The Dogs of the Dam
Student: Randy Rietdijk Student number: 6140246 Field: Economics
Specialization: Economics and Finance Supervisor: Mr. M.O. Hoyer
Abstract
In this paper there has been investigated whether there is statistically evidence the Dogs of the Dow strategy has outperformed the AEX index for the period 1990 till 2013. For each year of the sample period there had to be calculated which ten stocks of the AEX do have the highest dividend yield to apply the DoD strategy on the Dutch market. These were all called the dogs of the Dam and were for one year part of the portfolio. In each stock there was an equal amount invested. After one year the ten highest dividend yield stocks were derived again and the portfolio was rebalanced to maintain the composition of the ten highest dividend yield stocks.
The monthly stock returns, index returns and risk-free rates were used to calculate the abnormal returns and these monthly time series were also used for risk-adjustments. The abnormal returns and risk-adjustments were calculated on yearly bases, but also on larger periods to look if the strategy worked on the long term.
The results for this strategy applied on the AEX suggest there is no evidence of any outperformance on a yearly base or for a longer period. The mean annual market-adjusted return is 1.07%, but this result is not statistical significant. Other returns calculated over a larger period were also not significantly different than the market. Even if it were significant, the outperformance is that small it will hardly survive taxes or transactions costs.
Table of contents
1.
Introduction ... 4
2.
Literature review ... 5
3.
Data and Methodology ... 7
3.1 Data selection ... 7
3.2 Methodology ... 9
4.
Results ... 11
4.1 Abnormal returns ... 11
4.2 Risk-adjustments ... 16
4.3 Further adjustments ... 19
5.
Conclusions ... 20
6.
Reference list ... 21
1. Introduction
The Dogs of the Dow strategy is an investment strategy, where one holds a portfolio of ten stocks of the Dow Jones Industrial Average (DJIA) with the highest dividend yield. These stocks are called the Dogs of the Dow (DoD). One buys these stocks, each with an equal amount of the investment, at the beginning of the year. After that year, one rebalances it to maintain the portfolio with the ten highest dividend yield stocks.
They call these stocks the dogs, because they are, according to their dividend pay outs, undervalued. The theory is that they will increase again in value. This is the principle of the ‘value’ stocks. On the other side there are the ‘growth’ stocks, which are stocks with a lower value, but do have a higher growth rate for the earnings. And for that reason their stock price is overpriced. These stocks are however less profitable, cause eventually their prices will be corrected in the market. This principle, where the big firms outperform the small firms on the long term, is the so called value premium, which Fama and French (1995) have discovered. This is also supported by Capaul, Rowley and Sharpe (1993). They have done research about the returns of value and growth firms over the period 1981 till 1992 in France, Germany, Switzerland, UK, Japan and US. They concluded that ‘during this period, portfolios of stocks with low price-to-book ratios ("value" stocks) provided risk-adjusted performance superior to that of portfolios of stocks with high price-to-book ratios ("growth" stocks).’
The findings of Fama and French (1995) and Capaul, Rowley and Sharpe (1993) are based on the book-to-market value, but the Dogs of the Dow strategy is based on the dividend yield. However, the theory about the value premium holds also for the dividend yield. Fama and French (1998) discovered namely that there’s also a value premium for stocks with a high dividend yield. In that article they investigated also the Dutch market. And also for the Dutch market there is evidence for the value premium based on the dividend yield, which is
important for this paper. In this paper there will be examined whether there is statistically evidence the Dogs of the Dow strategy has outperformed the AEX for the period 1990-2013. So there will be investigated if the DoD outperforms in another market instead of the DJIA and also for a more recent period than previous papers have studied. There will be looked at the monthly returns of the ten DoD and the AEX and after risk adjustments, fair conclusions could be made about the strategy. Also will there be investigated whether the returns
contributes to the winner-loser effect determined by De Bondt and Thaler (1985).
Beside the studies on the DJIA, there are several researches done about the Dogs of the Dow strategy applied on other markets than the Dow Jones. Latin America, England,
Canada and Finland are examples of countries where the strategy is tested. For so far there’s no research done on the Dutch stock market, the AEX. So this paper examines for the first time the performance of the DoD strategy applied on the AEX. In the next section the literature of the DoD theory in the US and other countries will be discussed. In section 3 the data and methodology will be described and in section 4 the results of the DoD theory applied on the Dutch market will be showed and analyzed. At last, in section 5 the paper ends with the conclusion.
2. Literature review
Analyst John Slatter came as first up with the DoD theory in an article written by Dorfman in the Wall Street Journal. He discovered that this theory outperforms the market between 1972 and 1987 by 7.6% per year on average. The book ‘Beating the Dow’ written by O’Higgins and Downes (1991) made this strategy popular under investors. They reported the strategy also outperformed the market on average from about the same period as Dorfman reported, namely from 1973 till 1991 by a return of 6.2% per year. Knowles and Petty (1992) found evidence the strategy is even working for a longer period, namely between 1957 and 1990, but by a lower outperformance of the market of 3.8%. But there has to be made a note by these studies. There is no risk adjusted and no transaction costs and tax payments are taken in. The research of McQueen, Shields and Thorley (1997) does however takes all of this in account and it concludes after these adjustments, the strategy only outperforms the market in 20 years out of the 50 years they have investigated it. Furthermore they could not say whether the 20 years of outperformance just a result of luck was. The risk McQueen, Shields and Thorley (1997) mentioned is the higher standard deviation caused by less diversification.
