Bonds: Forward Returns Don't Look Good, But Not A Place For Shorts Either

by: Long/Short Investments


Historically, the 10-year Treasury tends to provide good value when its yield exceeds that of nominal growth. That isn't the case currently.

I explain why bonds are not likely to materially appreciate or depreciate over the next 6-12 months.

I provide potential bond trade ideas (for those with longer time horizons) with respect to broader themes.

Argument: Speaking in terms of broad asset classifications, "long bonds" is a difficult trade to make money from. Bonds of all risk characteristics are expensive, and Treasuries don't look favorable relative to current/future nominal growth rates. Also, this isn't an asset class that one would want to independently short either, in my estimation.

I do, however, believe that investors with longer time horizons could consider pair trades going long back-end Treasury yields (i.e., longer duration) and short front-end yields (i.e., shorter duration), while also going long safer bonds at the expense of riskier bonds. I explain the rationale behind each below.


Treasuries - and fixed-yield bonds more generally - tend to follow the path of inflation. This only makes sense. As inflation increases, this eats into fixed yields, and investors expect to be compensated for it accordingly. Conversely, as inflation decreases, real yields increase, which pushes more investors into bonds and drives down yields.

In the diagram below, I use PCEPI inflation - the Federal Reserve's preferred measure which factors out the noisier elements of inflation tied to commodity movements - and track it against the 10-year US Treasury.

(Source: St. Louis Federal Reserve)

To plot this as a difference (10-year minus inflation), we see that the 10-year nearly always trades somewhere above inflation. The exception is at certain points during recessions or "low sentiment" periods, when risk aversion increases and capital preservation becomes the main investment motivation. This causes real yields to become negative. It typically indicates negative real interest rates and/or investor overreaction.

(Source: St. Louis Federal Reserve)

The all-time high spread between the 10-year and PCEPI inflation of 966 bps came in June 1984 and has been in a downtrend since. The spread is currently only about 20 bps, with PCEPI inflation around 2.0% and the 10-year at 2.2%.

Since 1960, the average spread has been 289 bps, with 269 bps as the median.

Bond prices are also a function of economic growth. Over the long run, nominal growth should more or less roughly match the 10-year yield by virtue of the fact that the 10-year is considered a "risk-free rate." In other words, it's very safe and highly unlikely to default, thus mitigating investors' need for a premium.

From Q1 1960 to Q4 2016, the average spread between nominal growth (using PCEPI inflation) and the 10-year has been negative 17 bps. The median spread has been negative 25 bps.

(Source: St. Louis Federal Reserve)

If we assume we're more or less currently trending along at ~2% growth and ~2% inflation, this would put this spread at negative 180 bps. This suggests that bond prices are currently too high relative to the amount of compensation investors are receiving.

The macroeconomic factors of just real growth and inflation can explain a reasonably high percentage of where the 10-year Treasury should be.

If the 10-year yield exceeds nominal growth, chances are bond returns (as an asset class, not just Treasuries or the 10-year itself) will provide positive returns as the market equilibrates the difference between the two down to around zero.

If nominal growth is greater than the 10-year yield, it's less likely that bonds will provide favorable returns, as it suggests that bonds are relatively overpriced relative to the amount of growth and inflation in the economy. As the "reflation trade" after the November US elections demonstrated, higher growth and inflation expectations will cause bonds to sell off in anticipation of higher rates and capital costs ahead.

For the sake of simplicity, we can compare by decade.

In the 1960s and 1970s, when nominal growth came in higher than the 10-year yield, bond returns disappointed, providing negative returns. However, from roughly 1980 forward, bonds have been in a bull market. Since the financial crisis, their prices have been driven further forward through central bank policies of ultra-low rates and quantitative easing regimens, wherein these institutions buy government bonds (and other forms of securities if they need to venture further up the risk curve) in order to support the financial and economic conditions of their economies.

Given recent events, such as expectations of stimulatory fiscal policy coming out of the US, there was the expectation that bond yields would rise again in conjunction with nominal GDP. In several different articles, I asserted this was unlikely to happen, and that 2.2-2.8% was a more likely range for the 10-year yield over the next 6-12 months.

The main reasons boiled down to:

  1. Structurally embedded global and domestic disinflationary forces. Briefly, a pile-up of exorbitant amounts of public debt, underutilization of natural resources, supply overcapacity in India and China, a stronger US dollar, and pull-forward of a large amount of consumption through 8-9 years of ultra-low rates.
  2. The difficulty in driving GDP growth when aging demographics and slowing innovation are secular, rather than cyclical, forces. Three percent GDP growth was formerly the norm, even somewhat yawn-worthy 10-15 years ago. Now, averaging even 2.5% growth would be difficult to achieve. The long-run growth rate of the US economy is probably around 1.8%.

