BofA Looks At The "What If" Scenario For Bonds: Find A Surprising Result

One of the big fears among the bond market, where most participants now openly admit there is a "bubble in credit", is that an unwind in global bond yields would lead to substantial losses.To test this assumption, Bank of America's Ralf Preusser looks at the "what if" scenario, namely what would happen to total returns should government yields fully reverse their 30 year historical evolution. What he finds is surprising.

As Preusser observes, "central banks seem increasingly reluctant to cut rates into negative territory. If policy rates have reached their lower bound, then the distribution of forward rates becomes more asymmetric, skewed to the upside. With talk about a potential bond “bubble” more prevalent, we address the question of what a secular bear market in rates could look like."

He continues:

With rates at, or near record lows, it becomes increasingly hard to justify a constructive stance on the rates market. Economists’ consensus forecasts invariably point to higher rates across the US, UK and euro area. Historical total returns in fixed income cannot be repeated without rates falling ever further. Whilst central banks were willing to cut rates, longs could point to the option value of duration in case of more negative policy rates. However, with the debate on the merits of negative rates having shifted, this argument has lost support.

So is it time to start selling even assuming an unwind of the entire Great Moderation is upon us? Not according to BofA:

We believe it is too early to call for a secular bear market in rates. We expect government bond yields to stay at or below their current levels across major markets for the remainder of the year.However, it is interesting to address the question of what such a bear market would actually mean. Across the markets we analyze we find strongly positive total returns over our simulated 30y horizon. Income returns dominate price returns in most years. For total return investors, even a reversal of the secular bull market would justify a constant allocation to USTs.

To conduct its reversal test, Preusser assumes that history perfectly reverses itself. "To avoid any normative discussion about what the future beyond our forecast horizon looks like, government bond yields are assumed to retrace their path of the last 30 years in a mirror image. This approach implies that issuance patterns also reverse their historical evolution."

Whilst clearly simplistic it does allow us to very easily test what the return environment for bond investors looks like in that scenario. We use our BofA Merrill Lynch indices total return series for Treasuries (G0Q0), Gilts (G0L0), German government bonds (G0D0), and Euro government bonds (EG00). By subtracting historical price return from historical income return, we can easily calculate what total returns government bonds would deliver if rates were to increase back towards 10% over the next 30 years (see Chart 12).

And here is the surprising punchline: as total returns are dominated by income, BofA's results "bury the myth that a rising rate environment would deliver severe losses for bond investors over multiple decades.

Table 6 shows that whilst total returns clearly are lower over the next 30 years than the last 30 years, they remain decent and outperform USD cash.

To be sure, a sharp spike in yields would result in dramatic MTM losses, but that is not the focus of the BofA study, which instead looks at the long-term impact of rising yields. More details:

Portfolio rebalancing at rising rates soon starts to dominate, with income return generating total returns fairly quickly. Even now, the market weighted average coupon of the German government bond index is 2.52%. Hiding out in cash, is therefore a loss making proposition over the thirty year horizon we simulate here. Returns are also strictly positive on an inflation adjusted basis (again assuming inflation to reverse history). Portfolio rebalancing refers to the fact that as bonds drop out of the index and new bonds are issued, it creates a need to rebalance the portfolio to maintain index weights, and thereby the opportunity to accumulate bonds in the future with higher coupon levels.

However, as BofA correctly points out, most investors will not be reassured by the fact that over a thirty year horizon, returns will work out to be OK. Instead, investors will be much more concerned about how to trade the start of a bear market, where returns are likely to be much more limited. Table 6 shows some additional risk metrics for each market. Staying invested in US Treasuries not an outrageous proposition.

Preusser then calculates that USTs actually look attractive even through the start of this simulated bear market: the cumulative loss from today would be limited to -3.4%, with the total return index bottoming already in Feb-17. The worst 1y return would be -4.8%, compared to -4.5% over the last 30 years. In comparison, Gilts arguably look the least attractive. Gilts underperform on the back of the highest duration of these bond markets, thereby resulting in a much slower coupon reset, whilst USTs benefit from the lowest duration and therefore the fastest rebalancing cycle. For the euro area government bond index, it is important to recall that the simulation assumes a repeat of the euro area sovereign crisis as well, which will obviously skew the results and explains why the total return index troughs out the latest (in Mar-21).

So is it all smooth sailing, and nobody will rush to sell rates? Maybe, however there are several key assumptions.

Reversing history assumes curves bear steepen from here

While history may repeat itself, it certainly will not likely reverse itself as perfectly as this analysis assumes. We do not want to get into a discussion of how the world will evolve over the next thirty years, however, one key artefact of this analysis warrants discussion. For all markets, a reversal of the bull flattening we have seen over the last three years, implies a bear steepening as its mirror image. This implies a “normalization” of interest rates that is not led by central banks, but instead by a repricing of long-term breakevens or neutral policy rates.

Hedging rather than positioning for a secular bear market

We are not looking for a secular bear market in bonds. We expect bond yields at or below current levels in both the US and euro area for the remainder of the year. Negative yields will continue to power portfolio reallocations. The Brexit impact is yet to appear in the data. The ECB is expected to announce an extension of its QE programme. China is expected to lose momentum in the second half of the year. Our colleagues are looking for a correction in equity markets. Pension liability hedging will likely support demand for duration in the US, and demographics will likely lend support to fixed income generally

While not unexpected, BofA's conclusion is somewhat ironic: it is the same presented by equity advisors every day: "stay invested", specifically, Preusser says that "being invested is the most important thing."

One caveat from us: all of the above is, in theory, correct. In practice, with momentum and crowding by far the two most important "investing" factors in a world in which collateral scarcity means most bondholders will seek to sell first and ask questions later, we doubt the proposed yield trajecteory will look anything like what BofA has proposed. In fact, if indeed central banks make a coordinated push for unwinding rate exposure, it is feasible that HFTs, algos, and trigger happy 17 year old hedge fund managers will shorten the entire 30 year proposed period of blowing out yields to a dramatically shorter time horizon, one which will - as we have warned before - lead to staggering losses for those who mark their bond exposure to market (obviously not central banks).

Ultimately, what the shape and severity of any bond selloff will look like in practice will depend entirely on central banks, who are unlikely to stand inert should a major selloff materialize as they seek to continue nudging the market in any one given direction for the duration of this so-called "renormalization" phase.