Tuesday, August 22, 2006

Moving through Bernstein's The Four Pillars of Investing - chapter 2

I'm now a bit more than 1/3 of the way through Bernstein's wonderful book. I strongly recommend it to all investors.

One of my original goals was to post chapter by chapter summaries of the book as I read it, however, I'm drawn to maintain momentum by pushing through the book and mentioning some highlights as I move through.

PP 50 - 100 were fairly slow and tedious reading, but quite important in their discussion of whether investment value can be accurately predicted (using the Dividend Discount Method aka DDM).

The DDM helps to explain the value of a security - specifically, it calculates the present value of future dividends discounted to the present. The core driver of security value being the expected increase in earnings.

Here's the DDM formula: Market Value = Present Dividend / (Discount Rate - Dividend Growth Rate). Due to the significant impact of the Discount Rate on the formula and the entirely arbitrary means by which it can be chosen, few analysts ever achieve the same result and it's accuracy is unpredictable. This formula can be used for entire markets as much as individual equities - and may be more accurate for markets (with the caveat that it's still very arbitrary).

One question I have that was unclear to me, where does Mr. Bernstein obtain the Dividend Growth Rate - is this based on past history or projected future amounts - perhaps I have not read the text closely enough - does anyone know?

The important take away from the discussion of DDM is that the underlying assumptions of the formula are too arbitrary for it to be reliable or useful for long term investing. Even using the DDM or other techniques, few if any investment managers can accurately beat the market (relevent index) and those few who do, rarely do it consistently over long periods of time.

One interesting note was Bernstein's discussion of the Gordon Equation, which is derived from the DDM formula. It states that Market Return (Discount Rate) = Dividend Yield + Dividend Growth Rate. Again I understand that one can obtain the Dividend Yield for the Dow or S&P 500, but where does the Dividend Growth Rate come from? I think he derives it from long term averages.

Bernstein uses the Gordon equation to calculate expected future returns (long term) of different asset classes - I recommend reading the book to see his conclusions - the context of the book helps give meaning to the results. He also uses the formula to examine past periods and it's fairly accurate. I will note that his calculations see a convergence of large equity and bond returns (on a real / inflation adjusted basis). Other asset classes perform differently.

One important take away from the Gordon Equation is a discussion of cheapness and expensiveness of stocks. When stocks are expensive, their returns are lower - this initiates an ongoing discussion of value, value stocks and over/under priced markets. The equation does not incorporate changes in the dividend or PE multiple. Therefore, investor willingness to pay more for dividends (thus increasing the ratio between price and dividends) makes dividends more expensive and accounts for the 1% historical difference between the Gordon Equation and the actual results. Throught the 20th Century the dividend multiple tripled.

A final take away is that due to the convergence of large cap and bond returns investors should consider placing at least 20% of their portfolios into bonds.

I'll discuss chapter 3 in another post.

If you've read the book, I'd appreciate your posting comments on my take away and would be grateful if you could help explain the source of expected dividend growth used by Dr. Bernstein.

Have a wonderful day,

No comments: