A common sentiment among retail investors is that investing in Singapore is all about buying Banks and REITs. There is some folk wisdom to having this sentiment as both investment categories have been outperforming the rest of the STI index during the COVID crisis. Another reason is this general idea that rising interest rates are good for banks but bad for REITs with falling interest rates having an opposite effect.
For the 5 years from 1 January 2015 to 22 May 2020, banks have a positive annualised return of 0.85% against the -2.1% of losses inflicted on the STI ETF. Another reason why banks are attractive is that banks are now more willing to see themselves as yield instruments with DBS Group holdings now having a current yield of over 6% at a priced of $20.82. Finally, banks seemed to have suffered more during the market crash in March 2020 but are now experiencing a faster rebound.
Over a similar timeframe, we saw more pronounced outperformance of REITs found in the STI with the REITs component returning 8.99% compared to 2.1% of the STI. Investing in REITs in the STI have been so successful of late that their dividends are now less than or comparable to banking stocks.
But beyond REIT and Banks, other categories of blue-chips exist. There is the triad of real estate developers Capitaland, UOL and City Developments; there is also Jardine related counters consisting of Diary Farm, Hong Kong Land, Jardine C&C, Jardine Matheson Holdings and Jardine Strategic holdings. We found that both of these categories have lacklustre performance compared to the STI in recent days.
If you limit consideration to just the four categories of REITs, Banks, developer, and Jardine-related counters, you would still be accounting only for half of the STI counters. The question is whether hidden clusters of blue-chips exist that the remaining half?
Data science may be able to more answers to such questions.
Agglomerative clustering is a computer algorithm that can assist us in grouping and categorising objects together based on how similar their financial characteristics are. After getting this algorithm to work on some financial data of local stocks, I was able to create a computer programme to automatically pair similar companies up and visualise them, thereby exposing close relationships between blue-chip counters that even the most experienced investors may not know after a lifetime of investing in the financial markets.
The above diagram shows major blue-chips grouped with a “buddy that has the most similar characteristics. We have some confidence that the program works because it was able to automatically group Banks, REITs and three Jardine companies together without human supervision.
Above and beyond detecting the correct stock categories, we found a possible new cluster containing the following stocks:
- Venture Manufacturing (V03) – A contract manufacturing company
- SATS (S58) – A provider of food and gateway services
- Comfort Delgro (C52) – A transport company
- Visibly absent is Singapore Technologies Engineering from this list of stocks.
We could imagine these blue-chips as a cluster of companies doing some engineering work.
The next step after discovering a new cluster would be to observe the correlation between each counter. Correlation is a mathematical property between two stocks represented by a number between -1 and 1. A correlation close to one implies that the stocks move in tandem to each other. In such a case, owning one stock may be almost similar to owning the other. Local banks typically have correlations around 0.7 – 0.8 against each other, which is why I tell my students that, if they have insufficient capital to own all three banks in their portfolio, owning one of them is just fine. The correlation between these three counters in our new cluster is as follows:
The correlation between the three counters is observed to be quite low compared to banks. This relationship implies that these three stocks have a mind of their own and do not blindly follow each other in the markets. Thus, owning three counters at the same time would be significantly less risky than owning just one.
A look at their historical performance yields more surprising results.
A balanced portfolio of Venture, Comfort Delgro and SATS would have significantly outperformed the STI over the past five years, albeit with higher volatility. During the COVID-19 crisis, however, the three stocks suffered much more than the rest of the market. There also seems to be an indication that the stocks also rebounded faster than the STI index.
In summary, retail investors should keep an open mind when investing in the stock markets. When you encounter folk wisdom on the stock-markets, it is useful to find out empirically to what extent folk wisdom holds. The methods, used by data scientists on local financial data, can be repurposed to lead to novel insights often alien even to an experienced investor. In this case, we examine a new trio of stocks with similar fundamental characteristics to each other but are well diversified from each other and, when held in tandem, would have outperformed the rest of the markets.