- Wealth PMS (50L+)
In a nutshell:
Presenting the Capitalmind Low Volatility Portfolio
Capitalmind Low Vol Portfolio Introduction by Capitalmind
“Books on stock-picking are easier to write because there are always great stories when it comes to individual companies: fascinating tales of greatness and woe that end wonderfully for the sage stock picker who is the hero of this tale. Factors don’t have made-for-TV endings. Success is measured by less thrilling statements like, “and then the factor had a +1 standard deviation decade,” versus discretionary stock-picking stories that end with “and then the company I invested in invented the iPod!” – Foreword: Your Complete Guide to Factor-Based Investing
The no-jargon version:
Let’s say you were tasked with constructing a high-quality basketball team.
One way you might go about it is to ask around for the best basketball players, the folks with pedigree colleges with track records of producing great players, those who have scored the most points or been in the most winning games. Then look at recorded footage, observe them in live practice. You might even meet them in person, talk to their trainers. And only then decide whether to have them on your team.
You might be thinking “Wait, I don’t know anything about basketball.” The approach will be similar for any other team sport. In fact, it holds for other fields like how big companies recruit, how VCs fund start-ups
An alternate approach in our basketball team hypothetical would be to come up with a set of characteristics that might represent good basketball players. Thinking intuitively, you start with height. You’d go back into the past, and create teams of the tallest available players. Then you’d see how many points those teams scored and whether they outperformed the average team. If the data says a team of the tallest available players consistently scored 3 extra points per game compared to the average team, you could conclude that it might continue to work in the future.
Note how both approaches are rooted in pattern recognition. Relying on the past to tell you what might work in the future. The first is largely subjective but considers a broad set of hard and soft aspects. The second is objective and only considers what can be quantified and more importantly, can be tested with historical data.
It’s probably apparent, the first approach is the investing equivalent of fundamental stock selection, and the second approach is the analogy to factor investing.
Factor Investing is about defining a characteristic (also called factor), and consistently applying a set of criteria to buy a diversified portfolio of stocks that share that characteristic.
What factors offer the potential to beat the market?
We picked height as the “factor” for exploration in our illustration above. But what if there were other factors more effective at predicting outperformance?
Speed for instance. Or Age. How about Place of birth? Relationship status? You quickly realize the possibilities are endless. After all, all kinds of things could impact the performance of professional athletes.
For example, if you were considering NBA data from the 1990s, picking players from teams with domesticated agricultural animals in their names would deliver many points in excess of the average team. Wait, what?
Fact: The Chicago Bulls won the NBA championship 6 years out of the 10 between 1991 and 2000.
This brings us to the Bangladesh Butter Production problem. In 1995, David Leinweber, now the head of Center for Innovative Financial Technology, Lawrence Berkley National Lab, examined over 150 macro-economic indicators to find statistically significant relationships with the S&P500.
His intent, to identify any indicators that could predict the market. It was butter production in our neighbouring country that showed the tightest relationship with the US Equity Index. Leinweber published this finding as a joke which has since been part of most “correlation is not causation” lessons in courses the world over.
The simple takeaway for factor investors, if a factor does not make intuitive sense, it probably is a chance relationship applicable only to the historical data in question. Therefore unlikely to deliver excess return in the future.
In their book, Your Complete Guide to Factor-Based Investing, the authors lay out five tests to determine whether a factor offers the promise of superior returns. To merit consideration, a factor should be:
The more of the above five conditions a factor checks, the more likely it is to offer potential for outperformance in the future.
At Capitalmind, we started with factor-based Momentum portfolios in 2017.
Long before momentum portfolios became the rage in 2020 and 2021, it had been the subject of scrutiny for financial researchers. Studies spanning two centuries have found the prevalence of ‘the momentum premium”:
“The return premium is evident in 212 years (yes, this is not a typo, two hundred and twelve years of data from 1801 to 2012) of U.S. equity data, dating back to the Victorian age in U.K. equity data, in more than 20 years of out-of-sample evidence from its original discovery, in 40 other countries, and in more than a dozen other asset classes.” – Asness et al in Fact, Fiction and Momentum Investing
Our research, published in our whitepaper Does Momentum Investing in Indian Equities?, confirmed the prevalence of the momentum premium for the Indian market:
The Capitalmind Momentum Portfolio was published on smallcase in early 2019. We are the first (and probably only) PMS in India to run a discretionary Momentum Portfolio. Our momentum smallcase and the portfolio in the PMS are different portfolios built on the same underlying set of principles. Since launch, investor outcomes in both momentum portfolios have been satisfactory.
In the background, we have been working on identifying other factor-based strategies that meet the core criteria of offering meaningful diversification from the market and other factor strategies. This is in line with our emphasis on asset allocation and position sizing as the core of a healthy investment strategy you can stick with for the long term.
In fact, in their paper “The death of diversification has been greatly exaggerated“, authors Antti Ilmanen and Jared Kizer make the case that
“…factor diversification is more effective at reducing portfolio volatility and market directionality than asset class diversification”
We would not go so far as to dismiss the need for diverse assets, but we do see the benefits of exposure to meaningfully diverse factors in an investors’ portfolio.
