Longer bonds pay more with better diversification

Introduction

In this piece the following three main topics are covered:

• Part one of the piece suggests that the longer end of the Australian dollar (AUD) fixed interest rate swap curve is relatively cheap on an historical basis. After looking at the various reasons for this cheapness, we express a view with regard to the possible future direction of the longer end spread, known as the 30-10 swap spread, or the interest rate differential between 30 year and 10 year swap rates.
• Part two of the piece suggests that given the view that the spread could narrow, we look how 30 year swap has performed as an alternative to the usual UBS composite 0 + year index, within a portfolio asset allocation context. Here, we find some very interesting results. In particular, we find that while longer bonds can be very volatile, they can more than offset the volatility of equities; the longer the bond, the better it offsets equity market volatility. Longer bonds boost return because they pay the investor for extra risk, and if the 30-10 spread is cheap, then the prospects for superior performance should be sound, and
• Part three of the piece provides some possible options for taking advantage of longer bond cheapness.

Part one: Relative value of longer AUD swap curve

Thirty year swaps are now cheap to the ten year swap, as Figure 1 indicates,

Figure 1

Several reasons exist for the cheapness.

Reason 1: Steepening US curve

First, the US curve has steepened, as rates have fallen, as Figure 2 indicates,

Figure 2

Notice how the AUD 30-10 spread has now compressed to the US 30-10 spread, as shown by the green line, in Figure 3 below,

Figure 3

In other words, the AUD 30-10 spread is quite wide, relative to the US 30-10 spread, while also being wide on a historical basis.

Reason 2: Lower rates

As rates have declined, the market has become nervous about rates reverting to somewhat more “normal”, or higher, levels which has led to a steepening of the 30 year part of the swap curve, as shown in Figure 4 below.

Figure 4

Hence, participants have reduced duration and that has meant that there have been more sellers of the 30 year, and more buyers of the 10 year, leading to a steepening of the 30-10 spread in the Australian swap market.

Reason 3: Japanese Yen

Many commentators have argued that the swap curve steepness is mostly about the currency, specifically the AUD-JPY cross currency rate, as Figure 5 tends to indicate,

Figure 5

Here, the weakness in the AUD around 2009, forced investment in the longer end by the Japanese, so the long end became inverted. Equally, recent strength in the AUD is argued to have seen the long end steepen. However, the impact of the currency does not explain the situation before the GFC, where the long end was inverting. Also, the current levels of AUD/Yen are equivalent to those seen before the GFC, and the curve slope is almost the complete opposite of what it was before the GFC; steep now and inverted before the GFC.

Hence, we need to explain the inversion of the long end, before the collapse of the AUD. One might suspect that the inversion of the long end of the swap curve was mostly to do with the lack of Commonwealth Government issuance, before the GFC, as liabilities needed to be covered in a market with insufficient assets to allow such coverage to proceed.

In other words, the current steepness of the AUD 30-10 swap spread reflects a combination of the following:

• a steepening of the US 30-10 spread,
• strength in the AUD, relative to the Yen, as sellers of the currency lead to selling in the longer part of the AUD swap, and
• lower rates are making market participants nervous about the longer part of the AUD swap curve.

Forecast

Recent moves by the BOJ have flattened the long end of the JPY curve substantially, and that flattening will have an impact on global curves, including the US and AUD curves. As the below chart indicates, the JPY 30 year les 10 year swap spread has now flattened substantially, and the impact of such QE should be for a flatter AUD curve, as shown in Figure 6 below,

Figure 6

Part two: Portfolio context

Given that we have briefly established the relative cheapness of the 30 year part of the curve, as well as the reasons for this cheapness, one now needs to see how the longer part of the Australia curve fits into a portfolio context. Zero coupon bonds are one way to take advantage of the longer dated part of the curve, as well as the longer dated semi-governments, like the QTC 2033. Below, Figure 7 shows the history of the 30 year swap yield since 2001,

Figure 7

Here, we note that while we have provided much detail on portfolio allocations using the standard market benchmark; the UBS composite 0+ year index, one of the main problems with this benchmark is that it is relatively short duration, from a global perspective. This is because Australian issuance is somewhat shorter than global markets, and because the index includes securities between 0 and 1 year maturities, which are typically excluded in global fixed income benchmarks.

