Here's why investors shouldn't try to play stock market trends

It's coming! September is the worst month, on average, for stocks. Are you ready?

No, I haven't fallen for seasonality. But let's have fun. We've all heard the 'sell in May and go away' maxim. May through October is the weakest six-month stretch, so seasonal types recommend avoiding it.

Yet monthly returns show it isn't so simple. Since 1928, the S&P 500's best month is July - smack in the middle of that supposedly weak stretch! August historically rocked, too.

The only really bad month is September, down 1 per cent in price terms and 0.7 per cent when you include dividends. Even May, so widely feared, averages 0.2 per cent with dividends (-0.1 per cent without).

So if you believe in seasonality, here's how you really want to play it. Sell on May Day - the month averages weak returns and is down 44 per cent of the time. But buy back sometime in June to get some of its 1 per cent average total return.

Stay for all of July to get that 1.8 per cent average total return - and most of August, which averages another 1 per cent. But get out for September - yuck! Don't stay out too long, though, because October is up.

Again: Sell in May, buy in June, sell in August, buy in October, hope your trading costs don't make the whole thing moot. Got that?

But maybe May Day isn't it! Perhaps we must play with daily data to find the best day to do all this. Is selling on the May's first Monday better than waiting till the second Friday?

Should we buy back on 1 June or wait for the first Tuesday after the first Saturday? Do we want to skip all of September, or is that first week OK? So many decisions!

You could figure this out if you want, but I'm not sure it's worth it. One, you'd need an Excel spreadsheet with about 22,250 rows - and you'd pop a blood vessel in your eyes staring at that much data.

Two, it isn't necessary - none of this actually works! Averages aren't predictive. Nor are any of these months uniformly positive, negative, up big or whatever else. July has great average returns but was negative in 35 of 89 calendar years since 1928. Yucky September rose 51 times.

How can you tell the good Septembers from bad ones before they start? You can't! Stocks are too volatile in the short term. September could easily rally during a bear market or flop in a bull.

Four negative Septembers since 2009

This bull market, which started in March 2009, has four negative Septembers - 2011, 2014, 2015 and 2016. If you skipped them all, you'd save yourself from an 11.7 per cent compound drop. But if you skipped all this bull's Septembers, you'd miss a 4.7 per cent compound gain. The good outweighed the bad.

Back to my original question: The smartest way to be ready for a bull market September isn't to sell. Instead, grit your teeth and wait.

If September booms, smile all the way to the bank. If it sags, remember wobbles are normal and usually bounce high and fast.

If you sell at the wrong time, you miss the rebound - not something anyone wants to do! So think long term and own stocks. Here are two for your buy list:

Semiconductor chips are getting smaller and smaller, yet the capital required to make the latest and greatest chips continues to march higher - great for industry leaders like Japan's Tokyo Electron.

It should get a double boost from rising wafer fab equipment spending and leadership shifting to non-US stocks. Buy it cheap today at just 16 times my 2018 earnings estimate, and watch the profits get bigger as the chips get smaller.

As more of the world moves online, more businesses need Capgemini (CAP), the French IT services provider, to usher them into the digital age.

Forget old fears over declining demand for legacy IT services like application development and maintenance. The digital future shines brighter than Castor and Pollux, and few fathom Capgemini's transition to a digital leader.

A bargain relative to competitors at 16 times my 2018 earnings estimate, buy it now for an extra boost from Europe's accelerating economy and rising demand for tech services.

Add new comment