Sunday, February 24, 2019

MacBooks and Starbucks

I had mentioned about MacBooks and Starbucks in a one-liner here.  How this came about was my tendency to count the number of Apple MacBook users whenever I visited a Starbucks store. Out of 10 visits, on seven occasions I would see that MacBook users form the majority (>50%) of all laptops.

I posed this phenomenon to a few of my friends and colleagues, and one of them gave me a very interesting answer; MacBook users tend to be young adults, and Starbucks is considered a hipster joint, so these two things compliment each other. Short of carrying out a nationwide (or global wide) survey and research, from a presumptive point of view, it kind of makes sense.

So is there a positive correlation between MacBooks and Starbucks?

Good thing that the two companies, Apple and Starbucks, are publicly listed, so I can find out their market correlation using the Portfolio Visualizer (www.portfoliovisualizer.com). I will take the period of January 1993 to December 2018 for calculating the correlation, since Starbucks was listed somewhere in 1992, and the first MacBook (called the PowerBook) was released in 1991, to give a fair comparison.

Based on annual returns, the correlation is 0.161. While this number is positive, it is positive on the low end, and usually we see this correlation score between distinct asset classes.

So, does this mean my observations are wrong?

Short answer is, it depends.

It Depends #1

Both companies are in different businesses. Apple is in technology while Starbucks is in food and beverage. Also, in 2018 Mac computers (comprising of Mac desktops and MacBooks) formed only 8.9% of Apple’s total revenue2, as compared to Starbucks whose primary product is coffee. These reasons could explain why the correlation is low between them.

It Depends #2

On another note, my observation settings may be skewed. It is not wrong to look purely at MacBooks in Starbucks, since I am comparing these two variables. To make it more complete, we could see if MacBook users form the majority in other cafes or outlets that serve local coffee, or we could find out if MacBook users are at other non-café places such as fast food restaurants or the library.

It Depends #3

Back to the numbers game, we go by the law of averages and assumptions, and find out some interesting statistics. Assuming:

  • Average Lifespan of An Apple Device: 4 years3
  • Number of Macs sold between 2015 and 2018 (to fulfill the above 4 years): 76.53 million4
  • Assumed number of Macs sold that are MacBooks: 76.53 x 49.11%= 37.58 million
  • Total number of Starbucks outlets: 29,3246


If all MacBook users gather at Starbucks at the same time, there would be 37.58 million / 29,324 = approximately 1,281 per outlet!

Thus to form a majority, you would need 1,280 people or less using non-Mac laptops within each Starbucks. This is just “wow”. Of course, the last answer was given in a hypothetical situation which, in all probability, near to impossible to realize.  

What I had gone through above is a very simple exercise in fact finding based on figures and observations, which I could classify them as quantitative and qualitative data, respectively. 

So why am I mentioning this?

Simple. It is a way of doing fundamental analysis (FA).

You will encounter these two types of data when you conduct your FA; facts and opinions in the news, financial statements and management messages in annual reports, etc. Quantitative data is difficult to refute (aside from “figures fudging”) as they are past or “as at” data. Qualitative data is a bit on blurred grounds because it is obtained by non-quantified methods, such as observations, trend spotting and opinions. Biases from different individuals’ personalities and experiences could also play a part in qualitative analysis.

Contradicting as it may sound, we need both quantitative and qualitative data and information in order to give a more rounded picture in FA. I have always said that guesstimates (or calculated guesses) are better than pure wild-stab-in-the-dark guesswork, and these can be used for further analysis.

Sounds like hard work? Maybe for the first few times, but once you get the hang of it, it is easier. While you are at it, go grab a cup of coffee and enjoy it at the same time.



