Mark Twain said “There are three kinds of lies: lies, damned lies, and statistics.” His point was that through the judicious selection of data, statistics can be manipulated to prove just about any point a person wants to make. If this is true (and I wouldn’t want to argue with Mark Twain), investment performance falls firmly into the same category.
Timing the Market vs. Time In the Market
Arguments for and against the loosely-defined “market timing” is an area that takes the most liberties when it comes to quoting investment performance numbers. The most common case for a buy-and-hold strategy is that pulling money out of the market could cause an investor to miss a big market day. A quick search on Google reveals a long list of websites and articles that follow this general theme: If you had invested $10,000 at such and such a date and kept it invested, you could now retire early and send your kids and grandkids to Harvard. But if you missed the [10 best months/best month each year/5 best days each month/etc/etc], your investment would now be practically worthless. Obviously, Time In the Market is better than trying to Time the Market.
Standard & Poor’s Financial Library contains a chart that is often used in web articles to prove the case against market timing. The chart, titled “The Effect of Staying Invested vs. Missing Top Performance Days, 1997 to 2006″, shows the result of investing $10,000 and staying invested over 10 years, compared to missing the best 5, 10, 15, …, 30 days of the market. A buy-and-hold approach results in the investment growing to $22,451. Missing the best 30 days over that 10 year period results in the investment shrinking to $6,921. The conclusion stated on the chart, as well as most articles that use the chart, is that “missing the market’s top-performing days can prove costly.” Duh. Now let! ’s apply a little logic.
If your market timing system is so bad that it only misses the best days in the market, then a buy-and-hold approach is clearly the way to go. However, if a case can be made based on something as ridiculous as missing ONLY the 30 best days over 10 years, then it seems there would be an equal chance of missing only the 30 worst days. What happens in that case? Well, if you invested the same $10,000 and happened to miss only the 30 worst days (also clearly ridiculous), your investment would have grown to $69,879. Clearly, Timing the Market is much better than Time-In-the-Market. Both data points are hogwash, but only one of them tends to be used in what masquerades as a serious argument.
Historical Return Of the Stock Market is X%
Now let’s take a look at another way historical performance is often misused. The superior long-term performance (or lack of performance) of stocks is often used to justify portfolio allocations, indexing, actively-managed mutual funds, etc. After all, U.S. stocks have returned an average of 10.3% per year. Or they have been flat for over 4 decades when adjusted for inflation and excluding dividends. Or they have underperformed bonds. All of these statements are true in the right context and with enough disclosure. The long-term performance of stocks is a wonderful and dangerous tool because there are so many degrees of freedom with which to play. If I want a good long-term number, I can start my performance analysis in 1908 and end it in 2007. However, if I start in 1929 and end in 2008, I get a very different (and much worse) number. Going back to using stock performance to justify the buy-and-hold argument, the DJIA has had a couple of periods that work very well to support buy-and-hold. From 1943 to 1962, the DJIA had an average return of about 8.2%, and from 1982 to 2000 the return was around 12.9%. However, 1900-1943 (2.3% annual return), 1962-1982 (2.4% per year), and 1996 to 2009 (0%) didn’t work out so well, and could ! all be u sed to argue the exact opposite point. I recently saw a chart of the DJIA adjusted for inflation and excluding dividends. When looking at the market this way, the market is currently at about the same level it was in 1966. Even more interesting was that you could draw a straight line on the chart between 1929 and 1992. This data could be used to justify market timing, or a bond portfolio, or real estate investing…you name it. Many may argue that you can’t disregard dividends, and that the definition of inflation is up for interpretation, but however we massage the data, it can still be used to justify practically any argument…just like statistics.
How Should Performance Be Used?
The only data we have to go on is past performance, so obviously we shouldn’t throw out the data just because it can be easily manipulated. However, an awareness that data can be selectively chosen to justify our own biases is important, especially since this can be done subconsciously. In severe bear markets, it is easy to point back to the latest bull market and convince ourselves that things will quickly get back to “normal” if we just hang on a little longer. If normal is defined as consistent 10.3% returns, there are long periods in the market that don’t support this. Another thing to keep in mind is that it is just as easy to use the most recent market performance to justify the newest investing fad as it is to ignore recent market data in order to argue that traditional investing ideas will always work. U.S. and world economies evolve, and just because a strategy would have worked over the last 80 years does not mean it will work over the next 20. The key thing is to maintain a good dose of skepticism whenever performance data is used to justify an argument, and always ask “does this make sense”. Ignoring the best market performance days but including the worst days to “prove” a point should raise some red flags, and would certainly make Mark Twa! in think twice.
Jerry Verseput is Certified Financial Planner and Registered Investment Advisor in El Dorado Hills, CA. More information can be found at http://www.veripax.net
No comments:
Post a Comment