"It is the calm and silent water that drowns a man" - African Proverb

Very often in the financial markets historical patterns and beliefs that have persisted for a long period of time become well entrenched dogma that is taken for granted by many investors. Take for example the belief that interest rates or oil prices could never go negative or that equities have positive returns in the long run. However, these patterns are generally based on underlying assumptions which are regime dependent and are not static in time. This makes it very difficult for quant models that are calibrated to historical data to adapt on their own to new patterns when the regime changes. To illustrate this point we look at arguably one of the most important variables in modern portfolio construction: the equity versus bond correlation.

Over the past three decades the negative average realized correlation between equities and bonds has accentuated the diversification properties of bonds and has given them a substantial role in the construction of investors portfolios. Typically an asset with a positive carry in “good times” would likely lose money in “bad times”. Over the last 30 years (and in hindsight) bonds were the gift that rarely stopped giving to investors as they were both a source of positive carry and a good diversifier in "bad times" for equities. Bill Gross famously described this period: "All of us, even the old guys like Buffett, Soros, Fuss, yeah – me too, have cut our teeth during perhaps a most advantageous period of time, the most attractive epoch, that an investor could experience" [1]. However, the fact that we didn’t really experience bad times in bonds (high inflation) over this period doesn’t mean that the risk has disappeared and thus despite the historical benefits of bonds as a portfolio diversifier over this period it is fair to question their efficacy going forward now that interest rates are at historical lows.

To illustrate the implications of correlation assumptions on investor portfolios let us go over a simplified "optimal" portfolio construction example. Let us assume for simplicity that the sharpe ratio of equities and bonds is the same and that equities volatility is double the bonds volatility. In that case an "optimal" portfolio would have 2 times as much bonds as equities and the optimal ratio between bonds and equities is not actually sensitive to the correlation assumption. However, what is very sensitive to the correlation assumption is the overall amount of leverage that the portfolio can take on both equities and bonds to achieve the same level of risk/volatility. In other words at a macro level a negative or low correlation means that optimized portfolios and the system as a whole are more leveraged than it would have been possible otherwise

Thus, it’s no coincidence that investment strategies like risk parity emerged in the late 90’s when the current negative realized correlation regime between bonds and equities has started. Prior to this and for a period that extended over 3 decades a positive correlation regime prevailed (Chart 1). Within the current negative correlation regime it is not uncommon to witness instability in that relationship especially when it is observed over shorter time windows such as in 2018 (Chart 2) or the taper tantrum in 2013 where we did see investment strategies that are highly sensitive to the correlation suffer.

Using a simple discounted cash flow model we can see that the effect of nominal rate variations is pretty clear on bonds but is less so on equities because the cash flows (dividends and earnings) themselves might be sensitive to rate variations. To illustrate this, suppose that an increase in bond yields comes in response to an improvement in growth. In such a scenario investors would expect higher earnings growth which can potentially dominate the higher interest rates effect and thus actually boost equity prices. This would result in a negative correlation between equity and bonds as equity prices rise and bond prices fall. The situation is very different if the increase in bond yields is in response to an unexpected rise in inflation or inflation expectations and with no change to the growth prospects. In such a scenario there is no change to the future expected earnings/dividends and the higher interest rates would result in lower equity prices. This would lead to a positive correlation between equities and bonds as both equities and bond prices would fall.

In fact the path of realized inflation and inflation expectations as well as risk appetite and the evolution of the Federal Reserve monetary policy over the past two decades are important factors in understanding this shift in the historical relationship between equities and bonds [2][3]. In the period since the late 90s the Fed has become clearer and more outspoken about its monetary policy and thus making its policy decisions more predictable to investors. This along with well anchored inflation expectations (Chart 3 and 4) has caused the equity market in this period to be mainly driven by changes in growth prospects and risk preferences and less by runaway inflation or persistent hawkish fed surprises. Under such scenarios and like we explained above the correlation between bonds and equities tends to be negative.

Yet it is important to realize that statistical correlations don’t tell the full story. As far as diversification is concerned what’s more important is the performance of bonds during periods of significant equity market drawdowns which tend to happen during recessions [4].

So what drives bond performance? To understand the dynamics we can decompose the returns of a constant maturity bond investment where we keep rolling to the same maturity. The returns can be broken down into three components: carry, rate movement and roll down. Carry is the average interest rate one accrues by holding the bonds throughout time and so is a function of the average interest rate level during the holding period. Rate movement is related to the change in the mark to market value of the strategy due to changes in interest rates. Roll down is probably the least understood component and is related to the shape of the curve. If the curve is steep then a rolling constant maturity strategy will keep selling bonds with lower yields and shorter maturity and buying bonds with higher yields and longer maturity and thus overtime benefiting from this curve steepness. This is not just a theoretical observation and in fact explains a big portion of the returns of low yielding government bonds as Bunds and JGBs over the past few years. Take for example the period from Dec 2016 to Dec 2018 (not to cherry pick a period but just so that the point is easily made) a rolling 10 year Bunds strategy would have delivered an annualized excess return of 2.6%. The contribution of Carry is 42bps (the average yield during the period). The starting and ending yield is around 20bps so the contribution from rate movement is nil. The rolldown contributed 2.18%.

At a high level of starting interest rates there is of course more carry and there is more buffer for rates to move favourably during recessions as central banks typically slash rates. Going forward and given the current historically low starting level of interest rates it is likely that both the carry and rate movement engines will have very limited capacity to help in stress periods so it will be very hard for the roll down (if anything left there) to make up for these lost grounds. Separately, we could see a return of inflation over the medium to long term as a result of 1) the current global monetary and fiscal policy expansion 2) and the potential for the pendulum of globalization to reverse course in favor of bringing back manufacturing onshore at higher production costs.

Under such conditions is it reasonable to assume that government bonds will act as an effective equity diversifier as they have been over the last three decades? Can the inflation risk and its potential impact on the equity versus bond relationship be safely ignored as it has been the case in the recent past? Time will tell how we evolve from here but we can at least say that the diversification benefits have diminished and a careful assessment is warranted. This type of secular change if and when it occurs can potentially be a blind spot for a number of investment strategies. Thus, going forward it’s important to be conscious of the macro environment when constructing portfolios and there will be an ever more need for finding new uncorrelated sources of returns.


[1] "A Man in the Mirror", April 2013

[2] "Macroeconomics Determinants of the Correlation Between Stocks and Bonds", Nov. 2018

[3] "Positively Negative: Stock-Bond Correlation and Its Implications for Investors", Feb. 2019

[4] "Does the Stock‑Bond Correlation Really Matter?", Nov. 2018

This document serves, among other things, as marketing material. This document has been prepared by Quantum Investing Limited solely for the purpose of providing background information to the person to whom it has been delivered. The information contained herein is strictly confidential and is only for the use of the person to whom it is sent. Please refer to the important disclosures. Disclaimer

Prev Next Share: