Backtesting validates the effectiveness of DGI (Dividend Growth Investment) strategies. Over the past 10 years, an investor utilizing this approach enjoyed lower beta, higher Sharpe ratios, lower maximum drawdowns, and considerable Alpha. As a bonus, he beat the market.
However, DGI does not outperform other dividend strategies, as is demonstrated by comparing it with a strategy I have dubbed "Dividend Yield."
This article presents a series of backtests developed on StockScreen123, then proceeds to a discussion of the investment implications of the information developed.
A 10-Year Backtest
(click to enlarge)
The DGI Screen has the following parameters:
Member of S&P 500 (NYSEARCA:SPY)
Increased dividend each of the past 10 years
Above 50th percentile on the site's "Basic: Quality" rating system
The Dividend Yield Screen eliminates the increase requirement, but is otherwise identical. Stocks that pass either screen are sorted by Yield, and the 40 highest selected. The portfolios are rebalanced once a year, resulting in an average 16% turnover for DGI, and 30% for Dividend Yield.
The database utilized by the site is free from survivorship bias, and dividends are considered for the DGI portfolio, the Dividend Yield Portfolio and the S&P 500. Slippage is included. I checked current output of the DGI screen against the list of Dividend Champions and Contenders maintained by SA Contributor David Fish and all were included.
Testing Other Time Periods
(click to enlarge)
Dividend strategies don't uniformly outperform the market, as demonstrated by the current year's results and by the July 1999 - July 2000 time period. However, they display superior results in declining markets, and pass the cruelest test of all - from the dot.com bubble peak to the March 2009 trough - with flying colors.
Dealing with Taxes
DGI turnover, as mentioned earlier, averages 16% annually. For a Dividend Growth investor who simply bought and held as of 11/25/2002, the average return would have been 178%, compared with 87.5% for the S&P 500. Similar results were obtained using other dates and holding until the present.
If a strategy is utilized in a tax-deferred retirement account, the tax issue is less important.
First one thing works, then another. Past results are not indicative of future performance. That having been said, past results are impressive. Beating the market is not the primary concern here. It's more important to evaluate an investment strategy in terms of its goals.
David Van Knapp presents a simple explanation of the DGI strategy as follows:
Let me start with a brief description of dividend growth investing. It is a strategy to accumulate dividend growth stocks that provide "organic" income from rising dividends. The dividends are reinvested during your accumulation years to speed up the accumulation process. Then when you retire, you stop reinvesting and simply take the dividends as income. Depending on how much you have accumulated and other sources of income (such as a pension or Social Security), you may not have to sell anything to create sufficient retirement income. The income is generated naturally by your investments. Studies show that in most years, the rising dividends grow faster than inflation.
The objective is to fund retirement by accumulating a rising stream of dividend income. In the buy and hold example mentioned above, from 11/25/2002 to the present, the investor would have increased his dividend income by 12.76% annualized.
Retirement planning is difficult. Results from investing in equities can display substantial variation, to include drawdowns of more than 50%. Few and fortunate are those whose life situation and career develop smoothly during the accumulation years: job losses, career changes, health concerns and divorce can create situations where retirement savings may have to be used to support immediate needs. When these considerations are factored in, the lower beta and drawdowns displayed by the DGI strategy constitute a critical advantage.
It should be noted that the objective of living off dividends plus Social Security and pension (if any) is difficult. The 20 highest-yielding stocks that would pass the Dividend Growth screen are yielding an average 3.15% at today's share prices. A more conventional recommendation is to plan on withdrawing 4% of portfolio annually, which would reduce the required portfolio size proportionately.
However, the strategy as articulated by Van Knapp does not rely on capital appreciation. As a thought experiment, suppose the market falls off a cliff and the portfolio is cut in half. That happened during the period of time embraced by both the 10-year and 5-year backtests. Because the companies involved have a history of increasing the dividend even during difficult times, the investor most likely would continue to receive increasing dividends. As long as the dividends continue to increase equal to or greater than inflation, the investor is still meeting the stated objective.
(click to enlarge)
This is a different way of looking at the past 10 years. Using the data from the 10-year backtest, I computed the dividend income in dollars, in order to track the development of that amount over the period. It increases an average 12.57% per year. Some years it goes up more than others. However, the trend is clear - a line that slopes consistently upward from left to right.
The Dividend Yield strategy is a stalking horse, for the purposes of this article. Assuming the investor was attempting to live off the dividend income, it averaged 4% for the 10-year backtest period. Total return was 12.99%, so with capital appreciation of approximately 9% the Dividend Yield investor beat inflation even while drawing at 4%.
Performance of both screens is dependent on the inclusion of the "Basic: Quality" rating scheme in the selection criteria. Simply buying the cheapest stock that has a continuous history of dividend increases is not useful. Sometimes a company continues to incur debt and pay high and/or increasing dividends without addressing underlying weaknesses in the business operations or model.
The period of time covered by the backtest has been characterized by extreme turbulence in the market, two recessions, and decreasing interest rates. The capital appreciation component of dividend strategies over the past five years can be attributed to the reduced yields on competing investments, which is unlikely to continue indefinitely and can't go much lower.
The 40-stock maximum used by the screens is intended to provide diversification and make them closer to something that investors would actually do in real life.
If the Dividend Yield investor simply bought the five highest-yielding stocks that passed the quality test, the 10-year average return would have been 19.55%. Most screens achieve better performance if fewer stocks are selected. But as of today three of the five highest-yielding stocks are tobacco companies.
Additional disclosure: I'm a paid subscriber at StockScreen123. As a retail investor, I'm talking shop and not giving advice.