Sunday 29 December 2013

Who Says You Can't Beat the Index?

Numerous academia and literature have impressed upon us that active investors like us cannot beat the index. This large no. of academia and literature cannot be wrong. Nevertheless, after reviewing the returns of the Straits Times Index (STI) over the past 26 years, I have an inkling that we might be able to beat the index after all. 

Over the past 26 years from Jan 1988 till Nov 2013, the returns of the Dow Jones Industrial Average, Standard & Poor's 500, Nasdaq Composite and Hang Seng Index have averaged (computed as geometric average) about 8.5%, 7.8%, 10.0% and 9.3% respectively. The returns of STI, in contrast, averaged about 5.0%, similar to FTSE's return of 5.2%. The only index among those studied that performed worse than STI is the Nikkei index, which returned -1.6%. At a return of 5.0%, it should not be difficult in beating the index. 

To understand the poor performance of STI, I collected the price of the 30 component stocks of STI (the prices are available only from 2000 onwards from Yahoo! Finance) and computed their individual performances. It turned out that on an equal-weighted basis (i.e. all component stocks have equal weights in the index), the return since 2000 is actually much higher, at 9.7% compared to 3.9% for the STI. See the table below for the computation. (Note that there are gaps in the prices due to corporate actions. For example, SPH, prior to Jun 2004, was priced around $20. After a 5-for-1 stock split, the price became around $4. The data from Yahoo! Finance only recorded the price after the stock split but not before. Nevertheless, it does not significantly affect the conclusion below.)

Returns of STI Component Stocks (Without Dividends)

The reason for the lower performance of STI is because stocks with higher market capitalisation are given higher weights in the index. The 10 stocks with the highest weights are namely DBS, OCBC, Singtel, UOB, Keppel Corp, Jardine Matheson, Jardine Strategic, Genting Singapore, Global Logistics Properties and Hong Kong Land in that order. These stocks have a total weight of 65.7% in the index. If we consider only the first 4 heavyweight stocks (namely, DBS, OCBC, Singtel and UOB), which have a combined weight of 40.9%, their return since 2000 is only 2.9%. This pulled down the performance of the rest of the STI component stocks.

If we include dividends, the results would be similar. The average return of STI is 6.7%, which is slightly less than that of the 4 heavyweight stocks of 7.5%. In contrast, the average return for a portfolio of equal-weighted STI component stocks is 14.1%.

Returns of STI Component Stocks (With Dividends)

Thus, a simple way of beating the index is to construct a portfolio of equal-weighted STI component stocks. Considering dividends, this will return 14.1%, more than double the returns of STI of 6.7%. Alternatively, by omitting the 4 heavyweights of DBS, OCBC, Singtel and UOB, the portfolio will return an even higher 15.4%. On a dollar basis, over the 13 years since 2000, a $10,000 investment in STI with dividends reinvested would become $23,235. An equal-weighted portfolio of all STI component stocks would become $55,554. An equal-weighted portfolio of STI minus the heavyweights would become $64,369. So, who says you cannot beat the index?

Wish everybody a Happy, Prosperous and Healthy 2014!


See related blog posts:

14 comments:

  1. Interesting analysis, but I wonder if it would be better to include dividend returns for a proper comparison? Similar to what has been done on this post:
    http://www.fool.sg/2013/11/05/impact-of-dividends-on-long-term-sti-gains/

    Also, how practical is it for an ordinary investor with only $10,000 to invest in an "equal-weighted portfolio of STI minus the heavyweights"? Even with $50k, its a tall order due to the requirement to buy whole lots and the high trading costs in Singapore.

    I do agree with you that the STI is overly weighted on the local banks + Singtel though.

    ReplyDelete
    Replies
    1. Thanks for your comments and for sharing the Motley Fool post. The analysis was carried out without considering dividend yield.

      Agree with you that it is difficult to invest in an equal-weighted portfolio of STI component stocks minus the heavyweights. I've added the average returns of the individual component stocks for comparison. There are some stocks that have consistently outperformed the STI, such as SembCorp Industries. Nevertheless, I think the best course of action for investors is not to invest solely in one or a few stocks, but to invest in other indices, such as a global or US stock index, which returns more than the STI have done in the past 26 years.

