Rethinking the TINA argument

Relative to bonds, stocks still overvalued

Acommon refrain since the March 2009 market low has been “there is no alternative” (TINA) — a catchphrase used as justification for buying stocks amidst ever higher price-to-earnings multiples.

In essence, the argument is that with government bond yields at very low levels, prospective forward returns in the bond market are weaker than they have been in the past. As a result, in order to achieve the same return profile, investors will need to take on more equity exposure (at the expense of government bonds). While we do not disagree with the crux of this argument, we believe it has increasingly been used to rationalize overpaying for stocks.

We pulled annual data on the S&P 500 trailing P/E and the 10-year T-note yield since 1962 and found there is a clear negative relationship between Treasury yields and the P/E ratio. This means, all else being equal, a lower Treasury yield is associated with a higher P/E ratio whereas a higher Treasury yield implies a lower P/E ratio. This result is entirely expected and is supported by any traditional valuation model: lowering the discount rate, and holding all other variables constant, will push the “fair value” P/E higher.

As a result, with the 10-year Treasury yield up about 120 basis points year to date, it’s no surprise that P/E ratios have been adjusted lower. Indeed, in the case of the S&P 500, the trailing P/E has dropped to 22.5x from 24.5x and the forward P/E has declined to 18.8x from 21.4x. However, while this adjustment is directionally correct, we believe there is further to go.

One way to visualize the valuation of stocks relative to bonds is to look at the equity risk premium (ERP) — the difference between the estimated return on stocks versus bonds. Perhaps the simplest way to do this is to invert the P/E (thereby calculating the earnings yield) and subtract the 10-year Treasury yield. Currently, the ERP on this basis is just 2.6 per cent, considerably below its average since 2009 (4.2 per cent) and the lowest it has been since December 2007.

This tells us that despite the decline in the P/E ratio this year, it has been more than offset by the move higher in Treasury yields. Therefore, relative to bonds, stocks are the most overvalued they have been since December 2007, which was not exactly a good time to buy with the benefit of hindsight In order to return the ERP to its average since 2009 (absent a decline in Treasury yields), the forward P/E ratio would have to fall to 14.5x (consistent with an S&P 500 level of 3,400). If we instead use the average ERP over the past 20 years (3.6 per cent), the forward P/E would need to decline to 15.8x (implying an S&P 500 level of 3,700).

One other measure of the ERP we like to monitor is its rolling three-year z-score, which allows us to track large swings over a shorter horizon. Currently, this metric is two standard deviations below average, a level that rarely gets surpassed. Historically, the forward return implications of this development have not been positive: over the next year, the S&P 500 has risen just 3.6 per cent; for reference, this compares to a 12-month return in all other periods of 10 per cent.

The upward pressure on government bond yields has drastically reduced the validity of the TINA argument, which we believe was overstated to begin win. After all, the ERP — which allows us to compare the valuations of stocks versus bonds — is the least attractive it has been since 2007.

To get back towards an “average” ERP level — absent a fall in interest rates — will require a more substantial downward adjustment to P/E ratios. This has negative implications for forward returns, informing our cautious outlook on the U.S. equity market as a whole.