Efficiently Inefficient #11 : Quantitative Equity Investing.

This is the chapter that I have to read really carefully because I am in the process of semi-professionalizing my own portfolio.

As it stands, everything is going fine with my 200% leveraged portfolio, I basically collected every single REIT that yields between 6%-8% that is not in my main portfolio and built a leveraged portfolio out of it. I have about 13 counter as of now (it would have been 14 has the Cromwell IPO not been cancelled). The portfolio size is expected to reach over $200k next week and it yields around $11,000 a year. That's not bad considering that I only pumped about $110k into this portfolio so far.  The portfolio was also remarkably stable during the 7th month when all the other portfolios I have took a small dive.

The book confirmed that this is, in a limited extent, what some professional hedge fund managers do in practice - they apply leverage on lower beta stocks to magnify returns and, if they wish to stay market neutral, employ a much lower gearing on their riskier short positions.

Anyway, let's just review the 3 types of Quantitative Equity investing :

a) Fundamental Quantitative Investing

This is what I am trying to do with the exception of shorting overpriced stocks.

I backtest a series of different strategies that result in diversified portfolios which tend to outperform markets over time. Two strategies that tend to work over time are mentioned in the textbook is (a) Buying stocks with a high book value relative to price and shorting stocks with a low book value relative to price. (b) Buying the stocks which performed well over the last 12 months and holding them for a month. Selling the worst performing stocks over the last 12 months and holding those positions for a month.

I am doing just the long-only variant of (a) but (a) and (b) tend to result in different portfolios which have a low correlation with each other.

My next project is to see if something can be said about (b) for SGX stocks.

b) Statistical Arbitrage

This sound really fun but I am limited by my inability to short stocks on SGX. In statistical arbitrage, you hunt for Siamese twin counters and bet on them converging in price. An example of a  Siamese twin might be Unilever NV which is traded in the Netherlands and Unilever PLC that is traded in the UK. In this form of investing, you monitor a set of Siamese twin stocks and you bet on them converging in price when they start to deviate in value. You execute your strategy by buying the lower priced twin and selling the higher priced twin.

Just thinking loudly to myself, one local Siamese Twin counter might be the two HPH counters which are denominated in USD and SGD.

I'm going to sit this one because this strategy might generate a lot of trading costs.

c) High Frequency Trading

Sometimes I wish that I could get into this industry.

Imagine you have a big order of 1,000,000 shares. You split the order into 1,000 orders of 1,000 shares and use a computer algorithm to perform these trades for you. You earn a profit by being some kind of market maker, providing liquidity to buyers and sellers who arrive at the market at different times.

The theory is that if your orders arrive fast enough ( and this is done by co-locating your servers nearer to that of the exchange at a fee ), you can profit from the small mispricings that occur in every given day.