. As is typical and even natural for this kind of mind, her article includes some "clumsy statements" and no doubt errors too, but those are more than balanced by the new insights she offers and thoughts she provokes. It seems to me that she is not too wedded to any economic theory, so she is more free to describe what she actually sees; what really is happening in the real world -- quite literally so, as her focus on real interest rates shows.

But here I think Ms Coppola goes wrong (assuming I don't misinterpret her):

First, the investment function of government bonds is changing. Traditionally, they have been used as hedges in a diversified portfolio. But once yields are deeply negative and falling, they can no longer be regarded as hedges: since holding them incurs an inevitable loss, they must be regarded as risky, not risk-free, and therefore actively traded for a capital return. Their replacement would, of course, be physical cash and other zero-yielding assets, possibly including zero-coupon issues from blue-chip companies.

 First, to equate "risk-free" with "hedge" is hasty, to say the least. I'm sure she knows how risk-free was never really free of all risk. It's more like a heuristic, or at worst, conventional wisdom. For instance, holding these "hedges" has given investors very good returns since the early 1980s as bonds have appreciated. These returns, depending on the timeframe we choose, are often competing head-to-head with stock returns. In a high inflation environment, we would have seen these "hedges" bring considerable losses to investors.

Second, holding these bonds doesn't incur an inevitable loss. Buying them at more than face value and holding them to maturity does incur a nominal loss. Well, this is probably what she actually meant.

Here is my third, and most important, objection to Ms Coppola's statement: Why would these bonds be less risk-free than they used to be if the investor who buys them today is guaranteed a nominal loss if he holds them to maturity? I don't see much risk in a guaranteed nominal loss, the size of which is known ex ante. This statement is also in contrast with her showing earlier in the article how she understands how it is real, not nominal, outcomes which ultimately matter. We need to assess real risk, not nominal risk. Government bonds have always carried real risks. And we cannot assume that blue-chip companies would be a safer bet, even if their bonds yield more. Neither is physical cash safer for anyone with hundreds of millions, if not billions, to look after.

Next I delve into something Ms Coppola seems to omit in her article -- as I believe most of the other commentators have done too.

CORRECTION (MAR-10-2015 -- 20:50 GMT): Frances Coppola kindly advised me that she has discussed the "hot potato" effect (which I discuss below) already in another article (see here) which her FT post linked to. It seems I have a lot of relevant stuff to read tonight. My apologies & thanks for the feedback, Frances!

Commercial Banks and Hot Potatoes

As a consequence of QE, the commercial banking system as a whole ends up holding excess reserves. (For how this works, see this paper from New York Fed -- a paper I referred to in my first post in which I present a sketch on bank reserves and other "monies".) These excess reserves, which usually yield less than any alternatives (I'll get to these soon), become a "hot potato" which any individual bank would like to get rid of, but which banks in aggregate cannot get rid of. One could think that this "hot-potatonizement" is especially strong when the deposit rate "paid" on these reserves is negative -- as it is now at ECB --, but does it need to be so? What probably matters more is the "spread" between the yield on reserves and the yields on any alternative assets banks might want to hold.

If we assume that European banks are operating in a "risk-off mode" (they build buffers while they adapt to a stricter regulatory environment), we can conclude that these banks would prefer to hold, instead of excess reserves, any "near reserves" as long as those yield more than reserves. These "near reserves" can be German Bunds and other AAA/AA bonds which banks can, any day, exchange for reserves should they need them. By buying these bonds from investors with accounts in other banks (so that the bank ends up transferring reserves to the other bank), a bank can switch excess reserves to bonds with (slightly) higher yields -- until the yields on these converge with the yield on reserves.

Even if it won't make much sense for the bank to buy more of these "near reserves" when the yields on them have already converged with yields on reserves, it might make sense for a bank to hold on to these bonds as they are an asset that keeps on appreciating -- regardless of the reason it does so.

There is, of course, the "hot potato" problem: these reserves might eventually return to the bank if other banks try to pass them on in a similar fashion, although it might be mainly a problem of the largest banks where most of the deposit accounts reside. In any case, the U.S. commercial banks have increased their holdings of "near reserves" (Treasuries, and Agency Securities like MBS) significantly during QE1, QE2 and QE3:

As Bloomberg reports:

Investing in government bonds is proving to be a profitable move for banks. They’re making over a full percentage point more by purchasing five-year Treasuries instead of leaving the idle cash parked at the Fed, where they earn only 0.25 percent.

Now, I don't claim that this is the only, or even the most important dynamic behind the very low and in some cases even negative bond yields we are witnessing. No doubt German Bund yields are driven down also because those are one of the few euro-denominated "safe haven" securities -- especially if one factors in the risk of certain countries leaving the euro in the foreseeable future. The current yield on a German 2-year Bund speaks for these other yield-compressing factors, as it stands already at minus 0.24 percent. And like always, there is a lot of "front-running" by other investors involved in this. All this makes it very hard if not impossible to rigorously test any "causal hypothesis" against the market data.

No comments:

Post a Comment