Market timing is normally not my thing but after reading this book I accept that, Yes, you can time the market with Stein and DeMuth's method. The only thing is the method requires that you have the patience of Job and the lifespan of Methuselah. What investor would be willing to not buy / not invest in the market for 17 years waiting for the buy signal to begin flashing? The book's buy signal test was a red "no-go" for the whole time between 1984 and 2001, a period during which the S&P 500 (used as the US market proxy throughout the book) experienced huge unprecedented gains. Similarly, the twelve years between 1954 and 1966 was another long period for an investor to sit on cash or money market accounts according to Stein and DeMuth . The authors' goal to show conservative investors how to make money in the long run is truly vulnerable to Keynes famous quip that in the long run, we are all dead.
The Yes market timing method works as follows:
- objective - determine when the US market is over-valued, in which case don't buy, or under-valued, in which case buy
- assess over-/under-valuation with four real after-inflation metrics: price, price/earnings, dividend yield, earnings yield vs bond yields, price to sales, price to cash flow and Tobin's Q (a measure of fundamental value of companies; each individual metric works but several simultaneously saying Yes works even better
- the current value of the metric is compared against the trailing 15 year average
- valuation is applied against the market index only - the S&P 500; it is explicitly not proposed to be used for individual stocks
- wait ten years or more (the longer you wait, the better, though their testing only goes up to twenty years) to achieve far superior returns than you will get whether investing at random as an average of years or on a continual dollar cost averaging schedule.
- when signals are saying are saying No, it only means don't buy, it doesn't mean sell; you keep whatever stock investment you have and simply wait till the next buy signal.
For the most part, the analysis and comparisons make a convincing case for the authors' thesis. Using 100 years of market data, rolling periods and looking at results after 5, 10, 15 or 20 years of holding after purchase builds confidence that the data was properly compiled. There is also economic logic supplied as to why the indicators should work.
Where the argument isn't convincing is the comparison of dollar cost averaging vs their market timing using 1977 as a starting point. In 1977 the buy signals were flashing so the market timer got their money into the game a lot sooner than the DCA investor. With a subsequently rising market, the market timer was bound to win.
It is very useful for the book to contain all the year by year tables of past signals, both buy and don't buy, along with subsequent results of the 5-20 year holding periods. That reveals a key fact - the market timing system did not produce great returns every time it said buy (e.g. buying in 1973 produced only a 251% gain 20 years later) , just as buying in many years when the system said it wasn't propititious to buy produced outstanding returns in subsequent years (e.g. buying in 1982 gave off a gain of 582% after only 15 years). The system appears to produce better returns on average.
Chapter 8 titled Using Market Timing contains a lot of very sensible cautious advice for investors, the antithesis of a get rich quick mentality that one might suppose a book on market timing might present - e.g. "Never make a "bold" investment decision; Don't think big; Don't make any sudden moves".
The book was published in 2003 so the data series stop too soon for us to find out what any reader wants to know - what is the situation today, is the market over or under-priced and is it time to buy? Fear not, the authors have continued to update the metrics on the book website http://www.yesyoucantimethemarket.com/index.html. As of Oct.30th, 2009, it shows for the S&P 500 - Price - Green! BUY!; P/E Ratio Red! Don't Buy!; Dividend Yield - Green! Buy! (They say also that data for several other indicators is no longer available.) That should mean it's a good time to buy.
Stein and DeMuth posit that their method should work on other markets, though they haven't tested. It might/should given its essence is to identify times when a market is clearly under-priced in historical terms. Of course, the method also rests on the assumption that the future will be like the past, that reversion to past means will occur.
My rating: 3.5 out of 5 stars.