2 – Statista. The Mac’s Waning Relevance to Apple. https://www.statista.com/chart/8817/mac-sales-as-a-percentage-of-apples-revenue/ (accessed 23 Feb 2019)

3 – MacRumors. Average Apple Device Lifespan Estimated at Just Over Four Years by Analyst. https://www.macrumors.com/2018/03/02/average-apple-device-lifespan/ (accessed 24 Feb 2019)

4 – Statista. Global Apple Mac sales in the fiscal years from 2002 to 2018 (in million units). https://www.statista.com/statistics/276308/global-apple-mac-sales-since-fiscal-year-2002/(accessed 23 Feb 2019)

5 – GlobalStat StatCounter. Desktop vs Mobile vs Tablet Market Share Worldwide. Jan 2018 – Jan 2019. http://gs.statcounter.com/platform-market-share/desktop-mobile-tablet(accessed 24 Feb 2019)

6 – Statista. Number of Starbucks locations worldwide 2003-2018. https://www.statista.com/statistics/266465/number-of-starbucks-stores-worldwide/ (accessed 23 Feb 2019)

Saturday, February 9, 2019

Revisiting The Three REIT ETFs

A while ago, I had written about the three locally listed REIT ETFs here, so this post is some sort of a follow-up. To recap, the three REIT ETFs are the (my short form in brackets) Philip SGX APAC Dividend Leaders REIT ETF (Philip APAC), the NikkoAM-Straits Trading Asia ex Japan REIT ETF (Nikko-Straits Trading) and the Lion-Philip S-REIT ETF (Lion-Philip).

For this blog post, we shall see what would be the weightage of individual REITs and the sectors if we decide to buy all of the REIT ETFs, in equal share numbers, for our investment portfolio, based on the latest fund information.1,2,3

REITPhilip APACNikko-Straits TradingLion-PhilipOverall Weightage
CAPITALAND MALL TRUST3.6810.2010.007.96
ASCENDAS REIT4.6510.008.207.62
CAPITALAND COMMERCIAL TRUST 8.709.906.20
MAPLETREE COMMERCIAL TRUST 6.409.805.40
MAPLETREE LOGISTICS TRUST 6.205.904.03
MAPLETREE INDUSTRIAL TRUST 5.106.904.00
MAPLETREE NORTH ASIA COMMERCIAL TRUST  4.904.803.23

Fig.1 – Weightage of individual REITs (in %). REITs that are common in at least two REIT ETFs and among the top 10 holdings are selected.

SectorPhilip APACNikko-Straits TradingLion-PhilipOverall Weightage
RETAIL37.7838.0029.2035.08
INDUSTRIAL17.3721.3023.9020.82
DIVERSIFIED27.6417.008.4017.67
OFFICE13.5317.9019.3016.89
HOTEL & RESORT0.962.308.103.78

Fig. 2 – Weightage of REITs by sector (in %). Only top 5 sectors are shown.

Based on the above two tables, we can see that, if the three REIT ETFs are combined into one, Capitaland Mall Trust is the highest holding at 7.96%, with Ascendas coming in second at 7.62%. The biggest component, which is the retail sector, is slightly above 35%, with industrial coming in second at around 21%. However, this status does not remain stagnant as the weightages do change from time to time, though mostly not by that much. If you know some “Google-Fu”, you can check out the weightages from past fund information documents.

A lot of discussion points can be generated based on the above numbers, ranging from the superficial (e.g. which ETF to go to for an all-rounder sector coverage) to the complex that involves looking deeper at various degrees and from different dimensions (e.g. the high concentration of retail REITs in the ETFs, and how shopping habits and the impact of electronic commerce could erode their returns). I would stop at here for now, and I hope at least the above numbers can spark some analysis of your own.


1 – Philip SGX APAC Dividend Leaders REIT ETF. Product Info Sheet. December 2018. https://www.phillipfunds.com/uploads/funds_file/201812_Phillip_SGX_APAC_Dividend_Leaders_REIT_ETF_Product_Sheet.pdf (accessed 8 Feb 2019)

2 – NikkoAM-Straits Trading Asia ex Japan REIT ETF. Factsheet. 31 December 2018. https://www.nikkoam.com.sg/files/documents/funds/fact_sheet/axj_reit_etf_fs.pdf?v20170630 (accessed 8 Feb 2019)

3 – Lion-Philip S-REIT ETF. Fund Information. December 2018. https://lgi.nextview.com/doc/uploads/documents/index.php?type=FS&fid=LEPF&lang=EN (accessed 8 Feb 2019)

Saturday, February 2, 2019

SPDR GLD ETF Lot Size Change

OK, I may be “WOLS” (internet-speak for being slow on the news), but I would like to highlight that from 14 Jan 2019 onwards, you are able to purchase the State Street Global Advisors SPDR Gold Trust (or better known as the SPDR GLD ETF) in lot size of 5 shares, down from the previous lot size of 10.

