Wednesday, March 10, 2010

The 10-Year Stock Market Projection

This post presents my "official" projected stock market returns for the 10 years beginning in January 2012. It also compares previous 10-year projections with the actual results.

Note: I do this "formal" projection once a year when earnings are published.  For the most recent informal (& text-only) monthly update, see March, 2013 Stock Market Performance 


Projected 10-Year Stock Market Returns

Stock market (Dow Jones Index) forecast for next 10 Years performance /returns
Projected Dow 10-Year Returns

The above graph (click to expand) shows my projected returns for the DJIA (Dow Jones Industrial Average) for the twelve 10-year periods beginning end-of-year 2000 through end-of-year 2011. The blue dashed line shows the returns that I estimated that a hypothetical investor in the stock market would receive over the next ten years 
if he invested at the end of the starting year and sold 10 years later -- reinvesting dividends during the intervening ten years. The estimates are as they would have been calculated at the beginning of each 10-year period. 

An earlier post "backtested" the original methodology and showed projected vs. actual results from 1900 through 1999 (the last year for which actual results were available at the time). That was my "proof of concept."  Since 2010, in addition to making new projections, we've plotted actual results for the recently completed 10-year periods vs. the original projections. The solid blue line shows those actual results.  For perspective, the dotted lines show average, best-ever & worst-ever 10-year returns.

2001 10-Year Stock Market "Forecast" vs. Actual 2001-2011 Results

As was the case last year, the actual returns were better than the projection.  At the end of 2001, the model would have projected 10-year returns of 1.5%/year; the actual return was 4.5%/year.  So, the model underestimated future returns by 3%/year. Interestingly, the 2000 projection was also low by 3%/year (0.1% projected vs. 3.1% actual).  Furthermore, this continues a trend that was present in varying degrees in the final years of the backtest.

How accurate a forecast you think that is depends on your point of view. However, remember I'm not forecasting the future, I'm projecting the future.The reality is that in both 2000 and 2001 the model (would have) projected very modest returns -- in the lower range of historical returns -- and that's what happened.  As you can see from the chart, the actual returns were well below the average line; in fact, they were in the lowest 15% of historical returns.

Why was the 2001 projection too low? Well, not because it underestimated earnings.  In 2001, the model would have projected 2011 normalized earnings to be $698; I currently estimate 2011 normalized earnings at $697! It turns out that he bulk of the discrepancy is explained by valuation.  The model assumed that we would end 2011 (and every other year for that matter) with a normalized P/E ratio near the long-term average; instead the normalized P/E was around 17.5.  Had we ended 2011 at the average normalized P/E, actual returns would have been very close to the "predicted" 1.5% returns.

Projecting Returns for the Next 10 Years (Beginning Year-End 2011)

As you can see from the chart, since 2000 the model has consistently projected well below average, though still positive, returns. In fact, the 2000 projection was the second lowest ever; the 1999 forecast was for a loss of 1.1%/year. From 2000, prospective returns increased gradually, peaking after the 2008 crash.

If you look at my price/earnings(P/E) ratio history, you will see that historically, secular bear markets have bottomed with normalized P/E ratios below 10.  Those cheap valuations lay the foundation for the subsequent secular bull market, and are generally associated with superior projected returns -- 10-year returns as high as 20% or so.  As gut-wrenching as the 2008 crash may have seemed, valuations never reached the levels typically associated with secular bear market bottoms.  As a result, not only did prospective returns not climb into the superior range, they barely got back to the average range.

At any rate, the 2008 crash lowered prices enough to raise prospective returns somewhat, but not into the superior range. Since then, prospective returns have generally drifted downward such that by year-end 2011 -- the beginning of 2012 -- the projected returns stood at 5.6%.

The market started this year off on a tear -- increasing approximately 1,000 points through the end of April. Those above average market increases have in effect borrowed returns from the future and further reduced prospective returns.  As of the end of April, my informal monthly model projects 10-year returns of around 5%.

Conclusion

As I have said before, I prefer to describe this as a stock market projection rather than a stock market forecast. The reality is, no one knows for sure where the stock market will be 10 years from now -- it's the future! I think of the above graph as establishing a benchmark. It says IF normalized earnings grow at the historical rate*, and IF 10 years from now the market is applying average* valuations in terms of earnings multiples, THEN these are the results an average buy-and-hold investor can expect. The chances of this scenario playing out exactly are, I think, remote. However, a look at previous results suggests the chances of actual returns being in the projected "ballpark" are quite high. In short, with all its imperfections, I think it's a good place to start our thinking about future returns.

* Results based upon my Stock Market Analysis Model/Spreadsheet. Average earnings growth rates and p/e ratios are based on "full-cycle" results from 1941-1982.

Related Materials

Rolling 10-Year Returns shows all previous 10-year returns from 1900 to 1999 for reference.
Projecting Stock Market Returns: The initial post in this series. Compares previous projected to actual returns.
How the Stock Market Projection Model Works discusses the methodology in more detail, and contains links to all of the posts that lay the foundation for this methodology.
Where Are Interest Rates Headed? My unique methodology for forecasting interest rates.
Interest Rate Forecast for 5-Year Treasury Notes: Uses above methodology for forecast 5-year rates.
Starting P/E Ratio vs. 10-Year Returns: Shows that the initial P/E has a significant impact on future returns.
Borrowing Returns from the Future: some evidence that bull markets tend to be followed by poor returns.
The 2011 10-Year Stock Market Projection: The projection from early in 2011.
2010 10-Year Stock Market Projection: The projection from early in 2010.
Bogle on Mutual Funds, by Vanguard founder John Bogle, discusses a similar methodology based on price/dividend ratios (see chapter 12 The Allocation of Investment Assets).
Profit Margins: The Epicenter of the Valuation Debate: A cogent discussion of an alternate valuation methodology that yields substantially different results.
For a list of other popular posts see the sidebar to the left or the blog header.
Copyright © 2010.                         Last modified: 3/29/2014

Share This Article

Delicious
To share via Facebook, Twitter, etc., see below.

2 comments:

  1. Excellent analysis. It would even be better if you were to do this for the S&P500. Much more representative of the overall US market than the Dow. I've attempted to do this for the S&P500 and get about a 3.1% annualized return from end of 2011 to the end of 2011

    ReplyDelete
    Replies
    1. I agree the S&P would be "better" in some sense. If I were starting over, I'd almost certainly use the S&P. However, projections like these are really intended to be more qualitative than quantitative, so it's not clear what you've really gained. Bottom line either way is below average expected returns.

      Delete

No spam, please! Comment spam will not be published. See comment guidelines here.
Sorry, but I can no longer accept anonymous comments. They're 99% spam.