How To Build A Safer Dividend Strategy

by: InvestorsEdge

Summary

What makes a safe dividend stock?

Combining factors doesn't always improve your returns.

We build a strategy that would have produced 24% returns over 17 years.

In the first article in this series ("When Does Chasing Yield Become Dangerous To Your Wealth?"), we examined a strategy that focused solely on chasing yield, and confirmed that buying a stock simply because it has a high yield would have provided you with mediocre returns and an uncomfortable level of volatility over the past 17 years.

While that was possibly not the most shocking fact that dropped into your inbox that day, what we did find was that simply investing in the top 10 U.S. stocks with the highest sub-5% yields every year would have returned 14% in the same period, with far less volatility.

Today we are going search for factors that, when combined, result in a robust and profitable income-based trading strategy.

What makes a "safe" dividend stock

While we all have different factors that we examine when trying to ascertain the safety of a dividend, they tend to fall into the following broad categories:

  • Low payout ratios - Companies with lower payout ratios should have more ability to both increase their returns to shareholders and maintain payments during a crisis.
  • History of increasing dividends - A company with a long history of increasing distributions to shareholders should continue to do so.
  • Improving results - Companies that are growing their revenues and profits should be able to increase their returns to shareholders.
  • High efficiency - The more efficiently the company utilizes its assets, capital and equity, the more chances are that it will be able to maintain and increase its dividend payments.
  • Low debt - Companies with low debt should be more able to maintain and increase their dividend payments.

Let's start with the first factor: low payout ratios. To test these factors we are going to use the InvestorsEdge.net platform. You can see the results and statistics on the initial model here. To view all 23 models we have used to analyze our factors, simply click on the History tab and navigate between the different versions.

Our initial model tests the base trading strategy:

  • Our tradeable universe will include all common stocks and depository receipts (DRs) in the U.S. that have returned cash in the form of dividends or distributions to shareholders within the previous 12 months.
  • We will rebalance our portfolio on an annual basis.
  • We'll start off with an initial cash sum of US $10,000.
  • Each transaction will cost a flat fee of US $7.
  • We will simulate using a Market On Close order to buy stocks at their closing price on the next trading day after the rebalance.
  • At each rebalance point, we will buy the 30 stocks ranked equally by their trailing yield and payout ratios.

Here's how the model performed, benchmarked against the S&P 500:

Yield + Payout Ratio Strategy

You can see that our strategy would have returned 14% a year, with drawdowns that are quite comfortable for most of the time, with the exception of the financial crisis in 2008. Over the backtest period, dividends and spin-offs would have accounted for 40% of our returns.

We will use this as our baseline to help us analyze the impact of our new factors on our model.

Dividend History

Our second iteration equally ranks each stock in our universe by trailing yield, payout ratio and the number of years of increases in annual dividend payments over the past 25 years:

Dividend Increases Strategy

You can see here that including our dividend increase count has reduced the performance of our model by 1.5% annually, which was a surprise to us given the prominence that is accorded to Dividend Aristocrat status by many income investors. While this reduced cost was justified slightly by reduced volatility in the portfolio, we don't think there was a huge amount of risk in the original model in the first place.

When you examine the result statistics (in the History tab of the model, click on the Div Increases row), you can see that the major difference between the two models is in the dividend payments - basing the selection criteria on a count of dividend increases pushes us towards stocks that are "safer," where investors accept lower yields in return for lower risk.

When you substitute the number of increases for the average annual percentage increase in dividends over the previous 3, 5 and 10 years, the numbers are similarly depressed. This time, though, the dividends returned from the portfolio remain high, but the capital gains on the stocks go down.

Dividend 5Yr CAGR

In conclusion, for this model, selecting stocks with better track records of dividend rises doesn't lead to improved returns, nor does it improve any of the 23 versions of this strategy that we tested.

Improving Results

For our third group of factors, we have ranked our stock universe by trailing yield, payout ratio and the % increase from the previous quarter in the following factors:

Factor

CAGR

Total Return

Sharpe Ratio

Sales

14.7%

989%

0.76

EPS

13.0%

743%

0.71

Cash Flow

16.0%

1,224%

0.83

Free Cash Flow

16.3%

1,291%

0.86

Combined

13.6%

815%

0.76

You can see from the above table that the most influential factor of the set was Free Cash Flow. This is not surprising; Free Cash Flow is the amount of cash produced by a company after accounting for capital expenditures and represents excess cash available for expansion and dividends.

Efficiency

We have used Return on Assets (ROA), Return on Equity (ROE) and Return on Capital Employed (ROCE) to ascertain the efficiency of a company. As before, we are ranking our stocks on trailing yield, payout ratio and each of these factors:

Factor

CAGR

Total Return

Sharpe Ratio

ROA

15.5%

1,122%

0.86

ROE

13.3%

779%

0.71

ROCE

17.3%

1,500%

0.91

The ROCE figure is the most influential efficiency factor that we tested.

