One of the greatest inefficiencies in the supposedly efficient stock market is the incredible disparity in returns produced over the past half-century during the six-month favorable (November-April) and unfavorable (May-October) periods. When we first reported on annual seasonality back in August 2002, we quoted the Stock Trader's Almanac in showing that a $10,000 investment would have gained a whopping $425,890 if invested during only the six favorable months of each year 1950-2000, while the same amount invested during the six unfavorable months would have gained just $1,743.

We went on to show that those returns, impressive as they are, could have been improved further by buying a few days before the end of the month in October, and selling a few days into May. We quoted Sy Harding's research from 1964-1998 that showed how adding this "monthly seasonality" to annual seasonality would have caused $10,000 invested during the favorable period (plus the few extra days on each end) to grow to $787,200. (This assumes that during the unfavorable period the money was taken out of stocks and invested in short-term T-bills.) The result of leaving the money in stocks year round? Returns would have been just $425,200. Seasonality offered nearly twice the gains with half the risk, since the money was out of the market for half the year.

For our purposes in SMI, that was as far as we took the seasonality idea until October 2003. The numbers show how incredibly effective that application has been, plus it's very easy to understand, since the entry and exit dates are the same every year. But it's not the end of the story. There's another step that has produced even more incredible results.

This additional step adds a simple market-timing indicator called MACD (Moving Average Convergence Divergence) to determine the exact dates to buy and sell. Now that you've had a chance to get used to the idea of annual seasonality — initially a foreign concept to today's investors steeped in buy-and-hold orthodoxy — we're going to introduce you to the MACD indicator as a guide to annual seasonality decision-making.

The full details of what MACD is and how it works is more than I want to get into here (feel free to Google it if you're interested in the nitty-gritty detail). The short version is that MACD is a very short-term timing indicator that is useful at times like October and April when we're already looking for a shift in the market trend. It can help us buy or sell up to a month early or late based on the current market trend.

That extra little flexibility to buy stocks earlier in October or sell them later in May has produced a staggering improvement in the returns of the strategy (which were already pretty eye-popping). Earlier I mentioned how using static buy/sell dates, Harding reported that a $10,000 Dow investment in 1964 would have been worth $787,200 by 1998. Using MACD to signal the flexible buy/sell dates instead, that $10,000 investment would have grown to $1,144,100! That's nearly triple the amount $10,000 would have grown to if left in stocks that entire period ($399,400).

MACD can help us in two different ways. For example, in the Fall we start looking for a MACD Buy signal beginning the first of October. MACD can help us if the market starts turning up before October is over by providing an "early" buy signal prior to the "regular buy date" (five-trading-days before the end of the month). Or it may help us in a declining market by providing a "late" buy signal as late as a few weeks into November. The same is true in reverse regarding the end-of-April sell signal.

The main reason we haven't mentioned MACD before is that it adds a lot of complexity to what is otherwise a very simple seasonality strategy. The MACD signals themselves require a bit of interpretation, which is unappealing. For example, the two sources mentioned above, Hirsch and Harding, were a month apart in the spring of 2003 when they told their readers to sell. It turned out that Hirsch went early, costing its readers a bundle by selling early into what turned out to be a substantial rally.

A word of caution is in order. MACD is not a magic bullet that will transform you into an ace market timer, and we're not depending on it as a core signal flag. It's just a very short-term trend indicator. We start looking at it the first of April and October only because those are junctures where we're predisposed to buy/sell anyway according to our seasonality strategy, and want a little extra fine-tuning to help select exactly which day.

Don't be tempted to think that MACD will help you jump in and out of the market - that's the wrong use for this tool. However, if you know you're going to be buying or selling anyway, it can help fine-tune that decision a little bit. In other words, as always, start with the big picture of your overall plan, narrow your focus to the strategy level, then use indicators like MACD to fine-tune. Where people get in trouble is when they reverse that pyramid, and instead of starting broad in their thinking (at the plan level), they let the narrow end (the indicators) start driving their decision-making.