However Hirschey (2000) has concluded that over the period 1961-1998 the strategy did not significant outperformed. For a couple of years it did, but not for the whole period. He said ‘investors have enthusiastically embraced a plausible but ineffective investment
philosophy.’ People wanted the strategy would be working and they wanted the market to be inefficient, so they believed it. Hirschey (2000) suggested actually previous results are caused by data snooping. But Prather and Webb (2002) concluded in their research about the DoD this could not be proved and contributes it to the anomalies, like overreaction. With that saying, Prather and Webb (2002) cite to a conclusion of Fama (1998). Fama (1998)
are chance results, apparent overreaction of stock prices to information is about as common as under reaction. And post-event continuation of pre-event abnormal returns is about as
frequent as post-event reversal.’ This is based on the findings of De Bondt and Thaler (1985) about market overreaction. They conclude portfolios of former "losers" eventually outperform prior "winners", because people are inclined to overreact negative results. That means stock prices are lower than they should be. So after recovering of the stock, the returns are positive. But in addition to that, the price correction for the overreacted downturn makes the returns even more higher. This is called the winner-loser effect discovered by De Bondt and Thaler (1985). According to the DoD theory, it implies stocks with high dividend yields are prior losers, so their returns will raise more than the winners (low dividend yield stocks).
Beside studies about the DoD in the US, there is also done some research outside the US. Fillbeck and Visscher (1997) are the first one who investigated the results of the DoD strategy outside the US, namely in the UK. They have found the strategy was between 1984 and 1994 not very effective. It outperforms the FT-SE 100 only in 4 years. A reason the strategy worked not that well, is that the FT-SE 100 consists out of more companies and also more financial companies instead of typically American industrial companies. But the most important explanation comes from the fact the FT-SE 100 is a value weighted index, while the DJIA is a price weighted index. This means high dividend yield stocks from the FT-SE 100, which are undervalued stocks with a higher market value than average, have also a big
influence on the index and not only on the portfolio consisting out of dogs of the UK. And for that reason there could be not much difference between the portfolio and the index, according to Filbeck and Visscher (1997). This explanation is important, because the AEX is also value weighted instead of price weighted. So this could also be relevant if there is none statistically evidence the DoD strategy outperforms the AEX.
Da Silva (2001) has done research whether the DoD strategy holds for six Latin American countries. He concluded in five of the six countries the strategy outperforms the market, namely Argentina, Chili, Colombia, Mexico, Peru and Venezuela. Only the dogs of Brazil did not defeat the market. But his conclusions are not significant, because the period he investigated the strategy was between 1994 and 1999, which is a very short period.
The previous studies concluded there is not significant evidence the DoD strategy is working outside the US, but the following studies show there is. The first one is the study of Visscher and Fillbeck (2003). They have applied the DoD method on the Canadian stock market and looked if it has outperformed the Toronto-35 and TSE-300 between from 1988 till 1997. They conclude, even after risk adjustments, the average annual returns are high enough
to beat the market significantly. And with the excess return of 6.6%, it is also high enough to survive taxes and transaction costs.
Rinne and Vahamaa (2011) investigated whether the DoD theory also works on the Finnish stock market for the period 1988-2008. They conclude it did with an average annual abnormal return of 4.5%. After risk adjustment, the DoD still outperforms the market. But they are not sure if this return still holds after transaction costs and taxes. They also examined if the winner-loser effect contributes to the outperformance of the DoD theory applied on the Finnish market, and they conclude this could be the case. At last, Rinne and Vahamaa (2011) conclude the highest returns are achieved when the stock market is declining.
For so far there exists no earlier research about the DoD strategy applied on the AEX. And the findings on other markets outside the US are mixed. So there is no consensus
between earlier results. But according to the theory, there are some factors that could explain the results, positively and negatively. Transaction costs and taxes are higher for a DoD portfolio, the portfolio is less diversified, the AEX is value weighted, market overreaction could cause a winner-loser effect according to De Bondt and Thaler (1985) and the value premium, which is according to Capaul, Rowley and Sharpe (1993) also present in the Dutch market, could explain higher returns for the Dogs of the Dow strategy applied on the AEX.
3. Data and methodology
3.1 Data selection
The Amsterdam Exchange Index (AEX) exists out of 25 companies with the highest market value traded on the Amsterdam stock exchange. The index is weighted based on market value, while the DJIA is a price weighted index. For the AEX this means, fluctuations in the stock prices from big companies have a greater influence on the index than fluctuations of smaller firms. But the weight of a single company is limited to 15%. The AEX is founded in 1983, but since half way 1989 the AEX consists for the first time out of 25 companies. My purpose is to test the investment strategy for 24 years for the period December 1989 till December 2013, so the index has time to stabilize after the last added company half way 1989. That means the last trading day of the year is each year a rebalancing day. The 24 year sample period should be great enough to test for significance and this period covers also a period of booms and bursts of the economy. So there could be noticed if the strategy for example only outperforms in booms.
