Investors interested in high-yield stocks need to watch out for "lemons," Research Affiliates says.

Research Affiliates may perhaps be best known for the 2% return advantage by which its smart beta strategy is said to outperform cap-weighted indexes.

Now the data-driven investment firm, in its June newsletter, is proposing a dividend investing strategy that also coincidentally results in a 2% per year advantage over a more basic high-yield stock portfolio in a simulation looking back over the past 50 years.

The near zero interest-rate environment that has prevailed since the financial crisis has heightened interest in stock dividends as a potentially more reliable method of generating portfolio income, but one that is nevertheless fraught with its own unique risks.

Migrating from low-yielding bonds to high-yielding stocks might seem a no-brainer since dividends are persistent, less volatile than equity prices and since stock volatility provides opportunities to buy yield cheaply; what’s more, investors can spend the dividends while leaving their principal intact.

Yet as Research Affiliates execs Chris Brightman, Vitali Kalesnik and Engin Kose point out, dividend-paying stocks may resemble one another for their high yield but hidden within the basket containing these equities are “lemons” whose fundamental flaws will give their owners grief.

“As when buying a used car, buying bargain equities can produce nasty surprises,” the authors write.

So while cheap equities (whose dividend yield rises as the price of the stock declines) have historically outperformed expensive equities (and even lowered volatility), high-yield stocks carry the downside of a higher percentage of delisted stocks and slower future dividend growth.

The question Research Affiliates asks is: can an investor enjoy high equity income “cherries” without interference of these occasional, seemingly indistinguishable lemons?

And indeed the Newport Beach, Calif.-based firm suggests three ways to screen out the investment lemons, with each incrementally improving characteristics of the ensuring portfolio.

The first screen is profitability, based on the intuitive notion that a company with poor growth prospects is like used-car lemons that are always in and out of the auto shop.

Citing Blockbuster Video Entertainment as a classic case of a high-dividend-yielding company with poor profitability (after cable, satellite and broadband killed its business), the authors use return on assets as a proxy measure of profitable firms.

Their simulation shows that the high-yield, high-profitability portfolio gained 50 basis points per year (over 50 years) in comparison to a high-yield, low-profitability portfolio. What’s more, the high-quality portfolio had many other advantages: its volatility was lower; the number of delisted companies fell from 8 to 1 (out of 100 companies); and its subsequent five-year dividend growth rate shot up to 18.6% from 10.5%.

Investors only lost 0.1% in dividend yield (from 5.6% to 5.5%) to gain these quality advantages.

The profitability screen fueled the portfolio’s higher performance, not through higher yield (which declined slightly) but through a faster rate of company growth.

The Research Affiliates team applied their second filter with similar results, seeking to screen out distressed companies like General Motors, Lehman brothers, Washington Mutual and Fannie Mae that failed when they could not cover their short-term obligations.

The authors use the debt-coverage ratio as a proxy for such distressed companies, observing that “cheaply priced equities of companies with high distress risk are like the lemons that break down soon after you drive the car off of the lot.”

Thusly filtered, the high-yield, low-distress portfolio generates a 1.6% annual return advantage (13.3% vs. 11.7%), reduced volatility, zero delisted companies (versus 9 for the unfiltered high-yield portfolio) and future dividend growth of 17.8% versus 12.1%. Once again, the investor loses just 0.1% in dividend yield (5.5% vs. 5.6%).

The Research Affiliates team then uses its third filter seeking to avoid companies whose high yield masks poor accounting practices, or even fraud. Enron — a once highly admired company that turned out to be booking fake sales — would be the paradigmatic company the portfolio seeks to weed out.

The authors cite research showing that a high level of net operating assets is a red flag for a company whose bookings of receivables may be getting unsustainably ahead of its free cash flow.

Applying this high-accounting quality screen, the 50-year-period simulation shows a higher average annual return of 1.6%, slightly lower defaults and somewhat superior dividend growth rate.

The low-accounting-quality portfolio actually exhibited slightly reduced volatility and (as in in the previous cases) a somewhat higher dividend yield.

But the most impressive results occurred when the Research Affiliates team created a composite of all three screens.

Once again, the low-quality portfolio did have a slightly higher dividend yield, 5.7% compared to the high-quality portfolio’s still high 5.4% yield.

But the high-quality portfolio trounced its low-quality counterpart by a full 2 percentage points, with 13.4% vs. 11.4% average annual return. The former also exhibited lower volatility (13.6% vs. 15.3%) and a higher subsequent 5-year dividend growth rate (18% vs. 11.1%).

The authors’ conclusion is that the simplistic approach of paying the lowest price for a dividend — that is, buying the highest-yielding stocks — will cost investors in the end because the portfolio of cherries will contain more than a few lemons.

Just like a car’s past owners’ driving habits, maintenance practices and accident history are predictive of the new owner’s satisfaction, so too do profitability, distress and accounting practices provide a way to optimize a dividend portfolio.

— Related on ThinkAdvisor: