The case for dynamic asset allocation for insurance portfolios

19 April 2017

Eugene Dimitrou, head of insurance solutions at Columbia Threadneedle and Maya Bhandari, fund manager and a member of the asset allocation strategy group at Columbia Threadneedle, argue the case for dynamic asset allocation for insurance portfolios, rather than the ‘set and forget’ strategy of the past.

For many institutional investors a ‘set and forget’ strategic asset allocation strategy has worked well for the past 35 years. Equities have provided strong returns, and the portfolio efficiencies from owning core government bonds in conjunction with equity have been phenomenal, as long-dated bond yields fell steadily from around 14% in 1981 to 2% by the start of this year.

[This is based on a 10-year US Treasury bond yield, showing 14.74% on 10/8/1981 and 2.44% on 2/1/2017. Source: Macrotrends]

This bond market rally has been a steady and substantial tailwind, elevating valuations of all financial assets. However, a number of factors at play today mean that this ‘set and forget’ approach is no longer set to work for insurers.

In the last four decades, during which time the Bank of England and international central banks progressively cut the base rate of interest, combining equities with long-dated bonds at a typical 60:40 equity/bond mix resulted in healthy returns. The 40% side of that asset allocation strategy helped deliver bond returns that surpassed many people’s expectations, while keeping overall portfolio volatility low. Simply put, over the last three to four decades such an asset mix had around the same volatility as a pure 100% bond portfolio, but with a better return.

Changing times

For that strategy to continue working, institutions – including insurers – require the bond market rally to continue; but there are many reasons why that may no longer be the case, some of a cyclical nature, others more structural. And this will have important implications for how insurers manage their asset allocation going forward.

  • For 3½ decades now, we have enjoyed a “positive labour supply shock”, which started in 1970 with a sharp increase in the working age population and was bolstered considerably by the advent of China and E Europe in the 1990s.
  • The result of globalisation and associated labour market convergence => collapse in labour power in the West, with falling real wages, lower L/GDP, and companies incentivised to substitute capital for labour.
  • As neutral real CoC fell in West, PV of future cash flows went up and asset markets boomed. We think these demographic trends and lower real rates that we have enjoyed for 35 years now could soon be challenged as
     
  • a) China reaches Lewis turning point – when surplus rural labour dries up pushes up urban wages and labour bargaining power;
  •  
  • b) Folk in the west move into retirement and consume> save. Hard to call a turning point in a long run cycle

This seems likely to be exacerbated by an increase in protectionism and/or shift away from globalisation following the UK’s decision to leave the European Union and the election of Donald Trump to President of the United States.

There is also an increasing likelihood the discount rate will start to rise. Monetary policy saw discount rates move ever lower and QE-related bond purchases moving ever higher.

Less positive environment for bonds 

But as monetary policy gives way to fiscal stimulus, policy rates and inflationary expectations appear likely to rise, which is a less positive environment for bonds. Moreover, on a valuation basis bonds are now richly-valued and at current levels are unlikely to respond well if Trump’s tax proposals and infrastructure spending proposals are delivered.

That is not to say that bonds will no longer remain an important part of insurers’ portfolios. Bonds are key under the Solvency II framework in matching liability duration, as insurers look to find the right balance between asset class yields, duration management and their capital cost.

Active asset allocation

If the above factors look set to end the environment in which bonds can generate the returns insurers require – at a palatable level of volatility – what should asset allocators do about it? And how can insurers adopt an asset allocation approach that satisfies capital constraints relating to Solvency II?

We expect equities over the long-term to produce an average return of around 6% per year, though they are highly volatile and can experience draw-downs of up to 40%.

Typically, these drawdowns tend to occur at precisely the moment when you least need them to, when other assets in your portfolio are also falling in value. In a world where bonds are no longer the answer, asset allocators need to think about how they can get an equity-like return while also stemming volatility.

Insurers can benefit from an unconstrained, diversified approach, using dynamic asset allocation. Crucially, the Solvency II framework allows for this flexibility. Asset allocators must marry this flexibility with a more nimble asset allocation approach to progress beyond the traditional 60:40 blend.

 

Looking to the future

Ultimately, while insurers’ appetite for alternative assets has merit, the vast majority of investments will be represented by conventional asset classes.

With that in mind, where are we seeing opportunities in our unconstrained universe? Broadly, we sit at neutral equities, credit, and cash; we favour property, and commodities; and in general dislike government bonds.

Within equities we favour Japan, Asia ex-Japan and Europe; are neutral the UK and emerging markets, and are underweight the US. We are neutral credit, but retain a preference for high yield.

As things stand, volatility across asset classes remains relatively low but is likely to rise in our view across the course of this year and next year.

For our Threadneedle Dynamic Real Return Fund, volatility creates opportunities. Indeed, it is also worth remembering that volatility is not an asset manager’s enemy.

If we rewind to last year, markets were up and down in both January and February, and by the middle of February didn’t have a big risk appetite. We then took on more risk because volatility had returned and asset markets had cheapened, which opened up a number of opportunities for us. Volatility can be a positive for active asset allocators.

In conclusion, we believe asset allocation managed dynamically to adjust to changing correlations and current market conditions provides opportunity for better risk-adjusted returns than strategic asset allocation alone. But dynamic asset allocation requires a certain skillset, which may suit insurers equipped with large internal resources of experienced investment personnel.

For others, it may make more sense to outsource this expertise to external asset managers. Either way, insurers must be more active in how they approach asset allocation in the future if they are to generate the risk-adjusted returns they require.