fifth bloggiversary

Yes, another fiversary (though not work-related). I've not marked the prior millstones as I've ground through them, and I'll admit that the unevenly monthly posting of the past year has yielded up thin gruel lately. Not that the lack of clear direction is anything new; Stochastic Bookmark was inblogurated with writing samples pulled from the drawer, moved on to mostly litcritical ruminations before settling into a reading diary, which course was abandoned a couple of years ago to try a different tack, but the wind didn't hold; I've since indulged a broader range of interests, including professional (vs confessional self-indulgence), with gallimaufrian result. I'll not mark this occasion with a medley of greatest hits, only take the opportunity to update the table of contents (which bumped the first post into the second).

I will however revisit a post from a year ago, on yield curve dynamics (capsule summary: there's potential a few years out for a rare configuration, a inverted humped yield curve, as occurred in Germany in spring '94, but regardless there's relative value in the belly of the curve). Below I'll track how my recommendation of the 7-yr Treasury note is doing so far, but first I'll remark on some recent interesting developments:

Reading the economy from the curve: The Cleveland Fed released a research note suggesting the curve signals slow growth but no double-dip (cf reactions from PragCap, Felix Salmon, Greg Ip). The old saw was that an inverted yield curve has predicted 9 of the last 5 recessions. The simplistic economic reading using two points to proxy the curve is deceptive: The last inversion was minimal and equivocal, hardly indicating the severity of the damage to come, and I pointed out last year that the duration of the episode during which the curve was essentially flat, in unstable equilibrium, stored up energy like a spring (though I've yet to see an economic model that recognizes it). I also said that one of the curve features bugging me was: "The current yield curve, while 'normal' in shape, is abnormal not only in being so low (though Japan's has long been lower [a dead cat silhouette?]), but also in that its inflection point occurs farther out (2-3 yrs)." Since then, the same, only moreso (as with the Fed's "extended period" of low rates, ntm managing the curve with Treasury purchases). Japan's curve signalled its lost decade plus, but that occurred while the rest of the world wasn't losing any time. Turning Japanese? I don't think so, pace St Louis Fed prez's latest jawboning (pdf summary, paper), but ... after factoring out safety and liquidity etc, the US curve indicates that there's a lot of drag for some years (but no decade) to come. Long-term alarmist, if you will.

Benchmark bond swap spread goes negative: Again. This is more puzzling; my partial explanation at FTAlphaville the first time around (LIBOR/OIS misbehavior, convergence of 10- and 30-yr spreads) has since reversed, and now the 10-yr spread has breached zero again, following the 30-yr spread down, with little evident correlation to other market indicators. Negative swap spreads are nearly as anathematic as negative interest rates (why should inter-bank credit be better than the government's?), and while technical factors (e.g., structured exotica hedging) mitigate concern over the persistence of 30-yr spread negativity, that it should crop up against the most liquid security in the world is more troubling. I provisionally expect (without recourse to more recondite data, based on how the 30-yr spread behaved when it crossed zero) this episode to persist for about 2-3 months in negative single digits before normalizing to low positive single digits. But there's a lot that can operate at variance with this expectation.

Update 30.07: FTαville is onto this, striving to relate it to convexity trading, which in the past drove repo through zero (and elsewhere to asset swap demand for burgeoning corporate issuance), and noting repo misbehavior now as well. The parenthetical explanation seems to be the best try, but insufficient in and of itself. Interesting times ...

Back to prior prediction: The evolution of the curve over the past year is consistent with (but does not prove nor make more likely) a humped inversion. The belly of the curve (5-, 7-yr maturities) was cheapened by expectations of reversion to a normal shape, for which the slope is greatest across shorter maturities; instead, this maximal slope has drifted farther out the curve, benefitting the belly by both lower yields and roll-down (migration of fixed-date maturity to lower tenor). I chose the 7-yr, "that step-child among the on-the-runs", as the best value on the curve due both to these benefits and the lesser detriments of curve normalization, with the kicker that humped inversion would pay off halfway to its maturity. So, one year on, how's it doing?

Maturity (yrs)______ 1 __ 2 __ 3 __ 5 __ 7 ___ 10 __ 30
Yield 29.7.09 _____ 0.49 1.12 1.66 2.63 3.30 _ 3.72 4.56
Yield 29.7.10 _____ 0.30 0.61 0.95 1.75 2.43 _ 3.03 4.07
Change (bp) ________ -19. -51. -71. -88. -87 _ -69 _ -49
Yld w/1 yr roll ________ 0.30 0.61 1.35 2.09 _ 2.83 4.05
Change (bp) _____________ -82 -105 -128 -121 _ -89 _ -51
Price apprec (over par)_ 0.82 2.08 4.97 6.79 _ 7.03 8.67
Apprec incl cpn ___ 0.49 1.94 3.74 7.60 10.09 10.75 13.23
$Duration (start)__ 1.00 1.97 2.91 4.66 6.20 _ 8.29 16.26
Riskadj ret(/$Dur)_ 0.49 0.98 1.28 1.63 1.63 _ 1.30 0.81
notes: yields from FRB H15, roll lin interp, pricing for par as of 29.7.09

So, pretty much as advertised (though the risk adjustment by interest rate sensitivity is an outer bound, as readjustment over the course of the year would favor shorter maturities). The 7-yr may still be attractive, today's auction notwithstanding, but some of the juice has been squeezed out of it.

Sorry, bloggiversaries get me all wonky ...

Notable July reading:
Yves Bonnefoy, The Curved Planks: Nobel-worthy poetry
Adam Foulds, The Quickening Maze: redolent of Penelope Fitzgerald, high praise in my book