Not least of those benefits is increased chances of long-term outperformance by staying invested in meaningfully diverse fact0r-based portfolios.
Low Volatility is not a new concept in India. There is a NIFTY strategy index called the NIFTY 100 Low Volatility 30. In March 2021, ICICI Prudential launched an ETF based on this index. We already liked the idea of a passive instrument investing in low-volatility stocks and said as much in our NFO review.
Chart compares how ₹100 invested in the NIFTY100 Low Vol 30 index would do compared to the NIFTY.
The sizable difference in ending values over 16 years suggests there is something there for the investor looking to outperform the NIFTY.
We set out to explore whether we take the core concept of lower volatility stocks and get a better risk-return profile.
Clearly, an attempt at creating our own low volatility portfolio had to not only outperform the NIFTY but also compare favourably with the Low Volatility index.
Our Hypothesis: A diverse portfolio of stocks demonstrating relatively lower volatility than the market, rebalanced periodically, can deliver excess returns with lower downside volatility compared to the market (NIFTY 50) and the NIFTY100 Low Volatility 30 indices.
Sticking to the core principle of building portfolios of stocks showing relative lower volatility and rebalancing quarterly, we examined various rules that would help prove or disprove our hypothesis.
Table below shows summary of performance of the CM Low Vol Portfolio versus the NIFTY and the NIFTY100 Low Volatility 30.
Aggregated performance metrics condense long periods into easy-to-understand metrics. On the other hand, they lose out on the nuance of performance comparison over specific periods.
Chart below compares the 1 and 3 year rolling returns
Another way to visualise the rolling returns is to see which of the three was the best for what percentage of the 16 years from Apr 2005 to July 2021.
This chart makes an important point. Both the Low Volatility index and our version of the strategy underperform the NIFTY for stretches of time. On a 1-year trailing basis, both the Low Volatility index and our version of the Low Vol strategy underperform the NIFTY for 17% of the time. That’s 2.7 years out of 16. On a longer time frame, this shrinks significantly underscoring the importance of sticking with it to reap the eventual outperformance.
The CM Low Vol portfolio outperforms both the NIFTY and the NIFTY100 Low Vol Index 81% of the time over any 3-year basis and 52% of the time on any 1-year basis.
Slicing return performance every which way says buying a portfolio of low-volatility stocks with a few screening conditions helps outperform the index. But why might low volatility stocks offer an edge?
Looking at frequent portfolio holdings in the past offers some insight:
Note these are stocks that were frequently part of the portfolio over the backtest period and not necessarily current constituents.
Interesting how the list of companies looks a lot like lists of Quality stocks. In fact, there is significant overlap between the Low Volatility and Quality indices over time.
Two things seem to be common among the companies repeatedly picked by the low-volatility portfolio:
A possible explanation could be behavioural. The typical active investor builds positions in anticipation of positive earnings surprises. Such an investor would not be interested in holding companies with little such scope. It is logical then that in the first set of companies, some disappoint on the extent of earnings growth and therefore see significant price volatility as active investors exit them. This would automatically make the “no surprises” set of companies to be relatively less volatile.
Most low-volatility stocks tend to be highly liquid stocks with no shortage of supply or demand. Also, stocks like these have longer-term holders who don’t sell the stock based on short-term news flows which potentially acts as a dampener on rapid price moves both up and down.
Holding low volatility stocks seems to work well as a long-term investment strategy.
CM Low Vol can outperform the NIFTY but can it outperform CM Momentum on annualised returns? Over the long term, probably not. Then why bother?
For the same reason that having some debt in your portfolio can help you stay invested through deep drawdowns in equity. Refer back to the quote from the Ilmanen and Kizer paper – “Factor diversification reduces portfolio volatility.”
Take 2018 for instance:
The next two years were different.
Imagine concluding on Dec 31st 2018 that Momentum investing does not work.
We only know that every strategy outperforms and underperforms. We just don’t know when. This means being having the conviction to stay invested becomes a super power.
Sure point to point returns are not the best way to make investment decisions but try convincing yourself of that when in the middle of a stretch of underperformance.
Combining truly diverse factor portfolios has an interesting effect on overall portfolio performance. For instance, table below shows how a simple 50:50 combination of CM Momentum and CM Low Vol would have done from 2017 to 2021:
Notice how the volatility of the combined portfolio is lower than either of the portfolios taken alone. This happens anytime you combine two instruments with less than perfect correlation of 1.
In this case, because the correlation of daily returns between CM Momentum and CM Low Vol is 0.55, the combined volatility is lower than either of them.
A portfolio made up of diverse factors will see lower fluctuations which makes it easier for investors to stay invested when markets surprise.
Note that the benefits of combining portfolios only apply when the underlying investment philosophies are meaningfully different. Investing in four variations of momentum portfolios does not count.
The biggest case for holding the CM Low-Volatility factor portfolio in addition to the CM Momentum-factor portfolio is that they tend to hold stocks showing mutually exclusive characteristics and so will have negligible overlap in holdings at any given time.
Investing in the CM Low Vol Portfolio
Subscribe to Capitalmind Premium for immediate access to premium research, model portfolios, actionable strategies, and prolific member community
Reach out to us on twitter: @capitalmind_in or email us on premium [at] capitalmind [dot] in for any queries.