Datastream is used to provide the monthly prices of the AEX from December 1989 till December 2013. With these numbers the monthly returns are calculated and also the
compounded annual returns. http://www AEX.nl gives a historical recap about the
composition of the index since the beginning of the AEX in 1983. So that information is used to determine which companies are selected in the index at the last trading day of each year since 1989. There are over the 24 year sample period 64 candidates to become member of the DoD portfolio.1 With that list there will be searched on the last trading day of 1989 for the ten stocks in the AEX with the highest dividend yield. So it is clear which stocks have to be selected for the portfolio. This will be done by taking the dividends per share paid out over the year and divide them by the current stock price. Invest an equal amount in these stocks on the last trading day of 1989. Then, on the last trading day of 1990 the dogs of the AEX are calculated again and the portfolio will be rebalanced that day. So if necessary stocks will be replaced and in each dog there will be an equal amount invested again. This routine will be repeated till 2013. Till 1999 the invested amount for each dog will be 1000 guilder and after the introduction of the euro, this amount will be 1000 euro.
Table 1 shows for the period 1990-2013 the selected firms of the AEX. From the 64 candidates 47 of them were present in the DoD portfolio. 14 were only one year present during the 24 year sample period. 9 firms were at least 10 times selected for the portfolio. The most ‘experienced’ dogs are ABN AMRO, DSM, ING and Shell with respectively 17, 18, 17 and 19 appearances. The big absentee is Heineken. It never appears on the list, despite the constant dividend payments and the seize of the firm. During the years the portfolio composition is quite steady, because the average turnover rate is 3.13.
When the dogs are selected, the end of the month stock prices and dividends of each dog are required to derive the monthly and annual returns. These monthly returns are also provided by Datastream. The monthly returns consist out of the stock price changes and also the dividend pay outs. The dividends received were invested in the same stock that paid it out. When all the monthly returns of the dogs are collected, the monthly returns of the portfolio could be derived. This is simply done by adding the returns of the 10 dogs together and divide the total by 10, because every return replicates one tenth of the portfolio. After the collection of the monthly portfolio returns, the yearly compounded returns could be calculated. When a firm is removed from the AEX, it will also be removed out of the portfolio and replaced by an investment in the risk free rate.
1
The actual number of different company names is higher, because of name changes and mergers.
The monthly risk free return is measured by using the twelfth root of the 3 month Euribor rate, as recommended by Datastream and logically also provided by Datastream. The Dutch Treasury Certificate (DTC) is the Dutch Treasury bill, but this exists only since 1993. Also for that reason the recommendation by Datastream will be followed by using the 3 month Euribor rate as the monthly risk free return.
3.2 Methodology
After collecting these data, it’s time to test the theory. The returns of the portfolio will be compared with the index return. The mean difference and the standard deviation of the portfolio with the AEX will be calculated on a monthly basis. With a paired difference t-test there will be looked if there is statistically evidence for an outperformance. The abnormal yearly returns will be calculated with two methods. The first one will be calculated by simply subtracting the index return from the portfolio return to derive the market-adjusted return. The formula for that is as follows:
𝑅𝑅𝑀𝑀𝑀𝑀 = 𝑅𝑅𝐷𝐷𝐷𝐷𝐷𝐷− 𝑅𝑅𝑀𝑀𝐴𝐴𝐴𝐴
The second abnormal return is the Modigliani-squared or M² return, which is a risk-adjusted return based on the Sharpe ratio and described by Modigliani and Modigliani (1997). It measures the risk-adjusted return relatively to the AEX index and the formula is as follows:
𝑅𝑅𝑀𝑀2 = �𝑅𝑅𝐷𝐷𝐷𝐷𝐷𝐷− 𝑅𝑅𝑓𝑓� �𝜎𝜎𝜎𝜎𝜎𝜎𝜎𝜎
𝜎𝜎𝜎𝜎𝜎𝜎𝜎𝜎� − (𝑅𝑅𝑀𝑀− 𝑅𝑅𝑓𝑓)
𝜎𝜎𝜎𝜎𝜎𝜎𝜎𝜎 and 𝜎𝜎𝜎𝜎𝜎𝜎𝜎𝜎 are the standard deviations or volatility indicators for the AEX index and the DoD strategy respectively. The symbol 𝑅𝑅𝑓𝑓 stands for the risk-free interest rate, which the Euribor rate is in this case.
After calculating the two different abnormal returns there will be looked if there’s any statistical evidence to conclude if the Dogs of the Dow strategy outperform the market
consistent. For that reason the abnormal returns are not only calculated on a yearly base, but also on 5-year periods,10-year periods and for the whole 24 year sample period to see if the strategy is more effective on the long term than on yearly basis.
