As of this Past Fourth Quarter, the S&P 500 Remained Relatively Cheap
Now that the first quarter of 2018 has just ended, what could be more fitting than to look back at the relative valuation of the S&P 500 stock index as of last year’s fourth quarter? After all, isn’t that what we economists do best, look back? This commentary is an update to my November 29, 2017 commentary, “The S&P 500 Is Not Expensive According to the Kasriel Valuation Model”. Before reviewing my methodology for estimating the over/under valuation of the S&P 500 stock market index, let me give you the Q4:2017 results upfront. According to my methodology, the S&P 500 index in Q4:2017 was overvalued by 7.4% compared with a revised 4.1% overvaluation in Q3:2017. Given that the median value of over/under valuation of my methodology in the period Q1:1964 through Q4:2017 was 35.5% overvaluation, the 7.4% overvaluation of the S&P 500 in Q4:2017 seems trivial. The continued low level of the corporate bond yield is the principal factor keeping the S&P 500 from being more materially overvalued. Now for the important disclaimer. My relative valuation methodology involves only three observable variables – smoothed annualized reported earnings of corporations included in the S&P 500 equity index, the level of the corporate BAA/A-BBB bond yield and the actual market capitalization of the S&P 500. My methodology does not take into consideration expectations of corporate earnings or of bond yields. Nor does it take into consideration exogenous “shocks” including but not restricted to the likelihood of changes in the U.S. tax code, of geopolitical conflicts or of the imposition of U.S. trade protectionist policies. Of course, if I had a high degree of certainty about these exceptions to my “model”, I would be an extremely wealthy person and would not be inclined to share my “wisdom” with you!
To refresh your memory about my methodology, I calculate a quarterly theoretical market capitalization for the S&P 500 by discounting (dividing) smoothed annualized reported earnings of S&P 500 corporations by the yield on the lowest-rate investment-grade corporate bonds (BAA/A-BBB). I then compare my calculated theoretical market capitalization value with the actual market capitalization value. The percent of over/under value for the S&P 500 is calculated as follows:
((Actual Market Cap/Theoretical Market Cap)-1)*100
The technique I use to smooth reported corporate earnings is some high-falutin’ econometric technique called the Hodrick-Prescott filter. (Edward Prescott is a Nobel Prize winner in economics.) This smoothing technique is designed to remove the cyclical variation from a trending series. Another economics Nobel Prize winner, Robert Shiller, uses a 10-year moving average to smooth the S&P 500 price-to-earnings ratio in his stock market valuation research. He calls this the cyclically-adjusted P/E. It seems to me that a 10-year moving average is an arbitrary tool to use to remove the cyclical component from a time series. Why not use a technique specifically designed to remove cyclicality? While I’m on the subject of the Shiller cyclically-adjusted P/E, aside from the arbitrariness of using a 10-year moving average, a P/E in isolation tells you nothing about the stock market’s over/under valuation without taking into consideration the level of bond yields. A low P/E could be an indication of an overvalued stock market if the bond yield is relatively high. Conversely, a high P/E could be an indication of an undervalued stock market if the bond yield were relatively low, as has been the case for several years now.