      Delete
    2. I have further updated the post with dividends considered. Thanks for your suggestion.

      Delete
  2. Amended the figures above as the earlier post inadvertently excluded the returns in 2001. The conclusion that it is easy to beat the index remains unchanged. Apologies for the error.

    Rgds,
    (The) Boring Investor

    ReplyDelete
  3. That analysis is flawed. You are looking at past returns and constructing a portfolio with perfect hindsight. If the large caps did well during the period, your portfolio would have done worse than the market as you would not have sufficient exposure. Following your train of thought, you might as well pick the individual stock which has the highest return during the period and say you beat the market.

    http://earlyretirementsg.weebly.com/

    ReplyDelete
    Replies
    1. No doubt that the analysis was carried out based on past returns. But if the portfolio can't even stand the test of time in the past, how could we expect it to stand the test of time in the future? The best gauge of whether this portfolio can beat the market is really its performance in the future, which we can't tell for now.

      Delete
  4. I think the greatest adv of investing in index over investing in components of index is that while individual companies can go bust and fade away, an index with constantly updated components will never go bust. In other words, an index has survivorship bias that is working in your advantage as an investor and hence is more robust than a stock portfolio made up of the components of the index. Well, unless you also rebalance the portfolio according to the latest components of the index, but that will add up to the transaction costs.

    ReplyDelete
    Replies
    1. Yes, the idea is to rebalance the portfolio according to changes to the index component stocks. As such changes are quite infrequent, the trading costs should not eat too much into the profits.

      Delete
  5. Updated the figures to include the returns for 2013.

    Rgds,
    (The) Boring Investor

    ReplyDelete
  6. Hi Chin Wai,

    When we rebalance a 60/40 equity-bond portfolio, we believe in some form of "mean-reversion".

    So could a mean reversion happen to the Big 4 (Banks and Singtel)?

    I was also thinking the survivorship bias of STI could help to explain your findings. Stocks like NOL (with poor results) have been replaced by better and stronger companies and your findings have excluded these poor results?

    I was thinking maybe the outperformance of the equally weighted STI might not be so big if we include the "drop-outs".

    ReplyDelete
    Replies
    1. Hi 15HWW,

      I can't be sure if mean reversion will happen to the Big 4.

      Yes, it's possible that survivorship bias could be one of the reasons why the STI w/o heavyweights could outperform the STI.

      Delete
    2. Hi Chin Wai,

      Firstly, I appreciate your sharing of these technical work on your blog. =)

      To be honest, I am quite skeptical of "mechanical" strategies like Dogs of STI and Puppies of STI.

      I received a comment on my blog today that mentioned that it's easy to beat the market. According to him, selecting the lowest 30 P/E stocks every year can get one a return of 17%.

      And out of curiosity, I checked out the results of the Puppies of STI. Think as of today, there is an underperformance of 7% compared to the STI. Understandably, it's only a year and such strategies need to be tested over a much longer time horizon.

      Granted, I am actually interested in executing an equal-weighted STI since I do believe that it's easier for smaller companies to grow.

      Anyway, I was wondering if you have executed any of these strategies which you write about in your own portfolio?

      Delete
    3. Hi 15HWW,

      Yes, the posts on some of the strategies were written based on my experience. Examples would be dollar cost averaging, portfolio rebalancing and market timing.

      There are also posts written on some stragies that I investigated, but chose not to invest. Dogs of STI and Bounceback Portfolio belong to this category. It is actually worth investigating some of these strategies. As we discussed in another post, we only have 1 life and 1 holding period. There is no turning back if we were to find that our investment strategies are wrong after 30 years. So, I try to perform some back-testing before embarking on a new investment strategy.

      Lastly, there are other stuffs that I investigated for fun. These include the calendar effects and World Cup effects.

      Delete
    4. Read John Bogle's book on Mean Reversion in the long run, it applies to small vs mid vs large cap as well as different sectors in US stock market. Big4 are simply financials and largest caps in STI.

      Delete