Table 1. List of the ‘Dogs’ of the AEX for the period 1990-2013 Firm/Year 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 Total ABN x 1 ABN AMRO x x x x x x x x x x x x x x x x x 17 AEGON x x x x x x x x x x x x 12 Ahold x x 2 AIR France x x x x 4 AKZO x x x x x x x x x x x x 12 Amro x 1 Aperam x x 2 Arcelor Mittal x 1 BAM Groep x x 2 Boskalis x x 2 Bührmann-Tetterode x 1 Corio x x x x x 5 Corporate Express x x x x x x x x x 9 Corus x x x x x x x x x 9 DAF x x x 3 DSM x x x x x x x x x x x x x x x x x x 18 Fokker x 1 Fortis x x x x x x x x x x x x 12 Gist Brocade x x x x x x x x x 9 Gucci x 1 Hagemeyer x x x x x 5 ING x x x x x x x x x x x x x x x x x 17 KLM x x x x 4 Koninklijke BolsWessanen x x x x 4 KPN x x x x x x x x x x x 11 LogicaCMG x x 2 Nationale Nederlanden x 1 NMB x x 2 Numico x x 2 OCÉ x 1 Pakhoed x 1 Philips x 1 PostNL x 1 Randstad x x 2 RD Shell-A x x x x x x x x x x x x x x x x x x x 19 Reed Elsevier x x x x x x x x x x 10 SBM Offshore x x 2 Stork x 1 Unibail-Rodamco x x x x x x 6 Unilever x x x x x x x 7 USG People x 1 Van Ommeren x x x 3 Vd Moolen x x x 3 Vendex KBB x 1 Wereldhave x x x 3 Wolters Kluwer x x x x x x 6
Firms staying on the list 5 7 7 6 7 6 9 8 8 8 6 7 7 6 8 8 8 9 4 5 8 6 5 Turnover(average = 3.13) 5 3 3 4 3 4 1 2 2 2 4 3 3 4 2 2 2 1 6 5 2 4 5 47 firms total
The next step is to inspect if the returns of the portfolio are rewarded for any risk. This will be done by calculating the Sharpe ratio and the Treynor ratio. The Sharpe ratio is
calculated by deriving as first the mean of the excess returns. These are derived by subtracting the risk-free interest rate or in this case the Euribor rate from the portfolio returns. Derive the mean of these excess returns and then divide these excess returns by the standard deviation of the excess returns. In formula form this will looks like this:
𝑆𝑆 =𝑅𝑅�������������𝜎𝜎𝐷𝐷𝐷𝐷𝐷𝐷− 𝑅𝑅𝑓𝑓
𝑅𝑅𝐷𝐷𝐷𝐷𝐷𝐷−𝑅𝑅𝑓𝑓
The Sharpe ratio measures how much the excess returns compensate the unsystematic risk, which is risk that could be diversified away. In other words, the Sharpe ratio measures the return obtained per unit of total risk. The Sharpe ratio for the AEX is measured by replacing the returns of the DoD for AEX returns.
The Treynor ratio measures how much the excess returns compensate the systematic risk, which is risk that could not be diversified away. Therefore it is calculated by dividing the mean excess return by the beta of the portfolio. The formula is:
𝑇𝑇 = 𝑅𝑅�������������𝐷𝐷𝐷𝐷𝐷𝐷𝛽𝛽− 𝑅𝑅𝑓𝑓
The Treynor ratio for the AEX is calculated by replacing the returns of the DoD for the AEX and divide them by the beta. The beta for the market itself is 1, so the Treynor ratio for the AEX equals the excess returns.
4. Results
4.1 Abnormal returns
The average annual returns for the ‘dogs’, the market and abnormal returns are presented in table 2. The DoD mean return is 8.15%, but the market has a mean return of 7.08%. So the difference with the market or the market-adjusted return is just 1.07%. The median return for the DoD is higher than the mean return, namely 11.22%, while the median return for the AEX is lower than her mean return. The fact that for the DoD the median is higher than the mean
indicates the lower returns have a greater influence on the mean than the higher returns. For the AEX is thus the case. The standard deviation or volatility of the market (24.14%) is higher than the ‘dogs’ standard deviation (22.52%). The minimum return for the DoD portfolio is higher than the minimum for the index, while the maximums are roughly the same for the two returns. These are all indicators the DoD may have outperformed the AEX, only the number of positive years is higher for the AEX index.
Table 2. Annual DoD strategy details.
Figure 1. Cumulative returns for the ‘dogs of the Dam’ and the AEX index.
The cumulative returns for the DoD strategy and the AEX are displayed above in figure 1. In the first two years the DoD performed better, but from 1993 till 1998 the results were quite the same. Although the lines have the same patterns, the AEX performed better after 1998 a couple of years till 2001. Because of the crisis in 2000 both returns declined. The DoD returns dropped much more in 2003, but a year later the returns were about the same and both increased till the next crisis. In 2008 the crisis caused in both returns a great fall, but the
0% 25% 50% 75% 100% 125% 150% 175% 200% 225% 250% 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 DoD AEX DoD return AEX return MA-return M²-return RF- return RF-adjusted Mean 8.15 7.08 1.07 3.00 3.82 4.33 Standard deviation 22.52 24.14 10.08 11.73 2.56 23.13 Median 11.22 5.18 0.60 3.51 3.33 9.63 Minimum -40.65 -52.32 -20.72 -20.05 .21 -43.39 Maximum 44.38 44.93 19.67 23.69 9.50 40.32 Positive years 16 18 12 14 24 14
decline for the AEX was bigger. After that fall they both recovered and increased again with about the same pattern, but the DoD performed significantly better. Over the whole period the DoD outperformed with a cumulative return of 192.94%, while the AEX has a return of 154.76%.
The average annual market-adjusted return is positive with 1.07%. But is that
difference positive all of the years? Table 3 shows the returns and market-adjusted returns per year and also for multiple periods. If we look whether the strategy outperforms per year, we see there is not a consistent outperformance of the DoD. The portfolio is 12 years better, the AEX 11 years and one year the returns are roughly the same. To investigate if there is any significant difference between the portfolio and the market, there is done a paired difference t-test. The p-values are also presented in table 3. They show that only in 1990 and 1998 the differences were significant2. But the sample size, 12 monthly returns, is not that great. For that reason we take a look at the multiple periods differences. The DoD outperforms in 3 out of the 5 5-year periods3, 1 out of 2 10-year periods and also for the whole sample period. But these results are according to their high p-values also not significantly different.
