Neither a borrower nor a lender be.
Shakespeare, Hamlet
Bonds can provide a safe harbour in times of equity market volatility, and be a critical diversifying element in an investment portfolio.
Yet as the journey to financial independence has progressed, one of the least examined parts of the my FIRE portfolio has been the fixed interest and bond components. This is despite the fact they constituted around a quarter of portfolio assets at the start of the journey.
With the end of ‘big rebalance’ into Australian equities at least in sight, keeping to the target asset allocation may require purchasing bonds or fixed interest instruments over the coming year. This would represent the first significant direct purchase of bond assets since 2014. In turn, this means benign neglect of this part of the portfolio is no longer feasible.
To help establish what this potential future purchase should be, it felt critical to know what I had already in the bond part of the portfolio. The target allocation for bonds is currently set at 15 per cent of the total FIRE portfolio.
This longer read article explores these current bond and fixed interest holdings and seeks to reach a possible choice correct for my personal circumstances and goals. Its focus is not offering advice or fully explaining the operation of bonds (pdf), subjects recently (and better) covered by others.
History of bonds in the portfolio
Bonds have formed at least a small part of the portfolio since its inception. From 2009 to 2014 over $130 000 of new investments flowed into bonds through contributions to a number of Vanguard retail funds. This drove the portfolio allocation to bonds to a maximum of 29.5 per cent of total portfolio assets in July 2014.
There was no particularly deliberate logic or consideration behind this increase, in part it probably reflected inertia from having some regular automatic investments in place, and in part, it also likely appeared a relatively safe haven in the immediate aftermath of the Global Financial Crisis.
This lack of active choices around this asset class is also reflected in the short consideration of bonds in annual reviews, where I simply noted that they were included for diversification and to reduce portfolio volatility.
An overall perspective on changes in the absolute level of bond and fixed interest holdings in the portfolio during the past decade is given by the chart below.
In January 2018 bond and fixed interest holdings reached their highest absolute level of $280 000 or 20 per cent of total portfolio assets. In the portfolio review that year I set a bond allocation of 15 per cent – comprising 5 per cent Australian and 10 per cent international bonds, a division not informed by any particularly strong rationale.
A year later I replaced this with a ‘naive diversification’ approach of an equal split of 7.5 per cent each. As of October 2019, bond and fixed interest holdings are around $257 000 or 15 per cent of the total portfolio.
Examining the stores – what lies beneath the current bond holdings
For the journey so far, my knowledge of what was within the bond holdings of the portfolio was limited. From existing worksheets I knew the split between Australian and global bonds, and the absolute level of each. Beyond that was an undiscovered country.
Part of the issue was finding time to delve into what was a relatively small portion of the overall portfolio. Due to holding bonds through no less than four different Vanguard funds, as well as some other smaller fixed interest holdings through Ratesetter peer-to-peer lending and Raiz, establishing full visibility of the holdings was non-trivial.
Yet a rather complicated multi-step Excel sheet did enable a fuller picture to emerge. Each of the Vanguard retail funds actually invests slightly different proportions in just two underlying wholesale funds:
- Vanguard Global Aggregate Bond Fund (which is currency hedged); and
- Vanguard Australian Fixed Interest Index Fund
Both of these funds invest in a wide range of bonds and fixed interest investments. These include:
- Treasury notes and government bonds
- Corporate bonds (including financials, industrials and utilities)
- Mortgage-backed securities
- Government-related entity bonds
- Securitised bonds
These are both extremely diversified. Between them they include bond issuances from over 2 200 different issuing entities, and over 8 000 separate holdings (as issuers normally issue bonds regularly, on different terms or for different durations).
Distributions from the bond holdings through the period have been uneven. An example of this can be seen in the distributions from the Vanguard Diversified Bond fund, which saw no new contributions over the past four years.
Despite this stable balance, in this period distributions for this fund have bounced from as low as $140 per year, to a high of $5270. This is likely to be associated with the funds realising substantial capital growth in periods of lower interest rates, and distributing these gains as part of rebalancing to each funds target allocation.
Looking deeper into the holdings – analysing where and what
Once broken into their component parts, analysis of the actual balance of the bond and fixed interests holdings is possible.
The first thing to note is that while to the overall target allocation of 15 per cent has been reached, the balance of holdings is heavily tilted towards international bonds.
