How Far Can Bond Yields Rise Before Hurting Equities? Goldman Answers

Ever since "Trumpflation" emerged as a driver of risk-assets, a tension has emerged in capital markets: how much higher can rates rise (and by implication the US Dollar) before financial conditions become so tight that the equity rally reverses under the weight of the very reflation it is pricing in. Ten days ago, SocGen became the first to attempt an answer, by providing the following table, according to which the answer was roughly 2.6% on the 10Y Treasury: a level at which equities become rich relative to bond yields.

Of course, on a simple yield comparison basis, if looking at the 10Y yield, which today is just over 2.31%, relative to overall dividend yield of the S&P500, stocks are already overvalued...

... however that is a rather simplistic take. A full analysis would take additional variables into consideration, chief among which is the so-called growth/yield trade off.

Conveniently, an answer to that question, and more broadly, to the question of how far can equities rise alongside higher bond yields, was provided earlier today courtesy of Goldman Sachs.

Not surprisingly, Goldman's strategist Peter Oppenheimer writes that "this is a question we are asked a lot" and adds that "the inflection point this summer, where deflationary risks have faded and reflationary hopes have been rekindled, has caused the relationship between bonds and equities to change. In effect, the collapse of the 'deflation' risk premium has pushed up bond yields and has supported some equities, especially financials. But in reality this shift in the risk premium (and slightly improved earnings outlook) has not been sufficient to offset the speed and change in bond yields. The net effect has been a flat market but with a small de-rating in the valuation."

This is shown in the exhibit below which shows that the market P/E has moved down by a full PE point since bond yields have increased from July.

12-month forward IBES consensus P/E for SXXP


Goldman then provides a broader answer to the bigger question of the "The Growth/Yield trade off" as follows:

The answer to the question about the level of bond yields that could dampen equities is not a concrete one; answering it is more an art than a science. There are several factors to consider:

1) The relative movements between bond yields and expected growth

One way to think about the range of possible outcomes is to attempt to put some possible numbers behind the 'trade-off' between the discount rate and long-term nominal growth expectations. We do this by using our standard DDM in the exhibit below. Here we fix the equity risk premium (ERP) and vary growth and bond yields. As a rough rule of thumb a 50bp rise in bond yields would knock around 10% off equity values without any change in the long run expected growth rate.

SXXP sensitivity to risk-free rate and growth assumptions, with a fixed ERP

2) The level of bond yields

While the correlation between changes in bond prices and equity prices is a positive one under most 'normal' environments, as bond yields fall to very low levels (around 3% or so) the correlation flips. That is why falling bond yields post the financial crisis have often been combined with weaker equities as they both reflect deflation risks. By association, the reverse should be true, and the rise in bond yields from trough levels has been accompanied by a rally in risky assets since the July lows this year, although in Europe equities have remained flat as they have de-rated alongside the rise in yields. However, a rise to anything over c.3% on US treasuries (see Exhibit 12) could be a significant negative, particularly given high current valuations. In fact, the speed of adjustment is also important; note from Exhibit 12 that the sharp rise in bond yields during the so-called taper tantrum in 2013 was negative for equities as well as bonds. As Exhibit 11 shows, the same relationships exist in Europe.

Another way to show this is in Exhibits 13 and 14, described in GOAL, Reflation, equity/bond correlation and diversification desperation, November 14, 2016. This shows that the correlation between German 10-year bond returns and equities is negative when there is a low level of GDP growth and low levels of inflation. The 'tipping point' in the correlations is typically around real GDP of c.1.5%-2.0% and CPI inflation of around 2.0%-2.5%.

3) The valuations of bonds

In addition to levels, and the rate of change of bond yields and inflation expectations, the valuations of assets are also important in determining the impact of rising bond yields on equity prices. This is a crucial point. In recent years, given fears of deflation and the impact of QE, the valuation of bond markets has increased to levels well beyond those we believe are justified by fundamentals. It is as if bonds have benefited from the deflation risk premium (while of course equities have correspondingly been held back by this risk premium). We can get a sense of excess valuation by using our bond strategy team's so-called 'Sudoku' fair value model. Exhibit 15 shows that when this model shows overvaluation, the correlation with equity prices is negative: rising bond yields are good for equities. The reverse is true when the bond market is at fair value or is expensive. This is further illustrated in Exhibit 16 which shows a scatter plot of the valuation mismatch of bonds (in standard deviations from fair value) and the correlation of equities and bonds. The overvaluation of bond yields has meant that rising yields (and a diminishing deflation risk premium) has been good for equities.

But we have already moved away from the point at which rising bond yields were definitively good. Back in July, when concerns about the ramifications of Brexit on global growth and inflation had pushed yields down to extreme low levels, and consequently had pushed down equity markets, any rise in inflation expectations or growth helped to push up yields and see equities rise from the lows together. In July, the correlation between European equities and US bonds was about -80%, but even now this correlation has already moved to -40%.

* * *

All of which brings us to the one thing that really matters: the number, or rather Goldman's estimation, above which the 10Y yield would lead to selling in stocks, all else being equal. That number, according to Goldman is...

Taking all of these approaches together we believe that the equity market is still at a level that can cope with moderately rising bond yields. We estimate that a rise in US bond yields above 2.75% or probably between 0.75-1% in Germany would create a more serious problem for equity markets: at that point we would expect the correlation between bonds and equities to be more positive - i.e., any further rises in yields from there would be a negative for stock returns.

Taking a midpoint of Socgen's and Goldman's estimates, it means that the 10Y can rise another ~35 bps before it starts acting as a break to further gains in the S&P500 on expectations of the Trump "reflation" trade.