Disinflationary environments are bullish for bonds. If growth continues to disappoint ahead, this will continue to be support demand for safe-yield assets.

At the same time, the Fed is undergoing a tightening cycle in which it will likely bring up the overnight rate to 2.00-3.00%, as well as potentially running off fixed-income assets from its balance sheet, which currently includes about $2.5 trillion worth of US Treasuries and $1.8 trillion worth of mortgage-backed securities. This is bearish for bonds.

As a result of this direct conflict of both bullish and bearish forces, I've been arguing for the past several months that mid- and long-term Treasury yields will be stuck in a holding pattern. Short-term rates are likely to rise faster than mid- and long-term rates as the Fed increases the Federal Funds rate. And this is what we've been seeing since it hiked rates in two of its past three meetings, with a compression of the 10-year/overnight spread ("10-0 spread") of about 50-60 bps.

At some point, this spread will invert (i.e., overnight rates will be higher than 10-year yields), and this will more than likely indicate that the business cycle has peaked. Shortly after, there will be a recession.

If we look at the 10-0 spread graphically, we see that nearly every time it goes negative, a recession follows soon thereafter.

(Source: St. Louis Federal Reserve)

Habitual pre-recession inversion is also true of other spreads involving short-term to mid-/long-term Treasury yields.

Only when this occurs do I expect that we will see bond spreads finally widen back out - especially between low-risk and high-risk bonds, which are currently trading too tightly relative to long-term default expectations. In other words, I don't expect this to happen during the current business cycle.

If the current administration is successful in getting through stimulatory fiscal policy measures, this may cause yields to increase and spreads between short-term and long-term yields to widen - but it will only be a moderate boost. Demand for Treasuries is simply too high. During a recession, the widening in spreads between short-term and long-term yields and low-risk and high-risk assets occurs quickly and with a high magnitude.

This, too, will naturally result in a correction in stocks, and investors will once again be able to see cheaper prices in the market.

Main Themes to Get Out of This

1. As the Fed hikes rates, it is likely to increasingly flatten the curve due to diminished inflation expectations (relative to the last 1-5 months) and ongoing demand for US Treasuries. Longer-duration securities, in this case, would be the better buy relative to shorter-duration securities.

To play this via ETFs (instead of futures contracts), you can buy, for example, a long-duration Treasury security (TLT, TLO) and short a shorter-duration Treasuries fund (BIL, SHY). Note that this may take several years to play out, but comes with the added benefit that given it's a long/short pair trade, there is no capital outlay involved outside of commission fees and any ETF-related expenses. It also comes with carry, as the long-duration security will provide a higher dividend yield than the short-term instrument.

2. Spreads between high-yield and Treasuries are near, but not at, all-time lows. The spread between the 10-year and B-rated debt is currently 80-90 bps off all-time lows reached during June 2014 and in the mid-2000s.

(Source: St. Louis Federal Reserve)

For investors with longer time horizons, this can present an opportunity to recognize that long-run default expectations don't properly compensate for the risk taken on. B-rated debt should trade at better than mid-5% yields (mid-3% yield in real terms).

Accordingly, if one were to go the ETF route with this, one could short a liquid high-yield ETF (HYG), while going long a safer ETF with similar effective duration to hedge out interest rate risk. With HYG's effective duration of 4.29, the most appropriate Treasuries ETF would be ITE with an effective duration of 3.99.

Note that unlike the previous trade idea, this involves negative carry in that you would pay out the 5.22% yield (at the time of this writing) on HYG and only obtain 1.21% yield on ITE. That means each year you would pay out 4% net on an unleveraged position, in addition to any commissions and related expenses. This creates some level of time sensitivity with respect to this trade, even in the absence of any net capital outlay.


I'm personally not bearish on the economy, stocks, or bonds. Forward returns of stocks have been bid down pretty low, and earnings will need to churn out above-trend results to justify current market valuations. Accordingly, the risk/reward of stocks is slanted downward. But that shouldn't be interpreted as being bearish on the market itself.

With respect to the economy itself, it's possible that the current expansion in the business cycle can last a few more years or even beyond that. It's been eight years since we exited the last recession, but the amount of time that has elapsed since the previous down-cycle doesn't have much bearing on the time until the next contraction.

Bonds are stuck in a holding pattern. At the moment, I prefer longer-duration Treasuries over their shorter-duration counterparts, and higher credit-quality bonds relative to non-investment grade bonds, which on aggregate don't adequately compensate for long-term default expectations. Betting on these ideas, however, is only suitable for those with longer time horizons.

Disclosure: I am/we are long TLT, ITE.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: Short SHY, HYG.