Sunday 5 January 2014

The Dogs and Puppies of STI

Many academia and literature says that active investors cannot beat the index. However, if you believe that this is true but still want to beat the index, then it follows that the key to beating the index must lies within the index itself, i.e. the portfolio of stocks that beat the index must come from the index component stocks.

In my previous post, I mentioned that an equal-weighted portfolio of the Straits Times Index (STI) component stocks can beat the index easily. In this post, we will discuss whether a "Dogs of STI" investment strategy can beat the index and even outperform the equal-weighted portfolio of STI component stocks.

The Dogs of STI replicates the investment strategy of the Dogs of the Dow. In this strategy, the 10 highest yielding Dow Jones Industrial Average (DJIA) component stocks are selected for investment in the following year. This strategy has returned 10.8% per annum over the past 20 years, matching the returns of DJIA and beating the Standard and Poor's 500 index's return of 9.6%.

What is the rationale behind the Dogs of the Dow investment strategy? Proponents of the strategy argue that blue chip companies do not change their dividends regularly to reflect changes in business conditions whereas stock prices regularly fluctuate according to business conditions. Hence, a stock that is higher yield than average would most likely be near the bottom of the business cycle and the stock price is likely to recover more than that of lower-yielding stocks. It suggests that the management of the companies believes their stocks are oversold and are willing to back them up by paying a relatively high dividend. By investing in the Dogs of the Dow, investors gain both by the higher price appreciation and the higher dividend yield.

Besides the Dogs of the Dow, there is a related investment strategy known as the "Small Dogs of the Dow" or "Puppies of the Dow". In this strategy, the same 10 stocks are selected as in the Dogs of the Dow. From this list, the 5 lowest-priced stocks are selected as the Puppies of the Dow. This smaller portfolio outperforms even the Dogs of the Dow, returning 12.5% versus 10.8% for the Dogs of the Dow over a 20-year period. 

So, how would a "Dogs of STI" and "Puppies of STI" portfolio perform? The table below shows the Dogs and Puppies of STI over a 13-year period since 2000 (Please refer to the previous post on Who Says You Can't Beat the Index? for limitations of the stock price data). Stocks highlighted in orange are the Puppies of STI while stocks highlighted in yellow make up the rest of the Dogs of STI.

Dogs and Puppies of STI Selected at End of Year
 
Returns of Dogs and Puppies of STI in Following Year (W/o Dividends)

Returns of Dogs and Puppies of STI in Following Year (With Dividends)

As shown above, the Dogs of STI outperforms the STI and even the equal-weight portfolios of STI and STI without the heavy weights (i.e. DBS, OCBC, Singtel and UOB). The average return including dividends is 19.3%, higher than STI's 6.7%, equal-weighted STI's 14.1% and equal-weighted STI without the heavyweights of 15.4%. The Puppies of STI performed even better than the Dogs of STI, returning 21.0% over the same period.

The outperformance of both the Dogs and Puppies is not due to dividends alone. Without dividends, the price appreciation (12.7% for Dogs and 13.3% for Puppies) already outperformed that of the STI (3.9%), equal-weighted STI (9.7%) and equal-weight STI without the heavyweights (11.0%). The inclusion of dividends further increased the outperformance of the Dogs and Puppies. The table below shows the outperformance relative to the STI, equal-weighted STI and equal-weighted STI without the heavyweights with and without dividends included.


Dogs of STI Puppies of STI

No Dividends With Dividends No Dividends With Dividends
STI 8.8% 12.6% 9.4% 14.3%
Equal-weighted STI 3.0% 5.2% 3.6% 6.9%
Equal-weighted STI w/o Heavyweights 1.7% 3.8% 2.2% 5.5%

On a dollar basis, over a 13-year period ending in Dec 2013, a $10,000 investment in the Dogs of STI with dividends reinvested would become $99,158 while a $10,000 investment in the Puppies of STI would become $119,182. This compares very well with STI ($23,235), equal-weighted STI ($55,554) and equal-weighted STI without the heavyweights ($64,369). So, dogs (and puppies) are Man's best friends apply to investors as well :)


See related blog posts:

23 comments:

  1. your returns is based on equal lots each?

    ReplyDelete
    Replies
    1. Hi StevenC,

      It's based on equal amount of money in each of the stocks.

      Delete
    2. i dun quite get the calculation of overall returns from its component stocks?
      mind explaining how to get 6.6% returns in 2013 puppies as an example?

