Bps and Pieces

[yes, I copied this post title from a famous newsletter, what can I say…]

Last week Lorenzo Biagi (author, financial advisor and…finfluencer? ;)) posted the following picture, unfortunately without the source links:

These two graphs offered him the opportunity to stress how bonds might be unattractive at this point, especially as a hedge to stock declines.

I found a couple of similar interesting conversations about bonds on Twitter. As someone put it, it is amazing how puzzling an instrument as plain vanilla as a bond can be.

Back to Lorenzo’s graphs. My reading is that in the past, when stock/bond correlation was high, bonds were also offering a high-term risk premium, the compensation that investors require for bearing the risk that interest rates may change over the life of the bond. This time is different. With correlation at an all-time high, the term premium is zero (not entirely surprising considering the inverted yield curve).

I guess the conclusion is: why bother?

Not so easy. Correlation is a measure that by definition looks at the past, not exactly what we care about. Today, we want to know how correlation will be tomorrow, not yesterday. One might even argue that we past the worst and the future can only present a lower correlation. If the past can offer any insight, it is that it is indeed difficult to predict the exact level of future correlation but regimes are persistent: if today I see positive (negative) correlation, there are higher chances that also tomorrow’s correlation would be positive (negative). So my best guess is that tomorrow’s correlation would be lower than today but positive.

Is this an issue? It depends on why we hold bonds in our portfolio. As shown in this previous post, even when correlation is positive, bonds continue to offer crisis alpha. With the term premium so low, this protection simply becomes less cheap. There is a trade-off between the income you want to earn from your bonds and the protection you want to achieve. Shorter bonds offer higher yields and more flexibility but their lower duration means you need a higher allocation to hedge the same stock-crash risk amount.

There might be an advanced Risk Parity model out there that adjusts stock and bond portfolio weights based on realised correlation (as an input to infer future correlation) but I do not think it is material for us, poor bastards. So we are better off thinking stocks and bonds will generally be positively correlated and stop; however high this correlation is, it is not our problem.

There is a way to profit from the actual term risk premium level and I wrote about it in the Futures Yield post. Or better, by combining bonds with a carry strategy, our portfolio is neutral the term risk premium when it is not positive (or enough positive, depending on how the strategy is built). Problem solved…as long as the carry strategy re-positions itself fast enough when the yield curve changes and the portfolio needs the bond crisis alpha to kick in.

Overall, it is a narrative issue. As much as bonds weren’t the perfect diversifier in the ’00 to ’20, they are not the worst now. They have always been meh. A meh that happens to have positive EV, positive real returns (in the long term) and offer crisis alpha when combined with stocks. Today long-term bonds do not (seem to?) offer any inflation premium: if inflation continues to yo-yo up and down staying above what is forecasted, bonds will have negative real returns in the next decade. But can you time it?

If you do, where are you allocating other than bonds? Let’s not forget that there’s another half of the world screaming that stocks are overvalued (at least in the US). Cash might be a better allocation if things hum along (I assume you do not want to buy even more stocks) but if they are overvalued and crash, then bonds are back in first place.

The only sure thing is that in one year (or two, or three), I’ll have to hear many people who would state that it was “obvious” that stocks would crash and/or bonds would be trash.

Let’s continue to confound ourselves.

I do not remember where I found this picture:

1981 is the year when Treasury yields peaked (or prices bottomed, as you prefer) post the ’70s inflation shock. The following 40-year bond bull market warped the efficient frontier substantially. Using Long-Term bonds as a complement to stocks was overwhelmingly the best possible choice, so much so that a 40/60 portfolio generated a return that would have required a 70% to 80% stock allocation when paired with lower-duration instruments.

Just a reminder, between 1981 and 2001 stock-bond correlation was positive: half of the period. Correlation is not necessarily a bad thing if the price of both assets goes up.

It is a great picture to demonstrate that the stock-bond combo worked under different regimes. The loss in returns of a less than 100% stock portfolio was always compensated with a higher reduction in volatility. However the optimal configuration was highly regime-dependent.

@TrendingValue and @OptimizedPort had a nice back-and-forth on the same topic:

I'm trying to get to the bottom of bonds and diversification. @OptimizedPort argues for long bonds due to volatility and correlation benefits in an equity heavy portfolio. I originally saw these arguments (first 20% in long bonds) from famous Boglehead poster Vineviz and… pic.twitter.com/gQ6smkumIK

— Trending Value (@TrendingValue) June 24, 2024

1. I know you explicitly called yourself out on this but I'll reiterate – Why would we selectively cherrypick a single arbitrary 30 year period that supports the hypothesis? Letting it run to present shows long bonds being superior. But we'd likely want to look at rolling periods…

— Optimized Portfolio | John Williamson, APMA® (@OptimizedPort) June 24, 2024

Is it better to allocate to Intermediate or Long Term? Historically, past the 10-year maturity, investors got paltry term risk premia, especially if adjusted by volatility. On the other side, the longer the duration, the higher the convexity and related crisis alpha. Levered IT bonds might offer the best of both worlds, a la $TYA, but @OptimizedPort argues that the ETF has had wonky behaviour…

My conclusion is that aiming for the optimal allocation might be an impossible puzzle to solve; but diversification, in whatever form, is at least a guarantee of a fine result (and if you are lucky…). The more, in terms of risk premia, the merrier.

What I am reading now:

Follow me on Twitter @nprotasoni