Factor Investing in Practice: Building Robust Portfolios

Adding factor tilts to your portfolio is harder than it sounds: you have to decide which factors to invest in, in what proportion, how to combine them, and whether to vary the weights over time.

Problems with Single-Factor Bets

Most people get attracted to factor investing because of the strong historical backtests.
But even though factors reward in the long-run, they suffer from a few problems that make them challenging to hold.

Factors are prone to severe crashes. For instance, the momentum factor fell more than 60% during 2008-09. These drawdowns also often coincide with economic downturns—precisely when capital preservation matters most.

Also, all factors are cyclical: they perform well during specific phases of the market/economic cycle; as economies go through cycles of ups and downs, so do factors. For instance, momentum performs best in a rising/bull market, while value performs best during the market’s recovery phase. This inherent cyclicality means factors have underperformed the overall market for long periods. These periods, which can last several years, can make even the most iron-stomached investors question their investment philosophy.

Moving money across factors depending on the market phase is tempting but not a viable mitigation strategy: factor timing, similar to market timing, is notoriously difficult to get right ¹.

In short, betting all your money on a single factor is not prudent. Since all factors don’t react to the same underlying drivers, they tend to have low correlation to each other. Diversifying across multiple factors can therefore help reduce the drawdowns and shorten the periods of underperformance. But that raises a new question: how exactly should you combine them?

Combining Factors: Top-Down vs Bottom-Up Approach

Top-Down and Bottom-Up are two different schools of thought on how to diversify across factors.

In the Top-Down approach, you construct independent “sleeves” for each factor and then aggregate them at the portfolio level—essentially creating a ‘portfolio of factor portfolios’. For example, if you want to invest in momentum, value, and quality ², you would:

1. Calculate the momentum scores of all the stocks you want to select from, sort, and select the top X%.
2. Independently repeat the ranking and selection process for value and quality.
3. Build three distinct sub-portfolios (aka sleeves) from these selections.
4. Combine these independent sleeves at a predetermined ratio—say, 33% each—to form the final diversified portfolio.

This structure ensures that each sleeve represents a “pure” expression of its specific factor.

In the Bottom-Up or “integrated approach” you build the portfolio at an individual stock-level:

1. For each stock, calculate the scores for each factor (e.g., momentum, value, and quality).
2. Normalize the absolute factor scores to z-scores. Without normalization, a “Value” score (like P/E ratio) and a “Momentum” score (percentage price change) cannot be added or compared.
3. Calculate a weighted average of the factor z-scores to arrive at a composite score, for each stock. The weights reflect your diversification preferences across the factors.
4. Sort the stocks by their composite score and select the top X%.

There is significant flexibility for customisation in both the approaches. For example, after selection, you can weight the stocks in the portfolio by their composite score, company size, a combination, or equally (less common). You can also keep the factor weights fixed, or vary them over time depending on market conditions. The flexibility makes both approaches well-suited to active as well as passive investing styles.

Which One Works Better?

Both approaches have been found to outperform traditional market-cap weight based indices. The Top-Down approach is based on the academic definition of factors, and has been the traditional form of factor investing. However, multiple studies (State Street³, S&P DJI⁴, AQR⁵), have found that the integrated approach delivers higher absolute and risk-adjusted returns—especially for long-only ⁶ investors, and as the number of stocks decreases. This is because top-down portfolios tend to have stocks that score highly on one factor but average or even negative on others. An integrated approach ensures that every selected stock scores reasonably high across all factors. For a concentrated portfolio (e.g., 50 stocks selected from 500), the effect is quite pronounced as each selected stock is guaranteed to score well on multiple factors.

On the flip-side, the bottom-up approach often excludes—or underweights—the highest scorers of any factor, limiting the ability to capture that factor’s full upside. It also risks feeling like a “black box”: with the Top-Down approach, it is easy to attribute the portfolio performance to performance of individual factors but for Bottom-Up, it can be hard to tell why a portfolio is beating or underperforming the market from the holdings alone⁷. This difficulty in interpreting and attributing performance can make it hard to stick to the plan during tough times. Integrated portfolios also tend to have lower volatility, resulting in a “boring ride”. At times, you can end up feeling left out compared to others who went all in on a single factor like momentum, and consider switching investment style.

Factor Funds in the Indian Market

In principle, one can implement a factor investing strategy from the ground up themselves, by computing factor scores and directly investing in stocks. But that requires significant time and effort, and is also hard to get right.

Luckily, factor-focused mutual funds help avoid this complexity. The available options in India have expanded significantly over the last few years, and there are now several single and multi-factor funds available.

Single-Factor Funds

• Passive single-factor funds, like UTI Nifty 200 Momentum 30 Index Fund and Motilal Oswal BSE Enhanced Value Index Fund, are based on factor indices published by index providers, like NSE or BSE. Since NSE and BSE publish their methodologies publicly, anyone can analyse the scoring formulas.
• Active single factor funds, like Motilal Oswal Active Momentum Fund and Quant Momentum Fund, on the other hand, use proprietary factor scoring and weights. They also incorporate additional screening rules beyond simple scoring. These rules aim to improve risk-adjusted returns by excluding stocks that have poor fundamentals or low scores on other factors. Trusting such funds, however, requires faith in the fund manager’s process. Also, active funds have higher expense ratios than their passive counterparts.

Single-factor funds can serve as building blocks for a DIY Top-Down multi-factor allocation. The benefit is complete control over factor exposure: which factors and in what ratio. But there are important caveats to be aware of:

• Most people frequently re-evaluate individual investments (or funds) in their portfolio. This makes it likely that you will abandon a factor after a severe crash or prolonged underperformance, instead of treating your portfolio of multiple factors as a single entity. This risks wiping out years of gains, and potentially leaving you worse off than if you had avoided factors in the first place.
• Rebalancing amongst factors is tax-inefficient.

Multi-Factor Funds

Multi-factor funds are pre-packaged implementations of a top-down or bottom-up approach. You trade off control in favour of tax efficiency, and built in behavioural restraints. Similar to single-factor funds, these are also available in both active and passive flavours.

Top-Down

• Passive top-down funds are implemented as Fund of Funds (FoF), using multiple passive single-factor funds as constituents, and rebalanced frequently to maintain constant weights. For instance, Axis Multi-Factor Passive FOF aims to maintain a 25% exposure across momentum, quality, value, and low volatility.
• Their active counterparts, like Mirae Asset Multi-Factor FOF, use a similar structure but vary the factor weights based on market conditions.

The internal rebalancing in the top-down funds has no tax implications for investors.

Bottom-Up

• Passive bottom-up funds, like Edelweiss Nifty500 Multicap Momentum Quality 50 Index Fund and UTI Nifty Midsmallcap 400 Momentum Quality 100 Index Fund, mirror indices published by NSE or BSE. These indices are based on fixed weights of factors.
• Active multi-factor funds, like the Bandhan Multi-Factor Fund and Sundaram Multi-Factor Fund, use proprietary factor scores and vary the factor weights based on market conditions.

In the end, both the multi-factor allocation styles have their merits and trade-offs. I prefer the passive bottom-up approach, but the existing fund options are based on only two factors (e.g., Momentum + Quality). I’m waiting for a fund that replicates something like Nifty 500 Multifactor MQVLv 50, which is an index comprising of 50 stocks (selected from 500) based on their combined score across four well-researched factors: momentum, quality, value and low volatility.

Whichever strategy you choose, make sure you understand its practical implications well enough to stick with it for decades.

Disclaimer: This is for educational purposes only, not investment advice.