Our big issue to watch in 2016? Much the same as it was for 2015, and for a few years before that: what happens to the equity risk premium?
The equity risk premium is the apparent return you are offered for holding equities versus a ‘risk free’ asset*. With interest rates and bond yields very low today, this premium appears to be very high in much of the Western world.
There is a common bias in finance to assume mean reversion. When the short term doesn’t look as you’d expect, the natural human impulse is to assume that it is temporary; ask any Chelsea fan.
So when most of us look at the chart above our inclination is to believe that the gap must close, otherwise equity markets are offering a ‘free lunch.’
However, much of the discussion today in and around this issue reveals some stark inconsistencies and biases in our thought.
Bond yields to rise?
The knee-jerk reaction to figure one may be to assume that cash rates have to rise. After all, it is easy to view ultra-easy monetary policy as the cause of the wider risk premium and believe that this is the anomaly that needs to correct.
Many discussions I have had, and this is revealed in the yields of longer dated bonds, suggest an implicit belief among many that somewhere, just around the corner, is a ‘return to normal.’ Normal for many is seen as a return to bond yields of the mid 2000s or mid 1990s (or rather how they remember that period to have been).
And yet irrespective of the rise in US rates last week, many have a hard time articulating how such significant rate rises will come about. Many of us today find it hard to conceive of a world with much stronger growth or inflation.
On top of this, and as I mentioned in a blog earlier this year, many investors are displaying distinct tiredness with the bond debate. There seems to be a subconscious belief in a return to normal but decreasing willingness to articulate how or why this will happen.
The experience of betting on mean reversion in bond yields has been extremely painful for many. One of the best forecasters of US bond yields over recent history has been the continuation of the trend, why should this change?
Moreover, what concept of normal is relevant? Perhaps we have already reverted to the normal that prevailed prior to the 1960s
Equity valuations: more mean reverting than bonds?
In equity markets we are almost seeing the reverse. Investor beliefs in a return to ‘normal’ have arguably strengthened. We all find it very easy to picture the equity risk premium falling because the earnings assumptions built in to valuations today are shown to be overoptimistic (i.e. our experience of the financial crisis).
There are very real, and well-rehearsed reasons why this is a risk, but we would caution against believing it ‘must’ happen simply because valuations are above some long term mean.
It has long struck us how investors can fall into discussing equity markets being ‘cheap’ or ‘expensive’ using rules of thumb, often derived from historic averages. Indeed this is far more prevalent than in bond discussions. The Shiller p/e is the most common example of this, albeit one that makes the bear case for equities based on a reversion in rating rather than earnings.
The p/e certainly is higher than its mean. But how relevant is this? The chart above looks to us more like a series of regime than a stationary series:
- Fifteen years or so of de-rating from the mid 1960s to the mid 1970s
- Relatively unchanged valuations for ten years, followed by a re-rating between 1985 and 1995
- A period of more volatile valuations, including the tech bubble and crash and the earnings collapse (and subsequent investment opportunity) of the financial crisis
There are periods for over a decade where hoping for mean reversion would have been very painful, even if you had known the mean over the whole period ahead of time.
What do we mean by ‘expensive?’
As asset allocators we need to be clear about what we mean (no pun intended) when we say something is expensive. Do we think the asset will sell-off, or do we think that returns will be positive, but lower than we have been used to?
Ideas of mean reversion in valuation imply the former, in which case it is best to hold cash, even if it offers a negative return in its own right. However, if we believe the latter, as investors seem increasingly willing to do in the case of bonds, then the conclusion is a very difficult one. In that case longer dated bonds, property, and equities still offer the most attractive return of mainstream competing assets.
A behavioural interpretation of the risk premium
Our interpretation for the heightened risk premia today is that investors want more compensation for holding equities due to pessimism regarding earnings and an unwillingness to tolerate short term volatility. These are elements of investor beliefs that have been impacted by the experience of the last ten years; the mental scarring of successive equity crashes, weak global growth, and the superior capital returns on most bonds.
We believe that these behavioural forces as well as fundamentals can define a regime, and a question we should ask ourselves is how these may change.
Should it prove to be the case that we are in a regime of permanently lower interest rate structures, the better the prospects for equities and other assets to beat bonds. The longer this goes on, the more you may see the argument for equity mean reversion jettisoned, just as investors did in the late 1990s and have been increasingly willing to do for bonds today. This itself would represent a new regime, one in which investors will again have to be pragmatically flexible in how they interpret the relevance of the past.
*Note: this risk free asset is typically considered to be cash or local mainstream government bonds, even if you are guaranteed to make a real loss if you hold them to maturity. They are risk free only in the sense that there is no uncertainty about the return you will receive.
The value of investments will fluctuate, which will cause prices to fall as well as rise and you may not get back the original amount you invested. Past performance is not a guide to future performance.