I recently read a newsletter article titled “Dangers of Market Timing”. My first thought when I saw the title was hoping that the article would contain something new. Not only was this not the case, but the sloppy data was so bad I felt like I had to say something today before I could get some real work done.
The article states that market timing amounts to gambling because it is impossible to forecast market movements, and “you run the risk of missing periods of exceptional returns.” As an example (here comes the nonsense), a $1 investment in the S&P 500 over the last 20 years would have grown to $4.84, but if you had missed the best 10 months of stock returns, the ending value would have been only $2.04.
It’s pretty easy to grab these numbers from Yahoo and throw them into a spreadsheet. I used the Vanguard S&P500 Index fund assuming reinvested dividends. The author used the April, 1990 to April, 2010 timeframe, so I used the 20 years ending June 1 to capture the latest downturn, which ”proves” the case against market timing just as well. In the 20 years ending June 1, 2010, a $1 investment would have grown to $4.43, and missing the 10 best months would have resulted in only $1.91.
Now for some results that were not mentioned. What if you missed the 10 worst months, which is just as ridiculous as somehow missing the best months, but equally likely? In this case, your $1 turned into $13.48. Doesn’t that seem worth mentioning? How about something less ridiculous, like missing the 10 best and 10 worst months? Now, the $1 is $5.43, a 22% improvement!
Now let’s get a little smarter about the whole thing. If a big positive month happens in the middle of a string of positive months (bull market), why would you get out of the market? So let’s assume we miss positive months only when they happen after at least an 8% drop, when a portfolio would likely have been moved to cash. Similarly, if a negative month comes after a positive trend it is likely you would be fully invested, so we’ll assume you experience the full month’s loss. As a specific example, the monthly returns for June-Sept, 1998 were 4.08%, -1.06%, -14.47%, and 6.41%. We’ll count the big negative return in August because the previous month would not have triggered any selling, and we’ll miss the positive return in September because August would have caused us to move to cash. This is still incredibly unsophisticated, but it’s a heck of a lot better than the brain dead 10 best or 10 worst months. The result with just a little intelligence added results in $6.36, a 43% improvement over buy-and-hold.
The conclusion is that anyone who uses a “missing the best [days, weeks, months]” argument to prove a buy-and-hold point is either being incredibly sloppy with the data, or being disingenuous. I’m open to other arguments on the topic, but not that one.
