When it comes to investing, there’s no shortage of bad advice floating around out there. Among the worst, though, is the old saw, “You can’t go broke by taking a profit.”

The saying refers to the belief that if you have a stock that’s gone up in value, it’s hard to go wrong selling that stock and “locking in” the gains. But while the saying is technically true — it’s hard to picture a scenario where an investor is suddenly bankrupt after selling a stock at a profit — it’s a dangerous platitude for investors to follow.

There’s a name for that
The practice of selling winning stocks and hanging on to losing ones is a practice that’s familiar to behavioral-finance experts. It’s a behavioral bias known as the disposition effect and has been revealed to be quite harmful for investors. A number of academic papers have shed light on the subject, including Berkeley professor Terrance Odean’s 1998 study that concluded that individual investors’ “preference for selling winners and holding losers … leads, in fact, to lower returns.”

A possible explanation
If the long-term returns from stocks were distributed normally — that is, they formed the familiar bell-shaped curve and most stocks’ returns clustered around the average — selling winners and holding losers might actually work. If the returns from most individual stocks were likely to be right around the average for all stocks, then a big winner would be more likely to stall out after its winning streak than continue climbing. At the other end, it wouldn’t be unreasonable to expect a stock that’s been a big loser to climb back closer to the average.

But that’s not how it works.

I was reminded of this by a recent report by Shankar Vedantam for NPR, called “Put Away the Bell Curve: Most of Us Aren’t Average.  Vedantam reviewed the research and work of Ernest O’Boyle Jr. and Herman Aguinis, who studied the performance of 633,263 people involved in academia, sports, politics, and entertainment.

In short, the pair’s finding was that the performance distribution in these groups wasn’t bell-shaped. Instead, many participants clustered below the mathematical average, while a group of superstars produced results far above the average and pulled the overall average up.

Stock returns have a similar distaste for fitting to a bell curve. Over the past 10 years, 63% of the S&P 500 companies underperformed the average. Meanwhile, a large group of significant outperformers delivered returns that were well above the average.

As compared with the bell curve in the background, the data plotted here is a mess. And it should be. Stock returns are not normally distributed — which is what produces that nice bell-shaped curve. And though stats-stars who are much smarter than me often try to describe stock returns as “lognormal” — a mathematical transformation of the returns that gets them to more closely fit a bell curve — they’re not that, either. Stocks are typified by “fat tails” on either end — that is, more seriously outperforming and underperforming stocks than is easily captured by streamlined mathematical models.

So no matter how you look at stock returns, a surprising number of stocks end up returning far more and far less than the average. Practically, this means that the practice of “locking in gains” and hanging on to losers is a good way to miss out on the market’s huge outperformers, stay stuck with poor performers, and earn lackluster overall returns.

HT: iShares

EconomPicData Blog proved that relative strength can also be applied to home values:

How well would an investor have done applying momentum to the various cities that make up the Case Shiller Home Price Index, pretending (of course) that each city index was investable and liquid (i.e. things they aren’t).

First, a quick update on the Case Shiller Home Price Index.  The Huffington Post:

The Standard & Poor’s/Case-Shiller home-price index shows that prices dripped in February from January in 16 of the 20 cities it tracks.

The steady price declines have brought the nationwide index to its late 2002 level.  Home prices have fallen 35 percent since the housing bust.

Prices in nine cities fell to their lowest levels since the housing bust.  The average price in Atlanta fell 17.3 percent in February compared with a year earlier.  That’s the biggest annual drop in the history of the index for any city.

Yikes…let’s see what momentum can do with this mess.

Rules…

1) Take the 6-month rolling return for each city

2) Allocate the next month to the city that had the highest six month return

How well would we have done?

The chart below outlines the performance of this relative strength allocation, the composite-10, andPortland(which happened to be the best performing city over this time frame… who knew).

For those keeping track at home, that’s a 12.7% annualized return for the relative strength index vs. 3.3% for the composite-10 and 4.8% for Portland, despite there being no rule that an investor get out of the market.

Not too bad.

This demonstrates again that the best way to find future winners is to buy current winners and stay with them as long as they remain strong.

HT: Abnormal Returns

RS leaders and RS laggards have had similar performance over the past couple of years.  History would strongly suggest that we will eventually see RS leaders resume their outperformance.

The central truth of the investment business is that investment behavior is driven by career risk.  In the professional investment business we are all agents, managing other peoples’ money.  The prime directive, as Keynes knew so well, is first and last to keep your job.  To do this, he explained that you must never, ever be wrong on your own.  To prevent this calamity, professional investors pay ruthless attention to what other investors in general are doing.  The great majority “go with the flow,” either completely or partially.  This creates herding, or momentum, which drives prices far above or far below fair price.  There are many other inefficiencies in market pricing, but this is by far the largest.

