In focus

How do you construct portfolios in a world of zero interest rates?

Interest rates are at zero. Developed market bonds yield not much more. This has significant implications for returns over the next decade and for portfolio construction with listed assets.

So, do government bonds still have a place in portfolios?

For investors who require liquidity, we believe the answer is highly dependent on risk appetite. Some will have no choice whether to hold bonds. But for others we believe the portfolio can be made to work harder through the use of liquid alternatives, thematic investing and greater use of stock selection.

For many investors, a portion of assets with less liquidity will be suitable and we show how such assets can be part of the solution to the return challenge.

Why is there a portfolio construction problem?

In our paper ‘Can static 60/40 portfolios still deliver?’, we discussed the classic asset allocation theory. Under this theory, government bonds are put into a growth-oriented multi-asset portfolio to offset any short-term losses in equities, resulting in a smoother path of returns.

However, a zero interest rate environment leads us to reassess the roles that conventional assets play in our multi-asset portfolios. And it makes us consider whether we might need to look elsewhere to meet return objectives.

As can be seen from Figure 1, the efficient frontiers for traditional equity/bond portfolios in every decade (or every economic cycle) are different. Over the period following the global financial crisis (2010s), the world went through an atypical economic cycle, in which the US market exhibited exceptional performance and there was a monetary tailwind for bonds. So, what’s changed?


Short term US government bonds now yield around zero. Not only does this impact returns, but it meaningfully raises the cost of buying the US dollar, either for hedging or return-generating. We have included in Figure 1 our expectations for the 2020s and you can see this is much lower than the 2010s.

Looking ahead, things don’t appear as attractive. The main issue is the evaporation of yield from cash and bonds. For some investors targeting a specific return, such as defined benefit plans and endowments and foundations, this expected reduction in returns that affects all liquid asset classes will make achieving their target significantly more difficult.

It will possibly result in further contributions from the company or sponsor. For individuals in defined contribution plans, a significant reduction in returns is likely to require additional personal contributions or a delayed retirement date.

Additionally, bonds had a persistently negative correlation with risk assets, as can be seen in Figure 2, meaning that they protected the portfolio in the event of a growth shock impacting equities. However, we expect this relationship to be less stable and therefore less reliable as a hedge in a zero interest rate world with the lack of an interest rate cycle.


Federal Reserve (Fed) chair Jerome Powell recently commented that he was prepared to run the economy hot, and would not be moving to taper the Fed’s monthly bond purchases. Despite this, we are not currently concerned about a pick-up in inflation to levels that would require a sharp increase of interest rates to keep it under control. This means that the lack of an interest rate cycle could have an impact for some time to come.

Although we’re not currently concerned about inflation, we believe it is important to understand valuation of asset classes in the context of the cycle. And to be able to adjust the asset allocation appropriately at different stages of it.

Government bonds are unlikely to be as useful in portfolios as they were

With interest rates at zero in many developed markets and with a less stable correlation between equities and bonds, government bonds will no longer provide the cushion in a portfolio that they once did.

So, is there anything that can adequately replace bonds and cash in a portfolio? We think not.

Cash clearly provides protection in the sense that it is unlikely to fall in value in the environments when equities fall. However, cash rates are currently at historic lows and in periods where inflation is above the cash rate, cash is expected to deliver negative real returns.

That said, for the most conservative investors, cash and government bonds may continue to be the only solution.

Potential liquid replacements include:

  • Gold
  • Currencies

Gold is a potential hedging asset but it is heavily influenced by real interest rates and has no yield (see Figure 3). Ever since injections of central bank stimulus became commonplace, which have driven interest rates lower and inflation expectations higher, equities have been negatively correlated with real interest rates.

Since gold is also negatively correlated with real interest rates, this explains why it has been positively correlated with equities. This warns us against relying on gold to hedge us against the next equity market drawdown. In periods of extreme crisis, when market order breaks down, such as in 2008, there are many market participants who can be desperate for liquidity, irrespective of whether it is to rebalance, meet budget needs, de- risking or collateral.

This was also true in March 2020 when the initial wave of the Covid-19 pandemic hit developed markets. Figure 3 shows that gold is unreliable in protecting against equity drawdowns.


Currencies that generally rally when equities fall are few and far between. Yen, Swiss franc and US dollar are typically considered, but these can be inconsistent in their hedging properties (and the former two are also subject to idiosyncratic risks and government/central bank influence).

Shifting away from the idea of ‘hedging’ equity risk with defensive assets, we can instead look at ‘diversifying’ the return-generating part of the portfolio away from equities and towards alternative return-generating assets, such as credit. However, in the context of growth assets, while these assets prove it is possible to earn higher yield and diversify using fixed income, in order to do so you must take on more risk (see Figure 4).


In summary, we don’t think there is an adequate replacement for bonds and cash as a hedge against equity risk in the portfolio. So, depending on risk appetite, we either need to accept lower overall returns, or seek to diversify within the return generating and risk-reducing parts of the portfolio separately.


Using a broader mix of asset classes

In Figure 5 we show how a 60% allocation of a simple 60/40 portfolio can be improved (shown by the light blue line) by investing in a broader selection of liquid assets. This shows that aggregate returns from the past may be difficult to repeat. But diversifying equities into alternative liquid growth assets, and US Treasuries into some alternative liquid defensive assets, can result in an improved risk/return profile.