Debt

The last of our factor groups focuses on debt. Here are the results of ranking our universe by trailing yield, payout ratio and...:

Factor

CAGR

Total Return

Sharpe Ratio

Debt-to-Assets

11.6%

578%

0.64

Debt-to-Equity

13.6%

815%

0.75

Debt-to-Enterprise Value

15.2%

1,084.8%

0.77

For this model, none of the debt factors improved results dramatically when compared to the other groups considered above. Indeed, if you look at the results for each of them, you can see that including debt ratios actually increases the volatility of returns slightly.

Putting It All Together

Our analysis has concluded that of all the factors tested, ROCE and quarter-on-quarter changes in Free Cash Flow are the most influential ratios to use when considering one stock over another, whilst using debt ratios and dividend history as a consideration actually causes a drag on returns.

So, if we combine trailing yield, payout ratio, ROCE and % change in Free Cash Flow and re-run our backtest, we should get a pretty good increase in our returns, right? Wrong:
Putting it all together

You can see here what often happens when you combine factors - they don't play nicely together. This happens because when we rebalance our portfolio, we are not just using the data to quantitatively compare companies against each other, but we are also implying that now is a good time to open a position - and the timing of the different ratios just doesn't combine well.

So, how do we combine all this into a trading strategy that is more than the sum of its parts?

Final Cut

In previous articles, we have identified that two ratios that do play well together and provide accurate timing signals are Price-to-Sales and Price-to-Free Cash Flow.

Additionally, there is another factor that often goes under the radar screen - the Piotroski F score. In 2002, Joseph Piotroski published the paper "Value Investing: The use of Historical Financial Statement Information to Separate Winners From Losers," where he described his method for rating stocks. The F score analyses a company and returns a score of 0-9, with 9 being the best. The factors it uses should seem familiar:

  • Return on assets
  • Cash flow
  • Gross margin (sales less cost of goods sold)
  • Leverage (e.g. debt loads)
  • Assets and liabilities
  • Share float

The Piotroski F score provides us with an effective way of measuring improving results, debt and efficiency, and does so by combining these factors so that they complement each other.

So how does the final cut of our strategy perform?

Final cut

This version would have returned 24% a year for the past 17 years, and would have only lost money in 2008 and 2017, which we are only halfway through. From an income perspective, the strategy would have returned a yield of 3.6% per annum.

Final cut drawdowns

Ignoring 2008, the maximum drawdown - which essentially represents how well you would have slept at night following this strategy - would have been less than 20%, except for during 2016, when it went as low as 30% before rapidly recovering.

Final cut market cap

Finally, the distribution across sectors is fairly even, and the strategy mainly opens positions in small companies (those with a market cap between US $300 million and US $2 billion - we're not investing in tiny stocks with no day-to-day liquidity).

Would We Trade This Strategy?

For us to execute a trading strategy in real life, we need to have good reason to trust that it will continue to be profitable. To help us decide the likelihood of this, we have four-high level tests that our model should pass:

  • Investable - The strategy should be tradeable in real life and should scale. Our backtests include trading fees, so frictional costs are already taken into account. From a liquidity point of view, changing the starting cash to US $1 million slightly reduces profitability down to 21% per annum. Pass.
  • Intuitive - There should be logical risk- or behavioral-based reasons why the strategy works. We have attempted to create an income-producing strategy and have focused on factors that affect the payment of distributions to shareholders, namely Free Cash Flow, Yield, Payout Ratio, Debt and Return On Assets that identify companies that are likely to continue paying distributions and increase in value. Pass.
  • Persistent - The factors involved should work over long periods of time. Our tests have concentrated on the 2000-17 period. From an academic point of view, this would not be long enough to prove persistence. However, from our point of view at InvestorsEdge, we consider going through two major market shocks and a series of smaller downturns as enough to convince us the strategy works on a long-term basis. Pass.
  • Pervasive - Pervasiveness is more an ideal than a hard rule - our model should work across countries, regions and sectors. Testing across other individual countries shows returns of 12-17% a year with high Sharpe and Sortino ratios (which measure risk-adjusted returns). A key reason for this reduced return in other countries is almost certainly the value ratios (Price-to-Sales and Price-to-FCF) that we use - Price-to-Sales tends to work fairly effectively in the US but not as well elsewhere. Having said that, a test of all 10 countries currently represented within the InvestorsEdge platform returned an annual gain of 23% a year with a high Sharpe ratio of 1.26. Pass.

Your Takeaway

The aim of this article was to define a robust and profitable income-producing trading strategy, which we believe we have achieved. Our final model would have produced returns of 24% a year on average, 75% of which was capital gains and 25% was distributions to shareholders.

Along the way, we have seen that dividend history and debt can't be relied upon to predict capital gains, and in fact, act as a drag on returns for this model. While we are not suggesting that this research proves that history and leverage have no bearing on future price rises, what we have proved is that sometimes factors don't mix particularly well together.

What this article also highlights two important questions that are rarely asked - just because there's a perfectly logical reason to invest in stocks with low gearing, great dividend histories and high efficiency ratios, do you really understand how these factors combine, and do they provide you with an investing edge?

The article has discussed a lot of background information on the backtested models that is not shown on the graphs and tables - far more risk and return data, including complete lists of stocks that would have been bought and sold, is available for each version of this strategy if you click on the link at the start of this article.

We'll refine our strategy further in the next article in the series.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.