Rinne and Vahamaa (2011) suggested in their paper the Finnish dogs performed better than the market during market downturns. There is not any evidence to believe this may be the case on the Dutch market, because of the absence of any statistical proof. But in 6 years when the market dropped, the DoD performed better or less worse. The same thing holds for the multiple periods. For the periods 2000-2004 and 2005-2009 the market returns were negative and the portfolio returns better. Reasonable, this holds also for the period 2000-2009.
The risk-adjusted abnormal return or the M² return has an annual mean return of 3%, which is showed in table 2. This return is higher than the market-adjusted return, which could indicate the returns of the DoD include a risk premium. This is also suggested by the fact the annual average risk-free rate is not higher than the DoD returns. But risk is also included in the AEX index, because these returns are also higher than the risk-free rate. The M² returns are 14 year positive, which is 2 year more than the market-adjusted return. But if we look at table 4 at the M² return for the whole period, the return is just 3.56%. This indicates there is just a very small risk premium.
2
Investigated at a 5% significance level.
3 The last period is actually a 4-year period from 2010 till 2013.
Table 3. DoD strategy compounded returns compared with the AEX index for 1990-2013. Year DoD return AEX return MA-return* Winner Mean difference of portfolio versus AEX** Standard deviation difference of portfolio versus AEX p-value paired t-test 1990 -15.98 -23.85 7.87 DoD .78 1.20 0.0457 1991 30.92 20.87 10.05 DoD .70 1.91 0.2281 1992 -7.47 3.17 -10.64 AEX -.83 2.37 0.2492 1993 37.59 44.93 -7.34 AEX -.44 2.48 0.5509 1994 -1.61 1.06 -2.67 AEX -.16 1.67 0.7430 1995 17.87 1.05 16.82 DoD .06 .06 0.8986 1996 22.65 33.56 -10.91 AEX -.71 1.89 0.2216 1997 26.59 40.95 -14.36 AEX -.84 1.87 0.1461 1998 9.13 29.85 -20.72 AEX -1.53 2.40 0.0494 1999 44.38 24.71 19.67 DoD 1.32 3.10 0.1679 2000 4.06 -5.04 9.10 DoD .79 3.57 0.4618 2001 -12.29 -20.52 8.23 DoD .82 2.54 0.2849 2002 -38.38 -36.32 -2.06 AEX -.04 2.63 0.9616 2003 -2.39 4.62 -7.01 AEX -.10 7.38 0.9643 2004 13.11 3.09 10.02 DoD .78 1.24 0.0527 2005 25.24 25.48 -0.24 Tie -.02 1.21 0.9464 2006 14.86 13.41 1.45 DoD .12 1.04 0.6928 2007 0.02 4.12 -4.10 AEX -.34 1.36 0.4027 2008 -40.65 -52.32 11.67 DoD 1.82 3.58 0.1059 2009 41.79 36.35 5.44 DoD .56 4.46 0.6731 2010 9.34 5.74 3.60 DoD .32 2.31 0.6397 2011 -16.40 -11.87 -4.53 AEX -.50 2.18 0.4453 2012 18.22 9.68 8.54 DoD .74 3.60 0.4933 2013 14.96 17.24 -2.28 AEX -.15 1.56 0.7485 Multiple years 1990-94 37.78 37.70 0.08 Tie .01 2.02 0.9686 1995-99 188.35 256.98 -68.63 AEX -.34 2.37 0.2686 2000-04 -37.91 -48.16 10.25 DoD .45 3.94 0.3782 2005-09 21.08 -3.66 24.74 DoD .43 2.74 0.2320 2010-13 24.24 19.82 4.42 DoD .10 2.49 0.7762 1990-99 306.76 391.55 -84.79 AEX -.17 2.20 0.4110 2000-09 24.82 -50.06 74.88 DoD .44 3.37 0.1570 1990-13 205.80 194.16 11.64 DoD .13 2.80 0.4270 Summary
Category DoD wins AEX wins Tie
Per year 12 11 1
Multiple years 5 2 1
*MA-return stands for market-adjusted return, which is the DoD return minus AEX return. **Differences are calculated based on the monthly returns.
Table 4. DoD strategy M²-adjusted returns and the risk-free adjusted returns for 1990-2013. Year M²-return* RF-rate RF-adjusted return** 1990 6.37 8.43 24.41 1991 23.69 9.19 21.73 1992 -5.61 9.50 -16.97 1993 -8.06 6.70 30.89 1994 -6.96 5.01 -6.62 1995 16.51 4.25 13.62 1996 -13.49 2.95 19.70 1997 16.65 3.20 23.39 1998 -20.05 3.54 5.59 1999 9.11 2.63 41.75 2000 -0.31 4.48 -0.42 2001 9.13 4.47 -16.76 2002 4.55 3.47 -41.84 2003 -8.02 2.14 -4.53 2004 9.66 2.12 10.99 2005 1.17 2.12 23.12 2006 10.26 2.97 11.89 2007 -4.30 4.15 -4.13 2008 16.16 4.96 -45.61 2009 22.65 1.24 40.54 2010 2.47 0.73 8.61 2011 -13.12 1.50 -17.90 2012 7.70 0.66 17.56 2013 -4.08 0.21 14.75 Multiple years 1990-94 0.42 45.26 -6.62 1995-99 -80.53 17.70 151.17 2000-04 22.85 17.78 -48.92 2005-09 24.25 16.37 0.96 2010-13 1.90 3.13 20.29 1990-99 -97.43 70.97 134.53 2000-09 -19.69 4.48 -48.42 1990-13 3.56 141.68 45.51
*M²-return stands for the Modigliani-squared adjusted return. **The RF-adjusted return is the DoD return minus the risk-free rate.