This position is inconsistent with my current naive target of evenly splitting domestic and global holdings. This becomes even clearer when the overall breakdown of total holdings is considered below.
This shows that global treasury notes, corporate bonds and mortgage backed securities constitute the majority of bond and fixed interest holdings.
By contrast the two largest Australian holdings – government and government-related bonds – represent only around a quarter of the bond and fixed interest portfolio. Ratesetter peer-to-peer lending represents only 7 per cent of total bond and fixed interest holdings, and without specific action this will continue to diminish as the underlying loans are paid back.
It’s important to note that global bond holdings are themselves further diversified across multiple countries, with the largest holdings being exposed to United States, European Union and Japan.
The full split of all holdings can make it difficult to see clearly the relative weights of different types of holdings. The chart below removes the distinction between global and Australian holdings and simply examines the type of bond or fixed interest holding.
Here some observations can be made.
Government or government-related entity debt is the single largest component. Corporate bonds make up just under 20 per cent of holdings, with additional small holdings in mortgage-backed securities and peer-to-peer lending.
The question then looms, is this the right mixture of holdings, and how would an answer to this be established in current market conditions?
I have never been able to get this business of loans and interest into my head. I have never been able to understand it.
Philip II of Spain to his Minister of Finance, 1580
Becalmed – buying and holding bonds in a ‘zero rate’ world
If reaching the portfolio target asset allocation requires a purchase of bonds in the near term, in many ways this will be a difficult move to rationalise and execute.
There is a plausible case that Australia and the world are at the end of a four decade secular bull market – or even bubble – in bonds. Since the early 1980s across most developed markets bond yields have fallen to historically low levels.
Significant portions of government bonds around the world, and even some corporate bonds, are currently trading at negative yields. That is, borrowers are paying for the privilege of loaning their money out to debtors. One day, this extreme and unprecedented global trend could reverse. Bond yields could revert to levels closer to, or even above, their historical average.
There is, however, a significant chance that bond yields seen in the past will not be seen for decades to come. There is also the chance that further falls in yields could occur, leading to further gains in the capital value of bonds. Which of any of these potential futures will play out is difficult to forecast.
Yet this is not the only consideration. The original function of bonds in the portfolio is to reduce volatility, and therefore the primary consideration is not simply their absolute performance, but rather how their returns can be expected to move in relation to other parts of the portfolio.
Traditionally bonds have had low correlations to equity returns, which is a critical consideration given the set 75 per cent target allocation to equities in the portfolio. This has not been uniform, so for example I personally recall 1994 being an exceptional year in which both bonds and equities experienced sharply negative returns. An example of this record can be found in this calendar year return comparison of Australian bonds and equities here (see p.19).
It is possible that this correlation could break down into the future, although there are some underlying drivers for bonds being less volatile than equity, it is perfectly possible for these to be offset by external conditions for prolonged periods.
All bonds are not created equal
It is also worth noting that different types of bonds will have different correlations with risky equities. For example, a government with taxation powers and the option of printing local currency can technically never be forced to default on payments of domestic government bonds, regardless of equity market conditions.
By contrast, although corporate bond holders stand in front of equity holders in the queue for available company cash in difficult times, some of the same market forces could easily be expected to exert themselves on a corporations capacity to pay debt and provide equity returns, giving them higher correlations.
Similar considerations apply to the risk characteristics of peer-to-peer lending – which can often be used for car loans, small renovations or other similar purposes. Risks to repayments of these loans are likely to be significantly correlated to equity market conditions, and to changes in employment and wages conditions. Thus they may produce lower diversification benefits (even where the income is attractive) than traditional bonds.
Flags of convenience – are there benefits in holding offshore bonds?
A major finding from reviewing the detailed holding is the level and breadth of foreign bond exposure. Yet aside from the not insubstantial psychological satisfaction of apparently holding the equivalent around $200 in Romanian or Panamanian bonds, is there actually any financial benefit in holding foreign over domestic bonds?
This question is the focus of an excellent 2018 Vanguard research paper (pdf) Going global with bonds. This analysis uses 30 years of historical data to make observations on the value of global diversification in bonds. It finds that:
- Global exposure reduces volatility. Global bonds hedged into the local currency had significantly lower volatility than holding just domestic bonds, across major developed nations including Australia.
- In part by reducing individual bond risks. Global bond holdings reduced exposure to local market risk factors and gave better diversification to a range of interest rate and inflation risks that impact bond returns.