Many investors have looked increasingly to illiquid assets to improve returns given the challenging return forecasts from public markets. We can see in Figure 6 that inclusion of illiquid private assets in the 60% allocation allows the portfolio to get closer to the 2010s efficient frontier. Yet the need for investors to generate attractive returns without excessive risk remains in the liquid portions of their portfolios

For those institutions that require liquidity in their portfolios, as we do in our liquid multi-asset portfolios, a choice has to be made about whether bonds should feature in the portfolio at all at these levels.

If there is sufficient risk budget to be able to move up the risk spectrum and abandon bonds, the efficient frontier is no longer pinned on the left hand side. Instead there is a blurring between the defensive and growth portions of the portfolio. Effectively there is just one growth portfolio. In Figure 7 we show another frontier in pink, which only includes liquid assets but now doesn’t fix the return-seeking and defensive weights at 60% and 40% respectively.


Importantly, for those with a higher risk tolerance, there is a way to achieve returns equal to or higher than those predicted with a portfolio consisting solely of global equities. This can be done with considerably less risk by diversifying the portfolio (in an unconstrained manner) into liquid alternative assets.

Some investors may not have the governance structure to venture into some of the alternative return-seeking and defensive assets that we include in this modelling. In which case they’ll be constrained to portions of the light blue curve. Furthermore, the inherent unpredictability of alternative assets means that, alongside volatility, there are other risks that are unaccounted for in the efficient frontier. For the most risk-averse investors seeking precise management of risk, it might be the case that only the original dark blue line is appropriate.

In conclusion, while investors are unlikely to see the returns of the 2010 equity / bond frontier in the 2020s, expanding the universe and reducing allocations to nominal bonds appears likely to improve returns without being exposed to excessive risk for many investors.


As investors increasingly made use of passive strategies, they favoured lower costs and reliable beta versus expensive and uncertain alpha. Yet alpha may once again become useful to investors as returns purely from beta decline and fees for active management become more competitive. Stock selection is an important source of returns when betas are significantly reduced. This has also recently been highlighted by Willis Towers Watson, who said that “going forward, we expect a better environment for skilled stock pickers to generate alpha”.

Additionally, since bonds have now become less reliable as a diversifier, investors can make greater use of stock selection. Figure 8 shows that stock selection is relatively independent to market direction so will add diversifying qualities to a portfolio.

However, as we know, stock selection is a negative sum game after fees so superior selection of the manager and/or the theme is paramount. There are a number of ways in which an asset owner can stack the odds in their favour. Academic evidence points to concentrated portfolios being the most reliable source of positive alpha, so active equity exposure should be focused on concentrated strategies . This is evidenced in Figure 9, where we look at the average excess return of active global equity funds over their benchmark and group them by number of equity holdings.


Themes could also be a way in which to generate improved returns in the future, although it is difficult to model the likely impact of these on the portfolio, particularly in relation to diversification and risk. Using themes that have worked in the past is unlikely to be helpful to understand the themes that might be influential in the future. Future themes might include ‘disruption’. In the context of the global economy, disruption can be the result of:

  • Technological innovations (for example, the internet, mobile phones or artificial intelligence)
  • Changing consumer habits (such as the switch to internet shopping)
  • New regulations or government policy (such as the switch towards renewable energy sources or tougher regulations on the sale of products such as tobacco or alcohol).

Other themes could be carbon neutrality (through equities and bonds) or global city real estate in locations with strong infrastructure and a skilled labour force.

These themes often do not have a long track record that we can integrate into our efficient frontier analysis. We also recognise that some future successful themes may overlap and more than likely result in a growth bias to a portfolio. It may be appropriate to accept this style risk in order to benefit from these type of themes.

The skill in managing themes in a portfolio is to be able to identify the themes of the future that are distinct, investable and structural and to avoid those that have reached the “hype” stage. In addition, suitable resources are required to understand and manage these often concentrated portfolios and we will need to size these positions appropriately in diversified multi-asset portfolios.

Another area where active stock selection should be able to differentiate from a passive approach in the coming years is investing by integrating a sustainable approach into the investment process. 

The data used to construct passive indices is backward-looking and sustainability data is often only released annually, therefore the index allocation does not reflect what companies are doing to improve for the future. One important example is the move towards a low/no carbon economy globally. Significant and disruptive changes are needed to limit rises to the 1.5- 2°C commitment global leaders made in Paris. Companies that are likely to impact this transition may not register as scoring highly on their passive ‘environment’ score today but what matters is investing in companies that are improving or have plans to improve.


Returns for the next decade are likely to be considerably lower than for the 2010s, given that short term bond yields are now around zero. Additionally, hedging equity risk using bonds is expected to be less reliable than in the last 20 years, but we don’t think there is an adequate replacement for bonds and cash as the defensive asset. As a result, cash and government bonds will continue to be the only solution for the most conservative investors.

For those who are not constrained to holding government bonds and/or have a higher risk budget, we believe that portfolio construction needs to be reconsidered in a zero world. This means diversifying into other assets, investing in the themes that will play out over the next 10-20 years (or longer in the case of those saving for retirement) and leaning on stock selection as an additional source of return.

The 2010s might have been about passive investing and riding the wave of equities and bonds, but the 2020s are likely to need much more effort to even approach the returns of the 2010s. It’s going to require some radical rethinking about portfolio construction and using a wider set of tools.

This report is available as a PDF below.