Its all about risk – Part 3
Dr Stephen Nash
As each week passes, we see more and more support for the groundswell that is leading Australian investors back to a more conservative asset allocation. Dr. Ken Henry kicked off proceedings, while David Murray followed, and now Lindsay Tanner has confirmed the validity of concerns about questioning the motives of funds that chased return, without adequate concern for risk. As Tanner indicates,
The thing I found troubling were responses to the very legitimate questions being raised by, in many respects, the ultimate heavyweight crew of Ken Henry, David Murray, Jeremy Cooper and Mark Johnson. All essentially saying that our super fund system is overexposed to equities … One thing I did learn during my time in government was when people of this calibre and background assert something serious like that you pay a lot of attention. Comments from people of this stature need to be viewed as a serious thing and should be listened to not just by super funds but by governments and by society more broadly … On any measure, your typical Australian superannuation fund is massively overweight equities*.
In Part II we supported the argument for a greater consideration of risk by arguing that Australians are investing their life savings, via their superannuation nest eggs in risky assets, while not receiving adequate return. In fact, the default asset allocation of 70% equities was shown to deliver about the same ratio of return to risk, as the equity asset class, since such a high allocation effectively swamps the portfolio in equity risk (click here to read Part I of the series and here to read Part II). In contrast, we suggested that the FIIG recommended asset allocation cuts risk in half while delivering the same return as the default strategy, using the conservative benchmark of the UBS Composite Bond Index 0+ years. In summary, the following two portfolios were considered:
- The “default” portfolio, with 70% equities, 25% bonds, and 5% cash
- The FIIG recommended portfolio allocation with 25% equities, 70% bonds, and 5% cash
This week, we look at why bonds provide lower portfolio risk without sacrificing higher annual return**. As we saw last week, taking the FIIG asset allocation, in preference to the “default” allocation, can really cut risk dramatically, as Table 1 shows below.
Source: UBS. ASX, Bloomberg, FIIG Securities, 1989 to March 2012
Why risk is cut so much
In general, the main reason why bonds cut portfolio risk is because they expose a portfolio to interest rate risk. As interest rates fall, bond returns rise, as can be seen in Figure 1 in the aftermath of 1994, and the period during the GFC. Bonds soar in times of adversity, again for very good reasons. Understanding why bonds outperform cash in times of adversity can be obtained by considering how central banks react to a collapse of sentiment. Note how bonds, on average, outperform cash by around 2.50%, as they have more risk than cash.
Specifically, if perceptions of growth fall dramatically, as was the case during 2007 and 2008, the RBA tries to offset this fall by lowering the cash rate. Such lower cash rates make the purchase of long bonds easier, by forcing the cash rate lower, and bond prices typically rise, because the underlying cost of funding falls and allows longer rates to fall in sympathy. This means that, most of the time, when equity returns are negative, bonds outperform. In other words, bonds are negatively correlated with equities, which is a huge positive for portfolio diversification. In order to illustrate the correlation aspects of bonds with equities, Figure 2 below separates the annual correlation of bonds to equities, under two conditions:
(i) where the annual return of equities is positive, and
(ii) when the annual return of equities is negative.
Based on the above rationale we find that in times of adversity when equity returns are negative, the correlation to bonds should be significantly negative, meaning that bonds can be expected to increase in price when equities decrease in price. On the other hand, when equity returns are positive, the correlation becomes somewhat insignificant. Here, when equities increase in price, and the pressure is off the RBA to ease monetary policy, bonds may either rise, or fall in price, depending on the inflation outlook. Sometimes, the inflation outlook is favourable, meaning bonds can fall in yield, while sometimes the opposite may be the case and bonds rise in yield and fall in price. Hence, the correlation to bond returns becomes insignificant, as the inflation outlook varies significantly through time, and the correlation between bonds and equities becomes somewhat less reliable compared to when equity prices are under pressure.
Figure 2 confirms the above expectations, and indicates that when the annual return on equities is negative, the average correlation between equities and bonds is significantly negative. In other words, when equities fall in price, we can expect the bond sector of the portfolio to move in the other direction; to rise in price.
Hence, when push comes to shove, bonds deliver “the goods”, or positive returns. This correlation information largely explains why bonds cut overall portfolio risk so much, without taking the chainsaw to returns. Further, it is possible to look at the average annual correlation of bonds to equities as the annual return of equities fall, by different amounts. Here, one would expect that as equity returns fall further, the negative correlation to bonds would strengthen further. Figure 3 illustrates the instances where bond correlations with equities were negative during the period mentioned above; daily data from 1989 to 2012.
In Figure 3 the average annual correlation, between bonds and equities, is determined for periods when equities return an annual loss. What’s interesting is that as the negative annual equity return values increase, the correlation between equities and bonds becomes more negative. Among other things, this means the greater the equity loss, the better the performance of the bonds. So, if equity returns are less than negative 20%, the correlation between bonds and equities is just under negative 35%. Similarly, if annual equity returns are less than negative 30%, the correlation between bonds and equities diverges further, where it is close to negative 40%.
In summary, as equity returns worsen, investors can expect an even stronger negative correlation between bonds and equities to eventuate. In other words, the worse the equity market gets, the stronger the bond market becomes, on average, as shown in Figure 3 above.
As the first risk mentioned above, credit risk, increases, a bond becomes more equity-like, depending, of course, on the position of the bond in the capital structure. As the bond becomes more like an equity, the diversifying feature of bonds is diluted, which means that products, like hybrids, do not provide as much diversification as bonds.
While it is difficult to foresee who will join the list of luminaries advocating more bonds and less equities in Australian investment portfolios, the case seems fairly clear. Dr. Henry, in his recommendation of more conservative portfolios, has effectively uncovered a debate about a major macro-prudential risk for Australia. As our pension industry approaches 100% of GDP, a collapse in equities would pose a substantial risk to the entire economy. In other words, size breeds conservatism, and while we have the size, we do not yet have adequate conservatism in our pension system. It is a concern with these macro-prudential risks that is forcing Tanner, and others, to comment. It is good to see that the groundswell of opinion is now firmly behind a greater allocation to bonds, for some of the reasons outlined above.
*“Intervene in ‘overweight’ super: Tanner”, by James Frost, The Australian, April 04, 2012.
**Returns are sourced using the following indices:
- (Cash) UBS Bank Bill from 1989 to March 2012
- (Bonds) UBS Composite Bond 0+ years, from 1989 to March 2012
- (Equities) ASX All Ordinaries, accumulation, from 1989 to March 2012