In February 2012, I wrote about research that showed that high-yield, low-payout stocks outperformed the overall market in total return, as well as other yield-payout groups. I created the DG-HYLP model portfolio based on this research, and last rebalanced it in July 2012. Mathematically, the combination of high yield and low payout ratio identifies stocks with a low price-to-earnings [PE] ratio, which could be an indicator of value. Furthermore, the low payout ratio means there is room for the firm to increase its dividend. The use of the dividend growth rate [DGR] in the ranking system favors firms that don't just have the capacity to increase the dividend, but actually do raise it. These should be favorable characteristics for the stocks in this portfolio, in addition to the higher yield.
Over the last year, the DG-HYLP underperformed my other DG model portfolios. The portfolio had some rather volatile stocks, and because of its value focus, I did not initially implement my -20% stop-loss rule that the other portfolios use as a sell signal. This rule is now active for the DG-HYLP portfolio to help minimize losses. In addition, I tweaked the screening process to remove stocks with negative earnings growth projections, as earnings drive dividend growth and stock price. This caused some additional turnover, 18 of the 30 holdings were replaced in this rebalance.
Determining "High" and "Low"
In the research, stocks were sorted into high, medium, and low groups based on yield and payout, but no specific cutoff numbers were given. I sorted the combined CCC stock list [471 stocks] by payout and by yield, and checked each third to get an idea of what the cutoffs looked like based on the data set. This check yielded the following cutoffs:
2.0% <= x < 3.2%
35% <= x < 59%
>=59% or n/a
Compared to last summer, the yield cutoffs have shifted downward. 3.2% hardly feels "high", but relative to 0.5% CDs, I suppose it is. I didn't want to be too restrictive, as I apply additional filters later, and this universe also included MLPs and REITs, which were likely to be eliminated due to high payout ratios. I reduced the high yield cutoff, as this group probably would not have many 4%+ stocks, and allowed for a higher payout to capture more higher yield stocks. Readers are free to choose different cutoffs.
2.1% < x < 3.3%
45% < x < 65%
>=65% or n/a
I started with the combined CCC list [471 stocks] and added a column for "Modified Payout Ratio." Basically this value equals the EPS payout ratio when present or else the free cash flow [FCF] payout. For MLPs and REITs, I used the FCF payout. Next, I performed the following filters:
- Removed all firms with a market cap less than $250MM, which were mostly small-cap financials. [426 stocks remained]
- Removed 12 stocks with "n/a" for the Modified Payout Ratio. 
- Removed 19 stocks that were overdue for a dividend increase 
- Removed BHP Billiton (BHP), as BBL is the same stock and has no foreign dividend tax withholding. 
Using these cutoffs, I sorted the remaining 394 stocks into the nine combinations of low, medium, and high payout and yield. The table below shows how many stocks are in each classification. In general, I classify the top left squares as signaling an overvalued stock, based on the high payout and low yield. There are few stocks in that region. Stocks that freeze or cut their dividend are removed from the CCC list, so there is survivor bias. There are many stocks in the Low-Low group, which includes smaller firms with good earnings and dividend growth. The High-High group contains most of the MLPs, utility, tobacco, and REIT stocks. It has higher yield, but potentially less growth.
The lower right quadrant is the focus of this model. The Medium-Medium group contains the likes of McDonald's (MCD), Procter & Gamble (PG), and Colgate (CL). Nothing wrong with these stocks, but as we move to the right, we get similar yields with lower payout ratios, so there is more room to grow the dividend. Moving to the High Yield-Medium Payout group provides higher yields than MCD at comparable payout levels, which also seems like a good upgrade. The holy grail is the High Yield-Low Payout group, where just 11 of the original 471 stocks ended up. In an effort to achieve more value, this model will not utilize stocks from the Medium-Medium group. I combined the Hi-Med, Hi-Low, and Med-Low yield-payout groups to create the screened group [97 stocks] and proceeded to calculate sector weightings.
Screened Group Characteristics
I added sector tags and sorted the list by sector and industry. By count, the group is heavy on financials [29%], industrials [16%], and utility [12%] stocks. However, the financial and utility stocks were mainly small caps, so I also calculated market cap weightings, which favored energy [27%], info technology [13%], and industrials [13%]. To determine the portfolio weightings, I averaged the count and market cap weightings. Financials, industrials, and health care received the highest allotment of stocks [4 each]. The purpose of this weighting is to allocate the 30 stocks in this portfolio in a manner that represents the overall characteristics of the larger group. Those of you with lots of money could just buy all 97.
