Don't Get Stuck Reaching for the Last Crumbs

"A little word of advice my friend: sometimes you gotta let those hard-to-reach chips go."

- Dante Hicks, "Clerks"

If I'm going to embarrass myself by publicly admitting to fluent recall of bawdy-humor movies from the early Nineties, then hopefully it's in service of making an important point:

The yield on long-term Treasuries (ETF ticker: TLT, iShares 20+yr Treasury Bond) does not justify the principal risk inherent in owning them at these levels, despite the long-standing historic diversification benefits that arise from negative correlation between long bonds and equities.

And the fact that investors have enjoyed a 12.44% YTD return on TLT - despite rates that have churned around historic lows for several years - only tightens the noose around the trade; this return has come as 30yr Treasuries have declined from 2.98% to 2.39% since January.

Traditionally, the vast majority of bondholders' returns come from interest income. And when yields are generous (or even mediocre) they can offset periodic capital declines should yields move higher. But TLT's current 30-day SEC yield stands at 2.05%. What happened the last time rates approached these levels? From 1940-1950 bond yields posted negative real returns over a decade.

Speaking from a price perspective, Doubleline Capital's Jeffrey Gundlach - the reigning "bond king" - put it this way: "The yield on the 10-year yield may reverse and go lower again but I am not interested. You don't make any money. The risk-reward is horrific." The same holds true looking at the investment on the basis of its paltry yield.

And what about that correlation benefit to stocks? After all, equity earnings multiples are expensive as well. I may not like bonds, but shouldn't they still offer good diversification in the event of a broad-based selloff? Even if the correlation is less negative, won't the magnitude of my loss likely still be lower?

Probably. But the fact is that since 2002 (the data I was able to most readily find), the correlation of SHY (1-3yr Treasuries) to US Stocks was -0.36 versus -0.42 for TLT (20+yr Treasuries) - meaning, I achieved almost the same amount of diversification without taking on the duration risk.

Moreover, the correlation between long bonds and equities can certainly change. Per Vanguard:

"Assets with low and unchanging correlation can and do move in the same direction from time to time. In addition, correlations between asset class returns can and do change over time or in particular circumstances. Future correlations may also differ from those in the past because of changing economic and market regimes. Investors should take these factors into consideration when using correlation as a key input for constructing investment portfolios, not relying solely on statistical measures, but mixing in common sense and qualitative judgment as well."

"Why does measured correlation differ from its long- term average? The fact that observed correlation varies, even over relatively long periods of time, does not necessarily mean that “correlations are changing,” although this may be the case. It simply reflects randomness in the return variables themselves, which generally produces ex-post outcomes that differ from the “true” underlying statistic or longer-term average, particularly over shorter periods."

- Vanguard, "Dynamic correlations: The implications for portfolio construction", April 2012. [my emphasis]

And this is what common sense tell me: over the period 1972-2015, long Treasuries averaged an 8.64% yield with a 12.08% standard deviation (a Sharpe ratio of .715), as yields marched down from 15% Volcker-era highs to our current nadir. Along the way, investors also enjoyed the diversification benefits that came from a low/negative correlation to equities.

But the three decade bull market in Treasuries has driven yields towards their mathematical limits (let's ignore the insanity of negative yields in this argument). The coupon income currently available is insufficient to offset the likely scenario that yields drift higher over time.