Going with the flow in an unsystematic way is likely to lead to poor results, but capitalizing on this market inefficiency in a systematic manner has demonstrated the ability to provide superior performance over time.

The 10-day moving average of this indicator is 77% and the one-day reading is 43%.  The high RS universe has taken a hit over the last week.

RS leaders and RS laggards have had similar performance over the past couple of years.  History would strongly suggest that we will eventually see RS leaders resume their outperformance.

See www.powershares.com for more information.

The S&P 500 had a great first quarter ending up about 12% (price only).  Relative strength strategies did OK.  The best performing factors during Q1 were actually the factors that performed the worst over a long time horizon (see the white paper for details).  Several of the best long-term winning factors had a tough time in Q1.

(Click To Enlarge)

The graph above shows the returns for all 100 trials for each of the time-based RS factors we track.  A trailing 18 month and 36 month window to compute RS worked very well.  These worked well because those models didn’t rotate into low volatility names at the end of last year, and then rotate back out of them during Q1.  In effect, the long time horizon allowed them to capitalize on the laggard bounce that was so prevalent during the first part of the quarter.  The very short-term windows also did well.  They were able to quickly rotate into the high beta names that were the leadership.  But, more importantly, that trend was sustainable so the short-term mean reversion effect didn’t hurt those factors in Q1.  The 6 month and 9 month factors performed very poorly.  The main reason is these intermediate term factors rotated into low beta and high dividend stocks at the end of last year.  Those were the laggards during Q1, and it took some time for those models to rotate into the new leadership.  Keep in mind, however, that these two factors are two of the best performing over long time horizons.

The laggard bounce was most pronounced in January and February.  By March things had settled back down and the intermediate term factors were performing well.  The better performance was the result of the market rewarding intermediate term momentum, and the models having a chance to shed the laggards and re-position themselves into the current leadership.

January Performance

(Click To Enlarge)

February Performance

(Click To Enlarge)

March Performance

(Click To Enlarge)

The turnaround for intermediate term momentum strategies wasn’t enough to totally reverse the underperformance during the first two months of the year.  But it is very good to see the intermediate term factors getting back into gear!  We noticed the same thing in our managed portfolios too.  Things definitely picked up in the last part of the quarter for high RS stocks.

All of the factors in this post are simple, time based relative strength (momentum) factors.  These are the factors that match what we published in the white paper.  We do track other RS factors though.  It is interesting to note, that the underperformance of the intermediate term factors was most pronounced in the simple, time based factors.  Intermediate term factors we track that use some sort of smoothing or multiple time periods performed much better than the 6 and 9 month factors.  The only explanation I have for that is that the 6 month ranking window was the perfect time to maximize your whipsaw into low volatility and back out again.  The smoothed and compound factors did a much better job this quarter at avoiding that whipsaw.

Asset classes have become so highly correlated over the past few years that many traditional diversification strategies have lost their effectiveness.

For example, take the link between growth and value stocks.

For the decade ended December 2000, the correlation between the Russell 1000 Growth Index and the Russell 1000 Value Index was just 57%. During the decade ended this past December, it jumped to 92%.

For a more extreme case, compare the correlation of the MSCI Emerging Markets Index with the Russell growth index. The former was negatively correlated to the latter by 6% — which was great for those seeking diversification — in the decade through December 2000, but the correlation spiked to 89% in the following decade.

You can see the issue—drastically changing correlations will move your efficient frontier far from where you imagined it was.

Some of the observers Mr. Benjamin quoted were blunt:

“Traditional diversification is like a seat belt that only works when you’re not in a car accident,” said Michael Abelson, senior vice president of investments at Genworth Financial Wealth Management Inc.

“Depending on risk tolerance, we might recommend allocating half a portfolio to a diversified strategic strategy and then 30% to 35% to a tactical strategy and 15% to 20% to alternatives,” Mr. Abelson said.

Besides having a knack for a fine turn of phrase, Mr. Abelson mentions something that we have noticed more and more in recent years.  It used to be the case that tactical allocation was used as a satellite strategy and might get only a 10% slice of a portfolio.  Now, we often see the tactical strategy with a 35-50% weight.  Some advisors are even using the tactical allocation as the core strategy and arranging alternatives and other asset classes as strategic overweights.

With the rise of tactical allocation come new challenges.  Chief among them is how to manage the tactical portion of the portfolio.  All-in/all-out timing decisions are notoriously difficult to get right.  Overweighting and underweighting based on valuation requires sophisticated modeling that must be constantly updated.  In addition, many assets are resistant to traditional valuation methods.

One method that does work over time is tactical asset class rotation using relative strength.  We’ve chosen that path for our Global Macro strategy because it allows a very large and diversified universe to be ranked on the same metric.  That, and because it works.

Click here and here for disclosures.  Past performance is no guarantee of future returns.

The 10-day moving average of this indicator is 88% and the one-day reading is 88%.