For reference, the screened group had an average yield of 3.2%, a payout ratio of 42.9%, a beta of 0.85 and 1-, 3-, and 5-year dividend growth rates [DGRs] of 11.6%, 10.8%, and 11.9% respectively. While the yield is lower than my other income-focused models, the payout ratio is also lower and the DGRs are consistently in the double-digits.
Stock Selection Process
Because the focus of this model is on dividend yield and payout ratio, I created a scoring formula that utilized those two variables, in addition to the next year's earnings growth rate and the 1-year and 3-year DGRs. The formula ranked these metrics for each stock according to the table above. The earnings growth rate and two DGR values are averaged together and used as a proxy for the strength of future dividend growth. This value is combined with the payout and yield rankings. Since I wanted the model to be relevant for income investors, I double-weighted the yield score to emphasize yield, knowing that the average payout for this group is relatively low. I then searched for the highest scoring stocks in each sector, with preference given to those with higher yields in the case of a tie score. I also used the following rules:
- Exclude stocks with a zero ranking for [negative] earnings growth, as firms with positive earnings growth are preferable.
- Limit the Industrial sector to two stocks from the defense/aerospace industry. Defense firms made up 4 of the top 5 ranked stocks in this sector, which would not provide much diversification.
- Remove stocks that triggered the -20% stop-loss rule after their last dividend increase. There were four stocks in this category: NTT DoCoMo (DCM), Intel (INTC), Occidental Petroleum (OXY), and John Wiley & Sons (JW-A).
The resulting DG-HYLP portfolio is shown below, as well as a yield-payout matrix. It has a yield of 3.53%, a payout ratio of 45%, a 0.86 beta, and 1-, 3-, and 5-year DGRs of 15.2%, 13.7%, and 17.0%. Relative to the overall screened stock group, this portfolio has about the same beta and payout ratio, with higher yield and much higher DGRs.
Medium Payout (45% to 65%)
Low Payout (<=45%)
Nu Skin (NUS)
Exxon Mobil (XOM)
Cardinal Health (CAH)
Baxter Intl (BAX)
Harris Corp (HRS)
High Yield (3.3%+)
Darden Restaurants (DRI)
Sunoco Logistic Partners (SXL)
Arrow Financial (AROW)
Lockheed Martin (LMT)
Sonoco Products (SON)
Shaw Communications (SJR)
Rogers Communications (RCI)
Laclede Group (LG)
NV Energy (NVE)
Universal Corp (UVV)
Conoco Phillips (COP)
Universal Health REIT (UHT)
First of Long Island (FLIC)
Northrop Grumman (NOC)
Textainer Group (TGH)
BHP Billiton (BBL)
Given the run-up in the market and the lack of MLPs or REITs in this portfolio, I was not surprised to see the portfolio yield come in well below 4%. While I still really like the thesis behind a high-yield, low-payout model, I'm starting to think that there may be more value in applying these traits to the DG-IncomeGrowth and DG-LowBetaHighDGR portfolios. Low payout stocks typically don't have what most would consider "high yield" [4%+]; combining these traits does not provide a large group to choose from. The DG-IncomeGrowth model pursues higher yield stocks, including REITs and MLPs, without a hard payout ratio limit, but it checks cash flow for stocks with higher payout ratios. The DG-LowBetaHighDGR model doesn't necessarily require a low payout ratio, but typically stocks with a high DGR have lower payout ratios, and its beta comes in much lower than this model. The DG-HYLP model does a great job of finding low PE stocks with high dividend growth and moderate yields. I'm not giving up on it, but my hunch is that the other models will perform better. However, I only have one year of data for the DG-HYLP, so we'll see what returns its second year brings.
I updated the virtual portfolio with the new holdings as of the closing prices on March 3, 2013. I will continue to track this portfolio and report monthly on its performance relative to the S&P (SPY), S&P Dividend ETF (SDY), and the other DG model portfolios. The next rebalance will occur in early September 2013. Cash from dividends will be invested in a corporate bond fund at the end of each month.
I welcome feedback on this model and its process, as I continue to refine it based on comments, observations, and new learning. I hope SA members find the model and the recommended list useful for identifying potential candidates for their portfolios.