A few months ago, I wrote two articles (Part 1, Part 2) about a potential system for detecting dividend cuts for consistent dividend growth stocks. The purpose was to help DG investors to avoid the price declines that come when a company cuts its dividend. In fact, the price typically has already declined because the company’s financial difficulties are known prior to things getting so bad that a dividend cut occurs. My thought was that perhaps this price movement could signal the upcoming cut, based on the concepts of Efficient Market Theory, where price captures all known information.
Using a list of dividend cutters from the last 10 years or so, plus stocks in my model DG portfolios (Small Cap, Dividend Aristocrats, Income-Growth), I observed that DG stocks that underperformed the S&P 500 index (SPY) by 20 percentage points for four consecutive weeks (using weekly data), tended to either cut their dividend in the near future and/or experience a further price decline averaging near 20%. I tended to be rather aggressive when choosing my time periods, deliberately choosing starting points where the stock price was near a high versus the S&P. Over 90% of these stocks declined after the signal date, so from a total value standpoint, exiting these stocks and investing the funds elsewhere was a profitable decision. I have since included this test for stocks in my model portfolios to hopefully help them avoid major price declines.
When the prior articles were published, SDS provided me with a list of stocks from David Fish’s CCC lists that he found to meet the -20% gap criterion. Now that three months have passed, I pulled data for those stocks, David Fish’s most recent Overdue Dividend Increases list, and David Van Knapp’s Dividends in Danger list, to see what happened. The complete list is at the end of the article, and you can download my Excel spreadsheet with the data here. The stocks from their lists are in blue on the chart. Note that the prices are not adjusted for dividends, so you could add 2%-3% to the average returns to account for that. It’s difficult since the time periods are different for each stock, but most are around six months. Given the magnitude of the average decline, it doesn’t change the overall result.
The purpose of this article is to share these results and my analysis, and to continue the discussion about ways to avoid losses as part of a dividend growth investment strategy. Overall, the use of the -20% gap criterion as a dividend cut detector is still questionable, but its use as a stop-loss strategy appears stronger.
New Data Results Of the 40 stocks on my list, only 3 cut their dividend, so the -20% criterion as a dividend cut predictor appears less useful than in my prior sample data. The cutters were Getty Realty (GTY), Infosys (INFY), and Orrstown Financial (ORRF). About half of the stocks are overdue for a dividend increase though, so the theory may yet play out. I will need to check again in three months.
32.5% of stocks raised their dividend after the signal; 60 % are either overdue (42.5%) or will become overdue (18.5%) if they don’t raise it in the next quarter.
Since the signal date (fourth week of -20%) through January 11, 2012:
- 67.5% of the stocks declined, while 32.5% increased.
- On average, the percentage change in price was -11.1%.
- Only four stocks had price increases > 10%, while 18 had price declines > 10%. Meredith (MDP) changed its dividend policy, increasing dividends significantly. Murphy Oil (MUR) is overdue for a raise. StanCorp Financial (SFG) and SEI Investments (SEIC) pay dividends annually or semi-annually, which makes them different from the group. Perhaps they should be excluded from this strategy for that reason, as it is harder to correlate price to dividend changes due to less frequent dividend payments.
- On average, stocks in this group declined 25.5% at some point after the signal (i.e. lowest price point).
- 60% of the group had a low price point 20% or more below the signal date price.
- 90% of the group had a low price point at least 9% below the signal date price.
While this research originally started as an effort to identify dividend cuts, it appears that my “-20% rule” is more effective at signaling when to get out of a DG stock to avoid future losses. On average, this group lost around 11% if one held the stocks through January 11, 2012. In addition, at some point during the last year, the investor would have been down around 25% on average for these holdings, which is probably more volatility than some DG investors would care to experience. Only a few companies went positive after the signal, so from a total value standpoint, selling was the right move.
For those focused on the income stream, only three of 40 have cut their dividend, so perhaps holding on was not so bad. However, 24 of 40 are overdue or will be overdue if they don’t raise their dividend in the next quarter. While a frozen dividend is not a huge problem, it fails to meet the growth part of “dividend growth,” unless you believe the company will resume increases at some point in the future. I know some investors don’t worry as much about the price as the income, but if the dividend starts to falter, the price will decline, and then if you decide to sell, those unrealized price losses will become quite real.