      Delete
    3. The 5 puppies identified at the end of Dec 2012 are CapitaMall, HPH, SingTel, ST Engg and Starhub. Their returns for 2013 inclusive of dividends are: -5.9%, -3.8%, 16.0%, 8.1% and 18.5% respectively. The average return is 6.6%.

      Delete
    4. i think u mistakenly highlighted sph instead of starhub for puppies..

      anw thank you, appreciate wad you have done here.. keep it up!

      Delete
    5. Thks for highlighting. Have updated the colour alignment. The issue was the Dogs and Puppies selected at the end of 2012 were highlighted but their returns for 2013 were tracked.

      Delete
  2. Are you assuming that you will buy the dogs and the puppies at the beginning of each year? Have you considered the transaction fees? And how are the shares allocated?

    ReplyDelete
    Replies
    1. Hi Eugene,

      Yes, that's the assumption. It's based on equal amount of money in each of the stocks. Transaction fee should be minimal since it is rebalanced only once a year.

      Delete
    2. Hi Chin Wai, is it possible to mail me the spreadsheet? Trying to reproduce your findings. Thanks.

      Delete
    3. Hi Eugene,

      I think you might have difficulty following my spreadsheet. Nevertheless, I've added another screenshot on the dividend yield to explain how the Dogs and Puppies were selected at the end of each year. Hope that helps :)

      Delete
  3. Hi Chin Wai,

    I just discovered your blog today and I think it's awesome! Thank you for sharing so much of your knowledge.

    My recent studies have also pointed out to me the inefficiencies of a market-cap weighted index, so I am very pleasantly surprised to see that the equal weighted STI performed so much better than traditional STI.

    Have you considered further backtesting even beyond 2001? It would be interesting to see how robust the Eq-W, Eq-W minus Heavyweights, Dogs and Puppies strategy in an even further lookback period.

    Finally, it seems like the performance for earlier years are only using the current STI components now (at least the equal weighted portion). Is it not more accurate to include the other components that were part of the STI back then as well to be part of the equal weighted portfolio? Since we will not know which of the current components may no longer be in the STI in the future.

    Sorry to dig up an old thread! I hope to hear your thoughts on this!

    Regards

    ReplyDelete
    Replies
    1. Hi Cystic,

      The data available from Yahoo! Finance is only up till Jan 2000, so I could only backtest till then.

      Yes, you are right to point out that the stock components are based on the current STI components. I couldn't find the STI components of previous years.

      Delete
    2. Hi Chin Wai,

      Thanks for taking time to respond, appreciate it!

      I tried searching around too, the best that I could find is the semi-annual review from 2008 until present, which documents the inclusions and exclusions from the STI.

      http://www.ftse.com/Indices/FTSE_ST_Index_Series/Index_Reviews.jsp

      Just in case you are curious. If I have the time, I will try to replicate your study to show the STI returns from 2008 until now, with changing constituents.

      Even though your study has that slight survivorship bias, I am hopeful that the results from my proposed test will show that the STI as an index has certain fundamental flaws!

      I look forward for more of your insightful posts!

      Delete
    3. Hi Cystic,

      I'm sure you'll find something useful from your study!

      Delete
  4. would you mind looking into the Bounceback portfolio in sg context?

    "The idea of the Bounceback Portfolio is that a portfolio of the 10 worst performing FTSE 350 stocks in one year has historically beaten the index in the first three months of the following year."

    ReplyDelete
    Replies
    1. Thanks for your suggestion. I'll review this strategy at the opportune time.

      Delete
    2. I've posted a review of the Bounceback Portfolio in Singapore's context. It's available at http://boringinvestor.blogspot.sg/2014/08/the-bounceback-portfolio-of-sti.html.

      Delete
  5. is it possible for you to explore the puppies selection based on market cap since sti is a marketcap-weighted index. was wondering if the premise for puppies of dow is based on a price-weight dow...

    ReplyDelete
    Replies
    1. The index weightage methodology only affects the computation of the index value. It does not affect the selection of the dogs and puppies, which is based on the yield and price.

      Delete
  6. Hi,

    I stumbled on your blog thinking about applying the "Dogs of Dow" strategy on the SGX exchange.

    What are your thoughts of including all dividend paying large caps instead of just limiting to the STI?

    ReplyDelete
    Replies
    1. Hi Azrael,

      Yes, the theory behind the Dogs of Dow strategy should be applicable to an "index" of all dividend paying large caps.

      Delete
  7. Hi Chin Wai, when you say rebalance, do you mean selling dogs and puppies from last year and buy new ones for this year?

    ReplyDelete
    Replies
    1. Hi Ken, yes, replace the dogs and puppies at the start of every year.

      Delete