"As investors try to decide what to do now, I think it is useful to contrast two options. Let?s assume investors are making their decisions for a five-year time horizon. Were the time horizon any shorter, we would say stocks shouldn?t even be considered because short-term results can be too random. One option for an investor is to say ?no? to any risk, and invest in a five-year government bond. Many investors, stinging from recent losses, are making that very choice. The annual yield for that bond today is 1.8%. So, at the end of five years, that investor could be certain to have a 9% return (not considering either taxes or inflation). The other choice is to buy equities, anticipating a higher return in exchange for accepting uncertainty. Can history give us any guide as to what that return might be? One approach would be to assume that returns would simply match the historical average of 9% per year, or 54% compounded over a five-year period. That answer, however, ignores the effect of the starting price. I think the following is more useful.
First, the dividend yield of the S&P 500 is now about 3%, so over five years the equity investor should receive a 15% return plus or minus price change. We can estimate the S&P price five years from now by estimating both its earnings level and its P/E ratio. Over the past 80 years the median P/E ratio for the S&P 500 has been 15 times. I could argue that today?s very low rates on government bonds suggest future P/Es should be higher, but let?s not bother with that complexity. Earnings are trickier to forecast. Operating earnings for the S&P 500 peaked at $88 in 2006 but the consensus forecasts a trough at about $62 this year. Extrapolating either peak or trough earnings is not likely to be productive. Instead, let?s look back over the past thirty years. A regression analysis of the past thirty years shows that trend earnings for 2009 are about $84, or 5% below the peak achieved three years earlier. Further, that same regression analysis calculates that earnings growth has averaged between 6 and 7% per year. Extrapolating based on those numbers puts trend line earnings for 2014 at about $115. Multiplying $115 in earnings by a P/E of 15 produces a 2014 expected price for the S&P 500 of 1725, 116% higher than today?s price of 798. Were that to happen, the annualized return for the next five years would be about 20%, a little more than twice the historical average, and more than ten times the bond return.
What about the downside? By 2014, the S&P could fall to about 750 and still match the return on a five-year bond because the current dividend yield exceeds the bond?s interest rate. If the historical average P/E is attained, that means the S&P earnings would have to be more than 50% below trend for the stock investor to underperform the bond investor. Alternatively, if the earnings trend line proves accurate, the P/E would have to be below 7 times for the stock investor to underperform. While those outcomes are certainly possible, they would be extreme historical outliers. The possibility of historically high returns, combined with what we believe to be a low probability of loss, makes me excited about investing in stocks today. And that?s why I significantly added again to my personal Oakmark investment last quarter."