- Hedging global bonds smooths volatility. Whilst hedged and unhedged global bond returns will be similar over long periods, hedging historically smooths volatility – indeed, significant unexpected falls in the Australian dollar would be needed to mathematically justify not hedging.
- And offers more downside protection. Hedged bond portfolios have offered better protection in adverse market conditions than their unhedged equivalents.
- Which can help reduce risk even in broad portfolios. Adding hedged global bonds to a mixed equity and bond portfolio can offer some modest portfolio volatility reduction (of around one per cent).
A necessary caveat is that these findings are a function of the time period selected. Yet the analysis is also the best empirical review I have seen so far on the allocation issue which takes into account recent and prevailing bond market conditions.
Departing course – which types of bonds should the portfolio include?
Finally, there is the question of whether the portfolio should seek an alternative exposure to different types of bonds and fixed interest instruments than that currently held. In principle, under modern portfolio theory, there is a specific bond portfolio that would minimise risk and produce an optimal risk-adjusted return for every different equity portfolio and investment target.
Yet having adopted a market capitalisation weighting approach for equities, through equity index ETFs such as Betashares A200 and Vanguard’s Australian equities ETF (VAS), it would not be consistent to start seeking to make active sectoral ‘bets’ in types of bonds and fixed interest. Especially if there was no reason to believe a special insight or information would enable this active approach to reliably add value.
At present, therefore, being ‘market capitalisation weight neutral’ (i.e. holding different bonds in roughly the proportion of their total market value) is a more efficient approach given the low likelihood of outperforming professional bond market participants and higher risks.
At the time of writing, with a gradual reduction in Ratesetter balances expected to continue, the bond portfolio is allocated quite close to global capitalisation weighted benchmarks (i.e. it reflects the size and make up of bond and fixed interest markets globally). This is an automatic result of how Vanguard’s funds themselves generally seek to match Australian and global sector weightings.
Summary – applying the learnings
The exploration of the bond and fixed interest part of the FIRE portfolio has filled in a lot of knowledge gaps that should not have existed. Yet what I have found provides some confidence that thankfully this past neglect has not come at a significant cost or increased risk.
From the review a few useful points and lessons have been reinforced. To summarise:
- Reliance on Vanguard retail funds has built the foundations of a diversified portfolio. The bond portfolio is already well-diversified across different bond and fixed interest issuers, markets, countries and risk types.
- A ‘home bias’ for Australian bonds and fixed interest is not warranted. Based on the data and analysis above, there is no reason to target a level of Australian bond holdings at anything above around 1.5-2.0 per cent as there are no significant advantages of any ‘home bias’.
- Future bond investments will be made through market capitalisation weighted index vehicles. This could plausibly include Vanguard’s Hedged Global Aggregate Bond Index Fund (VGBND) or further investments in Vanguard’s Diversified Bond retail fund.
- Peer-to-peer lending should be considered as a separate type of fixed interest asset to traditional bonds and fixed interest. While peer-to-peer lending can have diversification and income benefits, it may not be a direct bond or fixed interest substitute. That is, it may not provide significant reduction in portfolio volatility should loan defaults rise in a downturn.
With these points in mind future decisions on investments in bonds will be made with increased knowledge and taking into account recent market evidence. And, importantly for the decision ahead, the contents of the stores below deck are known and accounted for.
Sources and further reading
Bernstein, W. The Intelligent Asset Allocator, McGraw-Hill, New York, 2000
Graeber, D. Debt: The First 5000 Years, Melville House, 2011
Macdonald, J A Free Nation Deep in Debt, Farrar Strauss and Giroux, New York, 2003
Vanguard Plain Talk Library Bond Investing (pdf)
Vanguard Research Going global with bonds: The benefits of a more global fixed income allocation, April 2018 (pdf)
Disclaimer
This article does not provide advice and is not a recommendation to invest in any specific bond or fixed interest instrument. Its sole purpose is to discuss bond and fixed interest investment issues relevant to my personal circumstances.
Even so, it is important to note that this narrow gap could stabilise or modestly rise once forthcoming December distributions form part of the average, replacing a higher placeholder assumption based on June’s figures.
There are significant boundaries of uncertainty around this estimate, and some simplifications. For example, it excludes Ratesetter and smaller individual shareholdings (which represent about 10 per cent of the holdings). It also assumes for simplicity equal ETF payments through the year.