Do shares and bonds still correlate?
We are nearing the end of a 30-year bull run in the bond market where many investors anticipate that rates will head higher, and the role of bonds in a diversified portfolio is under challenge. Jeff Rogers, ipac CIO, breaks down the historical correlation between shares and bonds, and expectations for what's to come.
There is an old saying that “diversification is the only free lunch in finance”. While this adage is often wrongly attributed to Harry Markowitz, it does a great job of articulating the essential insight in his Nobel Prize winning research.
What people who’ve followed Markowitz know is that the attractiveness of an asset in a portfolio doesn’t just depend on its own expected return and risk. You need to consider the degree of correlation of the asset’s risk with the rest of the portfolio.
We are near the end of a never-to-be-repeated 30-year bull run in bond market. Today, the yield on a 10-year Australian Government bond is 2.75 per cent. In 1986 the rate was 13.8 per cent and it traded down to 1.9 per cent in 2016. What a wonderful time to be a fixed income portfolio manager!
Many investors anticipate that rates will head higher from here, generating capital losses. So, it’s not surprising that the role of bonds in diversified portfolios is under challenge. It is important to consider the implications of Markowitz’s insight in addressing the question of whether bonds will indeed – as they have in the past – help to account for equity risks in diversified portfolios.
The key statistic to watch closely is the correlation coefficient between bond prices and share prices. The coefficient is always between -1 and 1. It is a positive number if the two markets tend to rise and fall together, while it is negative if they generally move in opposite directions. The smaller the coefficient the greater the diversification benefit.
The implication is that if the correlation between bonds and shares is negative, then bonds may still be an attractive asset class in a portfolio context, even if their prospective return appears unexciting. Remember that bond prices move in the opposite direction to yields. So, if yields are rising, and there is a negative correlation between bonds and share prices, then there is a tendency for share prices to be rising.
The past as a guide
Understanding what propelled the exceptional historical performance of bond markets and analysing how share markets behaved at the same time provides valuable information about the likely correlation in the period ahead.
There were three distinct phases in the market over the past thirty years.
In the first phase, 10-year bond yields trended down to around 6 per cent following a major fall in the inflation rate to 2.5 per cent. This reduction was engineered by central banks who adopted explicit inflation targets and ran monetary policy to tamp down on economic activity whenever there was risk of a rebound in inflation.
If there was too much good news about economic growth, central banks would raise cash rates and both bond prices and share prices would fall in unison. This positive correlation between the asset classes had in fact prevailed for most of the twentieth century. That era ended in 1998 as the Asian and Russian debt crises were unfolding.
During the second phase, Australian bond yields traded in a narrow range around 6 per cent. Both inflation expectations and real interest rates (ie: cash rates adjusted for inflation) remained stable. Central banks declared victory in the war on inflation and focused on moderating economic volatility, particularly after the US tech bubble burst.
Bouts of bad news about economic activity would serve to increase earnings uncertainty and depress share prices. This would create expectations of a lower path for cash rates, so bond prices would rise. A new era of negative correlation between bond prices and share prices had begun.
Then came the third phase. In the aftermath of the global financial crisis, bond yields plummeted to historic lows as real interest rates collapsed. Underlying economic activity failed to recover to previous levels, inflation remained stubbornly below target, and there was rising fear of deflation. Central banks pushed cash rates close to zero – or below – and began to engage in major asset buying programs to underwrite investor confidence.
The negative correlation between bond prices and share prices intensified, as bonds became the preferred liquid “flight-to-safety” asset and the period became known as the “risk-on risk-off” era. This very strong diversification effect, together with low asset market volatility, contributed to very low volatility of diversified funds.
Looking forward, the influence of the financial crisis is finally waning. We are likely to experience a new pattern of behaviour for bond yields, market returns and correlations.
When historians look back at interest rates over past thirty years, I expect they will regard the start of this 30-year period as more of an anomaly than the end. But given the expectations priced into bond markets today we should prepare for some increase in yields.
Thankfully, not all increases in bond yields must coincide with poor portfolio outcomes, especially if correlations between equites and bonds remains negative. It all depends on what drives the shift on rates.
I think we will see inflation revert to normal levels in the period ahead and companies will start to invest for growth again. We should then see a modest rise in real interest rates, partially reversing recent declines.
While returns from bond markets may be flat and interest rate sensitive sectors of the share market struggle to advance, the cyclical parts of the share market should see solid earnings growth, validating higher share prices.
I expect the correlation between bonds and shares to rise but remain negative. Bonds will still represent a good source of diversification, though overall portfolio volatility might rise due to the withdrawal of central bank asset buying programs.
Two alternative scenarios worry me. One unwelcome development would be a change of mandate for central banks, away from inflation-targeting to a different policy regime.
Another scenario would see a marked acceleration in inflation at a time that economic activity was decelerating. In these circumstances correlations would turn positive, bonds and shares would fall in sync, and portfolio volatility would rise.
While these possibilities don’t represent high probability scenarios we need to be watchful. In the meantime, remember that Government bonds, which may seem an unappealing asset on a stand-alone basis, can still provide valuable sustenance during times of famine in share markets.