One of the ways in which long-term market direction is often forecasted is by fundamental value. One such method is that of yield analysis. As a stock price rises, a stock paying a constant dividend will incur a lowering dividend yield as the denominator continues to increase with a mostly steady numerator (Dividend Yield = $Dividend/$Price). By compiling all the dividends in an index such as the S&P 500, we can calculate a composite yield for the "market." Below we use very crude data to prove a point. While the numbers can be obtained in much finer granularity and precision, it is unneccessary to make the point. The common view is that low yields show an undervalued market, one prime for bottom fishing. However, under closer surveillance, we can see that while lower ranges of the S&P 500 dividend yield do hold higher returns, they are not significantly different from the average return overall, and therefore of little value.
Above I have split up the historical yields over the last century or so to look at the validity of this view. The decision is a simple T-Test applied against H0 being the average return (since the market has risen in general over the 100 years, testing against a 0% return would prove most tests to be statistically significant). As can be seen, except for using a 3 year timeframe, yield analysis in this matter is not very useful. Basically we see that buying a cheap market according to yield will give a positive return while waiting out an entire bear market (3 years forward). However, by applying a few proprietary transforms, we can turn this into a very useful and powerful indicator for both short and long-term use.
As can be seen, we now have an indicator that can help us with forward returns with some degree of accuracy (the pentiles used here are arbitrary numbers according to our indicator and so are not important). With a more useful way to use equity yield, let's look at some of the results both in numbers and graphically.
While our new measure of finding market valuation does a great job at finding where the market is undervalued and bottoming, it fails to pick up the severly overvalued spots as well as the major financial crises. The issue is that tops and bottoms form differently. Since bonds often move ahead of equities, and rates begin to be lowered before stocks fall, this creates some specific spread activity between the 10 year and the equity dividend yield. By using this data series instead of the raw dividend yield, and performing the same transform as above, we arrive at a much better tool for finding markets that may have high downside risk attached to them over the longer term (3 year).
With this tool we can do what we believe is the key to long-term success and positive returns, we can avoid and/or lower exposure when we do no like the risks in a market. The more tools we have to avoid these risk areas on short and long-term time frames, the more our returns will naturally accumulate.