Contrary to the prevailing opinion, equity investors have been right to mostly ignore the bond market this past year. To understand why, it is worth considering how bond yields and equity prices ought to behave.
An oft-heard argument is that the relationship should reflect the relative value of stocks and bonds, for example by comparing the earnings yield on stocks relative to the bond yield. The intuition is pretty straightforward. If bond yields rise, so the thinking goes, equity valuations must fall to offer a more attractive earnings yield. This approach was popularised in the early 1990s as the so-called ‘Fed model’ (though it had little to do with the Fed).
If that approach were true then, all else being equal, equity values should always fall when bond yields rise and vice versa. Yet that outcome is not always observed, which strongly suggests ‘all else is not equal’.
Equity valuations can fluctuate for other reasons, specifically variations in earnings growth and in the equity risk premium, neither of which are explicitly captured by the ‘Fed model’.
Equity valuation can be viewed as a series of future dividends, appropriately discounted, where the discount rate reflects the sum of the bond yield and the equity risk premium. How equity valuations respond to bond market moves depends, therefore, on how expected future dividends and the equity risk premium adjust to changes in bond yields.
Consider the following example. Bond yields rise, reflecting better nominal GDP growth, which pushes up the sum of future dividend streams in equal fashion, without any impact on the equity risk premium. In that case, equity valuations are essentially unchanged as the increases in the numerator and denominator cancel each other out. That outcome contrasts with the ‘Fed model’, which would predict a compression of the market multiple.
Indeed, under ‘normal’ circumstances one should assume that bond yields, which reflect current information about future nominal GDP growth, and expected future dividends, which are in the aggregate strongly correlated to future nominal GDP growth, should largely move in tandem. If so, in normal times, changes in bond yields will not have meaningful, or even predictable, impacts on share prices.
But is it always correct to assume that the equity risk premium will be invariant to changes in bond yields, that is invariant to swings in cyclical outcomes? This is questionable, particularly when cyclical conditions move to extremes.
For instance, when economic activity is collapsing and inflation turns negative, bond yields collapse and equity prices typically plummet. Arguably, that outcome reflects a sharp rise in the equity risk premium as the risk of default, or total loss to the equity holder, jumps as deflation grips.
On the other hand, when inflation accelerates beyond conventional measures of ‘price stability’, cyclical risk increases. In an ‘overheating’ economy, rising bond yields are accompanied by a rising equity risk premium and a de-rating of equities, as seen during the inflationary 1970s.
The preceding two examples, albeit extreme cases, demonstrate that equity and bond prices are not always positively correlated, as predicted by the ‘Fed model’. Instead, during episodes of deflation, the correlation can be negative, whereas when the economy overheats the correlation is typically positive.
What about in between — a world of full employment and steady inflation — as is arguably the case today in the US, the UK, and northern Europe? In that state, the correlation between stocks and bonds is indeterminate. At the margin, however, equities do well when bond yields rise owing to expectations of stronger, non-inflationary real GDP growth. And that is pretty much what occurred from the low point in growth expectations in mid-2016 until the end of the first quarter 2017, as equities were bolstered by better-than-expected growth and by hopes for pro-growth policies in the US, given Republican rule in Washington.
Since March, however, global growth expectations have slipped, as confidence in the ability of Republicans to unite has waned and as growth surprises have peaked. In the past four months equities have consolidated as bond yields have retraced, with bonds supported by unexpected declines in inflation.
What now? So long as growth and inflation expectations move within ‘acceptable ranges’, equities will probably grind higher, accompanied by rotation among styles and regions. But ‘acceptable ranges’ may not endure. In particular, the prevailing consensus expectation for sustained low inflation could prove wrong. If, instead, inflation were to accelerate above the ‘zone of price stability’, bond and equity prices will fall sharply.
Such an outcome would put to rest another misconception, namely that equities are always an effective inflation hedge.