Also See: Forecasting the Stock Market (Part 1)

As the good news on the economy continues to roll in, investors are still scratching their heads over stock market performance. The market continues to go sideways despite the positive economic outlook. However, that’s just the short-term picture; in the long term, things may look a bit different.

Last week, we examined the widely varying opinions of others for that longer-term perspective. As I promised, this week, we will explore several ways to perform a do-it-yourself forecast.

The easiest way to develop a forecast is to compare the value of the stock market (using returns on large cap stocks computed by Ibbotson and Associates as representative of the market) measured in ten-year rolling periods from 1926–2003. That average turns out to be 11.1%.

Using that average with the year 2000 as the beginning of this decade, we could project a Dow of approximately 32,000 by the end of the ten-year period. For that to happen, the Dow would have to average 21.1% for the next 5 ½ years.

But there is a problem with this number, as anyone who follows markets knows—markets don’t move in a straight line. Some years are up and other years are down. This variation can be measured with what is known as a standard deviation, and in our example, it computes at 5.73%. In layman’s terms, this means that 68% of the time, the actual return will be within a range of plus or minus 5.73% from our average.

When applied to the decade 2000–2009, we can forecast a range of returns from a low of 19,396 using an average of 5.37% (11.1% – 5.73%) to a high of 54,489 using an average of 16.83% (11.1% + 5.73%). To hit the lower number, the Dow would have to average 10.9% from 2004 to 2009; the higher number requires an average of 31.7% for those years.

A second method of valuation involves comparing the earnings of the stock market to a risk-free rate of return. The risk-free rate commonly used in fundamental analysis is the ten-year Treasury bond, currently yielding 4.76%. An investor buying the ten-year bond is buying a financial instrument with a Price/Earnings ratio of 21. Typically, the market trades at a P/E similar to that of the ten-year note.

If that were the case today, the S&P 500 index would be undervalued by about 20%, trading at 1,110 when by the valuation model it should be 1,365 (based on forward earnings of $65 as reported on the S&P website).

A stock buyer is buying the expectations of future earnings from a stock. He/she might be willing to pay more for the future earnings if interest rates are lower (implying a higher P/E for the ten-year bond, lower inflation and more money supply) than he/she might if the reverse were true.

If you go to the S&P website, http://www.standardandpoors.com/, you’ll find the forecasted earnings projections for the S&P through 2009.

For instance, the forecasted earnings for 2005 = $73, scaling up to a forecast of $92.82 by 2009. Using the model, make a forecast for ten-year bond interest. Divide that number into 100 to determine the P/E on the bond. Multiply the S&P forecasted earnings by that P/E. The resulting figure is your forecast for future value of the S&P. For example, if the ten-year bond is yielding 5%, its P/E is 20, and in 2009, its forecasted earnings = $92.82. The value of the S&P forecast is $92.82 x 20 = 1,856. This implies a 9.98% annual increase in the S&P for the next 5 ½ years.

Try varying this forecast exercise with lower interest rate assumptions and compare the results. Voila—a do-it-yourself market forecast.