Re-Entering a Stock and the Impacts
Given that this signal is supposed to detect a dividend cut, it made sense to me to re-enter the stock if the company did the opposite and raised its dividend; that is, the signal was wrong. The chart includes the date and price for dividend raise announcements (13 of them), and the impact of making this trade. Effectively, the investor still owns the stock, but sold at the signal date and repurchased on the increase announcement date, so that difference is the savings or loss. I did not include transaction fees or tax implications, so assume this is in an IRA.
The “Impact” column of the chart contains the savings or loss from executing each trade. For stocks that did not raise their dividend, this is the difference between the signal date price and the current price. For stocks that raised their dividend, it is the difference between the signal price and the re-entry price. On average, following these rules resulted in an 11% savings to the investor relative to just holding on to all of the stocks.
For stocks that were repurchased, there was an average price decline of 0.4% since the repurchase date. This means that most of the 11% savings came from avoiding losses on the stocks that did not raise their dividends. However, there were two cases, Avon Products (AVP) and Old Republic (ORI), where the dividend increase was within one month of the signal date. In both of these situations, the stock declined further by over 20%. It makes me wonder if more time should elapse before re-entering a stock to ensure that its situation has improved. If these two stocks were not repurchased, the price change for re-entered stocks increased 5.2% on average, and the impact change increased to 12.6%.
Based on this set of data, the “-20% rule” may not be as good of a dividend cut predictor as it seemed, unless a majority of the stocks overdue for a raise end up cutting dividends. Of course, the rule could also be modified to detect no-raise or cut, as opposed to just cuts. That middle group (no-raise) was difficult to quantify in my prior project research too. On average though, the rule accurately predicts future declines of 10% or more in DG stocks, which can help investors avoid further losses.
Obviously if an investor knows more about the company, its financials, and its prospects, s/he may decide the decline will be temporary and perhaps be a time to buy more. For those investing in DG stocks based on their consistent dividend growth and who don’t like to experience large price swings, this strategy gives you a supporting reason to exit and move on, as it continues to identify stocks overdue for dividend increases and that have 10%+ price declines. Given that there are over 400 stocks on the CCC lists, most of which have not exhibited this signal (i.e. haven’t already lost 20% more than the S&P), switching to another stock as a replacement isn’t that difficult. As a total return investor, this strategy will help to avoid losses and to avoid initiating positions in stocks with this characteristic.
I plan to continue collecting data on stocks that exhibit the -20% trait. If you happen to notice one, please feel free to email me and let me know. There are just too many for me to check, so I stick more to the 80 or so in my model portfolios, and the overdue and potential dividend cut articles by David Fish and David Van Knapp. If anyone has ideas on other ratios that might help to improve accuracy, I’d be interested in your thoughts. Payout ratio is listed on the chart, but most of them are pretty reasonable, so I’m not sure that provides any help. Good luck to all in 2012!
Stocks in the Data Set:
Aflac (AFL), Atlantic Tele-Network (ATNI), Avon (AVP), Badger Meter (BMI), Eagle Financial (EFSI), Eaton Vance (EV), CenturyLink (CTL), Frontier Communications (FTR), Greif Inc (GEF), Getty Realty (GTY), Infosys (INFY), Communications Systems (JCS), Knight Transportation (KNX), Meredith Corp (MDP), Murphy Oil (MUR), NAACO Industries (NC), Inergy LP (NRGY), Corporate Office Properties (OFC), Old Republic Intl (ORI), Orrstown Financial (ORRF), Pacific Gas & Electric (PCG), PP&L Corp (PPL), PartnerRe Ltd (PRE), SEI Investments (SEIC), StanCorp Financial (SFG), Shenandoah Telecom (SHEN), Telephone and Data Systems (TDS), TEVA Pharmaceuticals (TEVA), Thomson Reuters (TRI), United Community Bancorp (UCBA), Westamerica Bancorporation (WABC), Weyco Group (WEYS), Washington REIT (WRE), Norwood Financial (NWFL), DeVry Inc (DV), American Greetings (AM), Arch Coal (ACI), and Strayer Education (STRA).
Disclosure: I am long AFL.