According to an announcement by State Street Global Advisors on 10 Dec 20181, this reduction allows “…more Singaporeans, including Central Provident Fund (CPF) Investment Scheme members, to efficiently invest in gold…”. It also emphasized that gold “…historically acted as a portfolio diversifier, particularly during times of market volatility”, which I agree as I am a believer of diversification.

So what does this mean for The Bedokian Portfolio investors? There are two issues which I will mention here.

Issue #1: Bedokian Portfolio Starting Amount

In my eBook, I used the SPDR GLD ETF as the basis to determine the starting amount for The Bedokian Portfolio2. Taking the ETF closing price of United States Dollars (USD) 124.693, and the exchange rate between USD and Singapore dollars (SGD) of 1.30854(both information as at 1 Feb 2019), one SPDR GLD ETF share is about SGD 163.16. If you have decided to allocate 5% of your Bedokian Portfolio to commodities, under the old 10-share lot, you would need to fork out SGD 32,632 (SGD 163.16 x 10 x 20) to start the portfolio (before any transaction costs and commission fees).

However, with the new 5-share lot size, your starting portfolio can be halved to SGD 16,316 (SGD 163.16 x 5 x 20). 

Issue #2: “Overshooting/Undershooting” Risk

My eBook also mentioned about the deviation tolerance of the asset class allocation5. For asset classes with 5% to 10% allocation, a deviation of 2.5% is acceptable. So in this case, if you allocate 5% to commodities, rebalancing is not necessary unless it goes below 2.5% or above 7.5%.

Depending on the size of your portfolio, when your commodities allocation goes below (or above) the thresholds stated in the previous paragraph, rebalancing with a larger lot size may mathematically cause the portion to go beyond the other limit. For example, if your commodities portion is at, say 2% of your portfolio, and by purchasing 10 SPDR GLD ETF shares, the allocation may go up to above 7.5%. A smaller lot size would reduce this “overshooting/undershooting” risk.

One More Thing…

The above two issues are based on the investor’s preference to start off and rebalance with the SPDR GLD ETF listed in the Singapore Exchange. However, these issues will be moot if you decide to get the SPDR GLD ETF from other markets. Yes, it is the same ETF, and it is listed not just in Singapore, but also in the United States, Mexico, Japan and Hong Kong6. Again as stated in my eBook7, you can buy one SPDR GLD ETF share in the United States market, thus bringing down your portfolio starting amount to a very affordable level. In addition, the overshoot/undershoot problems during rebalancing will be greatly reduced.

A caveat though, if you wish to invest your CPF savings, you could only get the SPDR GLD ETF from the SGX, not from overseas markets, as per CPF regulations.

With all the talk on gold, and since it is a good omen topic in the coming festivities, The Bedokian would like to wish you a Happy and Prosperous Lunar New Year!


1 – State Street Newsroom. SPDR Gold Shares Halves Board Lot Size on Singapore Exchange. 10 Dec 2018. https://newsroom.statestreet.com/press-release/spdr-gold-shares-halves-board-lot-size-singapore-exchange (accessed 2 Feb 2019)

2 – The Bedokian Portfolio, p72-73

3 – Singapore Exchange. https://www2.sgx.com (accessed 2 Feb 2019)

4 – Exchange-Rates.org. US Dollars (USD) to Singapore Dollars (SGD) exchange rate for February 1, 2018. https://www.exchange-rates.org/Rate/USD/SGD/2-1-2018 (accessed 2 Feb 2019)

5 – The Bedokian Portfolio, p82

6 – GLD SPDR Gold Shares. https://www.spdrgoldshares.com (accessed 2 Feb 2019)

7 – The Bedokian Portfolio, p73