To summarize the findings so far, the DoD outperforms the market with a market-adjusted return of 1.07%. But this result is not statistically significant. 24 annual returns is a small sample size, but also for the monthly returns for the 24 year sample period there is not significant evidence for any difference. Also the second abnormal return, the M² return is
positive with an annual average of 3%. But if we look at the period as a whole, the return is just 3.56%. This is for such a large period a small risk premium. Therefore the risk-adjusted return of the DoD does not outperform the market.
Capaul, Rowley and Sharpe (1993) found evidence for some value premium in the Dutch market, but the findings in the results above do not support that statement. The market-adjusted return of 1.07% is namely not significant.
4.2 Risk-adjusted returns
Table 5 shows the Sharpe ratio’s for the DoD strategy and for the AEX index per year and also for multiple periods. Also the values used to derive the Sharpe ratio are included, namely the mean differences of the excess returns and the standard deviations of these excess returns. If we compare the ratio’s per year there is no winner, because the DoD and AEX did it both 12 years better than the other one. For the multiple periods the portfolio did it better, but differences are not that great to call the DoD the big winner. For the yearly results both have in the same years a negative Sharpe ratio, which means it was better to put money in a risk-less investment. Almost the same holds for the multiple periods, except for the period 2000 till 2005. In this period the AEX has a negative ratio and the portfolio a positive ratio.
Looking at the mean differences and the standard deviations, we see that the numbers do not much differ from each other, which explains roughly the same Sharpe ratios. There is also not a period where the AEX or the DOD beat each other. In other words, it is just luck when one of them outperforms the other one. So based on the Sharpe ratio, the DOD does not outperforms the AEX.
Also the Treynor ratio is calculated. Table 6 shows the ratio’s for the DOD strategy and the Amsterdam Exchange Index for the sample period per year and again for the multiple periods. Also the beta, the measurement for the systematic risk and the denominator for the Treynor ratio formula, is provided. The mean excess returns, the numerator in the formula, are presented in table 5. The yearly Treynor ratios do not differ that much with each other, just like the Sharpe ratio. The DoD outperformed the AEX 13 years and the AEX outperformed the DoD 11 years. But again the numbers are that close to each other, so no big winner could be suggested. The same reason holds for the multiple periods. While the DoD outperformed in more multiple periods the market, but no winner could be noted. This is again caused by the fact the ratio’s are so close to each other and also the ratio’s for the AEX and the ‘dogs’ are
too small to suggest any outperformance. As a result there could be concluded that the Dogs of the Dow theory does not dominate the Dutch market based on the Treynor ratio.
Table 5. Sharpe ratio’s for the DoD strategy and the AEX index for the period 1990-2013.
Sharpe ratio Mean difference of*
Standard deviation difference of*
Year DoD AEX Winner DoD-RF AEX-RF DoD-RF AEX-RF
1990 -0.359 -0.470 DoD -1.99 -2.78 5.56 5.90 1991 0.359 0.241 DoD 1.61 0.91 4.50 3.78 1992 -0.223 -0.105 AEX -1.25 -0.41 5.59 3.94 1993 0.498 0.611 AEX 2.24 2.68 4.50 4.38 1994 -0.154 -0.087 AEX -0.51 -0.35 3.31 3.99 1995 0.322 0.312 DoD 1.08 1.02 3.35 3.27 1996 0.390 0.654 AEX 1.54 2.25 3.96 3.44 1997 0.252 0.413 AEX 2.00 2.84 7.93 6.89 1998 0.096 0.277 AEX 0.69 2.22 7.17 8.03 1999 0.496 0.375 DoD 3.03 1.70 6.10 4.55 2000 0.017 -0.160 DoD 0.08 -0.70 4.91 4.41 2001 -0.204 -0.357 DoD -1.25 -2.08 6.16 5.82 2002 -0.285 -0.335 DoD -3.49 -3.45 12.26 10.32 2003 0.027 0.059 AEX 0.34 0.44 12.59 7.48 2004 0.295 0.041 DoD 0.90 0.12 3.05 2.95 2005 0.505 0.482 DoD 1.77 1.79 3.50 3.71 2006 0.286 0.293 AEX 0.96 0.83 3.34 2.85 2007 -0.095 0.011 AEX -0.30 0.04 3.20 3.35 2008 -0.413 -0.660 DoD -4.14 -5.95 10.01 9.02 2009 0.338 0.404 AEX 3.29 2.73 9.72 6.75 2010 0.143 0.102 DoD 0.83 0.51 5.83 5.06 2011 -0.446 -0.204 AEX -1.54 -1.04 3.46 5.11 2012 0.255 0.217 DoD 1.53 0.79 5.99 3.65 2013 0.302 0.386 AEX 1.22 1.37 4.05 3.56 Multiple years 1990-94 0.004 0.002 DoD 0.02 0.01 4.85 4.68 1995-99 0.286 0.371 AEX 1.67 2.01 5.82 5.42 2000-04 -0.080 -0.171 DoD -0.68 -1.14 8.57 6.63 2005-09 0.045 -0.018 DoD 0.31 -0.11 7.00 6.27 2010-13 0.103 0.094 DoD 0.51 0.41 4.95 4.36 1990-99 0.156 0.197 AEX 0.84 1.01 5.42 5.14 2000-09 -0.024 -0.097 DoD -0.18 -0.62 7.80 6.45 1990-13 0.056 0.041 DoD 0.36 0.23 6.45 5.64 Summary
Category DoD wins AEX wins Tie
Per year 12 12 0
Multiple years 6 2 0
Table 6. Treynor ratio’s for the DoD strategy and the AEX index for the period 1990-2013.
