When thinking about a client’s stock portfolio, we believe that the Future is Unknowable and that Evidence, not Hope, should drive decision making. As such, we do not believe in timing the market or building concentrated portfolios that take unnecessary risks.
Our stock portfolios reflect this philosophy. If markets are efficient, then reading analyst reports and listening to earnings calls to pick individual stocks is a fool’s errand. That’s why we don’t do it. We believe that there are too many investors chasing returns and a theoretical edge, making it all but impossible to pick the right individual stock at the right time.
However, that doesn’t mean that there are zero inefficiencies in markets. Decades of academic research have proven that inefficiencies do exist but taking advantage of them requires a systemic and bias-free investment approach. That is what we practice at Leo Wealth and that is how we construct our Direct Equity Portfolios.
Our 3-step process to building a stock portfolio:
define the objective of the portfolio and identify all stocks that fit into the relevant universe. Universe can be broad or narrow. Examples: entire global equity universe, global consumer stocks, Asian manufacturing stocks, clean energy stocks, cyber security stocks.
Do not ignore investment principles to capture a popular stock. Score stocks via multiple ESG methods to identify outliers. Utilize a 30-factor valuation, quality, technical, sentiment screen to narrow universe.
Select highest scoring stocks from universe while paying attention to regional/sub-sector effects. Adjust as needed for areas that cannot be captured by systematic screens. Review universe, scoring framework and individual scores regularly and rebalance as needed.
We utilize a factor model that relies on decades of academic research into market drivers. If multiple papers have not been published and peer reviewed on a particular factor, it is not included. If only one metric works for a factor, it is not included. If it works only temporarily, or in a specific region/sector, it is not included.
In total, our process scores ~14,000 stocks on a daily basis across 7 categories:
Value – is a stock cheap relative to peers on a variety of measures?
Safety – how does it compare on debt, distress, volatility, liquidity and other defensive metrics?
Payout – what is the income/yield profile?
Quality – is it a consistently profitable, balance-sheet efficient company?
Technicals – does it show signs of price and earnings momentum, reversal or cycle skew?
Sentiment – are insiders buying or selling, what are analysts and institutions doing?
Macro – how has it historically fared in macro environment’s like today?
No systematic approach is perfect and therefore we constantly look for ways to improve. In addition, no model or process can quantify all information (such as fraud by a CEO), so we add ESG scores from multiple agencies and a qualitative review by our investment team.
By quantifying and combining these various “return factors”, we aim to deliver a consistent approach to finding stocks, sectors and markets with winning characteristics. Based on evidence, not hope. While past performance doesn’t guarantee future returns, we are confident that this approach can work long-term, as it has over the previous few decades.
Looking back to 1996, constantly prioritizing the best scoring stocks and avoiding the worst would have created a material improvement in returns. Below is a chart that shows 25 years of data, with the green lines showing performance of the top deciles of the market by factor score and the red lines showing performance of the bottom decile.
Learn more about our available Direct Equity Portfolios.
Chart shows the average return per score decile. Q1 contains a basket of stocks with the lowest scores (stocks that have a score between 0%-10%), and Q10 contains stocks with the highest scores (stocks that have a score between 90%-100%). The simulated portfolio for each decile is based on an equal-weighted basket of stocks rebalanced monthly. The key takeaway is that on average, the model is able to successfully separate baskets of winners from baskets of losers. Source: BCA Research, July 2021