If we substitute the 30 year swap for the UBS 0+ index, we can see that the volatility in the swap is more than a match for the All Ordinaries Index in terms of annual volatility, as shown below in Figure 8, where the annual return of the 30 year swap is shown by the light blue line, while equity annual return is shown with the dark blue line. The reason that the 30 year swap is so volatile is that it has a very long duration, and while it varies it has averaged 14.50 years. That means for every 100 basis points move in rates, the 30 year swap provides a capital movement of 14.50%, so a fall of 100 bps, means a mark-to-market capital gain of 14.50%, and the reverse is also true.

Figure 8

Since 2001, which is where the available data begins, the average annual return on the 30 year swap has been 9.66%, with an annual volatility of 13.65%, while equities has returned an annual average 9.07% over the same period, with an annual volatility of 19.17%. Now, if we combine the 30 year swap as an alternative to the use of the UBS composite index, and adjust the weighting to fixed income down to 50%, given the large volatility of the 30 year swap, then the results are very interesting, as shown below in Figure 9, where we look at the annual returns for two different portfolios:

• 50% 30 year swap and 50% All Ordinaries, (the dark blue line) and
• 75% UBS composite 0+ year and 25% All Ordinaries (the light blue line).

Figure 9

As the above analysis indicates, the swap portfolio tends to beat the UBS composite index. In fact, the average return of the 50% 30 year swap and 50% equity portfolio is 9.36%, with an annual volatility of 6.31%. In other words, we can improve on the annual return of two very volatile asset classes by combining them, as equities had an annual return of 9.07% with a volatility of 19.17%, while the 30 year swap had a return of 9.66% with a volatility of 13.65%. Importantly, such a combination of asset classes cuts risk dramatically. In searching for a reason for this result, the above time series of annual returns tells us why; negative correlation. As equities fall in price, the 30 year swap rises, and the reverse is also true.

We chart the outperformance of the 30 year swap portfolio, relative to the UBS composite portfolio below in Figure 10,

Figure 10

Several reasons, among others, are available for this outperformance:

• the 30 year swap yield is not dragged down in yield by the Commonwealth bond market,
• investors are being paid more to take more risk in longer bonds, and
• yields have fallen, although the period covered includes many times when rates were higher than the post GFC levels.

As we argued extensively at the start of this piece, the flattening of the 30-10 spread, as initiated by the BOJ should support the AUD 30-10 spread; forcing it to flatten. Two possible routes to this flattening are apparent:

• short rates rise and the 30 year does not rise as much, or
• the 30 year falls in yield, and short rates remain unchanged or go lower.

While we favour the latter route, the value of the longer part of the AUD curve should now be apparent, and the benefits of combining longer bonds in the portfolio should also be apparent.

Part three: Possible options to exploit the cheaper long end

Option one

Zero coupon bonds are a very effective mechanism to gain exposure to longer dated rates. Zero coupon bonds allow investors to gain access to long dated interest rates without allocating 100% of capital to the asset.

For example: $1m face value of NSWTC 23 Nov 2020 zero coupon has a price of $742,243.87, when settled on 17 April 2013, and the bond has a basis point value of around $542 (basis point value being the change in the price per basis point or 0.01% change in yield).

Instead of paying $100 for the bond, the discounting of the final payment brings the market value to around 75% of the face value of the bond. In other words, you pay less for a zero and yet you get the interest rate risk that helps you diversify your equity portfolio. Income can be provided from other parts of the portfolio, such as cash and equities.

Option two

Invest in longer term QTC 2033, or the longer term ILBs. In the longer end we recommend conservative credits, as the longer maturities are very sensitive to perceptions of credit worthiness.

Conclusion

AUD 30 year swap yields reflect the caution in global markets about low rates, and the prospect that this is an ephemeral or short-lived position, where rates will soon rise. However, recent moves by the BOJ, and continued guidance from the US FOMC, suggest that lower rates might be here for longer, and the gradual settlement of the market, in terms of longer term rate expectations, should see a compression of the 30-10 spread.

Looking at the 30 year swap from a portfolio asset allocation perspective, we find that it adds return and some risk to what the market benchmark typically suggests. When we compare the return of a 50% 30 year swap/50% equity portfolio to a 75% UBS composite/25% equity portfolio, we find that average portfolio return increases, from 7.14% to 9.36%, an increase of roughly 30%, while risk increases, from 3.53% to 6.31%, and increase of roughly 80%. Hence, the 30 swap portfolio lifts risk and return substantially, because longer interest rate risk provides higher returns, and higher risk. As we note, the prospect of higher returns from the 30 year part of the yield curve is now apparent.

Several options remain available for exploiting this cheapness of the longer part of the interest rate market, and zero-coupons are the most effective mechanism.