Treynor ratio
Year DoD AEX Winner DoD beta*
1990 -0.022 -0.028 DoD 0.923 1991 0.015 0.009 DoD 1.080 1992 -0.009 -0.004 AEX 1.330 1993 0.026 0.027 AEX 0.863 1994 -0.007 -0.003 AEX 0.756 1995 0.012 0.010 DoD 0.910 1996 0.015 0.022 AEX 1.011 1997 0.018 0.028 AEX 1.124 1998 0.008 0.022 AEX 0.854 1999 0.026 0.017 DoD 1.165 2000 0.001 -0.007 DoD 0.794 2001 -0.013 -0.021 DoD 0.963 2002 -0.030 -0.035 DoD 1.173 2003 0.002 0.004 AEX 1.431 2004 0.010 0.001 DoD 0.946 2005 0.020 0.018 DoD 0.891 2006 0.009 0.008 DoD 1.121 2007 -0.003 0.000 AEX 0.873 2008 -0.040 -0.060 DoD 1.036 2009 0.025 0.027 AEX 1.318 2010 0.008 0.005 DoD 1.059 2011 -0.024 -0.010 AEX 0.637 2012 0.011 0.008 DoD 1.363 2013 0.012 0.014 AEX 1.052 Multiple years 1990-94 0.000 0.000 Tie 0.954 1995-99 0.017 0.020 AEX 0.982 2000-04 -0.006 -0.011 DoD 1.159 2005-09 0.003 -0.001 DoD 1.028 2010-13 0.005 0.004 DoD 0.981 1990-99 0.009 0.010 AEX 0.965 2000-09 -0.002 -0.006 DoD 1.097 1990-13 0.003 0.002 DoD 1.031 Summary
Category DoD wins AEX wins Tie
Per year 13 11 0
Multiple years 5 2 1
The beta’s per year presented in table 6 range from 0.637 to 1.431. The mean annual beta is 1.028, which is close to 1. This means the systematic risk or the not-diversifiable risk for the DoD is quite the same as for the market index. The beta’s for the multiple periods are also around 1. The beta for the whole period is 1.031, which is close to the average of 1.028. So also for the multiple periods holds the conclusion that the systematic risk for the portfolio is quite the same as for the AEX index.
There can be conclude there exists no risk premium for the DoD strategy on the Dutch market, despite the portfolio is not diversified and thus should be exposed to more risk. This could be explained by the fact that the AEX index is value weighted. This means big firms, which are most of the time also the firms with the highest dividend yields, have a great influence on the movements of the price of the index. This reason is also supported by the mean annual beta, which is close to 1 and therefore says that the portfolio is highly correlated with the market. It is logical that they are highly correlated, but in this case they are too strongly correlated. Also the standard deviations showed in table 5 are close to each other, which means there is also no extra unsystematic risk involved. So there is no risk premium, because there is for the DoD no extra risk taken compared to the market.
4.3 Further adjustments
For so far there has been made adjustments for the risk-free rate, for the market return and for risk. But since the strategy implies that the portfolio every year needs to be rebalanced, also transaction costs have to be taken into account. In table 1 there was showed that the yearly average turnover rate for the stocks was 3.19. The turnover rate Filbeck and Visscher (1997) found for the dogs in England was on average 5. The average turnover rate Rinne and Vahamaa (2011) for the Finnish dogs was 4.9 and Visscher and Filbeck (2003) found for the Canadian dogs the turnover rate was 2.5. The last example is the research of McQueen, Shields and Thorley (1997), who find for the American dogs an average turnover rate of 2.96. So the Dutch turnover rate is quite average. Rinne and Vahamaa (2011) and McQueen, Shields and Thorley (1997) discussed the transaction costs decrease the returns with about 0.5%. If we follow that calculation the market-adjusted return is roughly 0.57%. But this has to be taken with reservation. Because when the portfolio is for example managed by a big investment firm, the transaction costs will be much lower than when the portfolio is managed by a single person.
In addition to the transaction costs also tax payments over the returns need to be subtracted from the returns to see if any money is left over. The dividend tax rate is 15% in
the Netherlands and is hold back by the firm who pays out the dividend. But if the profits of dividend payouts are marked as profits from saving and investing the taxes could be returned by request. So for that reason no dividend taxes are taken into account. Therefore the only tax paid is the ‘Vermogensrendementsheffing.’ This means 1.2% taxes are paid over the whole equity and this implies that the dividends received are also taxed with 1.2%, because the dividends are for the DoD strategy reinvested in the stocks. So if we multiply 1.2% with the 8.15% average annual return of the DoD strategy, we obtain an average yearly tax payment loss of 0.096%. For the AEX this means an average loss of 0.085%. If we derive the market-adjusted return after the tax-adjustments, we obtain an average yearly return of 1.06%.
Rinne and Vahamaa (2011) have investigated whether the Finnish dogs outperformed the market, because of the winner-loser effect described by De Bondt and Thaler (1985). They concluded this may be the case. For the Dutch market this further research is not that
necessary, because of the low market-adjusted return and the lack of statistical power of it. Basically the conclusion is that the DoD strategy did not outperform the AEX. More
accurately, the returns for the portfolio and the market index do not significantly differ from each other. Maybe this is caused by the fact the Dutch market does not overreact that much, however there is no need to test for a winner-loser effect.
5. Conclusions
In this paper there has been investigated whether there is statistically evidence the Dogs of the Dow strategy has outperformed the AEX index for the period 1990 till 2013. For each year of the sample period there had to be calculated which ten stocks of the AEX do have the highest dividend yield to apply the DoD strategy on the Dutch market. These were all called the dogs of the Dam and were for one year part of the portfolio. In each stock there was an equal amount invested. After one year the ten highest dividend yield stocks were derived again and the portfolio was rebalanced to maintain the composition of the ten highest dividend yield stocks.
The results for this strategy applied on the AEX suggest there is no evidence of any outperformance on a yearly base or for a longer period. So it is not more lucrative to invest in the dogs of the Dam instead of the AEX index itself. The cumulative return for the whole period is 192.94%, while the AEX has a return of 154.76%. Further, the mean annual market-adjusted return is 1.07%, but this result is not statistical significant. Other returns calculated over a larger period were also not significantly different than the market. Even if it were
significant, the outperformance is that small it will hardly survive taxes or transactions costs. As a result there is no evidence for a value premium in this return. There was also not any evidence of a risk premium, because no extra risk was taken by following this strategy compared with the market. Unless the portfolio was not well diversified, the volatility of the DoD did not differ from the market. This could be explained by the fact the AEX index is value weighted, which is supported by the result that the average beta of the portfolio is roughly 1. At last, it was not necessary to derive any winner-loser effect, because there is already concluded the dogs of the Dam did not significantly differ from the AEX index.
6. Reference list
De Bondt, W., & Thaler, R. (1985). Does the stock market overreact? Journal of
Finance, 40(3), pp. 793-808. doi:10.2307/2327804
Capaul, C., Rowley, I., & Sharpe, W.F. (1993) International Value and Growth Stock Returns. Financial Analysts Journal, 49(1), pp. 27-36. Retrieved from
http://www.jstor.org.proxy.uba.uva.nl:2048/stable/4479610
Dorfman, J.R. (1988). Study of industrial average’s finds stocks with high dividends are big winners. The Wall Street Journal, August 11.
Fama, E. F. (1998). Market efficiency, long-term returns, and behavioral finance. Journal of
Financial Economics, 49(3), pp. 283-306. doi:10.1016/S0304-405X(98)00026-9
Fama, E. F., & French, K. R. (1995). Size and book-to-market factors in earnings and returns.
Journal of Finance, 50(1), pp. 131-155. doi:10.2307/2329241
Fama, E. F., & French, K. R. (1998), Value versus Growth: The International Evidence. The
Journal of Finance, 53(1), pp. 1975–1999. doi:10.1111/0022-1082.00080
Filbeck, G., & Visscher, S. (1997). Dividend yield strategies in the British stock market.
European Journal of Finance 3(4), pp. 277–89. doi:10.1080/135184797337372
Hirschey, M. (2000). The “Dogs of the Dow” Myth. Financial Review, 35(1), pp. 1–16. doi: 10.1111/j.1540-6288.2000.tb01411.x
Knowles, H. C., & Petty, D. H. (1992). The dividend investor: A safe, sure way to beat the
market. Chicago, IL: Probus Publishing Co.
beat the Dow statistically and economically? Financial Analysts Journal, 53(4), pp. 66–72.
Modigliani, F., & Modigliani, L. (1997). Risk-adjusted performance. Journal of Financial
Economics, 23(2), pp. 45-54. Retrieved from
http://bbs.cenet.org.cn/uploadImages/20035715461865606.pdf
O’Higgins, M., & Downes, J. (1991). Beating the Dow. New York, NY: HarperCollins Prather, L., & WEBB, G. (2002). Window Dressing, Data Mining, Or Data Errors: A Re –
Examination Of The Dogs Of The Dow Theory. Journal of Applied Business
Research, 18(2), pp. 115 -124. Retrieved from
http://cluteinstitute.com/ojs/index.php/JABR/article/view/2122/2099
Rinne, E., & Vahamaa, S. (2011). The ‘Dogs of the Dow’ strategy revisited: Finnish evidence.
The European Journal of Finance, 10(5-6), pp. 451-469.
doi:10.1080/1351847X.2010.544951
Da Silva, A. L. C. (2001). Empirical tests of the Dogs of the Dow strategy in Latin American stock markets. International Review of Financial Analysis, 10(2), pp. 187-199.
doi:10.1016/S1057-5219(01)00047-3
Thomson Reuters. (2014). Stock prices, Euribor rates, AEX index prices [Time series]. Retrieved from Datastream database
Visscher, S., & Filbeck, G. (2003). Dividend-Yield Strategies in the Canadian Stock Market.
Financial Analysts Journal, 59(1), pp. 99-106. Retrieved from
